diff --git a/LICENSE b/LICENSE --- a/LICENSE +++ b/LICENSE @@ -1,6 +1,21 @@ +# CC0 1.0 Universal + +To the extent possible under law, 唐鳳 has waived all copyright +and related or neighboring rights to regex-pcre-builtin + +This work is published from Taiwan. + +http://creativecommons.org/publicdomain/zero/1.0 + +# Contains code from from pcre 8.31, under the BSD license: + +http://www.pcre.org/licence.txt + +# Derived from regex-pcre 0.94.4: + This modile is under this "3 clause" BSD license: -Copyright (c) 2007, Christopher Kuklewicz +Copyright (c) 2007-2012, Christopher Kuklewicz All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: diff --git a/Text/Regex/PCRE.hs b/Text/Regex/PCRE.hs --- a/Text/Regex/PCRE.hs +++ b/Text/Regex/PCRE.hs @@ -57,6 +57,6 @@ getVersion_Text_Regex_PCRE :: Version getVersion_Text_Regex_PCRE = - Version { versionBranch = [0,94,1] -- Keep in sync with regex-pcre.cabal + Version { versionBranch = [0,94,4,0,8,31] -- Keep in sync with regex-pcre.cabal , versionTags = ["unstable"] } diff --git a/pcre/config.h b/pcre/config.h --- a/pcre/config.h +++ b/pcre/config.h @@ -29,7 +29,10 @@ /* If you are compiling for a system that uses EBCDIC instead of ASCII character codes, define this macro as 1. On systems that can use - "configure", this can be done via --enable-ebcdic. */ + "configure", this can be done via --enable-ebcdic. PCRE will then assume + that all input strings are in EBCDIC. If you do not define this macro, PCRE + will assume input strings are ASCII or UTF-8/16 Unicode. It is not possible + to build a version of PCRE that supports both EBCDIC and UTF-8/16. */ /* #undef EBCDIC */ /* Define to 1 if you have the `bcopy' function. */ @@ -55,6 +58,12 @@ #define HAVE_DLFCN_H 1 #endif +/* Define to 1 if you have the header file. */ +/* #undef HAVE_EDITLINE_READLINE_H */ + +/* Define to 1 if you have the header file. */ +/* #undef HAVE_EDIT_READLINE_READLINE_H */ + /* Define to 1 if you have the header file. */ #ifndef HAVE_INTTYPES_H #define HAVE_INTTYPES_H 1 @@ -81,14 +90,10 @@ #endif /* Define to 1 if you have the header file. */ -#ifndef HAVE_READLINE_HISTORY_H -#define HAVE_READLINE_HISTORY_H 1 -#endif +/* #undef HAVE_READLINE_HISTORY_H */ /* Define to 1 if you have the header file. */ -#ifndef HAVE_READLINE_READLINE_H -#define HAVE_READLINE_READLINE_H 1 -#endif +/* #undef HAVE_READLINE_READLINE_H */ /* Define to 1 if you have the header file. */ #ifndef HAVE_STDINT_H @@ -120,10 +125,13 @@ #define HAVE_STRING_H 1 #endif -/* Define to 1 if you have the `strtoll' function. */ +/* Define to 1 if you have `strtoimax'. */ +/* #undef HAVE_STRTOIMAX */ + +/* Define to 1 if you have `strtoll'. */ /* #undef HAVE_STRTOLL */ -/* Define to 1 if you have the `strtoq' function. */ +/* Define to 1 if you have `strtoq'. */ #ifndef HAVE_STRTOQ #define HAVE_STRTOQ 1 #endif @@ -159,7 +167,7 @@ #define HAVE_ZLIB_H 1 #endif -/* Define to 1 if you have the `_strtoi64' function. */ +/* Define to 1 if you have `_strtoi64'. */ /* #undef HAVE__STRTOI64 */ /* The value of LINK_SIZE determines the number of bytes used to store links @@ -172,6 +180,12 @@ #define LINK_SIZE 2 #endif +/* Define to the sub-directory in which libtool stores uninstalled libraries. + */ +#ifndef LT_OBJDIR +#define LT_OBJDIR ".libs/" +#endif + /* The value of MATCH_LIMIT determines the default number of times the internal match() function can be called during a single execution of pcre_exec(). There is a runtime interface for setting a different limit. @@ -238,15 +252,28 @@ #define PACKAGE_NAME "PCRE" /* Define to the full name and version of this package. */ -#define PACKAGE_STRING "PCRE 7.7" +#define PACKAGE_STRING "PCRE 8.31" /* Define to the one symbol short name of this package. */ #define PACKAGE_TARNAME "pcre" +/* Define to the home page for this package. */ +#define PACKAGE_URL "" + /* Define to the version of this package. */ -#define PACKAGE_VERSION "7.7" +#define PACKAGE_VERSION "8.31" +/* The value of PCREGREP_BUFSIZE determines the size of buffer used by + pcregrep to hold parts of the file it is searching. On systems that support + it, "configure" can be used to override the default, which is 8192. This is + also the minimum value. The actual amount of memory used by pcregrep is + three times this number, because it allows for the buffering of "before" + and "after" lines. */ +#ifndef PCREGREP_BUFSIZE +#define PCREGREP_BUFSIZE 20480 +#endif + /* If you are compiling for a system other than a Unix-like system or Win32, and it needs some magic to be inserted before the definition of a function that is exported by the library, define this macro to @@ -278,10 +305,16 @@ #define STDC_HEADERS 1 #endif +/* Define to enable support for Just-In-Time compiling. */ +/* #undef SUPPORT_JIT */ + /* Define to allow pcregrep to be linked with libbz2, so that it is able to handle .bz2 files. */ /* #undef SUPPORT_LIBBZ2 */ +/* Define to allow pcretest to be linked with libedit. */ +/* #undef SUPPORT_LIBEDIT */ + /* Define to allow pcretest to be linked with libreadline. */ /* #undef SUPPORT_LIBREADLINE */ @@ -289,19 +322,37 @@ handle .gz files. */ /* #undef SUPPORT_LIBZ */ -/* Define to enable support for Unicode properties */ -#define SUPPORT_UCP +/* Define to enable the 16 bit PCRE library. */ +/* #undef SUPPORT_PCRE16 */ -/* Define to enable support for the UTF-8 Unicode encoding. */ -#define SUPPORT_UTF8 +/* Define to enable the 8 bit PCRE library. */ +#ifndef SUPPORT_PCRE8 +#define SUPPORT_PCRE8 /**/ +#endif +/* Define to enable JIT support in pcregrep. */ +/* #undef SUPPORT_PCREGREP_JIT */ + +/* Define to enable support for Unicode properties. */ +/* #undef SUPPORT_UCP */ + +/* Define to enable support for the UTF-8/16 Unicode encoding. This will work + even in an EBCDIC environment, but it is incompatible with the EBCDIC + macro. That is, PCRE can support *either* EBCDIC code *or* ASCII/UTF-8/16, + but not both at once. */ +/* #undef SUPPORT_UTF */ + /* Version number of package */ #ifndef VERSION -#define VERSION "7.7" +#define VERSION "8.31" #endif /* Define to empty if `const' does not conform to ANSI C. */ /* #undef const */ + +/* Define to the type of a signed integer type of width exactly 64 bits if + such a type exists and the standard includes do not define it. */ +/* #undef int64_t */ /* Define to `unsigned int' if does not define. */ /* #undef size_t */ diff --git a/pcre/dftables.c b/pcre/dftables.c deleted file mode 100644 --- a/pcre/dftables.c +++ /dev/null @@ -1,199 +0,0 @@ -/************************************************* -* Perl-Compatible Regular Expressions * -*************************************************/ - -/* PCRE is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. - - Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* This is a freestanding support program to generate a file containing -character tables for PCRE. The tables are built according to the current -locale. Now that pcre_maketables is a function visible to the outside world, we -make use of its code from here in order to be consistent. */ - -#ifdef HAVE_CONFIG_H -#include "config.h" -#endif - -#include -#include -#include -#include - -#include "pcre_internal.h" - -#define DFTABLES /* pcre_maketables.c notices this */ -#include "pcre_maketables.c" - - -int main(int argc, char **argv) -{ -FILE *f; -int i = 1; -const unsigned char *tables; -const unsigned char *base_of_tables; - -/* By default, the default C locale is used rather than what the building user -happens to have set. However, if the -L option is given, set the locale from -the LC_xxx environment variables. */ - -if (argc > 1 && strcmp(argv[1], "-L") == 0) - { - setlocale(LC_ALL, ""); /* Set from environment variables */ - i++; - } - -if (argc < i + 1) - { - fprintf(stderr, "dftables: one filename argument is required\n"); - return 1; - } - -tables = pcre_maketables(); -base_of_tables = tables; - -f = fopen(argv[i], "wb"); -if (f == NULL) - { - fprintf(stderr, "dftables: failed to open %s for writing\n", argv[1]); - return 1; - } - -/* There are several fprintf() calls here, because gcc in pedantic mode -complains about the very long string otherwise. */ - -fprintf(f, - "/*************************************************\n" - "* Perl-Compatible Regular Expressions *\n" - "*************************************************/\n\n" - "/* This file was automatically written by the dftables auxiliary\n" - "program. It contains character tables that are used when no external\n" - "tables are passed to PCRE by the application that calls it. The tables\n" - "are used only for characters whose code values are less than 256.\n\n"); -fprintf(f, - "The following #includes are present because without them gcc 4.x may remove\n" - "the array definition from the final binary if PCRE is built into a static\n" - "library and dead code stripping is activated. This leads to link errors.\n" - "Pulling in the header ensures that the array gets flagged as \"someone\n" - "outside this compilation unit might reference this\" and so it will always\n" - "be supplied to the linker. */\n\n" - "#ifdef HAVE_CONFIG_H\n" - "#include \"config.h\"\n" - "#endif\n\n" - "#include \"pcre_internal.h\"\n\n"); -fprintf(f, - "const unsigned char _pcre_default_tables[] = {\n\n" - "/* This table is a lower casing table. */\n\n"); - -fprintf(f, " "); -for (i = 0; i < 256; i++) - { - if ((i & 7) == 0 && i != 0) fprintf(f, "\n "); - fprintf(f, "%3d", *tables++); - if (i != 255) fprintf(f, ","); - } -fprintf(f, ",\n\n"); - -fprintf(f, "/* This table is a case flipping table. */\n\n"); - -fprintf(f, " "); -for (i = 0; i < 256; i++) - { - if ((i & 7) == 0 && i != 0) fprintf(f, "\n "); - fprintf(f, "%3d", *tables++); - if (i != 255) fprintf(f, ","); - } -fprintf(f, ",\n\n"); - -fprintf(f, - "/* This table contains bit maps for various character classes.\n" - "Each map is 32 bytes long and the bits run from the least\n" - "significant end of each byte. The classes that have their own\n" - "maps are: space, xdigit, digit, upper, lower, word, graph\n" - "print, punct, and cntrl. Other classes are built from combinations. */\n\n"); - -fprintf(f, " "); -for (i = 0; i < cbit_length; i++) - { - if ((i & 7) == 0 && i != 0) - { - if ((i & 31) == 0) fprintf(f, "\n"); - fprintf(f, "\n "); - } - fprintf(f, "0x%02x", *tables++); - if (i != cbit_length - 1) fprintf(f, ","); - } -fprintf(f, ",\n\n"); - -fprintf(f, - "/* This table identifies various classes of character by individual bits:\n" - " 0x%02x white space character\n" - " 0x%02x letter\n" - " 0x%02x decimal digit\n" - " 0x%02x hexadecimal digit\n" - " 0x%02x alphanumeric or '_'\n" - " 0x%02x regular expression metacharacter or binary zero\n*/\n\n", - ctype_space, ctype_letter, ctype_digit, ctype_xdigit, ctype_word, - ctype_meta); - -fprintf(f, " "); -for (i = 0; i < 256; i++) - { - if ((i & 7) == 0 && i != 0) - { - fprintf(f, " /* "); - if (isprint(i-8)) fprintf(f, " %c -", i-8); - else fprintf(f, "%3d-", i-8); - if (isprint(i-1)) fprintf(f, " %c ", i-1); - else fprintf(f, "%3d", i-1); - fprintf(f, " */\n "); - } - fprintf(f, "0x%02x", *tables++); - if (i != 255) fprintf(f, ","); - } - -fprintf(f, "};/* "); -if (isprint(i-8)) fprintf(f, " %c -", i-8); - else fprintf(f, "%3d-", i-8); -if (isprint(i-1)) fprintf(f, " %c ", i-1); - else fprintf(f, "%3d", i-1); -fprintf(f, " */\n\n/* End of pcre_chartables.c */\n"); - -fclose(f); -free((void *)base_of_tables); -return 0; -} - -/* End of dftables.c */ diff --git a/pcre/pcre.h b/pcre/pcre.h --- a/pcre/pcre.h +++ b/pcre/pcre.h @@ -5,7 +5,7 @@ /* This is the public header file for the PCRE library, to be #included by applications that call the PCRE functions. - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -41,10 +41,10 @@ /* The current PCRE version information. */ -#define PCRE_MAJOR 7 -#define PCRE_MINOR 7 +#define PCRE_MAJOR 8 +#define PCRE_MINOR 31 #define PCRE_PRERELEASE -#define PCRE_DATE 2008-05-07 +#define PCRE_DATE 2012-07-06 /* When an application links to a PCRE DLL in Windows, the symbols that are imported have to be identified as such. When building PCRE, the appropriate @@ -95,64 +95,125 @@ extern "C" { #endif -/* Options */ +/* Options. Some are compile-time only, some are run-time only, and some are +both, so we keep them all distinct. However, almost all the bits in the options +word are now used. In the long run, we may have to re-use some of the +compile-time only bits for runtime options, or vice versa. In the comments +below, "compile", "exec", and "DFA exec" mean that the option is permitted to +be set for those functions; "used in" means that an option may be set only for +compile, but is subsequently referenced in exec and/or DFA exec. Any of the +compile-time options may be inspected during studying (and therefore JIT +compiling). */ -#define PCRE_CASELESS 0x00000001 -#define PCRE_MULTILINE 0x00000002 -#define PCRE_DOTALL 0x00000004 -#define PCRE_EXTENDED 0x00000008 -#define PCRE_ANCHORED 0x00000010 -#define PCRE_DOLLAR_ENDONLY 0x00000020 -#define PCRE_EXTRA 0x00000040 -#define PCRE_NOTBOL 0x00000080 -#define PCRE_NOTEOL 0x00000100 -#define PCRE_UNGREEDY 0x00000200 -#define PCRE_NOTEMPTY 0x00000400 -#define PCRE_UTF8 0x00000800 -#define PCRE_NO_AUTO_CAPTURE 0x00001000 -#define PCRE_NO_UTF8_CHECK 0x00002000 -#define PCRE_AUTO_CALLOUT 0x00004000 -#define PCRE_PARTIAL 0x00008000 -#define PCRE_DFA_SHORTEST 0x00010000 -#define PCRE_DFA_RESTART 0x00020000 -#define PCRE_FIRSTLINE 0x00040000 -#define PCRE_DUPNAMES 0x00080000 -#define PCRE_NEWLINE_CR 0x00100000 -#define PCRE_NEWLINE_LF 0x00200000 -#define PCRE_NEWLINE_CRLF 0x00300000 -#define PCRE_NEWLINE_ANY 0x00400000 -#define PCRE_NEWLINE_ANYCRLF 0x00500000 -#define PCRE_BSR_ANYCRLF 0x00800000 -#define PCRE_BSR_UNICODE 0x01000000 -#define PCRE_JAVASCRIPT_COMPAT 0x02000000 +#define PCRE_CASELESS 0x00000001 /* Compile */ +#define PCRE_MULTILINE 0x00000002 /* Compile */ +#define PCRE_DOTALL 0x00000004 /* Compile */ +#define PCRE_EXTENDED 0x00000008 /* Compile */ +#define PCRE_ANCHORED 0x00000010 /* Compile, exec, DFA exec */ +#define PCRE_DOLLAR_ENDONLY 0x00000020 /* Compile, used in exec, DFA exec */ +#define PCRE_EXTRA 0x00000040 /* Compile */ +#define PCRE_NOTBOL 0x00000080 /* Exec, DFA exec */ +#define PCRE_NOTEOL 0x00000100 /* Exec, DFA exec */ +#define PCRE_UNGREEDY 0x00000200 /* Compile */ +#define PCRE_NOTEMPTY 0x00000400 /* Exec, DFA exec */ +/* The next two are also used in exec and DFA exec */ +#define PCRE_UTF8 0x00000800 /* Compile (same as PCRE_UTF16) */ +#define PCRE_UTF16 0x00000800 /* Compile (same as PCRE_UTF8) */ +#define PCRE_NO_AUTO_CAPTURE 0x00001000 /* Compile */ +/* The next two are also used in exec and DFA exec */ +#define PCRE_NO_UTF8_CHECK 0x00002000 /* Compile (same as PCRE_NO_UTF16_CHECK) */ +#define PCRE_NO_UTF16_CHECK 0x00002000 /* Compile (same as PCRE_NO_UTF8_CHECK) */ +#define PCRE_AUTO_CALLOUT 0x00004000 /* Compile */ +#define PCRE_PARTIAL_SOFT 0x00008000 /* Exec, DFA exec */ +#define PCRE_PARTIAL 0x00008000 /* Backwards compatible synonym */ +#define PCRE_DFA_SHORTEST 0x00010000 /* DFA exec */ +#define PCRE_DFA_RESTART 0x00020000 /* DFA exec */ +#define PCRE_FIRSTLINE 0x00040000 /* Compile, used in exec, DFA exec */ +#define PCRE_DUPNAMES 0x00080000 /* Compile */ +#define PCRE_NEWLINE_CR 0x00100000 /* Compile, exec, DFA exec */ +#define PCRE_NEWLINE_LF 0x00200000 /* Compile, exec, DFA exec */ +#define PCRE_NEWLINE_CRLF 0x00300000 /* Compile, exec, DFA exec */ +#define PCRE_NEWLINE_ANY 0x00400000 /* Compile, exec, DFA exec */ +#define PCRE_NEWLINE_ANYCRLF 0x00500000 /* Compile, exec, DFA exec */ +#define PCRE_BSR_ANYCRLF 0x00800000 /* Compile, exec, DFA exec */ +#define PCRE_BSR_UNICODE 0x01000000 /* Compile, exec, DFA exec */ +#define PCRE_JAVASCRIPT_COMPAT 0x02000000 /* Compile, used in exec */ +#define PCRE_NO_START_OPTIMIZE 0x04000000 /* Compile, exec, DFA exec */ +#define PCRE_NO_START_OPTIMISE 0x04000000 /* Synonym */ +#define PCRE_PARTIAL_HARD 0x08000000 /* Exec, DFA exec */ +#define PCRE_NOTEMPTY_ATSTART 0x10000000 /* Exec, DFA exec */ +#define PCRE_UCP 0x20000000 /* Compile, used in exec, DFA exec */ /* Exec-time and get/set-time error codes */ -#define PCRE_ERROR_NOMATCH (-1) -#define PCRE_ERROR_NULL (-2) -#define PCRE_ERROR_BADOPTION (-3) -#define PCRE_ERROR_BADMAGIC (-4) -#define PCRE_ERROR_UNKNOWN_OPCODE (-5) -#define PCRE_ERROR_UNKNOWN_NODE (-5) /* For backward compatibility */ -#define PCRE_ERROR_NOMEMORY (-6) -#define PCRE_ERROR_NOSUBSTRING (-7) -#define PCRE_ERROR_MATCHLIMIT (-8) -#define PCRE_ERROR_CALLOUT (-9) /* Never used by PCRE itself */ -#define PCRE_ERROR_BADUTF8 (-10) -#define PCRE_ERROR_BADUTF8_OFFSET (-11) -#define PCRE_ERROR_PARTIAL (-12) -#define PCRE_ERROR_BADPARTIAL (-13) -#define PCRE_ERROR_INTERNAL (-14) -#define PCRE_ERROR_BADCOUNT (-15) -#define PCRE_ERROR_DFA_UITEM (-16) -#define PCRE_ERROR_DFA_UCOND (-17) -#define PCRE_ERROR_DFA_UMLIMIT (-18) -#define PCRE_ERROR_DFA_WSSIZE (-19) -#define PCRE_ERROR_DFA_RECURSE (-20) -#define PCRE_ERROR_RECURSIONLIMIT (-21) -#define PCRE_ERROR_NULLWSLIMIT (-22) /* No longer actually used */ -#define PCRE_ERROR_BADNEWLINE (-23) +#define PCRE_ERROR_NOMATCH (-1) +#define PCRE_ERROR_NULL (-2) +#define PCRE_ERROR_BADOPTION (-3) +#define PCRE_ERROR_BADMAGIC (-4) +#define PCRE_ERROR_UNKNOWN_OPCODE (-5) +#define PCRE_ERROR_UNKNOWN_NODE (-5) /* For backward compatibility */ +#define PCRE_ERROR_NOMEMORY (-6) +#define PCRE_ERROR_NOSUBSTRING (-7) +#define PCRE_ERROR_MATCHLIMIT (-8) +#define PCRE_ERROR_CALLOUT (-9) /* Never used by PCRE itself */ +#define PCRE_ERROR_BADUTF8 (-10) /* Same for 8/16 */ +#define PCRE_ERROR_BADUTF16 (-10) /* Same for 8/16 */ +#define PCRE_ERROR_BADUTF8_OFFSET (-11) /* Same for 8/16 */ +#define PCRE_ERROR_BADUTF16_OFFSET (-11) /* Same for 8/16 */ +#define PCRE_ERROR_PARTIAL (-12) +#define PCRE_ERROR_BADPARTIAL (-13) +#define PCRE_ERROR_INTERNAL (-14) +#define PCRE_ERROR_BADCOUNT (-15) +#define PCRE_ERROR_DFA_UITEM (-16) +#define PCRE_ERROR_DFA_UCOND (-17) +#define PCRE_ERROR_DFA_UMLIMIT (-18) +#define PCRE_ERROR_DFA_WSSIZE (-19) +#define PCRE_ERROR_DFA_RECURSE (-20) +#define PCRE_ERROR_RECURSIONLIMIT (-21) +#define PCRE_ERROR_NULLWSLIMIT (-22) /* No longer actually used */ +#define PCRE_ERROR_BADNEWLINE (-23) +#define PCRE_ERROR_BADOFFSET (-24) +#define PCRE_ERROR_SHORTUTF8 (-25) +#define PCRE_ERROR_SHORTUTF16 (-25) /* Same for 8/16 */ +#define PCRE_ERROR_RECURSELOOP (-26) +#define PCRE_ERROR_JIT_STACKLIMIT (-27) +#define PCRE_ERROR_BADMODE (-28) +#define PCRE_ERROR_BADENDIANNESS (-29) +#define PCRE_ERROR_DFA_BADRESTART (-30) +/* Specific error codes for UTF-8 validity checks */ + +#define PCRE_UTF8_ERR0 0 +#define PCRE_UTF8_ERR1 1 +#define PCRE_UTF8_ERR2 2 +#define PCRE_UTF8_ERR3 3 +#define PCRE_UTF8_ERR4 4 +#define PCRE_UTF8_ERR5 5 +#define PCRE_UTF8_ERR6 6 +#define PCRE_UTF8_ERR7 7 +#define PCRE_UTF8_ERR8 8 +#define PCRE_UTF8_ERR9 9 +#define PCRE_UTF8_ERR10 10 +#define PCRE_UTF8_ERR11 11 +#define PCRE_UTF8_ERR12 12 +#define PCRE_UTF8_ERR13 13 +#define PCRE_UTF8_ERR14 14 +#define PCRE_UTF8_ERR15 15 +#define PCRE_UTF8_ERR16 16 +#define PCRE_UTF8_ERR17 17 +#define PCRE_UTF8_ERR18 18 +#define PCRE_UTF8_ERR19 19 +#define PCRE_UTF8_ERR20 20 +#define PCRE_UTF8_ERR21 21 + +/* Specific error codes for UTF-16 validity checks */ + +#define PCRE_UTF16_ERR0 0 +#define PCRE_UTF16_ERR1 1 +#define PCRE_UTF16_ERR2 2 +#define PCRE_UTF16_ERR3 3 +#define PCRE_UTF16_ERR4 4 + /* Request types for pcre_fullinfo() */ #define PCRE_INFO_OPTIONS 0 @@ -171,6 +232,10 @@ #define PCRE_INFO_OKPARTIAL 12 #define PCRE_INFO_JCHANGED 13 #define PCRE_INFO_HASCRORLF 14 +#define PCRE_INFO_MINLENGTH 15 +#define PCRE_INFO_JIT 16 +#define PCRE_INFO_JITSIZE 17 +#define PCRE_INFO_MAXLOOKBEHIND 18 /* Request types for pcre_config(). Do not re-arrange, in order to remain compatible. */ @@ -184,8 +249,18 @@ #define PCRE_CONFIG_UNICODE_PROPERTIES 6 #define PCRE_CONFIG_MATCH_LIMIT_RECURSION 7 #define PCRE_CONFIG_BSR 8 +#define PCRE_CONFIG_JIT 9 +#define PCRE_CONFIG_UTF16 10 +#define PCRE_CONFIG_JITTARGET 11 -/* Bit flags for the pcre_extra structure. Do not re-arrange or redefine +/* Request types for pcre_study(). Do not re-arrange, in order to remain +compatible. */ + +#define PCRE_STUDY_JIT_COMPILE 0x0001 +#define PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE 0x0002 +#define PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE 0x0004 + +/* Bit flags for the pcre[16]_extra structure. Do not re-arrange or redefine these bits, just add new ones on the end, in order to remain compatible. */ #define PCRE_EXTRA_STUDY_DATA 0x0001 @@ -193,12 +268,34 @@ #define PCRE_EXTRA_CALLOUT_DATA 0x0004 #define PCRE_EXTRA_TABLES 0x0008 #define PCRE_EXTRA_MATCH_LIMIT_RECURSION 0x0010 +#define PCRE_EXTRA_MARK 0x0020 +#define PCRE_EXTRA_EXECUTABLE_JIT 0x0040 /* Types */ struct real_pcre; /* declaration; the definition is private */ typedef struct real_pcre pcre; +struct real_pcre16; /* declaration; the definition is private */ +typedef struct real_pcre16 pcre16; + +struct real_pcre_jit_stack; /* declaration; the definition is private */ +typedef struct real_pcre_jit_stack pcre_jit_stack; + +struct real_pcre16_jit_stack; /* declaration; the definition is private */ +typedef struct real_pcre16_jit_stack pcre16_jit_stack; + +/* If PCRE is compiled with 16 bit character support, PCRE_UCHAR16 must contain +a 16 bit wide signed data type. Otherwise it can be a dummy data type since +pcre16 functions are not implemented. There is a check for this in pcre_internal.h. */ +#ifndef PCRE_UCHAR16 +#define PCRE_UCHAR16 unsigned short +#endif + +#ifndef PCRE_SPTR16 +#define PCRE_SPTR16 const PCRE_UCHAR16 * +#endif + /* When PCRE is compiled as a C++ library, the subject pointer type can be replaced with a custom type. For conventional use, the public interface is a const char *. */ @@ -218,8 +315,23 @@ void *callout_data; /* Data passed back in callouts */ const unsigned char *tables; /* Pointer to character tables */ unsigned long int match_limit_recursion; /* Max recursive calls to match() */ + unsigned char **mark; /* For passing back a mark pointer */ + void *executable_jit; /* Contains a pointer to a compiled jit code */ } pcre_extra; +/* Same structure as above, but with 16 bit char pointers. */ + +typedef struct pcre16_extra { + unsigned long int flags; /* Bits for which fields are set */ + void *study_data; /* Opaque data from pcre_study() */ + unsigned long int match_limit; /* Maximum number of calls to match() */ + void *callout_data; /* Data passed back in callouts */ + const unsigned char *tables; /* Pointer to character tables */ + unsigned long int match_limit_recursion; /* Max recursive calls to match() */ + PCRE_UCHAR16 **mark; /* For passing back a mark pointer */ + void *executable_jit; /* Contains a pointer to a compiled jit code */ +} pcre16_extra; + /* The structure for passing out data via the pcre_callout_function. We use a structure so that new fields can be added on the end in future versions, without changing the API of the function, thereby allowing old clients to work @@ -240,9 +352,33 @@ /* ------------------- Added for Version 1 -------------------------- */ int pattern_position; /* Offset to next item in the pattern */ int next_item_length; /* Length of next item in the pattern */ + /* ------------------- Added for Version 2 -------------------------- */ + const unsigned char *mark; /* Pointer to current mark or NULL */ /* ------------------------------------------------------------------ */ } pcre_callout_block; +/* Same structure as above, but with 16 bit char pointers. */ + +typedef struct pcre16_callout_block { + int version; /* Identifies version of block */ + /* ------------------------ Version 0 ------------------------------- */ + int callout_number; /* Number compiled into pattern */ + int *offset_vector; /* The offset vector */ + PCRE_SPTR16 subject; /* The subject being matched */ + int subject_length; /* The length of the subject */ + int start_match; /* Offset to start of this match attempt */ + int current_position; /* Where we currently are in the subject */ + int capture_top; /* Max current capture */ + int capture_last; /* Most recently closed capture */ + void *callout_data; /* Data passed in with the call */ + /* ------------------- Added for Version 1 -------------------------- */ + int pattern_position; /* Offset to next item in the pattern */ + int next_item_length; /* Length of next item in the pattern */ + /* ------------------- Added for Version 2 -------------------------- */ + const PCRE_UCHAR16 *mark; /* Pointer to current mark or NULL */ + /* ------------------------------------------------------------------ */ +} pcre16_callout_block; + /* Indirection for store get and free functions. These can be set to alternative malloc/free functions if required. Special ones are used in the non-recursive case for "frames". There is also an optional callout function @@ -255,47 +391,114 @@ PCRE_EXP_DECL void *(*pcre_stack_malloc)(size_t); PCRE_EXP_DECL void (*pcre_stack_free)(void *); PCRE_EXP_DECL int (*pcre_callout)(pcre_callout_block *); + +PCRE_EXP_DECL void *(*pcre16_malloc)(size_t); +PCRE_EXP_DECL void (*pcre16_free)(void *); +PCRE_EXP_DECL void *(*pcre16_stack_malloc)(size_t); +PCRE_EXP_DECL void (*pcre16_stack_free)(void *); +PCRE_EXP_DECL int (*pcre16_callout)(pcre16_callout_block *); #else /* VPCOMPAT */ PCRE_EXP_DECL void *pcre_malloc(size_t); PCRE_EXP_DECL void pcre_free(void *); PCRE_EXP_DECL void *pcre_stack_malloc(size_t); PCRE_EXP_DECL void pcre_stack_free(void *); PCRE_EXP_DECL int pcre_callout(pcre_callout_block *); + +PCRE_EXP_DECL void *pcre16_malloc(size_t); +PCRE_EXP_DECL void pcre16_free(void *); +PCRE_EXP_DECL void *pcre16_stack_malloc(size_t); +PCRE_EXP_DECL void pcre16_stack_free(void *); +PCRE_EXP_DECL int pcre16_callout(pcre16_callout_block *); #endif /* VPCOMPAT */ +/* User defined callback which provides a stack just before the match starts. */ + +typedef pcre_jit_stack *(*pcre_jit_callback)(void *); +typedef pcre16_jit_stack *(*pcre16_jit_callback)(void *); + /* Exported PCRE functions */ PCRE_EXP_DECL pcre *pcre_compile(const char *, int, const char **, int *, const unsigned char *); +PCRE_EXP_DECL pcre16 *pcre16_compile(PCRE_SPTR16, int, const char **, int *, + const unsigned char *); PCRE_EXP_DECL pcre *pcre_compile2(const char *, int, int *, const char **, int *, const unsigned char *); +PCRE_EXP_DECL pcre16 *pcre16_compile2(PCRE_SPTR16, int, int *, const char **, + int *, const unsigned char *); PCRE_EXP_DECL int pcre_config(int, void *); +PCRE_EXP_DECL int pcre16_config(int, void *); PCRE_EXP_DECL int pcre_copy_named_substring(const pcre *, const char *, int *, int, const char *, char *, int); -PCRE_EXP_DECL int pcre_copy_substring(const char *, int *, int, int, char *, - int); +PCRE_EXP_DECL int pcre16_copy_named_substring(const pcre16 *, PCRE_SPTR16, + int *, int, PCRE_SPTR16, PCRE_UCHAR16 *, int); +PCRE_EXP_DECL int pcre_copy_substring(const char *, int *, int, int, + char *, int); +PCRE_EXP_DECL int pcre16_copy_substring(PCRE_SPTR16, int *, int, int, + PCRE_UCHAR16 *, int); PCRE_EXP_DECL int pcre_dfa_exec(const pcre *, const pcre_extra *, const char *, int, int, int, int *, int , int *, int); +PCRE_EXP_DECL int pcre16_dfa_exec(const pcre16 *, const pcre16_extra *, + PCRE_SPTR16, int, int, int, int *, int , int *, int); PCRE_EXP_DECL int pcre_exec(const pcre *, const pcre_extra *, PCRE_SPTR, int, int, int, int *, int); +PCRE_EXP_DECL int pcre16_exec(const pcre16 *, const pcre16_extra *, + PCRE_SPTR16, int, int, int, int *, int); PCRE_EXP_DECL void pcre_free_substring(const char *); +PCRE_EXP_DECL void pcre16_free_substring(PCRE_SPTR16); PCRE_EXP_DECL void pcre_free_substring_list(const char **); +PCRE_EXP_DECL void pcre16_free_substring_list(PCRE_SPTR16 *); PCRE_EXP_DECL int pcre_fullinfo(const pcre *, const pcre_extra *, int, void *); +PCRE_EXP_DECL int pcre16_fullinfo(const pcre16 *, const pcre16_extra *, int, + void *); PCRE_EXP_DECL int pcre_get_named_substring(const pcre *, const char *, int *, int, const char *, const char **); +PCRE_EXP_DECL int pcre16_get_named_substring(const pcre16 *, PCRE_SPTR16, + int *, int, PCRE_SPTR16, PCRE_SPTR16 *); PCRE_EXP_DECL int pcre_get_stringnumber(const pcre *, const char *); +PCRE_EXP_DECL int pcre16_get_stringnumber(const pcre16 *, PCRE_SPTR16); PCRE_EXP_DECL int pcre_get_stringtable_entries(const pcre *, const char *, char **, char **); +PCRE_EXP_DECL int pcre16_get_stringtable_entries(const pcre16 *, PCRE_SPTR16, + PCRE_UCHAR16 **, PCRE_UCHAR16 **); PCRE_EXP_DECL int pcre_get_substring(const char *, int *, int, int, const char **); +PCRE_EXP_DECL int pcre16_get_substring(PCRE_SPTR16, int *, int, int, + PCRE_SPTR16 *); PCRE_EXP_DECL int pcre_get_substring_list(const char *, int *, int, const char ***); -PCRE_EXP_DECL int pcre_info(const pcre *, int *, int *); +PCRE_EXP_DECL int pcre16_get_substring_list(PCRE_SPTR16, int *, int, + PCRE_SPTR16 **); PCRE_EXP_DECL const unsigned char *pcre_maketables(void); +PCRE_EXP_DECL const unsigned char *pcre16_maketables(void); PCRE_EXP_DECL int pcre_refcount(pcre *, int); +PCRE_EXP_DECL int pcre16_refcount(pcre16 *, int); PCRE_EXP_DECL pcre_extra *pcre_study(const pcre *, int, const char **); +PCRE_EXP_DECL pcre16_extra *pcre16_study(const pcre16 *, int, const char **); +PCRE_EXP_DECL void pcre_free_study(pcre_extra *); +PCRE_EXP_DECL void pcre16_free_study(pcre16_extra *); PCRE_EXP_DECL const char *pcre_version(void); +PCRE_EXP_DECL const char *pcre16_version(void); + +/* Utility functions for byte order swaps. */ +PCRE_EXP_DECL int pcre_pattern_to_host_byte_order(pcre *, pcre_extra *, + const unsigned char *); +PCRE_EXP_DECL int pcre16_pattern_to_host_byte_order(pcre16 *, pcre16_extra *, + const unsigned char *); +PCRE_EXP_DECL int pcre16_utf16_to_host_byte_order(PCRE_UCHAR16 *, + PCRE_SPTR16, int, int *, int); + +/* JIT compiler related functions. */ + +PCRE_EXP_DECL pcre_jit_stack *pcre_jit_stack_alloc(int, int); +PCRE_EXP_DECL pcre16_jit_stack *pcre16_jit_stack_alloc(int, int); +PCRE_EXP_DECL void pcre_jit_stack_free(pcre_jit_stack *); +PCRE_EXP_DECL void pcre16_jit_stack_free(pcre16_jit_stack *); +PCRE_EXP_DECL void pcre_assign_jit_stack(pcre_extra *, + pcre_jit_callback, void *); +PCRE_EXP_DECL void pcre16_assign_jit_stack(pcre16_extra *, + pcre16_jit_callback, void *); #ifdef __cplusplus } /* extern "C" */ diff --git a/pcre/pcre_byte_order.c b/pcre/pcre_byte_order.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_byte_order.c @@ -0,0 +1,288 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Written by Philip Hazel + Copyright (c) 1997-2012 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + + +/* This module contains an internal function that tests a compiled pattern to +see if it was compiled with the opposite endianness. If so, it uses an +auxiliary local function to flip the appropriate bytes. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + +/************************************************* +* Swap byte functions * +*************************************************/ + +/* The following functions swap the bytes of a pcre_uint16 +and pcre_uint32 value. + +Arguments: + value any number + +Returns: the byte swapped value +*/ + +static pcre_uint32 +swap_uint32(pcre_uint32 value) +{ +return ((value & 0x000000ff) << 24) | + ((value & 0x0000ff00) << 8) | + ((value & 0x00ff0000) >> 8) | + (value >> 24); +} + +static pcre_uint16 +swap_uint16(pcre_uint16 value) +{ +return (value >> 8) | (value << 8); +} + + +/************************************************* +* Test for a byte-flipped compiled regex * +*************************************************/ + +/* This function swaps the bytes of a compiled pattern usually +loaded form the disk. It also sets the tables pointer, which +is likely an invalid pointer after reload. + +Arguments: + argument_re points to the compiled expression + extra_data points to extra data or is NULL + tables points to the character tables or NULL + +Returns: 0 if the swap is successful, negative on error +*/ + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DECL int pcre_pattern_to_host_byte_order(pcre *argument_re, + pcre_extra *extra_data, const unsigned char *tables) +#else +PCRE_EXP_DECL int pcre16_pattern_to_host_byte_order(pcre16 *argument_re, + pcre16_extra *extra_data, const unsigned char *tables) +#endif +{ +REAL_PCRE *re = (REAL_PCRE *)argument_re; +pcre_study_data *study; +#ifndef COMPILE_PCRE8 +pcre_uchar *ptr; +int length; +#ifdef SUPPORT_UTF +BOOL utf; +BOOL utf16_char; +#endif /* SUPPORT_UTF */ +#endif /* !COMPILE_PCRE8 */ + +if (re == NULL) return PCRE_ERROR_NULL; +if (re->magic_number == MAGIC_NUMBER) + { + if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE; + re->tables = tables; + return 0; + } + +if (re->magic_number != REVERSED_MAGIC_NUMBER) return PCRE_ERROR_BADMAGIC; +if ((swap_uint16(re->flags) & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE; + +re->magic_number = MAGIC_NUMBER; +re->size = swap_uint32(re->size); +re->options = swap_uint32(re->options); +re->flags = swap_uint16(re->flags); +re->top_bracket = swap_uint16(re->top_bracket); +re->top_backref = swap_uint16(re->top_backref); +re->first_char = swap_uint16(re->first_char); +re->req_char = swap_uint16(re->req_char); +re->name_table_offset = swap_uint16(re->name_table_offset); +re->name_entry_size = swap_uint16(re->name_entry_size); +re->name_count = swap_uint16(re->name_count); +re->ref_count = swap_uint16(re->ref_count); +re->tables = tables; + +if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_STUDY_DATA) != 0) + { + study = (pcre_study_data *)extra_data->study_data; + study->size = swap_uint32(study->size); + study->flags = swap_uint32(study->flags); + study->minlength = swap_uint32(study->minlength); + } + +#ifndef COMPILE_PCRE8 +ptr = (pcre_uchar *)re + re->name_table_offset; +length = re->name_count * re->name_entry_size; +#ifdef SUPPORT_UTF +utf = (re->options & PCRE_UTF16) != 0; +utf16_char = FALSE; +#endif + +while(TRUE) + { + /* Swap previous characters. */ + while (length-- > 0) + { + *ptr = swap_uint16(*ptr); + ptr++; + } +#ifdef SUPPORT_UTF + if (utf16_char) + { + if (HAS_EXTRALEN(ptr[-1])) + { + /* We know that there is only one extra character in UTF-16. */ + *ptr = swap_uint16(*ptr); + ptr++; + } + } + utf16_char = FALSE; +#endif /* SUPPORT_UTF */ + + /* Get next opcode. */ + length = 0; + *ptr = swap_uint16(*ptr); + switch (*ptr) + { + case OP_END: + return 0; + +#ifdef SUPPORT_UTF + case OP_CHAR: + case OP_CHARI: + case OP_NOT: + case OP_NOTI: + case OP_STAR: + case OP_MINSTAR: + case OP_PLUS: + case OP_MINPLUS: + case OP_QUERY: + case OP_MINQUERY: + case OP_UPTO: + case OP_MINUPTO: + case OP_EXACT: + case OP_POSSTAR: + case OP_POSPLUS: + case OP_POSQUERY: + case OP_POSUPTO: + case OP_STARI: + case OP_MINSTARI: + case OP_PLUSI: + case OP_MINPLUSI: + case OP_QUERYI: + case OP_MINQUERYI: + case OP_UPTOI: + case OP_MINUPTOI: + case OP_EXACTI: + case OP_POSSTARI: + case OP_POSPLUSI: + case OP_POSQUERYI: + case OP_POSUPTOI: + case OP_NOTSTAR: + case OP_NOTMINSTAR: + case OP_NOTPLUS: + case OP_NOTMINPLUS: + case OP_NOTQUERY: + case OP_NOTMINQUERY: + case OP_NOTUPTO: + case OP_NOTMINUPTO: + case OP_NOTEXACT: + case OP_NOTPOSSTAR: + case OP_NOTPOSPLUS: + case OP_NOTPOSQUERY: + case OP_NOTPOSUPTO: + case OP_NOTSTARI: + case OP_NOTMINSTARI: + case OP_NOTPLUSI: + case OP_NOTMINPLUSI: + case OP_NOTQUERYI: + case OP_NOTMINQUERYI: + case OP_NOTUPTOI: + case OP_NOTMINUPTOI: + case OP_NOTEXACTI: + case OP_NOTPOSSTARI: + case OP_NOTPOSPLUSI: + case OP_NOTPOSQUERYI: + case OP_NOTPOSUPTOI: + if (utf) utf16_char = TRUE; +#endif + /* Fall through. */ + + default: + length = PRIV(OP_lengths)[*ptr] - 1; + break; + + case OP_CLASS: + case OP_NCLASS: + /* Skip the character bit map. */ + ptr += 32/sizeof(pcre_uchar); + length = 0; + break; + + case OP_XCLASS: + /* Reverse the size of the XCLASS instance. */ + ptr++; + *ptr = swap_uint16(*ptr); + if (LINK_SIZE > 1) + { + /* LINK_SIZE can be 1 or 2 in 16 bit mode. */ + ptr++; + *ptr = swap_uint16(*ptr); + } + ptr++; + length = (GET(ptr, -LINK_SIZE)) - (1 + LINK_SIZE + 1); + *ptr = swap_uint16(*ptr); + if ((*ptr & XCL_MAP) != 0) + { + /* Skip the character bit map. */ + ptr += 32/sizeof(pcre_uchar); + length -= 32/sizeof(pcre_uchar); + } + break; + } + ptr++; + } +/* Control should never reach here in 16 bit mode. */ +#endif /* !COMPILE_PCRE8 */ + +return 0; +} + +/* End of pcre_byte_order.c */ diff --git a/pcre/pcre_chartables.c b/pcre/pcre_chartables.c deleted file mode 100644 --- a/pcre/pcre_chartables.c +++ /dev/null @@ -1,198 +0,0 @@ -/************************************************* -* Perl-Compatible Regular Expressions * -*************************************************/ - -/* This file contains character tables that are used when no external tables -are passed to PCRE by the application that calls it. The tables are used only -for characters whose code values are less than 256. - -This is a default version of the tables that assumes ASCII encoding. A program -called dftables (which is distributed with PCRE) can be used to build -alternative versions of this file. This is necessary if you are running in an -EBCDIC environment, or if you want to default to a different encoding, for -example ISO-8859-1. When dftables is run, it creates these tables in the -current locale. If PCRE is configured with --enable-rebuild-chartables, this -happens automatically. - -The following #includes are present because without the gcc 4.x may remove the -array definition from the final binary if PCRE is built into a static library -and dead code stripping is activated. This leads to link errors. Pulling in the -header ensures that the array gets flagged as "someone outside this compilation -unit might reference this" and so it will always be supplied to the linker. */ - -#ifdef HAVE_CONFIG_H -#include "config.h" -#endif - -#include "pcre_internal.h" - -const unsigned char _pcre_default_tables[] = { - -/* This table is a lower casing table. */ - - 0, 1, 2, 3, 4, 5, 6, 7, - 8, 9, 10, 11, 12, 13, 14, 15, - 16, 17, 18, 19, 20, 21, 22, 23, - 24, 25, 26, 27, 28, 29, 30, 31, - 32, 33, 34, 35, 36, 37, 38, 39, - 40, 41, 42, 43, 44, 45, 46, 47, - 48, 49, 50, 51, 52, 53, 54, 55, - 56, 57, 58, 59, 60, 61, 62, 63, - 64, 97, 98, 99,100,101,102,103, - 104,105,106,107,108,109,110,111, - 112,113,114,115,116,117,118,119, - 120,121,122, 91, 92, 93, 94, 95, - 96, 97, 98, 99,100,101,102,103, - 104,105,106,107,108,109,110,111, - 112,113,114,115,116,117,118,119, - 120,121,122,123,124,125,126,127, - 128,129,130,131,132,133,134,135, - 136,137,138,139,140,141,142,143, - 144,145,146,147,148,149,150,151, - 152,153,154,155,156,157,158,159, - 160,161,162,163,164,165,166,167, - 168,169,170,171,172,173,174,175, - 176,177,178,179,180,181,182,183, - 184,185,186,187,188,189,190,191, - 192,193,194,195,196,197,198,199, - 200,201,202,203,204,205,206,207, - 208,209,210,211,212,213,214,215, - 216,217,218,219,220,221,222,223, - 224,225,226,227,228,229,230,231, - 232,233,234,235,236,237,238,239, - 240,241,242,243,244,245,246,247, - 248,249,250,251,252,253,254,255, - -/* This table is a case flipping table. */ - - 0, 1, 2, 3, 4, 5, 6, 7, - 8, 9, 10, 11, 12, 13, 14, 15, - 16, 17, 18, 19, 20, 21, 22, 23, - 24, 25, 26, 27, 28, 29, 30, 31, - 32, 33, 34, 35, 36, 37, 38, 39, - 40, 41, 42, 43, 44, 45, 46, 47, - 48, 49, 50, 51, 52, 53, 54, 55, - 56, 57, 58, 59, 60, 61, 62, 63, - 64, 97, 98, 99,100,101,102,103, - 104,105,106,107,108,109,110,111, - 112,113,114,115,116,117,118,119, - 120,121,122, 91, 92, 93, 94, 95, - 96, 65, 66, 67, 68, 69, 70, 71, - 72, 73, 74, 75, 76, 77, 78, 79, - 80, 81, 82, 83, 84, 85, 86, 87, - 88, 89, 90,123,124,125,126,127, - 128,129,130,131,132,133,134,135, - 136,137,138,139,140,141,142,143, - 144,145,146,147,148,149,150,151, - 152,153,154,155,156,157,158,159, - 160,161,162,163,164,165,166,167, - 168,169,170,171,172,173,174,175, - 176,177,178,179,180,181,182,183, - 184,185,186,187,188,189,190,191, - 192,193,194,195,196,197,198,199, - 200,201,202,203,204,205,206,207, - 208,209,210,211,212,213,214,215, - 216,217,218,219,220,221,222,223, - 224,225,226,227,228,229,230,231, - 232,233,234,235,236,237,238,239, - 240,241,242,243,244,245,246,247, - 248,249,250,251,252,253,254,255, - -/* This table contains bit maps for various character classes. Each map is 32 -bytes long and the bits run from the least significant end of each byte. The -classes that have their own maps are: space, xdigit, digit, upper, lower, word, -graph, print, punct, and cntrl. Other classes are built from combinations. */ - - 0x00,0x3e,0x00,0x00,0x01,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - - 0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03, - 0x7e,0x00,0x00,0x00,0x7e,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - - 0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0xfe,0xff,0xff,0x07,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0x07, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - - 0x00,0x00,0x00,0x00,0x00,0x00,0xff,0x03, - 0xfe,0xff,0xff,0x87,0xfe,0xff,0xff,0x07, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - - 0x00,0x00,0x00,0x00,0xfe,0xff,0xff,0xff, - 0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - - 0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff, - 0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - - 0x00,0x00,0x00,0x00,0xfe,0xff,0x00,0xfc, - 0x01,0x00,0x00,0xf8,0x01,0x00,0x00,0x78, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - - 0xff,0xff,0xff,0xff,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, - -/* This table identifies various classes of character by individual bits: - 0x01 white space character - 0x02 letter - 0x04 decimal digit - 0x08 hexadecimal digit - 0x10 alphanumeric or '_' - 0x80 regular expression metacharacter or binary zero -*/ - - 0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 */ - 0x00,0x01,0x01,0x00,0x01,0x01,0x00,0x00, /* 8- 15 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ - 0x01,0x00,0x00,0x00,0x80,0x00,0x00,0x00, /* - ' */ - 0x80,0x80,0x80,0x80,0x00,0x00,0x80,0x00, /* ( - / */ - 0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /* 0 - 7 */ - 0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x80, /* 8 - ? */ - 0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* @ - G */ - 0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* H - O */ - 0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* P - W */ - 0x12,0x12,0x12,0x80,0x80,0x00,0x80,0x10, /* X - _ */ - 0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* ` - g */ - 0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* h - o */ - 0x12,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* p - w */ - 0x12,0x12,0x12,0x80,0x80,0x00,0x00,0x00, /* x -127 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */ - -/* End of pcre_chartables.c */ diff --git a/pcre/pcre_compile.c b/pcre/pcre_compile.c --- a/pcre/pcre_compile.c +++ b/pcre/pcre_compile.c @@ -6,6374 +6,8204 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* This module contains the external function pcre_compile(), along with -supporting internal functions that are not used by other modules. */ - - -#ifdef HAVE_CONFIG_H -#include "config.h" -#endif - -#define NLBLOCK cd /* Block containing newline information */ -#define PSSTART start_pattern /* Field containing processed string start */ -#define PSEND end_pattern /* Field containing processed string end */ - -#include "pcre_internal.h" - - -/* When DEBUG is defined, we need the pcre_printint() function, which is also -used by pcretest. DEBUG is not defined when building a production library. */ - -#ifdef DEBUG -#include "pcre_printint.src" -#endif - - -/* Macro for setting individual bits in class bitmaps. */ - -#define SETBIT(a,b) a[b/8] |= (1 << (b%8)) - -/* Maximum length value to check against when making sure that the integer that -holds the compiled pattern length does not overflow. We make it a bit less than -INT_MAX to allow for adding in group terminating bytes, so that we don't have -to check them every time. */ - -#define OFLOW_MAX (INT_MAX - 20) - - -/************************************************* -* Code parameters and static tables * -*************************************************/ - -/* This value specifies the size of stack workspace that is used during the -first pre-compile phase that determines how much memory is required. The regex -is partly compiled into this space, but the compiled parts are discarded as -soon as they can be, so that hopefully there will never be an overrun. The code -does, however, check for an overrun. The largest amount I've seen used is 218, -so this number is very generous. - -The same workspace is used during the second, actual compile phase for -remembering forward references to groups so that they can be filled in at the -end. Each entry in this list occupies LINK_SIZE bytes, so even when LINK_SIZE -is 4 there is plenty of room. */ - -#define COMPILE_WORK_SIZE (4096) - - -/* Table for handling escaped characters in the range '0'-'z'. Positive returns -are simple data values; negative values are for special things like \d and so -on. Zero means further processing is needed (for things like \x), or the escape -is invalid. */ - -#ifndef EBCDIC /* This is the "normal" table for ASCII systems */ -static const short int escapes[] = { - 0, 0, 0, 0, 0, 0, 0, 0, /* 0 - 7 */ - 0, 0, ':', ';', '<', '=', '>', '?', /* 8 - ? */ - '@', -ESC_A, -ESC_B, -ESC_C, -ESC_D, -ESC_E, 0, -ESC_G, /* @ - G */ --ESC_H, 0, 0, -ESC_K, 0, 0, 0, 0, /* H - O */ --ESC_P, -ESC_Q, -ESC_R, -ESC_S, 0, 0, -ESC_V, -ESC_W, /* P - W */ --ESC_X, 0, -ESC_Z, '[', '\\', ']', '^', '_', /* X - _ */ - '`', 7, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0, /* ` - g */ --ESC_h, 0, 0, -ESC_k, 0, 0, ESC_n, 0, /* h - o */ --ESC_p, 0, ESC_r, -ESC_s, ESC_tee, 0, -ESC_v, -ESC_w, /* p - w */ - 0, 0, -ESC_z /* x - z */ -}; - -#else /* This is the "abnormal" table for EBCDIC systems */ -static const short int escapes[] = { -/* 48 */ 0, 0, 0, '.', '<', '(', '+', '|', -/* 50 */ '&', 0, 0, 0, 0, 0, 0, 0, -/* 58 */ 0, 0, '!', '$', '*', ')', ';', '~', -/* 60 */ '-', '/', 0, 0, 0, 0, 0, 0, -/* 68 */ 0, 0, '|', ',', '%', '_', '>', '?', -/* 70 */ 0, 0, 0, 0, 0, 0, 0, 0, -/* 78 */ 0, '`', ':', '#', '@', '\'', '=', '"', -/* 80 */ 0, 7, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0, -/* 88 */-ESC_h, 0, 0, '{', 0, 0, 0, 0, -/* 90 */ 0, 0, -ESC_k, 'l', 0, ESC_n, 0, -ESC_p, -/* 98 */ 0, ESC_r, 0, '}', 0, 0, 0, 0, -/* A0 */ 0, '~', -ESC_s, ESC_tee, 0,-ESC_v, -ESC_w, 0, -/* A8 */ 0,-ESC_z, 0, 0, 0, '[', 0, 0, -/* B0 */ 0, 0, 0, 0, 0, 0, 0, 0, -/* B8 */ 0, 0, 0, 0, 0, ']', '=', '-', -/* C0 */ '{',-ESC_A, -ESC_B, -ESC_C, -ESC_D,-ESC_E, 0, -ESC_G, -/* C8 */-ESC_H, 0, 0, 0, 0, 0, 0, 0, -/* D0 */ '}', 0, -ESC_K, 0, 0, 0, 0, -ESC_P, -/* D8 */-ESC_Q,-ESC_R, 0, 0, 0, 0, 0, 0, -/* E0 */ '\\', 0, -ESC_S, 0, 0,-ESC_V, -ESC_W, -ESC_X, -/* E8 */ 0,-ESC_Z, 0, 0, 0, 0, 0, 0, -/* F0 */ 0, 0, 0, 0, 0, 0, 0, 0, -/* F8 */ 0, 0, 0, 0, 0, 0, 0, 0 -}; -#endif - - -/* Table of special "verbs" like (*PRUNE). This is a short table, so it is -searched linearly. Put all the names into a single string, in order to reduce -the number of relocations when a shared library is dynamically linked. */ - -typedef struct verbitem { - int len; - int op; -} verbitem; - -static const char verbnames[] = - "ACCEPT\0" - "COMMIT\0" - "F\0" - "FAIL\0" - "PRUNE\0" - "SKIP\0" - "THEN"; - -static const verbitem verbs[] = { - { 6, OP_ACCEPT }, - { 6, OP_COMMIT }, - { 1, OP_FAIL }, - { 4, OP_FAIL }, - { 5, OP_PRUNE }, - { 4, OP_SKIP }, - { 4, OP_THEN } -}; - -static const int verbcount = sizeof(verbs)/sizeof(verbitem); - - -/* Tables of names of POSIX character classes and their lengths. The names are -now all in a single string, to reduce the number of relocations when a shared -library is dynamically loaded. The list of lengths is terminated by a zero -length entry. The first three must be alpha, lower, upper, as this is assumed -for handling case independence. */ - -static const char posix_names[] = - "alpha\0" "lower\0" "upper\0" "alnum\0" "ascii\0" "blank\0" - "cntrl\0" "digit\0" "graph\0" "print\0" "punct\0" "space\0" - "word\0" "xdigit"; - -static const uschar posix_name_lengths[] = { - 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 6, 0 }; - -/* Table of class bit maps for each POSIX class. Each class is formed from a -base map, with an optional addition or removal of another map. Then, for some -classes, there is some additional tweaking: for [:blank:] the vertical space -characters are removed, and for [:alpha:] and [:alnum:] the underscore -character is removed. The triples in the table consist of the base map offset, -second map offset or -1 if no second map, and a non-negative value for map -addition or a negative value for map subtraction (if there are two maps). The -absolute value of the third field has these meanings: 0 => no tweaking, 1 => -remove vertical space characters, 2 => remove underscore. */ - -static const int posix_class_maps[] = { - cbit_word, cbit_digit, -2, /* alpha */ - cbit_lower, -1, 0, /* lower */ - cbit_upper, -1, 0, /* upper */ - cbit_word, -1, 2, /* alnum - word without underscore */ - cbit_print, cbit_cntrl, 0, /* ascii */ - cbit_space, -1, 1, /* blank - a GNU extension */ - cbit_cntrl, -1, 0, /* cntrl */ - cbit_digit, -1, 0, /* digit */ - cbit_graph, -1, 0, /* graph */ - cbit_print, -1, 0, /* print */ - cbit_punct, -1, 0, /* punct */ - cbit_space, -1, 0, /* space */ - cbit_word, -1, 0, /* word - a Perl extension */ - cbit_xdigit,-1, 0 /* xdigit */ -}; - - -#define STRING(a) # a -#define XSTRING(s) STRING(s) - -/* The texts of compile-time error messages. These are "char *" because they -are passed to the outside world. Do not ever re-use any error number, because -they are documented. Always add a new error instead. Messages marked DEAD below -are no longer used. This used to be a table of strings, but in order to reduce -the number of relocations needed when a shared library is loaded dynamically, -it is now one long string. We cannot use a table of offsets, because the -lengths of inserts such as XSTRING(MAX_NAME_SIZE) are not known. Instead, we -simply count through to the one we want - this isn't a performance issue -because these strings are used only when there is a compilation error. */ - -static const char error_texts[] = - "no error\0" - "\\ at end of pattern\0" - "\\c at end of pattern\0" - "unrecognized character follows \\\0" - "numbers out of order in {} quantifier\0" - /* 5 */ - "number too big in {} quantifier\0" - "missing terminating ] for character class\0" - "invalid escape sequence in character class\0" - "range out of order in character class\0" - "nothing to repeat\0" - /* 10 */ - "operand of unlimited repeat could match the empty string\0" /** DEAD **/ - "internal error: unexpected repeat\0" - "unrecognized character after (? or (?-\0" - "POSIX named classes are supported only within a class\0" - "missing )\0" - /* 15 */ - "reference to non-existent subpattern\0" - "erroffset passed as NULL\0" - "unknown option bit(s) set\0" - "missing ) after comment\0" - "parentheses nested too deeply\0" /** DEAD **/ - /* 20 */ - "regular expression is too large\0" - "failed to get memory\0" - "unmatched parentheses\0" - "internal error: code overflow\0" - "unrecognized character after (?<\0" - /* 25 */ - "lookbehind assertion is not fixed length\0" - "malformed number or name after (?(\0" - "conditional group contains more than two branches\0" - "assertion expected after (?(\0" - "(?R or (?[+-]digits must be followed by )\0" - /* 30 */ - "unknown POSIX class name\0" - "POSIX collating elements are not supported\0" - "this version of PCRE is not compiled with PCRE_UTF8 support\0" - "spare error\0" /** DEAD **/ - "character value in \\x{...} sequence is too large\0" - /* 35 */ - "invalid condition (?(0)\0" - "\\C not allowed in lookbehind assertion\0" - "PCRE does not support \\L, \\l, \\N, \\U, or \\u\0" - "number after (?C is > 255\0" - "closing ) for (?C expected\0" - /* 40 */ - "recursive call could loop indefinitely\0" - "unrecognized character after (?P\0" - "syntax error in subpattern name (missing terminator)\0" - "two named subpatterns have the same name\0" - "invalid UTF-8 string\0" - /* 45 */ - "support for \\P, \\p, and \\X has not been compiled\0" - "malformed \\P or \\p sequence\0" - "unknown property name after \\P or \\p\0" - "subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)\0" - "too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")\0" - /* 50 */ - "repeated subpattern is too long\0" /** DEAD **/ - "octal value is greater than \\377 (not in UTF-8 mode)\0" - "internal error: overran compiling workspace\0" - "internal error: previously-checked referenced subpattern not found\0" - "DEFINE group contains more than one branch\0" - /* 55 */ - "repeating a DEFINE group is not allowed\0" - "inconsistent NEWLINE options\0" - "\\g is not followed by a braced, angle-bracketed, or quoted name/number or by a plain number\0" - "a numbered reference must not be zero\0" - "(*VERB) with an argument is not supported\0" - /* 60 */ - "(*VERB) not recognized\0" - "number is too big\0" - "subpattern name expected\0" - "digit expected after (?+\0" - "] is an invalid data character in JavaScript compatibility mode"; - - -/* Table to identify digits and hex digits. This is used when compiling -patterns. Note that the tables in chartables are dependent on the locale, and -may mark arbitrary characters as digits - but the PCRE compiling code expects -to handle only 0-9, a-z, and A-Z as digits when compiling. That is why we have -a private table here. It costs 256 bytes, but it is a lot faster than doing -character value tests (at least in some simple cases I timed), and in some -applications one wants PCRE to compile efficiently as well as match -efficiently. - -For convenience, we use the same bit definitions as in chartables: - - 0x04 decimal digit - 0x08 hexadecimal digit - -Then we can use ctype_digit and ctype_xdigit in the code. */ - -#ifndef EBCDIC /* This is the "normal" case, for ASCII systems */ -static const unsigned char digitab[] = - { - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 8- 15 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - ' */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ( - / */ - 0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /* 0 - 7 */ - 0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00, /* 8 - ? */ - 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* @ - G */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* H - O */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* P - W */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* X - _ */ - 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* ` - g */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* h - o */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* p - w */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* x -127 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */ - -#else /* This is the "abnormal" case, for EBCDIC systems */ -static const unsigned char digitab[] = - { - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 0 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 8- 15 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 10 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 32- 39 20 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 40- 47 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 48- 55 30 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 56- 63 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - 71 40 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 72- | */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* & - 87 50 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 88- 95 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - -103 60 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 104- ? */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 112-119 70 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 120- " */ - 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* 128- g 80 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* h -143 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144- p 90 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* q -159 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160- x A0 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* y -175 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ^ -183 B0 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ - 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* { - G C0 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* H -207 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* } - P D0 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* Q -223 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* \ - X E0 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* Y -239 */ - 0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /* 0 - 7 F0 */ - 0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00};/* 8 -255 */ - -static const unsigned char ebcdic_chartab[] = { /* chartable partial dup */ - 0x80,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 0- 7 */ - 0x00,0x00,0x00,0x00,0x01,0x01,0x00,0x00, /* 8- 15 */ - 0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 16- 23 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ - 0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 32- 39 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 40- 47 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 48- 55 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 56- 63 */ - 0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - 71 */ - 0x00,0x00,0x00,0x80,0x00,0x80,0x80,0x80, /* 72- | */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* & - 87 */ - 0x00,0x00,0x00,0x80,0x80,0x80,0x00,0x00, /* 88- 95 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - -103 */ - 0x00,0x00,0x00,0x00,0x00,0x10,0x00,0x80, /* 104- ? */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 112-119 */ - 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 120- " */ - 0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* 128- g */ - 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* h -143 */ - 0x00,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* 144- p */ - 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* q -159 */ - 0x00,0x00,0x12,0x12,0x12,0x12,0x12,0x12, /* 160- x */ - 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* y -175 */ - 0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ^ -183 */ - 0x00,0x00,0x80,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ - 0x80,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* { - G */ - 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* H -207 */ - 0x00,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* } - P */ - 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* Q -223 */ - 0x00,0x00,0x12,0x12,0x12,0x12,0x12,0x12, /* \ - X */ - 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* Y -239 */ - 0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /* 0 - 7 */ - 0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x00};/* 8 -255 */ -#endif - - -/* Definition to allow mutual recursion */ - -static BOOL - compile_regex(int, int, uschar **, const uschar **, int *, BOOL, BOOL, int, - int *, int *, branch_chain *, compile_data *, int *); - - - -/************************************************* -* Find an error text * -*************************************************/ - -/* The error texts are now all in one long string, to save on relocations. As -some of the text is of unknown length, we can't use a table of offsets. -Instead, just count through the strings. This is not a performance issue -because it happens only when there has been a compilation error. - -Argument: the error number -Returns: pointer to the error string -*/ - -static const char * -find_error_text(int n) -{ -const char *s = error_texts; -for (; n > 0; n--) while (*s++ != 0); -return s; -} - - -/************************************************* -* Handle escapes * -*************************************************/ - -/* This function is called when a \ has been encountered. It either returns a -positive value for a simple escape such as \n, or a negative value which -encodes one of the more complicated things such as \d. A backreference to group -n is returned as -(ESC_REF + n); ESC_REF is the highest ESC_xxx macro. When -UTF-8 is enabled, a positive value greater than 255 may be returned. On entry, -ptr is pointing at the \. On exit, it is on the final character of the escape -sequence. - -Arguments: - ptrptr points to the pattern position pointer - errorcodeptr points to the errorcode variable - bracount number of previous extracting brackets - options the options bits - isclass TRUE if inside a character class - -Returns: zero or positive => a data character - negative => a special escape sequence - on error, errorcodeptr is set -*/ - -static int -check_escape(const uschar **ptrptr, int *errorcodeptr, int bracount, - int options, BOOL isclass) -{ -BOOL utf8 = (options & PCRE_UTF8) != 0; -const uschar *ptr = *ptrptr + 1; -int c, i; - -GETCHARINCTEST(c, ptr); /* Get character value, increment pointer */ -ptr--; /* Set pointer back to the last byte */ - -/* If backslash is at the end of the pattern, it's an error. */ - -if (c == 0) *errorcodeptr = ERR1; - -/* Non-alphanumerics are literals. For digits or letters, do an initial lookup -in a table. A non-zero result is something that can be returned immediately. -Otherwise further processing may be required. */ - -#ifndef EBCDIC /* ASCII coding */ -else if (c < '0' || c > 'z') {} /* Not alphanumeric */ -else if ((i = escapes[c - '0']) != 0) c = i; - -#else /* EBCDIC coding */ -else if (c < 'a' || (ebcdic_chartab[c] & 0x0E) == 0) {} /* Not alphanumeric */ -else if ((i = escapes[c - 0x48]) != 0) c = i; -#endif - -/* Escapes that need further processing, or are illegal. */ - -else - { - const uschar *oldptr; - BOOL braced, negated; - - switch (c) - { - /* A number of Perl escapes are not handled by PCRE. We give an explicit - error. */ - - case 'l': - case 'L': - case 'N': - case 'u': - case 'U': - *errorcodeptr = ERR37; - break; - - /* \g must be followed by one of a number of specific things: - - (1) A number, either plain or braced. If positive, it is an absolute - backreference. If negative, it is a relative backreference. This is a Perl - 5.10 feature. - - (2) Perl 5.10 also supports \g{name} as a reference to a named group. This - is part of Perl's movement towards a unified syntax for back references. As - this is synonymous with \k{name}, we fudge it up by pretending it really - was \k. - - (3) For Oniguruma compatibility we also support \g followed by a name or a - number either in angle brackets or in single quotes. However, these are - (possibly recursive) subroutine calls, _not_ backreferences. Just return - the -ESC_g code (cf \k). */ - - case 'g': - if (ptr[1] == '<' || ptr[1] == '\'') - { - c = -ESC_g; - break; - } - - /* Handle the Perl-compatible cases */ - - if (ptr[1] == '{') - { - const uschar *p; - for (p = ptr+2; *p != 0 && *p != '}'; p++) - if (*p != '-' && (digitab[*p] & ctype_digit) == 0) break; - if (*p != 0 && *p != '}') - { - c = -ESC_k; - break; - } - braced = TRUE; - ptr++; - } - else braced = FALSE; - - if (ptr[1] == '-') - { - negated = TRUE; - ptr++; - } - else negated = FALSE; - - c = 0; - while ((digitab[ptr[1]] & ctype_digit) != 0) - c = c * 10 + *(++ptr) - '0'; - - if (c < 0) /* Integer overflow */ - { - *errorcodeptr = ERR61; - break; - } - - if (braced && *(++ptr) != '}') - { - *errorcodeptr = ERR57; - break; - } - - if (c == 0) - { - *errorcodeptr = ERR58; - break; - } - - if (negated) - { - if (c > bracount) - { - *errorcodeptr = ERR15; - break; - } - c = bracount - (c - 1); - } - - c = -(ESC_REF + c); - break; - - /* The handling of escape sequences consisting of a string of digits - starting with one that is not zero is not straightforward. By experiment, - the way Perl works seems to be as follows: - - Outside a character class, the digits are read as a decimal number. If the - number is less than 10, or if there are that many previous extracting - left brackets, then it is a back reference. Otherwise, up to three octal - digits are read to form an escaped byte. Thus \123 is likely to be octal - 123 (cf \0123, which is octal 012 followed by the literal 3). If the octal - value is greater than 377, the least significant 8 bits are taken. Inside a - character class, \ followed by a digit is always an octal number. */ - - case '1': case '2': case '3': case '4': case '5': - case '6': case '7': case '8': case '9': - - if (!isclass) - { - oldptr = ptr; - c -= '0'; - while ((digitab[ptr[1]] & ctype_digit) != 0) - c = c * 10 + *(++ptr) - '0'; - if (c < 0) /* Integer overflow */ - { - *errorcodeptr = ERR61; - break; - } - if (c < 10 || c <= bracount) - { - c = -(ESC_REF + c); - break; - } - ptr = oldptr; /* Put the pointer back and fall through */ - } - - /* Handle an octal number following \. If the first digit is 8 or 9, Perl - generates a binary zero byte and treats the digit as a following literal. - Thus we have to pull back the pointer by one. */ - - if ((c = *ptr) >= '8') - { - ptr--; - c = 0; - break; - } - - /* \0 always starts an octal number, but we may drop through to here with a - larger first octal digit. The original code used just to take the least - significant 8 bits of octal numbers (I think this is what early Perls used - to do). Nowadays we allow for larger numbers in UTF-8 mode, but no more - than 3 octal digits. */ - - case '0': - c -= '0'; - while(i++ < 2 && ptr[1] >= '0' && ptr[1] <= '7') - c = c * 8 + *(++ptr) - '0'; - if (!utf8 && c > 255) *errorcodeptr = ERR51; - break; - - /* \x is complicated. \x{ddd} is a character number which can be greater - than 0xff in utf8 mode, but only if the ddd are hex digits. If not, { is - treated as a data character. */ - - case 'x': - if (ptr[1] == '{') - { - const uschar *pt = ptr + 2; - int count = 0; - - c = 0; - while ((digitab[*pt] & ctype_xdigit) != 0) - { - register int cc = *pt++; - if (c == 0 && cc == '0') continue; /* Leading zeroes */ - count++; - -#ifndef EBCDIC /* ASCII coding */ - if (cc >= 'a') cc -= 32; /* Convert to upper case */ - c = (c << 4) + cc - ((cc < 'A')? '0' : ('A' - 10)); -#else /* EBCDIC coding */ - if (cc >= 'a' && cc <= 'z') cc += 64; /* Convert to upper case */ - c = (c << 4) + cc - ((cc >= '0')? '0' : ('A' - 10)); -#endif - } - - if (*pt == '}') - { - if (c < 0 || count > (utf8? 8 : 2)) *errorcodeptr = ERR34; - ptr = pt; - break; - } - - /* If the sequence of hex digits does not end with '}', then we don't - recognize this construct; fall through to the normal \x handling. */ - } - - /* Read just a single-byte hex-defined char */ - - c = 0; - while (i++ < 2 && (digitab[ptr[1]] & ctype_xdigit) != 0) - { - int cc; /* Some compilers don't like ++ */ - cc = *(++ptr); /* in initializers */ -#ifndef EBCDIC /* ASCII coding */ - if (cc >= 'a') cc -= 32; /* Convert to upper case */ - c = c * 16 + cc - ((cc < 'A')? '0' : ('A' - 10)); -#else /* EBCDIC coding */ - if (cc <= 'z') cc += 64; /* Convert to upper case */ - c = c * 16 + cc - ((cc >= '0')? '0' : ('A' - 10)); -#endif - } - break; - - /* For \c, a following letter is upper-cased; then the 0x40 bit is flipped. - This coding is ASCII-specific, but then the whole concept of \cx is - ASCII-specific. (However, an EBCDIC equivalent has now been added.) */ - - case 'c': - c = *(++ptr); - if (c == 0) - { - *errorcodeptr = ERR2; - break; - } - -#ifndef EBCDIC /* ASCII coding */ - if (c >= 'a' && c <= 'z') c -= 32; - c ^= 0x40; -#else /* EBCDIC coding */ - if (c >= 'a' && c <= 'z') c += 64; - c ^= 0xC0; -#endif - break; - - /* PCRE_EXTRA enables extensions to Perl in the matter of escapes. Any - other alphanumeric following \ is an error if PCRE_EXTRA was set; - otherwise, for Perl compatibility, it is a literal. This code looks a bit - odd, but there used to be some cases other than the default, and there may - be again in future, so I haven't "optimized" it. */ - - default: - if ((options & PCRE_EXTRA) != 0) switch(c) - { - default: - *errorcodeptr = ERR3; - break; - } - break; - } - } - -*ptrptr = ptr; -return c; -} - - - -#ifdef SUPPORT_UCP -/************************************************* -* Handle \P and \p * -*************************************************/ - -/* This function is called after \P or \p has been encountered, provided that -PCRE is compiled with support for Unicode properties. On entry, ptrptr is -pointing at the P or p. On exit, it is pointing at the final character of the -escape sequence. - -Argument: - ptrptr points to the pattern position pointer - negptr points to a boolean that is set TRUE for negation else FALSE - dptr points to an int that is set to the detailed property value - errorcodeptr points to the error code variable - -Returns: type value from ucp_type_table, or -1 for an invalid type -*/ - -static int -get_ucp(const uschar **ptrptr, BOOL *negptr, int *dptr, int *errorcodeptr) -{ -int c, i, bot, top; -const uschar *ptr = *ptrptr; -char name[32]; - -c = *(++ptr); -if (c == 0) goto ERROR_RETURN; - -*negptr = FALSE; - -/* \P or \p can be followed by a name in {}, optionally preceded by ^ for -negation. */ - -if (c == '{') - { - if (ptr[1] == '^') - { - *negptr = TRUE; - ptr++; - } - for (i = 0; i < (int)sizeof(name) - 1; i++) - { - c = *(++ptr); - if (c == 0) goto ERROR_RETURN; - if (c == '}') break; - name[i] = c; - } - if (c !='}') goto ERROR_RETURN; - name[i] = 0; - } - -/* Otherwise there is just one following character */ - -else - { - name[0] = c; - name[1] = 0; - } - -*ptrptr = ptr; - -/* Search for a recognized property name using binary chop */ - -bot = 0; -top = _pcre_utt_size; - -while (bot < top) - { - i = (bot + top) >> 1; - c = strcmp(name, _pcre_utt_names + _pcre_utt[i].name_offset); - if (c == 0) - { - *dptr = _pcre_utt[i].value; - return _pcre_utt[i].type; - } - if (c > 0) bot = i + 1; else top = i; - } - -*errorcodeptr = ERR47; -*ptrptr = ptr; -return -1; - -ERROR_RETURN: -*errorcodeptr = ERR46; -*ptrptr = ptr; -return -1; -} -#endif - - - - -/************************************************* -* Check for counted repeat * -*************************************************/ - -/* This function is called when a '{' is encountered in a place where it might -start a quantifier. It looks ahead to see if it really is a quantifier or not. -It is only a quantifier if it is one of the forms {ddd} {ddd,} or {ddd,ddd} -where the ddds are digits. - -Arguments: - p pointer to the first char after '{' - -Returns: TRUE or FALSE -*/ - -static BOOL -is_counted_repeat(const uschar *p) -{ -if ((digitab[*p++] & ctype_digit) == 0) return FALSE; -while ((digitab[*p] & ctype_digit) != 0) p++; -if (*p == '}') return TRUE; - -if (*p++ != ',') return FALSE; -if (*p == '}') return TRUE; - -if ((digitab[*p++] & ctype_digit) == 0) return FALSE; -while ((digitab[*p] & ctype_digit) != 0) p++; - -return (*p == '}'); -} - - - -/************************************************* -* Read repeat counts * -*************************************************/ - -/* Read an item of the form {n,m} and return the values. This is called only -after is_counted_repeat() has confirmed that a repeat-count quantifier exists, -so the syntax is guaranteed to be correct, but we need to check the values. - -Arguments: - p pointer to first char after '{' - minp pointer to int for min - maxp pointer to int for max - returned as -1 if no max - errorcodeptr points to error code variable - -Returns: pointer to '}' on success; - current ptr on error, with errorcodeptr set non-zero -*/ - -static const uschar * -read_repeat_counts(const uschar *p, int *minp, int *maxp, int *errorcodeptr) -{ -int min = 0; -int max = -1; - -/* Read the minimum value and do a paranoid check: a negative value indicates -an integer overflow. */ - -while ((digitab[*p] & ctype_digit) != 0) min = min * 10 + *p++ - '0'; -if (min < 0 || min > 65535) - { - *errorcodeptr = ERR5; - return p; - } - -/* Read the maximum value if there is one, and again do a paranoid on its size. -Also, max must not be less than min. */ - -if (*p == '}') max = min; else - { - if (*(++p) != '}') - { - max = 0; - while((digitab[*p] & ctype_digit) != 0) max = max * 10 + *p++ - '0'; - if (max < 0 || max > 65535) - { - *errorcodeptr = ERR5; - return p; - } - if (max < min) - { - *errorcodeptr = ERR4; - return p; - } - } - } - -/* Fill in the required variables, and pass back the pointer to the terminating -'}'. */ - -*minp = min; -*maxp = max; -return p; -} - - - -/************************************************* -* Find forward referenced subpattern * -*************************************************/ - -/* This function scans along a pattern's text looking for capturing -subpatterns, and counting them. If it finds a named pattern that matches the -name it is given, it returns its number. Alternatively, if the name is NULL, it -returns when it reaches a given numbered subpattern. This is used for forward -references to subpatterns. We know that if (?P< is encountered, the name will -be terminated by '>' because that is checked in the first pass. - -Arguments: - ptr current position in the pattern - cd compile background data - name name to seek, or NULL if seeking a numbered subpattern - lorn name length, or subpattern number if name is NULL - xmode TRUE if we are in /x mode - -Returns: the number of the named subpattern, or -1 if not found -*/ - -static int -find_parens(const uschar *ptr, compile_data *cd, const uschar *name, int lorn, - BOOL xmode) -{ -const uschar *thisname; -int count = cd->bracount; - -for (; *ptr != 0; ptr++) - { - int term; - - /* Skip over backslashed characters and also entire \Q...\E */ - - if (*ptr == '\\') - { - if (*(++ptr) == 0) return -1; - if (*ptr == 'Q') for (;;) - { - while (*(++ptr) != 0 && *ptr != '\\'); - if (*ptr == 0) return -1; - if (*(++ptr) == 'E') break; - } - continue; - } - - /* Skip over character classes; this logic must be similar to the way they - are handled for real. If the first character is '^', skip it. Also, if the - first few characters (either before or after ^) are \Q\E or \E we skip them - too. This makes for compatibility with Perl. */ - - if (*ptr == '[') - { - BOOL negate_class = FALSE; - for (;;) - { - int c = *(++ptr); - if (c == '\\') - { - if (ptr[1] == 'E') ptr++; - else if (strncmp((const char *)ptr+1, "Q\\E", 3) == 0) ptr += 3; - else break; - } - else if (!negate_class && c == '^') - negate_class = TRUE; - else break; - } - - /* If the next character is ']', it is a data character that must be - skipped, except in JavaScript compatibility mode. */ - - if (ptr[1] == ']' && (cd->external_options & PCRE_JAVASCRIPT_COMPAT) == 0) - ptr++; - - while (*(++ptr) != ']') - { - if (*ptr == 0) return -1; - if (*ptr == '\\') - { - if (*(++ptr) == 0) return -1; - if (*ptr == 'Q') for (;;) - { - while (*(++ptr) != 0 && *ptr != '\\'); - if (*ptr == 0) return -1; - if (*(++ptr) == 'E') break; - } - continue; - } - } - continue; - } - - /* Skip comments in /x mode */ - - if (xmode && *ptr == '#') - { - while (*(++ptr) != 0 && *ptr != '\n'); - if (*ptr == 0) return -1; - continue; - } - - /* An opening parens must now be a real metacharacter */ - - if (*ptr != '(') continue; - if (ptr[1] != '?' && ptr[1] != '*') - { - count++; - if (name == NULL && count == lorn) return count; - continue; - } - - ptr += 2; - if (*ptr == 'P') ptr++; /* Allow optional P */ - - /* We have to disambiguate (? */ - - if ((*ptr != '<' || ptr[1] == '!' || ptr[1] == '=') && - *ptr != '\'') - continue; - - count++; - - if (name == NULL && count == lorn) return count; - term = *ptr++; - if (term == '<') term = '>'; - thisname = ptr; - while (*ptr != term) ptr++; - if (name != NULL && lorn == ptr - thisname && - strncmp((const char *)name, (const char *)thisname, lorn) == 0) - return count; - } - -return -1; -} - - - -/************************************************* -* Find first significant op code * -*************************************************/ - -/* This is called by several functions that scan a compiled expression looking -for a fixed first character, or an anchoring op code etc. It skips over things -that do not influence this. For some calls, a change of option is important. -For some calls, it makes sense to skip negative forward and all backward -assertions, and also the \b assertion; for others it does not. - -Arguments: - code pointer to the start of the group - options pointer to external options - optbit the option bit whose changing is significant, or - zero if none are - skipassert TRUE if certain assertions are to be skipped - -Returns: pointer to the first significant opcode -*/ - -static const uschar* -first_significant_code(const uschar *code, int *options, int optbit, - BOOL skipassert) -{ -for (;;) - { - switch ((int)*code) - { - case OP_OPT: - if (optbit > 0 && ((int)code[1] & optbit) != (*options & optbit)) - *options = (int)code[1]; - code += 2; - break; - - case OP_ASSERT_NOT: - case OP_ASSERTBACK: - case OP_ASSERTBACK_NOT: - if (!skipassert) return code; - do code += GET(code, 1); while (*code == OP_ALT); - code += _pcre_OP_lengths[*code]; - break; - - case OP_WORD_BOUNDARY: - case OP_NOT_WORD_BOUNDARY: - if (!skipassert) return code; - /* Fall through */ - - case OP_CALLOUT: - case OP_CREF: - case OP_RREF: - case OP_DEF: - code += _pcre_OP_lengths[*code]; - break; - - default: - return code; - } - } -/* Control never reaches here */ -} - - - - -/************************************************* -* Find the fixed length of a pattern * -*************************************************/ - -/* Scan a pattern and compute the fixed length of subject that will match it, -if the length is fixed. This is needed for dealing with backward assertions. -In UTF8 mode, the result is in characters rather than bytes. - -Arguments: - code points to the start of the pattern (the bracket) - options the compiling options - -Returns: the fixed length, or -1 if there is no fixed length, - or -2 if \C was encountered -*/ - -static int -find_fixedlength(uschar *code, int options) -{ -int length = -1; - -register int branchlength = 0; -register uschar *cc = code + 1 + LINK_SIZE; - -/* Scan along the opcodes for this branch. If we get to the end of the -branch, check the length against that of the other branches. */ - -for (;;) - { - int d; - register int op = *cc; - switch (op) - { - case OP_CBRA: - case OP_BRA: - case OP_ONCE: - case OP_COND: - d = find_fixedlength(cc + ((op == OP_CBRA)? 2:0), options); - if (d < 0) return d; - branchlength += d; - do cc += GET(cc, 1); while (*cc == OP_ALT); - cc += 1 + LINK_SIZE; - break; - - /* Reached end of a branch; if it's a ket it is the end of a nested - call. If it's ALT it is an alternation in a nested call. If it is - END it's the end of the outer call. All can be handled by the same code. */ - - case OP_ALT: - case OP_KET: - case OP_KETRMAX: - case OP_KETRMIN: - case OP_END: - if (length < 0) length = branchlength; - else if (length != branchlength) return -1; - if (*cc != OP_ALT) return length; - cc += 1 + LINK_SIZE; - branchlength = 0; - break; - - /* Skip over assertive subpatterns */ - - case OP_ASSERT: - case OP_ASSERT_NOT: - case OP_ASSERTBACK: - case OP_ASSERTBACK_NOT: - do cc += GET(cc, 1); while (*cc == OP_ALT); - /* Fall through */ - - /* Skip over things that don't match chars */ - - case OP_REVERSE: - case OP_CREF: - case OP_RREF: - case OP_DEF: - case OP_OPT: - case OP_CALLOUT: - case OP_SOD: - case OP_SOM: - case OP_EOD: - case OP_EODN: - case OP_CIRC: - case OP_DOLL: - case OP_NOT_WORD_BOUNDARY: - case OP_WORD_BOUNDARY: - cc += _pcre_OP_lengths[*cc]; - break; - - /* Handle literal characters */ - - case OP_CHAR: - case OP_CHARNC: - case OP_NOT: - branchlength++; - cc += 2; -#ifdef SUPPORT_UTF8 - if ((options & PCRE_UTF8) != 0) - { - while ((*cc & 0xc0) == 0x80) cc++; - } -#endif - break; - - /* Handle exact repetitions. The count is already in characters, but we - need to skip over a multibyte character in UTF8 mode. */ - - case OP_EXACT: - branchlength += GET2(cc,1); - cc += 4; -#ifdef SUPPORT_UTF8 - if ((options & PCRE_UTF8) != 0) - { - while((*cc & 0x80) == 0x80) cc++; - } -#endif - break; - - case OP_TYPEEXACT: - branchlength += GET2(cc,1); - if (cc[3] == OP_PROP || cc[3] == OP_NOTPROP) cc += 2; - cc += 4; - break; - - /* Handle single-char matchers */ - - case OP_PROP: - case OP_NOTPROP: - cc += 2; - /* Fall through */ - - case OP_NOT_DIGIT: - case OP_DIGIT: - case OP_NOT_WHITESPACE: - case OP_WHITESPACE: - case OP_NOT_WORDCHAR: - case OP_WORDCHAR: - case OP_ANY: - case OP_ALLANY: - branchlength++; - cc++; - break; - - /* The single-byte matcher isn't allowed */ - - case OP_ANYBYTE: - return -2; - - /* Check a class for variable quantification */ - -#ifdef SUPPORT_UTF8 - case OP_XCLASS: - cc += GET(cc, 1) - 33; - /* Fall through */ -#endif - - case OP_CLASS: - case OP_NCLASS: - cc += 33; - - switch (*cc) - { - case OP_CRSTAR: - case OP_CRMINSTAR: - case OP_CRQUERY: - case OP_CRMINQUERY: - return -1; - - case OP_CRRANGE: - case OP_CRMINRANGE: - if (GET2(cc,1) != GET2(cc,3)) return -1; - branchlength += GET2(cc,1); - cc += 5; - break; - - default: - branchlength++; - } - break; - - /* Anything else is variable length */ - - default: - return -1; - } - } -/* Control never gets here */ -} - - - - -/************************************************* -* Scan compiled regex for numbered bracket * -*************************************************/ - -/* This little function scans through a compiled pattern until it finds a -capturing bracket with the given number. - -Arguments: - code points to start of expression - utf8 TRUE in UTF-8 mode - number the required bracket number - -Returns: pointer to the opcode for the bracket, or NULL if not found -*/ - -static const uschar * -find_bracket(const uschar *code, BOOL utf8, int number) -{ -for (;;) - { - register int c = *code; - if (c == OP_END) return NULL; - - /* XCLASS is used for classes that cannot be represented just by a bit - map. This includes negated single high-valued characters. The length in - the table is zero; the actual length is stored in the compiled code. */ - - if (c == OP_XCLASS) code += GET(code, 1); - - /* Handle capturing bracket */ - - else if (c == OP_CBRA) - { - int n = GET2(code, 1+LINK_SIZE); - if (n == number) return (uschar *)code; - code += _pcre_OP_lengths[c]; - } - - /* Otherwise, we can get the item's length from the table, except that for - repeated character types, we have to test for \p and \P, which have an extra - two bytes of parameters. */ - - else - { - switch(c) - { - case OP_TYPESTAR: - case OP_TYPEMINSTAR: - case OP_TYPEPLUS: - case OP_TYPEMINPLUS: - case OP_TYPEQUERY: - case OP_TYPEMINQUERY: - case OP_TYPEPOSSTAR: - case OP_TYPEPOSPLUS: - case OP_TYPEPOSQUERY: - if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2; - break; - - case OP_TYPEUPTO: - case OP_TYPEMINUPTO: - case OP_TYPEEXACT: - case OP_TYPEPOSUPTO: - if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2; - break; - } - - /* Add in the fixed length from the table */ - - code += _pcre_OP_lengths[c]; - - /* In UTF-8 mode, opcodes that are followed by a character may be followed by - a multi-byte character. The length in the table is a minimum, so we have to - arrange to skip the extra bytes. */ - -#ifdef SUPPORT_UTF8 - if (utf8) switch(c) - { - case OP_CHAR: - case OP_CHARNC: - case OP_EXACT: - case OP_UPTO: - case OP_MINUPTO: - case OP_POSUPTO: - case OP_STAR: - case OP_MINSTAR: - case OP_POSSTAR: - case OP_PLUS: - case OP_MINPLUS: - case OP_POSPLUS: - case OP_QUERY: - case OP_MINQUERY: - case OP_POSQUERY: - if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f]; - break; - } -#endif - } - } -} - - - -/************************************************* -* Scan compiled regex for recursion reference * -*************************************************/ - -/* This little function scans through a compiled pattern until it finds an -instance of OP_RECURSE. - -Arguments: - code points to start of expression - utf8 TRUE in UTF-8 mode - -Returns: pointer to the opcode for OP_RECURSE, or NULL if not found -*/ - -static const uschar * -find_recurse(const uschar *code, BOOL utf8) -{ -for (;;) - { - register int c = *code; - if (c == OP_END) return NULL; - if (c == OP_RECURSE) return code; - - /* XCLASS is used for classes that cannot be represented just by a bit - map. This includes negated single high-valued characters. The length in - the table is zero; the actual length is stored in the compiled code. */ - - if (c == OP_XCLASS) code += GET(code, 1); - - /* Otherwise, we can get the item's length from the table, except that for - repeated character types, we have to test for \p and \P, which have an extra - two bytes of parameters. */ - - else - { - switch(c) - { - case OP_TYPESTAR: - case OP_TYPEMINSTAR: - case OP_TYPEPLUS: - case OP_TYPEMINPLUS: - case OP_TYPEQUERY: - case OP_TYPEMINQUERY: - case OP_TYPEPOSSTAR: - case OP_TYPEPOSPLUS: - case OP_TYPEPOSQUERY: - if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2; - break; - - case OP_TYPEPOSUPTO: - case OP_TYPEUPTO: - case OP_TYPEMINUPTO: - case OP_TYPEEXACT: - if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2; - break; - } - - /* Add in the fixed length from the table */ - - code += _pcre_OP_lengths[c]; - - /* In UTF-8 mode, opcodes that are followed by a character may be followed - by a multi-byte character. The length in the table is a minimum, so we have - to arrange to skip the extra bytes. */ - -#ifdef SUPPORT_UTF8 - if (utf8) switch(c) - { - case OP_CHAR: - case OP_CHARNC: - case OP_EXACT: - case OP_UPTO: - case OP_MINUPTO: - case OP_POSUPTO: - case OP_STAR: - case OP_MINSTAR: - case OP_POSSTAR: - case OP_PLUS: - case OP_MINPLUS: - case OP_POSPLUS: - case OP_QUERY: - case OP_MINQUERY: - case OP_POSQUERY: - if (code[-1] >= 0xc0) code += _pcre_utf8_table4[code[-1] & 0x3f]; - break; - } -#endif - } - } -} - - - -/************************************************* -* Scan compiled branch for non-emptiness * -*************************************************/ - -/* This function scans through a branch of a compiled pattern to see whether it -can match the empty string or not. It is called from could_be_empty() -below and from compile_branch() when checking for an unlimited repeat of a -group that can match nothing. Note that first_significant_code() skips over -backward and negative forward assertions when its final argument is TRUE. If we -hit an unclosed bracket, we return "empty" - this means we've struck an inner -bracket whose current branch will already have been scanned. - -Arguments: - code points to start of search - endcode points to where to stop - utf8 TRUE if in UTF8 mode - -Returns: TRUE if what is matched could be empty -*/ - -static BOOL -could_be_empty_branch(const uschar *code, const uschar *endcode, BOOL utf8) -{ -register int c; -for (code = first_significant_code(code + _pcre_OP_lengths[*code], NULL, 0, TRUE); - code < endcode; - code = first_significant_code(code + _pcre_OP_lengths[c], NULL, 0, TRUE)) - { - const uschar *ccode; - - c = *code; - - /* Skip over forward assertions; the other assertions are skipped by - first_significant_code() with a TRUE final argument. */ - - if (c == OP_ASSERT) - { - do code += GET(code, 1); while (*code == OP_ALT); - c = *code; - continue; - } - - /* Groups with zero repeats can of course be empty; skip them. */ - - if (c == OP_BRAZERO || c == OP_BRAMINZERO || c == OP_SKIPZERO) - { - code += _pcre_OP_lengths[c]; - do code += GET(code, 1); while (*code == OP_ALT); - c = *code; - continue; - } - - /* For other groups, scan the branches. */ - - if (c == OP_BRA || c == OP_CBRA || c == OP_ONCE || c == OP_COND) - { - BOOL empty_branch; - if (GET(code, 1) == 0) return TRUE; /* Hit unclosed bracket */ - - /* Scan a closed bracket */ - - empty_branch = FALSE; - do - { - if (!empty_branch && could_be_empty_branch(code, endcode, utf8)) - empty_branch = TRUE; - code += GET(code, 1); - } - while (*code == OP_ALT); - if (!empty_branch) return FALSE; /* All branches are non-empty */ - c = *code; - continue; - } - - /* Handle the other opcodes */ - - switch (c) - { - /* Check for quantifiers after a class. XCLASS is used for classes that - cannot be represented just by a bit map. This includes negated single - high-valued characters. The length in _pcre_OP_lengths[] is zero; the - actual length is stored in the compiled code, so we must update "code" - here. */ - -#ifdef SUPPORT_UTF8 - case OP_XCLASS: - ccode = code += GET(code, 1); - goto CHECK_CLASS_REPEAT; -#endif - - case OP_CLASS: - case OP_NCLASS: - ccode = code + 33; - -#ifdef SUPPORT_UTF8 - CHECK_CLASS_REPEAT: -#endif - - switch (*ccode) - { - case OP_CRSTAR: /* These could be empty; continue */ - case OP_CRMINSTAR: - case OP_CRQUERY: - case OP_CRMINQUERY: - break; - - default: /* Non-repeat => class must match */ - case OP_CRPLUS: /* These repeats aren't empty */ - case OP_CRMINPLUS: - return FALSE; - - case OP_CRRANGE: - case OP_CRMINRANGE: - if (GET2(ccode, 1) > 0) return FALSE; /* Minimum > 0 */ - break; - } - break; - - /* Opcodes that must match a character */ - - case OP_PROP: - case OP_NOTPROP: - case OP_EXTUNI: - case OP_NOT_DIGIT: - case OP_DIGIT: - case OP_NOT_WHITESPACE: - case OP_WHITESPACE: - case OP_NOT_WORDCHAR: - case OP_WORDCHAR: - case OP_ANY: - case OP_ALLANY: - case OP_ANYBYTE: - case OP_CHAR: - case OP_CHARNC: - case OP_NOT: - case OP_PLUS: - case OP_MINPLUS: - case OP_POSPLUS: - case OP_EXACT: - case OP_NOTPLUS: - case OP_NOTMINPLUS: - case OP_NOTPOSPLUS: - case OP_NOTEXACT: - case OP_TYPEPLUS: - case OP_TYPEMINPLUS: - case OP_TYPEPOSPLUS: - case OP_TYPEEXACT: - return FALSE; - - /* These are going to continue, as they may be empty, but we have to - fudge the length for the \p and \P cases. */ - - case OP_TYPESTAR: - case OP_TYPEMINSTAR: - case OP_TYPEPOSSTAR: - case OP_TYPEQUERY: - case OP_TYPEMINQUERY: - case OP_TYPEPOSQUERY: - if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2; - break; - - /* Same for these */ - - case OP_TYPEUPTO: - case OP_TYPEMINUPTO: - case OP_TYPEPOSUPTO: - if (code[3] == OP_PROP || code[3] == OP_NOTPROP) code += 2; - break; - - /* End of branch */ - - case OP_KET: - case OP_KETRMAX: - case OP_KETRMIN: - case OP_ALT: - return TRUE; - - /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO, - MINUPTO, and POSUPTO may be followed by a multibyte character */ - -#ifdef SUPPORT_UTF8 - case OP_STAR: - case OP_MINSTAR: - case OP_POSSTAR: - case OP_QUERY: - case OP_MINQUERY: - case OP_POSQUERY: - case OP_UPTO: - case OP_MINUPTO: - case OP_POSUPTO: - if (utf8) while ((code[2] & 0xc0) == 0x80) code++; - break; -#endif - } - } - -return TRUE; -} - - - -/************************************************* -* Scan compiled regex for non-emptiness * -*************************************************/ - -/* This function is called to check for left recursive calls. We want to check -the current branch of the current pattern to see if it could match the empty -string. If it could, we must look outwards for branches at other levels, -stopping when we pass beyond the bracket which is the subject of the recursion. - -Arguments: - code points to start of the recursion - endcode points to where to stop (current RECURSE item) - bcptr points to the chain of current (unclosed) branch starts - utf8 TRUE if in UTF-8 mode - -Returns: TRUE if what is matched could be empty -*/ - -static BOOL -could_be_empty(const uschar *code, const uschar *endcode, branch_chain *bcptr, - BOOL utf8) -{ -while (bcptr != NULL && bcptr->current >= code) - { - if (!could_be_empty_branch(bcptr->current, endcode, utf8)) return FALSE; - bcptr = bcptr->outer; - } -return TRUE; -} - - - -/************************************************* -* Check for POSIX class syntax * -*************************************************/ - -/* This function is called when the sequence "[:" or "[." or "[=" is -encountered in a character class. It checks whether this is followed by a -sequence of characters terminated by a matching ":]" or ".]" or "=]". If we -reach an unescaped ']' without the special preceding character, return FALSE. - -Originally, this function only recognized a sequence of letters between the -terminators, but it seems that Perl recognizes any sequence of characters, -though of course unknown POSIX names are subsequently rejected. Perl gives an -"Unknown POSIX class" error for [:f\oo:] for example, where previously PCRE -didn't consider this to be a POSIX class. Likewise for [:1234:]. - -The problem in trying to be exactly like Perl is in the handling of escapes. We -have to be sure that [abc[:x\]pqr] is *not* treated as containing a POSIX -class, but [abc[:x\]pqr:]] is (so that an error can be generated). The code -below handles the special case of \], but does not try to do any other escape -processing. This makes it different from Perl for cases such as [:l\ower:] -where Perl recognizes it as the POSIX class "lower" but PCRE does not recognize -"l\ower". This is a lesser evil that not diagnosing bad classes when Perl does, -I think. - -Arguments: - ptr pointer to the initial [ - endptr where to return the end pointer - -Returns: TRUE or FALSE -*/ - -static BOOL -check_posix_syntax(const uschar *ptr, const uschar **endptr) -{ -int terminator; /* Don't combine these lines; the Solaris cc */ -terminator = *(++ptr); /* compiler warns about "non-constant" initializer. */ -for (++ptr; *ptr != 0; ptr++) - { - if (*ptr == '\\' && ptr[1] == ']') ptr++; else - { - if (*ptr == ']') return FALSE; - if (*ptr == terminator && ptr[1] == ']') - { - *endptr = ptr; - return TRUE; - } - } - } -return FALSE; -} - - - - -/************************************************* -* Check POSIX class name * -*************************************************/ - -/* This function is called to check the name given in a POSIX-style class entry -such as [:alnum:]. - -Arguments: - ptr points to the first letter - len the length of the name - -Returns: a value representing the name, or -1 if unknown -*/ - -static int -check_posix_name(const uschar *ptr, int len) -{ -const char *pn = posix_names; -register int yield = 0; -while (posix_name_lengths[yield] != 0) - { - if (len == posix_name_lengths[yield] && - strncmp((const char *)ptr, pn, len) == 0) return yield; - pn += posix_name_lengths[yield] + 1; - yield++; - } -return -1; -} - - -/************************************************* -* Adjust OP_RECURSE items in repeated group * -*************************************************/ - -/* OP_RECURSE items contain an offset from the start of the regex to the group -that is referenced. This means that groups can be replicated for fixed -repetition simply by copying (because the recursion is allowed to refer to -earlier groups that are outside the current group). However, when a group is -optional (i.e. the minimum quantifier is zero), OP_BRAZERO or OP_SKIPZERO is -inserted before it, after it has been compiled. This means that any OP_RECURSE -items within it that refer to the group itself or any contained groups have to -have their offsets adjusted. That one of the jobs of this function. Before it -is called, the partially compiled regex must be temporarily terminated with -OP_END. - -This function has been extended with the possibility of forward references for -recursions and subroutine calls. It must also check the list of such references -for the group we are dealing with. If it finds that one of the recursions in -the current group is on this list, it adjusts the offset in the list, not the -value in the reference (which is a group number). - -Arguments: - group points to the start of the group - adjust the amount by which the group is to be moved - utf8 TRUE in UTF-8 mode - cd contains pointers to tables etc. - save_hwm the hwm forward reference pointer at the start of the group - -Returns: nothing -*/ - -static void -adjust_recurse(uschar *group, int adjust, BOOL utf8, compile_data *cd, - uschar *save_hwm) -{ -uschar *ptr = group; - -while ((ptr = (uschar *)find_recurse(ptr, utf8)) != NULL) - { - int offset; - uschar *hc; - - /* See if this recursion is on the forward reference list. If so, adjust the - reference. */ - - for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE) - { - offset = GET(hc, 0); - if (cd->start_code + offset == ptr + 1) - { - PUT(hc, 0, offset + adjust); - break; - } - } - - /* Otherwise, adjust the recursion offset if it's after the start of this - group. */ - - if (hc >= cd->hwm) - { - offset = GET(ptr, 1); - if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust); - } - - ptr += 1 + LINK_SIZE; - } -} - - - -/************************************************* -* Insert an automatic callout point * -*************************************************/ - -/* This function is called when the PCRE_AUTO_CALLOUT option is set, to insert -callout points before each pattern item. - -Arguments: - code current code pointer - ptr current pattern pointer - cd pointers to tables etc - -Returns: new code pointer -*/ - -static uschar * -auto_callout(uschar *code, const uschar *ptr, compile_data *cd) -{ -*code++ = OP_CALLOUT; -*code++ = 255; -PUT(code, 0, ptr - cd->start_pattern); /* Pattern offset */ -PUT(code, LINK_SIZE, 0); /* Default length */ -return code + 2*LINK_SIZE; -} - - - -/************************************************* -* Complete a callout item * -*************************************************/ - -/* A callout item contains the length of the next item in the pattern, which -we can't fill in till after we have reached the relevant point. This is used -for both automatic and manual callouts. - -Arguments: - previous_callout points to previous callout item - ptr current pattern pointer - cd pointers to tables etc - -Returns: nothing -*/ - -static void -complete_callout(uschar *previous_callout, const uschar *ptr, compile_data *cd) -{ -int length = ptr - cd->start_pattern - GET(previous_callout, 2); -PUT(previous_callout, 2 + LINK_SIZE, length); -} - - - -#ifdef SUPPORT_UCP -/************************************************* -* Get othercase range * -*************************************************/ - -/* This function is passed the start and end of a class range, in UTF-8 mode -with UCP support. It searches up the characters, looking for internal ranges of -characters in the "other" case. Each call returns the next one, updating the -start address. - -Arguments: - cptr points to starting character value; updated - d end value - ocptr where to put start of othercase range - odptr where to put end of othercase range - -Yield: TRUE when range returned; FALSE when no more -*/ - -static BOOL -get_othercase_range(unsigned int *cptr, unsigned int d, unsigned int *ocptr, - unsigned int *odptr) -{ -unsigned int c, othercase, next; - -for (c = *cptr; c <= d; c++) - { if ((othercase = _pcre_ucp_othercase(c)) != NOTACHAR) break; } - -if (c > d) return FALSE; - -*ocptr = othercase; -next = othercase + 1; - -for (++c; c <= d; c++) - { - if (_pcre_ucp_othercase(c) != next) break; - next++; - } - -*odptr = next - 1; -*cptr = c; - -return TRUE; -} -#endif /* SUPPORT_UCP */ - - - -/************************************************* -* Check if auto-possessifying is possible * -*************************************************/ - -/* This function is called for unlimited repeats of certain items, to see -whether the next thing could possibly match the repeated item. If not, it makes -sense to automatically possessify the repeated item. - -Arguments: - op_code the repeated op code - this data for this item, depends on the opcode - utf8 TRUE in UTF-8 mode - utf8_char used for utf8 character bytes, NULL if not relevant - ptr next character in pattern - options options bits - cd contains pointers to tables etc. - -Returns: TRUE if possessifying is wanted -*/ - -static BOOL -check_auto_possessive(int op_code, int item, BOOL utf8, uschar *utf8_char, - const uschar *ptr, int options, compile_data *cd) -{ -int next; - -/* Skip whitespace and comments in extended mode */ - -if ((options & PCRE_EXTENDED) != 0) - { - for (;;) - { - while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++; - if (*ptr == '#') - { - while (*(++ptr) != 0) - if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; } - } - else break; - } - } - -/* If the next item is one that we can handle, get its value. A non-negative -value is a character, a negative value is an escape value. */ - -if (*ptr == '\\') - { - int temperrorcode = 0; - next = check_escape(&ptr, &temperrorcode, cd->bracount, options, FALSE); - if (temperrorcode != 0) return FALSE; - ptr++; /* Point after the escape sequence */ - } - -else if ((cd->ctypes[*ptr] & ctype_meta) == 0) - { -#ifdef SUPPORT_UTF8 - if (utf8) { GETCHARINC(next, ptr); } else -#endif - next = *ptr++; - } - -else return FALSE; - -/* Skip whitespace and comments in extended mode */ - -if ((options & PCRE_EXTENDED) != 0) - { - for (;;) - { - while ((cd->ctypes[*ptr] & ctype_space) != 0) ptr++; - if (*ptr == '#') - { - while (*(++ptr) != 0) - if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; } - } - else break; - } - } - -/* If the next thing is itself optional, we have to give up. */ - -if (*ptr == '*' || *ptr == '?' || strncmp((char *)ptr, "{0,", 3) == 0) - return FALSE; - -/* Now compare the next item with the previous opcode. If the previous is a -positive single character match, "item" either contains the character or, if -"item" is greater than 127 in utf8 mode, the character's bytes are in -utf8_char. */ - - -/* Handle cases when the next item is a character. */ - -if (next >= 0) switch(op_code) - { - case OP_CHAR: -#ifdef SUPPORT_UTF8 - if (utf8 && item > 127) { GETCHAR(item, utf8_char); } -#endif - return item != next; - - /* For CHARNC (caseless character) we must check the other case. If we have - Unicode property support, we can use it to test the other case of - high-valued characters. */ - - case OP_CHARNC: -#ifdef SUPPORT_UTF8 - if (utf8 && item > 127) { GETCHAR(item, utf8_char); } -#endif - if (item == next) return FALSE; -#ifdef SUPPORT_UTF8 - if (utf8) - { - unsigned int othercase; - if (next < 128) othercase = cd->fcc[next]; else -#ifdef SUPPORT_UCP - othercase = _pcre_ucp_othercase((unsigned int)next); -#else - othercase = NOTACHAR; -#endif - return (unsigned int)item != othercase; - } - else -#endif /* SUPPORT_UTF8 */ - return (item != cd->fcc[next]); /* Non-UTF-8 mode */ - - /* For OP_NOT, "item" must be a single-byte character. */ - - case OP_NOT: - if (item == next) return TRUE; - if ((options & PCRE_CASELESS) == 0) return FALSE; -#ifdef SUPPORT_UTF8 - if (utf8) - { - unsigned int othercase; - if (next < 128) othercase = cd->fcc[next]; else -#ifdef SUPPORT_UCP - othercase = _pcre_ucp_othercase(next); -#else - othercase = NOTACHAR; -#endif - return (unsigned int)item == othercase; - } - else -#endif /* SUPPORT_UTF8 */ - return (item == cd->fcc[next]); /* Non-UTF-8 mode */ - - case OP_DIGIT: - return next > 127 || (cd->ctypes[next] & ctype_digit) == 0; - - case OP_NOT_DIGIT: - return next <= 127 && (cd->ctypes[next] & ctype_digit) != 0; - - case OP_WHITESPACE: - return next > 127 || (cd->ctypes[next] & ctype_space) == 0; - - case OP_NOT_WHITESPACE: - return next <= 127 && (cd->ctypes[next] & ctype_space) != 0; - - case OP_WORDCHAR: - return next > 127 || (cd->ctypes[next] & ctype_word) == 0; - - case OP_NOT_WORDCHAR: - return next <= 127 && (cd->ctypes[next] & ctype_word) != 0; - - case OP_HSPACE: - case OP_NOT_HSPACE: - switch(next) - { - case 0x09: - case 0x20: - case 0xa0: - case 0x1680: - case 0x180e: - case 0x2000: - case 0x2001: - case 0x2002: - case 0x2003: - case 0x2004: - case 0x2005: - case 0x2006: - case 0x2007: - case 0x2008: - case 0x2009: - case 0x200A: - case 0x202f: - case 0x205f: - case 0x3000: - return op_code != OP_HSPACE; - default: - return op_code == OP_HSPACE; - } - - case OP_VSPACE: - case OP_NOT_VSPACE: - switch(next) - { - case 0x0a: - case 0x0b: - case 0x0c: - case 0x0d: - case 0x85: - case 0x2028: - case 0x2029: - return op_code != OP_VSPACE; - default: - return op_code == OP_VSPACE; - } - - default: - return FALSE; - } - - -/* Handle the case when the next item is \d, \s, etc. */ - -switch(op_code) - { - case OP_CHAR: - case OP_CHARNC: -#ifdef SUPPORT_UTF8 - if (utf8 && item > 127) { GETCHAR(item, utf8_char); } -#endif - switch(-next) - { - case ESC_d: - return item > 127 || (cd->ctypes[item] & ctype_digit) == 0; - - case ESC_D: - return item <= 127 && (cd->ctypes[item] & ctype_digit) != 0; - - case ESC_s: - return item > 127 || (cd->ctypes[item] & ctype_space) == 0; - - case ESC_S: - return item <= 127 && (cd->ctypes[item] & ctype_space) != 0; - - case ESC_w: - return item > 127 || (cd->ctypes[item] & ctype_word) == 0; - - case ESC_W: - return item <= 127 && (cd->ctypes[item] & ctype_word) != 0; - - case ESC_h: - case ESC_H: - switch(item) - { - case 0x09: - case 0x20: - case 0xa0: - case 0x1680: - case 0x180e: - case 0x2000: - case 0x2001: - case 0x2002: - case 0x2003: - case 0x2004: - case 0x2005: - case 0x2006: - case 0x2007: - case 0x2008: - case 0x2009: - case 0x200A: - case 0x202f: - case 0x205f: - case 0x3000: - return -next != ESC_h; - default: - return -next == ESC_h; - } - - case ESC_v: - case ESC_V: - switch(item) - { - case 0x0a: - case 0x0b: - case 0x0c: - case 0x0d: - case 0x85: - case 0x2028: - case 0x2029: - return -next != ESC_v; - default: - return -next == ESC_v; - } - - default: - return FALSE; - } - - case OP_DIGIT: - return next == -ESC_D || next == -ESC_s || next == -ESC_W || - next == -ESC_h || next == -ESC_v; - - case OP_NOT_DIGIT: - return next == -ESC_d; - - case OP_WHITESPACE: - return next == -ESC_S || next == -ESC_d || next == -ESC_w; - - case OP_NOT_WHITESPACE: - return next == -ESC_s || next == -ESC_h || next == -ESC_v; - - case OP_HSPACE: - return next == -ESC_S || next == -ESC_H || next == -ESC_d || next == -ESC_w; - - case OP_NOT_HSPACE: - return next == -ESC_h; - - /* Can't have \S in here because VT matches \S (Perl anomaly) */ - case OP_VSPACE: - return next == -ESC_V || next == -ESC_d || next == -ESC_w; - - case OP_NOT_VSPACE: - return next == -ESC_v; - - case OP_WORDCHAR: - return next == -ESC_W || next == -ESC_s || next == -ESC_h || next == -ESC_v; - - case OP_NOT_WORDCHAR: - return next == -ESC_w || next == -ESC_d; - - default: - return FALSE; - } - -/* Control does not reach here */ -} - - - -/************************************************* -* Compile one branch * -*************************************************/ - -/* Scan the pattern, compiling it into the a vector. If the options are -changed during the branch, the pointer is used to change the external options -bits. This function is used during the pre-compile phase when we are trying -to find out the amount of memory needed, as well as during the real compile -phase. The value of lengthptr distinguishes the two phases. - -Arguments: - optionsptr pointer to the option bits - codeptr points to the pointer to the current code point - ptrptr points to the current pattern pointer - errorcodeptr points to error code variable - firstbyteptr set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE) - reqbyteptr set to the last literal character required, else < 0 - bcptr points to current branch chain - cd contains pointers to tables etc. - lengthptr NULL during the real compile phase - points to length accumulator during pre-compile phase - -Returns: TRUE on success - FALSE, with *errorcodeptr set non-zero on error -*/ - -static BOOL -compile_branch(int *optionsptr, uschar **codeptr, const uschar **ptrptr, - int *errorcodeptr, int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, - compile_data *cd, int *lengthptr) -{ -int repeat_type, op_type; -int repeat_min = 0, repeat_max = 0; /* To please picky compilers */ -int bravalue = 0; -int greedy_default, greedy_non_default; -int firstbyte, reqbyte; -int zeroreqbyte, zerofirstbyte; -int req_caseopt, reqvary, tempreqvary; -int options = *optionsptr; -int after_manual_callout = 0; -int length_prevgroup = 0; -register int c; -register uschar *code = *codeptr; -uschar *last_code = code; -uschar *orig_code = code; -uschar *tempcode; -BOOL inescq = FALSE; -BOOL groupsetfirstbyte = FALSE; -const uschar *ptr = *ptrptr; -const uschar *tempptr; -uschar *previous = NULL; -uschar *previous_callout = NULL; -uschar *save_hwm = NULL; -uschar classbits[32]; - -#ifdef SUPPORT_UTF8 -BOOL class_utf8; -BOOL utf8 = (options & PCRE_UTF8) != 0; -uschar *class_utf8data; -uschar *class_utf8data_base; -uschar utf8_char[6]; -#else -BOOL utf8 = FALSE; -uschar *utf8_char = NULL; -#endif - -#ifdef DEBUG -if (lengthptr != NULL) DPRINTF((">> start branch\n")); -#endif - -/* Set up the default and non-default settings for greediness */ - -greedy_default = ((options & PCRE_UNGREEDY) != 0); -greedy_non_default = greedy_default ^ 1; - -/* Initialize no first byte, no required byte. REQ_UNSET means "no char -matching encountered yet". It gets changed to REQ_NONE if we hit something that -matches a non-fixed char first char; reqbyte just remains unset if we never -find one. - -When we hit a repeat whose minimum is zero, we may have to adjust these values -to take the zero repeat into account. This is implemented by setting them to -zerofirstbyte and zeroreqbyte when such a repeat is encountered. The individual -item types that can be repeated set these backoff variables appropriately. */ - -firstbyte = reqbyte = zerofirstbyte = zeroreqbyte = REQ_UNSET; - -/* The variable req_caseopt contains either the REQ_CASELESS value or zero, -according to the current setting of the caseless flag. REQ_CASELESS is a bit -value > 255. It is added into the firstbyte or reqbyte variables to record the -case status of the value. This is used only for ASCII characters. */ - -req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS : 0; - -/* Switch on next character until the end of the branch */ - -for (;; ptr++) - { - BOOL negate_class; - BOOL should_flip_negation; - BOOL possessive_quantifier; - BOOL is_quantifier; - BOOL is_recurse; - BOOL reset_bracount; - int class_charcount; - int class_lastchar; - int newoptions; - int recno; - int refsign; - int skipbytes; - int subreqbyte; - int subfirstbyte; - int terminator; - int mclength; - uschar mcbuffer[8]; - - /* Get next byte in the pattern */ - - c = *ptr; - - /* If we are in the pre-compile phase, accumulate the length used for the - previous cycle of this loop. */ - - if (lengthptr != NULL) - { -#ifdef DEBUG - if (code > cd->hwm) cd->hwm = code; /* High water info */ -#endif - if (code > cd->start_workspace + COMPILE_WORK_SIZE) /* Check for overrun */ - { - *errorcodeptr = ERR52; - goto FAILED; - } - - /* There is at least one situation where code goes backwards: this is the - case of a zero quantifier after a class (e.g. [ab]{0}). At compile time, - the class is simply eliminated. However, it is created first, so we have to - allow memory for it. Therefore, don't ever reduce the length at this point. - */ - - if (code < last_code) code = last_code; - - /* Paranoid check for integer overflow */ - - if (OFLOW_MAX - *lengthptr < code - last_code) - { - *errorcodeptr = ERR20; - goto FAILED; - } - - *lengthptr += code - last_code; - DPRINTF(("length=%d added %d c=%c\n", *lengthptr, code - last_code, c)); - - /* If "previous" is set and it is not at the start of the work space, move - it back to there, in order to avoid filling up the work space. Otherwise, - if "previous" is NULL, reset the current code pointer to the start. */ - - if (previous != NULL) - { - if (previous > orig_code) - { - memmove(orig_code, previous, code - previous); - code -= previous - orig_code; - previous = orig_code; - } - } - else code = orig_code; - - /* Remember where this code item starts so we can pick up the length - next time round. */ - - last_code = code; - } - - /* In the real compile phase, just check the workspace used by the forward - reference list. */ - - else if (cd->hwm > cd->start_workspace + COMPILE_WORK_SIZE) - { - *errorcodeptr = ERR52; - goto FAILED; - } - - /* If in \Q...\E, check for the end; if not, we have a literal */ - - if (inescq && c != 0) - { - if (c == '\\' && ptr[1] == 'E') - { - inescq = FALSE; - ptr++; - continue; - } - else - { - if (previous_callout != NULL) - { - if (lengthptr == NULL) /* Don't attempt in pre-compile phase */ - complete_callout(previous_callout, ptr, cd); - previous_callout = NULL; - } - if ((options & PCRE_AUTO_CALLOUT) != 0) - { - previous_callout = code; - code = auto_callout(code, ptr, cd); - } - goto NORMAL_CHAR; - } - } - - /* Fill in length of a previous callout, except when the next thing is - a quantifier. */ - - is_quantifier = c == '*' || c == '+' || c == '?' || - (c == '{' && is_counted_repeat(ptr+1)); - - if (!is_quantifier && previous_callout != NULL && - after_manual_callout-- <= 0) - { - if (lengthptr == NULL) /* Don't attempt in pre-compile phase */ - complete_callout(previous_callout, ptr, cd); - previous_callout = NULL; - } - - /* In extended mode, skip white space and comments */ - - if ((options & PCRE_EXTENDED) != 0) - { - if ((cd->ctypes[c] & ctype_space) != 0) continue; - if (c == '#') - { - while (*(++ptr) != 0) - { - if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; } - } - if (*ptr != 0) continue; - - /* Else fall through to handle end of string */ - c = 0; - } - } - - /* No auto callout for quantifiers. */ - - if ((options & PCRE_AUTO_CALLOUT) != 0 && !is_quantifier) - { - previous_callout = code; - code = auto_callout(code, ptr, cd); - } - - switch(c) - { - /* ===================================================================*/ - case 0: /* The branch terminates at string end */ - case '|': /* or | or ) */ - case ')': - *firstbyteptr = firstbyte; - *reqbyteptr = reqbyte; - *codeptr = code; - *ptrptr = ptr; - if (lengthptr != NULL) - { - if (OFLOW_MAX - *lengthptr < code - last_code) - { - *errorcodeptr = ERR20; - goto FAILED; - } - *lengthptr += code - last_code; /* To include callout length */ - DPRINTF((">> end branch\n")); - } - return TRUE; - - - /* ===================================================================*/ - /* Handle single-character metacharacters. In multiline mode, ^ disables - the setting of any following char as a first character. */ - - case '^': - if ((options & PCRE_MULTILINE) != 0) - { - if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; - } - previous = NULL; - *code++ = OP_CIRC; - break; - - case '$': - previous = NULL; - *code++ = OP_DOLL; - break; - - /* There can never be a first char if '.' is first, whatever happens about - repeats. The value of reqbyte doesn't change either. */ - - case '.': - if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; - zerofirstbyte = firstbyte; - zeroreqbyte = reqbyte; - previous = code; - *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY; - break; - - - /* ===================================================================*/ - /* Character classes. If the included characters are all < 256, we build a - 32-byte bitmap of the permitted characters, except in the special case - where there is only one such character. For negated classes, we build the - map as usual, then invert it at the end. However, we use a different opcode - so that data characters > 255 can be handled correctly. - - If the class contains characters outside the 0-255 range, a different - opcode is compiled. It may optionally have a bit map for characters < 256, - but those above are are explicitly listed afterwards. A flag byte tells - whether the bitmap is present, and whether this is a negated class or not. - - In JavaScript compatibility mode, an isolated ']' causes an error. In - default (Perl) mode, it is treated as a data character. */ - - case ']': - if ((cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0) - { - *errorcodeptr = ERR64; - goto FAILED; - } - goto NORMAL_CHAR; - - case '[': - previous = code; - - /* PCRE supports POSIX class stuff inside a class. Perl gives an error if - they are encountered at the top level, so we'll do that too. */ - - if ((ptr[1] == ':' || ptr[1] == '.' || ptr[1] == '=') && - check_posix_syntax(ptr, &tempptr)) - { - *errorcodeptr = (ptr[1] == ':')? ERR13 : ERR31; - goto FAILED; - } - - /* If the first character is '^', set the negation flag and skip it. Also, - if the first few characters (either before or after ^) are \Q\E or \E we - skip them too. This makes for compatibility with Perl. */ - - negate_class = FALSE; - for (;;) - { - c = *(++ptr); - if (c == '\\') - { - if (ptr[1] == 'E') ptr++; - else if (strncmp((const char *)ptr+1, "Q\\E", 3) == 0) ptr += 3; - else break; - } - else if (!negate_class && c == '^') - negate_class = TRUE; - else break; - } - - /* Empty classes are allowed in JavaScript compatibility mode. Otherwise, - an initial ']' is taken as a data character -- the code below handles - that. In JS mode, [] must always fail, so generate OP_FAIL, whereas - [^] must match any character, so generate OP_ALLANY. */ - - if (c ==']' && (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0) - { - *code++ = negate_class? OP_ALLANY : OP_FAIL; - if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; - zerofirstbyte = firstbyte; - break; - } - - /* If a class contains a negative special such as \S, we need to flip the - negation flag at the end, so that support for characters > 255 works - correctly (they are all included in the class). */ - - should_flip_negation = FALSE; - - /* Keep a count of chars with values < 256 so that we can optimize the case - of just a single character (as long as it's < 256). However, For higher - valued UTF-8 characters, we don't yet do any optimization. */ - - class_charcount = 0; - class_lastchar = -1; - - /* Initialize the 32-char bit map to all zeros. We build the map in a - temporary bit of memory, in case the class contains only 1 character (less - than 256), because in that case the compiled code doesn't use the bit map. - */ - - memset(classbits, 0, 32 * sizeof(uschar)); - -#ifdef SUPPORT_UTF8 - class_utf8 = FALSE; /* No chars >= 256 */ - class_utf8data = code + LINK_SIZE + 2; /* For UTF-8 items */ - class_utf8data_base = class_utf8data; /* For resetting in pass 1 */ -#endif - - /* Process characters until ] is reached. By writing this as a "do" it - means that an initial ] is taken as a data character. At the start of the - loop, c contains the first byte of the character. */ - - if (c != 0) do - { - const uschar *oldptr; - -#ifdef SUPPORT_UTF8 - if (utf8 && c > 127) - { /* Braces are required because the */ - GETCHARLEN(c, ptr, ptr); /* macro generates multiple statements */ - } - - /* In the pre-compile phase, accumulate the length of any UTF-8 extra - data and reset the pointer. This is so that very large classes that - contain a zillion UTF-8 characters no longer overwrite the work space - (which is on the stack). */ - - if (lengthptr != NULL) - { - *lengthptr += class_utf8data - class_utf8data_base; - class_utf8data = class_utf8data_base; - } - -#endif - - /* Inside \Q...\E everything is literal except \E */ - - if (inescq) - { - if (c == '\\' && ptr[1] == 'E') /* If we are at \E */ - { - inescq = FALSE; /* Reset literal state */ - ptr++; /* Skip the 'E' */ - continue; /* Carry on with next */ - } - goto CHECK_RANGE; /* Could be range if \E follows */ - } - - /* Handle POSIX class names. Perl allows a negation extension of the - form [:^name:]. A square bracket that doesn't match the syntax is - treated as a literal. We also recognize the POSIX constructions - [.ch.] and [=ch=] ("collating elements") and fault them, as Perl - 5.6 and 5.8 do. */ - - if (c == '[' && - (ptr[1] == ':' || ptr[1] == '.' || ptr[1] == '=') && - check_posix_syntax(ptr, &tempptr)) - { - BOOL local_negate = FALSE; - int posix_class, taboffset, tabopt; - register const uschar *cbits = cd->cbits; - uschar pbits[32]; - - if (ptr[1] != ':') - { - *errorcodeptr = ERR31; - goto FAILED; - } - - ptr += 2; - if (*ptr == '^') - { - local_negate = TRUE; - should_flip_negation = TRUE; /* Note negative special */ - ptr++; - } - - posix_class = check_posix_name(ptr, tempptr - ptr); - if (posix_class < 0) - { - *errorcodeptr = ERR30; - goto FAILED; - } - - /* If matching is caseless, upper and lower are converted to - alpha. This relies on the fact that the class table starts with - alpha, lower, upper as the first 3 entries. */ - - if ((options & PCRE_CASELESS) != 0 && posix_class <= 2) - posix_class = 0; - - /* We build the bit map for the POSIX class in a chunk of local store - because we may be adding and subtracting from it, and we don't want to - subtract bits that may be in the main map already. At the end we or the - result into the bit map that is being built. */ - - posix_class *= 3; - - /* Copy in the first table (always present) */ - - memcpy(pbits, cbits + posix_class_maps[posix_class], - 32 * sizeof(uschar)); - - /* If there is a second table, add or remove it as required. */ - - taboffset = posix_class_maps[posix_class + 1]; - tabopt = posix_class_maps[posix_class + 2]; - - if (taboffset >= 0) - { - if (tabopt >= 0) - for (c = 0; c < 32; c++) pbits[c] |= cbits[c + taboffset]; - else - for (c = 0; c < 32; c++) pbits[c] &= ~cbits[c + taboffset]; - } - - /* Not see if we need to remove any special characters. An option - value of 1 removes vertical space and 2 removes underscore. */ - - if (tabopt < 0) tabopt = -tabopt; - if (tabopt == 1) pbits[1] &= ~0x3c; - else if (tabopt == 2) pbits[11] &= 0x7f; - - /* Add the POSIX table or its complement into the main table that is - being built and we are done. */ - - if (local_negate) - for (c = 0; c < 32; c++) classbits[c] |= ~pbits[c]; - else - for (c = 0; c < 32; c++) classbits[c] |= pbits[c]; - - ptr = tempptr + 1; - class_charcount = 10; /* Set > 1; assumes more than 1 per class */ - continue; /* End of POSIX syntax handling */ - } - - /* Backslash may introduce a single character, or it may introduce one - of the specials, which just set a flag. The sequence \b is a special - case. Inside a class (and only there) it is treated as backspace. - Elsewhere it marks a word boundary. Other escapes have preset maps ready - to 'or' into the one we are building. We assume they have more than one - character in them, so set class_charcount bigger than one. */ - - if (c == '\\') - { - c = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE); - if (*errorcodeptr != 0) goto FAILED; - - if (-c == ESC_b) c = '\b'; /* \b is backspace in a class */ - else if (-c == ESC_X) c = 'X'; /* \X is literal X in a class */ - else if (-c == ESC_R) c = 'R'; /* \R is literal R in a class */ - else if (-c == ESC_Q) /* Handle start of quoted string */ - { - if (ptr[1] == '\\' && ptr[2] == 'E') - { - ptr += 2; /* avoid empty string */ - } - else inescq = TRUE; - continue; - } - else if (-c == ESC_E) continue; /* Ignore orphan \E */ - - if (c < 0) - { - register const uschar *cbits = cd->cbits; - class_charcount += 2; /* Greater than 1 is what matters */ - - /* Save time by not doing this in the pre-compile phase. */ - - if (lengthptr == NULL) switch (-c) - { - case ESC_d: - for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_digit]; - continue; - - case ESC_D: - should_flip_negation = TRUE; - for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_digit]; - continue; - - case ESC_w: - for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_word]; - continue; - - case ESC_W: - should_flip_negation = TRUE; - for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word]; - continue; - - case ESC_s: - for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_space]; - classbits[1] &= ~0x08; /* Perl 5.004 onwards omits VT from \s */ - continue; - - case ESC_S: - should_flip_negation = TRUE; - for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space]; - classbits[1] |= 0x08; /* Perl 5.004 onwards omits VT from \s */ - continue; - - default: /* Not recognized; fall through */ - break; /* Need "default" setting to stop compiler warning. */ - } - - /* In the pre-compile phase, just do the recognition. */ - - else if (c == -ESC_d || c == -ESC_D || c == -ESC_w || - c == -ESC_W || c == -ESC_s || c == -ESC_S) continue; - - /* We need to deal with \H, \h, \V, and \v in both phases because - they use extra memory. */ - - if (-c == ESC_h) - { - SETBIT(classbits, 0x09); /* VT */ - SETBIT(classbits, 0x20); /* SPACE */ - SETBIT(classbits, 0xa0); /* NSBP */ -#ifdef SUPPORT_UTF8 - if (utf8) - { - class_utf8 = TRUE; - *class_utf8data++ = XCL_SINGLE; - class_utf8data += _pcre_ord2utf8(0x1680, class_utf8data); - *class_utf8data++ = XCL_SINGLE; - class_utf8data += _pcre_ord2utf8(0x180e, class_utf8data); - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(0x2000, class_utf8data); - class_utf8data += _pcre_ord2utf8(0x200A, class_utf8data); - *class_utf8data++ = XCL_SINGLE; - class_utf8data += _pcre_ord2utf8(0x202f, class_utf8data); - *class_utf8data++ = XCL_SINGLE; - class_utf8data += _pcre_ord2utf8(0x205f, class_utf8data); - *class_utf8data++ = XCL_SINGLE; - class_utf8data += _pcre_ord2utf8(0x3000, class_utf8data); - } -#endif - continue; - } - - if (-c == ESC_H) - { - for (c = 0; c < 32; c++) - { - int x = 0xff; - switch (c) - { - case 0x09/8: x ^= 1 << (0x09%8); break; - case 0x20/8: x ^= 1 << (0x20%8); break; - case 0xa0/8: x ^= 1 << (0xa0%8); break; - default: break; - } - classbits[c] |= x; - } - -#ifdef SUPPORT_UTF8 - if (utf8) - { - class_utf8 = TRUE; - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data); - class_utf8data += _pcre_ord2utf8(0x167f, class_utf8data); - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(0x1681, class_utf8data); - class_utf8data += _pcre_ord2utf8(0x180d, class_utf8data); - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(0x180f, class_utf8data); - class_utf8data += _pcre_ord2utf8(0x1fff, class_utf8data); - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(0x200B, class_utf8data); - class_utf8data += _pcre_ord2utf8(0x202e, class_utf8data); - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(0x2030, class_utf8data); - class_utf8data += _pcre_ord2utf8(0x205e, class_utf8data); - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(0x2060, class_utf8data); - class_utf8data += _pcre_ord2utf8(0x2fff, class_utf8data); - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(0x3001, class_utf8data); - class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data); - } -#endif - continue; - } - - if (-c == ESC_v) - { - SETBIT(classbits, 0x0a); /* LF */ - SETBIT(classbits, 0x0b); /* VT */ - SETBIT(classbits, 0x0c); /* FF */ - SETBIT(classbits, 0x0d); /* CR */ - SETBIT(classbits, 0x85); /* NEL */ -#ifdef SUPPORT_UTF8 - if (utf8) - { - class_utf8 = TRUE; - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(0x2028, class_utf8data); - class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data); - } -#endif - continue; - } - - if (-c == ESC_V) - { - for (c = 0; c < 32; c++) - { - int x = 0xff; - switch (c) - { - case 0x0a/8: x ^= 1 << (0x0a%8); - x ^= 1 << (0x0b%8); - x ^= 1 << (0x0c%8); - x ^= 1 << (0x0d%8); - break; - case 0x85/8: x ^= 1 << (0x85%8); break; - default: break; - } - classbits[c] |= x; - } - -#ifdef SUPPORT_UTF8 - if (utf8) - { - class_utf8 = TRUE; - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(0x0100, class_utf8data); - class_utf8data += _pcre_ord2utf8(0x2027, class_utf8data); - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(0x2029, class_utf8data); - class_utf8data += _pcre_ord2utf8(0x7fffffff, class_utf8data); - } -#endif - continue; - } - - /* We need to deal with \P and \p in both phases. */ - -#ifdef SUPPORT_UCP - if (-c == ESC_p || -c == ESC_P) - { - BOOL negated; - int pdata; - int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr); - if (ptype < 0) goto FAILED; - class_utf8 = TRUE; - *class_utf8data++ = ((-c == ESC_p) != negated)? - XCL_PROP : XCL_NOTPROP; - *class_utf8data++ = ptype; - *class_utf8data++ = pdata; - class_charcount -= 2; /* Not a < 256 character */ - continue; - } -#endif - /* Unrecognized escapes are faulted if PCRE is running in its - strict mode. By default, for compatibility with Perl, they are - treated as literals. */ - - if ((options & PCRE_EXTRA) != 0) - { - *errorcodeptr = ERR7; - goto FAILED; - } - - class_charcount -= 2; /* Undo the default count from above */ - c = *ptr; /* Get the final character and fall through */ - } - - /* Fall through if we have a single character (c >= 0). This may be - greater than 256 in UTF-8 mode. */ - - } /* End of backslash handling */ - - /* A single character may be followed by '-' to form a range. However, - Perl does not permit ']' to be the end of the range. A '-' character - at the end is treated as a literal. Perl ignores orphaned \E sequences - entirely. The code for handling \Q and \E is messy. */ - - CHECK_RANGE: - while (ptr[1] == '\\' && ptr[2] == 'E') - { - inescq = FALSE; - ptr += 2; - } - - oldptr = ptr; - - /* Remember \r or \n */ - - if (c == '\r' || c == '\n') cd->external_flags |= PCRE_HASCRORLF; - - /* Check for range */ - - if (!inescq && ptr[1] == '-') - { - int d; - ptr += 2; - while (*ptr == '\\' && ptr[1] == 'E') ptr += 2; - - /* If we hit \Q (not followed by \E) at this point, go into escaped - mode. */ - - while (*ptr == '\\' && ptr[1] == 'Q') - { - ptr += 2; - if (*ptr == '\\' && ptr[1] == 'E') { ptr += 2; continue; } - inescq = TRUE; - break; - } - - if (*ptr == 0 || (!inescq && *ptr == ']')) - { - ptr = oldptr; - goto LONE_SINGLE_CHARACTER; - } - -#ifdef SUPPORT_UTF8 - if (utf8) - { /* Braces are required because the */ - GETCHARLEN(d, ptr, ptr); /* macro generates multiple statements */ - } - else -#endif - d = *ptr; /* Not UTF-8 mode */ - - /* The second part of a range can be a single-character escape, but - not any of the other escapes. Perl 5.6 treats a hyphen as a literal - in such circumstances. */ - - if (!inescq && d == '\\') - { - d = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE); - if (*errorcodeptr != 0) goto FAILED; - - /* \b is backspace; \X is literal X; \R is literal R; any other - special means the '-' was literal */ - - if (d < 0) - { - if (d == -ESC_b) d = '\b'; - else if (d == -ESC_X) d = 'X'; - else if (d == -ESC_R) d = 'R'; else - { - ptr = oldptr; - goto LONE_SINGLE_CHARACTER; /* A few lines below */ - } - } - } - - /* Check that the two values are in the correct order. Optimize - one-character ranges */ - - if (d < c) - { - *errorcodeptr = ERR8; - goto FAILED; - } - - if (d == c) goto LONE_SINGLE_CHARACTER; /* A few lines below */ - - /* Remember \r or \n */ - - if (d == '\r' || d == '\n') cd->external_flags |= PCRE_HASCRORLF; - - /* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless - matching, we have to use an XCLASS with extra data items. Caseless - matching for characters > 127 is available only if UCP support is - available. */ - -#ifdef SUPPORT_UTF8 - if (utf8 && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127))) - { - class_utf8 = TRUE; - - /* With UCP support, we can find the other case equivalents of - the relevant characters. There may be several ranges. Optimize how - they fit with the basic range. */ - -#ifdef SUPPORT_UCP - if ((options & PCRE_CASELESS) != 0) - { - unsigned int occ, ocd; - unsigned int cc = c; - unsigned int origd = d; - while (get_othercase_range(&cc, origd, &occ, &ocd)) - { - if (occ >= (unsigned int)c && - ocd <= (unsigned int)d) - continue; /* Skip embedded ranges */ - - if (occ < (unsigned int)c && - ocd >= (unsigned int)c - 1) /* Extend the basic range */ - { /* if there is overlap, */ - c = occ; /* noting that if occ < c */ - continue; /* we can't have ocd > d */ - } /* because a subrange is */ - if (ocd > (unsigned int)d && - occ <= (unsigned int)d + 1) /* always shorter than */ - { /* the basic range. */ - d = ocd; - continue; - } - - if (occ == ocd) - { - *class_utf8data++ = XCL_SINGLE; - } - else - { - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(occ, class_utf8data); - } - class_utf8data += _pcre_ord2utf8(ocd, class_utf8data); - } - } -#endif /* SUPPORT_UCP */ - - /* Now record the original range, possibly modified for UCP caseless - overlapping ranges. */ - - *class_utf8data++ = XCL_RANGE; - class_utf8data += _pcre_ord2utf8(c, class_utf8data); - class_utf8data += _pcre_ord2utf8(d, class_utf8data); - - /* With UCP support, we are done. Without UCP support, there is no - caseless matching for UTF-8 characters > 127; we can use the bit map - for the smaller ones. */ - -#ifdef SUPPORT_UCP - continue; /* With next character in the class */ -#else - if ((options & PCRE_CASELESS) == 0 || c > 127) continue; - - /* Adjust upper limit and fall through to set up the map */ - - d = 127; - -#endif /* SUPPORT_UCP */ - } -#endif /* SUPPORT_UTF8 */ - - /* We use the bit map for all cases when not in UTF-8 mode; else - ranges that lie entirely within 0-127 when there is UCP support; else - for partial ranges without UCP support. */ - - class_charcount += d - c + 1; - class_lastchar = d; - - /* We can save a bit of time by skipping this in the pre-compile. */ - - if (lengthptr == NULL) for (; c <= d; c++) - { - classbits[c/8] |= (1 << (c&7)); - if ((options & PCRE_CASELESS) != 0) - { - int uc = cd->fcc[c]; /* flip case */ - classbits[uc/8] |= (1 << (uc&7)); - } - } - - continue; /* Go get the next char in the class */ - } - - /* Handle a lone single character - we can get here for a normal - non-escape char, or after \ that introduces a single character or for an - apparent range that isn't. */ - - LONE_SINGLE_CHARACTER: - - /* Handle a character that cannot go in the bit map */ - -#ifdef SUPPORT_UTF8 - if (utf8 && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127))) - { - class_utf8 = TRUE; - *class_utf8data++ = XCL_SINGLE; - class_utf8data += _pcre_ord2utf8(c, class_utf8data); - -#ifdef SUPPORT_UCP - if ((options & PCRE_CASELESS) != 0) - { - unsigned int othercase; - if ((othercase = _pcre_ucp_othercase(c)) != NOTACHAR) - { - *class_utf8data++ = XCL_SINGLE; - class_utf8data += _pcre_ord2utf8(othercase, class_utf8data); - } - } -#endif /* SUPPORT_UCP */ - - } - else -#endif /* SUPPORT_UTF8 */ - - /* Handle a single-byte character */ - { - classbits[c/8] |= (1 << (c&7)); - if ((options & PCRE_CASELESS) != 0) - { - c = cd->fcc[c]; /* flip case */ - classbits[c/8] |= (1 << (c&7)); - } - class_charcount++; - class_lastchar = c; - } - } - - /* Loop until ']' reached. This "while" is the end of the "do" above. */ - - while ((c = *(++ptr)) != 0 && (c != ']' || inescq)); - - if (c == 0) /* Missing terminating ']' */ - { - *errorcodeptr = ERR6; - goto FAILED; - } - - -/* This code has been disabled because it would mean that \s counts as -an explicit \r or \n reference, and that's not really what is wanted. Now -we set the flag only if there is a literal "\r" or "\n" in the class. */ - -#if 0 - /* Remember whether \r or \n are in this class */ - - if (negate_class) - { - if ((classbits[1] & 0x24) != 0x24) cd->external_flags |= PCRE_HASCRORLF; - } - else - { - if ((classbits[1] & 0x24) != 0) cd->external_flags |= PCRE_HASCRORLF; - } -#endif - - - /* If class_charcount is 1, we saw precisely one character whose value is - less than 256. As long as there were no characters >= 128 and there was no - use of \p or \P, in other words, no use of any XCLASS features, we can - optimize. - - In UTF-8 mode, we can optimize the negative case only if there were no - characters >= 128 because OP_NOT and the related opcodes like OP_NOTSTAR - operate on single-bytes only. This is an historical hangover. Maybe one day - we can tidy these opcodes to handle multi-byte characters. - - The optimization throws away the bit map. We turn the item into a - 1-character OP_CHAR[NC] if it's positive, or OP_NOT if it's negative. Note - that OP_NOT does not support multibyte characters. In the positive case, it - can cause firstbyte to be set. Otherwise, there can be no first char if - this item is first, whatever repeat count may follow. In the case of - reqbyte, save the previous value for reinstating. */ - -#ifdef SUPPORT_UTF8 - if (class_charcount == 1 && !class_utf8 && - (!utf8 || !negate_class || class_lastchar < 128)) -#else - if (class_charcount == 1) -#endif - { - zeroreqbyte = reqbyte; - - /* The OP_NOT opcode works on one-byte characters only. */ - - if (negate_class) - { - if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; - zerofirstbyte = firstbyte; - *code++ = OP_NOT; - *code++ = class_lastchar; - break; - } - - /* For a single, positive character, get the value into mcbuffer, and - then we can handle this with the normal one-character code. */ - -#ifdef SUPPORT_UTF8 - if (utf8 && class_lastchar > 127) - mclength = _pcre_ord2utf8(class_lastchar, mcbuffer); - else -#endif - { - mcbuffer[0] = class_lastchar; - mclength = 1; - } - goto ONE_CHAR; - } /* End of 1-char optimization */ - - /* The general case - not the one-char optimization. If this is the first - thing in the branch, there can be no first char setting, whatever the - repeat count. Any reqbyte setting must remain unchanged after any kind of - repeat. */ - - if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; - zerofirstbyte = firstbyte; - zeroreqbyte = reqbyte; - - /* If there are characters with values > 255, we have to compile an - extended class, with its own opcode, unless there was a negated special - such as \S in the class, because in that case all characters > 255 are in - the class, so any that were explicitly given as well can be ignored. If - (when there are explicit characters > 255 that must be listed) there are no - characters < 256, we can omit the bitmap in the actual compiled code. */ - -#ifdef SUPPORT_UTF8 - if (class_utf8 && !should_flip_negation) - { - *class_utf8data++ = XCL_END; /* Marks the end of extra data */ - *code++ = OP_XCLASS; - code += LINK_SIZE; - *code = negate_class? XCL_NOT : 0; - - /* If the map is required, move up the extra data to make room for it; - otherwise just move the code pointer to the end of the extra data. */ - - if (class_charcount > 0) - { - *code++ |= XCL_MAP; - memmove(code + 32, code, class_utf8data - code); - memcpy(code, classbits, 32); - code = class_utf8data + 32; - } - else code = class_utf8data; - - /* Now fill in the complete length of the item */ - - PUT(previous, 1, code - previous); - break; /* End of class handling */ - } -#endif - - /* If there are no characters > 255, set the opcode to OP_CLASS or - OP_NCLASS, depending on whether the whole class was negated and whether - there were negative specials such as \S in the class. Then copy the 32-byte - map into the code vector, negating it if necessary. */ - - *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS; - if (negate_class) - { - if (lengthptr == NULL) /* Save time in the pre-compile phase */ - for (c = 0; c < 32; c++) code[c] = ~classbits[c]; - } - else - { - memcpy(code, classbits, 32); - } - code += 32; - break; - - - /* ===================================================================*/ - /* Various kinds of repeat; '{' is not necessarily a quantifier, but this - has been tested above. */ - - case '{': - if (!is_quantifier) goto NORMAL_CHAR; - ptr = read_repeat_counts(ptr+1, &repeat_min, &repeat_max, errorcodeptr); - if (*errorcodeptr != 0) goto FAILED; - goto REPEAT; - - case '*': - repeat_min = 0; - repeat_max = -1; - goto REPEAT; - - case '+': - repeat_min = 1; - repeat_max = -1; - goto REPEAT; - - case '?': - repeat_min = 0; - repeat_max = 1; - - REPEAT: - if (previous == NULL) - { - *errorcodeptr = ERR9; - goto FAILED; - } - - if (repeat_min == 0) - { - firstbyte = zerofirstbyte; /* Adjust for zero repeat */ - reqbyte = zeroreqbyte; /* Ditto */ - } - - /* Remember whether this is a variable length repeat */ - - reqvary = (repeat_min == repeat_max)? 0 : REQ_VARY; - - op_type = 0; /* Default single-char op codes */ - possessive_quantifier = FALSE; /* Default not possessive quantifier */ - - /* Save start of previous item, in case we have to move it up to make space - for an inserted OP_ONCE for the additional '+' extension. */ - - tempcode = previous; - - /* If the next character is '+', we have a possessive quantifier. This - implies greediness, whatever the setting of the PCRE_UNGREEDY option. - If the next character is '?' this is a minimizing repeat, by default, - but if PCRE_UNGREEDY is set, it works the other way round. We change the - repeat type to the non-default. */ - - if (ptr[1] == '+') - { - repeat_type = 0; /* Force greedy */ - possessive_quantifier = TRUE; - ptr++; - } - else if (ptr[1] == '?') - { - repeat_type = greedy_non_default; - ptr++; - } - else repeat_type = greedy_default; - - /* If previous was a character match, abolish the item and generate a - repeat item instead. If a char item has a minumum of more than one, ensure - that it is set in reqbyte - it might not be if a sequence such as x{3} is - the first thing in a branch because the x will have gone into firstbyte - instead. */ - - if (*previous == OP_CHAR || *previous == OP_CHARNC) - { - /* Deal with UTF-8 characters that take up more than one byte. It's - easier to write this out separately than try to macrify it. Use c to - hold the length of the character in bytes, plus 0x80 to flag that it's a - length rather than a small character. */ - -#ifdef SUPPORT_UTF8 - if (utf8 && (code[-1] & 0x80) != 0) - { - uschar *lastchar = code - 1; - while((*lastchar & 0xc0) == 0x80) lastchar--; - c = code - lastchar; /* Length of UTF-8 character */ - memcpy(utf8_char, lastchar, c); /* Save the char */ - c |= 0x80; /* Flag c as a length */ - } - else -#endif - - /* Handle the case of a single byte - either with no UTF8 support, or - with UTF-8 disabled, or for a UTF-8 character < 128. */ - - { - c = code[-1]; - if (repeat_min > 1) reqbyte = c | req_caseopt | cd->req_varyopt; - } - - /* If the repetition is unlimited, it pays to see if the next thing on - the line is something that cannot possibly match this character. If so, - automatically possessifying this item gains some performance in the case - where the match fails. */ - - if (!possessive_quantifier && - repeat_max < 0 && - check_auto_possessive(*previous, c, utf8, utf8_char, ptr + 1, - options, cd)) - { - repeat_type = 0; /* Force greedy */ - possessive_quantifier = TRUE; - } - - goto OUTPUT_SINGLE_REPEAT; /* Code shared with single character types */ - } - - /* If previous was a single negated character ([^a] or similar), we use - one of the special opcodes, replacing it. The code is shared with single- - character repeats by setting opt_type to add a suitable offset into - repeat_type. We can also test for auto-possessification. OP_NOT is - currently used only for single-byte chars. */ - - else if (*previous == OP_NOT) - { - op_type = OP_NOTSTAR - OP_STAR; /* Use "not" opcodes */ - c = previous[1]; - if (!possessive_quantifier && - repeat_max < 0 && - check_auto_possessive(OP_NOT, c, utf8, NULL, ptr + 1, options, cd)) - { - repeat_type = 0; /* Force greedy */ - possessive_quantifier = TRUE; - } - goto OUTPUT_SINGLE_REPEAT; - } - - /* If previous was a character type match (\d or similar), abolish it and - create a suitable repeat item. The code is shared with single-character - repeats by setting op_type to add a suitable offset into repeat_type. Note - the the Unicode property types will be present only when SUPPORT_UCP is - defined, but we don't wrap the little bits of code here because it just - makes it horribly messy. */ - - else if (*previous < OP_EODN) - { - uschar *oldcode; - int prop_type, prop_value; - op_type = OP_TYPESTAR - OP_STAR; /* Use type opcodes */ - c = *previous; - - if (!possessive_quantifier && - repeat_max < 0 && - check_auto_possessive(c, 0, utf8, NULL, ptr + 1, options, cd)) - { - repeat_type = 0; /* Force greedy */ - possessive_quantifier = TRUE; - } - - OUTPUT_SINGLE_REPEAT: - if (*previous == OP_PROP || *previous == OP_NOTPROP) - { - prop_type = previous[1]; - prop_value = previous[2]; - } - else prop_type = prop_value = -1; - - oldcode = code; - code = previous; /* Usually overwrite previous item */ - - /* If the maximum is zero then the minimum must also be zero; Perl allows - this case, so we do too - by simply omitting the item altogether. */ - - if (repeat_max == 0) goto END_REPEAT; - - /* All real repeats make it impossible to handle partial matching (maybe - one day we will be able to remove this restriction). */ - - if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL; - - /* Combine the op_type with the repeat_type */ - - repeat_type += op_type; - - /* A minimum of zero is handled either as the special case * or ?, or as - an UPTO, with the maximum given. */ - - if (repeat_min == 0) - { - if (repeat_max == -1) *code++ = OP_STAR + repeat_type; - else if (repeat_max == 1) *code++ = OP_QUERY + repeat_type; - else - { - *code++ = OP_UPTO + repeat_type; - PUT2INC(code, 0, repeat_max); - } - } - - /* A repeat minimum of 1 is optimized into some special cases. If the - maximum is unlimited, we use OP_PLUS. Otherwise, the original item is - left in place and, if the maximum is greater than 1, we use OP_UPTO with - one less than the maximum. */ - - else if (repeat_min == 1) - { - if (repeat_max == -1) - *code++ = OP_PLUS + repeat_type; - else - { - code = oldcode; /* leave previous item in place */ - if (repeat_max == 1) goto END_REPEAT; - *code++ = OP_UPTO + repeat_type; - PUT2INC(code, 0, repeat_max - 1); - } - } - - /* The case {n,n} is just an EXACT, while the general case {n,m} is - handled as an EXACT followed by an UPTO. */ - - else - { - *code++ = OP_EXACT + op_type; /* NB EXACT doesn't have repeat_type */ - PUT2INC(code, 0, repeat_min); - - /* If the maximum is unlimited, insert an OP_STAR. Before doing so, - we have to insert the character for the previous code. For a repeated - Unicode property match, there are two extra bytes that define the - required property. In UTF-8 mode, long characters have their length in - c, with the 0x80 bit as a flag. */ - - if (repeat_max < 0) - { -#ifdef SUPPORT_UTF8 - if (utf8 && c >= 128) - { - memcpy(code, utf8_char, c & 7); - code += c & 7; - } - else -#endif - { - *code++ = c; - if (prop_type >= 0) - { - *code++ = prop_type; - *code++ = prop_value; - } - } - *code++ = OP_STAR + repeat_type; - } - - /* Else insert an UPTO if the max is greater than the min, again - preceded by the character, for the previously inserted code. If the - UPTO is just for 1 instance, we can use QUERY instead. */ - - else if (repeat_max != repeat_min) - { -#ifdef SUPPORT_UTF8 - if (utf8 && c >= 128) - { - memcpy(code, utf8_char, c & 7); - code += c & 7; - } - else -#endif - *code++ = c; - if (prop_type >= 0) - { - *code++ = prop_type; - *code++ = prop_value; - } - repeat_max -= repeat_min; - - if (repeat_max == 1) - { - *code++ = OP_QUERY + repeat_type; - } - else - { - *code++ = OP_UPTO + repeat_type; - PUT2INC(code, 0, repeat_max); - } - } - } - - /* The character or character type itself comes last in all cases. */ - -#ifdef SUPPORT_UTF8 - if (utf8 && c >= 128) - { - memcpy(code, utf8_char, c & 7); - code += c & 7; - } - else -#endif - *code++ = c; - - /* For a repeated Unicode property match, there are two extra bytes that - define the required property. */ - -#ifdef SUPPORT_UCP - if (prop_type >= 0) - { - *code++ = prop_type; - *code++ = prop_value; - } -#endif - } - - /* If previous was a character class or a back reference, we put the repeat - stuff after it, but just skip the item if the repeat was {0,0}. */ - - else if (*previous == OP_CLASS || - *previous == OP_NCLASS || -#ifdef SUPPORT_UTF8 - *previous == OP_XCLASS || -#endif - *previous == OP_REF) - { - if (repeat_max == 0) - { - code = previous; - goto END_REPEAT; - } - - /* All real repeats make it impossible to handle partial matching (maybe - one day we will be able to remove this restriction). */ - - if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL; - - if (repeat_min == 0 && repeat_max == -1) - *code++ = OP_CRSTAR + repeat_type; - else if (repeat_min == 1 && repeat_max == -1) - *code++ = OP_CRPLUS + repeat_type; - else if (repeat_min == 0 && repeat_max == 1) - *code++ = OP_CRQUERY + repeat_type; - else - { - *code++ = OP_CRRANGE + repeat_type; - PUT2INC(code, 0, repeat_min); - if (repeat_max == -1) repeat_max = 0; /* 2-byte encoding for max */ - PUT2INC(code, 0, repeat_max); - } - } - - /* If previous was a bracket group, we may have to replicate it in certain - cases. */ - - else if (*previous == OP_BRA || *previous == OP_CBRA || - *previous == OP_ONCE || *previous == OP_COND) - { - register int i; - int ketoffset = 0; - int len = code - previous; - uschar *bralink = NULL; - - /* Repeating a DEFINE group is pointless */ - - if (*previous == OP_COND && previous[LINK_SIZE+1] == OP_DEF) - { - *errorcodeptr = ERR55; - goto FAILED; - } - - /* If the maximum repeat count is unlimited, find the end of the bracket - by scanning through from the start, and compute the offset back to it - from the current code pointer. There may be an OP_OPT setting following - the final KET, so we can't find the end just by going back from the code - pointer. */ - - if (repeat_max == -1) - { - register uschar *ket = previous; - do ket += GET(ket, 1); while (*ket != OP_KET); - ketoffset = code - ket; - } - - /* The case of a zero minimum is special because of the need to stick - OP_BRAZERO in front of it, and because the group appears once in the - data, whereas in other cases it appears the minimum number of times. For - this reason, it is simplest to treat this case separately, as otherwise - the code gets far too messy. There are several special subcases when the - minimum is zero. */ - - if (repeat_min == 0) - { - /* If the maximum is also zero, we used to just omit the group from the - output altogether, like this: - - ** if (repeat_max == 0) - ** { - ** code = previous; - ** goto END_REPEAT; - ** } - - However, that fails when a group is referenced as a subroutine from - elsewhere in the pattern, so now we stick in OP_SKIPZERO in front of it - so that it is skipped on execution. As we don't have a list of which - groups are referenced, we cannot do this selectively. - - If the maximum is 1 or unlimited, we just have to stick in the BRAZERO - and do no more at this point. However, we do need to adjust any - OP_RECURSE calls inside the group that refer to the group itself or any - internal or forward referenced group, because the offset is from the - start of the whole regex. Temporarily terminate the pattern while doing - this. */ - - if (repeat_max <= 1) /* Covers 0, 1, and unlimited */ - { - *code = OP_END; - adjust_recurse(previous, 1, utf8, cd, save_hwm); - memmove(previous+1, previous, len); - code++; - if (repeat_max == 0) - { - *previous++ = OP_SKIPZERO; - goto END_REPEAT; - } - *previous++ = OP_BRAZERO + repeat_type; - } - - /* If the maximum is greater than 1 and limited, we have to replicate - in a nested fashion, sticking OP_BRAZERO before each set of brackets. - The first one has to be handled carefully because it's the original - copy, which has to be moved up. The remainder can be handled by code - that is common with the non-zero minimum case below. We have to - adjust the value or repeat_max, since one less copy is required. Once - again, we may have to adjust any OP_RECURSE calls inside the group. */ - - else - { - int offset; - *code = OP_END; - adjust_recurse(previous, 2 + LINK_SIZE, utf8, cd, save_hwm); - memmove(previous + 2 + LINK_SIZE, previous, len); - code += 2 + LINK_SIZE; - *previous++ = OP_BRAZERO + repeat_type; - *previous++ = OP_BRA; - - /* We chain together the bracket offset fields that have to be - filled in later when the ends of the brackets are reached. */ - - offset = (bralink == NULL)? 0 : previous - bralink; - bralink = previous; - PUTINC(previous, 0, offset); - } - - repeat_max--; - } - - /* If the minimum is greater than zero, replicate the group as many - times as necessary, and adjust the maximum to the number of subsequent - copies that we need. If we set a first char from the group, and didn't - set a required char, copy the latter from the former. If there are any - forward reference subroutine calls in the group, there will be entries on - the workspace list; replicate these with an appropriate increment. */ - - else - { - if (repeat_min > 1) - { - /* In the pre-compile phase, we don't actually do the replication. We - just adjust the length as if we had. Do some paranoid checks for - potential integer overflow. */ - - if (lengthptr != NULL) - { - int delta = (repeat_min - 1)*length_prevgroup; - if ((double)(repeat_min - 1)*(double)length_prevgroup > - (double)INT_MAX || - OFLOW_MAX - *lengthptr < delta) - { - *errorcodeptr = ERR20; - goto FAILED; - } - *lengthptr += delta; - } - - /* This is compiling for real */ - - else - { - if (groupsetfirstbyte && reqbyte < 0) reqbyte = firstbyte; - for (i = 1; i < repeat_min; i++) - { - uschar *hc; - uschar *this_hwm = cd->hwm; - memcpy(code, previous, len); - for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE) - { - PUT(cd->hwm, 0, GET(hc, 0) + len); - cd->hwm += LINK_SIZE; - } - save_hwm = this_hwm; - code += len; - } - } - } - - if (repeat_max > 0) repeat_max -= repeat_min; - } - - /* This code is common to both the zero and non-zero minimum cases. If - the maximum is limited, it replicates the group in a nested fashion, - remembering the bracket starts on a stack. In the case of a zero minimum, - the first one was set up above. In all cases the repeat_max now specifies - the number of additional copies needed. Again, we must remember to - replicate entries on the forward reference list. */ - - if (repeat_max >= 0) - { - /* In the pre-compile phase, we don't actually do the replication. We - just adjust the length as if we had. For each repetition we must add 1 - to the length for BRAZERO and for all but the last repetition we must - add 2 + 2*LINKSIZE to allow for the nesting that occurs. Do some - paranoid checks to avoid integer overflow. */ - - if (lengthptr != NULL && repeat_max > 0) - { - int delta = repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) - - 2 - 2*LINK_SIZE; /* Last one doesn't nest */ - if ((double)repeat_max * - (double)(length_prevgroup + 1 + 2 + 2*LINK_SIZE) - > (double)INT_MAX || - OFLOW_MAX - *lengthptr < delta) - { - *errorcodeptr = ERR20; - goto FAILED; - } - *lengthptr += delta; - } - - /* This is compiling for real */ - - else for (i = repeat_max - 1; i >= 0; i--) - { - uschar *hc; - uschar *this_hwm = cd->hwm; - - *code++ = OP_BRAZERO + repeat_type; - - /* All but the final copy start a new nesting, maintaining the - chain of brackets outstanding. */ - - if (i != 0) - { - int offset; - *code++ = OP_BRA; - offset = (bralink == NULL)? 0 : code - bralink; - bralink = code; - PUTINC(code, 0, offset); - } - - memcpy(code, previous, len); - for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE) - { - PUT(cd->hwm, 0, GET(hc, 0) + len + ((i != 0)? 2+LINK_SIZE : 1)); - cd->hwm += LINK_SIZE; - } - save_hwm = this_hwm; - code += len; - } - - /* Now chain through the pending brackets, and fill in their length - fields (which are holding the chain links pro tem). */ - - while (bralink != NULL) - { - int oldlinkoffset; - int offset = code - bralink + 1; - uschar *bra = code - offset; - oldlinkoffset = GET(bra, 1); - bralink = (oldlinkoffset == 0)? NULL : bralink - oldlinkoffset; - *code++ = OP_KET; - PUTINC(code, 0, offset); - PUT(bra, 1, offset); - } - } - - /* If the maximum is unlimited, set a repeater in the final copy. We - can't just offset backwards from the current code point, because we - don't know if there's been an options resetting after the ket. The - correct offset was computed above. - - Then, when we are doing the actual compile phase, check to see whether - this group is a non-atomic one that could match an empty string. If so, - convert the initial operator to the S form (e.g. OP_BRA -> OP_SBRA) so - that runtime checking can be done. [This check is also applied to - atomic groups at runtime, but in a different way.] */ - - else - { - uschar *ketcode = code - ketoffset; - uschar *bracode = ketcode - GET(ketcode, 1); - *ketcode = OP_KETRMAX + repeat_type; - if (lengthptr == NULL && *bracode != OP_ONCE) - { - uschar *scode = bracode; - do - { - if (could_be_empty_branch(scode, ketcode, utf8)) - { - *bracode += OP_SBRA - OP_BRA; - break; - } - scode += GET(scode, 1); - } - while (*scode == OP_ALT); - } - } - } - - /* If previous is OP_FAIL, it was generated by an empty class [] in - JavaScript mode. The other ways in which OP_FAIL can be generated, that is - by (*FAIL) or (?!) set previous to NULL, which gives a "nothing to repeat" - error above. We can just ignore the repeat in JS case. */ - - else if (*previous == OP_FAIL) goto END_REPEAT; - - /* Else there's some kind of shambles */ - - else - { - *errorcodeptr = ERR11; - goto FAILED; - } - - /* If the character following a repeat is '+', or if certain optimization - tests above succeeded, possessive_quantifier is TRUE. For some of the - simpler opcodes, there is an special alternative opcode for this. For - anything else, we wrap the entire repeated item inside OP_ONCE brackets. - The '+' notation is just syntactic sugar, taken from Sun's Java package, - but the special opcodes can optimize it a bit. The repeated item starts at - tempcode, not at previous, which might be the first part of a string whose - (former) last char we repeated. - - Possessifying an 'exact' quantifier has no effect, so we can ignore it. But - an 'upto' may follow. We skip over an 'exact' item, and then test the - length of what remains before proceeding. */ - - if (possessive_quantifier) - { - int len; - if (*tempcode == OP_EXACT || *tempcode == OP_TYPEEXACT || - *tempcode == OP_NOTEXACT) - tempcode += _pcre_OP_lengths[*tempcode] + - ((*tempcode == OP_TYPEEXACT && - (tempcode[3] == OP_PROP || tempcode[3] == OP_NOTPROP))? 2:0); - len = code - tempcode; - if (len > 0) switch (*tempcode) - { - case OP_STAR: *tempcode = OP_POSSTAR; break; - case OP_PLUS: *tempcode = OP_POSPLUS; break; - case OP_QUERY: *tempcode = OP_POSQUERY; break; - case OP_UPTO: *tempcode = OP_POSUPTO; break; - - case OP_TYPESTAR: *tempcode = OP_TYPEPOSSTAR; break; - case OP_TYPEPLUS: *tempcode = OP_TYPEPOSPLUS; break; - case OP_TYPEQUERY: *tempcode = OP_TYPEPOSQUERY; break; - case OP_TYPEUPTO: *tempcode = OP_TYPEPOSUPTO; break; - - case OP_NOTSTAR: *tempcode = OP_NOTPOSSTAR; break; - case OP_NOTPLUS: *tempcode = OP_NOTPOSPLUS; break; - case OP_NOTQUERY: *tempcode = OP_NOTPOSQUERY; break; - case OP_NOTUPTO: *tempcode = OP_NOTPOSUPTO; break; - - default: - memmove(tempcode + 1+LINK_SIZE, tempcode, len); - code += 1 + LINK_SIZE; - len += 1 + LINK_SIZE; - tempcode[0] = OP_ONCE; - *code++ = OP_KET; - PUTINC(code, 0, len); - PUT(tempcode, 1, len); - break; - } - } - - /* In all case we no longer have a previous item. We also set the - "follows varying string" flag for subsequently encountered reqbytes if - it isn't already set and we have just passed a varying length item. */ - - END_REPEAT: - previous = NULL; - cd->req_varyopt |= reqvary; - break; - - - /* ===================================================================*/ - /* Start of nested parenthesized sub-expression, or comment or lookahead or - lookbehind or option setting or condition or all the other extended - parenthesis forms. */ - - case '(': - newoptions = options; - skipbytes = 0; - bravalue = OP_CBRA; - save_hwm = cd->hwm; - reset_bracount = FALSE; - - /* First deal with various "verbs" that can be introduced by '*'. */ - - if (*(++ptr) == '*' && (cd->ctypes[ptr[1]] & ctype_letter) != 0) - { - int i, namelen; - const char *vn = verbnames; - const uschar *name = ++ptr; - previous = NULL; - while ((cd->ctypes[*++ptr] & ctype_letter) != 0); - if (*ptr == ':') - { - *errorcodeptr = ERR59; /* Not supported */ - goto FAILED; - } - if (*ptr != ')') - { - *errorcodeptr = ERR60; - goto FAILED; - } - namelen = ptr - name; - for (i = 0; i < verbcount; i++) - { - if (namelen == verbs[i].len && - strncmp((char *)name, vn, namelen) == 0) - { - *code = verbs[i].op; - if (*code++ == OP_ACCEPT) cd->had_accept = TRUE; - break; - } - vn += verbs[i].len + 1; - } - if (i < verbcount) continue; - *errorcodeptr = ERR60; - goto FAILED; - } - - /* Deal with the extended parentheses; all are introduced by '?', and the - appearance of any of them means that this is not a capturing group. */ - - else if (*ptr == '?') - { - int i, set, unset, namelen; - int *optset; - const uschar *name; - uschar *slot; - - switch (*(++ptr)) - { - case '#': /* Comment; skip to ket */ - ptr++; - while (*ptr != 0 && *ptr != ')') ptr++; - if (*ptr == 0) - { - *errorcodeptr = ERR18; - goto FAILED; - } - continue; - - - /* ------------------------------------------------------------ */ - case '|': /* Reset capture count for each branch */ - reset_bracount = TRUE; - /* Fall through */ - - /* ------------------------------------------------------------ */ - case ':': /* Non-capturing bracket */ - bravalue = OP_BRA; - ptr++; - break; - - - /* ------------------------------------------------------------ */ - case '(': - bravalue = OP_COND; /* Conditional group */ - - /* A condition can be an assertion, a number (referring to a numbered - group), a name (referring to a named group), or 'R', referring to - recursion. R and R&name are also permitted for recursion tests. - - There are several syntaxes for testing a named group: (?(name)) is used - by Python; Perl 5.10 onwards uses (?() or (?('name')). - - There are two unfortunate ambiguities, caused by history. (a) 'R' can - be the recursive thing or the name 'R' (and similarly for 'R' followed - by digits), and (b) a number could be a name that consists of digits. - In both cases, we look for a name first; if not found, we try the other - cases. */ - - /* For conditions that are assertions, check the syntax, and then exit - the switch. This will take control down to where bracketed groups, - including assertions, are processed. */ - - if (ptr[1] == '?' && (ptr[2] == '=' || ptr[2] == '!' || ptr[2] == '<')) - break; - - /* Most other conditions use OP_CREF (a couple change to OP_RREF - below), and all need to skip 3 bytes at the start of the group. */ - - code[1+LINK_SIZE] = OP_CREF; - skipbytes = 3; - refsign = -1; - - /* Check for a test for recursion in a named group. */ - - if (ptr[1] == 'R' && ptr[2] == '&') - { - terminator = -1; - ptr += 2; - code[1+LINK_SIZE] = OP_RREF; /* Change the type of test */ - } - - /* Check for a test for a named group's having been set, using the Perl - syntax (?() or (?('name') */ - - else if (ptr[1] == '<') - { - terminator = '>'; - ptr++; - } - else if (ptr[1] == '\'') - { - terminator = '\''; - ptr++; - } - else - { - terminator = 0; - if (ptr[1] == '-' || ptr[1] == '+') refsign = *(++ptr); - } - - /* We now expect to read a name; any thing else is an error */ - - if ((cd->ctypes[ptr[1]] & ctype_word) == 0) - { - ptr += 1; /* To get the right offset */ - *errorcodeptr = ERR28; - goto FAILED; - } - - /* Read the name, but also get it as a number if it's all digits */ - - recno = 0; - name = ++ptr; - while ((cd->ctypes[*ptr] & ctype_word) != 0) - { - if (recno >= 0) - recno = ((digitab[*ptr] & ctype_digit) != 0)? - recno * 10 + *ptr - '0' : -1; - ptr++; - } - namelen = ptr - name; - - if ((terminator > 0 && *ptr++ != terminator) || *ptr++ != ')') - { - ptr--; /* Error offset */ - *errorcodeptr = ERR26; - goto FAILED; - } - - /* Do no further checking in the pre-compile phase. */ - - if (lengthptr != NULL) break; - - /* In the real compile we do the work of looking for the actual - reference. If the string started with "+" or "-" we require the rest to - be digits, in which case recno will be set. */ - - if (refsign > 0) - { - if (recno <= 0) - { - *errorcodeptr = ERR58; - goto FAILED; - } - recno = (refsign == '-')? - cd->bracount - recno + 1 : recno +cd->bracount; - if (recno <= 0 || recno > cd->final_bracount) - { - *errorcodeptr = ERR15; - goto FAILED; - } - PUT2(code, 2+LINK_SIZE, recno); - break; - } - - /* Otherwise (did not start with "+" or "-"), start by looking for the - name. */ - - slot = cd->name_table; - for (i = 0; i < cd->names_found; i++) - { - if (strncmp((char *)name, (char *)slot+2, namelen) == 0) break; - slot += cd->name_entry_size; - } - - /* Found a previous named subpattern */ - - if (i < cd->names_found) - { - recno = GET2(slot, 0); - PUT2(code, 2+LINK_SIZE, recno); - } - - /* Search the pattern for a forward reference */ - - else if ((i = find_parens(ptr, cd, name, namelen, - (options & PCRE_EXTENDED) != 0)) > 0) - { - PUT2(code, 2+LINK_SIZE, i); - } - - /* If terminator == 0 it means that the name followed directly after - the opening parenthesis [e.g. (?(abc)...] and in this case there are - some further alternatives to try. For the cases where terminator != 0 - [things like (?(... or (?('name')... or (?(R&name)... ] we have - now checked all the possibilities, so give an error. */ - - else if (terminator != 0) - { - *errorcodeptr = ERR15; - goto FAILED; - } - - /* Check for (?(R) for recursion. Allow digits after R to specify a - specific group number. */ - - else if (*name == 'R') - { - recno = 0; - for (i = 1; i < namelen; i++) - { - if ((digitab[name[i]] & ctype_digit) == 0) - { - *errorcodeptr = ERR15; - goto FAILED; - } - recno = recno * 10 + name[i] - '0'; - } - if (recno == 0) recno = RREF_ANY; - code[1+LINK_SIZE] = OP_RREF; /* Change test type */ - PUT2(code, 2+LINK_SIZE, recno); - } - - /* Similarly, check for the (?(DEFINE) "condition", which is always - false. */ - - else if (namelen == 6 && strncmp((char *)name, "DEFINE", 6) == 0) - { - code[1+LINK_SIZE] = OP_DEF; - skipbytes = 1; - } - - /* Check for the "name" actually being a subpattern number. We are - in the second pass here, so final_bracount is set. */ - - else if (recno > 0 && recno <= cd->final_bracount) - { - PUT2(code, 2+LINK_SIZE, recno); - } - - /* Either an unidentified subpattern, or a reference to (?(0) */ - - else - { - *errorcodeptr = (recno == 0)? ERR35: ERR15; - goto FAILED; - } - break; - - - /* ------------------------------------------------------------ */ - case '=': /* Positive lookahead */ - bravalue = OP_ASSERT; - ptr++; - break; - - - /* ------------------------------------------------------------ */ - case '!': /* Negative lookahead */ - ptr++; - if (*ptr == ')') /* Optimize (?!) */ - { - *code++ = OP_FAIL; - previous = NULL; - continue; - } - bravalue = OP_ASSERT_NOT; - break; - - - /* ------------------------------------------------------------ */ - case '<': /* Lookbehind or named define */ - switch (ptr[1]) - { - case '=': /* Positive lookbehind */ - bravalue = OP_ASSERTBACK; - ptr += 2; - break; - - case '!': /* Negative lookbehind */ - bravalue = OP_ASSERTBACK_NOT; - ptr += 2; - break; - - default: /* Could be name define, else bad */ - if ((cd->ctypes[ptr[1]] & ctype_word) != 0) goto DEFINE_NAME; - ptr++; /* Correct offset for error */ - *errorcodeptr = ERR24; - goto FAILED; - } - break; - - - /* ------------------------------------------------------------ */ - case '>': /* One-time brackets */ - bravalue = OP_ONCE; - ptr++; - break; - - - /* ------------------------------------------------------------ */ - case 'C': /* Callout - may be followed by digits; */ - previous_callout = code; /* Save for later completion */ - after_manual_callout = 1; /* Skip one item before completing */ - *code++ = OP_CALLOUT; - { - int n = 0; - while ((digitab[*(++ptr)] & ctype_digit) != 0) - n = n * 10 + *ptr - '0'; - if (*ptr != ')') - { - *errorcodeptr = ERR39; - goto FAILED; - } - if (n > 255) - { - *errorcodeptr = ERR38; - goto FAILED; - } - *code++ = n; - PUT(code, 0, ptr - cd->start_pattern + 1); /* Pattern offset */ - PUT(code, LINK_SIZE, 0); /* Default length */ - code += 2 * LINK_SIZE; - } - previous = NULL; - continue; - - - /* ------------------------------------------------------------ */ - case 'P': /* Python-style named subpattern handling */ - if (*(++ptr) == '=' || *ptr == '>') /* Reference or recursion */ - { - is_recurse = *ptr == '>'; - terminator = ')'; - goto NAMED_REF_OR_RECURSE; - } - else if (*ptr != '<') /* Test for Python-style definition */ - { - *errorcodeptr = ERR41; - goto FAILED; - } - /* Fall through to handle (?P< as (?< is handled */ - - - /* ------------------------------------------------------------ */ - DEFINE_NAME: /* Come here from (?< handling */ - case '\'': - { - terminator = (*ptr == '<')? '>' : '\''; - name = ++ptr; - - while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++; - namelen = ptr - name; - - /* In the pre-compile phase, just do a syntax check. */ - - if (lengthptr != NULL) - { - if (*ptr != terminator) - { - *errorcodeptr = ERR42; - goto FAILED; - } - if (cd->names_found >= MAX_NAME_COUNT) - { - *errorcodeptr = ERR49; - goto FAILED; - } - if (namelen + 3 > cd->name_entry_size) - { - cd->name_entry_size = namelen + 3; - if (namelen > MAX_NAME_SIZE) - { - *errorcodeptr = ERR48; - goto FAILED; - } - } - } - - /* In the real compile, create the entry in the table */ - - else - { - slot = cd->name_table; - for (i = 0; i < cd->names_found; i++) - { - int crc = memcmp(name, slot+2, namelen); - if (crc == 0) - { - if (slot[2+namelen] == 0) - { - if ((options & PCRE_DUPNAMES) == 0) - { - *errorcodeptr = ERR43; - goto FAILED; - } - } - else crc = -1; /* Current name is substring */ - } - if (crc < 0) - { - memmove(slot + cd->name_entry_size, slot, - (cd->names_found - i) * cd->name_entry_size); - break; - } - slot += cd->name_entry_size; - } - - PUT2(slot, 0, cd->bracount + 1); - memcpy(slot + 2, name, namelen); - slot[2+namelen] = 0; - } - } - - /* In both cases, count the number of names we've encountered. */ - - ptr++; /* Move past > or ' */ - cd->names_found++; - goto NUMBERED_GROUP; - - - /* ------------------------------------------------------------ */ - case '&': /* Perl recursion/subroutine syntax */ - terminator = ')'; - is_recurse = TRUE; - /* Fall through */ - - /* We come here from the Python syntax above that handles both - references (?P=name) and recursion (?P>name), as well as falling - through from the Perl recursion syntax (?&name). We also come here from - the Perl \k or \k'name' back reference syntax and the \k{name} - .NET syntax, and the Oniguruma \g<...> and \g'...' subroutine syntax. */ - - NAMED_REF_OR_RECURSE: - name = ++ptr; - while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++; - namelen = ptr - name; - - /* In the pre-compile phase, do a syntax check and set a dummy - reference number. */ - - if (lengthptr != NULL) - { - if (namelen == 0) - { - *errorcodeptr = ERR62; - goto FAILED; - } - if (*ptr != terminator) - { - *errorcodeptr = ERR42; - goto FAILED; - } - if (namelen > MAX_NAME_SIZE) - { - *errorcodeptr = ERR48; - goto FAILED; - } - recno = 0; - } - - /* In the real compile, seek the name in the table. We check the name - first, and then check that we have reached the end of the name in the - table. That way, if the name that is longer than any in the table, - the comparison will fail without reading beyond the table entry. */ - - else - { - slot = cd->name_table; - for (i = 0; i < cd->names_found; i++) - { - if (strncmp((char *)name, (char *)slot+2, namelen) == 0 && - slot[2+namelen] == 0) - break; - slot += cd->name_entry_size; - } - - if (i < cd->names_found) /* Back reference */ - { - recno = GET2(slot, 0); - } - else if ((recno = /* Forward back reference */ - find_parens(ptr, cd, name, namelen, - (options & PCRE_EXTENDED) != 0)) <= 0) - { - *errorcodeptr = ERR15; - goto FAILED; - } - } - - /* In both phases, we can now go to the code than handles numerical - recursion or backreferences. */ - - if (is_recurse) goto HANDLE_RECURSION; - else goto HANDLE_REFERENCE; - - - /* ------------------------------------------------------------ */ - case 'R': /* Recursion */ - ptr++; /* Same as (?0) */ - /* Fall through */ - - - /* ------------------------------------------------------------ */ - case '-': case '+': - case '0': case '1': case '2': case '3': case '4': /* Recursion or */ - case '5': case '6': case '7': case '8': case '9': /* subroutine */ - { - const uschar *called; - terminator = ')'; - - /* Come here from the \g<...> and \g'...' code (Oniguruma - compatibility). However, the syntax has been checked to ensure that - the ... are a (signed) number, so that neither ERR63 nor ERR29 will - be called on this path, nor with the jump to OTHER_CHAR_AFTER_QUERY - ever be taken. */ - - HANDLE_NUMERICAL_RECURSION: - - if ((refsign = *ptr) == '+') - { - ptr++; - if ((digitab[*ptr] & ctype_digit) == 0) - { - *errorcodeptr = ERR63; - goto FAILED; - } - } - else if (refsign == '-') - { - if ((digitab[ptr[1]] & ctype_digit) == 0) - goto OTHER_CHAR_AFTER_QUERY; - ptr++; - } - - recno = 0; - while((digitab[*ptr] & ctype_digit) != 0) - recno = recno * 10 + *ptr++ - '0'; - - if (*ptr != terminator) - { - *errorcodeptr = ERR29; - goto FAILED; - } - - if (refsign == '-') - { - if (recno == 0) - { - *errorcodeptr = ERR58; - goto FAILED; - } - recno = cd->bracount - recno + 1; - if (recno <= 0) - { - *errorcodeptr = ERR15; - goto FAILED; - } - } - else if (refsign == '+') - { - if (recno == 0) - { - *errorcodeptr = ERR58; - goto FAILED; - } - recno += cd->bracount; - } - - /* Come here from code above that handles a named recursion */ - - HANDLE_RECURSION: - - previous = code; - called = cd->start_code; - - /* When we are actually compiling, find the bracket that is being - referenced. Temporarily end the regex in case it doesn't exist before - this point. If we end up with a forward reference, first check that - the bracket does occur later so we can give the error (and position) - now. Then remember this forward reference in the workspace so it can - be filled in at the end. */ - - if (lengthptr == NULL) - { - *code = OP_END; - if (recno != 0) called = find_bracket(cd->start_code, utf8, recno); - - /* Forward reference */ - - if (called == NULL) - { - if (find_parens(ptr, cd, NULL, recno, - (options & PCRE_EXTENDED) != 0) < 0) - { - *errorcodeptr = ERR15; - goto FAILED; - } - called = cd->start_code + recno; - PUTINC(cd->hwm, 0, code + 2 + LINK_SIZE - cd->start_code); - } - - /* If not a forward reference, and the subpattern is still open, - this is a recursive call. We check to see if this is a left - recursion that could loop for ever, and diagnose that case. */ - - else if (GET(called, 1) == 0 && - could_be_empty(called, code, bcptr, utf8)) - { - *errorcodeptr = ERR40; - goto FAILED; - } - } - - /* Insert the recursion/subroutine item, automatically wrapped inside - "once" brackets. Set up a "previous group" length so that a - subsequent quantifier will work. */ - - *code = OP_ONCE; - PUT(code, 1, 2 + 2*LINK_SIZE); - code += 1 + LINK_SIZE; - - *code = OP_RECURSE; - PUT(code, 1, called - cd->start_code); - code += 1 + LINK_SIZE; - - *code = OP_KET; - PUT(code, 1, 2 + 2*LINK_SIZE); - code += 1 + LINK_SIZE; - - length_prevgroup = 3 + 3*LINK_SIZE; - } - - /* Can't determine a first byte now */ - - if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; - continue; - - - /* ------------------------------------------------------------ */ - default: /* Other characters: check option setting */ - OTHER_CHAR_AFTER_QUERY: - set = unset = 0; - optset = &set; - - while (*ptr != ')' && *ptr != ':') - { - switch (*ptr++) - { - case '-': optset = &unset; break; - - case 'J': /* Record that it changed in the external options */ - *optset |= PCRE_DUPNAMES; - cd->external_flags |= PCRE_JCHANGED; - break; - - case 'i': *optset |= PCRE_CASELESS; break; - case 'm': *optset |= PCRE_MULTILINE; break; - case 's': *optset |= PCRE_DOTALL; break; - case 'x': *optset |= PCRE_EXTENDED; break; - case 'U': *optset |= PCRE_UNGREEDY; break; - case 'X': *optset |= PCRE_EXTRA; break; - - default: *errorcodeptr = ERR12; - ptr--; /* Correct the offset */ - goto FAILED; - } - } - - /* Set up the changed option bits, but don't change anything yet. */ - - newoptions = (options | set) & (~unset); - - /* If the options ended with ')' this is not the start of a nested - group with option changes, so the options change at this level. If this - item is right at the start of the pattern, the options can be - abstracted and made external in the pre-compile phase, and ignored in - the compile phase. This can be helpful when matching -- for instance in - caseless checking of required bytes. - - If the code pointer is not (cd->start_code + 1 + LINK_SIZE), we are - definitely *not* at the start of the pattern because something has been - compiled. In the pre-compile phase, however, the code pointer can have - that value after the start, because it gets reset as code is discarded - during the pre-compile. However, this can happen only at top level - if - we are within parentheses, the starting BRA will still be present. At - any parenthesis level, the length value can be used to test if anything - has been compiled at that level. Thus, a test for both these conditions - is necessary to ensure we correctly detect the start of the pattern in - both phases. - - If we are not at the pattern start, compile code to change the ims - options if this setting actually changes any of them. We also pass the - new setting back so that it can be put at the start of any following - branches, and when this group ends (if we are in a group), a resetting - item can be compiled. */ - - if (*ptr == ')') - { - if (code == cd->start_code + 1 + LINK_SIZE && - (lengthptr == NULL || *lengthptr == 2 + 2*LINK_SIZE)) - { - cd->external_options = newoptions; - options = newoptions; - } - else - { - if ((options & PCRE_IMS) != (newoptions & PCRE_IMS)) - { - *code++ = OP_OPT; - *code++ = newoptions & PCRE_IMS; - } - - /* Change options at this level, and pass them back for use - in subsequent branches. Reset the greedy defaults and the case - value for firstbyte and reqbyte. */ - - *optionsptr = options = newoptions; - greedy_default = ((newoptions & PCRE_UNGREEDY) != 0); - greedy_non_default = greedy_default ^ 1; - req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS : 0; - } - - previous = NULL; /* This item can't be repeated */ - continue; /* It is complete */ - } - - /* If the options ended with ':' we are heading into a nested group - with possible change of options. Such groups are non-capturing and are - not assertions of any kind. All we need to do is skip over the ':'; - the newoptions value is handled below. */ - - bravalue = OP_BRA; - ptr++; - } /* End of switch for character following (? */ - } /* End of (? handling */ - - /* Opening parenthesis not followed by '?'. If PCRE_NO_AUTO_CAPTURE is set, - all unadorned brackets become non-capturing and behave like (?:...) - brackets. */ - - else if ((options & PCRE_NO_AUTO_CAPTURE) != 0) - { - bravalue = OP_BRA; - } - - /* Else we have a capturing group. */ - - else - { - NUMBERED_GROUP: - cd->bracount += 1; - PUT2(code, 1+LINK_SIZE, cd->bracount); - skipbytes = 2; - } - - /* Process nested bracketed regex. Assertions may not be repeated, but - other kinds can be. All their opcodes are >= OP_ONCE. We copy code into a - non-register variable in order to be able to pass its address because some - compilers complain otherwise. Pass in a new setting for the ims options if - they have changed. */ - - previous = (bravalue >= OP_ONCE)? code : NULL; - *code = bravalue; - tempcode = code; - tempreqvary = cd->req_varyopt; /* Save value before bracket */ - length_prevgroup = 0; /* Initialize for pre-compile phase */ - - if (!compile_regex( - newoptions, /* The complete new option state */ - options & PCRE_IMS, /* The previous ims option state */ - &tempcode, /* Where to put code (updated) */ - &ptr, /* Input pointer (updated) */ - errorcodeptr, /* Where to put an error message */ - (bravalue == OP_ASSERTBACK || - bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */ - reset_bracount, /* True if (?| group */ - skipbytes, /* Skip over bracket number */ - &subfirstbyte, /* For possible first char */ - &subreqbyte, /* For possible last char */ - bcptr, /* Current branch chain */ - cd, /* Tables block */ - (lengthptr == NULL)? NULL : /* Actual compile phase */ - &length_prevgroup /* Pre-compile phase */ - )) - goto FAILED; - - /* At the end of compiling, code is still pointing to the start of the - group, while tempcode has been updated to point past the end of the group - and any option resetting that may follow it. The pattern pointer (ptr) - is on the bracket. */ - - /* If this is a conditional bracket, check that there are no more than - two branches in the group, or just one if it's a DEFINE group. We do this - in the real compile phase, not in the pre-pass, where the whole group may - not be available. */ - - if (bravalue == OP_COND && lengthptr == NULL) - { - uschar *tc = code; - int condcount = 0; - - do { - condcount++; - tc += GET(tc,1); - } - while (*tc != OP_KET); - - /* A DEFINE group is never obeyed inline (the "condition" is always - false). It must have only one branch. */ - - if (code[LINK_SIZE+1] == OP_DEF) - { - if (condcount > 1) - { - *errorcodeptr = ERR54; - goto FAILED; - } - bravalue = OP_DEF; /* Just a flag to suppress char handling below */ - } - - /* A "normal" conditional group. If there is just one branch, we must not - make use of its firstbyte or reqbyte, because this is equivalent to an - empty second branch. */ - - else - { - if (condcount > 2) - { - *errorcodeptr = ERR27; - goto FAILED; - } - if (condcount == 1) subfirstbyte = subreqbyte = REQ_NONE; - } - } - - /* Error if hit end of pattern */ - - if (*ptr != ')') - { - *errorcodeptr = ERR14; - goto FAILED; - } - - /* In the pre-compile phase, update the length by the length of the group, - less the brackets at either end. Then reduce the compiled code to just a - set of non-capturing brackets so that it doesn't use much memory if it is - duplicated by a quantifier.*/ - - if (lengthptr != NULL) - { - if (OFLOW_MAX - *lengthptr < length_prevgroup - 2 - 2*LINK_SIZE) - { - *errorcodeptr = ERR20; - goto FAILED; - } - *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE; - *code++ = OP_BRA; - PUTINC(code, 0, 1 + LINK_SIZE); - *code++ = OP_KET; - PUTINC(code, 0, 1 + LINK_SIZE); - break; /* No need to waste time with special character handling */ - } - - /* Otherwise update the main code pointer to the end of the group. */ - - code = tempcode; - - /* For a DEFINE group, required and first character settings are not - relevant. */ - - if (bravalue == OP_DEF) break; - - /* Handle updating of the required and first characters for other types of - group. Update for normal brackets of all kinds, and conditions with two - branches (see code above). If the bracket is followed by a quantifier with - zero repeat, we have to back off. Hence the definition of zeroreqbyte and - zerofirstbyte outside the main loop so that they can be accessed for the - back off. */ - - zeroreqbyte = reqbyte; - zerofirstbyte = firstbyte; - groupsetfirstbyte = FALSE; - - if (bravalue >= OP_ONCE) - { - /* If we have not yet set a firstbyte in this branch, take it from the - subpattern, remembering that it was set here so that a repeat of more - than one can replicate it as reqbyte if necessary. If the subpattern has - no firstbyte, set "none" for the whole branch. In both cases, a zero - repeat forces firstbyte to "none". */ - - if (firstbyte == REQ_UNSET) - { - if (subfirstbyte >= 0) - { - firstbyte = subfirstbyte; - groupsetfirstbyte = TRUE; - } - else firstbyte = REQ_NONE; - zerofirstbyte = REQ_NONE; - } - - /* If firstbyte was previously set, convert the subpattern's firstbyte - into reqbyte if there wasn't one, using the vary flag that was in - existence beforehand. */ - - else if (subfirstbyte >= 0 && subreqbyte < 0) - subreqbyte = subfirstbyte | tempreqvary; - - /* If the subpattern set a required byte (or set a first byte that isn't - really the first byte - see above), set it. */ - - if (subreqbyte >= 0) reqbyte = subreqbyte; - } - - /* For a forward assertion, we take the reqbyte, if set. This can be - helpful if the pattern that follows the assertion doesn't set a different - char. For example, it's useful for /(?=abcde).+/. We can't set firstbyte - for an assertion, however because it leads to incorrect effect for patterns - such as /(?=a)a.+/ when the "real" "a" would then become a reqbyte instead - of a firstbyte. This is overcome by a scan at the end if there's no - firstbyte, looking for an asserted first char. */ - - else if (bravalue == OP_ASSERT && subreqbyte >= 0) reqbyte = subreqbyte; - break; /* End of processing '(' */ - - - /* ===================================================================*/ - /* Handle metasequences introduced by \. For ones like \d, the ESC_ values - are arranged to be the negation of the corresponding OP_values. For the - back references, the values are ESC_REF plus the reference number. Only - back references and those types that consume a character may be repeated. - We can test for values between ESC_b and ESC_Z for the latter; this may - have to change if any new ones are ever created. */ - - case '\\': - tempptr = ptr; - c = check_escape(&ptr, errorcodeptr, cd->bracount, options, FALSE); - if (*errorcodeptr != 0) goto FAILED; - - if (c < 0) - { - if (-c == ESC_Q) /* Handle start of quoted string */ - { - if (ptr[1] == '\\' && ptr[2] == 'E') ptr += 2; /* avoid empty string */ - else inescq = TRUE; - continue; - } - - if (-c == ESC_E) continue; /* Perl ignores an orphan \E */ - - /* For metasequences that actually match a character, we disable the - setting of a first character if it hasn't already been set. */ - - if (firstbyte == REQ_UNSET && -c > ESC_b && -c < ESC_Z) - firstbyte = REQ_NONE; - - /* Set values to reset to if this is followed by a zero repeat. */ - - zerofirstbyte = firstbyte; - zeroreqbyte = reqbyte; - - /* \g or \g'name' is a subroutine call by name and \g or \g'n' - is a subroutine call by number (Oniguruma syntax). In fact, the value - -ESC_g is returned only for these cases. So we don't need to check for < - or ' if the value is -ESC_g. For the Perl syntax \g{n} the value is - -ESC_REF+n, and for the Perl syntax \g{name} the result is -ESC_k (as - that is a synonym for a named back reference). */ - - if (-c == ESC_g) - { - const uschar *p; - save_hwm = cd->hwm; /* Normally this is set when '(' is read */ - terminator = (*(++ptr) == '<')? '>' : '\''; - - /* These two statements stop the compiler for warning about possibly - unset variables caused by the jump to HANDLE_NUMERICAL_RECURSION. In - fact, because we actually check for a number below, the paths that - would actually be in error are never taken. */ - - skipbytes = 0; - reset_bracount = FALSE; - - /* Test for a name */ - - if (ptr[1] != '+' && ptr[1] != '-') - { - BOOL isnumber = TRUE; - for (p = ptr + 1; *p != 0 && *p != terminator; p++) - { - if ((cd->ctypes[*p] & ctype_digit) == 0) isnumber = FALSE; - if ((cd->ctypes[*p] & ctype_word) == 0) break; - } - if (*p != terminator) - { - *errorcodeptr = ERR57; - break; - } - if (isnumber) - { - ptr++; - goto HANDLE_NUMERICAL_RECURSION; - } - is_recurse = TRUE; - goto NAMED_REF_OR_RECURSE; - } - - /* Test a signed number in angle brackets or quotes. */ - - p = ptr + 2; - while ((digitab[*p] & ctype_digit) != 0) p++; - if (*p != terminator) - { - *errorcodeptr = ERR57; - break; - } - ptr++; - goto HANDLE_NUMERICAL_RECURSION; - } - - /* \k or \k'name' is a back reference by name (Perl syntax). - We also support \k{name} (.NET syntax) */ - - if (-c == ESC_k && (ptr[1] == '<' || ptr[1] == '\'' || ptr[1] == '{')) - { - is_recurse = FALSE; - terminator = (*(++ptr) == '<')? '>' : (*ptr == '\'')? '\'' : '}'; - goto NAMED_REF_OR_RECURSE; - } - - /* Back references are handled specially; must disable firstbyte if - not set to cope with cases like (?=(\w+))\1: which would otherwise set - ':' later. */ - - if (-c >= ESC_REF) - { - recno = -c - ESC_REF; - - HANDLE_REFERENCE: /* Come here from named backref handling */ - if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE; - previous = code; - *code++ = OP_REF; - PUT2INC(code, 0, recno); - cd->backref_map |= (recno < 32)? (1 << recno) : 1; - if (recno > cd->top_backref) cd->top_backref = recno; - } - - /* So are Unicode property matches, if supported. */ - -#ifdef SUPPORT_UCP - else if (-c == ESC_P || -c == ESC_p) - { - BOOL negated; - int pdata; - int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr); - if (ptype < 0) goto FAILED; - previous = code; - *code++ = ((-c == ESC_p) != negated)? OP_PROP : OP_NOTPROP; - *code++ = ptype; - *code++ = pdata; - } -#else - - /* If Unicode properties are not supported, \X, \P, and \p are not - allowed. */ - - else if (-c == ESC_X || -c == ESC_P || -c == ESC_p) - { - *errorcodeptr = ERR45; - goto FAILED; - } -#endif - - /* For the rest (including \X when Unicode properties are supported), we - can obtain the OP value by negating the escape value. */ - - else - { - previous = (-c > ESC_b && -c < ESC_Z)? code : NULL; - *code++ = -c; - } - continue; - } - - /* We have a data character whose value is in c. In UTF-8 mode it may have - a value > 127. We set its representation in the length/buffer, and then - handle it as a data character. */ - -#ifdef SUPPORT_UTF8 - if (utf8 && c > 127) - mclength = _pcre_ord2utf8(c, mcbuffer); - else -#endif - - { - mcbuffer[0] = c; - mclength = 1; - } - goto ONE_CHAR; - - - /* ===================================================================*/ - /* Handle a literal character. It is guaranteed not to be whitespace or # - when the extended flag is set. If we are in UTF-8 mode, it may be a - multi-byte literal character. */ - - default: - NORMAL_CHAR: - mclength = 1; - mcbuffer[0] = c; - -#ifdef SUPPORT_UTF8 - if (utf8 && c >= 0xc0) - { - while ((ptr[1] & 0xc0) == 0x80) - mcbuffer[mclength++] = *(++ptr); - } -#endif - - /* At this point we have the character's bytes in mcbuffer, and the length - in mclength. When not in UTF-8 mode, the length is always 1. */ - - ONE_CHAR: - previous = code; - *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARNC : OP_CHAR; - for (c = 0; c < mclength; c++) *code++ = mcbuffer[c]; - - /* Remember if \r or \n were seen */ - - if (mcbuffer[0] == '\r' || mcbuffer[0] == '\n') - cd->external_flags |= PCRE_HASCRORLF; - - /* Set the first and required bytes appropriately. If no previous first - byte, set it from this character, but revert to none on a zero repeat. - Otherwise, leave the firstbyte value alone, and don't change it on a zero - repeat. */ - - if (firstbyte == REQ_UNSET) - { - zerofirstbyte = REQ_NONE; - zeroreqbyte = reqbyte; - - /* If the character is more than one byte long, we can set firstbyte - only if it is not to be matched caselessly. */ - - if (mclength == 1 || req_caseopt == 0) - { - firstbyte = mcbuffer[0] | req_caseopt; - if (mclength != 1) reqbyte = code[-1] | cd->req_varyopt; - } - else firstbyte = reqbyte = REQ_NONE; - } - - /* firstbyte was previously set; we can set reqbyte only the length is - 1 or the matching is caseful. */ - - else - { - zerofirstbyte = firstbyte; - zeroreqbyte = reqbyte; - if (mclength == 1 || req_caseopt == 0) - reqbyte = code[-1] | req_caseopt | cd->req_varyopt; - } - - break; /* End of literal character handling */ - } - } /* end of big loop */ - - -/* Control never reaches here by falling through, only by a goto for all the -error states. Pass back the position in the pattern so that it can be displayed -to the user for diagnosing the error. */ - -FAILED: -*ptrptr = ptr; -return FALSE; -} - - - - -/************************************************* -* Compile sequence of alternatives * -*************************************************/ - -/* On entry, ptr is pointing past the bracket character, but on return it -points to the closing bracket, or vertical bar, or end of string. The code -variable is pointing at the byte into which the BRA operator has been stored. -If the ims options are changed at the start (for a (?ims: group) or during any -branch, we need to insert an OP_OPT item at the start of every following branch -to ensure they get set correctly at run time, and also pass the new options -into every subsequent branch compile. - -This function is used during the pre-compile phase when we are trying to find -out the amount of memory needed, as well as during the real compile phase. The -value of lengthptr distinguishes the two phases. - -Arguments: - options option bits, including any changes for this subpattern - oldims previous settings of ims option bits - codeptr -> the address of the current code pointer - ptrptr -> the address of the current pattern pointer - errorcodeptr -> pointer to error code variable - lookbehind TRUE if this is a lookbehind assertion - reset_bracount TRUE to reset the count for each branch - skipbytes skip this many bytes at start (for brackets and OP_COND) - firstbyteptr place to put the first required character, or a negative number - reqbyteptr place to put the last required character, or a negative number - bcptr pointer to the chain of currently open branches - cd points to the data block with tables pointers etc. - lengthptr NULL during the real compile phase - points to length accumulator during pre-compile phase - -Returns: TRUE on success -*/ - -static BOOL -compile_regex(int options, int oldims, uschar **codeptr, const uschar **ptrptr, - int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes, - int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd, - int *lengthptr) -{ -const uschar *ptr = *ptrptr; -uschar *code = *codeptr; -uschar *last_branch = code; -uschar *start_bracket = code; -uschar *reverse_count = NULL; -int firstbyte, reqbyte; -int branchfirstbyte, branchreqbyte; -int length; -int orig_bracount; -int max_bracount; -branch_chain bc; - -bc.outer = bcptr; -bc.current = code; - -firstbyte = reqbyte = REQ_UNSET; - -/* Accumulate the length for use in the pre-compile phase. Start with the -length of the BRA and KET and any extra bytes that are required at the -beginning. We accumulate in a local variable to save frequent testing of -lenthptr for NULL. We cannot do this by looking at the value of code at the -start and end of each alternative, because compiled items are discarded during -the pre-compile phase so that the work space is not exceeded. */ - -length = 2 + 2*LINK_SIZE + skipbytes; - -/* WARNING: If the above line is changed for any reason, you must also change -the code that abstracts option settings at the start of the pattern and makes -them global. It tests the value of length for (2 + 2*LINK_SIZE) in the -pre-compile phase to find out whether anything has yet been compiled or not. */ - -/* Offset is set zero to mark that this bracket is still open */ - -PUT(code, 1, 0); -code += 1 + LINK_SIZE + skipbytes; - -/* Loop for each alternative branch */ - -orig_bracount = max_bracount = cd->bracount; -for (;;) - { - /* For a (?| group, reset the capturing bracket count so that each branch - uses the same numbers. */ - - if (reset_bracount) cd->bracount = orig_bracount; - - /* Handle a change of ims options at the start of the branch */ - - if ((options & PCRE_IMS) != oldims) - { - *code++ = OP_OPT; - *code++ = options & PCRE_IMS; - length += 2; - } - - /* Set up dummy OP_REVERSE if lookbehind assertion */ - - if (lookbehind) - { - *code++ = OP_REVERSE; - reverse_count = code; - PUTINC(code, 0, 0); - length += 1 + LINK_SIZE; - } - - /* Now compile the branch; in the pre-compile phase its length gets added - into the length. */ - - if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstbyte, - &branchreqbyte, &bc, cd, (lengthptr == NULL)? NULL : &length)) - { - *ptrptr = ptr; - return FALSE; - } - - /* Keep the highest bracket count in case (?| was used and some branch - has fewer than the rest. */ - - if (cd->bracount > max_bracount) max_bracount = cd->bracount; - - /* In the real compile phase, there is some post-processing to be done. */ - - if (lengthptr == NULL) - { - /* If this is the first branch, the firstbyte and reqbyte values for the - branch become the values for the regex. */ - - if (*last_branch != OP_ALT) - { - firstbyte = branchfirstbyte; - reqbyte = branchreqbyte; - } - - /* If this is not the first branch, the first char and reqbyte have to - match the values from all the previous branches, except that if the - previous value for reqbyte didn't have REQ_VARY set, it can still match, - and we set REQ_VARY for the regex. */ - - else - { - /* If we previously had a firstbyte, but it doesn't match the new branch, - we have to abandon the firstbyte for the regex, but if there was - previously no reqbyte, it takes on the value of the old firstbyte. */ - - if (firstbyte >= 0 && firstbyte != branchfirstbyte) - { - if (reqbyte < 0) reqbyte = firstbyte; - firstbyte = REQ_NONE; - } - - /* If we (now or from before) have no firstbyte, a firstbyte from the - branch becomes a reqbyte if there isn't a branch reqbyte. */ - - if (firstbyte < 0 && branchfirstbyte >= 0 && branchreqbyte < 0) - branchreqbyte = branchfirstbyte; - - /* Now ensure that the reqbytes match */ - - if ((reqbyte & ~REQ_VARY) != (branchreqbyte & ~REQ_VARY)) - reqbyte = REQ_NONE; - else reqbyte |= branchreqbyte; /* To "or" REQ_VARY */ - } - - /* If lookbehind, check that this branch matches a fixed-length string, and - put the length into the OP_REVERSE item. Temporarily mark the end of the - branch with OP_END. */ - - if (lookbehind) - { - int fixed_length; - *code = OP_END; - fixed_length = find_fixedlength(last_branch, options); - DPRINTF(("fixed length = %d\n", fixed_length)); - if (fixed_length < 0) - { - *errorcodeptr = (fixed_length == -2)? ERR36 : ERR25; - *ptrptr = ptr; - return FALSE; - } - PUT(reverse_count, 0, fixed_length); - } - } - - /* Reached end of expression, either ')' or end of pattern. In the real - compile phase, go back through the alternative branches and reverse the chain - of offsets, with the field in the BRA item now becoming an offset to the - first alternative. If there are no alternatives, it points to the end of the - group. The length in the terminating ket is always the length of the whole - bracketed item. If any of the ims options were changed inside the group, - compile a resetting op-code following, except at the very end of the pattern. - Return leaving the pointer at the terminating char. */ - - if (*ptr != '|') - { - if (lengthptr == NULL) - { - int branch_length = code - last_branch; - do - { - int prev_length = GET(last_branch, 1); - PUT(last_branch, 1, branch_length); - branch_length = prev_length; - last_branch -= branch_length; - } - while (branch_length > 0); - } - - /* Fill in the ket */ - - *code = OP_KET; - PUT(code, 1, code - start_bracket); - code += 1 + LINK_SIZE; - - /* Resetting option if needed */ - - if ((options & PCRE_IMS) != oldims && *ptr == ')') - { - *code++ = OP_OPT; - *code++ = oldims; - length += 2; - } - - /* Retain the highest bracket number, in case resetting was used. */ - - cd->bracount = max_bracount; - - /* Set values to pass back */ - - *codeptr = code; - *ptrptr = ptr; - *firstbyteptr = firstbyte; - *reqbyteptr = reqbyte; - if (lengthptr != NULL) - { - if (OFLOW_MAX - *lengthptr < length) - { - *errorcodeptr = ERR20; - return FALSE; - } - *lengthptr += length; - } - return TRUE; - } - - /* Another branch follows. In the pre-compile phase, we can move the code - pointer back to where it was for the start of the first branch. (That is, - pretend that each branch is the only one.) - - In the real compile phase, insert an ALT node. Its length field points back - to the previous branch while the bracket remains open. At the end the chain - is reversed. It's done like this so that the start of the bracket has a - zero offset until it is closed, making it possible to detect recursion. */ - - if (lengthptr != NULL) - { - code = *codeptr + 1 + LINK_SIZE + skipbytes; - length += 1 + LINK_SIZE; - } - else - { - *code = OP_ALT; - PUT(code, 1, code - last_branch); - bc.current = last_branch = code; - code += 1 + LINK_SIZE; - } - - ptr++; - } -/* Control never reaches here */ -} - - - - -/************************************************* -* Check for anchored expression * -*************************************************/ - -/* Try to find out if this is an anchored regular expression. Consider each -alternative branch. If they all start with OP_SOD or OP_CIRC, or with a bracket -all of whose alternatives start with OP_SOD or OP_CIRC (recurse ad lib), then -it's anchored. However, if this is a multiline pattern, then only OP_SOD -counts, since OP_CIRC can match in the middle. - -We can also consider a regex to be anchored if OP_SOM starts all its branches. -This is the code for \G, which means "match at start of match position, taking -into account the match offset". - -A branch is also implicitly anchored if it starts with .* and DOTALL is set, -because that will try the rest of the pattern at all possible matching points, -so there is no point trying again.... er .... - -.... except when the .* appears inside capturing parentheses, and there is a -subsequent back reference to those parentheses. We haven't enough information -to catch that case precisely. - -At first, the best we could do was to detect when .* was in capturing brackets -and the highest back reference was greater than or equal to that level. -However, by keeping a bitmap of the first 31 back references, we can catch some -of the more common cases more precisely. - -Arguments: - code points to start of expression (the bracket) - options points to the options setting - bracket_map a bitmap of which brackets we are inside while testing; this - handles up to substring 31; after that we just have to take - the less precise approach - backref_map the back reference bitmap - -Returns: TRUE or FALSE -*/ - -static BOOL -is_anchored(register const uschar *code, int *options, unsigned int bracket_map, - unsigned int backref_map) -{ -do { - const uschar *scode = first_significant_code(code + _pcre_OP_lengths[*code], - options, PCRE_MULTILINE, FALSE); - register int op = *scode; - - /* Non-capturing brackets */ - - if (op == OP_BRA) - { - if (!is_anchored(scode, options, bracket_map, backref_map)) return FALSE; - } - - /* Capturing brackets */ - - else if (op == OP_CBRA) - { - int n = GET2(scode, 1+LINK_SIZE); - int new_map = bracket_map | ((n < 32)? (1 << n) : 1); - if (!is_anchored(scode, options, new_map, backref_map)) return FALSE; - } - - /* Other brackets */ - - else if (op == OP_ASSERT || op == OP_ONCE || op == OP_COND) - { - if (!is_anchored(scode, options, bracket_map, backref_map)) return FALSE; - } - - /* .* is not anchored unless DOTALL is set (which generates OP_ALLANY) and - it isn't in brackets that are or may be referenced. */ - - else if ((op == OP_TYPESTAR || op == OP_TYPEMINSTAR || - op == OP_TYPEPOSSTAR)) - { - if (scode[1] != OP_ALLANY || (bracket_map & backref_map) != 0) - return FALSE; - } - - /* Check for explicit anchoring */ - - else if (op != OP_SOD && op != OP_SOM && - ((*options & PCRE_MULTILINE) != 0 || op != OP_CIRC)) - return FALSE; - code += GET(code, 1); - } -while (*code == OP_ALT); /* Loop for each alternative */ -return TRUE; -} - - - -/************************************************* -* Check for starting with ^ or .* * -*************************************************/ - -/* This is called to find out if every branch starts with ^ or .* so that -"first char" processing can be done to speed things up in multiline -matching and for non-DOTALL patterns that start with .* (which must start at -the beginning or after \n). As in the case of is_anchored() (see above), we -have to take account of back references to capturing brackets that contain .* -because in that case we can't make the assumption. - -Arguments: - code points to start of expression (the bracket) - bracket_map a bitmap of which brackets we are inside while testing; this - handles up to substring 31; after that we just have to take - the less precise approach - backref_map the back reference bitmap - -Returns: TRUE or FALSE -*/ - -static BOOL -is_startline(const uschar *code, unsigned int bracket_map, - unsigned int backref_map) -{ -do { - const uschar *scode = first_significant_code(code + _pcre_OP_lengths[*code], - NULL, 0, FALSE); - register int op = *scode; - - /* Non-capturing brackets */ - - if (op == OP_BRA) - { - if (!is_startline(scode, bracket_map, backref_map)) return FALSE; - } - - /* Capturing brackets */ - - else if (op == OP_CBRA) - { - int n = GET2(scode, 1+LINK_SIZE); - int new_map = bracket_map | ((n < 32)? (1 << n) : 1); - if (!is_startline(scode, new_map, backref_map)) return FALSE; - } - - /* Other brackets */ - - else if (op == OP_ASSERT || op == OP_ONCE || op == OP_COND) - { if (!is_startline(scode, bracket_map, backref_map)) return FALSE; } - - /* .* means "start at start or after \n" if it isn't in brackets that - may be referenced. */ - - else if (op == OP_TYPESTAR || op == OP_TYPEMINSTAR || op == OP_TYPEPOSSTAR) - { - if (scode[1] != OP_ANY || (bracket_map & backref_map) != 0) return FALSE; - } - - /* Check for explicit circumflex */ - - else if (op != OP_CIRC) return FALSE; - - /* Move on to the next alternative */ - - code += GET(code, 1); - } -while (*code == OP_ALT); /* Loop for each alternative */ -return TRUE; -} - - - -/************************************************* -* Check for asserted fixed first char * -*************************************************/ - -/* During compilation, the "first char" settings from forward assertions are -discarded, because they can cause conflicts with actual literals that follow. -However, if we end up without a first char setting for an unanchored pattern, -it is worth scanning the regex to see if there is an initial asserted first -char. If all branches start with the same asserted char, or with a bracket all -of whose alternatives start with the same asserted char (recurse ad lib), then -we return that char, otherwise -1. - -Arguments: - code points to start of expression (the bracket) - options pointer to the options (used to check casing changes) - inassert TRUE if in an assertion - -Returns: -1 or the fixed first char -*/ - -static int -find_firstassertedchar(const uschar *code, int *options, BOOL inassert) -{ -register int c = -1; -do { - int d; - const uschar *scode = - first_significant_code(code + 1+LINK_SIZE, options, PCRE_CASELESS, TRUE); - register int op = *scode; - - switch(op) - { - default: - return -1; - - case OP_BRA: - case OP_CBRA: - case OP_ASSERT: - case OP_ONCE: - case OP_COND: - if ((d = find_firstassertedchar(scode, options, op == OP_ASSERT)) < 0) - return -1; - if (c < 0) c = d; else if (c != d) return -1; - break; - - case OP_EXACT: /* Fall through */ - scode += 2; - - case OP_CHAR: - case OP_CHARNC: - case OP_PLUS: - case OP_MINPLUS: - case OP_POSPLUS: - if (!inassert) return -1; - if (c < 0) - { - c = scode[1]; - if ((*options & PCRE_CASELESS) != 0) c |= REQ_CASELESS; - } - else if (c != scode[1]) return -1; - break; - } - - code += GET(code, 1); - } -while (*code == OP_ALT); -return c; -} - - - -/************************************************* -* Compile a Regular Expression * -*************************************************/ - -/* This function takes a string and returns a pointer to a block of store -holding a compiled version of the expression. The original API for this -function had no error code return variable; it is retained for backwards -compatibility. The new function is given a new name. - -Arguments: - pattern the regular expression - options various option bits - errorcodeptr pointer to error code variable (pcre_compile2() only) - can be NULL if you don't want a code value - errorptr pointer to pointer to error text - erroroffset ptr offset in pattern where error was detected - tables pointer to character tables or NULL - -Returns: pointer to compiled data block, or NULL on error, - with errorptr and erroroffset set -*/ - -PCRE_EXP_DEFN pcre * -pcre_compile(const char *pattern, int options, const char **errorptr, - int *erroroffset, const unsigned char *tables) -{ -return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables); -} - - -PCRE_EXP_DEFN pcre * -pcre_compile2(const char *pattern, int options, int *errorcodeptr, - const char **errorptr, int *erroroffset, const unsigned char *tables) -{ -real_pcre *re; -int length = 1; /* For final END opcode */ -int firstbyte, reqbyte, newline; -int errorcode = 0; -int skipatstart = 0; -#ifdef SUPPORT_UTF8 -BOOL utf8; -#endif -size_t size; -uschar *code; -const uschar *codestart; -const uschar *ptr; -compile_data compile_block; -compile_data *cd = &compile_block; - -/* This space is used for "compiling" into during the first phase, when we are -computing the amount of memory that is needed. Compiled items are thrown away -as soon as possible, so that a fairly large buffer should be sufficient for -this purpose. The same space is used in the second phase for remembering where -to fill in forward references to subpatterns. */ - -uschar cworkspace[COMPILE_WORK_SIZE]; - -/* Set this early so that early errors get offset 0. */ - -ptr = (const uschar *)pattern; - -/* We can't pass back an error message if errorptr is NULL; I guess the best we -can do is just return NULL, but we can set a code value if there is a code -pointer. */ - -if (errorptr == NULL) - { - if (errorcodeptr != NULL) *errorcodeptr = 99; - return NULL; - } - -*errorptr = NULL; -if (errorcodeptr != NULL) *errorcodeptr = ERR0; - -/* However, we can give a message for this error */ - -if (erroroffset == NULL) - { - errorcode = ERR16; - goto PCRE_EARLY_ERROR_RETURN2; - } - -*erroroffset = 0; - -/* Can't support UTF8 unless PCRE has been compiled to include the code. */ - -#ifdef SUPPORT_UTF8 -utf8 = (options & PCRE_UTF8) != 0; -if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0 && - (*erroroffset = _pcre_valid_utf8((uschar *)pattern, -1)) >= 0) - { - errorcode = ERR44; - goto PCRE_EARLY_ERROR_RETURN2; - } -#else -if ((options & PCRE_UTF8) != 0) - { - errorcode = ERR32; - goto PCRE_EARLY_ERROR_RETURN; - } -#endif - -if ((options & ~PUBLIC_OPTIONS) != 0) - { - errorcode = ERR17; - goto PCRE_EARLY_ERROR_RETURN; - } - -/* Set up pointers to the individual character tables */ - -if (tables == NULL) tables = _pcre_default_tables; -cd->lcc = tables + lcc_offset; -cd->fcc = tables + fcc_offset; -cd->cbits = tables + cbits_offset; -cd->ctypes = tables + ctypes_offset; - -/* Check for global one-time settings at the start of the pattern, and remember -the offset for later. */ - -while (ptr[skipatstart] == '(' && ptr[skipatstart+1] == '*') - { - int newnl = 0; - int newbsr = 0; - - if (strncmp((char *)(ptr+skipatstart+2), "CR)", 3) == 0) - { skipatstart += 5; newnl = PCRE_NEWLINE_CR; } - else if (strncmp((char *)(ptr+skipatstart+2), "LF)", 3) == 0) - { skipatstart += 5; newnl = PCRE_NEWLINE_LF; } - else if (strncmp((char *)(ptr+skipatstart+2), "CRLF)", 5) == 0) - { skipatstart += 7; newnl = PCRE_NEWLINE_CR + PCRE_NEWLINE_LF; } - else if (strncmp((char *)(ptr+skipatstart+2), "ANY)", 4) == 0) - { skipatstart += 6; newnl = PCRE_NEWLINE_ANY; } - else if (strncmp((char *)(ptr+skipatstart+2), "ANYCRLF)", 8) == 0) - { skipatstart += 10; newnl = PCRE_NEWLINE_ANYCRLF; } - - else if (strncmp((char *)(ptr+skipatstart+2), "BSR_ANYCRLF)", 12) == 0) - { skipatstart += 14; newbsr = PCRE_BSR_ANYCRLF; } - else if (strncmp((char *)(ptr+skipatstart+2), "BSR_UNICODE)", 12) == 0) - { skipatstart += 14; newbsr = PCRE_BSR_UNICODE; } - - if (newnl != 0) - options = (options & ~PCRE_NEWLINE_BITS) | newnl; - else if (newbsr != 0) - options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr; - else break; - } - -/* Check validity of \R options. */ - -switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) - { - case 0: - case PCRE_BSR_ANYCRLF: - case PCRE_BSR_UNICODE: - break; - default: errorcode = ERR56; goto PCRE_EARLY_ERROR_RETURN; - } - -/* Handle different types of newline. The three bits give seven cases. The -current code allows for fixed one- or two-byte sequences, plus "any" and -"anycrlf". */ - -switch (options & PCRE_NEWLINE_BITS) - { - case 0: newline = NEWLINE; break; /* Build-time default */ - case PCRE_NEWLINE_CR: newline = '\r'; break; - case PCRE_NEWLINE_LF: newline = '\n'; break; - case PCRE_NEWLINE_CR+ - PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break; - case PCRE_NEWLINE_ANY: newline = -1; break; - case PCRE_NEWLINE_ANYCRLF: newline = -2; break; - default: errorcode = ERR56; goto PCRE_EARLY_ERROR_RETURN; - } - -if (newline == -2) - { - cd->nltype = NLTYPE_ANYCRLF; - } -else if (newline < 0) - { - cd->nltype = NLTYPE_ANY; - } -else - { - cd->nltype = NLTYPE_FIXED; - if (newline > 255) - { - cd->nllen = 2; - cd->nl[0] = (newline >> 8) & 255; - cd->nl[1] = newline & 255; - } - else - { - cd->nllen = 1; - cd->nl[0] = newline; - } - } - -/* Maximum back reference and backref bitmap. The bitmap records up to 31 back -references to help in deciding whether (.*) can be treated as anchored or not. -*/ - -cd->top_backref = 0; -cd->backref_map = 0; - -/* Reflect pattern for debugging output */ - -DPRINTF(("------------------------------------------------------------------\n")); -DPRINTF(("%s\n", pattern)); - -/* Pretend to compile the pattern while actually just accumulating the length -of memory required. This behaviour is triggered by passing a non-NULL final -argument to compile_regex(). We pass a block of workspace (cworkspace) for it -to compile parts of the pattern into; the compiled code is discarded when it is -no longer needed, so hopefully this workspace will never overflow, though there -is a test for its doing so. */ - -cd->bracount = cd->final_bracount = 0; -cd->names_found = 0; -cd->name_entry_size = 0; -cd->name_table = NULL; -cd->start_workspace = cworkspace; -cd->start_code = cworkspace; -cd->hwm = cworkspace; -cd->start_pattern = (const uschar *)pattern; -cd->end_pattern = (const uschar *)(pattern + strlen(pattern)); -cd->req_varyopt = 0; -cd->external_options = options; -cd->external_flags = 0; - -/* Now do the pre-compile. On error, errorcode will be set non-zero, so we -don't need to look at the result of the function here. The initial options have -been put into the cd block so that they can be changed if an option setting is -found within the regex right at the beginning. Bringing initial option settings -outside can help speed up starting point checks. */ - -ptr += skipatstart; -code = cworkspace; -*code = OP_BRA; -(void)compile_regex(cd->external_options, cd->external_options & PCRE_IMS, - &code, &ptr, &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd, - &length); -if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN; - -DPRINTF(("end pre-compile: length=%d workspace=%d\n", length, - cd->hwm - cworkspace)); - -if (length > MAX_PATTERN_SIZE) - { - errorcode = ERR20; - goto PCRE_EARLY_ERROR_RETURN; - } - -/* Compute the size of data block needed and get it, either from malloc or -externally provided function. Integer overflow should no longer be possible -because nowadays we limit the maximum value of cd->names_found and -cd->name_entry_size. */ - -size = length + sizeof(real_pcre) + cd->names_found * (cd->name_entry_size + 3); -re = (real_pcre *)(pcre_malloc)(size); - -if (re == NULL) - { - errorcode = ERR21; - goto PCRE_EARLY_ERROR_RETURN; - } - -/* Put in the magic number, and save the sizes, initial options, internal -flags, and character table pointer. NULL is used for the default character -tables. The nullpad field is at the end; it's there to help in the case when a -regex compiled on a system with 4-byte pointers is run on another with 8-byte -pointers. */ - -re->magic_number = MAGIC_NUMBER; -re->size = size; -re->options = cd->external_options; -re->flags = cd->external_flags; -re->dummy1 = 0; -re->first_byte = 0; -re->req_byte = 0; -re->name_table_offset = sizeof(real_pcre); -re->name_entry_size = cd->name_entry_size; -re->name_count = cd->names_found; -re->ref_count = 0; -re->tables = (tables == _pcre_default_tables)? NULL : tables; -re->nullpad = NULL; - -/* The starting points of the name/number translation table and of the code are -passed around in the compile data block. The start/end pattern and initial -options are already set from the pre-compile phase, as is the name_entry_size -field. Reset the bracket count and the names_found field. Also reset the hwm -field; this time it's used for remembering forward references to subpatterns. -*/ - -cd->final_bracount = cd->bracount; /* Save for checking forward references */ -cd->bracount = 0; -cd->names_found = 0; -cd->name_table = (uschar *)re + re->name_table_offset; -codestart = cd->name_table + re->name_entry_size * re->name_count; -cd->start_code = codestart; -cd->hwm = cworkspace; -cd->req_varyopt = 0; -cd->had_accept = FALSE; - -/* Set up a starting, non-extracting bracket, then compile the expression. On -error, errorcode will be set non-zero, so we don't need to look at the result -of the function here. */ - -ptr = (const uschar *)pattern + skipatstart; -code = (uschar *)codestart; -*code = OP_BRA; -(void)compile_regex(re->options, re->options & PCRE_IMS, &code, &ptr, - &errorcode, FALSE, FALSE, 0, &firstbyte, &reqbyte, NULL, cd, NULL); -re->top_bracket = cd->bracount; -re->top_backref = cd->top_backref; -re->flags = cd->external_flags; - -if (cd->had_accept) reqbyte = -1; /* Must disable after (*ACCEPT) */ - -/* If not reached end of pattern on success, there's an excess bracket. */ - -if (errorcode == 0 && *ptr != 0) errorcode = ERR22; - -/* Fill in the terminating state and check for disastrous overflow, but -if debugging, leave the test till after things are printed out. */ - -*code++ = OP_END; - -#ifndef DEBUG -if (code - codestart > length) errorcode = ERR23; -#endif - -/* Fill in any forward references that are required. */ - -while (errorcode == 0 && cd->hwm > cworkspace) - { - int offset, recno; - const uschar *groupptr; - cd->hwm -= LINK_SIZE; - offset = GET(cd->hwm, 0); - recno = GET(codestart, offset); - groupptr = find_bracket(codestart, (re->options & PCRE_UTF8) != 0, recno); - if (groupptr == NULL) errorcode = ERR53; - else PUT(((uschar *)codestart), offset, groupptr - codestart); - } - -/* Give an error if there's back reference to a non-existent capturing -subpattern. */ - -if (errorcode == 0 && re->top_backref > re->top_bracket) errorcode = ERR15; - -/* Failed to compile, or error while post-processing */ - -if (errorcode != 0) - { - (pcre_free)(re); - PCRE_EARLY_ERROR_RETURN: - *erroroffset = ptr - (const uschar *)pattern; - PCRE_EARLY_ERROR_RETURN2: - *errorptr = find_error_text(errorcode); - if (errorcodeptr != NULL) *errorcodeptr = errorcode; - return NULL; - } - -/* If the anchored option was not passed, set the flag if we can determine that -the pattern is anchored by virtue of ^ characters or \A or anything else (such -as starting with .* when DOTALL is set). - -Otherwise, if we know what the first byte has to be, save it, because that -speeds up unanchored matches no end. If not, see if we can set the -PCRE_STARTLINE flag. This is helpful for multiline matches when all branches -start with ^. and also when all branches start with .* for non-DOTALL matches. -*/ - -if ((re->options & PCRE_ANCHORED) == 0) - { - int temp_options = re->options; /* May get changed during these scans */ - if (is_anchored(codestart, &temp_options, 0, cd->backref_map)) - re->options |= PCRE_ANCHORED; - else - { - if (firstbyte < 0) - firstbyte = find_firstassertedchar(codestart, &temp_options, FALSE); - if (firstbyte >= 0) /* Remove caseless flag for non-caseable chars */ - { - int ch = firstbyte & 255; - re->first_byte = ((firstbyte & REQ_CASELESS) != 0 && - cd->fcc[ch] == ch)? ch : firstbyte; - re->flags |= PCRE_FIRSTSET; - } - else if (is_startline(codestart, 0, cd->backref_map)) - re->flags |= PCRE_STARTLINE; - } - } - -/* For an anchored pattern, we use the "required byte" only if it follows a -variable length item in the regex. Remove the caseless flag for non-caseable -bytes. */ - -if (reqbyte >= 0 && - ((re->options & PCRE_ANCHORED) == 0 || (reqbyte & REQ_VARY) != 0)) - { - int ch = reqbyte & 255; - re->req_byte = ((reqbyte & REQ_CASELESS) != 0 && - cd->fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte; - re->flags |= PCRE_REQCHSET; - } - -/* Print out the compiled data if debugging is enabled. This is never the -case when building a production library. */ - -#ifdef DEBUG - -printf("Length = %d top_bracket = %d top_backref = %d\n", - length, re->top_bracket, re->top_backref); - -printf("Options=%08x\n", re->options); - -if ((re->flags & PCRE_FIRSTSET) != 0) - { - int ch = re->first_byte & 255; - const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)? - "" : " (caseless)"; - if (isprint(ch)) printf("First char = %c%s\n", ch, caseless); - else printf("First char = \\x%02x%s\n", ch, caseless); - } - -if ((re->flags & PCRE_REQCHSET) != 0) - { - int ch = re->req_byte & 255; - const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)? - "" : " (caseless)"; - if (isprint(ch)) printf("Req char = %c%s\n", ch, caseless); - else printf("Req char = \\x%02x%s\n", ch, caseless); - } - -pcre_printint(re, stdout, TRUE); - -/* This check is done here in the debugging case so that the code that -was compiled can be seen. */ - -if (code - codestart > length) - { - (pcre_free)(re); - *errorptr = find_error_text(ERR23); - *erroroffset = ptr - (uschar *)pattern; - if (errorcodeptr != NULL) *errorcodeptr = ERR23; - return NULL; - } -#endif /* DEBUG */ - -return (pcre *)re; + Copyright (c) 1997-2012 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + + +/* This module contains the external function pcre_compile(), along with +supporting internal functions that are not used by other modules. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#define NLBLOCK cd /* Block containing newline information */ +#define PSSTART start_pattern /* Field containing processed string start */ +#define PSEND end_pattern /* Field containing processed string end */ + +#include "pcre_internal.h" + + +/* When PCRE_DEBUG is defined, we need the pcre(16)_printint() function, which +is also used by pcretest. PCRE_DEBUG is not defined when building a production +library. We do not need to select pcre16_printint.c specially, because the +COMPILE_PCREx macro will already be appropriately set. */ + +#ifdef PCRE_DEBUG +/* pcre_printint.c should not include any headers */ +#define PCRE_INCLUDED +#include "pcre_printint.c" +#undef PCRE_INCLUDED +#endif + + +/* Macro for setting individual bits in class bitmaps. */ + +#define SETBIT(a,b) a[b/8] |= (1 << (b%8)) + +/* Maximum length value to check against when making sure that the integer that +holds the compiled pattern length does not overflow. We make it a bit less than +INT_MAX to allow for adding in group terminating bytes, so that we don't have +to check them every time. */ + +#define OFLOW_MAX (INT_MAX - 20) + + +/************************************************* +* Code parameters and static tables * +*************************************************/ + +/* This value specifies the size of stack workspace that is used during the +first pre-compile phase that determines how much memory is required. The regex +is partly compiled into this space, but the compiled parts are discarded as +soon as they can be, so that hopefully there will never be an overrun. The code +does, however, check for an overrun. The largest amount I've seen used is 218, +so this number is very generous. + +The same workspace is used during the second, actual compile phase for +remembering forward references to groups so that they can be filled in at the +end. Each entry in this list occupies LINK_SIZE bytes, so even when LINK_SIZE +is 4 there is plenty of room for most patterns. However, the memory can get +filled up by repetitions of forward references, for example patterns like +/(?1){0,1999}(b)/, and one user did hit the limit. The code has been changed so +that the workspace is expanded using malloc() in this situation. The value +below is therefore a minimum, and we put a maximum on it for safety. The +minimum is now also defined in terms of LINK_SIZE so that the use of malloc() +kicks in at the same number of forward references in all cases. */ + +#define COMPILE_WORK_SIZE (2048*LINK_SIZE) +#define COMPILE_WORK_SIZE_MAX (100*COMPILE_WORK_SIZE) + +/* The overrun tests check for a slightly smaller size so that they detect the +overrun before it actually does run off the end of the data block. */ + +#define WORK_SIZE_SAFETY_MARGIN (100) + +/* Private flags added to firstchar and reqchar. */ + +#define REQ_CASELESS 0x10000000l /* Indicates caselessness */ +#define REQ_VARY 0x20000000l /* Reqchar followed non-literal item */ + +/* Repeated character flags. */ + +#define UTF_LENGTH 0x10000000l /* The char contains its length. */ + +/* Table for handling escaped characters in the range '0'-'z'. Positive returns +are simple data values; negative values are for special things like \d and so +on. Zero means further processing is needed (for things like \x), or the escape +is invalid. */ + +#ifndef EBCDIC + +/* This is the "normal" table for ASCII systems or for EBCDIC systems running +in UTF-8 mode. */ + +static const short int escapes[] = { + 0, 0, + 0, 0, + 0, 0, + 0, 0, + 0, 0, + CHAR_COLON, CHAR_SEMICOLON, + CHAR_LESS_THAN_SIGN, CHAR_EQUALS_SIGN, + CHAR_GREATER_THAN_SIGN, CHAR_QUESTION_MARK, + CHAR_COMMERCIAL_AT, -ESC_A, + -ESC_B, -ESC_C, + -ESC_D, -ESC_E, + 0, -ESC_G, + -ESC_H, 0, + 0, -ESC_K, + 0, 0, + -ESC_N, 0, + -ESC_P, -ESC_Q, + -ESC_R, -ESC_S, + 0, 0, + -ESC_V, -ESC_W, + -ESC_X, 0, + -ESC_Z, CHAR_LEFT_SQUARE_BRACKET, + CHAR_BACKSLASH, CHAR_RIGHT_SQUARE_BRACKET, + CHAR_CIRCUMFLEX_ACCENT, CHAR_UNDERSCORE, + CHAR_GRAVE_ACCENT, 7, + -ESC_b, 0, + -ESC_d, ESC_e, + ESC_f, 0, + -ESC_h, 0, + 0, -ESC_k, + 0, 0, + ESC_n, 0, + -ESC_p, 0, + ESC_r, -ESC_s, + ESC_tee, 0, + -ESC_v, -ESC_w, + 0, 0, + -ESC_z +}; + +#else + +/* This is the "abnormal" table for EBCDIC systems without UTF-8 support. */ + +static const short int escapes[] = { +/* 48 */ 0, 0, 0, '.', '<', '(', '+', '|', +/* 50 */ '&', 0, 0, 0, 0, 0, 0, 0, +/* 58 */ 0, 0, '!', '$', '*', ')', ';', '~', +/* 60 */ '-', '/', 0, 0, 0, 0, 0, 0, +/* 68 */ 0, 0, '|', ',', '%', '_', '>', '?', +/* 70 */ 0, 0, 0, 0, 0, 0, 0, 0, +/* 78 */ 0, '`', ':', '#', '@', '\'', '=', '"', +/* 80 */ 0, 7, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0, +/* 88 */-ESC_h, 0, 0, '{', 0, 0, 0, 0, +/* 90 */ 0, 0, -ESC_k, 'l', 0, ESC_n, 0, -ESC_p, +/* 98 */ 0, ESC_r, 0, '}', 0, 0, 0, 0, +/* A0 */ 0, '~', -ESC_s, ESC_tee, 0,-ESC_v, -ESC_w, 0, +/* A8 */ 0,-ESC_z, 0, 0, 0, '[', 0, 0, +/* B0 */ 0, 0, 0, 0, 0, 0, 0, 0, +/* B8 */ 0, 0, 0, 0, 0, ']', '=', '-', +/* C0 */ '{',-ESC_A, -ESC_B, -ESC_C, -ESC_D,-ESC_E, 0, -ESC_G, +/* C8 */-ESC_H, 0, 0, 0, 0, 0, 0, 0, +/* D0 */ '}', 0, -ESC_K, 0, 0,-ESC_N, 0, -ESC_P, +/* D8 */-ESC_Q,-ESC_R, 0, 0, 0, 0, 0, 0, +/* E0 */ '\\', 0, -ESC_S, 0, 0,-ESC_V, -ESC_W, -ESC_X, +/* E8 */ 0,-ESC_Z, 0, 0, 0, 0, 0, 0, +/* F0 */ 0, 0, 0, 0, 0, 0, 0, 0, +/* F8 */ 0, 0, 0, 0, 0, 0, 0, 0 +}; +#endif + + +/* Table of special "verbs" like (*PRUNE). This is a short table, so it is +searched linearly. Put all the names into a single string, in order to reduce +the number of relocations when a shared library is dynamically linked. The +string is built from string macros so that it works in UTF-8 mode on EBCDIC +platforms. */ + +typedef struct verbitem { + int len; /* Length of verb name */ + int op; /* Op when no arg, or -1 if arg mandatory */ + int op_arg; /* Op when arg present, or -1 if not allowed */ +} verbitem; + +static const char verbnames[] = + "\0" /* Empty name is a shorthand for MARK */ + STRING_MARK0 + STRING_ACCEPT0 + STRING_COMMIT0 + STRING_F0 + STRING_FAIL0 + STRING_PRUNE0 + STRING_SKIP0 + STRING_THEN; + +static const verbitem verbs[] = { + { 0, -1, OP_MARK }, + { 4, -1, OP_MARK }, + { 6, OP_ACCEPT, -1 }, + { 6, OP_COMMIT, -1 }, + { 1, OP_FAIL, -1 }, + { 4, OP_FAIL, -1 }, + { 5, OP_PRUNE, OP_PRUNE_ARG }, + { 4, OP_SKIP, OP_SKIP_ARG }, + { 4, OP_THEN, OP_THEN_ARG } +}; + +static const int verbcount = sizeof(verbs)/sizeof(verbitem); + + +/* Tables of names of POSIX character classes and their lengths. The names are +now all in a single string, to reduce the number of relocations when a shared +library is dynamically loaded. The list of lengths is terminated by a zero +length entry. The first three must be alpha, lower, upper, as this is assumed +for handling case independence. */ + +static const char posix_names[] = + STRING_alpha0 STRING_lower0 STRING_upper0 STRING_alnum0 + STRING_ascii0 STRING_blank0 STRING_cntrl0 STRING_digit0 + STRING_graph0 STRING_print0 STRING_punct0 STRING_space0 + STRING_word0 STRING_xdigit; + +static const pcre_uint8 posix_name_lengths[] = { + 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 6, 0 }; + +/* Table of class bit maps for each POSIX class. Each class is formed from a +base map, with an optional addition or removal of another map. Then, for some +classes, there is some additional tweaking: for [:blank:] the vertical space +characters are removed, and for [:alpha:] and [:alnum:] the underscore +character is removed. The triples in the table consist of the base map offset, +second map offset or -1 if no second map, and a non-negative value for map +addition or a negative value for map subtraction (if there are two maps). The +absolute value of the third field has these meanings: 0 => no tweaking, 1 => +remove vertical space characters, 2 => remove underscore. */ + +static const int posix_class_maps[] = { + cbit_word, cbit_digit, -2, /* alpha */ + cbit_lower, -1, 0, /* lower */ + cbit_upper, -1, 0, /* upper */ + cbit_word, -1, 2, /* alnum - word without underscore */ + cbit_print, cbit_cntrl, 0, /* ascii */ + cbit_space, -1, 1, /* blank - a GNU extension */ + cbit_cntrl, -1, 0, /* cntrl */ + cbit_digit, -1, 0, /* digit */ + cbit_graph, -1, 0, /* graph */ + cbit_print, -1, 0, /* print */ + cbit_punct, -1, 0, /* punct */ + cbit_space, -1, 0, /* space */ + cbit_word, -1, 0, /* word - a Perl extension */ + cbit_xdigit,-1, 0 /* xdigit */ +}; + +/* Table of substitutes for \d etc when PCRE_UCP is set. The POSIX class +substitutes must be in the order of the names, defined above, and there are +both positive and negative cases. NULL means no substitute. */ + +#ifdef SUPPORT_UCP +static const pcre_uchar string_PNd[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_N, CHAR_d, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_pNd[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_N, CHAR_d, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_PXsp[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_s, CHAR_p, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_pXsp[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_s, CHAR_p, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_PXwd[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_w, CHAR_d, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_pXwd[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_w, CHAR_d, CHAR_RIGHT_CURLY_BRACKET, '\0' }; + +static const pcre_uchar *substitutes[] = { + string_PNd, /* \D */ + string_pNd, /* \d */ + string_PXsp, /* \S */ /* NOTE: Xsp is Perl space */ + string_pXsp, /* \s */ + string_PXwd, /* \W */ + string_pXwd /* \w */ +}; + +static const pcre_uchar string_pL[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_L, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_pLl[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_L, CHAR_l, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_pLu[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_L, CHAR_u, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_pXan[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_a, CHAR_n, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_h[] = { + CHAR_BACKSLASH, CHAR_h, '\0' }; +static const pcre_uchar string_pXps[] = { + CHAR_BACKSLASH, CHAR_p, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_p, CHAR_s, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_PL[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_L, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_PLl[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_L, CHAR_l, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_PLu[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_L, CHAR_u, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_PXan[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_a, CHAR_n, CHAR_RIGHT_CURLY_BRACKET, '\0' }; +static const pcre_uchar string_H[] = { + CHAR_BACKSLASH, CHAR_H, '\0' }; +static const pcre_uchar string_PXps[] = { + CHAR_BACKSLASH, CHAR_P, CHAR_LEFT_CURLY_BRACKET, + CHAR_X, CHAR_p, CHAR_s, CHAR_RIGHT_CURLY_BRACKET, '\0' }; + +static const pcre_uchar *posix_substitutes[] = { + string_pL, /* alpha */ + string_pLl, /* lower */ + string_pLu, /* upper */ + string_pXan, /* alnum */ + NULL, /* ascii */ + string_h, /* blank */ + NULL, /* cntrl */ + string_pNd, /* digit */ + NULL, /* graph */ + NULL, /* print */ + NULL, /* punct */ + string_pXps, /* space */ /* NOTE: Xps is POSIX space */ + string_pXwd, /* word */ + NULL, /* xdigit */ + /* Negated cases */ + string_PL, /* ^alpha */ + string_PLl, /* ^lower */ + string_PLu, /* ^upper */ + string_PXan, /* ^alnum */ + NULL, /* ^ascii */ + string_H, /* ^blank */ + NULL, /* ^cntrl */ + string_PNd, /* ^digit */ + NULL, /* ^graph */ + NULL, /* ^print */ + NULL, /* ^punct */ + string_PXps, /* ^space */ /* NOTE: Xps is POSIX space */ + string_PXwd, /* ^word */ + NULL /* ^xdigit */ +}; +#define POSIX_SUBSIZE (sizeof(posix_substitutes) / sizeof(pcre_uchar *)) +#endif + +#define STRING(a) # a +#define XSTRING(s) STRING(s) + +/* The texts of compile-time error messages. These are "char *" because they +are passed to the outside world. Do not ever re-use any error number, because +they are documented. Always add a new error instead. Messages marked DEAD below +are no longer used. This used to be a table of strings, but in order to reduce +the number of relocations needed when a shared library is loaded dynamically, +it is now one long string. We cannot use a table of offsets, because the +lengths of inserts such as XSTRING(MAX_NAME_SIZE) are not known. Instead, we +simply count through to the one we want - this isn't a performance issue +because these strings are used only when there is a compilation error. + +Each substring ends with \0 to insert a null character. This includes the final +substring, so that the whole string ends with \0\0, which can be detected when +counting through. */ + +static const char error_texts[] = + "no error\0" + "\\ at end of pattern\0" + "\\c at end of pattern\0" + "unrecognized character follows \\\0" + "numbers out of order in {} quantifier\0" + /* 5 */ + "number too big in {} quantifier\0" + "missing terminating ] for character class\0" + "invalid escape sequence in character class\0" + "range out of order in character class\0" + "nothing to repeat\0" + /* 10 */ + "operand of unlimited repeat could match the empty string\0" /** DEAD **/ + "internal error: unexpected repeat\0" + "unrecognized character after (? or (?-\0" + "POSIX named classes are supported only within a class\0" + "missing )\0" + /* 15 */ + "reference to non-existent subpattern\0" + "erroffset passed as NULL\0" + "unknown option bit(s) set\0" + "missing ) after comment\0" + "parentheses nested too deeply\0" /** DEAD **/ + /* 20 */ + "regular expression is too large\0" + "failed to get memory\0" + "unmatched parentheses\0" + "internal error: code overflow\0" + "unrecognized character after (?<\0" + /* 25 */ + "lookbehind assertion is not fixed length\0" + "malformed number or name after (?(\0" + "conditional group contains more than two branches\0" + "assertion expected after (?(\0" + "(?R or (?[+-]digits must be followed by )\0" + /* 30 */ + "unknown POSIX class name\0" + "POSIX collating elements are not supported\0" + "this version of PCRE is compiled without UTF support\0" + "spare error\0" /** DEAD **/ + "character value in \\x{...} sequence is too large\0" + /* 35 */ + "invalid condition (?(0)\0" + "\\C not allowed in lookbehind assertion\0" + "PCRE does not support \\L, \\l, \\N{name}, \\U, or \\u\0" + "number after (?C is > 255\0" + "closing ) for (?C expected\0" + /* 40 */ + "recursive call could loop indefinitely\0" + "unrecognized character after (?P\0" + "syntax error in subpattern name (missing terminator)\0" + "two named subpatterns have the same name\0" + "invalid UTF-8 string\0" + /* 45 */ + "support for \\P, \\p, and \\X has not been compiled\0" + "malformed \\P or \\p sequence\0" + "unknown property name after \\P or \\p\0" + "subpattern name is too long (maximum " XSTRING(MAX_NAME_SIZE) " characters)\0" + "too many named subpatterns (maximum " XSTRING(MAX_NAME_COUNT) ")\0" + /* 50 */ + "repeated subpattern is too long\0" /** DEAD **/ + "octal value is greater than \\377 in 8-bit non-UTF-8 mode\0" + "internal error: overran compiling workspace\0" + "internal error: previously-checked referenced subpattern not found\0" + "DEFINE group contains more than one branch\0" + /* 55 */ + "repeating a DEFINE group is not allowed\0" /** DEAD **/ + "inconsistent NEWLINE options\0" + "\\g is not followed by a braced, angle-bracketed, or quoted name/number or by a plain number\0" + "a numbered reference must not be zero\0" + "an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)\0" + /* 60 */ + "(*VERB) not recognized\0" + "number is too big\0" + "subpattern name expected\0" + "digit expected after (?+\0" + "] is an invalid data character in JavaScript compatibility mode\0" + /* 65 */ + "different names for subpatterns of the same number are not allowed\0" + "(*MARK) must have an argument\0" + "this version of PCRE is not compiled with Unicode property support\0" + "\\c must be followed by an ASCII character\0" + "\\k is not followed by a braced, angle-bracketed, or quoted name\0" + /* 70 */ + "internal error: unknown opcode in find_fixedlength()\0" + "\\N is not supported in a class\0" + "too many forward references\0" + "disallowed Unicode code point (>= 0xd800 && <= 0xdfff)\0" + "invalid UTF-16 string\0" + /* 75 */ + "name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)\0" + "character value in \\u.... sequence is too large\0" + ; + +/* Table to identify digits and hex digits. This is used when compiling +patterns. Note that the tables in chartables are dependent on the locale, and +may mark arbitrary characters as digits - but the PCRE compiling code expects +to handle only 0-9, a-z, and A-Z as digits when compiling. That is why we have +a private table here. It costs 256 bytes, but it is a lot faster than doing +character value tests (at least in some simple cases I timed), and in some +applications one wants PCRE to compile efficiently as well as match +efficiently. + +For convenience, we use the same bit definitions as in chartables: + + 0x04 decimal digit + 0x08 hexadecimal digit + +Then we can use ctype_digit and ctype_xdigit in the code. */ + +/* Using a simple comparison for decimal numbers rather than a memory read +is much faster, and the resulting code is simpler (the compiler turns it +into a subtraction and unsigned comparison). */ + +#define IS_DIGIT(x) ((x) >= CHAR_0 && (x) <= CHAR_9) + +#ifndef EBCDIC + +/* This is the "normal" case, for ASCII systems, and EBCDIC systems running in +UTF-8 mode. */ + +static const pcre_uint8 digitab[] = + { + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 8- 15 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - ' */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ( - / */ + 0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /* 0 - 7 */ + 0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00, /* 8 - ? */ + 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* @ - G */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* H - O */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* P - W */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* X - _ */ + 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* ` - g */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* h - o */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* p - w */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* x -127 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */ + +#else + +/* This is the "abnormal" case, for EBCDIC systems not running in UTF-8 mode. */ + +static const pcre_uint8 digitab[] = + { + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 0 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 8- 15 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 10 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 32- 39 20 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 40- 47 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 48- 55 30 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 56- 63 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - 71 40 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 72- | */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* & - 87 50 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 88- 95 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - -103 60 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 104- ? */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 112-119 70 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 120- " */ + 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* 128- g 80 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* h -143 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144- p 90 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* q -159 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160- x A0 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* y -175 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ^ -183 B0 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ + 0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* { - G C0 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* H -207 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* } - P D0 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* Q -223 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* \ - X E0 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* Y -239 */ + 0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /* 0 - 7 F0 */ + 0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00};/* 8 -255 */ + +static const pcre_uint8 ebcdic_chartab[] = { /* chartable partial dup */ + 0x80,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 0- 7 */ + 0x00,0x00,0x00,0x00,0x01,0x01,0x00,0x00, /* 8- 15 */ + 0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 16- 23 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */ + 0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 32- 39 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 40- 47 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 48- 55 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 56- 63 */ + 0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - 71 */ + 0x00,0x00,0x00,0x80,0x00,0x80,0x80,0x80, /* 72- | */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* & - 87 */ + 0x00,0x00,0x00,0x80,0x80,0x80,0x00,0x00, /* 88- 95 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - -103 */ + 0x00,0x00,0x00,0x00,0x00,0x10,0x00,0x80, /* 104- ? */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 112-119 */ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 120- " */ + 0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* 128- g */ + 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* h -143 */ + 0x00,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* 144- p */ + 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* q -159 */ + 0x00,0x00,0x12,0x12,0x12,0x12,0x12,0x12, /* 160- x */ + 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* y -175 */ + 0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ^ -183 */ + 0x00,0x00,0x80,0x00,0x00,0x00,0x00,0x00, /* 184-191 */ + 0x80,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* { - G */ + 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* H -207 */ + 0x00,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* } - P */ + 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* Q -223 */ + 0x00,0x00,0x12,0x12,0x12,0x12,0x12,0x12, /* \ - X */ + 0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* Y -239 */ + 0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /* 0 - 7 */ + 0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x00};/* 8 -255 */ +#endif + + +/* Definition to allow mutual recursion */ + +static BOOL + compile_regex(int, pcre_uchar **, const pcre_uchar **, int *, BOOL, BOOL, int, int, + int *, int *, branch_chain *, compile_data *, int *); + + + +/************************************************* +* Find an error text * +*************************************************/ + +/* The error texts are now all in one long string, to save on relocations. As +some of the text is of unknown length, we can't use a table of offsets. +Instead, just count through the strings. This is not a performance issue +because it happens only when there has been a compilation error. + +Argument: the error number +Returns: pointer to the error string +*/ + +static const char * +find_error_text(int n) +{ +const char *s = error_texts; +for (; n > 0; n--) + { + while (*s++ != 0) {}; + if (*s == 0) return "Error text not found (please report)"; + } +return s; +} + + +/************************************************* +* Expand the workspace * +*************************************************/ + +/* This function is called during the second compiling phase, if the number of +forward references fills the existing workspace, which is originally a block on +the stack. A larger block is obtained from malloc() unless the ultimate limit +has been reached or the increase will be rather small. + +Argument: pointer to the compile data block +Returns: 0 if all went well, else an error number +*/ + +static int +expand_workspace(compile_data *cd) +{ +pcre_uchar *newspace; +int newsize = cd->workspace_size * 2; + +if (newsize > COMPILE_WORK_SIZE_MAX) newsize = COMPILE_WORK_SIZE_MAX; +if (cd->workspace_size >= COMPILE_WORK_SIZE_MAX || + newsize - cd->workspace_size < WORK_SIZE_SAFETY_MARGIN) + return ERR72; + +newspace = (PUBL(malloc))(IN_UCHARS(newsize)); +if (newspace == NULL) return ERR21; +memcpy(newspace, cd->start_workspace, cd->workspace_size * sizeof(pcre_uchar)); +cd->hwm = (pcre_uchar *)newspace + (cd->hwm - cd->start_workspace); +if (cd->workspace_size > COMPILE_WORK_SIZE) + (PUBL(free))((void *)cd->start_workspace); +cd->start_workspace = newspace; +cd->workspace_size = newsize; +return 0; +} + + + +/************************************************* +* Check for counted repeat * +*************************************************/ + +/* This function is called when a '{' is encountered in a place where it might +start a quantifier. It looks ahead to see if it really is a quantifier or not. +It is only a quantifier if it is one of the forms {ddd} {ddd,} or {ddd,ddd} +where the ddds are digits. + +Arguments: + p pointer to the first char after '{' + +Returns: TRUE or FALSE +*/ + +static BOOL +is_counted_repeat(const pcre_uchar *p) +{ +if (!IS_DIGIT(*p)) return FALSE; +p++; +while (IS_DIGIT(*p)) p++; +if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE; + +if (*p++ != CHAR_COMMA) return FALSE; +if (*p == CHAR_RIGHT_CURLY_BRACKET) return TRUE; + +if (!IS_DIGIT(*p)) return FALSE; +p++; +while (IS_DIGIT(*p)) p++; + +return (*p == CHAR_RIGHT_CURLY_BRACKET); +} + + + +/************************************************* +* Handle escapes * +*************************************************/ + +/* This function is called when a \ has been encountered. It either returns a +positive value for a simple escape such as \n, or a negative value which +encodes one of the more complicated things such as \d. A backreference to group +n is returned as -(ESC_REF + n); ESC_REF is the highest ESC_xxx macro. When +UTF-8 is enabled, a positive value greater than 255 may be returned. On entry, +ptr is pointing at the \. On exit, it is on the final character of the escape +sequence. + +Arguments: + ptrptr points to the pattern position pointer + errorcodeptr points to the errorcode variable + bracount number of previous extracting brackets + options the options bits + isclass TRUE if inside a character class + +Returns: zero or positive => a data character + negative => a special escape sequence + on error, errorcodeptr is set +*/ + +static int +check_escape(const pcre_uchar **ptrptr, int *errorcodeptr, int bracount, + int options, BOOL isclass) +{ +/* PCRE_UTF16 has the same value as PCRE_UTF8. */ +BOOL utf = (options & PCRE_UTF8) != 0; +const pcre_uchar *ptr = *ptrptr + 1; +pcre_int32 c; +int i; + +GETCHARINCTEST(c, ptr); /* Get character value, increment pointer */ +ptr--; /* Set pointer back to the last byte */ + +/* If backslash is at the end of the pattern, it's an error. */ + +if (c == 0) *errorcodeptr = ERR1; + +/* Non-alphanumerics are literals. For digits or letters, do an initial lookup +in a table. A non-zero result is something that can be returned immediately. +Otherwise further processing may be required. */ + +#ifndef EBCDIC /* ASCII/UTF-8 coding */ +/* Not alphanumeric */ +else if (c < CHAR_0 || c > CHAR_z) {} +else if ((i = escapes[c - CHAR_0]) != 0) c = i; + +#else /* EBCDIC coding */ +/* Not alphanumeric */ +else if (c < 'a' || (!MAX_255(c) || (ebcdic_chartab[c] & 0x0E) == 0)) {} +else if ((i = escapes[c - 0x48]) != 0) c = i; +#endif + +/* Escapes that need further processing, or are illegal. */ + +else + { + const pcre_uchar *oldptr; + BOOL braced, negated; + + switch (c) + { + /* A number of Perl escapes are not handled by PCRE. We give an explicit + error. */ + + case CHAR_l: + case CHAR_L: + *errorcodeptr = ERR37; + break; + + case CHAR_u: + if ((options & PCRE_JAVASCRIPT_COMPAT) != 0) + { + /* In JavaScript, \u must be followed by four hexadecimal numbers. + Otherwise it is a lowercase u letter. */ + if (MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0 + && MAX_255(ptr[2]) && (digitab[ptr[2]] & ctype_xdigit) != 0 + && MAX_255(ptr[3]) && (digitab[ptr[3]] & ctype_xdigit) != 0 + && MAX_255(ptr[4]) && (digitab[ptr[4]] & ctype_xdigit) != 0) + { + c = 0; + for (i = 0; i < 4; ++i) + { + register int cc = *(++ptr); +#ifndef EBCDIC /* ASCII/UTF-8 coding */ + if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */ + c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10)); +#else /* EBCDIC coding */ + if (cc >= CHAR_a && cc <= CHAR_z) cc += 64; /* Convert to upper case */ + c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10)); +#endif + } + +#ifdef COMPILE_PCRE8 + if (c > (utf ? 0x10ffff : 0xff)) +#else +#ifdef COMPILE_PCRE16 + if (c > (utf ? 0x10ffff : 0xffff)) +#endif +#endif + { + *errorcodeptr = ERR76; + } + else if (utf && c >= 0xd800 && c <= 0xdfff) *errorcodeptr = ERR73; + } + } + else + *errorcodeptr = ERR37; + break; + + case CHAR_U: + /* In JavaScript, \U is an uppercase U letter. */ + if ((options & PCRE_JAVASCRIPT_COMPAT) == 0) *errorcodeptr = ERR37; + break; + + /* In a character class, \g is just a literal "g". Outside a character + class, \g must be followed by one of a number of specific things: + + (1) A number, either plain or braced. If positive, it is an absolute + backreference. If negative, it is a relative backreference. This is a Perl + 5.10 feature. + + (2) Perl 5.10 also supports \g{name} as a reference to a named group. This + is part of Perl's movement towards a unified syntax for back references. As + this is synonymous with \k{name}, we fudge it up by pretending it really + was \k. + + (3) For Oniguruma compatibility we also support \g followed by a name or a + number either in angle brackets or in single quotes. However, these are + (possibly recursive) subroutine calls, _not_ backreferences. Just return + the -ESC_g code (cf \k). */ + + case CHAR_g: + if (isclass) break; + if (ptr[1] == CHAR_LESS_THAN_SIGN || ptr[1] == CHAR_APOSTROPHE) + { + c = -ESC_g; + break; + } + + /* Handle the Perl-compatible cases */ + + if (ptr[1] == CHAR_LEFT_CURLY_BRACKET) + { + const pcre_uchar *p; + for (p = ptr+2; *p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET; p++) + if (*p != CHAR_MINUS && !IS_DIGIT(*p)) break; + if (*p != 0 && *p != CHAR_RIGHT_CURLY_BRACKET) + { + c = -ESC_k; + break; + } + braced = TRUE; + ptr++; + } + else braced = FALSE; + + if (ptr[1] == CHAR_MINUS) + { + negated = TRUE; + ptr++; + } + else negated = FALSE; + + /* The integer range is limited by the machine's int representation. */ + c = 0; + while (IS_DIGIT(ptr[1])) + { + if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */ + { + c = -1; + break; + } + c = c * 10 + *(++ptr) - CHAR_0; + } + if (((unsigned int)c) > INT_MAX) /* Integer overflow */ + { + while (IS_DIGIT(ptr[1])) + ptr++; + *errorcodeptr = ERR61; + break; + } + + if (braced && *(++ptr) != CHAR_RIGHT_CURLY_BRACKET) + { + *errorcodeptr = ERR57; + break; + } + + if (c == 0) + { + *errorcodeptr = ERR58; + break; + } + + if (negated) + { + if (c > bracount) + { + *errorcodeptr = ERR15; + break; + } + c = bracount - (c - 1); + } + + c = -(ESC_REF + c); + break; + + /* The handling of escape sequences consisting of a string of digits + starting with one that is not zero is not straightforward. By experiment, + the way Perl works seems to be as follows: + + Outside a character class, the digits are read as a decimal number. If the + number is less than 10, or if there are that many previous extracting + left brackets, then it is a back reference. Otherwise, up to three octal + digits are read to form an escaped byte. Thus \123 is likely to be octal + 123 (cf \0123, which is octal 012 followed by the literal 3). If the octal + value is greater than 377, the least significant 8 bits are taken. Inside a + character class, \ followed by a digit is always an octal number. */ + + case CHAR_1: case CHAR_2: case CHAR_3: case CHAR_4: case CHAR_5: + case CHAR_6: case CHAR_7: case CHAR_8: case CHAR_9: + + if (!isclass) + { + oldptr = ptr; + /* The integer range is limited by the machine's int representation. */ + c -= CHAR_0; + while (IS_DIGIT(ptr[1])) + { + if (((unsigned int)c) > INT_MAX / 10) /* Integer overflow */ + { + c = -1; + break; + } + c = c * 10 + *(++ptr) - CHAR_0; + } + if (((unsigned int)c) > INT_MAX) /* Integer overflow */ + { + while (IS_DIGIT(ptr[1])) + ptr++; + *errorcodeptr = ERR61; + break; + } + if (c < 10 || c <= bracount) + { + c = -(ESC_REF + c); + break; + } + ptr = oldptr; /* Put the pointer back and fall through */ + } + + /* Handle an octal number following \. If the first digit is 8 or 9, Perl + generates a binary zero byte and treats the digit as a following literal. + Thus we have to pull back the pointer by one. */ + + if ((c = *ptr) >= CHAR_8) + { + ptr--; + c = 0; + break; + } + + /* \0 always starts an octal number, but we may drop through to here with a + larger first octal digit. The original code used just to take the least + significant 8 bits of octal numbers (I think this is what early Perls used + to do). Nowadays we allow for larger numbers in UTF-8 mode and 16-bit mode, + but no more than 3 octal digits. */ + + case CHAR_0: + c -= CHAR_0; + while(i++ < 2 && ptr[1] >= CHAR_0 && ptr[1] <= CHAR_7) + c = c * 8 + *(++ptr) - CHAR_0; +#ifdef COMPILE_PCRE8 + if (!utf && c > 0xff) *errorcodeptr = ERR51; +#endif + break; + + /* \x is complicated. \x{ddd} is a character number which can be greater + than 0xff in utf or non-8bit mode, but only if the ddd are hex digits. + If not, { is treated as a data character. */ + + case CHAR_x: + if ((options & PCRE_JAVASCRIPT_COMPAT) != 0) + { + /* In JavaScript, \x must be followed by two hexadecimal numbers. + Otherwise it is a lowercase x letter. */ + if (MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0 + && MAX_255(ptr[2]) && (digitab[ptr[2]] & ctype_xdigit) != 0) + { + c = 0; + for (i = 0; i < 2; ++i) + { + register int cc = *(++ptr); +#ifndef EBCDIC /* ASCII/UTF-8 coding */ + if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */ + c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10)); +#else /* EBCDIC coding */ + if (cc >= CHAR_a && cc <= CHAR_z) cc += 64; /* Convert to upper case */ + c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10)); +#endif + } + } + break; + } + + if (ptr[1] == CHAR_LEFT_CURLY_BRACKET) + { + const pcre_uchar *pt = ptr + 2; + + c = 0; + while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) + { + register int cc = *pt++; + if (c == 0 && cc == CHAR_0) continue; /* Leading zeroes */ + +#ifndef EBCDIC /* ASCII/UTF-8 coding */ + if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */ + c = (c << 4) + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10)); +#else /* EBCDIC coding */ + if (cc >= CHAR_a && cc <= CHAR_z) cc += 64; /* Convert to upper case */ + c = (c << 4) + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10)); +#endif + +#ifdef COMPILE_PCRE8 + if (c > (utf ? 0x10ffff : 0xff)) { c = -1; break; } +#else +#ifdef COMPILE_PCRE16 + if (c > (utf ? 0x10ffff : 0xffff)) { c = -1; break; } +#endif +#endif + } + + if (c < 0) + { + while (MAX_255(*pt) && (digitab[*pt] & ctype_xdigit) != 0) pt++; + *errorcodeptr = ERR34; + } + + if (*pt == CHAR_RIGHT_CURLY_BRACKET) + { + if (utf && c >= 0xd800 && c <= 0xdfff) *errorcodeptr = ERR73; + ptr = pt; + break; + } + + /* If the sequence of hex digits does not end with '}', then we don't + recognize this construct; fall through to the normal \x handling. */ + } + + /* Read just a single-byte hex-defined char */ + + c = 0; + while (i++ < 2 && MAX_255(ptr[1]) && (digitab[ptr[1]] & ctype_xdigit) != 0) + { + int cc; /* Some compilers don't like */ + cc = *(++ptr); /* ++ in initializers */ +#ifndef EBCDIC /* ASCII/UTF-8 coding */ + if (cc >= CHAR_a) cc -= 32; /* Convert to upper case */ + c = c * 16 + cc - ((cc < CHAR_A)? CHAR_0 : (CHAR_A - 10)); +#else /* EBCDIC coding */ + if (cc <= CHAR_z) cc += 64; /* Convert to upper case */ + c = c * 16 + cc - ((cc >= CHAR_0)? CHAR_0 : (CHAR_A - 10)); +#endif + } + break; + + /* For \c, a following letter is upper-cased; then the 0x40 bit is flipped. + An error is given if the byte following \c is not an ASCII character. This + coding is ASCII-specific, but then the whole concept of \cx is + ASCII-specific. (However, an EBCDIC equivalent has now been added.) */ + + case CHAR_c: + c = *(++ptr); + if (c == 0) + { + *errorcodeptr = ERR2; + break; + } +#ifndef EBCDIC /* ASCII/UTF-8 coding */ + if (c > 127) /* Excludes all non-ASCII in either mode */ + { + *errorcodeptr = ERR68; + break; + } + if (c >= CHAR_a && c <= CHAR_z) c -= 32; + c ^= 0x40; +#else /* EBCDIC coding */ + if (c >= CHAR_a && c <= CHAR_z) c += 64; + c ^= 0xC0; +#endif + break; + + /* PCRE_EXTRA enables extensions to Perl in the matter of escapes. Any + other alphanumeric following \ is an error if PCRE_EXTRA was set; + otherwise, for Perl compatibility, it is a literal. This code looks a bit + odd, but there used to be some cases other than the default, and there may + be again in future, so I haven't "optimized" it. */ + + default: + if ((options & PCRE_EXTRA) != 0) switch(c) + { + default: + *errorcodeptr = ERR3; + break; + } + break; + } + } + +/* Perl supports \N{name} for character names, as well as plain \N for "not +newline". PCRE does not support \N{name}. However, it does support +quantification such as \N{2,3}. */ + +if (c == -ESC_N && ptr[1] == CHAR_LEFT_CURLY_BRACKET && + !is_counted_repeat(ptr+2)) + *errorcodeptr = ERR37; + +/* If PCRE_UCP is set, we change the values for \d etc. */ + +if ((options & PCRE_UCP) != 0 && c <= -ESC_D && c >= -ESC_w) + c -= (ESC_DU - ESC_D); + +/* Set the pointer to the final character before returning. */ + +*ptrptr = ptr; +return c; +} + + + +#ifdef SUPPORT_UCP +/************************************************* +* Handle \P and \p * +*************************************************/ + +/* This function is called after \P or \p has been encountered, provided that +PCRE is compiled with support for Unicode properties. On entry, ptrptr is +pointing at the P or p. On exit, it is pointing at the final character of the +escape sequence. + +Argument: + ptrptr points to the pattern position pointer + negptr points to a boolean that is set TRUE for negation else FALSE + dptr points to an int that is set to the detailed property value + errorcodeptr points to the error code variable + +Returns: type value from ucp_type_table, or -1 for an invalid type +*/ + +static int +get_ucp(const pcre_uchar **ptrptr, BOOL *negptr, int *dptr, int *errorcodeptr) +{ +int c, i, bot, top; +const pcre_uchar *ptr = *ptrptr; +pcre_uchar name[32]; + +c = *(++ptr); +if (c == 0) goto ERROR_RETURN; + +*negptr = FALSE; + +/* \P or \p can be followed by a name in {}, optionally preceded by ^ for +negation. */ + +if (c == CHAR_LEFT_CURLY_BRACKET) + { + if (ptr[1] == CHAR_CIRCUMFLEX_ACCENT) + { + *negptr = TRUE; + ptr++; + } + for (i = 0; i < (int)(sizeof(name) / sizeof(pcre_uchar)) - 1; i++) + { + c = *(++ptr); + if (c == 0) goto ERROR_RETURN; + if (c == CHAR_RIGHT_CURLY_BRACKET) break; + name[i] = c; + } + if (c != CHAR_RIGHT_CURLY_BRACKET) goto ERROR_RETURN; + name[i] = 0; + } + +/* Otherwise there is just one following character */ + +else + { + name[0] = c; + name[1] = 0; + } + +*ptrptr = ptr; + +/* Search for a recognized property name using binary chop */ + +bot = 0; +top = PRIV(utt_size); + +while (bot < top) + { + i = (bot + top) >> 1; + c = STRCMP_UC_C8(name, PRIV(utt_names) + PRIV(utt)[i].name_offset); + if (c == 0) + { + *dptr = PRIV(utt)[i].value; + return PRIV(utt)[i].type; + } + if (c > 0) bot = i + 1; else top = i; + } + +*errorcodeptr = ERR47; +*ptrptr = ptr; +return -1; + +ERROR_RETURN: +*errorcodeptr = ERR46; +*ptrptr = ptr; +return -1; +} +#endif + + + + +/************************************************* +* Read repeat counts * +*************************************************/ + +/* Read an item of the form {n,m} and return the values. This is called only +after is_counted_repeat() has confirmed that a repeat-count quantifier exists, +so the syntax is guaranteed to be correct, but we need to check the values. + +Arguments: + p pointer to first char after '{' + minp pointer to int for min + maxp pointer to int for max + returned as -1 if no max + errorcodeptr points to error code variable + +Returns: pointer to '}' on success; + current ptr on error, with errorcodeptr set non-zero +*/ + +static const pcre_uchar * +read_repeat_counts(const pcre_uchar *p, int *minp, int *maxp, int *errorcodeptr) +{ +int min = 0; +int max = -1; + +/* Read the minimum value and do a paranoid check: a negative value indicates +an integer overflow. */ + +while (IS_DIGIT(*p)) min = min * 10 + *p++ - CHAR_0; +if (min < 0 || min > 65535) + { + *errorcodeptr = ERR5; + return p; + } + +/* Read the maximum value if there is one, and again do a paranoid on its size. +Also, max must not be less than min. */ + +if (*p == CHAR_RIGHT_CURLY_BRACKET) max = min; else + { + if (*(++p) != CHAR_RIGHT_CURLY_BRACKET) + { + max = 0; + while(IS_DIGIT(*p)) max = max * 10 + *p++ - CHAR_0; + if (max < 0 || max > 65535) + { + *errorcodeptr = ERR5; + return p; + } + if (max < min) + { + *errorcodeptr = ERR4; + return p; + } + } + } + +/* Fill in the required variables, and pass back the pointer to the terminating +'}'. */ + +*minp = min; +*maxp = max; +return p; +} + + + +/************************************************* +* Subroutine for finding forward reference * +*************************************************/ + +/* This recursive function is called only from find_parens() below. The +top-level call starts at the beginning of the pattern. All other calls must +start at a parenthesis. It scans along a pattern's text looking for capturing +subpatterns, and counting them. If it finds a named pattern that matches the +name it is given, it returns its number. Alternatively, if the name is NULL, it +returns when it reaches a given numbered subpattern. Recursion is used to keep +track of subpatterns that reset the capturing group numbers - the (?| feature. + +This function was originally called only from the second pass, in which we know +that if (?< or (?' or (?P< is encountered, the name will be correctly +terminated because that is checked in the first pass. There is now one call to +this function in the first pass, to check for a recursive back reference by +name (so that we can make the whole group atomic). In this case, we need check +only up to the current position in the pattern, and that is still OK because +and previous occurrences will have been checked. To make this work, the test +for "end of pattern" is a check against cd->end_pattern in the main loop, +instead of looking for a binary zero. This means that the special first-pass +call can adjust cd->end_pattern temporarily. (Checks for binary zero while +processing items within the loop are OK, because afterwards the main loop will +terminate.) + +Arguments: + ptrptr address of the current character pointer (updated) + cd compile background data + name name to seek, or NULL if seeking a numbered subpattern + lorn name length, or subpattern number if name is NULL + xmode TRUE if we are in /x mode + utf TRUE if we are in UTF-8 / UTF-16 mode + count pointer to the current capturing subpattern number (updated) + +Returns: the number of the named subpattern, or -1 if not found +*/ + +static int +find_parens_sub(pcre_uchar **ptrptr, compile_data *cd, const pcre_uchar *name, int lorn, + BOOL xmode, BOOL utf, int *count) +{ +pcre_uchar *ptr = *ptrptr; +int start_count = *count; +int hwm_count = start_count; +BOOL dup_parens = FALSE; + +/* If the first character is a parenthesis, check on the type of group we are +dealing with. The very first call may not start with a parenthesis. */ + +if (ptr[0] == CHAR_LEFT_PARENTHESIS) + { + /* Handle specials such as (*SKIP) or (*UTF8) etc. */ + + if (ptr[1] == CHAR_ASTERISK) ptr += 2; + + /* Handle a normal, unnamed capturing parenthesis. */ + + else if (ptr[1] != CHAR_QUESTION_MARK) + { + *count += 1; + if (name == NULL && *count == lorn) return *count; + ptr++; + } + + /* All cases now have (? at the start. Remember when we are in a group + where the parenthesis numbers are duplicated. */ + + else if (ptr[2] == CHAR_VERTICAL_LINE) + { + ptr += 3; + dup_parens = TRUE; + } + + /* Handle comments; all characters are allowed until a ket is reached. */ + + else if (ptr[2] == CHAR_NUMBER_SIGN) + { + for (ptr += 3; *ptr != 0; ptr++) if (*ptr == CHAR_RIGHT_PARENTHESIS) break; + goto FAIL_EXIT; + } + + /* Handle a condition. If it is an assertion, just carry on so that it + is processed as normal. If not, skip to the closing parenthesis of the + condition (there can't be any nested parens). */ + + else if (ptr[2] == CHAR_LEFT_PARENTHESIS) + { + ptr += 2; + if (ptr[1] != CHAR_QUESTION_MARK) + { + while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++; + if (*ptr != 0) ptr++; + } + } + + /* Start with (? but not a condition. */ + + else + { + ptr += 2; + if (*ptr == CHAR_P) ptr++; /* Allow optional P */ + + /* We have to disambiguate (? for named groups */ + + if ((*ptr == CHAR_LESS_THAN_SIGN && ptr[1] != CHAR_EXCLAMATION_MARK && + ptr[1] != CHAR_EQUALS_SIGN) || *ptr == CHAR_APOSTROPHE) + { + int term; + const pcre_uchar *thisname; + *count += 1; + if (name == NULL && *count == lorn) return *count; + term = *ptr++; + if (term == CHAR_LESS_THAN_SIGN) term = CHAR_GREATER_THAN_SIGN; + thisname = ptr; + while (*ptr != term) ptr++; + if (name != NULL && lorn == ptr - thisname && + STRNCMP_UC_UC(name, thisname, lorn) == 0) + return *count; + term++; + } + } + } + +/* Past any initial parenthesis handling, scan for parentheses or vertical +bars. Stop if we get to cd->end_pattern. Note that this is important for the +first-pass call when this value is temporarily adjusted to stop at the current +position. So DO NOT change this to a test for binary zero. */ + +for (; ptr < cd->end_pattern; ptr++) + { + /* Skip over backslashed characters and also entire \Q...\E */ + + if (*ptr == CHAR_BACKSLASH) + { + if (*(++ptr) == 0) goto FAIL_EXIT; + if (*ptr == CHAR_Q) for (;;) + { + while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {}; + if (*ptr == 0) goto FAIL_EXIT; + if (*(++ptr) == CHAR_E) break; + } + continue; + } + + /* Skip over character classes; this logic must be similar to the way they + are handled for real. If the first character is '^', skip it. Also, if the + first few characters (either before or after ^) are \Q\E or \E we skip them + too. This makes for compatibility with Perl. Note the use of STR macros to + encode "Q\\E" so that it works in UTF-8 on EBCDIC platforms. */ + + if (*ptr == CHAR_LEFT_SQUARE_BRACKET) + { + BOOL negate_class = FALSE; + for (;;) + { + if (ptr[1] == CHAR_BACKSLASH) + { + if (ptr[2] == CHAR_E) + ptr+= 2; + else if (STRNCMP_UC_C8(ptr + 2, + STR_Q STR_BACKSLASH STR_E, 3) == 0) + ptr += 4; + else + break; + } + else if (!negate_class && ptr[1] == CHAR_CIRCUMFLEX_ACCENT) + { + negate_class = TRUE; + ptr++; + } + else break; + } + + /* If the next character is ']', it is a data character that must be + skipped, except in JavaScript compatibility mode. */ + + if (ptr[1] == CHAR_RIGHT_SQUARE_BRACKET && + (cd->external_options & PCRE_JAVASCRIPT_COMPAT) == 0) + ptr++; + + while (*(++ptr) != CHAR_RIGHT_SQUARE_BRACKET) + { + if (*ptr == 0) return -1; + if (*ptr == CHAR_BACKSLASH) + { + if (*(++ptr) == 0) goto FAIL_EXIT; + if (*ptr == CHAR_Q) for (;;) + { + while (*(++ptr) != 0 && *ptr != CHAR_BACKSLASH) {}; + if (*ptr == 0) goto FAIL_EXIT; + if (*(++ptr) == CHAR_E) break; + } + continue; + } + } + continue; + } + + /* Skip comments in /x mode */ + + if (xmode && *ptr == CHAR_NUMBER_SIGN) + { + ptr++; + while (*ptr != 0) + { + if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; } + ptr++; +#ifdef SUPPORT_UTF + if (utf) FORWARDCHAR(ptr); +#endif + } + if (*ptr == 0) goto FAIL_EXIT; + continue; + } + + /* Check for the special metacharacters */ + + if (*ptr == CHAR_LEFT_PARENTHESIS) + { + int rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, count); + if (rc > 0) return rc; + if (*ptr == 0) goto FAIL_EXIT; + } + + else if (*ptr == CHAR_RIGHT_PARENTHESIS) + { + if (dup_parens && *count < hwm_count) *count = hwm_count; + goto FAIL_EXIT; + } + + else if (*ptr == CHAR_VERTICAL_LINE && dup_parens) + { + if (*count > hwm_count) hwm_count = *count; + *count = start_count; + } + } + +FAIL_EXIT: +*ptrptr = ptr; +return -1; +} + + + + +/************************************************* +* Find forward referenced subpattern * +*************************************************/ + +/* This function scans along a pattern's text looking for capturing +subpatterns, and counting them. If it finds a named pattern that matches the +name it is given, it returns its number. Alternatively, if the name is NULL, it +returns when it reaches a given numbered subpattern. This is used for forward +references to subpatterns. We used to be able to start this scan from the +current compiling point, using the current count value from cd->bracount, and +do it all in a single loop, but the addition of the possibility of duplicate +subpattern numbers means that we have to scan from the very start, in order to +take account of such duplicates, and to use a recursive function to keep track +of the different types of group. + +Arguments: + cd compile background data + name name to seek, or NULL if seeking a numbered subpattern + lorn name length, or subpattern number if name is NULL + xmode TRUE if we are in /x mode + utf TRUE if we are in UTF-8 / UTF-16 mode + +Returns: the number of the found subpattern, or -1 if not found +*/ + +static int +find_parens(compile_data *cd, const pcre_uchar *name, int lorn, BOOL xmode, + BOOL utf) +{ +pcre_uchar *ptr = (pcre_uchar *)cd->start_pattern; +int count = 0; +int rc; + +/* If the pattern does not start with an opening parenthesis, the first call +to find_parens_sub() will scan right to the end (if necessary). However, if it +does start with a parenthesis, find_parens_sub() will return when it hits the +matching closing parens. That is why we have to have a loop. */ + +for (;;) + { + rc = find_parens_sub(&ptr, cd, name, lorn, xmode, utf, &count); + if (rc > 0 || *ptr++ == 0) break; + } + +return rc; +} + + + + +/************************************************* +* Find first significant op code * +*************************************************/ + +/* This is called by several functions that scan a compiled expression looking +for a fixed first character, or an anchoring op code etc. It skips over things +that do not influence this. For some calls, it makes sense to skip negative +forward and all backward assertions, and also the \b assertion; for others it +does not. + +Arguments: + code pointer to the start of the group + skipassert TRUE if certain assertions are to be skipped + +Returns: pointer to the first significant opcode +*/ + +static const pcre_uchar* +first_significant_code(const pcre_uchar *code, BOOL skipassert) +{ +for (;;) + { + switch ((int)*code) + { + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + if (!skipassert) return code; + do code += GET(code, 1); while (*code == OP_ALT); + code += PRIV(OP_lengths)[*code]; + break; + + case OP_WORD_BOUNDARY: + case OP_NOT_WORD_BOUNDARY: + if (!skipassert) return code; + /* Fall through */ + + case OP_CALLOUT: + case OP_CREF: + case OP_NCREF: + case OP_RREF: + case OP_NRREF: + case OP_DEF: + code += PRIV(OP_lengths)[*code]; + break; + + default: + return code; + } + } +/* Control never reaches here */ +} + + + + +/************************************************* +* Find the fixed length of a branch * +*************************************************/ + +/* Scan a branch and compute the fixed length of subject that will match it, +if the length is fixed. This is needed for dealing with backward assertions. +In UTF8 mode, the result is in characters rather than bytes. The branch is +temporarily terminated with OP_END when this function is called. + +This function is called when a backward assertion is encountered, so that if it +fails, the error message can point to the correct place in the pattern. +However, we cannot do this when the assertion contains subroutine calls, +because they can be forward references. We solve this by remembering this case +and doing the check at the end; a flag specifies which mode we are running in. + +Arguments: + code points to the start of the pattern (the bracket) + utf TRUE in UTF-8 / UTF-16 mode + atend TRUE if called when the pattern is complete + cd the "compile data" structure + +Returns: the fixed length, + or -1 if there is no fixed length, + or -2 if \C was encountered (in UTF-8 mode only) + or -3 if an OP_RECURSE item was encountered and atend is FALSE + or -4 if an unknown opcode was encountered (internal error) +*/ + +static int +find_fixedlength(pcre_uchar *code, BOOL utf, BOOL atend, compile_data *cd) +{ +int length = -1; + +register int branchlength = 0; +register pcre_uchar *cc = code + 1 + LINK_SIZE; + +/* Scan along the opcodes for this branch. If we get to the end of the +branch, check the length against that of the other branches. */ + +for (;;) + { + int d; + pcre_uchar *ce, *cs; + register int op = *cc; + + switch (op) + { + /* We only need to continue for OP_CBRA (normal capturing bracket) and + OP_BRA (normal non-capturing bracket) because the other variants of these + opcodes are all concerned with unlimited repeated groups, which of course + are not of fixed length. */ + + case OP_CBRA: + case OP_BRA: + case OP_ONCE: + case OP_ONCE_NC: + case OP_COND: + d = find_fixedlength(cc + ((op == OP_CBRA)? IMM2_SIZE : 0), utf, atend, cd); + if (d < 0) return d; + branchlength += d; + do cc += GET(cc, 1); while (*cc == OP_ALT); + cc += 1 + LINK_SIZE; + break; + + /* Reached end of a branch; if it's a ket it is the end of a nested call. + If it's ALT it is an alternation in a nested call. An ACCEPT is effectively + an ALT. If it is END it's the end of the outer call. All can be handled by + the same code. Note that we must not include the OP_KETRxxx opcodes here, + because they all imply an unlimited repeat. */ + + case OP_ALT: + case OP_KET: + case OP_END: + case OP_ACCEPT: + case OP_ASSERT_ACCEPT: + if (length < 0) length = branchlength; + else if (length != branchlength) return -1; + if (*cc != OP_ALT) return length; + cc += 1 + LINK_SIZE; + branchlength = 0; + break; + + /* A true recursion implies not fixed length, but a subroutine call may + be OK. If the subroutine is a forward reference, we can't deal with + it until the end of the pattern, so return -3. */ + + case OP_RECURSE: + if (!atend) return -3; + cs = ce = (pcre_uchar *)cd->start_code + GET(cc, 1); /* Start subpattern */ + do ce += GET(ce, 1); while (*ce == OP_ALT); /* End subpattern */ + if (cc > cs && cc < ce) return -1; /* Recursion */ + d = find_fixedlength(cs + IMM2_SIZE, utf, atend, cd); + if (d < 0) return d; + branchlength += d; + cc += 1 + LINK_SIZE; + break; + + /* Skip over assertive subpatterns */ + + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + do cc += GET(cc, 1); while (*cc == OP_ALT); + cc += PRIV(OP_lengths)[*cc]; + break; + + /* Skip over things that don't match chars */ + + case OP_MARK: + case OP_PRUNE_ARG: + case OP_SKIP_ARG: + case OP_THEN_ARG: + cc += cc[1] + PRIV(OP_lengths)[*cc]; + break; + + case OP_CALLOUT: + case OP_CIRC: + case OP_CIRCM: + case OP_CLOSE: + case OP_COMMIT: + case OP_CREF: + case OP_DEF: + case OP_DOLL: + case OP_DOLLM: + case OP_EOD: + case OP_EODN: + case OP_FAIL: + case OP_NCREF: + case OP_NRREF: + case OP_NOT_WORD_BOUNDARY: + case OP_PRUNE: + case OP_REVERSE: + case OP_RREF: + case OP_SET_SOM: + case OP_SKIP: + case OP_SOD: + case OP_SOM: + case OP_THEN: + case OP_WORD_BOUNDARY: + cc += PRIV(OP_lengths)[*cc]; + break; + + /* Handle literal characters */ + + case OP_CHAR: + case OP_CHARI: + case OP_NOT: + case OP_NOTI: + branchlength++; + cc += 2; +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]); +#endif + break; + + /* Handle exact repetitions. The count is already in characters, but we + need to skip over a multibyte character in UTF8 mode. */ + + case OP_EXACT: + case OP_EXACTI: + case OP_NOTEXACT: + case OP_NOTEXACTI: + branchlength += GET2(cc,1); + cc += 2 + IMM2_SIZE; +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]); +#endif + break; + + case OP_TYPEEXACT: + branchlength += GET2(cc,1); + if (cc[1 + IMM2_SIZE] == OP_PROP || cc[1 + IMM2_SIZE] == OP_NOTPROP) cc += 2; + cc += 1 + IMM2_SIZE + 1; + break; + + /* Handle single-char matchers */ + + case OP_PROP: + case OP_NOTPROP: + cc += 2; + /* Fall through */ + + case OP_HSPACE: + case OP_VSPACE: + case OP_NOT_HSPACE: + case OP_NOT_VSPACE: + case OP_NOT_DIGIT: + case OP_DIGIT: + case OP_NOT_WHITESPACE: + case OP_WHITESPACE: + case OP_NOT_WORDCHAR: + case OP_WORDCHAR: + case OP_ANY: + case OP_ALLANY: + branchlength++; + cc++; + break; + + /* The single-byte matcher isn't allowed. This only happens in UTF-8 mode; + otherwise \C is coded as OP_ALLANY. */ + + case OP_ANYBYTE: + return -2; + + /* Check a class for variable quantification */ + +#if defined SUPPORT_UTF || defined COMPILE_PCRE16 + case OP_XCLASS: + cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS]; + /* Fall through */ +#endif + + case OP_CLASS: + case OP_NCLASS: + cc += PRIV(OP_lengths)[OP_CLASS]; + + switch (*cc) + { + case OP_CRPLUS: + case OP_CRMINPLUS: + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRQUERY: + case OP_CRMINQUERY: + return -1; + + case OP_CRRANGE: + case OP_CRMINRANGE: + if (GET2(cc,1) != GET2(cc,1+IMM2_SIZE)) return -1; + branchlength += GET2(cc,1); + cc += 1 + 2 * IMM2_SIZE; + break; + + default: + branchlength++; + } + break; + + /* Anything else is variable length */ + + case OP_ANYNL: + case OP_BRAMINZERO: + case OP_BRAPOS: + case OP_BRAPOSZERO: + case OP_BRAZERO: + case OP_CBRAPOS: + case OP_EXTUNI: + case OP_KETRMAX: + case OP_KETRMIN: + case OP_KETRPOS: + case OP_MINPLUS: + case OP_MINPLUSI: + case OP_MINQUERY: + case OP_MINQUERYI: + case OP_MINSTAR: + case OP_MINSTARI: + case OP_MINUPTO: + case OP_MINUPTOI: + case OP_NOTMINPLUS: + case OP_NOTMINPLUSI: + case OP_NOTMINQUERY: + case OP_NOTMINQUERYI: + case OP_NOTMINSTAR: + case OP_NOTMINSTARI: + case OP_NOTMINUPTO: + case OP_NOTMINUPTOI: + case OP_NOTPLUS: + case OP_NOTPLUSI: + case OP_NOTPOSPLUS: + case OP_NOTPOSPLUSI: + case OP_NOTPOSQUERY: + case OP_NOTPOSQUERYI: + case OP_NOTPOSSTAR: + case OP_NOTPOSSTARI: + case OP_NOTPOSUPTO: + case OP_NOTPOSUPTOI: + case OP_NOTQUERY: + case OP_NOTQUERYI: + case OP_NOTSTAR: + case OP_NOTSTARI: + case OP_NOTUPTO: + case OP_NOTUPTOI: + case OP_PLUS: + case OP_PLUSI: + case OP_POSPLUS: + case OP_POSPLUSI: + case OP_POSQUERY: + case OP_POSQUERYI: + case OP_POSSTAR: + case OP_POSSTARI: + case OP_POSUPTO: + case OP_POSUPTOI: + case OP_QUERY: + case OP_QUERYI: + case OP_REF: + case OP_REFI: + case OP_SBRA: + case OP_SBRAPOS: + case OP_SCBRA: + case OP_SCBRAPOS: + case OP_SCOND: + case OP_SKIPZERO: + case OP_STAR: + case OP_STARI: + case OP_TYPEMINPLUS: + case OP_TYPEMINQUERY: + case OP_TYPEMINSTAR: + case OP_TYPEMINUPTO: + case OP_TYPEPLUS: + case OP_TYPEPOSPLUS: + case OP_TYPEPOSQUERY: + case OP_TYPEPOSSTAR: + case OP_TYPEPOSUPTO: + case OP_TYPEQUERY: + case OP_TYPESTAR: + case OP_TYPEUPTO: + case OP_UPTO: + case OP_UPTOI: + return -1; + + /* Catch unrecognized opcodes so that when new ones are added they + are not forgotten, as has happened in the past. */ + + default: + return -4; + } + } +/* Control never gets here */ +} + + + + +/************************************************* +* Scan compiled regex for specific bracket * +*************************************************/ + +/* This little function scans through a compiled pattern until it finds a +capturing bracket with the given number, or, if the number is negative, an +instance of OP_REVERSE for a lookbehind. The function is global in the C sense +so that it can be called from pcre_study() when finding the minimum matching +length. + +Arguments: + code points to start of expression + utf TRUE in UTF-8 / UTF-16 mode + number the required bracket number or negative to find a lookbehind + +Returns: pointer to the opcode for the bracket, or NULL if not found +*/ + +const pcre_uchar * +PRIV(find_bracket)(const pcre_uchar *code, BOOL utf, int number) +{ +for (;;) + { + register int c = *code; + + if (c == OP_END) return NULL; + + /* XCLASS is used for classes that cannot be represented just by a bit + map. This includes negated single high-valued characters. The length in + the table is zero; the actual length is stored in the compiled code. */ + + if (c == OP_XCLASS) code += GET(code, 1); + + /* Handle recursion */ + + else if (c == OP_REVERSE) + { + if (number < 0) return (pcre_uchar *)code; + code += PRIV(OP_lengths)[c]; + } + + /* Handle capturing bracket */ + + else if (c == OP_CBRA || c == OP_SCBRA || + c == OP_CBRAPOS || c == OP_SCBRAPOS) + { + int n = GET2(code, 1+LINK_SIZE); + if (n == number) return (pcre_uchar *)code; + code += PRIV(OP_lengths)[c]; + } + + /* Otherwise, we can get the item's length from the table, except that for + repeated character types, we have to test for \p and \P, which have an extra + two bytes of parameters, and for MARK/PRUNE/SKIP/THEN with an argument, we + must add in its length. */ + + else + { + switch(c) + { + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSSTAR: + case OP_TYPEPOSPLUS: + case OP_TYPEPOSQUERY: + if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2; + break; + + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEEXACT: + case OP_TYPEPOSUPTO: + if (code[1 + IMM2_SIZE] == OP_PROP + || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2; + break; + + case OP_MARK: + case OP_PRUNE_ARG: + case OP_SKIP_ARG: + code += code[1]; + break; + + case OP_THEN_ARG: + code += code[1]; + break; + } + + /* Add in the fixed length from the table */ + + code += PRIV(OP_lengths)[c]; + + /* In UTF-8 mode, opcodes that are followed by a character may be followed by + a multi-byte character. The length in the table is a minimum, so we have to + arrange to skip the extra bytes. */ + +#ifdef SUPPORT_UTF + if (utf) switch(c) + { + case OP_CHAR: + case OP_CHARI: + case OP_EXACT: + case OP_EXACTI: + case OP_UPTO: + case OP_UPTOI: + case OP_MINUPTO: + case OP_MINUPTOI: + case OP_POSUPTO: + case OP_POSUPTOI: + case OP_STAR: + case OP_STARI: + case OP_MINSTAR: + case OP_MINSTARI: + case OP_POSSTAR: + case OP_POSSTARI: + case OP_PLUS: + case OP_PLUSI: + case OP_MINPLUS: + case OP_MINPLUSI: + case OP_POSPLUS: + case OP_POSPLUSI: + case OP_QUERY: + case OP_QUERYI: + case OP_MINQUERY: + case OP_MINQUERYI: + case OP_POSQUERY: + case OP_POSQUERYI: + if (HAS_EXTRALEN(code[-1])) code += GET_EXTRALEN(code[-1]); + break; + } +#else + (void)(utf); /* Keep compiler happy by referencing function argument */ +#endif + } + } +} + + + +/************************************************* +* Scan compiled regex for recursion reference * +*************************************************/ + +/* This little function scans through a compiled pattern until it finds an +instance of OP_RECURSE. + +Arguments: + code points to start of expression + utf TRUE in UTF-8 / UTF-16 mode + +Returns: pointer to the opcode for OP_RECURSE, or NULL if not found +*/ + +static const pcre_uchar * +find_recurse(const pcre_uchar *code, BOOL utf) +{ +for (;;) + { + register int c = *code; + if (c == OP_END) return NULL; + if (c == OP_RECURSE) return code; + + /* XCLASS is used for classes that cannot be represented just by a bit + map. This includes negated single high-valued characters. The length in + the table is zero; the actual length is stored in the compiled code. */ + + if (c == OP_XCLASS) code += GET(code, 1); + + /* Otherwise, we can get the item's length from the table, except that for + repeated character types, we have to test for \p and \P, which have an extra + two bytes of parameters, and for MARK/PRUNE/SKIP/THEN with an argument, we + must add in its length. */ + + else + { + switch(c) + { + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSSTAR: + case OP_TYPEPOSPLUS: + case OP_TYPEPOSQUERY: + if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2; + break; + + case OP_TYPEPOSUPTO: + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEEXACT: + if (code[1 + IMM2_SIZE] == OP_PROP + || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2; + break; + + case OP_MARK: + case OP_PRUNE_ARG: + case OP_SKIP_ARG: + code += code[1]; + break; + + case OP_THEN_ARG: + code += code[1]; + break; + } + + /* Add in the fixed length from the table */ + + code += PRIV(OP_lengths)[c]; + + /* In UTF-8 mode, opcodes that are followed by a character may be followed + by a multi-byte character. The length in the table is a minimum, so we have + to arrange to skip the extra bytes. */ + +#ifdef SUPPORT_UTF + if (utf) switch(c) + { + case OP_CHAR: + case OP_CHARI: + case OP_NOT: + case OP_NOTI: + case OP_EXACT: + case OP_EXACTI: + case OP_NOTEXACT: + case OP_NOTEXACTI: + case OP_UPTO: + case OP_UPTOI: + case OP_NOTUPTO: + case OP_NOTUPTOI: + case OP_MINUPTO: + case OP_MINUPTOI: + case OP_NOTMINUPTO: + case OP_NOTMINUPTOI: + case OP_POSUPTO: + case OP_POSUPTOI: + case OP_NOTPOSUPTO: + case OP_NOTPOSUPTOI: + case OP_STAR: + case OP_STARI: + case OP_NOTSTAR: + case OP_NOTSTARI: + case OP_MINSTAR: + case OP_MINSTARI: + case OP_NOTMINSTAR: + case OP_NOTMINSTARI: + case OP_POSSTAR: + case OP_POSSTARI: + case OP_NOTPOSSTAR: + case OP_NOTPOSSTARI: + case OP_PLUS: + case OP_PLUSI: + case OP_NOTPLUS: + case OP_NOTPLUSI: + case OP_MINPLUS: + case OP_MINPLUSI: + case OP_NOTMINPLUS: + case OP_NOTMINPLUSI: + case OP_POSPLUS: + case OP_POSPLUSI: + case OP_NOTPOSPLUS: + case OP_NOTPOSPLUSI: + case OP_QUERY: + case OP_QUERYI: + case OP_NOTQUERY: + case OP_NOTQUERYI: + case OP_MINQUERY: + case OP_MINQUERYI: + case OP_NOTMINQUERY: + case OP_NOTMINQUERYI: + case OP_POSQUERY: + case OP_POSQUERYI: + case OP_NOTPOSQUERY: + case OP_NOTPOSQUERYI: + if (HAS_EXTRALEN(code[-1])) code += GET_EXTRALEN(code[-1]); + break; + } +#else + (void)(utf); /* Keep compiler happy by referencing function argument */ +#endif + } + } +} + + + +/************************************************* +* Scan compiled branch for non-emptiness * +*************************************************/ + +/* This function scans through a branch of a compiled pattern to see whether it +can match the empty string or not. It is called from could_be_empty() +below and from compile_branch() when checking for an unlimited repeat of a +group that can match nothing. Note that first_significant_code() skips over +backward and negative forward assertions when its final argument is TRUE. If we +hit an unclosed bracket, we return "empty" - this means we've struck an inner +bracket whose current branch will already have been scanned. + +Arguments: + code points to start of search + endcode points to where to stop + utf TRUE if in UTF-8 / UTF-16 mode + cd contains pointers to tables etc. + +Returns: TRUE if what is matched could be empty +*/ + +static BOOL +could_be_empty_branch(const pcre_uchar *code, const pcre_uchar *endcode, + BOOL utf, compile_data *cd) +{ +register int c; +for (code = first_significant_code(code + PRIV(OP_lengths)[*code], TRUE); + code < endcode; + code = first_significant_code(code + PRIV(OP_lengths)[c], TRUE)) + { + const pcre_uchar *ccode; + + c = *code; + + /* Skip over forward assertions; the other assertions are skipped by + first_significant_code() with a TRUE final argument. */ + + if (c == OP_ASSERT) + { + do code += GET(code, 1); while (*code == OP_ALT); + c = *code; + continue; + } + + /* For a recursion/subroutine call, if its end has been reached, which + implies a backward reference subroutine call, we can scan it. If it's a + forward reference subroutine call, we can't. To detect forward reference + we have to scan up the list that is kept in the workspace. This function is + called only when doing the real compile, not during the pre-compile that + measures the size of the compiled pattern. */ + + if (c == OP_RECURSE) + { + const pcre_uchar *scode; + BOOL empty_branch; + + /* Test for forward reference */ + + for (scode = cd->start_workspace; scode < cd->hwm; scode += LINK_SIZE) + if (GET(scode, 0) == code + 1 - cd->start_code) return TRUE; + + /* Not a forward reference, test for completed backward reference */ + + empty_branch = FALSE; + scode = cd->start_code + GET(code, 1); + if (GET(scode, 1) == 0) return TRUE; /* Unclosed */ + + /* Completed backwards reference */ + + do + { + if (could_be_empty_branch(scode, endcode, utf, cd)) + { + empty_branch = TRUE; + break; + } + scode += GET(scode, 1); + } + while (*scode == OP_ALT); + + if (!empty_branch) return FALSE; /* All branches are non-empty */ + continue; + } + + /* Groups with zero repeats can of course be empty; skip them. */ + + if (c == OP_BRAZERO || c == OP_BRAMINZERO || c == OP_SKIPZERO || + c == OP_BRAPOSZERO) + { + code += PRIV(OP_lengths)[c]; + do code += GET(code, 1); while (*code == OP_ALT); + c = *code; + continue; + } + + /* A nested group that is already marked as "could be empty" can just be + skipped. */ + + if (c == OP_SBRA || c == OP_SBRAPOS || + c == OP_SCBRA || c == OP_SCBRAPOS) + { + do code += GET(code, 1); while (*code == OP_ALT); + c = *code; + continue; + } + + /* For other groups, scan the branches. */ + + if (c == OP_BRA || c == OP_BRAPOS || + c == OP_CBRA || c == OP_CBRAPOS || + c == OP_ONCE || c == OP_ONCE_NC || + c == OP_COND) + { + BOOL empty_branch; + if (GET(code, 1) == 0) return TRUE; /* Hit unclosed bracket */ + + /* If a conditional group has only one branch, there is a second, implied, + empty branch, so just skip over the conditional, because it could be empty. + Otherwise, scan the individual branches of the group. */ + + if (c == OP_COND && code[GET(code, 1)] != OP_ALT) + code += GET(code, 1); + else + { + empty_branch = FALSE; + do + { + if (!empty_branch && could_be_empty_branch(code, endcode, utf, cd)) + empty_branch = TRUE; + code += GET(code, 1); + } + while (*code == OP_ALT); + if (!empty_branch) return FALSE; /* All branches are non-empty */ + } + + c = *code; + continue; + } + + /* Handle the other opcodes */ + + switch (c) + { + /* Check for quantifiers after a class. XCLASS is used for classes that + cannot be represented just by a bit map. This includes negated single + high-valued characters. The length in PRIV(OP_lengths)[] is zero; the + actual length is stored in the compiled code, so we must update "code" + here. */ + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: + ccode = code += GET(code, 1); + goto CHECK_CLASS_REPEAT; +#endif + + case OP_CLASS: + case OP_NCLASS: + ccode = code + PRIV(OP_lengths)[OP_CLASS]; + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + CHECK_CLASS_REPEAT: +#endif + + switch (*ccode) + { + case OP_CRSTAR: /* These could be empty; continue */ + case OP_CRMINSTAR: + case OP_CRQUERY: + case OP_CRMINQUERY: + break; + + default: /* Non-repeat => class must match */ + case OP_CRPLUS: /* These repeats aren't empty */ + case OP_CRMINPLUS: + return FALSE; + + case OP_CRRANGE: + case OP_CRMINRANGE: + if (GET2(ccode, 1) > 0) return FALSE; /* Minimum > 0 */ + break; + } + break; + + /* Opcodes that must match a character */ + + case OP_PROP: + case OP_NOTPROP: + case OP_EXTUNI: + case OP_NOT_DIGIT: + case OP_DIGIT: + case OP_NOT_WHITESPACE: + case OP_WHITESPACE: + case OP_NOT_WORDCHAR: + case OP_WORDCHAR: + case OP_ANY: + case OP_ALLANY: + case OP_ANYBYTE: + case OP_CHAR: + case OP_CHARI: + case OP_NOT: + case OP_NOTI: + case OP_PLUS: + case OP_MINPLUS: + case OP_POSPLUS: + case OP_EXACT: + case OP_NOTPLUS: + case OP_NOTMINPLUS: + case OP_NOTPOSPLUS: + case OP_NOTEXACT: + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEPOSPLUS: + case OP_TYPEEXACT: + return FALSE; + + /* These are going to continue, as they may be empty, but we have to + fudge the length for the \p and \P cases. */ + + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPOSSTAR: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSQUERY: + if (code[1] == OP_PROP || code[1] == OP_NOTPROP) code += 2; + break; + + /* Same for these */ + + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEPOSUPTO: + if (code[1 + IMM2_SIZE] == OP_PROP + || code[1 + IMM2_SIZE] == OP_NOTPROP) code += 2; + break; + + /* End of branch */ + + case OP_KET: + case OP_KETRMAX: + case OP_KETRMIN: + case OP_KETRPOS: + case OP_ALT: + return TRUE; + + /* In UTF-8 mode, STAR, MINSTAR, POSSTAR, QUERY, MINQUERY, POSQUERY, UPTO, + MINUPTO, and POSUPTO may be followed by a multibyte character */ + +#ifdef SUPPORT_UTF + case OP_STAR: + case OP_STARI: + case OP_MINSTAR: + case OP_MINSTARI: + case OP_POSSTAR: + case OP_POSSTARI: + case OP_QUERY: + case OP_QUERYI: + case OP_MINQUERY: + case OP_MINQUERYI: + case OP_POSQUERY: + case OP_POSQUERYI: + if (utf && HAS_EXTRALEN(code[1])) code += GET_EXTRALEN(code[1]); + break; + + case OP_UPTO: + case OP_UPTOI: + case OP_MINUPTO: + case OP_MINUPTOI: + case OP_POSUPTO: + case OP_POSUPTOI: + if (utf && HAS_EXTRALEN(code[1 + IMM2_SIZE])) code += GET_EXTRALEN(code[1 + IMM2_SIZE]); + break; +#endif + + /* MARK, and PRUNE/SKIP/THEN with an argument must skip over the argument + string. */ + + case OP_MARK: + case OP_PRUNE_ARG: + case OP_SKIP_ARG: + code += code[1]; + break; + + case OP_THEN_ARG: + code += code[1]; + break; + + /* None of the remaining opcodes are required to match a character. */ + + default: + break; + } + } + +return TRUE; +} + + + +/************************************************* +* Scan compiled regex for non-emptiness * +*************************************************/ + +/* This function is called to check for left recursive calls. We want to check +the current branch of the current pattern to see if it could match the empty +string. If it could, we must look outwards for branches at other levels, +stopping when we pass beyond the bracket which is the subject of the recursion. +This function is called only during the real compile, not during the +pre-compile. + +Arguments: + code points to start of the recursion + endcode points to where to stop (current RECURSE item) + bcptr points to the chain of current (unclosed) branch starts + utf TRUE if in UTF-8 / UTF-16 mode + cd pointers to tables etc + +Returns: TRUE if what is matched could be empty +*/ + +static BOOL +could_be_empty(const pcre_uchar *code, const pcre_uchar *endcode, + branch_chain *bcptr, BOOL utf, compile_data *cd) +{ +while (bcptr != NULL && bcptr->current_branch >= code) + { + if (!could_be_empty_branch(bcptr->current_branch, endcode, utf, cd)) + return FALSE; + bcptr = bcptr->outer; + } +return TRUE; +} + + + +/************************************************* +* Check for POSIX class syntax * +*************************************************/ + +/* This function is called when the sequence "[:" or "[." or "[=" is +encountered in a character class. It checks whether this is followed by a +sequence of characters terminated by a matching ":]" or ".]" or "=]". If we +reach an unescaped ']' without the special preceding character, return FALSE. + +Originally, this function only recognized a sequence of letters between the +terminators, but it seems that Perl recognizes any sequence of characters, +though of course unknown POSIX names are subsequently rejected. Perl gives an +"Unknown POSIX class" error for [:f\oo:] for example, where previously PCRE +didn't consider this to be a POSIX class. Likewise for [:1234:]. + +The problem in trying to be exactly like Perl is in the handling of escapes. We +have to be sure that [abc[:x\]pqr] is *not* treated as containing a POSIX +class, but [abc[:x\]pqr:]] is (so that an error can be generated). The code +below handles the special case of \], but does not try to do any other escape +processing. This makes it different from Perl for cases such as [:l\ower:] +where Perl recognizes it as the POSIX class "lower" but PCRE does not recognize +"l\ower". This is a lesser evil that not diagnosing bad classes when Perl does, +I think. + +A user pointed out that PCRE was rejecting [:a[:digit:]] whereas Perl was not. +It seems that the appearance of a nested POSIX class supersedes an apparent +external class. For example, [:a[:digit:]b:] matches "a", "b", ":", or +a digit. + +In Perl, unescaped square brackets may also appear as part of class names. For +example, [:a[:abc]b:] gives unknown POSIX class "[:abc]b:]". However, for +[:a[:abc]b][b:] it gives unknown POSIX class "[:abc]b][b:]", which does not +seem right at all. PCRE does not allow closing square brackets in POSIX class +names. + +Arguments: + ptr pointer to the initial [ + endptr where to return the end pointer + +Returns: TRUE or FALSE +*/ + +static BOOL +check_posix_syntax(const pcre_uchar *ptr, const pcre_uchar **endptr) +{ +int terminator; /* Don't combine these lines; the Solaris cc */ +terminator = *(++ptr); /* compiler warns about "non-constant" initializer. */ +for (++ptr; *ptr != 0; ptr++) + { + if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) + ptr++; + else if (*ptr == CHAR_RIGHT_SQUARE_BRACKET) return FALSE; + else + { + if (*ptr == terminator && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) + { + *endptr = ptr; + return TRUE; + } + if (*ptr == CHAR_LEFT_SQUARE_BRACKET && + (ptr[1] == CHAR_COLON || ptr[1] == CHAR_DOT || + ptr[1] == CHAR_EQUALS_SIGN) && + check_posix_syntax(ptr, endptr)) + return FALSE; + } + } +return FALSE; +} + + + + +/************************************************* +* Check POSIX class name * +*************************************************/ + +/* This function is called to check the name given in a POSIX-style class entry +such as [:alnum:]. + +Arguments: + ptr points to the first letter + len the length of the name + +Returns: a value representing the name, or -1 if unknown +*/ + +static int +check_posix_name(const pcre_uchar *ptr, int len) +{ +const char *pn = posix_names; +register int yield = 0; +while (posix_name_lengths[yield] != 0) + { + if (len == posix_name_lengths[yield] && + STRNCMP_UC_C8(ptr, pn, len) == 0) return yield; + pn += posix_name_lengths[yield] + 1; + yield++; + } +return -1; +} + + +/************************************************* +* Adjust OP_RECURSE items in repeated group * +*************************************************/ + +/* OP_RECURSE items contain an offset from the start of the regex to the group +that is referenced. This means that groups can be replicated for fixed +repetition simply by copying (because the recursion is allowed to refer to +earlier groups that are outside the current group). However, when a group is +optional (i.e. the minimum quantifier is zero), OP_BRAZERO or OP_SKIPZERO is +inserted before it, after it has been compiled. This means that any OP_RECURSE +items within it that refer to the group itself or any contained groups have to +have their offsets adjusted. That one of the jobs of this function. Before it +is called, the partially compiled regex must be temporarily terminated with +OP_END. + +This function has been extended with the possibility of forward references for +recursions and subroutine calls. It must also check the list of such references +for the group we are dealing with. If it finds that one of the recursions in +the current group is on this list, it adjusts the offset in the list, not the +value in the reference (which is a group number). + +Arguments: + group points to the start of the group + adjust the amount by which the group is to be moved + utf TRUE in UTF-8 / UTF-16 mode + cd contains pointers to tables etc. + save_hwm the hwm forward reference pointer at the start of the group + +Returns: nothing +*/ + +static void +adjust_recurse(pcre_uchar *group, int adjust, BOOL utf, compile_data *cd, + pcre_uchar *save_hwm) +{ +pcre_uchar *ptr = group; + +while ((ptr = (pcre_uchar *)find_recurse(ptr, utf)) != NULL) + { + int offset; + pcre_uchar *hc; + + /* See if this recursion is on the forward reference list. If so, adjust the + reference. */ + + for (hc = save_hwm; hc < cd->hwm; hc += LINK_SIZE) + { + offset = GET(hc, 0); + if (cd->start_code + offset == ptr + 1) + { + PUT(hc, 0, offset + adjust); + break; + } + } + + /* Otherwise, adjust the recursion offset if it's after the start of this + group. */ + + if (hc >= cd->hwm) + { + offset = GET(ptr, 1); + if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust); + } + + ptr += 1 + LINK_SIZE; + } +} + + + +/************************************************* +* Insert an automatic callout point * +*************************************************/ + +/* This function is called when the PCRE_AUTO_CALLOUT option is set, to insert +callout points before each pattern item. + +Arguments: + code current code pointer + ptr current pattern pointer + cd pointers to tables etc + +Returns: new code pointer +*/ + +static pcre_uchar * +auto_callout(pcre_uchar *code, const pcre_uchar *ptr, compile_data *cd) +{ +*code++ = OP_CALLOUT; +*code++ = 255; +PUT(code, 0, (int)(ptr - cd->start_pattern)); /* Pattern offset */ +PUT(code, LINK_SIZE, 0); /* Default length */ +return code + 2 * LINK_SIZE; +} + + + +/************************************************* +* Complete a callout item * +*************************************************/ + +/* A callout item contains the length of the next item in the pattern, which +we can't fill in till after we have reached the relevant point. This is used +for both automatic and manual callouts. + +Arguments: + previous_callout points to previous callout item + ptr current pattern pointer + cd pointers to tables etc + +Returns: nothing +*/ + +static void +complete_callout(pcre_uchar *previous_callout, const pcre_uchar *ptr, compile_data *cd) +{ +int length = (int)(ptr - cd->start_pattern - GET(previous_callout, 2)); +PUT(previous_callout, 2 + LINK_SIZE, length); +} + + + +#ifdef SUPPORT_UCP +/************************************************* +* Get othercase range * +*************************************************/ + +/* This function is passed the start and end of a class range, in UTF-8 mode +with UCP support. It searches up the characters, looking for internal ranges of +characters in the "other" case. Each call returns the next one, updating the +start address. + +Arguments: + cptr points to starting character value; updated + d end value + ocptr where to put start of othercase range + odptr where to put end of othercase range + +Yield: TRUE when range returned; FALSE when no more +*/ + +static BOOL +get_othercase_range(unsigned int *cptr, unsigned int d, unsigned int *ocptr, + unsigned int *odptr) +{ +unsigned int c, othercase, next; + +for (c = *cptr; c <= d; c++) + { if ((othercase = UCD_OTHERCASE(c)) != c) break; } + +if (c > d) return FALSE; + +*ocptr = othercase; +next = othercase + 1; + +for (++c; c <= d; c++) + { + if (UCD_OTHERCASE(c) != next) break; + next++; + } + +*odptr = next - 1; +*cptr = c; + +return TRUE; +} + + + +/************************************************* +* Check a character and a property * +*************************************************/ + +/* This function is called by check_auto_possessive() when a property item +is adjacent to a fixed character. + +Arguments: + c the character + ptype the property type + pdata the data for the type + negated TRUE if it's a negated property (\P or \p{^) + +Returns: TRUE if auto-possessifying is OK +*/ + +static BOOL +check_char_prop(int c, int ptype, int pdata, BOOL negated) +{ +const ucd_record *prop = GET_UCD(c); +switch(ptype) + { + case PT_LAMP: + return (prop->chartype == ucp_Lu || + prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt) == negated; + + case PT_GC: + return (pdata == PRIV(ucp_gentype)[prop->chartype]) == negated; + + case PT_PC: + return (pdata == prop->chartype) == negated; + + case PT_SC: + return (pdata == prop->script) == negated; + + /* These are specials */ + + case PT_ALNUM: + return (PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N) == negated; + + case PT_SPACE: /* Perl space */ + return (PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR) + == negated; + + case PT_PXSPACE: /* POSIX space */ + return (PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_VT || + c == CHAR_FF || c == CHAR_CR) + == negated; + + case PT_WORD: + return (PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || + c == CHAR_UNDERSCORE) == negated; + } +return FALSE; +} +#endif /* SUPPORT_UCP */ + + + +/************************************************* +* Check if auto-possessifying is possible * +*************************************************/ + +/* This function is called for unlimited repeats of certain items, to see +whether the next thing could possibly match the repeated item. If not, it makes +sense to automatically possessify the repeated item. + +Arguments: + previous pointer to the repeated opcode + utf TRUE in UTF-8 / UTF-16 mode + ptr next character in pattern + options options bits + cd contains pointers to tables etc. + +Returns: TRUE if possessifying is wanted +*/ + +static BOOL +check_auto_possessive(const pcre_uchar *previous, BOOL utf, + const pcre_uchar *ptr, int options, compile_data *cd) +{ +pcre_int32 c, next; +int op_code = *previous++; + +/* Skip whitespace and comments in extended mode */ + +if ((options & PCRE_EXTENDED) != 0) + { + for (;;) + { + while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_space) != 0) ptr++; + if (*ptr == CHAR_NUMBER_SIGN) + { + ptr++; + while (*ptr != 0) + { + if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; } + ptr++; +#ifdef SUPPORT_UTF + if (utf) FORWARDCHAR(ptr); +#endif + } + } + else break; + } + } + +/* If the next item is one that we can handle, get its value. A non-negative +value is a character, a negative value is an escape value. */ + +if (*ptr == CHAR_BACKSLASH) + { + int temperrorcode = 0; + next = check_escape(&ptr, &temperrorcode, cd->bracount, options, FALSE); + if (temperrorcode != 0) return FALSE; + ptr++; /* Point after the escape sequence */ + } +else if (!MAX_255(*ptr) || (cd->ctypes[*ptr] & ctype_meta) == 0) + { +#ifdef SUPPORT_UTF + if (utf) { GETCHARINC(next, ptr); } else +#endif + next = *ptr++; + } +else return FALSE; + +/* Skip whitespace and comments in extended mode */ + +if ((options & PCRE_EXTENDED) != 0) + { + for (;;) + { + while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_space) != 0) ptr++; + if (*ptr == CHAR_NUMBER_SIGN) + { + ptr++; + while (*ptr != 0) + { + if (IS_NEWLINE(ptr)) { ptr += cd->nllen; break; } + ptr++; +#ifdef SUPPORT_UTF + if (utf) FORWARDCHAR(ptr); +#endif + } + } + else break; + } + } + +/* If the next thing is itself optional, we have to give up. */ + +if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK || + STRNCMP_UC_C8(ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0) + return FALSE; + +/* Now compare the next item with the previous opcode. First, handle cases when +the next item is a character. */ + +if (next >= 0) switch(op_code) + { + case OP_CHAR: +#ifdef SUPPORT_UTF + GETCHARTEST(c, previous); +#else + c = *previous; +#endif + return c != next; + + /* For CHARI (caseless character) we must check the other case. If we have + Unicode property support, we can use it to test the other case of + high-valued characters. */ + + case OP_CHARI: +#ifdef SUPPORT_UTF + GETCHARTEST(c, previous); +#else + c = *previous; +#endif + if (c == next) return FALSE; +#ifdef SUPPORT_UTF + if (utf) + { + unsigned int othercase; + if (next < 128) othercase = cd->fcc[next]; else +#ifdef SUPPORT_UCP + othercase = UCD_OTHERCASE((unsigned int)next); +#else + othercase = NOTACHAR; +#endif + return (unsigned int)c != othercase; + } + else +#endif /* SUPPORT_UTF */ + return (c != TABLE_GET((unsigned int)next, cd->fcc, next)); /* Non-UTF-8 mode */ + + case OP_NOT: +#ifdef SUPPORT_UTF + GETCHARTEST(c, previous); +#else + c = *previous; +#endif + return c == next; + + case OP_NOTI: +#ifdef SUPPORT_UTF + GETCHARTEST(c, previous); +#else + c = *previous; +#endif + if (c == next) return TRUE; +#ifdef SUPPORT_UTF + if (utf) + { + unsigned int othercase; + if (next < 128) othercase = cd->fcc[next]; else +#ifdef SUPPORT_UCP + othercase = UCD_OTHERCASE((unsigned int)next); +#else + othercase = NOTACHAR; +#endif + return (unsigned int)c == othercase; + } + else +#endif /* SUPPORT_UTF */ + return (c == TABLE_GET((unsigned int)next, cd->fcc, next)); /* Non-UTF-8 mode */ + + /* Note that OP_DIGIT etc. are generated only when PCRE_UCP is *not* set. + When it is set, \d etc. are converted into OP_(NOT_)PROP codes. */ + + case OP_DIGIT: + return next > 255 || (cd->ctypes[next] & ctype_digit) == 0; + + case OP_NOT_DIGIT: + return next <= 255 && (cd->ctypes[next] & ctype_digit) != 0; + + case OP_WHITESPACE: + return next > 255 || (cd->ctypes[next] & ctype_space) == 0; + + case OP_NOT_WHITESPACE: + return next <= 255 && (cd->ctypes[next] & ctype_space) != 0; + + case OP_WORDCHAR: + return next > 255 || (cd->ctypes[next] & ctype_word) == 0; + + case OP_NOT_WORDCHAR: + return next <= 255 && (cd->ctypes[next] & ctype_word) != 0; + + case OP_HSPACE: + case OP_NOT_HSPACE: + switch(next) + { + case 0x09: + case 0x20: + case 0xa0: + case 0x1680: + case 0x180e: + case 0x2000: + case 0x2001: + case 0x2002: + case 0x2003: + case 0x2004: + case 0x2005: + case 0x2006: + case 0x2007: + case 0x2008: + case 0x2009: + case 0x200A: + case 0x202f: + case 0x205f: + case 0x3000: + return op_code == OP_NOT_HSPACE; + default: + return op_code != OP_NOT_HSPACE; + } + + case OP_ANYNL: + case OP_VSPACE: + case OP_NOT_VSPACE: + switch(next) + { + case 0x0a: + case 0x0b: + case 0x0c: + case 0x0d: + case 0x85: + case 0x2028: + case 0x2029: + return op_code == OP_NOT_VSPACE; + default: + return op_code != OP_NOT_VSPACE; + } + +#ifdef SUPPORT_UCP + case OP_PROP: + return check_char_prop(next, previous[0], previous[1], FALSE); + + case OP_NOTPROP: + return check_char_prop(next, previous[0], previous[1], TRUE); +#endif + + default: + return FALSE; + } + + +/* Handle the case when the next item is \d, \s, etc. Note that when PCRE_UCP +is set, \d turns into ESC_du rather than ESC_d, etc., so ESC_d etc. are +generated only when PCRE_UCP is *not* set, that is, when only ASCII +characteristics are recognized. Similarly, the opcodes OP_DIGIT etc. are +replaced by OP_PROP codes when PCRE_UCP is set. */ + +switch(op_code) + { + case OP_CHAR: + case OP_CHARI: +#ifdef SUPPORT_UTF + GETCHARTEST(c, previous); +#else + c = *previous; +#endif + switch(-next) + { + case ESC_d: + return c > 255 || (cd->ctypes[c] & ctype_digit) == 0; + + case ESC_D: + return c <= 255 && (cd->ctypes[c] & ctype_digit) != 0; + + case ESC_s: + return c > 255 || (cd->ctypes[c] & ctype_space) == 0; + + case ESC_S: + return c <= 255 && (cd->ctypes[c] & ctype_space) != 0; + + case ESC_w: + return c > 255 || (cd->ctypes[c] & ctype_word) == 0; + + case ESC_W: + return c <= 255 && (cd->ctypes[c] & ctype_word) != 0; + + case ESC_h: + case ESC_H: + switch(c) + { + case 0x09: + case 0x20: + case 0xa0: + case 0x1680: + case 0x180e: + case 0x2000: + case 0x2001: + case 0x2002: + case 0x2003: + case 0x2004: + case 0x2005: + case 0x2006: + case 0x2007: + case 0x2008: + case 0x2009: + case 0x200A: + case 0x202f: + case 0x205f: + case 0x3000: + return -next != ESC_h; + default: + return -next == ESC_h; + } + + case ESC_v: + case ESC_V: + switch(c) + { + case 0x0a: + case 0x0b: + case 0x0c: + case 0x0d: + case 0x85: + case 0x2028: + case 0x2029: + return -next != ESC_v; + default: + return -next == ESC_v; + } + + /* When PCRE_UCP is set, these values get generated for \d etc. Find + their substitutions and process them. The result will always be either + -ESC_p or -ESC_P. Then fall through to process those values. */ + +#ifdef SUPPORT_UCP + case ESC_du: + case ESC_DU: + case ESC_wu: + case ESC_WU: + case ESC_su: + case ESC_SU: + { + int temperrorcode = 0; + ptr = substitutes[-next - ESC_DU]; + next = check_escape(&ptr, &temperrorcode, 0, options, FALSE); + if (temperrorcode != 0) return FALSE; + ptr++; /* For compatibility */ + } + /* Fall through */ + + case ESC_p: + case ESC_P: + { + int ptype, pdata, errorcodeptr; + BOOL negated; + + ptr--; /* Make ptr point at the p or P */ + ptype = get_ucp(&ptr, &negated, &pdata, &errorcodeptr); + if (ptype < 0) return FALSE; + ptr++; /* Point past the final curly ket */ + + /* If the property item is optional, we have to give up. (When generated + from \d etc by PCRE_UCP, this test will have been applied much earlier, + to the original \d etc. At this point, ptr will point to a zero byte. */ + + if (*ptr == CHAR_ASTERISK || *ptr == CHAR_QUESTION_MARK || + STRNCMP_UC_C8(ptr, STR_LEFT_CURLY_BRACKET STR_0 STR_COMMA, 3) == 0) + return FALSE; + + /* Do the property check. */ + + return check_char_prop(c, ptype, pdata, (next == -ESC_P) != negated); + } +#endif + + default: + return FALSE; + } + + /* In principle, support for Unicode properties should be integrated here as + well. It means re-organizing the above code so as to get hold of the property + values before switching on the op-code. However, I wonder how many patterns + combine ASCII \d etc with Unicode properties? (Note that if PCRE_UCP is set, + these op-codes are never generated.) */ + + case OP_DIGIT: + return next == -ESC_D || next == -ESC_s || next == -ESC_W || + next == -ESC_h || next == -ESC_v || next == -ESC_R; + + case OP_NOT_DIGIT: + return next == -ESC_d; + + case OP_WHITESPACE: + return next == -ESC_S || next == -ESC_d || next == -ESC_w; + + case OP_NOT_WHITESPACE: + return next == -ESC_s || next == -ESC_h || next == -ESC_v || next == -ESC_R; + + case OP_HSPACE: + return next == -ESC_S || next == -ESC_H || next == -ESC_d || + next == -ESC_w || next == -ESC_v || next == -ESC_R; + + case OP_NOT_HSPACE: + return next == -ESC_h; + + /* Can't have \S in here because VT matches \S (Perl anomaly) */ + case OP_ANYNL: + case OP_VSPACE: + return next == -ESC_V || next == -ESC_d || next == -ESC_w; + + case OP_NOT_VSPACE: + return next == -ESC_v || next == -ESC_R; + + case OP_WORDCHAR: + return next == -ESC_W || next == -ESC_s || next == -ESC_h || + next == -ESC_v || next == -ESC_R; + + case OP_NOT_WORDCHAR: + return next == -ESC_w || next == -ESC_d; + + default: + return FALSE; + } + +/* Control does not reach here */ +} + + + +/************************************************* +* Compile one branch * +*************************************************/ + +/* Scan the pattern, compiling it into the a vector. If the options are +changed during the branch, the pointer is used to change the external options +bits. This function is used during the pre-compile phase when we are trying +to find out the amount of memory needed, as well as during the real compile +phase. The value of lengthptr distinguishes the two phases. + +Arguments: + optionsptr pointer to the option bits + codeptr points to the pointer to the current code point + ptrptr points to the current pattern pointer + errorcodeptr points to error code variable + firstcharptr set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE) + reqcharptr set to the last literal character required, else < 0 + bcptr points to current branch chain + cond_depth conditional nesting depth + cd contains pointers to tables etc. + lengthptr NULL during the real compile phase + points to length accumulator during pre-compile phase + +Returns: TRUE on success + FALSE, with *errorcodeptr set non-zero on error +*/ + +static BOOL +compile_branch(int *optionsptr, pcre_uchar **codeptr, + const pcre_uchar **ptrptr, int *errorcodeptr, pcre_int32 *firstcharptr, + pcre_int32 *reqcharptr, branch_chain *bcptr, int cond_depth, + compile_data *cd, int *lengthptr) +{ +int repeat_type, op_type; +int repeat_min = 0, repeat_max = 0; /* To please picky compilers */ +int bravalue = 0; +int greedy_default, greedy_non_default; +pcre_int32 firstchar, reqchar; +pcre_int32 zeroreqchar, zerofirstchar; +pcre_int32 req_caseopt, reqvary, tempreqvary; +int options = *optionsptr; /* May change dynamically */ +int after_manual_callout = 0; +int length_prevgroup = 0; +register int c; +register pcre_uchar *code = *codeptr; +pcre_uchar *last_code = code; +pcre_uchar *orig_code = code; +pcre_uchar *tempcode; +BOOL inescq = FALSE; +BOOL groupsetfirstchar = FALSE; +const pcre_uchar *ptr = *ptrptr; +const pcre_uchar *tempptr; +const pcre_uchar *nestptr = NULL; +pcre_uchar *previous = NULL; +pcre_uchar *previous_callout = NULL; +pcre_uchar *save_hwm = NULL; +pcre_uint8 classbits[32]; + +/* We can fish out the UTF-8 setting once and for all into a BOOL, but we +must not do this for other options (e.g. PCRE_EXTENDED) because they may change +dynamically as we process the pattern. */ + +#ifdef SUPPORT_UTF +/* PCRE_UTF16 has the same value as PCRE_UTF8. */ +BOOL utf = (options & PCRE_UTF8) != 0; +pcre_uchar utf_chars[6]; +#else +BOOL utf = FALSE; +#endif + +/* Helper variables for OP_XCLASS opcode (for characters > 255). */ + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 +BOOL xclass; +pcre_uchar *class_uchardata; +pcre_uchar *class_uchardata_base; +#endif + +#ifdef PCRE_DEBUG +if (lengthptr != NULL) DPRINTF((">> start branch\n")); +#endif + +/* Set up the default and non-default settings for greediness */ + +greedy_default = ((options & PCRE_UNGREEDY) != 0); +greedy_non_default = greedy_default ^ 1; + +/* Initialize no first byte, no required byte. REQ_UNSET means "no char +matching encountered yet". It gets changed to REQ_NONE if we hit something that +matches a non-fixed char first char; reqchar just remains unset if we never +find one. + +When we hit a repeat whose minimum is zero, we may have to adjust these values +to take the zero repeat into account. This is implemented by setting them to +zerofirstbyte and zeroreqchar when such a repeat is encountered. The individual +item types that can be repeated set these backoff variables appropriately. */ + +firstchar = reqchar = zerofirstchar = zeroreqchar = REQ_UNSET; + +/* The variable req_caseopt contains either the REQ_CASELESS value +or zero, according to the current setting of the caseless flag. The +REQ_CASELESS leaves the lower 28 bit empty. It is added into the +firstchar or reqchar variables to record the case status of the +value. This is used only for ASCII characters. */ + +req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS:0; + +/* Switch on next character until the end of the branch */ + +for (;; ptr++) + { + BOOL negate_class; + BOOL should_flip_negation; + BOOL possessive_quantifier; + BOOL is_quantifier; + BOOL is_recurse; + BOOL reset_bracount; + int class_has_8bitchar; + int class_single_char; + int newoptions; + int recno; + int refsign; + int skipbytes; + int subreqchar; + int subfirstchar; + int terminator; + int mclength; + int tempbracount; + pcre_uchar mcbuffer[8]; + + /* Get next character in the pattern */ + + c = *ptr; + + /* If we are at the end of a nested substitution, revert to the outer level + string. Nesting only happens one level deep. */ + + if (c == 0 && nestptr != NULL) + { + ptr = nestptr; + nestptr = NULL; + c = *ptr; + } + + /* If we are in the pre-compile phase, accumulate the length used for the + previous cycle of this loop. */ + + if (lengthptr != NULL) + { +#ifdef PCRE_DEBUG + if (code > cd->hwm) cd->hwm = code; /* High water info */ +#endif + if (code > cd->start_workspace + cd->workspace_size - + WORK_SIZE_SAFETY_MARGIN) /* Check for overrun */ + { + *errorcodeptr = ERR52; + goto FAILED; + } + + /* There is at least one situation where code goes backwards: this is the + case of a zero quantifier after a class (e.g. [ab]{0}). At compile time, + the class is simply eliminated. However, it is created first, so we have to + allow memory for it. Therefore, don't ever reduce the length at this point. + */ + + if (code < last_code) code = last_code; + + /* Paranoid check for integer overflow */ + + if (OFLOW_MAX - *lengthptr < code - last_code) + { + *errorcodeptr = ERR20; + goto FAILED; + } + + *lengthptr += (int)(code - last_code); + DPRINTF(("length=%d added %d c=%c (0x%x)\n", *lengthptr, + (int)(code - last_code), c, c)); + + /* If "previous" is set and it is not at the start of the work space, move + it back to there, in order to avoid filling up the work space. Otherwise, + if "previous" is NULL, reset the current code pointer to the start. */ + + if (previous != NULL) + { + if (previous > orig_code) + { + memmove(orig_code, previous, IN_UCHARS(code - previous)); + code -= previous - orig_code; + previous = orig_code; + } + } + else code = orig_code; + + /* Remember where this code item starts so we can pick up the length + next time round. */ + + last_code = code; + } + + /* In the real compile phase, just check the workspace used by the forward + reference list. */ + + else if (cd->hwm > cd->start_workspace + cd->workspace_size - + WORK_SIZE_SAFETY_MARGIN) + { + *errorcodeptr = ERR52; + goto FAILED; + } + + /* If in \Q...\E, check for the end; if not, we have a literal */ + + if (inescq && c != 0) + { + if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E) + { + inescq = FALSE; + ptr++; + continue; + } + else + { + if (previous_callout != NULL) + { + if (lengthptr == NULL) /* Don't attempt in pre-compile phase */ + complete_callout(previous_callout, ptr, cd); + previous_callout = NULL; + } + if ((options & PCRE_AUTO_CALLOUT) != 0) + { + previous_callout = code; + code = auto_callout(code, ptr, cd); + } + goto NORMAL_CHAR; + } + } + + /* Fill in length of a previous callout, except when the next thing is + a quantifier. */ + + is_quantifier = + c == CHAR_ASTERISK || c == CHAR_PLUS || c == CHAR_QUESTION_MARK || + (c == CHAR_LEFT_CURLY_BRACKET && is_counted_repeat(ptr+1)); + + if (!is_quantifier && previous_callout != NULL && + after_manual_callout-- <= 0) + { + if (lengthptr == NULL) /* Don't attempt in pre-compile phase */ + complete_callout(previous_callout, ptr, cd); + previous_callout = NULL; + } + + /* In extended mode, skip white space and comments. */ + + if ((options & PCRE_EXTENDED) != 0) + { + if (MAX_255(*ptr) && (cd->ctypes[c] & ctype_space) != 0) continue; + if (c == CHAR_NUMBER_SIGN) + { + ptr++; + while (*ptr != 0) + { + if (IS_NEWLINE(ptr)) { ptr += cd->nllen - 1; break; } + ptr++; +#ifdef SUPPORT_UTF + if (utf) FORWARDCHAR(ptr); +#endif + } + if (*ptr != 0) continue; + + /* Else fall through to handle end of string */ + c = 0; + } + } + + /* No auto callout for quantifiers. */ + + if ((options & PCRE_AUTO_CALLOUT) != 0 && !is_quantifier) + { + previous_callout = code; + code = auto_callout(code, ptr, cd); + } + + switch(c) + { + /* ===================================================================*/ + case 0: /* The branch terminates at string end */ + case CHAR_VERTICAL_LINE: /* or | or ) */ + case CHAR_RIGHT_PARENTHESIS: + *firstcharptr = firstchar; + *reqcharptr = reqchar; + *codeptr = code; + *ptrptr = ptr; + if (lengthptr != NULL) + { + if (OFLOW_MAX - *lengthptr < code - last_code) + { + *errorcodeptr = ERR20; + goto FAILED; + } + *lengthptr += (int)(code - last_code); /* To include callout length */ + DPRINTF((">> end branch\n")); + } + return TRUE; + + + /* ===================================================================*/ + /* Handle single-character metacharacters. In multiline mode, ^ disables + the setting of any following char as a first character. */ + + case CHAR_CIRCUMFLEX_ACCENT: + previous = NULL; + if ((options & PCRE_MULTILINE) != 0) + { + if (firstchar == REQ_UNSET) firstchar = REQ_NONE; + *code++ = OP_CIRCM; + } + else *code++ = OP_CIRC; + break; + + case CHAR_DOLLAR_SIGN: + previous = NULL; + *code++ = ((options & PCRE_MULTILINE) != 0)? OP_DOLLM : OP_DOLL; + break; + + /* There can never be a first char if '.' is first, whatever happens about + repeats. The value of reqchar doesn't change either. */ + + case CHAR_DOT: + if (firstchar == REQ_UNSET) firstchar = REQ_NONE; + zerofirstchar = firstchar; + zeroreqchar = reqchar; + previous = code; + *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY; + break; + + + /* ===================================================================*/ + /* Character classes. If the included characters are all < 256, we build a + 32-byte bitmap of the permitted characters, except in the special case + where there is only one such character. For negated classes, we build the + map as usual, then invert it at the end. However, we use a different opcode + so that data characters > 255 can be handled correctly. + + If the class contains characters outside the 0-255 range, a different + opcode is compiled. It may optionally have a bit map for characters < 256, + but those above are are explicitly listed afterwards. A flag byte tells + whether the bitmap is present, and whether this is a negated class or not. + + In JavaScript compatibility mode, an isolated ']' causes an error. In + default (Perl) mode, it is treated as a data character. */ + + case CHAR_RIGHT_SQUARE_BRACKET: + if ((cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0) + { + *errorcodeptr = ERR64; + goto FAILED; + } + goto NORMAL_CHAR; + + case CHAR_LEFT_SQUARE_BRACKET: + previous = code; + + /* PCRE supports POSIX class stuff inside a class. Perl gives an error if + they are encountered at the top level, so we'll do that too. */ + + if ((ptr[1] == CHAR_COLON || ptr[1] == CHAR_DOT || + ptr[1] == CHAR_EQUALS_SIGN) && + check_posix_syntax(ptr, &tempptr)) + { + *errorcodeptr = (ptr[1] == CHAR_COLON)? ERR13 : ERR31; + goto FAILED; + } + + /* If the first character is '^', set the negation flag and skip it. Also, + if the first few characters (either before or after ^) are \Q\E or \E we + skip them too. This makes for compatibility with Perl. */ + + negate_class = FALSE; + for (;;) + { + c = *(++ptr); + if (c == CHAR_BACKSLASH) + { + if (ptr[1] == CHAR_E) + ptr++; + else if (STRNCMP_UC_C8(ptr + 1, STR_Q STR_BACKSLASH STR_E, 3) == 0) + ptr += 3; + else + break; + } + else if (!negate_class && c == CHAR_CIRCUMFLEX_ACCENT) + negate_class = TRUE; + else break; + } + + /* Empty classes are allowed in JavaScript compatibility mode. Otherwise, + an initial ']' is taken as a data character -- the code below handles + that. In JS mode, [] must always fail, so generate OP_FAIL, whereas + [^] must match any character, so generate OP_ALLANY. */ + + if (c == CHAR_RIGHT_SQUARE_BRACKET && + (cd->external_options & PCRE_JAVASCRIPT_COMPAT) != 0) + { + *code++ = negate_class? OP_ALLANY : OP_FAIL; + if (firstchar == REQ_UNSET) firstchar = REQ_NONE; + zerofirstchar = firstchar; + break; + } + + /* If a class contains a negative special such as \S, we need to flip the + negation flag at the end, so that support for characters > 255 works + correctly (they are all included in the class). */ + + should_flip_negation = FALSE; + + /* For optimization purposes, we track some properties of the class. + class_has_8bitchar will be non-zero, if the class contains at least one + < 256 character. class_single_char will be 1 if the class contains only + a single character. */ + + class_has_8bitchar = 0; + class_single_char = 0; + + /* Initialize the 32-char bit map to all zeros. We build the map in a + temporary bit of memory, in case the class contains only 1 character (less + than 256), because in that case the compiled code doesn't use the bit map. + */ + + memset(classbits, 0, 32 * sizeof(pcre_uint8)); + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + xclass = FALSE; /* No chars >= 256 */ + class_uchardata = code + LINK_SIZE + 2; /* For UTF-8 items */ + class_uchardata_base = class_uchardata; /* For resetting in pass 1 */ +#endif + + /* Process characters until ] is reached. By writing this as a "do" it + means that an initial ] is taken as a data character. At the start of the + loop, c contains the first byte of the character. */ + + if (c != 0) do + { + const pcre_uchar *oldptr; + +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(c)) + { /* Braces are required because the */ + GETCHARLEN(c, ptr, ptr); /* macro generates multiple statements */ + } +#endif + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + /* In the pre-compile phase, accumulate the length of any extra + data and reset the pointer. This is so that very large classes that + contain a zillion > 255 characters no longer overwrite the work space + (which is on the stack). */ + + if (lengthptr != NULL) + { + *lengthptr += class_uchardata - class_uchardata_base; + class_uchardata = class_uchardata_base; + } +#endif + + /* Inside \Q...\E everything is literal except \E */ + + if (inescq) + { + if (c == CHAR_BACKSLASH && ptr[1] == CHAR_E) /* If we are at \E */ + { + inescq = FALSE; /* Reset literal state */ + ptr++; /* Skip the 'E' */ + continue; /* Carry on with next */ + } + goto CHECK_RANGE; /* Could be range if \E follows */ + } + + /* Handle POSIX class names. Perl allows a negation extension of the + form [:^name:]. A square bracket that doesn't match the syntax is + treated as a literal. We also recognize the POSIX constructions + [.ch.] and [=ch=] ("collating elements") and fault them, as Perl + 5.6 and 5.8 do. */ + + if (c == CHAR_LEFT_SQUARE_BRACKET && + (ptr[1] == CHAR_COLON || ptr[1] == CHAR_DOT || + ptr[1] == CHAR_EQUALS_SIGN) && check_posix_syntax(ptr, &tempptr)) + { + BOOL local_negate = FALSE; + int posix_class, taboffset, tabopt; + register const pcre_uint8 *cbits = cd->cbits; + pcre_uint8 pbits[32]; + + if (ptr[1] != CHAR_COLON) + { + *errorcodeptr = ERR31; + goto FAILED; + } + + ptr += 2; + if (*ptr == CHAR_CIRCUMFLEX_ACCENT) + { + local_negate = TRUE; + should_flip_negation = TRUE; /* Note negative special */ + ptr++; + } + + posix_class = check_posix_name(ptr, (int)(tempptr - ptr)); + if (posix_class < 0) + { + *errorcodeptr = ERR30; + goto FAILED; + } + + /* If matching is caseless, upper and lower are converted to + alpha. This relies on the fact that the class table starts with + alpha, lower, upper as the first 3 entries. */ + + if ((options & PCRE_CASELESS) != 0 && posix_class <= 2) + posix_class = 0; + + /* When PCRE_UCP is set, some of the POSIX classes are converted to + different escape sequences that use Unicode properties. */ + +#ifdef SUPPORT_UCP + if ((options & PCRE_UCP) != 0) + { + int pc = posix_class + ((local_negate)? POSIX_SUBSIZE/2 : 0); + if (posix_substitutes[pc] != NULL) + { + nestptr = tempptr + 1; + ptr = posix_substitutes[pc] - 1; + continue; + } + } +#endif + /* In the non-UCP case, we build the bit map for the POSIX class in a + chunk of local store because we may be adding and subtracting from it, + and we don't want to subtract bits that may be in the main map already. + At the end we or the result into the bit map that is being built. */ + + posix_class *= 3; + + /* Copy in the first table (always present) */ + + memcpy(pbits, cbits + posix_class_maps[posix_class], + 32 * sizeof(pcre_uint8)); + + /* If there is a second table, add or remove it as required. */ + + taboffset = posix_class_maps[posix_class + 1]; + tabopt = posix_class_maps[posix_class + 2]; + + if (taboffset >= 0) + { + if (tabopt >= 0) + for (c = 0; c < 32; c++) pbits[c] |= cbits[c + taboffset]; + else + for (c = 0; c < 32; c++) pbits[c] &= ~cbits[c + taboffset]; + } + + /* Not see if we need to remove any special characters. An option + value of 1 removes vertical space and 2 removes underscore. */ + + if (tabopt < 0) tabopt = -tabopt; + if (tabopt == 1) pbits[1] &= ~0x3c; + else if (tabopt == 2) pbits[11] &= 0x7f; + + /* Add the POSIX table or its complement into the main table that is + being built and we are done. */ + + if (local_negate) + for (c = 0; c < 32; c++) classbits[c] |= ~pbits[c]; + else + for (c = 0; c < 32; c++) classbits[c] |= pbits[c]; + + ptr = tempptr + 1; + /* Every class contains at least one < 256 characters. */ + class_has_8bitchar = 1; + /* Every class contains at least two characters. */ + class_single_char = 2; + continue; /* End of POSIX syntax handling */ + } + + /* Backslash may introduce a single character, or it may introduce one + of the specials, which just set a flag. The sequence \b is a special + case. Inside a class (and only there) it is treated as backspace. We + assume that other escapes have more than one character in them, so + speculatively set both class_has_8bitchar and class_single_char bigger + than one. Unrecognized escapes fall through and are either treated + as literal characters (by default), or are faulted if + PCRE_EXTRA is set. */ + + if (c == CHAR_BACKSLASH) + { + c = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE); + if (*errorcodeptr != 0) goto FAILED; + + if (-c == ESC_b) c = CHAR_BS; /* \b is backspace in a class */ + else if (-c == ESC_N) /* \N is not supported in a class */ + { + *errorcodeptr = ERR71; + goto FAILED; + } + else if (-c == ESC_Q) /* Handle start of quoted string */ + { + if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E) + { + ptr += 2; /* avoid empty string */ + } + else inescq = TRUE; + continue; + } + else if (-c == ESC_E) continue; /* Ignore orphan \E */ + + if (c < 0) + { + register const pcre_uint8 *cbits = cd->cbits; + /* Every class contains at least two < 256 characters. */ + class_has_8bitchar++; + /* Every class contains at least two characters. */ + class_single_char += 2; + + switch (-c) + { +#ifdef SUPPORT_UCP + case ESC_du: /* These are the values given for \d etc */ + case ESC_DU: /* when PCRE_UCP is set. We replace the */ + case ESC_wu: /* escape sequence with an appropriate \p */ + case ESC_WU: /* or \P to test Unicode properties instead */ + case ESC_su: /* of the default ASCII testing. */ + case ESC_SU: + nestptr = ptr; + ptr = substitutes[-c - ESC_DU] - 1; /* Just before substitute */ + class_has_8bitchar--; /* Undo! */ + continue; +#endif + case ESC_d: + for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_digit]; + continue; + + case ESC_D: + should_flip_negation = TRUE; + for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_digit]; + continue; + + case ESC_w: + for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_word]; + continue; + + case ESC_W: + should_flip_negation = TRUE; + for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word]; + continue; + + /* Perl 5.004 onwards omits VT from \s, but we must preserve it + if it was previously set by something earlier in the character + class. */ + + case ESC_s: + classbits[0] |= cbits[cbit_space]; + classbits[1] |= cbits[cbit_space+1] & ~0x08; + for (c = 2; c < 32; c++) classbits[c] |= cbits[c+cbit_space]; + continue; + + case ESC_S: + should_flip_negation = TRUE; + for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space]; + classbits[1] |= 0x08; /* Perl 5.004 onwards omits VT from \s */ + continue; + + case ESC_h: + SETBIT(classbits, 0x09); /* VT */ + SETBIT(classbits, 0x20); /* SPACE */ + SETBIT(classbits, 0xa0); /* NSBP */ +#ifndef COMPILE_PCRE8 + xclass = TRUE; + *class_uchardata++ = XCL_SINGLE; + *class_uchardata++ = 0x1680; + *class_uchardata++ = XCL_SINGLE; + *class_uchardata++ = 0x180e; + *class_uchardata++ = XCL_RANGE; + *class_uchardata++ = 0x2000; + *class_uchardata++ = 0x200a; + *class_uchardata++ = XCL_SINGLE; + *class_uchardata++ = 0x202f; + *class_uchardata++ = XCL_SINGLE; + *class_uchardata++ = 0x205f; + *class_uchardata++ = XCL_SINGLE; + *class_uchardata++ = 0x3000; +#elif defined SUPPORT_UTF + if (utf) + { + xclass = TRUE; + *class_uchardata++ = XCL_SINGLE; + class_uchardata += PRIV(ord2utf)(0x1680, class_uchardata); + *class_uchardata++ = XCL_SINGLE; + class_uchardata += PRIV(ord2utf)(0x180e, class_uchardata); + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x2000, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x200a, class_uchardata); + *class_uchardata++ = XCL_SINGLE; + class_uchardata += PRIV(ord2utf)(0x202f, class_uchardata); + *class_uchardata++ = XCL_SINGLE; + class_uchardata += PRIV(ord2utf)(0x205f, class_uchardata); + *class_uchardata++ = XCL_SINGLE; + class_uchardata += PRIV(ord2utf)(0x3000, class_uchardata); + } +#endif + continue; + + case ESC_H: + for (c = 0; c < 32; c++) + { + int x = 0xff; + switch (c) + { + case 0x09/8: x ^= 1 << (0x09%8); break; + case 0x20/8: x ^= 1 << (0x20%8); break; + case 0xa0/8: x ^= 1 << (0xa0%8); break; + default: break; + } + classbits[c] |= x; + } +#ifndef COMPILE_PCRE8 + xclass = TRUE; + *class_uchardata++ = XCL_RANGE; + *class_uchardata++ = 0x0100; + *class_uchardata++ = 0x167f; + *class_uchardata++ = XCL_RANGE; + *class_uchardata++ = 0x1681; + *class_uchardata++ = 0x180d; + *class_uchardata++ = XCL_RANGE; + *class_uchardata++ = 0x180f; + *class_uchardata++ = 0x1fff; + *class_uchardata++ = XCL_RANGE; + *class_uchardata++ = 0x200b; + *class_uchardata++ = 0x202e; + *class_uchardata++ = XCL_RANGE; + *class_uchardata++ = 0x2030; + *class_uchardata++ = 0x205e; + *class_uchardata++ = XCL_RANGE; + *class_uchardata++ = 0x2060; + *class_uchardata++ = 0x2fff; + *class_uchardata++ = XCL_RANGE; + *class_uchardata++ = 0x3001; +#ifdef SUPPORT_UTF + if (utf) + class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata); + else +#endif + *class_uchardata++ = 0xffff; +#elif defined SUPPORT_UTF + if (utf) + { + xclass = TRUE; + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x0100, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x167f, class_uchardata); + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x1681, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x180d, class_uchardata); + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x180f, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x1fff, class_uchardata); + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x200b, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x202e, class_uchardata); + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x2030, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x205e, class_uchardata); + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x2060, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x2fff, class_uchardata); + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x3001, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata); + } +#endif + continue; + + case ESC_v: + SETBIT(classbits, 0x0a); /* LF */ + SETBIT(classbits, 0x0b); /* VT */ + SETBIT(classbits, 0x0c); /* FF */ + SETBIT(classbits, 0x0d); /* CR */ + SETBIT(classbits, 0x85); /* NEL */ +#ifndef COMPILE_PCRE8 + xclass = TRUE; + *class_uchardata++ = XCL_RANGE; + *class_uchardata++ = 0x2028; + *class_uchardata++ = 0x2029; +#elif defined SUPPORT_UTF + if (utf) + { + xclass = TRUE; + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x2028, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x2029, class_uchardata); + } +#endif + continue; + + case ESC_V: + for (c = 0; c < 32; c++) + { + int x = 0xff; + switch (c) + { + case 0x0a/8: x ^= 1 << (0x0a%8); + x ^= 1 << (0x0b%8); + x ^= 1 << (0x0c%8); + x ^= 1 << (0x0d%8); + break; + case 0x85/8: x ^= 1 << (0x85%8); break; + default: break; + } + classbits[c] |= x; + } + +#ifndef COMPILE_PCRE8 + xclass = TRUE; + *class_uchardata++ = XCL_RANGE; + *class_uchardata++ = 0x0100; + *class_uchardata++ = 0x2027; + *class_uchardata++ = XCL_RANGE; + *class_uchardata++ = 0x202a; +#ifdef SUPPORT_UTF + if (utf) + class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata); + else +#endif + *class_uchardata++ = 0xffff; +#elif defined SUPPORT_UTF + if (utf) + { + xclass = TRUE; + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x0100, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x2027, class_uchardata); + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x202a, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata); + } +#endif + continue; + +#ifdef SUPPORT_UCP + case ESC_p: + case ESC_P: + { + BOOL negated; + int pdata; + int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr); + if (ptype < 0) goto FAILED; + xclass = TRUE; + *class_uchardata++ = ((-c == ESC_p) != negated)? + XCL_PROP : XCL_NOTPROP; + *class_uchardata++ = ptype; + *class_uchardata++ = pdata; + class_has_8bitchar--; /* Undo! */ + continue; + } +#endif + /* Unrecognized escapes are faulted if PCRE is running in its + strict mode. By default, for compatibility with Perl, they are + treated as literals. */ + + default: + if ((options & PCRE_EXTRA) != 0) + { + *errorcodeptr = ERR7; + goto FAILED; + } + class_has_8bitchar--; /* Undo the speculative increase. */ + class_single_char -= 2; /* Undo the speculative increase. */ + c = *ptr; /* Get the final character and fall through */ + break; + } + } + + /* Fall through if we have a single character (c >= 0). This may be + greater than 256. */ + + } /* End of backslash handling */ + + /* A single character may be followed by '-' to form a range. However, + Perl does not permit ']' to be the end of the range. A '-' character + at the end is treated as a literal. Perl ignores orphaned \E sequences + entirely. The code for handling \Q and \E is messy. */ + + CHECK_RANGE: + while (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E) + { + inescq = FALSE; + ptr += 2; + } + + oldptr = ptr; + + /* Remember \r or \n */ + + if (c == CHAR_CR || c == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF; + + /* Check for range */ + + if (!inescq && ptr[1] == CHAR_MINUS) + { + int d; + ptr += 2; + while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) ptr += 2; + + /* If we hit \Q (not followed by \E) at this point, go into escaped + mode. */ + + while (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_Q) + { + ptr += 2; + if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_E) + { ptr += 2; continue; } + inescq = TRUE; + break; + } + + if (*ptr == 0 || (!inescq && *ptr == CHAR_RIGHT_SQUARE_BRACKET)) + { + ptr = oldptr; + goto LONE_SINGLE_CHARACTER; + } + +#ifdef SUPPORT_UTF + if (utf) + { /* Braces are required because the */ + GETCHARLEN(d, ptr, ptr); /* macro generates multiple statements */ + } + else +#endif + d = *ptr; /* Not UTF-8 mode */ + + /* The second part of a range can be a single-character escape, but + not any of the other escapes. Perl 5.6 treats a hyphen as a literal + in such circumstances. */ + + if (!inescq && d == CHAR_BACKSLASH) + { + d = check_escape(&ptr, errorcodeptr, cd->bracount, options, TRUE); + if (*errorcodeptr != 0) goto FAILED; + + /* \b is backspace; any other special means the '-' was literal */ + + if (d < 0) + { + if (d == -ESC_b) d = CHAR_BS; else + { + ptr = oldptr; + goto LONE_SINGLE_CHARACTER; /* A few lines below */ + } + } + } + + /* Check that the two values are in the correct order. Optimize + one-character ranges */ + + if (d < c) + { + *errorcodeptr = ERR8; + goto FAILED; + } + + if (d == c) goto LONE_SINGLE_CHARACTER; /* A few lines below */ + + /* Remember \r or \n */ + + if (d == CHAR_CR || d == CHAR_NL) cd->external_flags |= PCRE_HASCRORLF; + + /* Since we found a character range, single character optimizations + cannot be done anymore. */ + class_single_char = 2; + + /* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless + matching, we have to use an XCLASS with extra data items. Caseless + matching for characters > 127 is available only if UCP support is + available. */ + +#if defined SUPPORT_UTF && !(defined COMPILE_PCRE8) + if ((d > 255) || (utf && ((options & PCRE_CASELESS) != 0 && d > 127))) +#elif defined SUPPORT_UTF + if (utf && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127))) +#elif !(defined COMPILE_PCRE8) + if (d > 255) +#endif +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + { + xclass = TRUE; + + /* With UCP support, we can find the other case equivalents of + the relevant characters. There may be several ranges. Optimize how + they fit with the basic range. */ + +#ifdef SUPPORT_UCP +#ifndef COMPILE_PCRE8 + if (utf && (options & PCRE_CASELESS) != 0) +#else + if ((options & PCRE_CASELESS) != 0) +#endif + { + unsigned int occ, ocd; + unsigned int cc = c; + unsigned int origd = d; + while (get_othercase_range(&cc, origd, &occ, &ocd)) + { + if (occ >= (unsigned int)c && + ocd <= (unsigned int)d) + continue; /* Skip embedded ranges */ + + if (occ < (unsigned int)c && + ocd >= (unsigned int)c - 1) /* Extend the basic range */ + { /* if there is overlap, */ + c = occ; /* noting that if occ < c */ + continue; /* we can't have ocd > d */ + } /* because a subrange is */ + if (ocd > (unsigned int)d && + occ <= (unsigned int)d + 1) /* always shorter than */ + { /* the basic range. */ + d = ocd; + continue; + } + + if (occ == ocd) + { + *class_uchardata++ = XCL_SINGLE; + } + else + { + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(occ, class_uchardata); + } + class_uchardata += PRIV(ord2utf)(ocd, class_uchardata); + } + } +#endif /* SUPPORT_UCP */ + + /* Now record the original range, possibly modified for UCP caseless + overlapping ranges. */ + + *class_uchardata++ = XCL_RANGE; +#ifdef SUPPORT_UTF +#ifndef COMPILE_PCRE8 + if (utf) + { + class_uchardata += PRIV(ord2utf)(c, class_uchardata); + class_uchardata += PRIV(ord2utf)(d, class_uchardata); + } + else + { + *class_uchardata++ = c; + *class_uchardata++ = d; + } +#else + class_uchardata += PRIV(ord2utf)(c, class_uchardata); + class_uchardata += PRIV(ord2utf)(d, class_uchardata); +#endif +#else /* SUPPORT_UTF */ + *class_uchardata++ = c; + *class_uchardata++ = d; +#endif /* SUPPORT_UTF */ + + /* With UCP support, we are done. Without UCP support, there is no + caseless matching for UTF characters > 127; we can use the bit map + for the smaller ones. As for 16 bit characters without UTF, we + can still use */ + +#ifdef SUPPORT_UCP +#ifndef COMPILE_PCRE8 + if (utf) +#endif + continue; /* With next character in the class */ +#endif /* SUPPORT_UCP */ + +#if defined SUPPORT_UTF && !defined(SUPPORT_UCP) && !(defined COMPILE_PCRE8) + if (utf) + { + if ((options & PCRE_CASELESS) == 0 || c > 127) continue; + /* Adjust upper limit and fall through to set up the map */ + d = 127; + } + else + { + if (c > 255) continue; + /* Adjust upper limit and fall through to set up the map */ + d = 255; + } +#elif defined SUPPORT_UTF && !defined(SUPPORT_UCP) + if ((options & PCRE_CASELESS) == 0 || c > 127) continue; + /* Adjust upper limit and fall through to set up the map */ + d = 127; +#else + if (c > 255) continue; + /* Adjust upper limit and fall through to set up the map */ + d = 255; +#endif /* SUPPORT_UTF && !SUPPORT_UCP && !COMPILE_PCRE8 */ + } +#endif /* SUPPORT_UTF || !COMPILE_PCRE8 */ + + /* We use the bit map for 8 bit mode, or when the characters fall + partially or entirely to [0-255] ([0-127] for UCP) ranges. */ + + class_has_8bitchar = 1; + + /* We can save a bit of time by skipping this in the pre-compile. */ + + if (lengthptr == NULL) for (; c <= d; c++) + { + classbits[c/8] |= (1 << (c&7)); + if ((options & PCRE_CASELESS) != 0) + { + int uc = cd->fcc[c]; /* flip case */ + classbits[uc/8] |= (1 << (uc&7)); + } + } + + continue; /* Go get the next char in the class */ + } + + /* Handle a lone single character - we can get here for a normal + non-escape char, or after \ that introduces a single character or for an + apparent range that isn't. */ + + LONE_SINGLE_CHARACTER: + + /* Only the value of 1 matters for class_single_char. */ + + if (class_single_char < 2) class_single_char++; + + /* If class_charcount is 1, we saw precisely one character. As long as + there was no use of \p or \P, in other words, no use of any XCLASS + features, we can optimize. + + The optimization throws away the bit map. We turn the item into a + 1-character OP_CHAR[I] if it's positive, or OP_NOT[I] if it's negative. + In the positive case, it can cause firstchar to be set. Otherwise, there + can be no first char if this item is first, whatever repeat count may + follow. In the case of reqchar, save the previous value for reinstating. */ + + if (class_single_char == 1 && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) + { + ptr++; + zeroreqchar = reqchar; + + if (negate_class) + { + if (firstchar == REQ_UNSET) firstchar = REQ_NONE; + zerofirstchar = firstchar; + *code++ = ((options & PCRE_CASELESS) != 0)? OP_NOTI: OP_NOT; +#ifdef SUPPORT_UTF + if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR) + code += PRIV(ord2utf)(c, code); + else +#endif + *code++ = c; + goto NOT_CHAR; + } + + /* For a single, positive character, get the value into mcbuffer, and + then we can handle this with the normal one-character code. */ + +#ifdef SUPPORT_UTF + if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR) + mclength = PRIV(ord2utf)(c, mcbuffer); + else +#endif + { + mcbuffer[0] = c; + mclength = 1; + } + goto ONE_CHAR; + } /* End of 1-char optimization */ + + /* Handle a character that cannot go in the bit map. */ + +#if defined SUPPORT_UTF && !(defined COMPILE_PCRE8) + if ((c > 255) || (utf && ((options & PCRE_CASELESS) != 0 && c > 127))) +#elif defined SUPPORT_UTF + if (utf && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127))) +#elif !(defined COMPILE_PCRE8) + if (c > 255) +#endif + +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + { + xclass = TRUE; + *class_uchardata++ = XCL_SINGLE; +#ifdef SUPPORT_UTF +#ifndef COMPILE_PCRE8 + /* In non 8 bit mode, we can get here even if we are not in UTF mode. */ + if (!utf) + *class_uchardata++ = c; + else +#endif + class_uchardata += PRIV(ord2utf)(c, class_uchardata); +#else /* SUPPORT_UTF */ + *class_uchardata++ = c; +#endif /* SUPPORT_UTF */ + +#ifdef SUPPORT_UCP +#ifdef COMPILE_PCRE8 + if ((options & PCRE_CASELESS) != 0) +#else + /* In non 8 bit mode, we can get here even if we are not in UTF mode. */ + if (utf && (options & PCRE_CASELESS) != 0) +#endif + { + unsigned int othercase; + if ((int)(othercase = UCD_OTHERCASE(c)) != c) + { + *class_uchardata++ = XCL_SINGLE; + class_uchardata += PRIV(ord2utf)(othercase, class_uchardata); + } + } +#endif /* SUPPORT_UCP */ + + } + else +#endif /* SUPPORT_UTF || COMPILE_PCRE16 */ + + /* Handle a single-byte character */ + { + class_has_8bitchar = 1; + classbits[c/8] |= (1 << (c&7)); + if ((options & PCRE_CASELESS) != 0) + { + c = cd->fcc[c]; /* flip case */ + classbits[c/8] |= (1 << (c&7)); + } + } + } + + /* Loop until ']' reached. This "while" is the end of the "do" far above. + If we are at the end of an internal nested string, revert to the outer + string. */ + + while (((c = *(++ptr)) != 0 || + (nestptr != NULL && + (ptr = nestptr, nestptr = NULL, c = *(++ptr)) != 0)) && + (c != CHAR_RIGHT_SQUARE_BRACKET || inescq)); + + /* Check for missing terminating ']' */ + + if (c == 0) + { + *errorcodeptr = ERR6; + goto FAILED; + } + + /* If this is the first thing in the branch, there can be no first char + setting, whatever the repeat count. Any reqchar setting must remain + unchanged after any kind of repeat. */ + + if (firstchar == REQ_UNSET) firstchar = REQ_NONE; + zerofirstchar = firstchar; + zeroreqchar = reqchar; + + /* If there are characters with values > 255, we have to compile an + extended class, with its own opcode, unless there was a negated special + such as \S in the class, and PCRE_UCP is not set, because in that case all + characters > 255 are in the class, so any that were explicitly given as + well can be ignored. If (when there are explicit characters > 255 that must + be listed) there are no characters < 256, we can omit the bitmap in the + actual compiled code. */ + +#ifdef SUPPORT_UTF + if (xclass && (!should_flip_negation || (options & PCRE_UCP) != 0)) +#elif !defined COMPILE_PCRE8 + if (xclass && !should_flip_negation) +#endif +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + { + *class_uchardata++ = XCL_END; /* Marks the end of extra data */ + *code++ = OP_XCLASS; + code += LINK_SIZE; + *code = negate_class? XCL_NOT:0; + + /* If the map is required, move up the extra data to make room for it; + otherwise just move the code pointer to the end of the extra data. */ + + if (class_has_8bitchar > 0) + { + *code++ |= XCL_MAP; + memmove(code + (32 / sizeof(pcre_uchar)), code, + IN_UCHARS(class_uchardata - code)); + memcpy(code, classbits, 32); + code = class_uchardata + (32 / sizeof(pcre_uchar)); + } + else code = class_uchardata; + + /* Now fill in the complete length of the item */ + + PUT(previous, 1, (int)(code - previous)); + break; /* End of class handling */ + } +#endif + + /* If there are no characters > 255, or they are all to be included or + excluded, set the opcode to OP_CLASS or OP_NCLASS, depending on whether the + whole class was negated and whether there were negative specials such as \S + (non-UCP) in the class. Then copy the 32-byte map into the code vector, + negating it if necessary. */ + + *code++ = (negate_class == should_flip_negation) ? OP_CLASS : OP_NCLASS; + if (lengthptr == NULL) /* Save time in the pre-compile phase */ + { + if (negate_class) + for (c = 0; c < 32; c++) classbits[c] = ~classbits[c]; + memcpy(code, classbits, 32); + } + code += 32 / sizeof(pcre_uchar); + NOT_CHAR: + break; + + + /* ===================================================================*/ + /* Various kinds of repeat; '{' is not necessarily a quantifier, but this + has been tested above. */ + + case CHAR_LEFT_CURLY_BRACKET: + if (!is_quantifier) goto NORMAL_CHAR; + ptr = read_repeat_counts(ptr+1, &repeat_min, &repeat_max, errorcodeptr); + if (*errorcodeptr != 0) goto FAILED; + goto REPEAT; + + case CHAR_ASTERISK: + repeat_min = 0; + repeat_max = -1; + goto REPEAT; + + case CHAR_PLUS: + repeat_min = 1; + repeat_max = -1; + goto REPEAT; + + case CHAR_QUESTION_MARK: + repeat_min = 0; + repeat_max = 1; + + REPEAT: + if (previous == NULL) + { + *errorcodeptr = ERR9; + goto FAILED; + } + + if (repeat_min == 0) + { + firstchar = zerofirstchar; /* Adjust for zero repeat */ + reqchar = zeroreqchar; /* Ditto */ + } + + /* Remember whether this is a variable length repeat */ + + reqvary = (repeat_min == repeat_max)? 0 : REQ_VARY; + + op_type = 0; /* Default single-char op codes */ + possessive_quantifier = FALSE; /* Default not possessive quantifier */ + + /* Save start of previous item, in case we have to move it up in order to + insert something before it. */ + + tempcode = previous; + + /* If the next character is '+', we have a possessive quantifier. This + implies greediness, whatever the setting of the PCRE_UNGREEDY option. + If the next character is '?' this is a minimizing repeat, by default, + but if PCRE_UNGREEDY is set, it works the other way round. We change the + repeat type to the non-default. */ + + if (ptr[1] == CHAR_PLUS) + { + repeat_type = 0; /* Force greedy */ + possessive_quantifier = TRUE; + ptr++; + } + else if (ptr[1] == CHAR_QUESTION_MARK) + { + repeat_type = greedy_non_default; + ptr++; + } + else repeat_type = greedy_default; + + /* If previous was a recursion call, wrap it in atomic brackets so that + previous becomes the atomic group. All recursions were so wrapped in the + past, but it no longer happens for non-repeated recursions. In fact, the + repeated ones could be re-implemented independently so as not to need this, + but for the moment we rely on the code for repeating groups. */ + + if (*previous == OP_RECURSE) + { + memmove(previous + 1 + LINK_SIZE, previous, IN_UCHARS(1 + LINK_SIZE)); + *previous = OP_ONCE; + PUT(previous, 1, 2 + 2*LINK_SIZE); + previous[2 + 2*LINK_SIZE] = OP_KET; + PUT(previous, 3 + 2*LINK_SIZE, 2 + 2*LINK_SIZE); + code += 2 + 2 * LINK_SIZE; + length_prevgroup = 3 + 3*LINK_SIZE; + + /* When actually compiling, we need to check whether this was a forward + reference, and if so, adjust the offset. */ + + if (lengthptr == NULL && cd->hwm >= cd->start_workspace + LINK_SIZE) + { + int offset = GET(cd->hwm, -LINK_SIZE); + if (offset == previous + 1 - cd->start_code) + PUT(cd->hwm, -LINK_SIZE, offset + 1 + LINK_SIZE); + } + } + + /* Now handle repetition for the different types of item. */ + + /* If previous was a character or negated character match, abolish the item + and generate a repeat item instead. If a char item has a minimum of more + than one, ensure that it is set in reqchar - it might not be if a sequence + such as x{3} is the first thing in a branch because the x will have gone + into firstchar instead. */ + + if (*previous == OP_CHAR || *previous == OP_CHARI + || *previous == OP_NOT || *previous == OP_NOTI) + { + switch (*previous) + { + default: /* Make compiler happy. */ + case OP_CHAR: op_type = OP_STAR - OP_STAR; break; + case OP_CHARI: op_type = OP_STARI - OP_STAR; break; + case OP_NOT: op_type = OP_NOTSTAR - OP_STAR; break; + case OP_NOTI: op_type = OP_NOTSTARI - OP_STAR; break; + } + + /* Deal with UTF characters that take up more than one character. It's + easier to write this out separately than try to macrify it. Use c to + hold the length of the character in bytes, plus UTF_LENGTH to flag that + it's a length rather than a small character. */ + +#ifdef SUPPORT_UTF + if (utf && NOT_FIRSTCHAR(code[-1])) + { + pcre_uchar *lastchar = code - 1; + BACKCHAR(lastchar); + c = (int)(code - lastchar); /* Length of UTF-8 character */ + memcpy(utf_chars, lastchar, IN_UCHARS(c)); /* Save the char */ + c |= UTF_LENGTH; /* Flag c as a length */ + } + else +#endif /* SUPPORT_UTF */ + + /* Handle the case of a single charater - either with no UTF support, or + with UTF disabled, or for a single character UTF character. */ + { + c = code[-1]; + if (*previous <= OP_CHARI && repeat_min > 1) + reqchar = c | req_caseopt | cd->req_varyopt; + } + + /* If the repetition is unlimited, it pays to see if the next thing on + the line is something that cannot possibly match this character. If so, + automatically possessifying this item gains some performance in the case + where the match fails. */ + + if (!possessive_quantifier && + repeat_max < 0 && + check_auto_possessive(previous, utf, ptr + 1, options, cd)) + { + repeat_type = 0; /* Force greedy */ + possessive_quantifier = TRUE; + } + + goto OUTPUT_SINGLE_REPEAT; /* Code shared with single character types */ + } + + /* If previous was a character type match (\d or similar), abolish it and + create a suitable repeat item. The code is shared with single-character + repeats by setting op_type to add a suitable offset into repeat_type. Note + the the Unicode property types will be present only when SUPPORT_UCP is + defined, but we don't wrap the little bits of code here because it just + makes it horribly messy. */ + + else if (*previous < OP_EODN) + { + pcre_uchar *oldcode; + int prop_type, prop_value; + op_type = OP_TYPESTAR - OP_STAR; /* Use type opcodes */ + c = *previous; + + if (!possessive_quantifier && + repeat_max < 0 && + check_auto_possessive(previous, utf, ptr + 1, options, cd)) + { + repeat_type = 0; /* Force greedy */ + possessive_quantifier = TRUE; + } + + OUTPUT_SINGLE_REPEAT: + if (*previous == OP_PROP || *previous == OP_NOTPROP) + { + prop_type = previous[1]; + prop_value = previous[2]; + } + else prop_type = prop_value = -1; + + oldcode = code; + code = previous; /* Usually overwrite previous item */ + + /* If the maximum is zero then the minimum must also be zero; Perl allows + this case, so we do too - by simply omitting the item altogether. */ + + if (repeat_max == 0) goto END_REPEAT; + + /*--------------------------------------------------------------------*/ + /* This code is obsolete from release 8.00; the restriction was finally + removed: */ + + /* All real repeats make it impossible to handle partial matching (maybe + one day we will be able to remove this restriction). */ + + /* if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL; */ + /*--------------------------------------------------------------------*/ + + /* Combine the op_type with the repeat_type */ + + repeat_type += op_type; + + /* A minimum of zero is handled either as the special case * or ?, or as + an UPTO, with the maximum given. */ + + if (repeat_min == 0) + { + if (repeat_max == -1) *code++ = OP_STAR + repeat_type; + else if (repeat_max == 1) *code++ = OP_QUERY + repeat_type; + else + { + *code++ = OP_UPTO + repeat_type; + PUT2INC(code, 0, repeat_max); + } + } + + /* A repeat minimum of 1 is optimized into some special cases. If the + maximum is unlimited, we use OP_PLUS. Otherwise, the original item is + left in place and, if the maximum is greater than 1, we use OP_UPTO with + one less than the maximum. */ + + else if (repeat_min == 1) + { + if (repeat_max == -1) + *code++ = OP_PLUS + repeat_type; + else + { + code = oldcode; /* leave previous item in place */ + if (repeat_max == 1) goto END_REPEAT; + *code++ = OP_UPTO + repeat_type; + PUT2INC(code, 0, repeat_max - 1); + } + } + + /* The case {n,n} is just an EXACT, while the general case {n,m} is + handled as an EXACT followed by an UPTO. */ + + else + { + *code++ = OP_EXACT + op_type; /* NB EXACT doesn't have repeat_type */ + PUT2INC(code, 0, repeat_min); + + /* If the maximum is unlimited, insert an OP_STAR. Before doing so, + we have to insert the character for the previous code. For a repeated + Unicode property match, there are two extra bytes that define the + required property. In UTF-8 mode, long characters have their length in + c, with the UTF_LENGTH bit as a flag. */ + + if (repeat_max < 0) + { +#ifdef SUPPORT_UTF + if (utf && (c & UTF_LENGTH) != 0) + { + memcpy(code, utf_chars, IN_UCHARS(c & 7)); + code += c & 7; + } + else +#endif + { + *code++ = c; + if (prop_type >= 0) + { + *code++ = prop_type; + *code++ = prop_value; + } + } + *code++ = OP_STAR + repeat_type; + } + + /* Else insert an UPTO if the max is greater than the min, again + preceded by the character, for the previously inserted code. If the + UPTO is just for 1 instance, we can use QUERY instead. */ + + else if (repeat_max != repeat_min) + { +#ifdef SUPPORT_UTF + if (utf && (c & UTF_LENGTH) != 0) + { + memcpy(code, utf_chars, IN_UCHARS(c & 7)); + code += c & 7; + } + else +#endif + *code++ = c; + if (prop_type >= 0) + { + *code++ = prop_type; + *code++ = prop_value; + } + repeat_max -= repeat_min; + + if (repeat_max == 1) + { + *code++ = OP_QUERY + repeat_type; + } + else + { + *code++ = OP_UPTO + repeat_type; + PUT2INC(code, 0, repeat_max); + } + } + } + + /* The character or character type itself comes last in all cases. */ + +#ifdef SUPPORT_UTF + if (utf && (c & UTF_LENGTH) != 0) + { + memcpy(code, utf_chars, IN_UCHARS(c & 7)); + code += c & 7; + } + else +#endif + *code++ = c; + + /* For a repeated Unicode property match, there are two extra bytes that + define the required property. */ + +#ifdef SUPPORT_UCP + if (prop_type >= 0) + { + *code++ = prop_type; + *code++ = prop_value; + } +#endif + } + + /* If previous was a character class or a back reference, we put the repeat + stuff after it, but just skip the item if the repeat was {0,0}. */ + + else if (*previous == OP_CLASS || + *previous == OP_NCLASS || +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + *previous == OP_XCLASS || +#endif + *previous == OP_REF || + *previous == OP_REFI) + { + if (repeat_max == 0) + { + code = previous; + goto END_REPEAT; + } + + /*--------------------------------------------------------------------*/ + /* This code is obsolete from release 8.00; the restriction was finally + removed: */ + + /* All real repeats make it impossible to handle partial matching (maybe + one day we will be able to remove this restriction). */ + + /* if (repeat_max != 1) cd->external_flags |= PCRE_NOPARTIAL; */ + /*--------------------------------------------------------------------*/ + + if (repeat_min == 0 && repeat_max == -1) + *code++ = OP_CRSTAR + repeat_type; + else if (repeat_min == 1 && repeat_max == -1) + *code++ = OP_CRPLUS + repeat_type; + else if (repeat_min == 0 && repeat_max == 1) + *code++ = OP_CRQUERY + repeat_type; + else + { + *code++ = OP_CRRANGE + repeat_type; + PUT2INC(code, 0, repeat_min); + if (repeat_max == -1) repeat_max = 0; /* 2-byte encoding for max */ + PUT2INC(code, 0, repeat_max); + } + } + + /* If previous was a bracket group, we may have to replicate it in certain + cases. Note that at this point we can encounter only the "basic" bracket + opcodes such as BRA and CBRA, as this is the place where they get converted + into the more special varieties such as BRAPOS and SBRA. A test for >= + OP_ASSERT and <= OP_COND includes ASSERT, ASSERT_NOT, ASSERTBACK, + ASSERTBACK_NOT, ONCE, BRA, CBRA, and COND. Originally, PCRE did not allow + repetition of assertions, but now it does, for Perl compatibility. */ + + else if (*previous >= OP_ASSERT && *previous <= OP_COND) + { + register int i; + int len = (int)(code - previous); + pcre_uchar *bralink = NULL; + pcre_uchar *brazeroptr = NULL; + + /* Repeating a DEFINE group is pointless, but Perl allows the syntax, so + we just ignore the repeat. */ + + if (*previous == OP_COND && previous[LINK_SIZE+1] == OP_DEF) + goto END_REPEAT; + + /* There is no sense in actually repeating assertions. The only potential + use of repetition is in cases when the assertion is optional. Therefore, + if the minimum is greater than zero, just ignore the repeat. If the + maximum is not not zero or one, set it to 1. */ + + if (*previous < OP_ONCE) /* Assertion */ + { + if (repeat_min > 0) goto END_REPEAT; + if (repeat_max < 0 || repeat_max > 1) repeat_max = 1; + } + + /* The case of a zero minimum is special because of the need to stick + OP_BRAZERO in front of it, and because the group appears once in the + data, whereas in other cases it appears the minimum number of times. For + this reason, it is simplest to treat this case separately, as otherwise + the code gets far too messy. There are several special subcases when the + minimum is zero. */ + + if (repeat_min == 0) + { + /* If the maximum is also zero, we used to just omit the group from the + output altogether, like this: + + ** if (repeat_max == 0) + ** { + ** code = previous; + ** goto END_REPEAT; + ** } + + However, that fails when a group or a subgroup within it is referenced + as a subroutine from elsewhere in the pattern, so now we stick in + OP_SKIPZERO in front of it so that it is skipped on execution. As we + don't have a list of which groups are referenced, we cannot do this + selectively. + + If the maximum is 1 or unlimited, we just have to stick in the BRAZERO + and do no more at this point. However, we do need to adjust any + OP_RECURSE calls inside the group that refer to the group itself or any + internal or forward referenced group, because the offset is from the + start of the whole regex. Temporarily terminate the pattern while doing + this. */ + + if (repeat_max <= 1) /* Covers 0, 1, and unlimited */ + { + *code = OP_END; + adjust_recurse(previous, 1, utf, cd, save_hwm); + memmove(previous + 1, previous, IN_UCHARS(len)); + code++; + if (repeat_max == 0) + { + *previous++ = OP_SKIPZERO; + goto END_REPEAT; + } + brazeroptr = previous; /* Save for possessive optimizing */ + *previous++ = OP_BRAZERO + repeat_type; + } + + /* If the maximum is greater than 1 and limited, we have to replicate + in a nested fashion, sticking OP_BRAZERO before each set of brackets. + The first one has to be handled carefully because it's the original + copy, which has to be moved up. The remainder can be handled by code + that is common with the non-zero minimum case below. We have to + adjust the value or repeat_max, since one less copy is required. Once + again, we may have to adjust any OP_RECURSE calls inside the group. */ + + else + { + int offset; + *code = OP_END; + adjust_recurse(previous, 2 + LINK_SIZE, utf, cd, save_hwm); + memmove(previous + 2 + LINK_SIZE, previous, IN_UCHARS(len)); + code += 2 + LINK_SIZE; + *previous++ = OP_BRAZERO + repeat_type; + *previous++ = OP_BRA; + + /* We chain together the bracket offset fields that have to be + filled in later when the ends of the brackets are reached. */ + + offset = (bralink == NULL)? 0 : (int)(previous - bralink); + bralink = previous; + PUTINC(previous, 0, offset); + } + + repeat_max--; + } + + /* If the minimum is greater than zero, replicate the group as many + times as necessary, and adjust the maximum to the number of subsequent + copies that we need. If we set a first char from the group, and didn't + set a required char, copy the latter from the former. If there are any + forward reference subroutine calls in the group, there will be entries on + the workspace list; replicate these with an appropriate increment. */ + + else + { + if (repeat_min > 1) + { + /* In the pre-compile phase, we don't actually do the replication. We + just adjust the length as if we had. Do some paranoid checks for + potential integer overflow. The INT64_OR_DOUBLE type is a 64-bit + integer type when available, otherwise double. */ + + if (lengthptr != NULL) + { + int delta = (repeat_min - 1)*length_prevgroup; + if ((INT64_OR_DOUBLE)(repeat_min - 1)* + (INT64_OR_DOUBLE)length_prevgroup > + (INT64_OR_DOUBLE)INT_MAX || + OFLOW_MAX - *lengthptr < delta) + { + *errorcodeptr = ERR20; + goto FAILED; + } + *lengthptr += delta; + } + + /* This is compiling for real. If there is a set first byte for + the group, and we have not yet set a "required byte", set it. Make + sure there is enough workspace for copying forward references before + doing the copy. */ + + else + { + if (groupsetfirstchar && reqchar < 0) reqchar = firstchar; + + for (i = 1; i < repeat_min; i++) + { + pcre_uchar *hc; + pcre_uchar *this_hwm = cd->hwm; + memcpy(code, previous, IN_UCHARS(len)); + + while (cd->hwm > cd->start_workspace + cd->workspace_size - + WORK_SIZE_SAFETY_MARGIN - (this_hwm - save_hwm)) + { + int save_offset = save_hwm - cd->start_workspace; + int this_offset = this_hwm - cd->start_workspace; + *errorcodeptr = expand_workspace(cd); + if (*errorcodeptr != 0) goto FAILED; + save_hwm = (pcre_uchar *)cd->start_workspace + save_offset; + this_hwm = (pcre_uchar *)cd->start_workspace + this_offset; + } + + for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE) + { + PUT(cd->hwm, 0, GET(hc, 0) + len); + cd->hwm += LINK_SIZE; + } + save_hwm = this_hwm; + code += len; + } + } + } + + if (repeat_max > 0) repeat_max -= repeat_min; + } + + /* This code is common to both the zero and non-zero minimum cases. If + the maximum is limited, it replicates the group in a nested fashion, + remembering the bracket starts on a stack. In the case of a zero minimum, + the first one was set up above. In all cases the repeat_max now specifies + the number of additional copies needed. Again, we must remember to + replicate entries on the forward reference list. */ + + if (repeat_max >= 0) + { + /* In the pre-compile phase, we don't actually do the replication. We + just adjust the length as if we had. For each repetition we must add 1 + to the length for BRAZERO and for all but the last repetition we must + add 2 + 2*LINKSIZE to allow for the nesting that occurs. Do some + paranoid checks to avoid integer overflow. The INT64_OR_DOUBLE type is + a 64-bit integer type when available, otherwise double. */ + + if (lengthptr != NULL && repeat_max > 0) + { + int delta = repeat_max * (length_prevgroup + 1 + 2 + 2*LINK_SIZE) - + 2 - 2*LINK_SIZE; /* Last one doesn't nest */ + if ((INT64_OR_DOUBLE)repeat_max * + (INT64_OR_DOUBLE)(length_prevgroup + 1 + 2 + 2*LINK_SIZE) + > (INT64_OR_DOUBLE)INT_MAX || + OFLOW_MAX - *lengthptr < delta) + { + *errorcodeptr = ERR20; + goto FAILED; + } + *lengthptr += delta; + } + + /* This is compiling for real */ + + else for (i = repeat_max - 1; i >= 0; i--) + { + pcre_uchar *hc; + pcre_uchar *this_hwm = cd->hwm; + + *code++ = OP_BRAZERO + repeat_type; + + /* All but the final copy start a new nesting, maintaining the + chain of brackets outstanding. */ + + if (i != 0) + { + int offset; + *code++ = OP_BRA; + offset = (bralink == NULL)? 0 : (int)(code - bralink); + bralink = code; + PUTINC(code, 0, offset); + } + + memcpy(code, previous, IN_UCHARS(len)); + + /* Ensure there is enough workspace for forward references before + copying them. */ + + while (cd->hwm > cd->start_workspace + cd->workspace_size - + WORK_SIZE_SAFETY_MARGIN - (this_hwm - save_hwm)) + { + int save_offset = save_hwm - cd->start_workspace; + int this_offset = this_hwm - cd->start_workspace; + *errorcodeptr = expand_workspace(cd); + if (*errorcodeptr != 0) goto FAILED; + save_hwm = (pcre_uchar *)cd->start_workspace + save_offset; + this_hwm = (pcre_uchar *)cd->start_workspace + this_offset; + } + + for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE) + { + PUT(cd->hwm, 0, GET(hc, 0) + len + ((i != 0)? 2+LINK_SIZE : 1)); + cd->hwm += LINK_SIZE; + } + save_hwm = this_hwm; + code += len; + } + + /* Now chain through the pending brackets, and fill in their length + fields (which are holding the chain links pro tem). */ + + while (bralink != NULL) + { + int oldlinkoffset; + int offset = (int)(code - bralink + 1); + pcre_uchar *bra = code - offset; + oldlinkoffset = GET(bra, 1); + bralink = (oldlinkoffset == 0)? NULL : bralink - oldlinkoffset; + *code++ = OP_KET; + PUTINC(code, 0, offset); + PUT(bra, 1, offset); + } + } + + /* If the maximum is unlimited, set a repeater in the final copy. For + ONCE brackets, that's all we need to do. However, possessively repeated + ONCE brackets can be converted into non-capturing brackets, as the + behaviour of (?:xx)++ is the same as (?>xx)++ and this saves having to + deal with possessive ONCEs specially. + + Otherwise, when we are doing the actual compile phase, check to see + whether this group is one that could match an empty string. If so, + convert the initial operator to the S form (e.g. OP_BRA -> OP_SBRA) so + that runtime checking can be done. [This check is also applied to ONCE + groups at runtime, but in a different way.] + + Then, if the quantifier was possessive and the bracket is not a + conditional, we convert the BRA code to the POS form, and the KET code to + KETRPOS. (It turns out to be convenient at runtime to detect this kind of + subpattern at both the start and at the end.) The use of special opcodes + makes it possible to reduce greatly the stack usage in pcre_exec(). If + the group is preceded by OP_BRAZERO, convert this to OP_BRAPOSZERO. + + Then, if the minimum number of matches is 1 or 0, cancel the possessive + flag so that the default action below, of wrapping everything inside + atomic brackets, does not happen. When the minimum is greater than 1, + there will be earlier copies of the group, and so we still have to wrap + the whole thing. */ + + else + { + pcre_uchar *ketcode = code - 1 - LINK_SIZE; + pcre_uchar *bracode = ketcode - GET(ketcode, 1); + + /* Convert possessive ONCE brackets to non-capturing */ + + if ((*bracode == OP_ONCE || *bracode == OP_ONCE_NC) && + possessive_quantifier) *bracode = OP_BRA; + + /* For non-possessive ONCE brackets, all we need to do is to + set the KET. */ + + if (*bracode == OP_ONCE || *bracode == OP_ONCE_NC) + *ketcode = OP_KETRMAX + repeat_type; + + /* Handle non-ONCE brackets and possessive ONCEs (which have been + converted to non-capturing above). */ + + else + { + /* In the compile phase, check for empty string matching. */ + + if (lengthptr == NULL) + { + pcre_uchar *scode = bracode; + do + { + if (could_be_empty_branch(scode, ketcode, utf, cd)) + { + *bracode += OP_SBRA - OP_BRA; + break; + } + scode += GET(scode, 1); + } + while (*scode == OP_ALT); + } + + /* Handle possessive quantifiers. */ + + if (possessive_quantifier) + { + /* For COND brackets, we wrap the whole thing in a possessively + repeated non-capturing bracket, because we have not invented POS + versions of the COND opcodes. Because we are moving code along, we + must ensure that any pending recursive references are updated. */ + + if (*bracode == OP_COND || *bracode == OP_SCOND) + { + int nlen = (int)(code - bracode); + *code = OP_END; + adjust_recurse(bracode, 1 + LINK_SIZE, utf, cd, save_hwm); + memmove(bracode + 1 + LINK_SIZE, bracode, IN_UCHARS(nlen)); + code += 1 + LINK_SIZE; + nlen += 1 + LINK_SIZE; + *bracode = OP_BRAPOS; + *code++ = OP_KETRPOS; + PUTINC(code, 0, nlen); + PUT(bracode, 1, nlen); + } + + /* For non-COND brackets, we modify the BRA code and use KETRPOS. */ + + else + { + *bracode += 1; /* Switch to xxxPOS opcodes */ + *ketcode = OP_KETRPOS; + } + + /* If the minimum is zero, mark it as possessive, then unset the + possessive flag when the minimum is 0 or 1. */ + + if (brazeroptr != NULL) *brazeroptr = OP_BRAPOSZERO; + if (repeat_min < 2) possessive_quantifier = FALSE; + } + + /* Non-possessive quantifier */ + + else *ketcode = OP_KETRMAX + repeat_type; + } + } + } + + /* If previous is OP_FAIL, it was generated by an empty class [] in + JavaScript mode. The other ways in which OP_FAIL can be generated, that is + by (*FAIL) or (?!) set previous to NULL, which gives a "nothing to repeat" + error above. We can just ignore the repeat in JS case. */ + + else if (*previous == OP_FAIL) goto END_REPEAT; + + /* Else there's some kind of shambles */ + + else + { + *errorcodeptr = ERR11; + goto FAILED; + } + + /* If the character following a repeat is '+', or if certain optimization + tests above succeeded, possessive_quantifier is TRUE. For some opcodes, + there are special alternative opcodes for this case. For anything else, we + wrap the entire repeated item inside OP_ONCE brackets. Logically, the '+' + notation is just syntactic sugar, taken from Sun's Java package, but the + special opcodes can optimize it. + + Some (but not all) possessively repeated subpatterns have already been + completely handled in the code just above. For them, possessive_quantifier + is always FALSE at this stage. + + Note that the repeated item starts at tempcode, not at previous, which + might be the first part of a string whose (former) last char we repeated. + + Possessifying an 'exact' quantifier has no effect, so we can ignore it. But + an 'upto' may follow. We skip over an 'exact' item, and then test the + length of what remains before proceeding. */ + + if (possessive_quantifier) + { + int len; + + if (*tempcode == OP_TYPEEXACT) + tempcode += PRIV(OP_lengths)[*tempcode] + + ((tempcode[1 + IMM2_SIZE] == OP_PROP + || tempcode[1 + IMM2_SIZE] == OP_NOTPROP)? 2 : 0); + + else if (*tempcode == OP_EXACT || *tempcode == OP_NOTEXACT) + { + tempcode += PRIV(OP_lengths)[*tempcode]; +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(tempcode[-1])) + tempcode += GET_EXTRALEN(tempcode[-1]); +#endif + } + + len = (int)(code - tempcode); + if (len > 0) switch (*tempcode) + { + case OP_STAR: *tempcode = OP_POSSTAR; break; + case OP_PLUS: *tempcode = OP_POSPLUS; break; + case OP_QUERY: *tempcode = OP_POSQUERY; break; + case OP_UPTO: *tempcode = OP_POSUPTO; break; + + case OP_STARI: *tempcode = OP_POSSTARI; break; + case OP_PLUSI: *tempcode = OP_POSPLUSI; break; + case OP_QUERYI: *tempcode = OP_POSQUERYI; break; + case OP_UPTOI: *tempcode = OP_POSUPTOI; break; + + case OP_NOTSTAR: *tempcode = OP_NOTPOSSTAR; break; + case OP_NOTPLUS: *tempcode = OP_NOTPOSPLUS; break; + case OP_NOTQUERY: *tempcode = OP_NOTPOSQUERY; break; + case OP_NOTUPTO: *tempcode = OP_NOTPOSUPTO; break; + + case OP_NOTSTARI: *tempcode = OP_NOTPOSSTARI; break; + case OP_NOTPLUSI: *tempcode = OP_NOTPOSPLUSI; break; + case OP_NOTQUERYI: *tempcode = OP_NOTPOSQUERYI; break; + case OP_NOTUPTOI: *tempcode = OP_NOTPOSUPTOI; break; + + case OP_TYPESTAR: *tempcode = OP_TYPEPOSSTAR; break; + case OP_TYPEPLUS: *tempcode = OP_TYPEPOSPLUS; break; + case OP_TYPEQUERY: *tempcode = OP_TYPEPOSQUERY; break; + case OP_TYPEUPTO: *tempcode = OP_TYPEPOSUPTO; break; + + /* Because we are moving code along, we must ensure that any + pending recursive references are updated. */ + + default: + *code = OP_END; + adjust_recurse(tempcode, 1 + LINK_SIZE, utf, cd, save_hwm); + memmove(tempcode + 1 + LINK_SIZE, tempcode, IN_UCHARS(len)); + code += 1 + LINK_SIZE; + len += 1 + LINK_SIZE; + tempcode[0] = OP_ONCE; + *code++ = OP_KET; + PUTINC(code, 0, len); + PUT(tempcode, 1, len); + break; + } + } + + /* In all case we no longer have a previous item. We also set the + "follows varying string" flag for subsequently encountered reqchars if + it isn't already set and we have just passed a varying length item. */ + + END_REPEAT: + previous = NULL; + cd->req_varyopt |= reqvary; + break; + + + /* ===================================================================*/ + /* Start of nested parenthesized sub-expression, or comment or lookahead or + lookbehind or option setting or condition or all the other extended + parenthesis forms. */ + + case CHAR_LEFT_PARENTHESIS: + newoptions = options; + skipbytes = 0; + bravalue = OP_CBRA; + save_hwm = cd->hwm; + reset_bracount = FALSE; + + /* First deal with various "verbs" that can be introduced by '*'. */ + + ptr++; + if (ptr[0] == CHAR_ASTERISK && (ptr[1] == ':' + || (MAX_255(ptr[1]) && ((cd->ctypes[ptr[1]] & ctype_letter) != 0)))) + { + int i, namelen; + int arglen = 0; + const char *vn = verbnames; + const pcre_uchar *name = ptr + 1; + const pcre_uchar *arg = NULL; + previous = NULL; + ptr++; + while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_letter) != 0) ptr++; + namelen = (int)(ptr - name); + + /* It appears that Perl allows any characters whatsoever, other than + a closing parenthesis, to appear in arguments, so we no longer insist on + letters, digits, and underscores. */ + + if (*ptr == CHAR_COLON) + { + arg = ++ptr; + while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++; + arglen = (int)(ptr - arg); + if (arglen > (int)MAX_MARK) + { + *errorcodeptr = ERR75; + goto FAILED; + } + } + + if (*ptr != CHAR_RIGHT_PARENTHESIS) + { + *errorcodeptr = ERR60; + goto FAILED; + } + + /* Scan the table of verb names */ + + for (i = 0; i < verbcount; i++) + { + if (namelen == verbs[i].len && + STRNCMP_UC_C8(name, vn, namelen) == 0) + { + /* Check for open captures before ACCEPT and convert it to + ASSERT_ACCEPT if in an assertion. */ + + if (verbs[i].op == OP_ACCEPT) + { + open_capitem *oc; + if (arglen != 0) + { + *errorcodeptr = ERR59; + goto FAILED; + } + cd->had_accept = TRUE; + for (oc = cd->open_caps; oc != NULL; oc = oc->next) + { + *code++ = OP_CLOSE; + PUT2INC(code, 0, oc->number); + } + *code++ = (cd->assert_depth > 0)? OP_ASSERT_ACCEPT : OP_ACCEPT; + + /* Do not set firstchar after *ACCEPT */ + if (firstchar == REQ_UNSET) firstchar = REQ_NONE; + } + + /* Handle other cases with/without an argument */ + + else if (arglen == 0) + { + if (verbs[i].op < 0) /* Argument is mandatory */ + { + *errorcodeptr = ERR66; + goto FAILED; + } + *code = verbs[i].op; + if (*code++ == OP_THEN) cd->external_flags |= PCRE_HASTHEN; + } + + else + { + if (verbs[i].op_arg < 0) /* Argument is forbidden */ + { + *errorcodeptr = ERR59; + goto FAILED; + } + *code = verbs[i].op_arg; + if (*code++ == OP_THEN_ARG) cd->external_flags |= PCRE_HASTHEN; + *code++ = arglen; + memcpy(code, arg, IN_UCHARS(arglen)); + code += arglen; + *code++ = 0; + } + + break; /* Found verb, exit loop */ + } + + vn += verbs[i].len + 1; + } + + if (i < verbcount) continue; /* Successfully handled a verb */ + *errorcodeptr = ERR60; /* Verb not recognized */ + goto FAILED; + } + + /* Deal with the extended parentheses; all are introduced by '?', and the + appearance of any of them means that this is not a capturing group. */ + + else if (*ptr == CHAR_QUESTION_MARK) + { + int i, set, unset, namelen; + int *optset; + const pcre_uchar *name; + pcre_uchar *slot; + + switch (*(++ptr)) + { + case CHAR_NUMBER_SIGN: /* Comment; skip to ket */ + ptr++; + while (*ptr != 0 && *ptr != CHAR_RIGHT_PARENTHESIS) ptr++; + if (*ptr == 0) + { + *errorcodeptr = ERR18; + goto FAILED; + } + continue; + + + /* ------------------------------------------------------------ */ + case CHAR_VERTICAL_LINE: /* Reset capture count for each branch */ + reset_bracount = TRUE; + /* Fall through */ + + /* ------------------------------------------------------------ */ + case CHAR_COLON: /* Non-capturing bracket */ + bravalue = OP_BRA; + ptr++; + break; + + + /* ------------------------------------------------------------ */ + case CHAR_LEFT_PARENTHESIS: + bravalue = OP_COND; /* Conditional group */ + + /* A condition can be an assertion, a number (referring to a numbered + group), a name (referring to a named group), or 'R', referring to + recursion. R and R&name are also permitted for recursion tests. + + There are several syntaxes for testing a named group: (?(name)) is used + by Python; Perl 5.10 onwards uses (?() or (?('name')). + + There are two unfortunate ambiguities, caused by history. (a) 'R' can + be the recursive thing or the name 'R' (and similarly for 'R' followed + by digits), and (b) a number could be a name that consists of digits. + In both cases, we look for a name first; if not found, we try the other + cases. */ + + /* For conditions that are assertions, check the syntax, and then exit + the switch. This will take control down to where bracketed groups, + including assertions, are processed. */ + + if (ptr[1] == CHAR_QUESTION_MARK && (ptr[2] == CHAR_EQUALS_SIGN || + ptr[2] == CHAR_EXCLAMATION_MARK || ptr[2] == CHAR_LESS_THAN_SIGN)) + break; + + /* Most other conditions use OP_CREF (a couple change to OP_RREF + below), and all need to skip 1+IMM2_SIZE bytes at the start of the group. */ + + code[1+LINK_SIZE] = OP_CREF; + skipbytes = 1+IMM2_SIZE; + refsign = -1; + + /* Check for a test for recursion in a named group. */ + + if (ptr[1] == CHAR_R && ptr[2] == CHAR_AMPERSAND) + { + terminator = -1; + ptr += 2; + code[1+LINK_SIZE] = OP_RREF; /* Change the type of test */ + } + + /* Check for a test for a named group's having been set, using the Perl + syntax (?() or (?('name') */ + + else if (ptr[1] == CHAR_LESS_THAN_SIGN) + { + terminator = CHAR_GREATER_THAN_SIGN; + ptr++; + } + else if (ptr[1] == CHAR_APOSTROPHE) + { + terminator = CHAR_APOSTROPHE; + ptr++; + } + else + { + terminator = 0; + if (ptr[1] == CHAR_MINUS || ptr[1] == CHAR_PLUS) refsign = *(++ptr); + } + + /* We now expect to read a name; any thing else is an error */ + + if (!MAX_255(ptr[1]) || (cd->ctypes[ptr[1]] & ctype_word) == 0) + { + ptr += 1; /* To get the right offset */ + *errorcodeptr = ERR28; + goto FAILED; + } + + /* Read the name, but also get it as a number if it's all digits */ + + recno = 0; + name = ++ptr; + while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0) + { + if (recno >= 0) + recno = (IS_DIGIT(*ptr))? recno * 10 + *ptr - CHAR_0 : -1; + ptr++; + } + namelen = (int)(ptr - name); + + if ((terminator > 0 && *ptr++ != terminator) || + *ptr++ != CHAR_RIGHT_PARENTHESIS) + { + ptr--; /* Error offset */ + *errorcodeptr = ERR26; + goto FAILED; + } + + /* Do no further checking in the pre-compile phase. */ + + if (lengthptr != NULL) break; + + /* In the real compile we do the work of looking for the actual + reference. If the string started with "+" or "-" we require the rest to + be digits, in which case recno will be set. */ + + if (refsign > 0) + { + if (recno <= 0) + { + *errorcodeptr = ERR58; + goto FAILED; + } + recno = (refsign == CHAR_MINUS)? + cd->bracount - recno + 1 : recno +cd->bracount; + if (recno <= 0 || recno > cd->final_bracount) + { + *errorcodeptr = ERR15; + goto FAILED; + } + PUT2(code, 2+LINK_SIZE, recno); + break; + } + + /* Otherwise (did not start with "+" or "-"), start by looking for the + name. If we find a name, add one to the opcode to change OP_CREF or + OP_RREF into OP_NCREF or OP_NRREF. These behave exactly the same, + except they record that the reference was originally to a name. The + information is used to check duplicate names. */ + + slot = cd->name_table; + for (i = 0; i < cd->names_found; i++) + { + if (STRNCMP_UC_UC(name, slot+IMM2_SIZE, namelen) == 0) break; + slot += cd->name_entry_size; + } + + /* Found a previous named subpattern */ + + if (i < cd->names_found) + { + recno = GET2(slot, 0); + PUT2(code, 2+LINK_SIZE, recno); + code[1+LINK_SIZE]++; + } + + /* Search the pattern for a forward reference */ + + else if ((i = find_parens(cd, name, namelen, + (options & PCRE_EXTENDED) != 0, utf)) > 0) + { + PUT2(code, 2+LINK_SIZE, i); + code[1+LINK_SIZE]++; + } + + /* If terminator == 0 it means that the name followed directly after + the opening parenthesis [e.g. (?(abc)...] and in this case there are + some further alternatives to try. For the cases where terminator != 0 + [things like (?(... or (?('name')... or (?(R&name)... ] we have + now checked all the possibilities, so give an error. */ + + else if (terminator != 0) + { + *errorcodeptr = ERR15; + goto FAILED; + } + + /* Check for (?(R) for recursion. Allow digits after R to specify a + specific group number. */ + + else if (*name == CHAR_R) + { + recno = 0; + for (i = 1; i < namelen; i++) + { + if (!IS_DIGIT(name[i])) + { + *errorcodeptr = ERR15; + goto FAILED; + } + recno = recno * 10 + name[i] - CHAR_0; + } + if (recno == 0) recno = RREF_ANY; + code[1+LINK_SIZE] = OP_RREF; /* Change test type */ + PUT2(code, 2+LINK_SIZE, recno); + } + + /* Similarly, check for the (?(DEFINE) "condition", which is always + false. */ + + else if (namelen == 6 && STRNCMP_UC_C8(name, STRING_DEFINE, 6) == 0) + { + code[1+LINK_SIZE] = OP_DEF; + skipbytes = 1; + } + + /* Check for the "name" actually being a subpattern number. We are + in the second pass here, so final_bracount is set. */ + + else if (recno > 0 && recno <= cd->final_bracount) + { + PUT2(code, 2+LINK_SIZE, recno); + } + + /* Either an unidentified subpattern, or a reference to (?(0) */ + + else + { + *errorcodeptr = (recno == 0)? ERR35: ERR15; + goto FAILED; + } + break; + + + /* ------------------------------------------------------------ */ + case CHAR_EQUALS_SIGN: /* Positive lookahead */ + bravalue = OP_ASSERT; + cd->assert_depth += 1; + ptr++; + break; + + + /* ------------------------------------------------------------ */ + case CHAR_EXCLAMATION_MARK: /* Negative lookahead */ + ptr++; + if (*ptr == CHAR_RIGHT_PARENTHESIS) /* Optimize (?!) */ + { + *code++ = OP_FAIL; + previous = NULL; + continue; + } + bravalue = OP_ASSERT_NOT; + cd->assert_depth += 1; + break; + + + /* ------------------------------------------------------------ */ + case CHAR_LESS_THAN_SIGN: /* Lookbehind or named define */ + switch (ptr[1]) + { + case CHAR_EQUALS_SIGN: /* Positive lookbehind */ + bravalue = OP_ASSERTBACK; + cd->assert_depth += 1; + ptr += 2; + break; + + case CHAR_EXCLAMATION_MARK: /* Negative lookbehind */ + bravalue = OP_ASSERTBACK_NOT; + cd->assert_depth += 1; + ptr += 2; + break; + + default: /* Could be name define, else bad */ + if (MAX_255(ptr[1]) && (cd->ctypes[ptr[1]] & ctype_word) != 0) + goto DEFINE_NAME; + ptr++; /* Correct offset for error */ + *errorcodeptr = ERR24; + goto FAILED; + } + break; + + + /* ------------------------------------------------------------ */ + case CHAR_GREATER_THAN_SIGN: /* One-time brackets */ + bravalue = OP_ONCE; + ptr++; + break; + + + /* ------------------------------------------------------------ */ + case CHAR_C: /* Callout - may be followed by digits; */ + previous_callout = code; /* Save for later completion */ + after_manual_callout = 1; /* Skip one item before completing */ + *code++ = OP_CALLOUT; + { + int n = 0; + ptr++; + while(IS_DIGIT(*ptr)) + n = n * 10 + *ptr++ - CHAR_0; + if (*ptr != CHAR_RIGHT_PARENTHESIS) + { + *errorcodeptr = ERR39; + goto FAILED; + } + if (n > 255) + { + *errorcodeptr = ERR38; + goto FAILED; + } + *code++ = n; + PUT(code, 0, (int)(ptr - cd->start_pattern + 1)); /* Pattern offset */ + PUT(code, LINK_SIZE, 0); /* Default length */ + code += 2 * LINK_SIZE; + } + previous = NULL; + continue; + + + /* ------------------------------------------------------------ */ + case CHAR_P: /* Python-style named subpattern handling */ + if (*(++ptr) == CHAR_EQUALS_SIGN || + *ptr == CHAR_GREATER_THAN_SIGN) /* Reference or recursion */ + { + is_recurse = *ptr == CHAR_GREATER_THAN_SIGN; + terminator = CHAR_RIGHT_PARENTHESIS; + goto NAMED_REF_OR_RECURSE; + } + else if (*ptr != CHAR_LESS_THAN_SIGN) /* Test for Python-style defn */ + { + *errorcodeptr = ERR41; + goto FAILED; + } + /* Fall through to handle (?P< as (?< is handled */ + + + /* ------------------------------------------------------------ */ + DEFINE_NAME: /* Come here from (?< handling */ + case CHAR_APOSTROPHE: + { + terminator = (*ptr == CHAR_LESS_THAN_SIGN)? + CHAR_GREATER_THAN_SIGN : CHAR_APOSTROPHE; + name = ++ptr; + + while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0) ptr++; + namelen = (int)(ptr - name); + + /* In the pre-compile phase, just do a syntax check. */ + + if (lengthptr != NULL) + { + if (*ptr != terminator) + { + *errorcodeptr = ERR42; + goto FAILED; + } + if (cd->names_found >= MAX_NAME_COUNT) + { + *errorcodeptr = ERR49; + goto FAILED; + } + if (namelen + IMM2_SIZE + 1 > cd->name_entry_size) + { + cd->name_entry_size = namelen + IMM2_SIZE + 1; + if (namelen > MAX_NAME_SIZE) + { + *errorcodeptr = ERR48; + goto FAILED; + } + } + } + + /* In the real compile, create the entry in the table, maintaining + alphabetical order. Duplicate names for different numbers are + permitted only if PCRE_DUPNAMES is set. Duplicate names for the same + number are always OK. (An existing number can be re-used if (?| + appears in the pattern.) In either event, a duplicate name results in + a duplicate entry in the table, even if the number is the same. This + is because the number of names, and hence the table size, is computed + in the pre-compile, and it affects various numbers and pointers which + would all have to be modified, and the compiled code moved down, if + duplicates with the same number were omitted from the table. This + doesn't seem worth the hassle. However, *different* names for the + same number are not permitted. */ + + else + { + BOOL dupname = FALSE; + slot = cd->name_table; + + for (i = 0; i < cd->names_found; i++) + { + int crc = memcmp(name, slot+IMM2_SIZE, IN_UCHARS(namelen)); + if (crc == 0) + { + if (slot[IMM2_SIZE+namelen] == 0) + { + if (GET2(slot, 0) != cd->bracount + 1 && + (options & PCRE_DUPNAMES) == 0) + { + *errorcodeptr = ERR43; + goto FAILED; + } + else dupname = TRUE; + } + else crc = -1; /* Current name is a substring */ + } + + /* Make space in the table and break the loop for an earlier + name. For a duplicate or later name, carry on. We do this for + duplicates so that in the simple case (when ?(| is not used) they + are in order of their numbers. */ + + if (crc < 0) + { + memmove(slot + cd->name_entry_size, slot, + IN_UCHARS((cd->names_found - i) * cd->name_entry_size)); + break; + } + + /* Continue the loop for a later or duplicate name */ + + slot += cd->name_entry_size; + } + + /* For non-duplicate names, check for a duplicate number before + adding the new name. */ + + if (!dupname) + { + pcre_uchar *cslot = cd->name_table; + for (i = 0; i < cd->names_found; i++) + { + if (cslot != slot) + { + if (GET2(cslot, 0) == cd->bracount + 1) + { + *errorcodeptr = ERR65; + goto FAILED; + } + } + else i--; + cslot += cd->name_entry_size; + } + } + + PUT2(slot, 0, cd->bracount + 1); + memcpy(slot + IMM2_SIZE, name, IN_UCHARS(namelen)); + slot[IMM2_SIZE + namelen] = 0; + } + } + + /* In both pre-compile and compile, count the number of names we've + encountered. */ + + cd->names_found++; + ptr++; /* Move past > or ' */ + goto NUMBERED_GROUP; + + + /* ------------------------------------------------------------ */ + case CHAR_AMPERSAND: /* Perl recursion/subroutine syntax */ + terminator = CHAR_RIGHT_PARENTHESIS; + is_recurse = TRUE; + /* Fall through */ + + /* We come here from the Python syntax above that handles both + references (?P=name) and recursion (?P>name), as well as falling + through from the Perl recursion syntax (?&name). We also come here from + the Perl \k or \k'name' back reference syntax and the \k{name} + .NET syntax, and the Oniguruma \g<...> and \g'...' subroutine syntax. */ + + NAMED_REF_OR_RECURSE: + name = ++ptr; + while (MAX_255(*ptr) && (cd->ctypes[*ptr] & ctype_word) != 0) ptr++; + namelen = (int)(ptr - name); + + /* In the pre-compile phase, do a syntax check. We used to just set + a dummy reference number, because it was not used in the first pass. + However, with the change of recursive back references to be atomic, + we have to look for the number so that this state can be identified, as + otherwise the incorrect length is computed. If it's not a backwards + reference, the dummy number will do. */ + + if (lengthptr != NULL) + { + const pcre_uchar *temp; + + if (namelen == 0) + { + *errorcodeptr = ERR62; + goto FAILED; + } + if (*ptr != terminator) + { + *errorcodeptr = ERR42; + goto FAILED; + } + if (namelen > MAX_NAME_SIZE) + { + *errorcodeptr = ERR48; + goto FAILED; + } + + /* The name table does not exist in the first pass, so we cannot + do a simple search as in the code below. Instead, we have to scan the + pattern to find the number. It is important that we scan it only as + far as we have got because the syntax of named subpatterns has not + been checked for the rest of the pattern, and find_parens() assumes + correct syntax. In any case, it's a waste of resources to scan + further. We stop the scan at the current point by temporarily + adjusting the value of cd->endpattern. */ + + temp = cd->end_pattern; + cd->end_pattern = ptr; + recno = find_parens(cd, name, namelen, + (options & PCRE_EXTENDED) != 0, utf); + cd->end_pattern = temp; + if (recno < 0) recno = 0; /* Forward ref; set dummy number */ + } + + /* In the real compile, seek the name in the table. We check the name + first, and then check that we have reached the end of the name in the + table. That way, if the name that is longer than any in the table, + the comparison will fail without reading beyond the table entry. */ + + else + { + slot = cd->name_table; + for (i = 0; i < cd->names_found; i++) + { + if (STRNCMP_UC_UC(name, slot+IMM2_SIZE, namelen) == 0 && + slot[IMM2_SIZE+namelen] == 0) + break; + slot += cd->name_entry_size; + } + + if (i < cd->names_found) /* Back reference */ + { + recno = GET2(slot, 0); + } + else if ((recno = /* Forward back reference */ + find_parens(cd, name, namelen, + (options & PCRE_EXTENDED) != 0, utf)) <= 0) + { + *errorcodeptr = ERR15; + goto FAILED; + } + } + + /* In both phases, we can now go to the code than handles numerical + recursion or backreferences. */ + + if (is_recurse) goto HANDLE_RECURSION; + else goto HANDLE_REFERENCE; + + + /* ------------------------------------------------------------ */ + case CHAR_R: /* Recursion */ + ptr++; /* Same as (?0) */ + /* Fall through */ + + + /* ------------------------------------------------------------ */ + case CHAR_MINUS: case CHAR_PLUS: /* Recursion or subroutine */ + case CHAR_0: case CHAR_1: case CHAR_2: case CHAR_3: case CHAR_4: + case CHAR_5: case CHAR_6: case CHAR_7: case CHAR_8: case CHAR_9: + { + const pcre_uchar *called; + terminator = CHAR_RIGHT_PARENTHESIS; + + /* Come here from the \g<...> and \g'...' code (Oniguruma + compatibility). However, the syntax has been checked to ensure that + the ... are a (signed) number, so that neither ERR63 nor ERR29 will + be called on this path, nor with the jump to OTHER_CHAR_AFTER_QUERY + ever be taken. */ + + HANDLE_NUMERICAL_RECURSION: + + if ((refsign = *ptr) == CHAR_PLUS) + { + ptr++; + if (!IS_DIGIT(*ptr)) + { + *errorcodeptr = ERR63; + goto FAILED; + } + } + else if (refsign == CHAR_MINUS) + { + if (!IS_DIGIT(ptr[1])) + goto OTHER_CHAR_AFTER_QUERY; + ptr++; + } + + recno = 0; + while(IS_DIGIT(*ptr)) + recno = recno * 10 + *ptr++ - CHAR_0; + + if (*ptr != terminator) + { + *errorcodeptr = ERR29; + goto FAILED; + } + + if (refsign == CHAR_MINUS) + { + if (recno == 0) + { + *errorcodeptr = ERR58; + goto FAILED; + } + recno = cd->bracount - recno + 1; + if (recno <= 0) + { + *errorcodeptr = ERR15; + goto FAILED; + } + } + else if (refsign == CHAR_PLUS) + { + if (recno == 0) + { + *errorcodeptr = ERR58; + goto FAILED; + } + recno += cd->bracount; + } + + /* Come here from code above that handles a named recursion */ + + HANDLE_RECURSION: + + previous = code; + called = cd->start_code; + + /* When we are actually compiling, find the bracket that is being + referenced. Temporarily end the regex in case it doesn't exist before + this point. If we end up with a forward reference, first check that + the bracket does occur later so we can give the error (and position) + now. Then remember this forward reference in the workspace so it can + be filled in at the end. */ + + if (lengthptr == NULL) + { + *code = OP_END; + if (recno != 0) + called = PRIV(find_bracket)(cd->start_code, utf, recno); + + /* Forward reference */ + + if (called == NULL) + { + if (find_parens(cd, NULL, recno, + (options & PCRE_EXTENDED) != 0, utf) < 0) + { + *errorcodeptr = ERR15; + goto FAILED; + } + + /* Fudge the value of "called" so that when it is inserted as an + offset below, what it actually inserted is the reference number + of the group. Then remember the forward reference. */ + + called = cd->start_code + recno; + if (cd->hwm >= cd->start_workspace + cd->workspace_size - + WORK_SIZE_SAFETY_MARGIN) + { + *errorcodeptr = expand_workspace(cd); + if (*errorcodeptr != 0) goto FAILED; + } + PUTINC(cd->hwm, 0, (int)(code + 1 - cd->start_code)); + } + + /* If not a forward reference, and the subpattern is still open, + this is a recursive call. We check to see if this is a left + recursion that could loop for ever, and diagnose that case. We + must not, however, do this check if we are in a conditional + subpattern because the condition might be testing for recursion in + a pattern such as /(?(R)a+|(?R)b)/, which is perfectly valid. + Forever loops are also detected at runtime, so those that occur in + conditional subpatterns will be picked up then. */ + + else if (GET(called, 1) == 0 && cond_depth <= 0 && + could_be_empty(called, code, bcptr, utf, cd)) + { + *errorcodeptr = ERR40; + goto FAILED; + } + } + + /* Insert the recursion/subroutine item. It does not have a set first + character (relevant if it is repeated, because it will then be + wrapped with ONCE brackets). */ + + *code = OP_RECURSE; + PUT(code, 1, (int)(called - cd->start_code)); + code += 1 + LINK_SIZE; + groupsetfirstchar = FALSE; + } + + /* Can't determine a first byte now */ + + if (firstchar == REQ_UNSET) firstchar = REQ_NONE; + continue; + + + /* ------------------------------------------------------------ */ + default: /* Other characters: check option setting */ + OTHER_CHAR_AFTER_QUERY: + set = unset = 0; + optset = &set; + + while (*ptr != CHAR_RIGHT_PARENTHESIS && *ptr != CHAR_COLON) + { + switch (*ptr++) + { + case CHAR_MINUS: optset = &unset; break; + + case CHAR_J: /* Record that it changed in the external options */ + *optset |= PCRE_DUPNAMES; + cd->external_flags |= PCRE_JCHANGED; + break; + + case CHAR_i: *optset |= PCRE_CASELESS; break; + case CHAR_m: *optset |= PCRE_MULTILINE; break; + case CHAR_s: *optset |= PCRE_DOTALL; break; + case CHAR_x: *optset |= PCRE_EXTENDED; break; + case CHAR_U: *optset |= PCRE_UNGREEDY; break; + case CHAR_X: *optset |= PCRE_EXTRA; break; + + default: *errorcodeptr = ERR12; + ptr--; /* Correct the offset */ + goto FAILED; + } + } + + /* Set up the changed option bits, but don't change anything yet. */ + + newoptions = (options | set) & (~unset); + + /* If the options ended with ')' this is not the start of a nested + group with option changes, so the options change at this level. If this + item is right at the start of the pattern, the options can be + abstracted and made external in the pre-compile phase, and ignored in + the compile phase. This can be helpful when matching -- for instance in + caseless checking of required bytes. + + If the code pointer is not (cd->start_code + 1 + LINK_SIZE), we are + definitely *not* at the start of the pattern because something has been + compiled. In the pre-compile phase, however, the code pointer can have + that value after the start, because it gets reset as code is discarded + during the pre-compile. However, this can happen only at top level - if + we are within parentheses, the starting BRA will still be present. At + any parenthesis level, the length value can be used to test if anything + has been compiled at that level. Thus, a test for both these conditions + is necessary to ensure we correctly detect the start of the pattern in + both phases. + + If we are not at the pattern start, reset the greedy defaults and the + case value for firstchar and reqchar. */ + + if (*ptr == CHAR_RIGHT_PARENTHESIS) + { + if (code == cd->start_code + 1 + LINK_SIZE && + (lengthptr == NULL || *lengthptr == 2 + 2*LINK_SIZE)) + { + cd->external_options = newoptions; + } + else + { + greedy_default = ((newoptions & PCRE_UNGREEDY) != 0); + greedy_non_default = greedy_default ^ 1; + req_caseopt = ((newoptions & PCRE_CASELESS) != 0)? REQ_CASELESS:0; + } + + /* Change options at this level, and pass them back for use + in subsequent branches. */ + + *optionsptr = options = newoptions; + previous = NULL; /* This item can't be repeated */ + continue; /* It is complete */ + } + + /* If the options ended with ':' we are heading into a nested group + with possible change of options. Such groups are non-capturing and are + not assertions of any kind. All we need to do is skip over the ':'; + the newoptions value is handled below. */ + + bravalue = OP_BRA; + ptr++; + } /* End of switch for character following (? */ + } /* End of (? handling */ + + /* Opening parenthesis not followed by '*' or '?'. If PCRE_NO_AUTO_CAPTURE + is set, all unadorned brackets become non-capturing and behave like (?:...) + brackets. */ + + else if ((options & PCRE_NO_AUTO_CAPTURE) != 0) + { + bravalue = OP_BRA; + } + + /* Else we have a capturing group. */ + + else + { + NUMBERED_GROUP: + cd->bracount += 1; + PUT2(code, 1+LINK_SIZE, cd->bracount); + skipbytes = IMM2_SIZE; + } + + /* Process nested bracketed regex. Assertions used not to be repeatable, + but this was changed for Perl compatibility, so all kinds can now be + repeated. We copy code into a non-register variable (tempcode) in order to + be able to pass its address because some compilers complain otherwise. */ + + previous = code; /* For handling repetition */ + *code = bravalue; + tempcode = code; + tempreqvary = cd->req_varyopt; /* Save value before bracket */ + tempbracount = cd->bracount; /* Save value before bracket */ + length_prevgroup = 0; /* Initialize for pre-compile phase */ + + if (!compile_regex( + newoptions, /* The complete new option state */ + &tempcode, /* Where to put code (updated) */ + &ptr, /* Input pointer (updated) */ + errorcodeptr, /* Where to put an error message */ + (bravalue == OP_ASSERTBACK || + bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */ + reset_bracount, /* True if (?| group */ + skipbytes, /* Skip over bracket number */ + cond_depth + + ((bravalue == OP_COND)?1:0), /* Depth of condition subpatterns */ + &subfirstchar, /* For possible first char */ + &subreqchar, /* For possible last char */ + bcptr, /* Current branch chain */ + cd, /* Tables block */ + (lengthptr == NULL)? NULL : /* Actual compile phase */ + &length_prevgroup /* Pre-compile phase */ + )) + goto FAILED; + + /* If this was an atomic group and there are no capturing groups within it, + generate OP_ONCE_NC instead of OP_ONCE. */ + + if (bravalue == OP_ONCE && cd->bracount <= tempbracount) + *code = OP_ONCE_NC; + + if (bravalue >= OP_ASSERT && bravalue <= OP_ASSERTBACK_NOT) + cd->assert_depth -= 1; + + /* At the end of compiling, code is still pointing to the start of the + group, while tempcode has been updated to point past the end of the group. + The pattern pointer (ptr) is on the bracket. + + If this is a conditional bracket, check that there are no more than + two branches in the group, or just one if it's a DEFINE group. We do this + in the real compile phase, not in the pre-pass, where the whole group may + not be available. */ + + if (bravalue == OP_COND && lengthptr == NULL) + { + pcre_uchar *tc = code; + int condcount = 0; + + do { + condcount++; + tc += GET(tc,1); + } + while (*tc != OP_KET); + + /* A DEFINE group is never obeyed inline (the "condition" is always + false). It must have only one branch. */ + + if (code[LINK_SIZE+1] == OP_DEF) + { + if (condcount > 1) + { + *errorcodeptr = ERR54; + goto FAILED; + } + bravalue = OP_DEF; /* Just a flag to suppress char handling below */ + } + + /* A "normal" conditional group. If there is just one branch, we must not + make use of its firstchar or reqchar, because this is equivalent to an + empty second branch. */ + + else + { + if (condcount > 2) + { + *errorcodeptr = ERR27; + goto FAILED; + } + if (condcount == 1) subfirstchar = subreqchar = REQ_NONE; + } + } + + /* Error if hit end of pattern */ + + if (*ptr != CHAR_RIGHT_PARENTHESIS) + { + *errorcodeptr = ERR14; + goto FAILED; + } + + /* In the pre-compile phase, update the length by the length of the group, + less the brackets at either end. Then reduce the compiled code to just a + set of non-capturing brackets so that it doesn't use much memory if it is + duplicated by a quantifier.*/ + + if (lengthptr != NULL) + { + if (OFLOW_MAX - *lengthptr < length_prevgroup - 2 - 2*LINK_SIZE) + { + *errorcodeptr = ERR20; + goto FAILED; + } + *lengthptr += length_prevgroup - 2 - 2*LINK_SIZE; + code++; /* This already contains bravalue */ + PUTINC(code, 0, 1 + LINK_SIZE); + *code++ = OP_KET; + PUTINC(code, 0, 1 + LINK_SIZE); + break; /* No need to waste time with special character handling */ + } + + /* Otherwise update the main code pointer to the end of the group. */ + + code = tempcode; + + /* For a DEFINE group, required and first character settings are not + relevant. */ + + if (bravalue == OP_DEF) break; + + /* Handle updating of the required and first characters for other types of + group. Update for normal brackets of all kinds, and conditions with two + branches (see code above). If the bracket is followed by a quantifier with + zero repeat, we have to back off. Hence the definition of zeroreqchar and + zerofirstchar outside the main loop so that they can be accessed for the + back off. */ + + zeroreqchar = reqchar; + zerofirstchar = firstchar; + groupsetfirstchar = FALSE; + + if (bravalue >= OP_ONCE) + { + /* If we have not yet set a firstchar in this branch, take it from the + subpattern, remembering that it was set here so that a repeat of more + than one can replicate it as reqchar if necessary. If the subpattern has + no firstchar, set "none" for the whole branch. In both cases, a zero + repeat forces firstchar to "none". */ + + if (firstchar == REQ_UNSET) + { + if (subfirstchar >= 0) + { + firstchar = subfirstchar; + groupsetfirstchar = TRUE; + } + else firstchar = REQ_NONE; + zerofirstchar = REQ_NONE; + } + + /* If firstchar was previously set, convert the subpattern's firstchar + into reqchar if there wasn't one, using the vary flag that was in + existence beforehand. */ + + else if (subfirstchar >= 0 && subreqchar < 0) + subreqchar = subfirstchar | tempreqvary; + + /* If the subpattern set a required byte (or set a first byte that isn't + really the first byte - see above), set it. */ + + if (subreqchar >= 0) reqchar = subreqchar; + } + + /* For a forward assertion, we take the reqchar, if set. This can be + helpful if the pattern that follows the assertion doesn't set a different + char. For example, it's useful for /(?=abcde).+/. We can't set firstchar + for an assertion, however because it leads to incorrect effect for patterns + such as /(?=a)a.+/ when the "real" "a" would then become a reqchar instead + of a firstchar. This is overcome by a scan at the end if there's no + firstchar, looking for an asserted first char. */ + + else if (bravalue == OP_ASSERT && subreqchar >= 0) reqchar = subreqchar; + break; /* End of processing '(' */ + + + /* ===================================================================*/ + /* Handle metasequences introduced by \. For ones like \d, the ESC_ values + are arranged to be the negation of the corresponding OP_values in the + default case when PCRE_UCP is not set. For the back references, the values + are ESC_REF plus the reference number. Only back references and those types + that consume a character may be repeated. We can test for values between + ESC_b and ESC_Z for the latter; this may have to change if any new ones are + ever created. */ + + case CHAR_BACKSLASH: + tempptr = ptr; + c = check_escape(&ptr, errorcodeptr, cd->bracount, options, FALSE); + if (*errorcodeptr != 0) goto FAILED; + + if (c < 0) + { + if (-c == ESC_Q) /* Handle start of quoted string */ + { + if (ptr[1] == CHAR_BACKSLASH && ptr[2] == CHAR_E) + ptr += 2; /* avoid empty string */ + else inescq = TRUE; + continue; + } + + if (-c == ESC_E) continue; /* Perl ignores an orphan \E */ + + /* For metasequences that actually match a character, we disable the + setting of a first character if it hasn't already been set. */ + + if (firstchar == REQ_UNSET && -c > ESC_b && -c < ESC_Z) + firstchar = REQ_NONE; + + /* Set values to reset to if this is followed by a zero repeat. */ + + zerofirstchar = firstchar; + zeroreqchar = reqchar; + + /* \g or \g'name' is a subroutine call by name and \g or \g'n' + is a subroutine call by number (Oniguruma syntax). In fact, the value + -ESC_g is returned only for these cases. So we don't need to check for < + or ' if the value is -ESC_g. For the Perl syntax \g{n} the value is + -ESC_REF+n, and for the Perl syntax \g{name} the result is -ESC_k (as + that is a synonym for a named back reference). */ + + if (-c == ESC_g) + { + const pcre_uchar *p; + save_hwm = cd->hwm; /* Normally this is set when '(' is read */ + terminator = (*(++ptr) == CHAR_LESS_THAN_SIGN)? + CHAR_GREATER_THAN_SIGN : CHAR_APOSTROPHE; + + /* These two statements stop the compiler for warning about possibly + unset variables caused by the jump to HANDLE_NUMERICAL_RECURSION. In + fact, because we actually check for a number below, the paths that + would actually be in error are never taken. */ + + skipbytes = 0; + reset_bracount = FALSE; + + /* Test for a name */ + + if (ptr[1] != CHAR_PLUS && ptr[1] != CHAR_MINUS) + { + BOOL is_a_number = TRUE; + for (p = ptr + 1; *p != 0 && *p != terminator; p++) + { + if (!MAX_255(*p)) { is_a_number = FALSE; break; } + if ((cd->ctypes[*p] & ctype_digit) == 0) is_a_number = FALSE; + if ((cd->ctypes[*p] & ctype_word) == 0) break; + } + if (*p != terminator) + { + *errorcodeptr = ERR57; + break; + } + if (is_a_number) + { + ptr++; + goto HANDLE_NUMERICAL_RECURSION; + } + is_recurse = TRUE; + goto NAMED_REF_OR_RECURSE; + } + + /* Test a signed number in angle brackets or quotes. */ + + p = ptr + 2; + while (IS_DIGIT(*p)) p++; + if (*p != terminator) + { + *errorcodeptr = ERR57; + break; + } + ptr++; + goto HANDLE_NUMERICAL_RECURSION; + } + + /* \k or \k'name' is a back reference by name (Perl syntax). + We also support \k{name} (.NET syntax). */ + + if (-c == ESC_k) + { + if ((ptr[1] != CHAR_LESS_THAN_SIGN && + ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET)) + { + *errorcodeptr = ERR69; + break; + } + is_recurse = FALSE; + terminator = (*(++ptr) == CHAR_LESS_THAN_SIGN)? + CHAR_GREATER_THAN_SIGN : (*ptr == CHAR_APOSTROPHE)? + CHAR_APOSTROPHE : CHAR_RIGHT_CURLY_BRACKET; + goto NAMED_REF_OR_RECURSE; + } + + /* Back references are handled specially; must disable firstchar if + not set to cope with cases like (?=(\w+))\1: which would otherwise set + ':' later. */ + + if (-c >= ESC_REF) + { + open_capitem *oc; + recno = -c - ESC_REF; + + HANDLE_REFERENCE: /* Come here from named backref handling */ + if (firstchar == REQ_UNSET) firstchar = REQ_NONE; + previous = code; + *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF; + PUT2INC(code, 0, recno); + cd->backref_map |= (recno < 32)? (1 << recno) : 1; + if (recno > cd->top_backref) cd->top_backref = recno; + + /* Check to see if this back reference is recursive, that it, it + is inside the group that it references. A flag is set so that the + group can be made atomic. */ + + for (oc = cd->open_caps; oc != NULL; oc = oc->next) + { + if (oc->number == recno) + { + oc->flag = TRUE; + break; + } + } + } + + /* So are Unicode property matches, if supported. */ + +#ifdef SUPPORT_UCP + else if (-c == ESC_P || -c == ESC_p) + { + BOOL negated; + int pdata; + int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr); + if (ptype < 0) goto FAILED; + previous = code; + *code++ = ((-c == ESC_p) != negated)? OP_PROP : OP_NOTPROP; + *code++ = ptype; + *code++ = pdata; + } +#else + + /* If Unicode properties are not supported, \X, \P, and \p are not + allowed. */ + + else if (-c == ESC_X || -c == ESC_P || -c == ESC_p) + { + *errorcodeptr = ERR45; + goto FAILED; + } +#endif + + /* For the rest (including \X when Unicode properties are supported), we + can obtain the OP value by negating the escape value in the default + situation when PCRE_UCP is not set. When it *is* set, we substitute + Unicode property tests. Note that \b and \B do a one-character + lookbehind. */ + + else + { + if ((-c == ESC_b || -c == ESC_B) && cd->max_lookbehind == 0) + cd->max_lookbehind = 1; +#ifdef SUPPORT_UCP + if (-c >= ESC_DU && -c <= ESC_wu) + { + nestptr = ptr + 1; /* Where to resume */ + ptr = substitutes[-c - ESC_DU] - 1; /* Just before substitute */ + } + else +#endif + /* In non-UTF-8 mode, we turn \C into OP_ALLANY instead of OP_ANYBYTE + so that it works in DFA mode and in lookbehinds. */ + + { + previous = (-c > ESC_b && -c < ESC_Z)? code : NULL; + *code++ = (!utf && c == -ESC_C)? OP_ALLANY : -c; + } + } + continue; + } + + /* We have a data character whose value is in c. In UTF-8 mode it may have + a value > 127. We set its representation in the length/buffer, and then + handle it as a data character. */ + +#ifdef SUPPORT_UTF + if (utf && c > MAX_VALUE_FOR_SINGLE_CHAR) + mclength = PRIV(ord2utf)(c, mcbuffer); + else +#endif + + { + mcbuffer[0] = c; + mclength = 1; + } + goto ONE_CHAR; + + + /* ===================================================================*/ + /* Handle a literal character. It is guaranteed not to be whitespace or # + when the extended flag is set. If we are in UTF-8 mode, it may be a + multi-byte literal character. */ + + default: + NORMAL_CHAR: + mclength = 1; + mcbuffer[0] = c; + +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(c)) + ACROSSCHAR(TRUE, ptr[1], mcbuffer[mclength++] = *(++ptr)); +#endif + + /* At this point we have the character's bytes in mcbuffer, and the length + in mclength. When not in UTF-8 mode, the length is always 1. */ + + ONE_CHAR: + previous = code; + *code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARI : OP_CHAR; + for (c = 0; c < mclength; c++) *code++ = mcbuffer[c]; + + /* Remember if \r or \n were seen */ + + if (mcbuffer[0] == CHAR_CR || mcbuffer[0] == CHAR_NL) + cd->external_flags |= PCRE_HASCRORLF; + + /* Set the first and required bytes appropriately. If no previous first + byte, set it from this character, but revert to none on a zero repeat. + Otherwise, leave the firstchar value alone, and don't change it on a zero + repeat. */ + + if (firstchar == REQ_UNSET) + { + zerofirstchar = REQ_NONE; + zeroreqchar = reqchar; + + /* If the character is more than one byte long, we can set firstchar + only if it is not to be matched caselessly. */ + + if (mclength == 1 || req_caseopt == 0) + { + firstchar = mcbuffer[0] | req_caseopt; + if (mclength != 1) reqchar = code[-1] | cd->req_varyopt; + } + else firstchar = reqchar = REQ_NONE; + } + + /* firstchar was previously set; we can set reqchar only if the length is + 1 or the matching is caseful. */ + + else + { + zerofirstchar = firstchar; + zeroreqchar = reqchar; + if (mclength == 1 || req_caseopt == 0) + reqchar = code[-1] | req_caseopt | cd->req_varyopt; + } + + break; /* End of literal character handling */ + } + } /* end of big loop */ + + +/* Control never reaches here by falling through, only by a goto for all the +error states. Pass back the position in the pattern so that it can be displayed +to the user for diagnosing the error. */ + +FAILED: +*ptrptr = ptr; +return FALSE; +} + + + + +/************************************************* +* Compile sequence of alternatives * +*************************************************/ + +/* On entry, ptr is pointing past the bracket character, but on return it +points to the closing bracket, or vertical bar, or end of string. The code +variable is pointing at the byte into which the BRA operator has been stored. +This function is used during the pre-compile phase when we are trying to find +out the amount of memory needed, as well as during the real compile phase. The +value of lengthptr distinguishes the two phases. + +Arguments: + options option bits, including any changes for this subpattern + codeptr -> the address of the current code pointer + ptrptr -> the address of the current pattern pointer + errorcodeptr -> pointer to error code variable + lookbehind TRUE if this is a lookbehind assertion + reset_bracount TRUE to reset the count for each branch + skipbytes skip this many bytes at start (for brackets and OP_COND) + cond_depth depth of nesting for conditional subpatterns + firstcharptr place to put the first required character, or a negative number + reqcharptr place to put the last required character, or a negative number + bcptr pointer to the chain of currently open branches + cd points to the data block with tables pointers etc. + lengthptr NULL during the real compile phase + points to length accumulator during pre-compile phase + +Returns: TRUE on success +*/ + +static BOOL +compile_regex(int options, pcre_uchar **codeptr, const pcre_uchar **ptrptr, + int *errorcodeptr, BOOL lookbehind, BOOL reset_bracount, int skipbytes, + int cond_depth, pcre_int32 *firstcharptr, pcre_int32 *reqcharptr, + branch_chain *bcptr, compile_data *cd, int *lengthptr) +{ +const pcre_uchar *ptr = *ptrptr; +pcre_uchar *code = *codeptr; +pcre_uchar *last_branch = code; +pcre_uchar *start_bracket = code; +pcre_uchar *reverse_count = NULL; +open_capitem capitem; +int capnumber = 0; +pcre_int32 firstchar, reqchar; +pcre_int32 branchfirstchar, branchreqchar; +int length; +int orig_bracount; +int max_bracount; +branch_chain bc; + +bc.outer = bcptr; +bc.current_branch = code; + +firstchar = reqchar = REQ_UNSET; + +/* Accumulate the length for use in the pre-compile phase. Start with the +length of the BRA and KET and any extra bytes that are required at the +beginning. We accumulate in a local variable to save frequent testing of +lenthptr for NULL. We cannot do this by looking at the value of code at the +start and end of each alternative, because compiled items are discarded during +the pre-compile phase so that the work space is not exceeded. */ + +length = 2 + 2*LINK_SIZE + skipbytes; + +/* WARNING: If the above line is changed for any reason, you must also change +the code that abstracts option settings at the start of the pattern and makes +them global. It tests the value of length for (2 + 2*LINK_SIZE) in the +pre-compile phase to find out whether anything has yet been compiled or not. */ + +/* If this is a capturing subpattern, add to the chain of open capturing items +so that we can detect them if (*ACCEPT) is encountered. This is also used to +detect groups that contain recursive back references to themselves. Note that +only OP_CBRA need be tested here; changing this opcode to one of its variants, +e.g. OP_SCBRAPOS, happens later, after the group has been compiled. */ + +if (*code == OP_CBRA) + { + capnumber = GET2(code, 1 + LINK_SIZE); + capitem.number = capnumber; + capitem.next = cd->open_caps; + capitem.flag = FALSE; + cd->open_caps = &capitem; + } + +/* Offset is set zero to mark that this bracket is still open */ + +PUT(code, 1, 0); +code += 1 + LINK_SIZE + skipbytes; + +/* Loop for each alternative branch */ + +orig_bracount = max_bracount = cd->bracount; +for (;;) + { + /* For a (?| group, reset the capturing bracket count so that each branch + uses the same numbers. */ + + if (reset_bracount) cd->bracount = orig_bracount; + + /* Set up dummy OP_REVERSE if lookbehind assertion */ + + if (lookbehind) + { + *code++ = OP_REVERSE; + reverse_count = code; + PUTINC(code, 0, 0); + length += 1 + LINK_SIZE; + } + + /* Now compile the branch; in the pre-compile phase its length gets added + into the length. */ + + if (!compile_branch(&options, &code, &ptr, errorcodeptr, &branchfirstchar, + &branchreqchar, &bc, cond_depth, cd, + (lengthptr == NULL)? NULL : &length)) + { + *ptrptr = ptr; + return FALSE; + } + + /* Keep the highest bracket count in case (?| was used and some branch + has fewer than the rest. */ + + if (cd->bracount > max_bracount) max_bracount = cd->bracount; + + /* In the real compile phase, there is some post-processing to be done. */ + + if (lengthptr == NULL) + { + /* If this is the first branch, the firstchar and reqchar values for the + branch become the values for the regex. */ + + if (*last_branch != OP_ALT) + { + firstchar = branchfirstchar; + reqchar = branchreqchar; + } + + /* If this is not the first branch, the first char and reqchar have to + match the values from all the previous branches, except that if the + previous value for reqchar didn't have REQ_VARY set, it can still match, + and we set REQ_VARY for the regex. */ + + else + { + /* If we previously had a firstchar, but it doesn't match the new branch, + we have to abandon the firstchar for the regex, but if there was + previously no reqchar, it takes on the value of the old firstchar. */ + + if (firstchar >= 0 && firstchar != branchfirstchar) + { + if (reqchar < 0) reqchar = firstchar; + firstchar = REQ_NONE; + } + + /* If we (now or from before) have no firstchar, a firstchar from the + branch becomes a reqchar if there isn't a branch reqchar. */ + + if (firstchar < 0 && branchfirstchar >= 0 && branchreqchar < 0) + branchreqchar = branchfirstchar; + + /* Now ensure that the reqchars match */ + + if ((reqchar & ~REQ_VARY) != (branchreqchar & ~REQ_VARY)) + reqchar = REQ_NONE; + else reqchar |= branchreqchar; /* To "or" REQ_VARY */ + } + + /* If lookbehind, check that this branch matches a fixed-length string, and + put the length into the OP_REVERSE item. Temporarily mark the end of the + branch with OP_END. If the branch contains OP_RECURSE, the result is -3 + because there may be forward references that we can't check here. Set a + flag to cause another lookbehind check at the end. Why not do it all at the + end? Because common, erroneous checks are picked up here and the offset of + the problem can be shown. */ + + if (lookbehind) + { + int fixed_length; + *code = OP_END; + fixed_length = find_fixedlength(last_branch, (options & PCRE_UTF8) != 0, + FALSE, cd); + DPRINTF(("fixed length = %d\n", fixed_length)); + if (fixed_length == -3) + { + cd->check_lookbehind = TRUE; + } + else if (fixed_length < 0) + { + *errorcodeptr = (fixed_length == -2)? ERR36 : + (fixed_length == -4)? ERR70: ERR25; + *ptrptr = ptr; + return FALSE; + } + else + { + if (fixed_length > cd->max_lookbehind) + cd->max_lookbehind = fixed_length; + PUT(reverse_count, 0, fixed_length); + } + } + } + + /* Reached end of expression, either ')' or end of pattern. In the real + compile phase, go back through the alternative branches and reverse the chain + of offsets, with the field in the BRA item now becoming an offset to the + first alternative. If there are no alternatives, it points to the end of the + group. The length in the terminating ket is always the length of the whole + bracketed item. Return leaving the pointer at the terminating char. */ + + if (*ptr != CHAR_VERTICAL_LINE) + { + if (lengthptr == NULL) + { + int branch_length = (int)(code - last_branch); + do + { + int prev_length = GET(last_branch, 1); + PUT(last_branch, 1, branch_length); + branch_length = prev_length; + last_branch -= branch_length; + } + while (branch_length > 0); + } + + /* Fill in the ket */ + + *code = OP_KET; + PUT(code, 1, (int)(code - start_bracket)); + code += 1 + LINK_SIZE; + + /* If it was a capturing subpattern, check to see if it contained any + recursive back references. If so, we must wrap it in atomic brackets. + In any event, remove the block from the chain. */ + + if (capnumber > 0) + { + if (cd->open_caps->flag) + { + memmove(start_bracket + 1 + LINK_SIZE, start_bracket, + IN_UCHARS(code - start_bracket)); + *start_bracket = OP_ONCE; + code += 1 + LINK_SIZE; + PUT(start_bracket, 1, (int)(code - start_bracket)); + *code = OP_KET; + PUT(code, 1, (int)(code - start_bracket)); + code += 1 + LINK_SIZE; + length += 2 + 2*LINK_SIZE; + } + cd->open_caps = cd->open_caps->next; + } + + /* Retain the highest bracket number, in case resetting was used. */ + + cd->bracount = max_bracount; + + /* Set values to pass back */ + + *codeptr = code; + *ptrptr = ptr; + *firstcharptr = firstchar; + *reqcharptr = reqchar; + if (lengthptr != NULL) + { + if (OFLOW_MAX - *lengthptr < length) + { + *errorcodeptr = ERR20; + return FALSE; + } + *lengthptr += length; + } + return TRUE; + } + + /* Another branch follows. In the pre-compile phase, we can move the code + pointer back to where it was for the start of the first branch. (That is, + pretend that each branch is the only one.) + + In the real compile phase, insert an ALT node. Its length field points back + to the previous branch while the bracket remains open. At the end the chain + is reversed. It's done like this so that the start of the bracket has a + zero offset until it is closed, making it possible to detect recursion. */ + + if (lengthptr != NULL) + { + code = *codeptr + 1 + LINK_SIZE + skipbytes; + length += 1 + LINK_SIZE; + } + else + { + *code = OP_ALT; + PUT(code, 1, (int)(code - last_branch)); + bc.current_branch = last_branch = code; + code += 1 + LINK_SIZE; + } + + ptr++; + } +/* Control never reaches here */ +} + + + + +/************************************************* +* Check for anchored expression * +*************************************************/ + +/* Try to find out if this is an anchored regular expression. Consider each +alternative branch. If they all start with OP_SOD or OP_CIRC, or with a bracket +all of whose alternatives start with OP_SOD or OP_CIRC (recurse ad lib), then +it's anchored. However, if this is a multiline pattern, then only OP_SOD will +be found, because ^ generates OP_CIRCM in that mode. + +We can also consider a regex to be anchored if OP_SOM starts all its branches. +This is the code for \G, which means "match at start of match position, taking +into account the match offset". + +A branch is also implicitly anchored if it starts with .* and DOTALL is set, +because that will try the rest of the pattern at all possible matching points, +so there is no point trying again.... er .... + +.... except when the .* appears inside capturing parentheses, and there is a +subsequent back reference to those parentheses. We haven't enough information +to catch that case precisely. + +At first, the best we could do was to detect when .* was in capturing brackets +and the highest back reference was greater than or equal to that level. +However, by keeping a bitmap of the first 31 back references, we can catch some +of the more common cases more precisely. + +Arguments: + code points to start of expression (the bracket) + bracket_map a bitmap of which brackets we are inside while testing; this + handles up to substring 31; after that we just have to take + the less precise approach + backref_map the back reference bitmap + +Returns: TRUE or FALSE +*/ + +static BOOL +is_anchored(register const pcre_uchar *code, unsigned int bracket_map, + unsigned int backref_map) +{ +do { + const pcre_uchar *scode = first_significant_code( + code + PRIV(OP_lengths)[*code], FALSE); + register int op = *scode; + + /* Non-capturing brackets */ + + if (op == OP_BRA || op == OP_BRAPOS || + op == OP_SBRA || op == OP_SBRAPOS) + { + if (!is_anchored(scode, bracket_map, backref_map)) return FALSE; + } + + /* Capturing brackets */ + + else if (op == OP_CBRA || op == OP_CBRAPOS || + op == OP_SCBRA || op == OP_SCBRAPOS) + { + int n = GET2(scode, 1+LINK_SIZE); + int new_map = bracket_map | ((n < 32)? (1 << n) : 1); + if (!is_anchored(scode, new_map, backref_map)) return FALSE; + } + + /* Other brackets */ + + else if (op == OP_ASSERT || op == OP_ONCE || op == OP_ONCE_NC || + op == OP_COND) + { + if (!is_anchored(scode, bracket_map, backref_map)) return FALSE; + } + + /* .* is not anchored unless DOTALL is set (which generates OP_ALLANY) and + it isn't in brackets that are or may be referenced. */ + + else if ((op == OP_TYPESTAR || op == OP_TYPEMINSTAR || + op == OP_TYPEPOSSTAR)) + { + if (scode[1] != OP_ALLANY || (bracket_map & backref_map) != 0) + return FALSE; + } + + /* Check for explicit anchoring */ + + else if (op != OP_SOD && op != OP_SOM && op != OP_CIRC) return FALSE; + code += GET(code, 1); + } +while (*code == OP_ALT); /* Loop for each alternative */ +return TRUE; +} + + + +/************************************************* +* Check for starting with ^ or .* * +*************************************************/ + +/* This is called to find out if every branch starts with ^ or .* so that +"first char" processing can be done to speed things up in multiline +matching and for non-DOTALL patterns that start with .* (which must start at +the beginning or after \n). As in the case of is_anchored() (see above), we +have to take account of back references to capturing brackets that contain .* +because in that case we can't make the assumption. + +Arguments: + code points to start of expression (the bracket) + bracket_map a bitmap of which brackets we are inside while testing; this + handles up to substring 31; after that we just have to take + the less precise approach + backref_map the back reference bitmap + +Returns: TRUE or FALSE +*/ + +static BOOL +is_startline(const pcre_uchar *code, unsigned int bracket_map, + unsigned int backref_map) +{ +do { + const pcre_uchar *scode = first_significant_code( + code + PRIV(OP_lengths)[*code], FALSE); + register int op = *scode; + + /* If we are at the start of a conditional assertion group, *both* the + conditional assertion *and* what follows the condition must satisfy the test + for start of line. Other kinds of condition fail. Note that there may be an + auto-callout at the start of a condition. */ + + if (op == OP_COND) + { + scode += 1 + LINK_SIZE; + if (*scode == OP_CALLOUT) scode += PRIV(OP_lengths)[OP_CALLOUT]; + switch (*scode) + { + case OP_CREF: + case OP_NCREF: + case OP_RREF: + case OP_NRREF: + case OP_DEF: + return FALSE; + + default: /* Assertion */ + if (!is_startline(scode, bracket_map, backref_map)) return FALSE; + do scode += GET(scode, 1); while (*scode == OP_ALT); + scode += 1 + LINK_SIZE; + break; + } + scode = first_significant_code(scode, FALSE); + op = *scode; + } + + /* Non-capturing brackets */ + + if (op == OP_BRA || op == OP_BRAPOS || + op == OP_SBRA || op == OP_SBRAPOS) + { + if (!is_startline(scode, bracket_map, backref_map)) return FALSE; + } + + /* Capturing brackets */ + + else if (op == OP_CBRA || op == OP_CBRAPOS || + op == OP_SCBRA || op == OP_SCBRAPOS) + { + int n = GET2(scode, 1+LINK_SIZE); + int new_map = bracket_map | ((n < 32)? (1 << n) : 1); + if (!is_startline(scode, new_map, backref_map)) return FALSE; + } + + /* Other brackets */ + + else if (op == OP_ASSERT || op == OP_ONCE || op == OP_ONCE_NC) + { + if (!is_startline(scode, bracket_map, backref_map)) return FALSE; + } + + /* .* means "start at start or after \n" if it isn't in brackets that + may be referenced. */ + + else if (op == OP_TYPESTAR || op == OP_TYPEMINSTAR || op == OP_TYPEPOSSTAR) + { + if (scode[1] != OP_ANY || (bracket_map & backref_map) != 0) return FALSE; + } + + /* Check for explicit circumflex */ + + else if (op != OP_CIRC && op != OP_CIRCM) return FALSE; + + /* Move on to the next alternative */ + + code += GET(code, 1); + } +while (*code == OP_ALT); /* Loop for each alternative */ +return TRUE; +} + + + +/************************************************* +* Check for asserted fixed first char * +*************************************************/ + +/* During compilation, the "first char" settings from forward assertions are +discarded, because they can cause conflicts with actual literals that follow. +However, if we end up without a first char setting for an unanchored pattern, +it is worth scanning the regex to see if there is an initial asserted first +char. If all branches start with the same asserted char, or with a bracket all +of whose alternatives start with the same asserted char (recurse ad lib), then +we return that char, otherwise -1. + +Arguments: + code points to start of expression (the bracket) + inassert TRUE if in an assertion + +Returns: -1 or the fixed first char +*/ + +static int +find_firstassertedchar(const pcre_uchar *code, BOOL inassert) +{ +register int c = -1; +do { + int d; + int xl = (*code == OP_CBRA || *code == OP_SCBRA || + *code == OP_CBRAPOS || *code == OP_SCBRAPOS)? IMM2_SIZE:0; + const pcre_uchar *scode = first_significant_code(code + 1+LINK_SIZE + xl, + TRUE); + register int op = *scode; + + switch(op) + { + default: + return -1; + + case OP_BRA: + case OP_BRAPOS: + case OP_CBRA: + case OP_SCBRA: + case OP_CBRAPOS: + case OP_SCBRAPOS: + case OP_ASSERT: + case OP_ONCE: + case OP_ONCE_NC: + case OP_COND: + if ((d = find_firstassertedchar(scode, op == OP_ASSERT)) < 0) + return -1; + if (c < 0) c = d; else if (c != d) return -1; + break; + + case OP_EXACT: + scode += IMM2_SIZE; + /* Fall through */ + + case OP_CHAR: + case OP_PLUS: + case OP_MINPLUS: + case OP_POSPLUS: + if (!inassert) return -1; + if (c < 0) c = scode[1]; + else if (c != scode[1]) return -1; + break; + + case OP_EXACTI: + scode += IMM2_SIZE; + /* Fall through */ + + case OP_CHARI: + case OP_PLUSI: + case OP_MINPLUSI: + case OP_POSPLUSI: + if (!inassert) return -1; + if (c < 0) c = scode[1] | REQ_CASELESS; + else if (c != scode[1]) return -1; + break; + } + + code += GET(code, 1); + } +while (*code == OP_ALT); +return c; +} + + + +/************************************************* +* Compile a Regular Expression * +*************************************************/ + +/* This function takes a string and returns a pointer to a block of store +holding a compiled version of the expression. The original API for this +function had no error code return variable; it is retained for backwards +compatibility. The new function is given a new name. + +Arguments: + pattern the regular expression + options various option bits + errorcodeptr pointer to error code variable (pcre_compile2() only) + can be NULL if you don't want a code value + errorptr pointer to pointer to error text + erroroffset ptr offset in pattern where error was detected + tables pointer to character tables or NULL + +Returns: pointer to compiled data block, or NULL on error, + with errorptr and erroroffset set +*/ + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION +pcre_compile(const char *pattern, int options, const char **errorptr, + int *erroroffset, const unsigned char *tables) +#else +PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION +pcre16_compile(PCRE_SPTR16 pattern, int options, const char **errorptr, + int *erroroffset, const unsigned char *tables) +#endif +{ +#ifdef COMPILE_PCRE8 +return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables); +#else +return pcre16_compile2(pattern, options, NULL, errorptr, erroroffset, tables); +#endif +} + + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN pcre * PCRE_CALL_CONVENTION +pcre_compile2(const char *pattern, int options, int *errorcodeptr, + const char **errorptr, int *erroroffset, const unsigned char *tables) +#else +PCRE_EXP_DEFN pcre16 * PCRE_CALL_CONVENTION +pcre16_compile2(PCRE_SPTR16 pattern, int options, int *errorcodeptr, + const char **errorptr, int *erroroffset, const unsigned char *tables) +#endif +{ +REAL_PCRE *re; +int length = 1; /* For final END opcode */ +pcre_int32 firstchar, reqchar; +int newline; +int errorcode = 0; +int skipatstart = 0; +BOOL utf; +size_t size; +pcre_uchar *code; +const pcre_uchar *codestart; +const pcre_uchar *ptr; +compile_data compile_block; +compile_data *cd = &compile_block; + +/* This space is used for "compiling" into during the first phase, when we are +computing the amount of memory that is needed. Compiled items are thrown away +as soon as possible, so that a fairly large buffer should be sufficient for +this purpose. The same space is used in the second phase for remembering where +to fill in forward references to subpatterns. That may overflow, in which case +new memory is obtained from malloc(). */ + +pcre_uchar cworkspace[COMPILE_WORK_SIZE]; + +/* Set this early so that early errors get offset 0. */ + +ptr = (const pcre_uchar *)pattern; + +/* We can't pass back an error message if errorptr is NULL; I guess the best we +can do is just return NULL, but we can set a code value if there is a code +pointer. */ + +if (errorptr == NULL) + { + if (errorcodeptr != NULL) *errorcodeptr = 99; + return NULL; + } + +*errorptr = NULL; +if (errorcodeptr != NULL) *errorcodeptr = ERR0; + +/* However, we can give a message for this error */ + +if (erroroffset == NULL) + { + errorcode = ERR16; + goto PCRE_EARLY_ERROR_RETURN2; + } + +*erroroffset = 0; + +/* Set up pointers to the individual character tables */ + +if (tables == NULL) tables = PRIV(default_tables); +cd->lcc = tables + lcc_offset; +cd->fcc = tables + fcc_offset; +cd->cbits = tables + cbits_offset; +cd->ctypes = tables + ctypes_offset; + +/* Check that all undefined public option bits are zero */ + +if ((options & ~PUBLIC_COMPILE_OPTIONS) != 0) + { + errorcode = ERR17; + goto PCRE_EARLY_ERROR_RETURN; + } + +/* Check for global one-time settings at the start of the pattern, and remember +the offset for later. */ + +while (ptr[skipatstart] == CHAR_LEFT_PARENTHESIS && + ptr[skipatstart+1] == CHAR_ASTERISK) + { + int newnl = 0; + int newbsr = 0; + +#ifdef COMPILE_PCRE8 + if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 5) == 0) + { skipatstart += 7; options |= PCRE_UTF8; continue; } +#endif +#ifdef COMPILE_PCRE16 + if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UTF_RIGHTPAR, 6) == 0) + { skipatstart += 8; options |= PCRE_UTF16; continue; } +#endif + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_UCP_RIGHTPAR, 4) == 0) + { skipatstart += 6; options |= PCRE_UCP; continue; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_NO_START_OPT_RIGHTPAR, 13) == 0) + { skipatstart += 15; options |= PCRE_NO_START_OPTIMIZE; continue; } + + if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_CR_RIGHTPAR, 3) == 0) + { skipatstart += 5; newnl = PCRE_NEWLINE_CR; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_LF_RIGHTPAR, 3) == 0) + { skipatstart += 5; newnl = PCRE_NEWLINE_LF; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_CRLF_RIGHTPAR, 5) == 0) + { skipatstart += 7; newnl = PCRE_NEWLINE_CR + PCRE_NEWLINE_LF; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_ANY_RIGHTPAR, 4) == 0) + { skipatstart += 6; newnl = PCRE_NEWLINE_ANY; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_ANYCRLF_RIGHTPAR, 8) == 0) + { skipatstart += 10; newnl = PCRE_NEWLINE_ANYCRLF; } + + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_BSR_ANYCRLF_RIGHTPAR, 12) == 0) + { skipatstart += 14; newbsr = PCRE_BSR_ANYCRLF; } + else if (STRNCMP_UC_C8(ptr+skipatstart+2, STRING_BSR_UNICODE_RIGHTPAR, 12) == 0) + { skipatstart += 14; newbsr = PCRE_BSR_UNICODE; } + + if (newnl != 0) + options = (options & ~PCRE_NEWLINE_BITS) | newnl; + else if (newbsr != 0) + options = (options & ~(PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) | newbsr; + else break; + } + +/* PCRE_UTF16 has the same value as PCRE_UTF8. */ +utf = (options & PCRE_UTF8) != 0; + +/* Can't support UTF unless PCRE has been compiled to include the code. The +return of an error code from PRIV(valid_utf)() is a new feature, introduced in +release 8.13. It is passed back from pcre_[dfa_]exec(), but at the moment is +not used here. */ + +#ifdef SUPPORT_UTF +if (utf && (options & PCRE_NO_UTF8_CHECK) == 0 && + (errorcode = PRIV(valid_utf)((PCRE_PUCHAR)pattern, -1, erroroffset)) != 0) + { +#ifdef COMPILE_PCRE8 + errorcode = ERR44; +#else + errorcode = ERR74; +#endif + goto PCRE_EARLY_ERROR_RETURN2; + } +#else +if (utf) + { + errorcode = ERR32; + goto PCRE_EARLY_ERROR_RETURN; + } +#endif + +/* Can't support UCP unless PCRE has been compiled to include the code. */ + +#ifndef SUPPORT_UCP +if ((options & PCRE_UCP) != 0) + { + errorcode = ERR67; + goto PCRE_EARLY_ERROR_RETURN; + } +#endif + +/* Check validity of \R options. */ + +if ((options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) == + (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) + { + errorcode = ERR56; + goto PCRE_EARLY_ERROR_RETURN; + } + +/* Handle different types of newline. The three bits give seven cases. The +current code allows for fixed one- or two-byte sequences, plus "any" and +"anycrlf". */ + +switch (options & PCRE_NEWLINE_BITS) + { + case 0: newline = NEWLINE; break; /* Build-time default */ + case PCRE_NEWLINE_CR: newline = CHAR_CR; break; + case PCRE_NEWLINE_LF: newline = CHAR_NL; break; + case PCRE_NEWLINE_CR+ + PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break; + case PCRE_NEWLINE_ANY: newline = -1; break; + case PCRE_NEWLINE_ANYCRLF: newline = -2; break; + default: errorcode = ERR56; goto PCRE_EARLY_ERROR_RETURN; + } + +if (newline == -2) + { + cd->nltype = NLTYPE_ANYCRLF; + } +else if (newline < 0) + { + cd->nltype = NLTYPE_ANY; + } +else + { + cd->nltype = NLTYPE_FIXED; + if (newline > 255) + { + cd->nllen = 2; + cd->nl[0] = (newline >> 8) & 255; + cd->nl[1] = newline & 255; + } + else + { + cd->nllen = 1; + cd->nl[0] = newline; + } + } + +/* Maximum back reference and backref bitmap. The bitmap records up to 31 back +references to help in deciding whether (.*) can be treated as anchored or not. +*/ + +cd->top_backref = 0; +cd->backref_map = 0; + +/* Reflect pattern for debugging output */ + +DPRINTF(("------------------------------------------------------------------\n")); +#ifdef PCRE_DEBUG +print_puchar(stdout, (PCRE_PUCHAR)pattern); +#endif +DPRINTF(("\n")); + +/* Pretend to compile the pattern while actually just accumulating the length +of memory required. This behaviour is triggered by passing a non-NULL final +argument to compile_regex(). We pass a block of workspace (cworkspace) for it +to compile parts of the pattern into; the compiled code is discarded when it is +no longer needed, so hopefully this workspace will never overflow, though there +is a test for its doing so. */ + +cd->bracount = cd->final_bracount = 0; +cd->names_found = 0; +cd->name_entry_size = 0; +cd->name_table = NULL; +cd->start_code = cworkspace; +cd->hwm = cworkspace; +cd->start_workspace = cworkspace; +cd->workspace_size = COMPILE_WORK_SIZE; +cd->start_pattern = (const pcre_uchar *)pattern; +cd->end_pattern = (const pcre_uchar *)(pattern + STRLEN_UC((const pcre_uchar *)pattern)); +cd->req_varyopt = 0; +cd->assert_depth = 0; +cd->max_lookbehind = 0; +cd->external_options = options; +cd->external_flags = 0; +cd->open_caps = NULL; + +/* Now do the pre-compile. On error, errorcode will be set non-zero, so we +don't need to look at the result of the function here. The initial options have +been put into the cd block so that they can be changed if an option setting is +found within the regex right at the beginning. Bringing initial option settings +outside can help speed up starting point checks. */ + +ptr += skipatstart; +code = cworkspace; +*code = OP_BRA; +(void)compile_regex(cd->external_options, &code, &ptr, &errorcode, FALSE, + FALSE, 0, 0, &firstchar, &reqchar, NULL, cd, &length); +if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN; + +DPRINTF(("end pre-compile: length=%d workspace=%d\n", length, + (int)(cd->hwm - cworkspace))); + +if (length > MAX_PATTERN_SIZE) + { + errorcode = ERR20; + goto PCRE_EARLY_ERROR_RETURN; + } + +/* Compute the size of data block needed and get it, either from malloc or +externally provided function. Integer overflow should no longer be possible +because nowadays we limit the maximum value of cd->names_found and +cd->name_entry_size. */ + +size = sizeof(REAL_PCRE) + (length + cd->names_found * cd->name_entry_size) * sizeof(pcre_uchar); +re = (REAL_PCRE *)(PUBL(malloc))(size); + +if (re == NULL) + { + errorcode = ERR21; + goto PCRE_EARLY_ERROR_RETURN; + } + +/* Put in the magic number, and save the sizes, initial options, internal +flags, and character table pointer. NULL is used for the default character +tables. The nullpad field is at the end; it's there to help in the case when a +regex compiled on a system with 4-byte pointers is run on another with 8-byte +pointers. */ + +re->magic_number = MAGIC_NUMBER; +re->size = (int)size; +re->options = cd->external_options; +re->flags = cd->external_flags; +re->first_char = 0; +re->req_char = 0; +re->name_table_offset = sizeof(REAL_PCRE) / sizeof(pcre_uchar); +re->name_entry_size = cd->name_entry_size; +re->name_count = cd->names_found; +re->ref_count = 0; +re->tables = (tables == PRIV(default_tables))? NULL : tables; +re->nullpad = NULL; + +/* The starting points of the name/number translation table and of the code are +passed around in the compile data block. The start/end pattern and initial +options are already set from the pre-compile phase, as is the name_entry_size +field. Reset the bracket count and the names_found field. Also reset the hwm +field; this time it's used for remembering forward references to subpatterns. +*/ + +cd->final_bracount = cd->bracount; /* Save for checking forward references */ +cd->assert_depth = 0; +cd->bracount = 0; +cd->max_lookbehind = 0; +cd->names_found = 0; +cd->name_table = (pcre_uchar *)re + re->name_table_offset; +codestart = cd->name_table + re->name_entry_size * re->name_count; +cd->start_code = codestart; +cd->hwm = (pcre_uchar *)(cd->start_workspace); +cd->req_varyopt = 0; +cd->had_accept = FALSE; +cd->check_lookbehind = FALSE; +cd->open_caps = NULL; + +/* Set up a starting, non-extracting bracket, then compile the expression. On +error, errorcode will be set non-zero, so we don't need to look at the result +of the function here. */ + +ptr = (const pcre_uchar *)pattern + skipatstart; +code = (pcre_uchar *)codestart; +*code = OP_BRA; +(void)compile_regex(re->options, &code, &ptr, &errorcode, FALSE, FALSE, 0, 0, + &firstchar, &reqchar, NULL, cd, NULL); +re->top_bracket = cd->bracount; +re->top_backref = cd->top_backref; +re->max_lookbehind = cd->max_lookbehind; +re->flags = cd->external_flags | PCRE_MODE; + +if (cd->had_accept) reqchar = REQ_NONE; /* Must disable after (*ACCEPT) */ + +/* If not reached end of pattern on success, there's an excess bracket. */ + +if (errorcode == 0 && *ptr != 0) errorcode = ERR22; + +/* Fill in the terminating state and check for disastrous overflow, but +if debugging, leave the test till after things are printed out. */ + +*code++ = OP_END; + +#ifndef PCRE_DEBUG +if (code - codestart > length) errorcode = ERR23; +#endif + +/* Fill in any forward references that are required. There may be repeated +references; optimize for them, as searching a large regex takes time. */ + +if (cd->hwm > cd->start_workspace) + { + int prev_recno = -1; + const pcre_uchar *groupptr = NULL; + while (errorcode == 0 && cd->hwm > cd->start_workspace) + { + int offset, recno; + cd->hwm -= LINK_SIZE; + offset = GET(cd->hwm, 0); + recno = GET(codestart, offset); + if (recno != prev_recno) + { + groupptr = PRIV(find_bracket)(codestart, utf, recno); + prev_recno = recno; + } + if (groupptr == NULL) errorcode = ERR53; + else PUT(((pcre_uchar *)codestart), offset, (int)(groupptr - codestart)); + } + } + +/* If the workspace had to be expanded, free the new memory. */ + +if (cd->workspace_size > COMPILE_WORK_SIZE) + (PUBL(free))((void *)cd->start_workspace); + +/* Give an error if there's back reference to a non-existent capturing +subpattern. */ + +if (errorcode == 0 && re->top_backref > re->top_bracket) errorcode = ERR15; + +/* If there were any lookbehind assertions that contained OP_RECURSE +(recursions or subroutine calls), a flag is set for them to be checked here, +because they may contain forward references. Actual recursions can't be fixed +length, but subroutine calls can. It is done like this so that those without +OP_RECURSE that are not fixed length get a diagnosic with a useful offset. The +exceptional ones forgo this. We scan the pattern to check that they are fixed +length, and set their lengths. */ + +if (cd->check_lookbehind) + { + pcre_uchar *cc = (pcre_uchar *)codestart; + + /* Loop, searching for OP_REVERSE items, and process those that do not have + their length set. (Actually, it will also re-process any that have a length + of zero, but that is a pathological case, and it does no harm.) When we find + one, we temporarily terminate the branch it is in while we scan it. */ + + for (cc = (pcre_uchar *)PRIV(find_bracket)(codestart, utf, -1); + cc != NULL; + cc = (pcre_uchar *)PRIV(find_bracket)(cc, utf, -1)) + { + if (GET(cc, 1) == 0) + { + int fixed_length; + pcre_uchar *be = cc - 1 - LINK_SIZE + GET(cc, -LINK_SIZE); + int end_op = *be; + *be = OP_END; + fixed_length = find_fixedlength(cc, (re->options & PCRE_UTF8) != 0, TRUE, + cd); + *be = end_op; + DPRINTF(("fixed length = %d\n", fixed_length)); + if (fixed_length < 0) + { + errorcode = (fixed_length == -2)? ERR36 : + (fixed_length == -4)? ERR70 : ERR25; + break; + } + if (fixed_length > cd->max_lookbehind) cd->max_lookbehind = fixed_length; + PUT(cc, 1, fixed_length); + } + cc += 1 + LINK_SIZE; + } + } + +/* Failed to compile, or error while post-processing */ + +if (errorcode != 0) + { + (PUBL(free))(re); + PCRE_EARLY_ERROR_RETURN: + *erroroffset = (int)(ptr - (const pcre_uchar *)pattern); + PCRE_EARLY_ERROR_RETURN2: + *errorptr = find_error_text(errorcode); + if (errorcodeptr != NULL) *errorcodeptr = errorcode; + return NULL; + } + +/* If the anchored option was not passed, set the flag if we can determine that +the pattern is anchored by virtue of ^ characters or \A or anything else (such +as starting with .* when DOTALL is set). + +Otherwise, if we know what the first byte has to be, save it, because that +speeds up unanchored matches no end. If not, see if we can set the +PCRE_STARTLINE flag. This is helpful for multiline matches when all branches +start with ^. and also when all branches start with .* for non-DOTALL matches. +*/ + +if ((re->options & PCRE_ANCHORED) == 0) + { + if (is_anchored(codestart, 0, cd->backref_map)) + re->options |= PCRE_ANCHORED; + else + { + if (firstchar < 0) + firstchar = find_firstassertedchar(codestart, FALSE); + if (firstchar >= 0) /* Remove caseless flag for non-caseable chars */ + { +#ifdef COMPILE_PCRE8 + re->first_char = firstchar & 0xff; +#else +#ifdef COMPILE_PCRE16 + re->first_char = firstchar & 0xffff; +#endif +#endif + if ((firstchar & REQ_CASELESS) != 0) + { +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + /* We ignore non-ASCII first chars in 8 bit mode. */ + if (utf) + { + if (re->first_char < 128) + { + if (cd->fcc[re->first_char] != re->first_char) + re->flags |= PCRE_FCH_CASELESS; + } + else if (UCD_OTHERCASE(re->first_char) != re->first_char) + re->flags |= PCRE_FCH_CASELESS; + } + else +#endif + if (MAX_255(re->first_char) + && cd->fcc[re->first_char] != re->first_char) + re->flags |= PCRE_FCH_CASELESS; + } + + re->flags |= PCRE_FIRSTSET; + } + else if (is_startline(codestart, 0, cd->backref_map)) + re->flags |= PCRE_STARTLINE; + } + } + +/* For an anchored pattern, we use the "required byte" only if it follows a +variable length item in the regex. Remove the caseless flag for non-caseable +bytes. */ + +if (reqchar >= 0 && + ((re->options & PCRE_ANCHORED) == 0 || (reqchar & REQ_VARY) != 0)) + { +#ifdef COMPILE_PCRE8 + re->req_char = reqchar & 0xff; +#else +#ifdef COMPILE_PCRE16 + re->req_char = reqchar & 0xffff; +#endif +#endif + if ((reqchar & REQ_CASELESS) != 0) + { +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + /* We ignore non-ASCII first chars in 8 bit mode. */ + if (utf) + { + if (re->req_char < 128) + { + if (cd->fcc[re->req_char] != re->req_char) + re->flags |= PCRE_RCH_CASELESS; + } + else if (UCD_OTHERCASE(re->req_char) != re->req_char) + re->flags |= PCRE_RCH_CASELESS; + } + else +#endif + if (MAX_255(re->req_char) && cd->fcc[re->req_char] != re->req_char) + re->flags |= PCRE_RCH_CASELESS; + } + + re->flags |= PCRE_REQCHSET; + } + +/* Print out the compiled data if debugging is enabled. This is never the +case when building a production library. */ + +#ifdef PCRE_DEBUG +printf("Length = %d top_bracket = %d top_backref = %d\n", + length, re->top_bracket, re->top_backref); + +printf("Options=%08x\n", re->options); + +if ((re->flags & PCRE_FIRSTSET) != 0) + { + pcre_uchar ch = re->first_char; + const char *caseless = + ((re->flags & PCRE_FCH_CASELESS) == 0)? "" : " (caseless)"; + if (PRINTABLE(ch)) printf("First char = %c%s\n", ch, caseless); + else printf("First char = \\x%02x%s\n", ch, caseless); + } + +if ((re->flags & PCRE_REQCHSET) != 0) + { + pcre_uchar ch = re->req_char; + const char *caseless = + ((re->flags & PCRE_RCH_CASELESS) == 0)? "" : " (caseless)"; + if (PRINTABLE(ch)) printf("Req char = %c%s\n", ch, caseless); + else printf("Req char = \\x%02x%s\n", ch, caseless); + } + +#ifdef COMPILE_PCRE8 +pcre_printint((pcre *)re, stdout, TRUE); +#else +pcre16_printint((pcre *)re, stdout, TRUE); +#endif + +/* This check is done here in the debugging case so that the code that +was compiled can be seen. */ + +if (code - codestart > length) + { + (PUBL(free))(re); + *errorptr = find_error_text(ERR23); + *erroroffset = ptr - (pcre_uchar *)pattern; + if (errorcodeptr != NULL) *errorcodeptr = ERR23; + return NULL; + } +#endif /* PCRE_DEBUG */ + +#ifdef COMPILE_PCRE8 +return (pcre *)re; +#else +return (pcre16 *)re; +#endif } /* End of pcre_compile.c */ diff --git a/pcre/pcre_config.c b/pcre/pcre_config.c --- a/pcre/pcre_config.c +++ b/pcre/pcre_config.c @@ -6,7 +6,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -45,6 +45,9 @@ #include "config.h" #endif +/* Keep the original link size. */ +static int real_link_size = LINK_SIZE; + #include "pcre_internal.h" @@ -62,19 +65,42 @@ Returns: 0 if data returned, negative on error */ -PCRE_EXP_DEFN int +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION pcre_config(int what, void *where) +#else +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_config(int what, void *where) +#endif { switch (what) { case PCRE_CONFIG_UTF8: -#ifdef SUPPORT_UTF8 +#if defined COMPILE_PCRE16 + *((int *)where) = 0; + return PCRE_ERROR_BADOPTION; +#else +#if defined SUPPORT_UTF *((int *)where) = 1; #else *((int *)where) = 0; #endif break; +#endif + case PCRE_CONFIG_UTF16: +#if defined COMPILE_PCRE8 + *((int *)where) = 0; + return PCRE_ERROR_BADOPTION; +#else +#if defined SUPPORT_UTF + *((int *)where) = 1; +#else + *((int *)where) = 0; +#endif + break; +#endif + case PCRE_CONFIG_UNICODE_PROPERTIES: #ifdef SUPPORT_UCP *((int *)where) = 1; @@ -83,6 +109,22 @@ #endif break; + case PCRE_CONFIG_JIT: +#ifdef SUPPORT_JIT + *((int *)where) = 1; +#else + *((int *)where) = 0; +#endif + break; + + case PCRE_CONFIG_JITTARGET: +#ifdef SUPPORT_JIT + *((const char **)where) = PRIV(jit_get_target)(); +#else + *((const char **)where) = NULL; +#endif + break; + case PCRE_CONFIG_NEWLINE: *((int *)where) = NEWLINE; break; @@ -96,7 +138,7 @@ break; case PCRE_CONFIG_LINK_SIZE: - *((int *)where) = LINK_SIZE; + *((int *)where) = real_link_size; break; case PCRE_CONFIG_POSIX_MALLOC_THRESHOLD: @@ -104,11 +146,11 @@ break; case PCRE_CONFIG_MATCH_LIMIT: - *((unsigned int *)where) = MATCH_LIMIT; + *((unsigned long int *)where) = MATCH_LIMIT; break; case PCRE_CONFIG_MATCH_LIMIT_RECURSION: - *((unsigned int *)where) = MATCH_LIMIT_RECURSION; + *((unsigned long int *)where) = MATCH_LIMIT_RECURSION; break; case PCRE_CONFIG_STACKRECURSE: diff --git a/pcre/pcre_dfa_exec.c b/pcre/pcre_dfa_exec.c --- a/pcre/pcre_dfa_exec.c +++ b/pcre/pcre_dfa_exec.c @@ -3,2888 +3,3602 @@ *************************************************/ /* PCRE is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. - - Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* This module contains the external function pcre_dfa_exec(), which is an -alternative matching function that uses a sort of DFA algorithm (not a true -FSM). This is NOT Perl- compatible, but it has advantages in certain -applications. */ - - -#ifdef HAVE_CONFIG_H -#include "config.h" -#endif - -#define NLBLOCK md /* Block containing newline information */ -#define PSSTART start_subject /* Field containing processed string start */ -#define PSEND end_subject /* Field containing processed string end */ - -#include "pcre_internal.h" - - -/* For use to indent debugging output */ - -#define SP " " - - - -/************************************************* -* Code parameters and static tables * -*************************************************/ - -/* These are offsets that are used to turn the OP_TYPESTAR and friends opcodes -into others, under special conditions. A gap of 20 between the blocks should be -enough. The resulting opcodes don't have to be less than 256 because they are -never stored, so we push them well clear of the normal opcodes. */ - -#define OP_PROP_EXTRA 300 -#define OP_EXTUNI_EXTRA 320 -#define OP_ANYNL_EXTRA 340 -#define OP_HSPACE_EXTRA 360 -#define OP_VSPACE_EXTRA 380 - - -/* This table identifies those opcodes that are followed immediately by a -character that is to be tested in some way. This makes is possible to -centralize the loading of these characters. In the case of Type * etc, the -"character" is the opcode for \D, \d, \S, \s, \W, or \w, which will always be a -small value. ***NOTE*** If the start of this table is modified, the two tables -that follow must also be modified. */ - -static const uschar coptable[] = { - 0, /* End */ - 0, 0, 0, 0, 0, /* \A, \G, \K, \B, \b */ - 0, 0, 0, 0, 0, 0, /* \D, \d, \S, \s, \W, \w */ - 0, 0, 0, /* Any, AllAny, Anybyte */ - 0, 0, 0, /* NOTPROP, PROP, EXTUNI */ - 0, 0, 0, 0, 0, /* \R, \H, \h, \V, \v */ - 0, 0, 0, 0, 0, /* \Z, \z, Opt, ^, $ */ - 1, /* Char */ - 1, /* Charnc */ - 1, /* not */ - /* Positive single-char repeats */ - 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ - 3, 3, 3, /* upto, minupto, exact */ - 1, 1, 1, 3, /* *+, ++, ?+, upto+ */ - /* Negative single-char repeats - only for chars < 256 */ - 1, 1, 1, 1, 1, 1, /* NOT *, *?, +, +?, ?, ?? */ - 3, 3, 3, /* NOT upto, minupto, exact */ - 1, 1, 1, 3, /* NOT *+, ++, ?+, updo+ */ - /* Positive type repeats */ - 1, 1, 1, 1, 1, 1, /* Type *, *?, +, +?, ?, ?? */ - 3, 3, 3, /* Type upto, minupto, exact */ - 1, 1, 1, 3, /* Type *+, ++, ?+, upto+ */ - /* Character class & ref repeats */ - 0, 0, 0, 0, 0, 0, /* *, *?, +, +?, ?, ?? */ - 0, 0, /* CRRANGE, CRMINRANGE */ - 0, /* CLASS */ - 0, /* NCLASS */ - 0, /* XCLASS - variable length */ - 0, /* REF */ - 0, /* RECURSE */ - 0, /* CALLOUT */ - 0, /* Alt */ - 0, /* Ket */ - 0, /* KetRmax */ - 0, /* KetRmin */ - 0, /* Assert */ - 0, /* Assert not */ - 0, /* Assert behind */ - 0, /* Assert behind not */ - 0, /* Reverse */ - 0, 0, 0, 0, /* ONCE, BRA, CBRA, COND */ - 0, 0, 0, /* SBRA, SCBRA, SCOND */ - 0, /* CREF */ - 0, /* RREF */ - 0, /* DEF */ - 0, 0, /* BRAZERO, BRAMINZERO */ - 0, 0, 0, 0, /* PRUNE, SKIP, THEN, COMMIT */ - 0, 0, 0 /* FAIL, ACCEPT, SKIPZERO */ -}; - -/* These 2 tables allow for compact code for testing for \D, \d, \S, \s, \W, -and \w */ - -static const uschar toptable1[] = { - 0, 0, 0, 0, 0, 0, - ctype_digit, ctype_digit, - ctype_space, ctype_space, - ctype_word, ctype_word, - 0, 0 /* OP_ANY, OP_ALLANY */ -}; - -static const uschar toptable2[] = { - 0, 0, 0, 0, 0, 0, - ctype_digit, 0, - ctype_space, 0, - ctype_word, 0, - 1, 1 /* OP_ANY, OP_ALLANY */ -}; - - -/* Structure for holding data about a particular state, which is in effect the -current data for an active path through the match tree. It must consist -entirely of ints because the working vector we are passed, and which we put -these structures in, is a vector of ints. */ - -typedef struct stateblock { - int offset; /* Offset to opcode */ - int count; /* Count for repeats */ - int ims; /* ims flag bits */ - int data; /* Some use extra data */ -} stateblock; - -#define INTS_PER_STATEBLOCK (sizeof(stateblock)/sizeof(int)) - - -#ifdef DEBUG -/************************************************* -* Print character string * -*************************************************/ - -/* Character string printing function for debugging. - -Arguments: - p points to string - length number of bytes - f where to print - -Returns: nothing -*/ - -static void -pchars(unsigned char *p, int length, FILE *f) -{ -int c; -while (length-- > 0) - { - if (isprint(c = *(p++))) - fprintf(f, "%c", c); - else - fprintf(f, "\\x%02x", c); - } -} -#endif - - - -/************************************************* -* Execute a Regular Expression - DFA engine * -*************************************************/ - -/* This internal function applies a compiled pattern to a subject string, -starting at a given point, using a DFA engine. This function is called from the -external one, possibly multiple times if the pattern is not anchored. The -function calls itself recursively for some kinds of subpattern. - -Arguments: - md the match_data block with fixed information - this_start_code the opening bracket of this subexpression's code - current_subject where we currently are in the subject string - start_offset start offset in the subject string - offsets vector to contain the matching string offsets - offsetcount size of same - workspace vector of workspace - wscount size of same - ims the current ims flags - rlevel function call recursion level - recursing regex recursive call level - -Returns: > 0 => number of match offset pairs placed in offsets - = 0 => offsets overflowed; longest matches are present - -1 => failed to match - < -1 => some kind of unexpected problem - -The following macros are used for adding states to the two state vectors (one -for the current character, one for the following character). */ - -#define ADD_ACTIVE(x,y) \ - if (active_count++ < wscount) \ - { \ - next_active_state->offset = (x); \ - next_active_state->count = (y); \ - next_active_state->ims = ims; \ - next_active_state++; \ - DPRINTF(("%.*sADD_ACTIVE(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \ - } \ - else return PCRE_ERROR_DFA_WSSIZE - -#define ADD_ACTIVE_DATA(x,y,z) \ - if (active_count++ < wscount) \ - { \ - next_active_state->offset = (x); \ - next_active_state->count = (y); \ - next_active_state->ims = ims; \ - next_active_state->data = (z); \ - next_active_state++; \ - DPRINTF(("%.*sADD_ACTIVE_DATA(%d,%d,%d)\n", rlevel*2-2, SP, (x), (y), (z))); \ - } \ - else return PCRE_ERROR_DFA_WSSIZE - -#define ADD_NEW(x,y) \ - if (new_count++ < wscount) \ - { \ - next_new_state->offset = (x); \ - next_new_state->count = (y); \ - next_new_state->ims = ims; \ - next_new_state++; \ - DPRINTF(("%.*sADD_NEW(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \ - } \ - else return PCRE_ERROR_DFA_WSSIZE - -#define ADD_NEW_DATA(x,y,z) \ - if (new_count++ < wscount) \ - { \ - next_new_state->offset = (x); \ - next_new_state->count = (y); \ - next_new_state->ims = ims; \ - next_new_state->data = (z); \ - next_new_state++; \ - DPRINTF(("%.*sADD_NEW_DATA(%d,%d,%d)\n", rlevel*2-2, SP, (x), (y), (z))); \ - } \ - else return PCRE_ERROR_DFA_WSSIZE - -/* And now, here is the code */ - -static int -internal_dfa_exec( - dfa_match_data *md, - const uschar *this_start_code, - const uschar *current_subject, - int start_offset, - int *offsets, - int offsetcount, - int *workspace, - int wscount, - int ims, - int rlevel, - int recursing) -{ -stateblock *active_states, *new_states, *temp_states; -stateblock *next_active_state, *next_new_state; - -const uschar *ctypes, *lcc, *fcc; -const uschar *ptr; -const uschar *end_code, *first_op; - -int active_count, new_count, match_count; - -/* Some fields in the md block are frequently referenced, so we load them into -independent variables in the hope that this will perform better. */ - -const uschar *start_subject = md->start_subject; -const uschar *end_subject = md->end_subject; -const uschar *start_code = md->start_code; - -#ifdef SUPPORT_UTF8 -BOOL utf8 = (md->poptions & PCRE_UTF8) != 0; -#else -BOOL utf8 = FALSE; -#endif - -rlevel++; -offsetcount &= (-2); - -wscount -= 2; -wscount = (wscount - (wscount % (INTS_PER_STATEBLOCK * 2))) / - (2 * INTS_PER_STATEBLOCK); - -DPRINTF(("\n%.*s---------------------\n" - "%.*sCall to internal_dfa_exec f=%d r=%d\n", - rlevel*2-2, SP, rlevel*2-2, SP, rlevel, recursing)); - -ctypes = md->tables + ctypes_offset; -lcc = md->tables + lcc_offset; -fcc = md->tables + fcc_offset; - -match_count = PCRE_ERROR_NOMATCH; /* A negative number */ - -active_states = (stateblock *)(workspace + 2); -next_new_state = new_states = active_states + wscount; -new_count = 0; - -first_op = this_start_code + 1 + LINK_SIZE + - ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA)? 2:0); - -/* The first thing in any (sub) pattern is a bracket of some sort. Push all -the alternative states onto the list, and find out where the end is. This -makes is possible to use this function recursively, when we want to stop at a -matching internal ket rather than at the end. - -If the first opcode in the first alternative is OP_REVERSE, we are dealing with -a backward assertion. In that case, we have to find out the maximum amount to -move back, and set up each alternative appropriately. */ - -if (*first_op == OP_REVERSE) - { - int max_back = 0; - int gone_back; - - end_code = this_start_code; - do - { - int back = GET(end_code, 2+LINK_SIZE); - if (back > max_back) max_back = back; - end_code += GET(end_code, 1); - } - while (*end_code == OP_ALT); - - /* If we can't go back the amount required for the longest lookbehind - pattern, go back as far as we can; some alternatives may still be viable. */ - -#ifdef SUPPORT_UTF8 - /* In character mode we have to step back character by character */ - - if (utf8) - { - for (gone_back = 0; gone_back < max_back; gone_back++) - { - if (current_subject <= start_subject) break; - current_subject--; - while (current_subject > start_subject && - (*current_subject & 0xc0) == 0x80) - current_subject--; - } - } - else -#endif - - /* In byte-mode we can do this quickly. */ - - { - gone_back = (current_subject - max_back < start_subject)? - current_subject - start_subject : max_back; - current_subject -= gone_back; - } - - /* Now we can process the individual branches. */ - - end_code = this_start_code; - do - { - int back = GET(end_code, 2+LINK_SIZE); - if (back <= gone_back) - { - int bstate = end_code - start_code + 2 + 2*LINK_SIZE; - ADD_NEW_DATA(-bstate, 0, gone_back - back); - } - end_code += GET(end_code, 1); - } - while (*end_code == OP_ALT); - } - -/* This is the code for a "normal" subpattern (not a backward assertion). The -start of a whole pattern is always one of these. If we are at the top level, -we may be asked to restart matching from the same point that we reached for a -previous partial match. We still have to scan through the top-level branches to -find the end state. */ - -else - { - end_code = this_start_code; - - /* Restarting */ - - if (rlevel == 1 && (md->moptions & PCRE_DFA_RESTART) != 0) - { - do { end_code += GET(end_code, 1); } while (*end_code == OP_ALT); - new_count = workspace[1]; - if (!workspace[0]) - memcpy(new_states, active_states, new_count * sizeof(stateblock)); - } - - /* Not restarting */ - - else - { - int length = 1 + LINK_SIZE + - ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA)? 2:0); - do - { - ADD_NEW(end_code - start_code + length, 0); - end_code += GET(end_code, 1); - length = 1 + LINK_SIZE; - } - while (*end_code == OP_ALT); - } - } - -workspace[0] = 0; /* Bit indicating which vector is current */ - -DPRINTF(("%.*sEnd state = %d\n", rlevel*2-2, SP, end_code - start_code)); - -/* Loop for scanning the subject */ - -ptr = current_subject; -for (;;) - { - int i, j; - int clen, dlen; - unsigned int c, d; - - /* Make the new state list into the active state list and empty the - new state list. */ - - temp_states = active_states; - active_states = new_states; - new_states = temp_states; - active_count = new_count; - new_count = 0; - - workspace[0] ^= 1; /* Remember for the restarting feature */ - workspace[1] = active_count; - -#ifdef DEBUG - printf("%.*sNext character: rest of subject = \"", rlevel*2-2, SP); - pchars((uschar *)ptr, strlen((char *)ptr), stdout); - printf("\"\n"); - - printf("%.*sActive states: ", rlevel*2-2, SP); - for (i = 0; i < active_count; i++) - printf("%d/%d ", active_states[i].offset, active_states[i].count); - printf("\n"); -#endif - - /* Set the pointers for adding new states */ - - next_active_state = active_states + active_count; - next_new_state = new_states; - - /* Load the current character from the subject outside the loop, as many - different states may want to look at it, and we assume that at least one - will. */ - - if (ptr < end_subject) - { - clen = 1; /* Number of bytes in the character */ -#ifdef SUPPORT_UTF8 - if (utf8) { GETCHARLEN(c, ptr, clen); } else -#endif /* SUPPORT_UTF8 */ - c = *ptr; - } - else - { - clen = 0; /* This indicates the end of the subject */ - c = NOTACHAR; /* This value should never actually be used */ - } - - /* Scan up the active states and act on each one. The result of an action - may be to add more states to the currently active list (e.g. on hitting a - parenthesis) or it may be to put states on the new list, for considering - when we move the character pointer on. */ - - for (i = 0; i < active_count; i++) - { - stateblock *current_state = active_states + i; - const uschar *code; - int state_offset = current_state->offset; - int count, codevalue; -#ifdef SUPPORT_UCP - int chartype, script; -#endif - -#ifdef DEBUG - printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset); - if (clen == 0) printf("EOL\n"); - else if (c > 32 && c < 127) printf("'%c'\n", c); - else printf("0x%02x\n", c); -#endif - - /* This variable is referred to implicity in the ADD_xxx macros. */ - - ims = current_state->ims; - - /* A negative offset is a special case meaning "hold off going to this - (negated) state until the number of characters in the data field have - been skipped". */ - - if (state_offset < 0) - { - if (current_state->data > 0) - { - DPRINTF(("%.*sSkipping this character\n", rlevel*2-2, SP)); - ADD_NEW_DATA(state_offset, current_state->count, - current_state->data - 1); - continue; - } - else - { - current_state->offset = state_offset = -state_offset; - } - } - - /* Check for a duplicate state with the same count, and skip if found. */ - - for (j = 0; j < i; j++) - { - if (active_states[j].offset == state_offset && - active_states[j].count == current_state->count) - { - DPRINTF(("%.*sDuplicate state: skipped\n", rlevel*2-2, SP)); - goto NEXT_ACTIVE_STATE; - } - } - - /* The state offset is the offset to the opcode */ - - code = start_code + state_offset; - codevalue = *code; - - /* If this opcode is followed by an inline character, load it. It is - tempting to test for the presence of a subject character here, but that - is wrong, because sometimes zero repetitions of the subject are - permitted. - - We also use this mechanism for opcodes such as OP_TYPEPLUS that take an - argument that is not a data character - but is always one byte long. We - have to take special action to deal with \P, \p, \H, \h, \V, \v and \X in - this case. To keep the other cases fast, convert these ones to new opcodes. - */ - - if (coptable[codevalue] > 0) - { - dlen = 1; -#ifdef SUPPORT_UTF8 - if (utf8) { GETCHARLEN(d, (code + coptable[codevalue]), dlen); } else -#endif /* SUPPORT_UTF8 */ - d = code[coptable[codevalue]]; - if (codevalue >= OP_TYPESTAR) - { - switch(d) - { - case OP_ANYBYTE: return PCRE_ERROR_DFA_UITEM; - case OP_NOTPROP: - case OP_PROP: codevalue += OP_PROP_EXTRA; break; - case OP_ANYNL: codevalue += OP_ANYNL_EXTRA; break; - case OP_EXTUNI: codevalue += OP_EXTUNI_EXTRA; break; - case OP_NOT_HSPACE: - case OP_HSPACE: codevalue += OP_HSPACE_EXTRA; break; - case OP_NOT_VSPACE: - case OP_VSPACE: codevalue += OP_VSPACE_EXTRA; break; - default: break; - } - } - } - else - { - dlen = 0; /* Not strictly necessary, but compilers moan */ - d = NOTACHAR; /* if these variables are not set. */ - } - - - /* Now process the individual opcodes */ - - switch (codevalue) - { - -/* ========================================================================== */ - /* Reached a closing bracket. If not at the end of the pattern, carry - on with the next opcode. Otherwise, unless we have an empty string and - PCRE_NOTEMPTY is set, save the match data, shifting up all previous - matches so we always have the longest first. */ - - case OP_KET: - case OP_KETRMIN: - case OP_KETRMAX: - if (code != end_code) - { - ADD_ACTIVE(state_offset + 1 + LINK_SIZE, 0); - if (codevalue != OP_KET) - { - ADD_ACTIVE(state_offset - GET(code, 1), 0); - } - } - else if (ptr > current_subject || (md->moptions & PCRE_NOTEMPTY) == 0) - { - if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0; - else if (match_count > 0 && ++match_count * 2 >= offsetcount) - match_count = 0; - count = ((match_count == 0)? offsetcount : match_count * 2) - 2; - if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int)); - if (offsetcount >= 2) - { - offsets[0] = current_subject - start_subject; - offsets[1] = ptr - start_subject; - DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP, - offsets[1] - offsets[0], current_subject)); - } - if ((md->moptions & PCRE_DFA_SHORTEST) != 0) - { - DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n" - "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel, - match_count, rlevel*2-2, SP)); - return match_count; - } - } - break; - -/* ========================================================================== */ - /* These opcodes add to the current list of states without looking - at the current character. */ - - /*-----------------------------------------------------------------*/ - case OP_ALT: - do { code += GET(code, 1); } while (*code == OP_ALT); - ADD_ACTIVE(code - start_code, 0); - break; - - /*-----------------------------------------------------------------*/ - case OP_BRA: - case OP_SBRA: - do - { - ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0); - code += GET(code, 1); - } - while (*code == OP_ALT); - break; - - /*-----------------------------------------------------------------*/ - case OP_CBRA: - case OP_SCBRA: - ADD_ACTIVE(code - start_code + 3 + LINK_SIZE, 0); - code += GET(code, 1); - while (*code == OP_ALT) - { - ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0); - code += GET(code, 1); - } - break; - - /*-----------------------------------------------------------------*/ - case OP_BRAZERO: - case OP_BRAMINZERO: - ADD_ACTIVE(state_offset + 1, 0); - code += 1 + GET(code, 2); - while (*code == OP_ALT) code += GET(code, 1); - ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0); - break; - - /*-----------------------------------------------------------------*/ - case OP_SKIPZERO: - code += 1 + GET(code, 2); - while (*code == OP_ALT) code += GET(code, 1); - ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0); - break; - - /*-----------------------------------------------------------------*/ - case OP_CIRC: - if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) || - ((ims & PCRE_MULTILINE) != 0 && - ptr != end_subject && - WAS_NEWLINE(ptr))) - { ADD_ACTIVE(state_offset + 1, 0); } - break; - - /*-----------------------------------------------------------------*/ - case OP_EOD: - if (ptr >= end_subject) { ADD_ACTIVE(state_offset + 1, 0); } - break; - - /*-----------------------------------------------------------------*/ - case OP_OPT: - ims = code[1]; - ADD_ACTIVE(state_offset + 2, 0); - break; - - /*-----------------------------------------------------------------*/ - case OP_SOD: - if (ptr == start_subject) { ADD_ACTIVE(state_offset + 1, 0); } - break; - - /*-----------------------------------------------------------------*/ - case OP_SOM: - if (ptr == start_subject + start_offset) { ADD_ACTIVE(state_offset + 1, 0); } - break; - - -/* ========================================================================== */ - /* These opcodes inspect the next subject character, and sometimes - the previous one as well, but do not have an argument. The variable - clen contains the length of the current character and is zero if we are - at the end of the subject. */ - - /*-----------------------------------------------------------------*/ - case OP_ANY: - if (clen > 0 && !IS_NEWLINE(ptr)) - { ADD_NEW(state_offset + 1, 0); } - break; - - /*-----------------------------------------------------------------*/ - case OP_ALLANY: - if (clen > 0) - { ADD_NEW(state_offset + 1, 0); } - break; - - /*-----------------------------------------------------------------*/ - case OP_EODN: - if (clen == 0 || (IS_NEWLINE(ptr) && ptr == end_subject - md->nllen)) - { ADD_ACTIVE(state_offset + 1, 0); } - break; - - /*-----------------------------------------------------------------*/ - case OP_DOLL: - if ((md->moptions & PCRE_NOTEOL) == 0) - { - if (clen == 0 || - (IS_NEWLINE(ptr) && - ((ims & PCRE_MULTILINE) != 0 || ptr == end_subject - md->nllen) - )) - { ADD_ACTIVE(state_offset + 1, 0); } - } - else if ((ims & PCRE_MULTILINE) != 0 && IS_NEWLINE(ptr)) - { ADD_ACTIVE(state_offset + 1, 0); } - break; - - /*-----------------------------------------------------------------*/ - - case OP_DIGIT: - case OP_WHITESPACE: - case OP_WORDCHAR: - if (clen > 0 && c < 256 && - ((ctypes[c] & toptable1[codevalue]) ^ toptable2[codevalue]) != 0) - { ADD_NEW(state_offset + 1, 0); } - break; - - /*-----------------------------------------------------------------*/ - case OP_NOT_DIGIT: - case OP_NOT_WHITESPACE: - case OP_NOT_WORDCHAR: - if (clen > 0 && (c >= 256 || - ((ctypes[c] & toptable1[codevalue]) ^ toptable2[codevalue]) != 0)) - { ADD_NEW(state_offset + 1, 0); } - break; - - /*-----------------------------------------------------------------*/ - case OP_WORD_BOUNDARY: - case OP_NOT_WORD_BOUNDARY: - { - int left_word, right_word; - - if (ptr > start_subject) - { - const uschar *temp = ptr - 1; -#ifdef SUPPORT_UTF8 - if (utf8) BACKCHAR(temp); -#endif - GETCHARTEST(d, temp); - left_word = d < 256 && (ctypes[d] & ctype_word) != 0; - } - else left_word = 0; - - if (clen > 0) right_word = c < 256 && (ctypes[c] & ctype_word) != 0; - else right_word = 0; - - if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY)) - { ADD_ACTIVE(state_offset + 1, 0); } - } - break; - - - /*-----------------------------------------------------------------*/ - /* Check the next character by Unicode property. We will get here only - if the support is in the binary; otherwise a compile-time error occurs. - */ - -#ifdef SUPPORT_UCP - case OP_PROP: - case OP_NOTPROP: - if (clen > 0) - { - BOOL OK; - int category = _pcre_ucp_findprop(c, &chartype, &script); - switch(code[1]) - { - case PT_ANY: - OK = TRUE; - break; - - case PT_LAMP: - OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt; - break; - - case PT_GC: - OK = category == code[2]; - break; - - case PT_PC: - OK = chartype == code[2]; - break; - - case PT_SC: - OK = script == code[2]; - break; - - /* Should never occur, but keep compilers from grumbling. */ - - default: - OK = codevalue != OP_PROP; - break; - } - - if (OK == (codevalue == OP_PROP)) { ADD_NEW(state_offset + 3, 0); } - } - break; -#endif - - - -/* ========================================================================== */ - /* These opcodes likewise inspect the subject character, but have an - argument that is not a data character. It is one of these opcodes: - OP_ANY, OP_ALLANY, OP_DIGIT, OP_NOT_DIGIT, OP_WHITESPACE, OP_NOT_SPACE, - OP_WORDCHAR, OP_NOT_WORDCHAR. The value is loaded into d. */ - - case OP_TYPEPLUS: - case OP_TYPEMINPLUS: - case OP_TYPEPOSPLUS: - count = current_state->count; /* Already matched */ - if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } - if (clen > 0) - { - if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || - (c < 256 && - (d != OP_ANY || !IS_NEWLINE(ptr)) && - ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) - { - if (count > 0 && codevalue == OP_TYPEPOSPLUS) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - count++; - ADD_NEW(state_offset, count); - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_TYPEQUERY: - case OP_TYPEMINQUERY: - case OP_TYPEPOSQUERY: - ADD_ACTIVE(state_offset + 2, 0); - if (clen > 0) - { - if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || - (c < 256 && - (d != OP_ANY || !IS_NEWLINE(ptr)) && - ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) - { - if (codevalue == OP_TYPEPOSQUERY) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - ADD_NEW(state_offset + 2, 0); - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_TYPESTAR: - case OP_TYPEMINSTAR: - case OP_TYPEPOSSTAR: - ADD_ACTIVE(state_offset + 2, 0); - if (clen > 0) - { - if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || - (c < 256 && - (d != OP_ANY || !IS_NEWLINE(ptr)) && - ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) - { - if (codevalue == OP_TYPEPOSSTAR) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - ADD_NEW(state_offset, 0); - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_TYPEEXACT: - count = current_state->count; /* Number already matched */ - if (clen > 0) - { - if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || - (c < 256 && - (d != OP_ANY || !IS_NEWLINE(ptr)) && - ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) - { - if (++count >= GET2(code, 1)) - { ADD_NEW(state_offset + 4, 0); } - else - { ADD_NEW(state_offset, count); } - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_TYPEUPTO: - case OP_TYPEMINUPTO: - case OP_TYPEPOSUPTO: - ADD_ACTIVE(state_offset + 4, 0); - count = current_state->count; /* Number already matched */ - if (clen > 0) - { - if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || - (c < 256 && - (d != OP_ANY || !IS_NEWLINE(ptr)) && - ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) - { - if (codevalue == OP_TYPEPOSUPTO) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - if (++count >= GET2(code, 1)) - { ADD_NEW(state_offset + 4, 0); } - else - { ADD_NEW(state_offset, count); } - } - } - break; - -/* ========================================================================== */ - /* These are virtual opcodes that are used when something like - OP_TYPEPLUS has OP_PROP, OP_NOTPROP, OP_ANYNL, or OP_EXTUNI as its - argument. It keeps the code above fast for the other cases. The argument - is in the d variable. */ - -#ifdef SUPPORT_UCP - case OP_PROP_EXTRA + OP_TYPEPLUS: - case OP_PROP_EXTRA + OP_TYPEMINPLUS: - case OP_PROP_EXTRA + OP_TYPEPOSPLUS: - count = current_state->count; /* Already matched */ - if (count > 0) { ADD_ACTIVE(state_offset + 4, 0); } - if (clen > 0) - { - BOOL OK; - int category = _pcre_ucp_findprop(c, &chartype, &script); - switch(code[2]) - { - case PT_ANY: - OK = TRUE; - break; - - case PT_LAMP: - OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt; - break; - - case PT_GC: - OK = category == code[3]; - break; - - case PT_PC: - OK = chartype == code[3]; - break; - - case PT_SC: - OK = script == code[3]; - break; - - /* Should never occur, but keep compilers from grumbling. */ - - default: - OK = codevalue != OP_PROP; - break; - } - - if (OK == (d == OP_PROP)) - { - if (count > 0 && codevalue == OP_PROP_EXTRA + OP_TYPEPOSPLUS) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - count++; - ADD_NEW(state_offset, count); - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_EXTUNI_EXTRA + OP_TYPEPLUS: - case OP_EXTUNI_EXTRA + OP_TYPEMINPLUS: - case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS: - count = current_state->count; /* Already matched */ - if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } - if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M) - { - const uschar *nptr = ptr + clen; - int ncount = 0; - if (count > 0 && codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - while (nptr < end_subject) - { - int nd; - int ndlen = 1; - GETCHARLEN(nd, nptr, ndlen); - if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break; - ncount++; - nptr += ndlen; - } - count++; - ADD_NEW_DATA(-state_offset, count, ncount); - } - break; -#endif - - /*-----------------------------------------------------------------*/ - case OP_ANYNL_EXTRA + OP_TYPEPLUS: - case OP_ANYNL_EXTRA + OP_TYPEMINPLUS: - case OP_ANYNL_EXTRA + OP_TYPEPOSPLUS: - count = current_state->count; /* Already matched */ - if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } - if (clen > 0) - { - int ncount = 0; - switch (c) - { - case 0x000b: - case 0x000c: - case 0x0085: - case 0x2028: - case 0x2029: - if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; - goto ANYNL01; - - case 0x000d: - if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; - /* Fall through */ - - ANYNL01: - case 0x000a: - if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - count++; - ADD_NEW_DATA(-state_offset, count, ncount); - break; - - default: - break; - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_VSPACE_EXTRA + OP_TYPEPLUS: - case OP_VSPACE_EXTRA + OP_TYPEMINPLUS: - case OP_VSPACE_EXTRA + OP_TYPEPOSPLUS: - count = current_state->count; /* Already matched */ - if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } - if (clen > 0) - { - BOOL OK; - switch (c) - { - case 0x000a: - case 0x000b: - case 0x000c: - case 0x000d: - case 0x0085: - case 0x2028: - case 0x2029: - OK = TRUE; - break; - - default: - OK = FALSE; - break; - } - - if (OK == (d == OP_VSPACE)) - { - if (count > 0 && codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSPLUS) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - count++; - ADD_NEW_DATA(-state_offset, count, 0); - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_HSPACE_EXTRA + OP_TYPEPLUS: - case OP_HSPACE_EXTRA + OP_TYPEMINPLUS: - case OP_HSPACE_EXTRA + OP_TYPEPOSPLUS: - count = current_state->count; /* Already matched */ - if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } - if (clen > 0) - { - BOOL OK; - switch (c) - { - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - case 0x1680: /* OGHAM SPACE MARK */ - case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ - case 0x2000: /* EN QUAD */ - case 0x2001: /* EM QUAD */ - case 0x2002: /* EN SPACE */ - case 0x2003: /* EM SPACE */ - case 0x2004: /* THREE-PER-EM SPACE */ - case 0x2005: /* FOUR-PER-EM SPACE */ - case 0x2006: /* SIX-PER-EM SPACE */ - case 0x2007: /* FIGURE SPACE */ - case 0x2008: /* PUNCTUATION SPACE */ - case 0x2009: /* THIN SPACE */ - case 0x200A: /* HAIR SPACE */ - case 0x202f: /* NARROW NO-BREAK SPACE */ - case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ - case 0x3000: /* IDEOGRAPHIC SPACE */ - OK = TRUE; - break; - - default: - OK = FALSE; - break; - } - - if (OK == (d == OP_HSPACE)) - { - if (count > 0 && codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSPLUS) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - count++; - ADD_NEW_DATA(-state_offset, count, 0); - } - } - break; - - /*-----------------------------------------------------------------*/ -#ifdef SUPPORT_UCP - case OP_PROP_EXTRA + OP_TYPEQUERY: - case OP_PROP_EXTRA + OP_TYPEMINQUERY: - case OP_PROP_EXTRA + OP_TYPEPOSQUERY: - count = 4; - goto QS1; - - case OP_PROP_EXTRA + OP_TYPESTAR: - case OP_PROP_EXTRA + OP_TYPEMINSTAR: - case OP_PROP_EXTRA + OP_TYPEPOSSTAR: - count = 0; - - QS1: - - ADD_ACTIVE(state_offset + 4, 0); - if (clen > 0) - { - BOOL OK; - int category = _pcre_ucp_findprop(c, &chartype, &script); - switch(code[2]) - { - case PT_ANY: - OK = TRUE; - break; - - case PT_LAMP: - OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt; - break; - - case PT_GC: - OK = category == code[3]; - break; - - case PT_PC: - OK = chartype == code[3]; - break; - - case PT_SC: - OK = script == code[3]; - break; - - /* Should never occur, but keep compilers from grumbling. */ - - default: - OK = codevalue != OP_PROP; - break; - } - - if (OK == (d == OP_PROP)) - { - if (codevalue == OP_PROP_EXTRA + OP_TYPEPOSSTAR || - codevalue == OP_PROP_EXTRA + OP_TYPEPOSQUERY) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - ADD_NEW(state_offset + count, 0); - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_EXTUNI_EXTRA + OP_TYPEQUERY: - case OP_EXTUNI_EXTRA + OP_TYPEMINQUERY: - case OP_EXTUNI_EXTRA + OP_TYPEPOSQUERY: - count = 2; - goto QS2; - - case OP_EXTUNI_EXTRA + OP_TYPESTAR: - case OP_EXTUNI_EXTRA + OP_TYPEMINSTAR: - case OP_EXTUNI_EXTRA + OP_TYPEPOSSTAR: - count = 0; - - QS2: - - ADD_ACTIVE(state_offset + 2, 0); - if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M) - { - const uschar *nptr = ptr + clen; - int ncount = 0; - if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSSTAR || - codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSQUERY) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - while (nptr < end_subject) - { - int nd; - int ndlen = 1; - GETCHARLEN(nd, nptr, ndlen); - if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break; - ncount++; - nptr += ndlen; - } - ADD_NEW_DATA(-(state_offset + count), 0, ncount); - } - break; -#endif - - /*-----------------------------------------------------------------*/ - case OP_ANYNL_EXTRA + OP_TYPEQUERY: - case OP_ANYNL_EXTRA + OP_TYPEMINQUERY: - case OP_ANYNL_EXTRA + OP_TYPEPOSQUERY: - count = 2; - goto QS3; - - case OP_ANYNL_EXTRA + OP_TYPESTAR: - case OP_ANYNL_EXTRA + OP_TYPEMINSTAR: - case OP_ANYNL_EXTRA + OP_TYPEPOSSTAR: - count = 0; - - QS3: - ADD_ACTIVE(state_offset + 2, 0); - if (clen > 0) - { - int ncount = 0; - switch (c) - { - case 0x000b: - case 0x000c: - case 0x0085: - case 0x2028: - case 0x2029: - if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; - goto ANYNL02; - - case 0x000d: - if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; - /* Fall through */ - - ANYNL02: - case 0x000a: - if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR || - codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - ADD_NEW_DATA(-(state_offset + count), 0, ncount); - break; - - default: - break; - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_VSPACE_EXTRA + OP_TYPEQUERY: - case OP_VSPACE_EXTRA + OP_TYPEMINQUERY: - case OP_VSPACE_EXTRA + OP_TYPEPOSQUERY: - count = 2; - goto QS4; - - case OP_VSPACE_EXTRA + OP_TYPESTAR: - case OP_VSPACE_EXTRA + OP_TYPEMINSTAR: - case OP_VSPACE_EXTRA + OP_TYPEPOSSTAR: - count = 0; - - QS4: - ADD_ACTIVE(state_offset + 2, 0); - if (clen > 0) - { - BOOL OK; - switch (c) - { - case 0x000a: - case 0x000b: - case 0x000c: - case 0x000d: - case 0x0085: - case 0x2028: - case 0x2029: - OK = TRUE; - break; - - default: - OK = FALSE; - break; - } - if (OK == (d == OP_VSPACE)) - { - if (codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSSTAR || - codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSQUERY) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - ADD_NEW_DATA(-(state_offset + count), 0, 0); - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_HSPACE_EXTRA + OP_TYPEQUERY: - case OP_HSPACE_EXTRA + OP_TYPEMINQUERY: - case OP_HSPACE_EXTRA + OP_TYPEPOSQUERY: - count = 2; - goto QS5; - - case OP_HSPACE_EXTRA + OP_TYPESTAR: - case OP_HSPACE_EXTRA + OP_TYPEMINSTAR: - case OP_HSPACE_EXTRA + OP_TYPEPOSSTAR: - count = 0; - - QS5: - ADD_ACTIVE(state_offset + 2, 0); - if (clen > 0) - { - BOOL OK; - switch (c) - { - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - case 0x1680: /* OGHAM SPACE MARK */ - case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ - case 0x2000: /* EN QUAD */ - case 0x2001: /* EM QUAD */ - case 0x2002: /* EN SPACE */ - case 0x2003: /* EM SPACE */ - case 0x2004: /* THREE-PER-EM SPACE */ - case 0x2005: /* FOUR-PER-EM SPACE */ - case 0x2006: /* SIX-PER-EM SPACE */ - case 0x2007: /* FIGURE SPACE */ - case 0x2008: /* PUNCTUATION SPACE */ - case 0x2009: /* THIN SPACE */ - case 0x200A: /* HAIR SPACE */ - case 0x202f: /* NARROW NO-BREAK SPACE */ - case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ - case 0x3000: /* IDEOGRAPHIC SPACE */ - OK = TRUE; - break; - - default: - OK = FALSE; - break; - } - - if (OK == (d == OP_HSPACE)) - { - if (codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSSTAR || - codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSQUERY) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - ADD_NEW_DATA(-(state_offset + count), 0, 0); - } - } - break; - - /*-----------------------------------------------------------------*/ -#ifdef SUPPORT_UCP - case OP_PROP_EXTRA + OP_TYPEEXACT: - case OP_PROP_EXTRA + OP_TYPEUPTO: - case OP_PROP_EXTRA + OP_TYPEMINUPTO: - case OP_PROP_EXTRA + OP_TYPEPOSUPTO: - if (codevalue != OP_PROP_EXTRA + OP_TYPEEXACT) - { ADD_ACTIVE(state_offset + 6, 0); } - count = current_state->count; /* Number already matched */ - if (clen > 0) - { - BOOL OK; - int category = _pcre_ucp_findprop(c, &chartype, &script); - switch(code[4]) - { - case PT_ANY: - OK = TRUE; - break; - - case PT_LAMP: - OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt; - break; - - case PT_GC: - OK = category == code[5]; - break; - - case PT_PC: - OK = chartype == code[5]; - break; - - case PT_SC: - OK = script == code[5]; - break; - - /* Should never occur, but keep compilers from grumbling. */ - - default: - OK = codevalue != OP_PROP; - break; - } - - if (OK == (d == OP_PROP)) - { - if (codevalue == OP_PROP_EXTRA + OP_TYPEPOSUPTO) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - if (++count >= GET2(code, 1)) - { ADD_NEW(state_offset + 6, 0); } - else - { ADD_NEW(state_offset, count); } - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_EXTUNI_EXTRA + OP_TYPEEXACT: - case OP_EXTUNI_EXTRA + OP_TYPEUPTO: - case OP_EXTUNI_EXTRA + OP_TYPEMINUPTO: - case OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO: - if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT) - { ADD_ACTIVE(state_offset + 4, 0); } - count = current_state->count; /* Number already matched */ - if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M) - { - const uschar *nptr = ptr + clen; - int ncount = 0; - if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - while (nptr < end_subject) - { - int nd; - int ndlen = 1; - GETCHARLEN(nd, nptr, ndlen); - if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break; - ncount++; - nptr += ndlen; - } - if (++count >= GET2(code, 1)) - { ADD_NEW_DATA(-(state_offset + 4), 0, ncount); } - else - { ADD_NEW_DATA(-state_offset, count, ncount); } - } - break; -#endif - - /*-----------------------------------------------------------------*/ - case OP_ANYNL_EXTRA + OP_TYPEEXACT: - case OP_ANYNL_EXTRA + OP_TYPEUPTO: - case OP_ANYNL_EXTRA + OP_TYPEMINUPTO: - case OP_ANYNL_EXTRA + OP_TYPEPOSUPTO: - if (codevalue != OP_ANYNL_EXTRA + OP_TYPEEXACT) - { ADD_ACTIVE(state_offset + 4, 0); } - count = current_state->count; /* Number already matched */ - if (clen > 0) - { - int ncount = 0; - switch (c) - { - case 0x000b: - case 0x000c: - case 0x0085: - case 0x2028: - case 0x2029: - if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; - goto ANYNL03; - - case 0x000d: - if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; - /* Fall through */ - - ANYNL03: - case 0x000a: - if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - if (++count >= GET2(code, 1)) - { ADD_NEW_DATA(-(state_offset + 4), 0, ncount); } - else - { ADD_NEW_DATA(-state_offset, count, ncount); } - break; - - default: - break; - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_VSPACE_EXTRA + OP_TYPEEXACT: - case OP_VSPACE_EXTRA + OP_TYPEUPTO: - case OP_VSPACE_EXTRA + OP_TYPEMINUPTO: - case OP_VSPACE_EXTRA + OP_TYPEPOSUPTO: - if (codevalue != OP_VSPACE_EXTRA + OP_TYPEEXACT) - { ADD_ACTIVE(state_offset + 4, 0); } - count = current_state->count; /* Number already matched */ - if (clen > 0) - { - BOOL OK; - switch (c) - { - case 0x000a: - case 0x000b: - case 0x000c: - case 0x000d: - case 0x0085: - case 0x2028: - case 0x2029: - OK = TRUE; - break; - - default: - OK = FALSE; - } - - if (OK == (d == OP_VSPACE)) - { - if (codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSUPTO) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - if (++count >= GET2(code, 1)) - { ADD_NEW_DATA(-(state_offset + 4), 0, 0); } - else - { ADD_NEW_DATA(-state_offset, count, 0); } - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_HSPACE_EXTRA + OP_TYPEEXACT: - case OP_HSPACE_EXTRA + OP_TYPEUPTO: - case OP_HSPACE_EXTRA + OP_TYPEMINUPTO: - case OP_HSPACE_EXTRA + OP_TYPEPOSUPTO: - if (codevalue != OP_HSPACE_EXTRA + OP_TYPEEXACT) - { ADD_ACTIVE(state_offset + 4, 0); } - count = current_state->count; /* Number already matched */ - if (clen > 0) - { - BOOL OK; - switch (c) - { - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - case 0x1680: /* OGHAM SPACE MARK */ - case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ - case 0x2000: /* EN QUAD */ - case 0x2001: /* EM QUAD */ - case 0x2002: /* EN SPACE */ - case 0x2003: /* EM SPACE */ - case 0x2004: /* THREE-PER-EM SPACE */ - case 0x2005: /* FOUR-PER-EM SPACE */ - case 0x2006: /* SIX-PER-EM SPACE */ - case 0x2007: /* FIGURE SPACE */ - case 0x2008: /* PUNCTUATION SPACE */ - case 0x2009: /* THIN SPACE */ - case 0x200A: /* HAIR SPACE */ - case 0x202f: /* NARROW NO-BREAK SPACE */ - case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ - case 0x3000: /* IDEOGRAPHIC SPACE */ - OK = TRUE; - break; - - default: - OK = FALSE; - break; - } - - if (OK == (d == OP_HSPACE)) - { - if (codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSUPTO) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - if (++count >= GET2(code, 1)) - { ADD_NEW_DATA(-(state_offset + 4), 0, 0); } - else - { ADD_NEW_DATA(-state_offset, count, 0); } - } - } - break; - -/* ========================================================================== */ - /* These opcodes are followed by a character that is usually compared - to the current subject character; it is loaded into d. We still get - here even if there is no subject character, because in some cases zero - repetitions are permitted. */ - - /*-----------------------------------------------------------------*/ - case OP_CHAR: - if (clen > 0 && c == d) { ADD_NEW(state_offset + dlen + 1, 0); } - break; - - /*-----------------------------------------------------------------*/ - case OP_CHARNC: - if (clen == 0) break; - -#ifdef SUPPORT_UTF8 - if (utf8) - { - if (c == d) { ADD_NEW(state_offset + dlen + 1, 0); } else - { - unsigned int othercase; - if (c < 128) othercase = fcc[c]; else - - /* If we have Unicode property support, we can use it to test the - other case of the character. */ - -#ifdef SUPPORT_UCP - othercase = _pcre_ucp_othercase(c); -#else - othercase = NOTACHAR; -#endif - - if (d == othercase) { ADD_NEW(state_offset + dlen + 1, 0); } - } - } - else -#endif /* SUPPORT_UTF8 */ - - /* Non-UTF-8 mode */ - { - if (lcc[c] == lcc[d]) { ADD_NEW(state_offset + 2, 0); } - } - break; - - -#ifdef SUPPORT_UCP - /*-----------------------------------------------------------------*/ - /* This is a tricky one because it can match more than one character. - Find out how many characters to skip, and then set up a negative state - to wait for them to pass before continuing. */ - - case OP_EXTUNI: - if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M) - { - const uschar *nptr = ptr + clen; - int ncount = 0; - while (nptr < end_subject) - { - int nclen = 1; - GETCHARLEN(c, nptr, nclen); - if (_pcre_ucp_findprop(c, &chartype, &script) != ucp_M) break; - ncount++; - nptr += nclen; - } - ADD_NEW_DATA(-(state_offset + 1), 0, ncount); - } - break; -#endif - - /*-----------------------------------------------------------------*/ - /* This is a tricky like EXTUNI because it too can match more than one - character (when CR is followed by LF). In this case, set up a negative - state to wait for one character to pass before continuing. */ - - case OP_ANYNL: - if (clen > 0) switch(c) - { - case 0x000b: - case 0x000c: - case 0x0085: - case 0x2028: - case 0x2029: - if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; - - case 0x000a: - ADD_NEW(state_offset + 1, 0); - break; - - case 0x000d: - if (ptr + 1 < end_subject && ptr[1] == 0x0a) - { - ADD_NEW_DATA(-(state_offset + 1), 0, 1); - } - else - { - ADD_NEW(state_offset + 1, 0); - } - break; - } - break; - - /*-----------------------------------------------------------------*/ - case OP_NOT_VSPACE: - if (clen > 0) switch(c) - { - case 0x000a: - case 0x000b: - case 0x000c: - case 0x000d: - case 0x0085: - case 0x2028: - case 0x2029: - break; - - default: - ADD_NEW(state_offset + 1, 0); - break; - } - break; - - /*-----------------------------------------------------------------*/ - case OP_VSPACE: - if (clen > 0) switch(c) - { - case 0x000a: - case 0x000b: - case 0x000c: - case 0x000d: - case 0x0085: - case 0x2028: - case 0x2029: - ADD_NEW(state_offset + 1, 0); - break; - - default: break; - } - break; - - /*-----------------------------------------------------------------*/ - case OP_NOT_HSPACE: - if (clen > 0) switch(c) - { - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - case 0x1680: /* OGHAM SPACE MARK */ - case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ - case 0x2000: /* EN QUAD */ - case 0x2001: /* EM QUAD */ - case 0x2002: /* EN SPACE */ - case 0x2003: /* EM SPACE */ - case 0x2004: /* THREE-PER-EM SPACE */ - case 0x2005: /* FOUR-PER-EM SPACE */ - case 0x2006: /* SIX-PER-EM SPACE */ - case 0x2007: /* FIGURE SPACE */ - case 0x2008: /* PUNCTUATION SPACE */ - case 0x2009: /* THIN SPACE */ - case 0x200A: /* HAIR SPACE */ - case 0x202f: /* NARROW NO-BREAK SPACE */ - case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ - case 0x3000: /* IDEOGRAPHIC SPACE */ - break; - - default: - ADD_NEW(state_offset + 1, 0); - break; - } - break; - - /*-----------------------------------------------------------------*/ - case OP_HSPACE: - if (clen > 0) switch(c) - { - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - case 0x1680: /* OGHAM SPACE MARK */ - case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ - case 0x2000: /* EN QUAD */ - case 0x2001: /* EM QUAD */ - case 0x2002: /* EN SPACE */ - case 0x2003: /* EM SPACE */ - case 0x2004: /* THREE-PER-EM SPACE */ - case 0x2005: /* FOUR-PER-EM SPACE */ - case 0x2006: /* SIX-PER-EM SPACE */ - case 0x2007: /* FIGURE SPACE */ - case 0x2008: /* PUNCTUATION SPACE */ - case 0x2009: /* THIN SPACE */ - case 0x200A: /* HAIR SPACE */ - case 0x202f: /* NARROW NO-BREAK SPACE */ - case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ - case 0x3000: /* IDEOGRAPHIC SPACE */ - ADD_NEW(state_offset + 1, 0); - break; - } - break; - - /*-----------------------------------------------------------------*/ - /* Match a negated single character. This is only used for one-byte - characters, that is, we know that d < 256. The character we are - checking (c) can be multibyte. */ - - case OP_NOT: - if (clen > 0) - { - unsigned int otherd = ((ims & PCRE_CASELESS) != 0)? fcc[d] : d; - if (c != d && c != otherd) { ADD_NEW(state_offset + dlen + 1, 0); } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_PLUS: - case OP_MINPLUS: - case OP_POSPLUS: - case OP_NOTPLUS: - case OP_NOTMINPLUS: - case OP_NOTPOSPLUS: - count = current_state->count; /* Already matched */ - if (count > 0) { ADD_ACTIVE(state_offset + dlen + 1, 0); } - if (clen > 0) - { - unsigned int otherd = NOTACHAR; - if ((ims & PCRE_CASELESS) != 0) - { -#ifdef SUPPORT_UTF8 - if (utf8 && d >= 128) - { -#ifdef SUPPORT_UCP - otherd = _pcre_ucp_othercase(d); -#endif /* SUPPORT_UCP */ - } - else -#endif /* SUPPORT_UTF8 */ - otherd = fcc[d]; - } - if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) - { - if (count > 0 && - (codevalue == OP_POSPLUS || codevalue == OP_NOTPOSPLUS)) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - count++; - ADD_NEW(state_offset, count); - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_QUERY: - case OP_MINQUERY: - case OP_POSQUERY: - case OP_NOTQUERY: - case OP_NOTMINQUERY: - case OP_NOTPOSQUERY: - ADD_ACTIVE(state_offset + dlen + 1, 0); - if (clen > 0) - { - unsigned int otherd = NOTACHAR; - if ((ims & PCRE_CASELESS) != 0) - { -#ifdef SUPPORT_UTF8 - if (utf8 && d >= 128) - { -#ifdef SUPPORT_UCP - otherd = _pcre_ucp_othercase(d); -#endif /* SUPPORT_UCP */ - } - else -#endif /* SUPPORT_UTF8 */ - otherd = fcc[d]; - } - if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) - { - if (codevalue == OP_POSQUERY || codevalue == OP_NOTPOSQUERY) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - ADD_NEW(state_offset + dlen + 1, 0); - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_STAR: - case OP_MINSTAR: - case OP_POSSTAR: - case OP_NOTSTAR: - case OP_NOTMINSTAR: - case OP_NOTPOSSTAR: - ADD_ACTIVE(state_offset + dlen + 1, 0); - if (clen > 0) - { - unsigned int otherd = NOTACHAR; - if ((ims & PCRE_CASELESS) != 0) - { -#ifdef SUPPORT_UTF8 - if (utf8 && d >= 128) - { -#ifdef SUPPORT_UCP - otherd = _pcre_ucp_othercase(d); -#endif /* SUPPORT_UCP */ - } - else -#endif /* SUPPORT_UTF8 */ - otherd = fcc[d]; - } - if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) - { - if (codevalue == OP_POSSTAR || codevalue == OP_NOTPOSSTAR) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - ADD_NEW(state_offset, 0); - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_EXACT: - case OP_NOTEXACT: - count = current_state->count; /* Number already matched */ - if (clen > 0) - { - unsigned int otherd = NOTACHAR; - if ((ims & PCRE_CASELESS) != 0) - { -#ifdef SUPPORT_UTF8 - if (utf8 && d >= 128) - { -#ifdef SUPPORT_UCP - otherd = _pcre_ucp_othercase(d); -#endif /* SUPPORT_UCP */ - } - else -#endif /* SUPPORT_UTF8 */ - otherd = fcc[d]; - } - if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) - { - if (++count >= GET2(code, 1)) - { ADD_NEW(state_offset + dlen + 3, 0); } - else - { ADD_NEW(state_offset, count); } - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_UPTO: - case OP_MINUPTO: - case OP_POSUPTO: - case OP_NOTUPTO: - case OP_NOTMINUPTO: - case OP_NOTPOSUPTO: - ADD_ACTIVE(state_offset + dlen + 3, 0); - count = current_state->count; /* Number already matched */ - if (clen > 0) - { - unsigned int otherd = NOTACHAR; - if ((ims & PCRE_CASELESS) != 0) - { -#ifdef SUPPORT_UTF8 - if (utf8 && d >= 128) - { -#ifdef SUPPORT_UCP - otherd = _pcre_ucp_othercase(d); -#endif /* SUPPORT_UCP */ - } - else -#endif /* SUPPORT_UTF8 */ - otherd = fcc[d]; - } - if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) - { - if (codevalue == OP_POSUPTO || codevalue == OP_NOTPOSUPTO) - { - active_count--; /* Remove non-match possibility */ - next_active_state--; - } - if (++count >= GET2(code, 1)) - { ADD_NEW(state_offset + dlen + 3, 0); } - else - { ADD_NEW(state_offset, count); } - } - } - break; - - -/* ========================================================================== */ - /* These are the class-handling opcodes */ - - case OP_CLASS: - case OP_NCLASS: - case OP_XCLASS: - { - BOOL isinclass = FALSE; - int next_state_offset; - const uschar *ecode; - - /* For a simple class, there is always just a 32-byte table, and we - can set isinclass from it. */ - - if (codevalue != OP_XCLASS) - { - ecode = code + 33; - if (clen > 0) - { - isinclass = (c > 255)? (codevalue == OP_NCLASS) : - ((code[1 + c/8] & (1 << (c&7))) != 0); - } - } - - /* An extended class may have a table or a list of single characters, - ranges, or both, and it may be positive or negative. There's a - function that sorts all this out. */ - - else - { - ecode = code + GET(code, 1); - if (clen > 0) isinclass = _pcre_xclass(c, code + 1 + LINK_SIZE); - } - - /* At this point, isinclass is set for all kinds of class, and ecode - points to the byte after the end of the class. If there is a - quantifier, this is where it will be. */ - - next_state_offset = ecode - start_code; - - switch (*ecode) - { - case OP_CRSTAR: - case OP_CRMINSTAR: - ADD_ACTIVE(next_state_offset + 1, 0); - if (isinclass) { ADD_NEW(state_offset, 0); } - break; - - case OP_CRPLUS: - case OP_CRMINPLUS: - count = current_state->count; /* Already matched */ - if (count > 0) { ADD_ACTIVE(next_state_offset + 1, 0); } - if (isinclass) { count++; ADD_NEW(state_offset, count); } - break; - - case OP_CRQUERY: - case OP_CRMINQUERY: - ADD_ACTIVE(next_state_offset + 1, 0); - if (isinclass) { ADD_NEW(next_state_offset + 1, 0); } - break; - - case OP_CRRANGE: - case OP_CRMINRANGE: - count = current_state->count; /* Already matched */ - if (count >= GET2(ecode, 1)) - { ADD_ACTIVE(next_state_offset + 5, 0); } - if (isinclass) - { - int max = GET2(ecode, 3); - if (++count >= max && max != 0) /* Max 0 => no limit */ - { ADD_NEW(next_state_offset + 5, 0); } - else - { ADD_NEW(state_offset, count); } - } - break; - - default: - if (isinclass) { ADD_NEW(next_state_offset, 0); } - break; - } - } - break; - -/* ========================================================================== */ - /* These are the opcodes for fancy brackets of various kinds. We have - to use recursion in order to handle them. The "always failing" assersion - (?!) is optimised when compiling to OP_FAIL, so we have to support that, - though the other "backtracking verbs" are not supported. */ - - case OP_FAIL: - break; - - case OP_ASSERT: - case OP_ASSERT_NOT: - case OP_ASSERTBACK: - case OP_ASSERTBACK_NOT: - { - int rc; - int local_offsets[2]; - int local_workspace[1000]; - const uschar *endasscode = code + GET(code, 1); - - while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1); - - rc = internal_dfa_exec( - md, /* static match data */ - code, /* this subexpression's code */ - ptr, /* where we currently are */ - ptr - start_subject, /* start offset */ - local_offsets, /* offset vector */ - sizeof(local_offsets)/sizeof(int), /* size of same */ - local_workspace, /* workspace vector */ - sizeof(local_workspace)/sizeof(int), /* size of same */ - ims, /* the current ims flags */ - rlevel, /* function recursion level */ - recursing); /* pass on regex recursion */ - - if ((rc >= 0) == (codevalue == OP_ASSERT || codevalue == OP_ASSERTBACK)) - { ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_COND: - case OP_SCOND: - { - int local_offsets[1000]; - int local_workspace[1000]; - int condcode = code[LINK_SIZE+1]; - - /* Back reference conditions are not supported */ - - if (condcode == OP_CREF) return PCRE_ERROR_DFA_UCOND; - - /* The DEFINE condition is always false */ - - if (condcode == OP_DEF) - { - ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); - } - - /* The only supported version of OP_RREF is for the value RREF_ANY, - which means "test if in any recursion". We can't test for specifically - recursed groups. */ - - else if (condcode == OP_RREF) - { - int value = GET2(code, LINK_SIZE+2); - if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND; - if (recursing > 0) { ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); } - else { ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); } - } - - /* Otherwise, the condition is an assertion */ - - else - { - int rc; - const uschar *asscode = code + LINK_SIZE + 1; - const uschar *endasscode = asscode + GET(asscode, 1); - - while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1); - - rc = internal_dfa_exec( - md, /* fixed match data */ - asscode, /* this subexpression's code */ - ptr, /* where we currently are */ - ptr - start_subject, /* start offset */ - local_offsets, /* offset vector */ - sizeof(local_offsets)/sizeof(int), /* size of same */ - local_workspace, /* workspace vector */ - sizeof(local_workspace)/sizeof(int), /* size of same */ - ims, /* the current ims flags */ - rlevel, /* function recursion level */ - recursing); /* pass on regex recursion */ - - if ((rc >= 0) == - (condcode == OP_ASSERT || condcode == OP_ASSERTBACK)) - { ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); } - else - { ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); } - } - } - break; - - /*-----------------------------------------------------------------*/ - case OP_RECURSE: - { - int local_offsets[1000]; - int local_workspace[1000]; - int rc; - - DPRINTF(("%.*sStarting regex recursion %d\n", rlevel*2-2, SP, - recursing + 1)); - - rc = internal_dfa_exec( - md, /* fixed match data */ - start_code + GET(code, 1), /* this subexpression's code */ - ptr, /* where we currently are */ - ptr - start_subject, /* start offset */ - local_offsets, /* offset vector */ - sizeof(local_offsets)/sizeof(int), /* size of same */ - local_workspace, /* workspace vector */ - sizeof(local_workspace)/sizeof(int), /* size of same */ - ims, /* the current ims flags */ - rlevel, /* function recursion level */ - recursing + 1); /* regex recurse level */ - - DPRINTF(("%.*sReturn from regex recursion %d: rc=%d\n", rlevel*2-2, SP, - recursing + 1, rc)); - - /* Ran out of internal offsets */ - - if (rc == 0) return PCRE_ERROR_DFA_RECURSE; - - /* For each successful matched substring, set up the next state with a - count of characters to skip before trying it. Note that the count is in - characters, not bytes. */ - - if (rc > 0) - { - for (rc = rc*2 - 2; rc >= 0; rc -= 2) - { - const uschar *p = start_subject + local_offsets[rc]; - const uschar *pp = start_subject + local_offsets[rc+1]; - int charcount = local_offsets[rc+1] - local_offsets[rc]; - while (p < pp) if ((*p++ & 0xc0) == 0x80) charcount--; - if (charcount > 0) - { - ADD_NEW_DATA(-(state_offset + LINK_SIZE + 1), 0, (charcount - 1)); - } - else - { - ADD_ACTIVE(state_offset + LINK_SIZE + 1, 0); - } - } - } - else if (rc != PCRE_ERROR_NOMATCH) return rc; - } - break; - - /*-----------------------------------------------------------------*/ - case OP_ONCE: - { - int local_offsets[2]; - int local_workspace[1000]; - - int rc = internal_dfa_exec( - md, /* fixed match data */ - code, /* this subexpression's code */ - ptr, /* where we currently are */ - ptr - start_subject, /* start offset */ - local_offsets, /* offset vector */ - sizeof(local_offsets)/sizeof(int), /* size of same */ - local_workspace, /* workspace vector */ - sizeof(local_workspace)/sizeof(int), /* size of same */ - ims, /* the current ims flags */ - rlevel, /* function recursion level */ - recursing); /* pass on regex recursion */ - - if (rc >= 0) - { - const uschar *end_subpattern = code; - int charcount = local_offsets[1] - local_offsets[0]; - int next_state_offset, repeat_state_offset; - - do { end_subpattern += GET(end_subpattern, 1); } - while (*end_subpattern == OP_ALT); - next_state_offset = end_subpattern - start_code + LINK_SIZE + 1; - - /* If the end of this subpattern is KETRMAX or KETRMIN, we must - arrange for the repeat state also to be added to the relevant list. - Calculate the offset, or set -1 for no repeat. */ - - repeat_state_offset = (*end_subpattern == OP_KETRMAX || - *end_subpattern == OP_KETRMIN)? - end_subpattern - start_code - GET(end_subpattern, 1) : -1; - - /* If we have matched an empty string, add the next state at the - current character pointer. This is important so that the duplicate - checking kicks in, which is what breaks infinite loops that match an - empty string. */ - - if (charcount == 0) - { - ADD_ACTIVE(next_state_offset, 0); - } - - /* Optimization: if there are no more active states, and there - are no new states yet set up, then skip over the subject string - right here, to save looping. Otherwise, set up the new state to swing - into action when the end of the substring is reached. */ - - else if (i + 1 >= active_count && new_count == 0) - { - ptr += charcount; - clen = 0; - ADD_NEW(next_state_offset, 0); - - /* If we are adding a repeat state at the new character position, - we must fudge things so that it is the only current state. - Otherwise, it might be a duplicate of one we processed before, and - that would cause it to be skipped. */ - - if (repeat_state_offset >= 0) - { - next_active_state = active_states; - active_count = 0; - i = -1; - ADD_ACTIVE(repeat_state_offset, 0); - } - } - else - { - const uschar *p = start_subject + local_offsets[0]; - const uschar *pp = start_subject + local_offsets[1]; - while (p < pp) if ((*p++ & 0xc0) == 0x80) charcount--; - ADD_NEW_DATA(-next_state_offset, 0, (charcount - 1)); - if (repeat_state_offset >= 0) - { ADD_NEW_DATA(-repeat_state_offset, 0, (charcount - 1)); } - } - - } - else if (rc != PCRE_ERROR_NOMATCH) return rc; - } - break; - - -/* ========================================================================== */ - /* Handle callouts */ - - case OP_CALLOUT: - if (pcre_callout != NULL) - { - int rrc; - pcre_callout_block cb; - cb.version = 1; /* Version 1 of the callout block */ - cb.callout_number = code[1]; - cb.offset_vector = offsets; - cb.subject = (PCRE_SPTR)start_subject; - cb.subject_length = end_subject - start_subject; - cb.start_match = current_subject - start_subject; - cb.current_position = ptr - start_subject; - cb.pattern_position = GET(code, 2); - cb.next_item_length = GET(code, 2 + LINK_SIZE); - cb.capture_top = 1; - cb.capture_last = -1; - cb.callout_data = md->callout_data; - if ((rrc = (*pcre_callout)(&cb)) < 0) return rrc; /* Abandon */ - if (rrc == 0) { ADD_ACTIVE(state_offset + 2 + 2*LINK_SIZE, 0); } - } - break; - - -/* ========================================================================== */ - default: /* Unsupported opcode */ - return PCRE_ERROR_DFA_UITEM; - } - - NEXT_ACTIVE_STATE: continue; - - } /* End of loop scanning active states */ - - /* We have finished the processing at the current subject character. If no - new states have been set for the next character, we have found all the - matches that we are going to find. If we are at the top level and partial - matching has been requested, check for appropriate conditions. */ - - if (new_count <= 0) - { - if (match_count < 0 && /* No matches found */ - rlevel == 1 && /* Top level match function */ - (md->moptions & PCRE_PARTIAL) != 0 && /* Want partial matching */ - ptr >= end_subject && /* Reached end of subject */ - ptr > current_subject) /* Matched non-empty string */ - { - if (offsetcount >= 2) - { - offsets[0] = current_subject - start_subject; - offsets[1] = end_subject - start_subject; - } - match_count = PCRE_ERROR_PARTIAL; - } - - DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n" - "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel, match_count, - rlevel*2-2, SP)); - break; /* In effect, "return", but see the comment below */ - } - - /* One or more states are active for the next character. */ - - ptr += clen; /* Advance to next subject character */ - } /* Loop to move along the subject string */ - -/* Control gets here from "break" a few lines above. We do it this way because -if we use "return" above, we have compiler trouble. Some compilers warn if -there's nothing here because they think the function doesn't return a value. On -the other hand, if we put a dummy statement here, some more clever compilers -complain that it can't be reached. Sigh. */ - -return match_count; -} - - - - -/************************************************* -* Execute a Regular Expression - DFA engine * -*************************************************/ - -/* This external function applies a compiled re to a subject string using a DFA -engine. This function calls the internal function multiple times if the pattern -is not anchored. - -Arguments: - argument_re points to the compiled expression - extra_data points to extra data or is NULL - subject points to the subject string - length length of subject string (may contain binary zeros) - start_offset where to start in the subject string - options option bits - offsets vector of match offsets - offsetcount size of same - workspace workspace vector - wscount size of same - -Returns: > 0 => number of match offset pairs placed in offsets - = 0 => offsets overflowed; longest matches are present - -1 => failed to match - < -1 => some kind of unexpected problem -*/ - -PCRE_EXP_DEFN int -pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data, - const char *subject, int length, int start_offset, int options, int *offsets, - int offsetcount, int *workspace, int wscount) -{ -real_pcre *re = (real_pcre *)argument_re; -dfa_match_data match_block; -dfa_match_data *md = &match_block; -BOOL utf8, anchored, startline, firstline; -const uschar *current_subject, *end_subject, *lcc; - -pcre_study_data internal_study; -const pcre_study_data *study = NULL; -real_pcre internal_re; - -const uschar *req_byte_ptr; -const uschar *start_bits = NULL; -BOOL first_byte_caseless = FALSE; -BOOL req_byte_caseless = FALSE; -int first_byte = -1; -int req_byte = -1; -int req_byte2 = -1; -int newline; - -/* Plausibility checks */ - -if ((options & ~PUBLIC_DFA_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION; -if (re == NULL || subject == NULL || workspace == NULL || - (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL; -if (offsetcount < 0) return PCRE_ERROR_BADCOUNT; -if (wscount < 20) return PCRE_ERROR_DFA_WSSIZE; - -/* We need to find the pointer to any study data before we test for byte -flipping, so we scan the extra_data block first. This may set two fields in the -match block, so we must initialize them beforehand. However, the other fields -in the match block must not be set until after the byte flipping. */ - -md->tables = re->tables; -md->callout_data = NULL; - -if (extra_data != NULL) - { - unsigned int flags = extra_data->flags; - if ((flags & PCRE_EXTRA_STUDY_DATA) != 0) - study = (const pcre_study_data *)extra_data->study_data; - if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0) return PCRE_ERROR_DFA_UMLIMIT; - if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0) - return PCRE_ERROR_DFA_UMLIMIT; - if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0) - md->callout_data = extra_data->callout_data; - if ((flags & PCRE_EXTRA_TABLES) != 0) - md->tables = extra_data->tables; - } - -/* Check that the first field in the block is the magic number. If it is not, -test for a regex that was compiled on a host of opposite endianness. If this is -the case, flipped values are put in internal_re and internal_study if there was -study data too. */ - -if (re->magic_number != MAGIC_NUMBER) - { - re = _pcre_try_flipped(re, &internal_re, study, &internal_study); - if (re == NULL) return PCRE_ERROR_BADMAGIC; - if (study != NULL) study = &internal_study; - } - -/* Set some local values */ - -current_subject = (const unsigned char *)subject + start_offset; -end_subject = (const unsigned char *)subject + length; -req_byte_ptr = current_subject - 1; - -#ifdef SUPPORT_UTF8 -utf8 = (re->options & PCRE_UTF8) != 0; -#else -utf8 = FALSE; -#endif - -anchored = (options & (PCRE_ANCHORED|PCRE_DFA_RESTART)) != 0 || - (re->options & PCRE_ANCHORED) != 0; - -/* The remaining fixed data for passing around. */ - -md->start_code = (const uschar *)argument_re + - re->name_table_offset + re->name_count * re->name_entry_size; -md->start_subject = (const unsigned char *)subject; -md->end_subject = end_subject; -md->moptions = options; -md->poptions = re->options; - -/* If the BSR option is not set at match time, copy what was set -at compile time. */ - -if ((md->moptions & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) == 0) - { - if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0) - md->moptions |= re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE); -#ifdef BSR_ANYCRLF - else md->moptions |= PCRE_BSR_ANYCRLF; -#endif - } - -/* Handle different types of newline. The three bits give eight cases. If -nothing is set at run time, whatever was used at compile time applies. */ - -switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : (pcre_uint32)options) & - PCRE_NEWLINE_BITS) - { - case 0: newline = NEWLINE; break; /* Compile-time default */ - case PCRE_NEWLINE_CR: newline = '\r'; break; - case PCRE_NEWLINE_LF: newline = '\n'; break; - case PCRE_NEWLINE_CR+ - PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break; - case PCRE_NEWLINE_ANY: newline = -1; break; - case PCRE_NEWLINE_ANYCRLF: newline = -2; break; - default: return PCRE_ERROR_BADNEWLINE; - } - -if (newline == -2) - { - md->nltype = NLTYPE_ANYCRLF; - } -else if (newline < 0) - { - md->nltype = NLTYPE_ANY; - } -else - { - md->nltype = NLTYPE_FIXED; - if (newline > 255) - { - md->nllen = 2; - md->nl[0] = (newline >> 8) & 255; - md->nl[1] = newline & 255; - } - else - { - md->nllen = 1; - md->nl[0] = newline; - } - } - -/* Check a UTF-8 string if required. Unfortunately there's no way of passing -back the character offset. */ - -#ifdef SUPPORT_UTF8 -if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0) - { - if (_pcre_valid_utf8((uschar *)subject, length) >= 0) - return PCRE_ERROR_BADUTF8; - if (start_offset > 0 && start_offset < length) - { - int tb = ((uschar *)subject)[start_offset]; - if (tb > 127) - { - tb &= 0xc0; - if (tb != 0 && tb != 0xc0) return PCRE_ERROR_BADUTF8_OFFSET; - } - } - } -#endif - -/* If the exec call supplied NULL for tables, use the inbuilt ones. This -is a feature that makes it possible to save compiled regex and re-use them -in other programs later. */ - -if (md->tables == NULL) md->tables = _pcre_default_tables; - -/* The lower casing table and the "must be at the start of a line" flag are -used in a loop when finding where to start. */ - -lcc = md->tables + lcc_offset; -startline = (re->flags & PCRE_STARTLINE) != 0; -firstline = (re->options & PCRE_FIRSTLINE) != 0; - -/* Set up the first character to match, if available. The first_byte value is -never set for an anchored regular expression, but the anchoring may be forced -at run time, so we have to test for anchoring. The first char may be unset for -an unanchored pattern, of course. If there's no first char and the pattern was -studied, there may be a bitmap of possible first characters. */ - -if (!anchored) - { - if ((re->flags & PCRE_FIRSTSET) != 0) - { - first_byte = re->first_byte & 255; - if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE) - first_byte = lcc[first_byte]; - } - else - { - if (startline && study != NULL && - (study->options & PCRE_STUDY_MAPPED) != 0) - start_bits = study->start_bits; - } - } - -/* For anchored or unanchored matches, there may be a "last known required -character" set. */ - -if ((re->flags & PCRE_REQCHSET) != 0) - { - req_byte = re->req_byte & 255; - req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0; - req_byte2 = (md->tables + fcc_offset)[req_byte]; /* case flipped */ - } - -/* Call the main matching function, looping for a non-anchored regex after a -failed match. Unless restarting, optimize by moving to the first match -character if possible, when not anchored. Then unless wanting a partial match, -check for a required later character. */ - -for (;;) - { - int rc; - - if ((options & PCRE_DFA_RESTART) == 0) - { - const uschar *save_end_subject = end_subject; - - /* Advance to a unique first char if possible. If firstline is TRUE, the - start of the match is constrained to the first line of a multiline string. - Implement this by temporarily adjusting end_subject so that we stop - scanning at a newline. If the match fails at the newline, later code breaks - this loop. */ - - if (firstline) - { - const uschar *t = current_subject; - while (t < md->end_subject && !IS_NEWLINE(t)) t++; - end_subject = t; - } - - if (first_byte >= 0) - { - if (first_byte_caseless) - while (current_subject < end_subject && - lcc[*current_subject] != first_byte) - current_subject++; - else - while (current_subject < end_subject && *current_subject != first_byte) - current_subject++; - } - - /* Or to just after a linebreak for a multiline match if possible */ - - else if (startline) - { - if (current_subject > md->start_subject + start_offset) - { - while (current_subject <= end_subject && !WAS_NEWLINE(current_subject)) - current_subject++; - - /* If we have just passed a CR and the newline option is ANY or - ANYCRLF, and we are now at a LF, advance the match position by one more - character. */ - - if (current_subject[-1] == '\r' && - (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) && - current_subject < end_subject && - *current_subject == '\n') - current_subject++; - } - } - - /* Or to a non-unique first char after study */ - - else if (start_bits != NULL) - { - while (current_subject < end_subject) - { - register unsigned int c = *current_subject; - if ((start_bits[c/8] & (1 << (c&7))) == 0) current_subject++; - else break; - } - } - - /* Restore fudged end_subject */ - - end_subject = save_end_subject; - } - - /* If req_byte is set, we know that that character must appear in the subject - for the match to succeed. If the first character is set, req_byte must be - later in the subject; otherwise the test starts at the match point. This - optimization can save a huge amount of work in patterns with nested unlimited - repeats that aren't going to match. Writing separate code for cased/caseless - versions makes it go faster, as does using an autoincrement and backing off - on a match. - - HOWEVER: when the subject string is very, very long, searching to its end can - take a long time, and give bad performance on quite ordinary patterns. This - showed up when somebody was matching /^C/ on a 32-megabyte string... so we - don't do this when the string is sufficiently long. - - ALSO: this processing is disabled when partial matching is requested. - */ - - if (req_byte >= 0 && - end_subject - current_subject < REQ_BYTE_MAX && - (options & PCRE_PARTIAL) == 0) - { - register const uschar *p = current_subject + ((first_byte >= 0)? 1 : 0); - - /* We don't need to repeat the search if we haven't yet reached the - place we found it at last time. */ - - if (p > req_byte_ptr) - { - if (req_byte_caseless) - { - while (p < end_subject) - { - register int pp = *p++; - if (pp == req_byte || pp == req_byte2) { p--; break; } - } - } - else - { - while (p < end_subject) - { - if (*p++ == req_byte) { p--; break; } - } - } - - /* If we can't find the required character, break the matching loop, - which will cause a return or PCRE_ERROR_NOMATCH. */ - - if (p >= end_subject) break; - - /* If we have found the required character, save the point where we - found it, so that we don't search again next time round the loop if - the start hasn't passed this character yet. */ - - req_byte_ptr = p; - } - } - - /* OK, now we can do the business */ - - rc = internal_dfa_exec( - md, /* fixed match data */ - md->start_code, /* this subexpression's code */ - current_subject, /* where we currently are */ - start_offset, /* start offset in subject */ - offsets, /* offset vector */ - offsetcount, /* size of same */ - workspace, /* workspace vector */ - wscount, /* size of same */ - re->options & (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL), /* ims flags */ - 0, /* function recurse level */ - 0); /* regex recurse level */ - - /* Anything other than "no match" means we are done, always; otherwise, carry - on only if not anchored. */ - - if (rc != PCRE_ERROR_NOMATCH || anchored) return rc; - - /* Advance to the next subject character unless we are at the end of a line - and firstline is set. */ - - if (firstline && IS_NEWLINE(current_subject)) break; - current_subject++; - if (utf8) - { - while (current_subject < end_subject && (*current_subject & 0xc0) == 0x80) - current_subject++; - } - if (current_subject > end_subject) break; - - /* If we have just passed a CR and we are now at a LF, and the pattern does - not contain any explicit matches for \r or \n, and the newline option is CRLF - or ANY or ANYCRLF, advance the match position by one more character. */ - - if (current_subject[-1] == '\r' && - current_subject < end_subject && - *current_subject == '\n' && +and semantics are as close as possible to those of the Perl 5 language (but see +below for why this module is different). + + Written by Philip Hazel + Copyright (c) 1997-2012 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + +/* This module contains the external function pcre_dfa_exec(), which is an +alternative matching function that uses a sort of DFA algorithm (not a true +FSM). This is NOT Perl-compatible, but it has advantages in certain +applications. */ + + +/* NOTE ABOUT PERFORMANCE: A user of this function sent some code that improved +the performance of his patterns greatly. I could not use it as it stood, as it +was not thread safe, and made assumptions about pattern sizes. Also, it caused +test 7 to loop, and test 9 to crash with a segfault. + +The issue is the check for duplicate states, which is done by a simple linear +search up the state list. (Grep for "duplicate" below to find the code.) For +many patterns, there will never be many states active at one time, so a simple +linear search is fine. In patterns that have many active states, it might be a +bottleneck. The suggested code used an indexing scheme to remember which states +had previously been used for each character, and avoided the linear search when +it knew there was no chance of a duplicate. This was implemented when adding +states to the state lists. + +I wrote some thread-safe, not-limited code to try something similar at the time +of checking for duplicates (instead of when adding states), using index vectors +on the stack. It did give a 13% improvement with one specially constructed +pattern for certain subject strings, but on other strings and on many of the +simpler patterns in the test suite it did worse. The major problem, I think, +was the extra time to initialize the index. This had to be done for each call +of internal_dfa_exec(). (The supplied patch used a static vector, initialized +only once - I suspect this was the cause of the problems with the tests.) + +Overall, I concluded that the gains in some cases did not outweigh the losses +in others, so I abandoned this code. */ + + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#define NLBLOCK md /* Block containing newline information */ +#define PSSTART start_subject /* Field containing processed string start */ +#define PSEND end_subject /* Field containing processed string end */ + +#include "pcre_internal.h" + + +/* For use to indent debugging output */ + +#define SP " " + + +/************************************************* +* Code parameters and static tables * +*************************************************/ + +/* These are offsets that are used to turn the OP_TYPESTAR and friends opcodes +into others, under special conditions. A gap of 20 between the blocks should be +enough. The resulting opcodes don't have to be less than 256 because they are +never stored, so we push them well clear of the normal opcodes. */ + +#define OP_PROP_EXTRA 300 +#define OP_EXTUNI_EXTRA 320 +#define OP_ANYNL_EXTRA 340 +#define OP_HSPACE_EXTRA 360 +#define OP_VSPACE_EXTRA 380 + + +/* This table identifies those opcodes that are followed immediately by a +character that is to be tested in some way. This makes it possible to +centralize the loading of these characters. In the case of Type * etc, the +"character" is the opcode for \D, \d, \S, \s, \W, or \w, which will always be a +small value. Non-zero values in the table are the offsets from the opcode where +the character is to be found. ***NOTE*** If the start of this table is +modified, the three tables that follow must also be modified. */ + +static const pcre_uint8 coptable[] = { + 0, /* End */ + 0, 0, 0, 0, 0, /* \A, \G, \K, \B, \b */ + 0, 0, 0, 0, 0, 0, /* \D, \d, \S, \s, \W, \w */ + 0, 0, 0, /* Any, AllAny, Anybyte */ + 0, 0, /* \P, \p */ + 0, 0, 0, 0, 0, /* \R, \H, \h, \V, \v */ + 0, /* \X */ + 0, 0, 0, 0, 0, 0, /* \Z, \z, ^, ^M, $, $M */ + 1, /* Char */ + 1, /* Chari */ + 1, /* not */ + 1, /* noti */ + /* Positive single-char repeats */ + 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* upto, minupto */ + 1+IMM2_SIZE, /* exact */ + 1, 1, 1, 1+IMM2_SIZE, /* *+, ++, ?+, upto+ */ + 1, 1, 1, 1, 1, 1, /* *I, *?I, +I, +?I, ?I, ??I */ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* upto I, minupto I */ + 1+IMM2_SIZE, /* exact I */ + 1, 1, 1, 1+IMM2_SIZE, /* *+I, ++I, ?+I, upto+I */ + /* Negative single-char repeats - only for chars < 256 */ + 1, 1, 1, 1, 1, 1, /* NOT *, *?, +, +?, ?, ?? */ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* NOT upto, minupto */ + 1+IMM2_SIZE, /* NOT exact */ + 1, 1, 1, 1+IMM2_SIZE, /* NOT *+, ++, ?+, upto+ */ + 1, 1, 1, 1, 1, 1, /* NOT *I, *?I, +I, +?I, ?I, ??I */ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* NOT upto I, minupto I */ + 1+IMM2_SIZE, /* NOT exact I */ + 1, 1, 1, 1+IMM2_SIZE, /* NOT *+I, ++I, ?+I, upto+I */ + /* Positive type repeats */ + 1, 1, 1, 1, 1, 1, /* Type *, *?, +, +?, ?, ?? */ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* Type upto, minupto */ + 1+IMM2_SIZE, /* Type exact */ + 1, 1, 1, 1+IMM2_SIZE, /* Type *+, ++, ?+, upto+ */ + /* Character class & ref repeats */ + 0, 0, 0, 0, 0, 0, /* *, *?, +, +?, ?, ?? */ + 0, 0, /* CRRANGE, CRMINRANGE */ + 0, /* CLASS */ + 0, /* NCLASS */ + 0, /* XCLASS - variable length */ + 0, /* REF */ + 0, /* REFI */ + 0, /* RECURSE */ + 0, /* CALLOUT */ + 0, /* Alt */ + 0, /* Ket */ + 0, /* KetRmax */ + 0, /* KetRmin */ + 0, /* KetRpos */ + 0, /* Reverse */ + 0, /* Assert */ + 0, /* Assert not */ + 0, /* Assert behind */ + 0, /* Assert behind not */ + 0, 0, /* ONCE, ONCE_NC */ + 0, 0, 0, 0, 0, /* BRA, BRAPOS, CBRA, CBRAPOS, COND */ + 0, 0, 0, 0, 0, /* SBRA, SBRAPOS, SCBRA, SCBRAPOS, SCOND */ + 0, 0, /* CREF, NCREF */ + 0, 0, /* RREF, NRREF */ + 0, /* DEF */ + 0, 0, 0, /* BRAZERO, BRAMINZERO, BRAPOSZERO */ + 0, 0, 0, /* MARK, PRUNE, PRUNE_ARG */ + 0, 0, 0, 0, /* SKIP, SKIP_ARG, THEN, THEN_ARG */ + 0, 0, 0, 0, /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT */ + 0, 0 /* CLOSE, SKIPZERO */ +}; + +/* This table identifies those opcodes that inspect a character. It is used to +remember the fact that a character could have been inspected when the end of +the subject is reached. ***NOTE*** If the start of this table is modified, the +two tables that follow must also be modified. */ + +static const pcre_uint8 poptable[] = { + 0, /* End */ + 0, 0, 0, 1, 1, /* \A, \G, \K, \B, \b */ + 1, 1, 1, 1, 1, 1, /* \D, \d, \S, \s, \W, \w */ + 1, 1, 1, /* Any, AllAny, Anybyte */ + 1, 1, /* \P, \p */ + 1, 1, 1, 1, 1, /* \R, \H, \h, \V, \v */ + 1, /* \X */ + 0, 0, 0, 0, 0, 0, /* \Z, \z, ^, ^M, $, $M */ + 1, /* Char */ + 1, /* Chari */ + 1, /* not */ + 1, /* noti */ + /* Positive single-char repeats */ + 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ + 1, 1, 1, /* upto, minupto, exact */ + 1, 1, 1, 1, /* *+, ++, ?+, upto+ */ + 1, 1, 1, 1, 1, 1, /* *I, *?I, +I, +?I, ?I, ??I */ + 1, 1, 1, /* upto I, minupto I, exact I */ + 1, 1, 1, 1, /* *+I, ++I, ?+I, upto+I */ + /* Negative single-char repeats - only for chars < 256 */ + 1, 1, 1, 1, 1, 1, /* NOT *, *?, +, +?, ?, ?? */ + 1, 1, 1, /* NOT upto, minupto, exact */ + 1, 1, 1, 1, /* NOT *+, ++, ?+, upto+ */ + 1, 1, 1, 1, 1, 1, /* NOT *I, *?I, +I, +?I, ?I, ??I */ + 1, 1, 1, /* NOT upto I, minupto I, exact I */ + 1, 1, 1, 1, /* NOT *+I, ++I, ?+I, upto+I */ + /* Positive type repeats */ + 1, 1, 1, 1, 1, 1, /* Type *, *?, +, +?, ?, ?? */ + 1, 1, 1, /* Type upto, minupto, exact */ + 1, 1, 1, 1, /* Type *+, ++, ?+, upto+ */ + /* Character class & ref repeats */ + 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ + 1, 1, /* CRRANGE, CRMINRANGE */ + 1, /* CLASS */ + 1, /* NCLASS */ + 1, /* XCLASS - variable length */ + 0, /* REF */ + 0, /* REFI */ + 0, /* RECURSE */ + 0, /* CALLOUT */ + 0, /* Alt */ + 0, /* Ket */ + 0, /* KetRmax */ + 0, /* KetRmin */ + 0, /* KetRpos */ + 0, /* Reverse */ + 0, /* Assert */ + 0, /* Assert not */ + 0, /* Assert behind */ + 0, /* Assert behind not */ + 0, 0, /* ONCE, ONCE_NC */ + 0, 0, 0, 0, 0, /* BRA, BRAPOS, CBRA, CBRAPOS, COND */ + 0, 0, 0, 0, 0, /* SBRA, SBRAPOS, SCBRA, SCBRAPOS, SCOND */ + 0, 0, /* CREF, NCREF */ + 0, 0, /* RREF, NRREF */ + 0, /* DEF */ + 0, 0, 0, /* BRAZERO, BRAMINZERO, BRAPOSZERO */ + 0, 0, 0, /* MARK, PRUNE, PRUNE_ARG */ + 0, 0, 0, 0, /* SKIP, SKIP_ARG, THEN, THEN_ARG */ + 0, 0, 0, 0, /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT */ + 0, 0 /* CLOSE, SKIPZERO */ +}; + +/* These 2 tables allow for compact code for testing for \D, \d, \S, \s, \W, +and \w */ + +static const pcre_uint8 toptable1[] = { + 0, 0, 0, 0, 0, 0, + ctype_digit, ctype_digit, + ctype_space, ctype_space, + ctype_word, ctype_word, + 0, 0 /* OP_ANY, OP_ALLANY */ +}; + +static const pcre_uint8 toptable2[] = { + 0, 0, 0, 0, 0, 0, + ctype_digit, 0, + ctype_space, 0, + ctype_word, 0, + 1, 1 /* OP_ANY, OP_ALLANY */ +}; + + +/* Structure for holding data about a particular state, which is in effect the +current data for an active path through the match tree. It must consist +entirely of ints because the working vector we are passed, and which we put +these structures in, is a vector of ints. */ + +typedef struct stateblock { + int offset; /* Offset to opcode */ + int count; /* Count for repeats */ + int data; /* Some use extra data */ +} stateblock; + +#define INTS_PER_STATEBLOCK (int)(sizeof(stateblock)/sizeof(int)) + + +#ifdef PCRE_DEBUG +/************************************************* +* Print character string * +*************************************************/ + +/* Character string printing function for debugging. + +Arguments: + p points to string + length number of bytes + f where to print + +Returns: nothing +*/ + +static void +pchars(const pcre_uchar *p, int length, FILE *f) +{ +int c; +while (length-- > 0) + { + if (isprint(c = *(p++))) + fprintf(f, "%c", c); + else + fprintf(f, "\\x%02x", c); + } +} +#endif + + + +/************************************************* +* Execute a Regular Expression - DFA engine * +*************************************************/ + +/* This internal function applies a compiled pattern to a subject string, +starting at a given point, using a DFA engine. This function is called from the +external one, possibly multiple times if the pattern is not anchored. The +function calls itself recursively for some kinds of subpattern. + +Arguments: + md the match_data block with fixed information + this_start_code the opening bracket of this subexpression's code + current_subject where we currently are in the subject string + start_offset start offset in the subject string + offsets vector to contain the matching string offsets + offsetcount size of same + workspace vector of workspace + wscount size of same + rlevel function call recursion level + +Returns: > 0 => number of match offset pairs placed in offsets + = 0 => offsets overflowed; longest matches are present + -1 => failed to match + < -1 => some kind of unexpected problem + +The following macros are used for adding states to the two state vectors (one +for the current character, one for the following character). */ + +#define ADD_ACTIVE(x,y) \ + if (active_count++ < wscount) \ + { \ + next_active_state->offset = (x); \ + next_active_state->count = (y); \ + next_active_state++; \ + DPRINTF(("%.*sADD_ACTIVE(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \ + } \ + else return PCRE_ERROR_DFA_WSSIZE + +#define ADD_ACTIVE_DATA(x,y,z) \ + if (active_count++ < wscount) \ + { \ + next_active_state->offset = (x); \ + next_active_state->count = (y); \ + next_active_state->data = (z); \ + next_active_state++; \ + DPRINTF(("%.*sADD_ACTIVE_DATA(%d,%d,%d)\n", rlevel*2-2, SP, (x), (y), (z))); \ + } \ + else return PCRE_ERROR_DFA_WSSIZE + +#define ADD_NEW(x,y) \ + if (new_count++ < wscount) \ + { \ + next_new_state->offset = (x); \ + next_new_state->count = (y); \ + next_new_state++; \ + DPRINTF(("%.*sADD_NEW(%d,%d)\n", rlevel*2-2, SP, (x), (y))); \ + } \ + else return PCRE_ERROR_DFA_WSSIZE + +#define ADD_NEW_DATA(x,y,z) \ + if (new_count++ < wscount) \ + { \ + next_new_state->offset = (x); \ + next_new_state->count = (y); \ + next_new_state->data = (z); \ + next_new_state++; \ + DPRINTF(("%.*sADD_NEW_DATA(%d,%d,%d) line %d\n", rlevel*2-2, SP, \ + (x), (y), (z), __LINE__)); \ + } \ + else return PCRE_ERROR_DFA_WSSIZE + +/* And now, here is the code */ + +static int +internal_dfa_exec( + dfa_match_data *md, + const pcre_uchar *this_start_code, + const pcre_uchar *current_subject, + int start_offset, + int *offsets, + int offsetcount, + int *workspace, + int wscount, + int rlevel) +{ +stateblock *active_states, *new_states, *temp_states; +stateblock *next_active_state, *next_new_state; + +const pcre_uint8 *ctypes, *lcc, *fcc; +const pcre_uchar *ptr; +const pcre_uchar *end_code, *first_op; + +dfa_recursion_info new_recursive; + +int active_count, new_count, match_count; + +/* Some fields in the md block are frequently referenced, so we load them into +independent variables in the hope that this will perform better. */ + +const pcre_uchar *start_subject = md->start_subject; +const pcre_uchar *end_subject = md->end_subject; +const pcre_uchar *start_code = md->start_code; + +#ifdef SUPPORT_UTF +BOOL utf = (md->poptions & PCRE_UTF8) != 0; +#else +BOOL utf = FALSE; +#endif + +BOOL reset_could_continue = FALSE; + +rlevel++; +offsetcount &= (-2); + +wscount -= 2; +wscount = (wscount - (wscount % (INTS_PER_STATEBLOCK * 2))) / + (2 * INTS_PER_STATEBLOCK); + +DPRINTF(("\n%.*s---------------------\n" + "%.*sCall to internal_dfa_exec f=%d\n", + rlevel*2-2, SP, rlevel*2-2, SP, rlevel)); + +ctypes = md->tables + ctypes_offset; +lcc = md->tables + lcc_offset; +fcc = md->tables + fcc_offset; + +match_count = PCRE_ERROR_NOMATCH; /* A negative number */ + +active_states = (stateblock *)(workspace + 2); +next_new_state = new_states = active_states + wscount; +new_count = 0; + +first_op = this_start_code + 1 + LINK_SIZE + + ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA || + *this_start_code == OP_CBRAPOS || *this_start_code == OP_SCBRAPOS) + ? IMM2_SIZE:0); + +/* The first thing in any (sub) pattern is a bracket of some sort. Push all +the alternative states onto the list, and find out where the end is. This +makes is possible to use this function recursively, when we want to stop at a +matching internal ket rather than at the end. + +If the first opcode in the first alternative is OP_REVERSE, we are dealing with +a backward assertion. In that case, we have to find out the maximum amount to +move back, and set up each alternative appropriately. */ + +if (*first_op == OP_REVERSE) + { + int max_back = 0; + int gone_back; + + end_code = this_start_code; + do + { + int back = GET(end_code, 2+LINK_SIZE); + if (back > max_back) max_back = back; + end_code += GET(end_code, 1); + } + while (*end_code == OP_ALT); + + /* If we can't go back the amount required for the longest lookbehind + pattern, go back as far as we can; some alternatives may still be viable. */ + +#ifdef SUPPORT_UTF + /* In character mode we have to step back character by character */ + + if (utf) + { + for (gone_back = 0; gone_back < max_back; gone_back++) + { + if (current_subject <= start_subject) break; + current_subject--; + ACROSSCHAR(current_subject > start_subject, *current_subject, current_subject--); + } + } + else +#endif + + /* In byte-mode we can do this quickly. */ + + { + gone_back = (current_subject - max_back < start_subject)? + (int)(current_subject - start_subject) : max_back; + current_subject -= gone_back; + } + + /* Save the earliest consulted character */ + + if (current_subject < md->start_used_ptr) + md->start_used_ptr = current_subject; + + /* Now we can process the individual branches. */ + + end_code = this_start_code; + do + { + int back = GET(end_code, 2+LINK_SIZE); + if (back <= gone_back) + { + int bstate = (int)(end_code - start_code + 2 + 2*LINK_SIZE); + ADD_NEW_DATA(-bstate, 0, gone_back - back); + } + end_code += GET(end_code, 1); + } + while (*end_code == OP_ALT); + } + +/* This is the code for a "normal" subpattern (not a backward assertion). The +start of a whole pattern is always one of these. If we are at the top level, +we may be asked to restart matching from the same point that we reached for a +previous partial match. We still have to scan through the top-level branches to +find the end state. */ + +else + { + end_code = this_start_code; + + /* Restarting */ + + if (rlevel == 1 && (md->moptions & PCRE_DFA_RESTART) != 0) + { + do { end_code += GET(end_code, 1); } while (*end_code == OP_ALT); + new_count = workspace[1]; + if (!workspace[0]) + memcpy(new_states, active_states, new_count * sizeof(stateblock)); + } + + /* Not restarting */ + + else + { + int length = 1 + LINK_SIZE + + ((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA || + *this_start_code == OP_CBRAPOS || *this_start_code == OP_SCBRAPOS) + ? IMM2_SIZE:0); + do + { + ADD_NEW((int)(end_code - start_code + length), 0); + end_code += GET(end_code, 1); + length = 1 + LINK_SIZE; + } + while (*end_code == OP_ALT); + } + } + +workspace[0] = 0; /* Bit indicating which vector is current */ + +DPRINTF(("%.*sEnd state = %d\n", rlevel*2-2, SP, (int)(end_code - start_code))); + +/* Loop for scanning the subject */ + +ptr = current_subject; +for (;;) + { + int i, j; + int clen, dlen; + unsigned int c, d; + int forced_fail = 0; + BOOL partial_newline = FALSE; + BOOL could_continue = reset_could_continue; + reset_could_continue = FALSE; + + /* Make the new state list into the active state list and empty the + new state list. */ + + temp_states = active_states; + active_states = new_states; + new_states = temp_states; + active_count = new_count; + new_count = 0; + + workspace[0] ^= 1; /* Remember for the restarting feature */ + workspace[1] = active_count; + +#ifdef PCRE_DEBUG + printf("%.*sNext character: rest of subject = \"", rlevel*2-2, SP); + pchars(ptr, STRLEN_UC(ptr), stdout); + printf("\"\n"); + + printf("%.*sActive states: ", rlevel*2-2, SP); + for (i = 0; i < active_count; i++) + printf("%d/%d ", active_states[i].offset, active_states[i].count); + printf("\n"); +#endif + + /* Set the pointers for adding new states */ + + next_active_state = active_states + active_count; + next_new_state = new_states; + + /* Load the current character from the subject outside the loop, as many + different states may want to look at it, and we assume that at least one + will. */ + + if (ptr < end_subject) + { + clen = 1; /* Number of data items in the character */ +#ifdef SUPPORT_UTF + if (utf) { GETCHARLEN(c, ptr, clen); } else +#endif /* SUPPORT_UTF */ + c = *ptr; + } + else + { + clen = 0; /* This indicates the end of the subject */ + c = NOTACHAR; /* This value should never actually be used */ + } + + /* Scan up the active states and act on each one. The result of an action + may be to add more states to the currently active list (e.g. on hitting a + parenthesis) or it may be to put states on the new list, for considering + when we move the character pointer on. */ + + for (i = 0; i < active_count; i++) + { + stateblock *current_state = active_states + i; + BOOL caseless = FALSE; + const pcre_uchar *code; + int state_offset = current_state->offset; + int count, codevalue, rrc; + +#ifdef PCRE_DEBUG + printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset); + if (clen == 0) printf("EOL\n"); + else if (c > 32 && c < 127) printf("'%c'\n", c); + else printf("0x%02x\n", c); +#endif + + /* A negative offset is a special case meaning "hold off going to this + (negated) state until the number of characters in the data field have + been skipped". If the could_continue flag was passed over from a previous + state, arrange for it to passed on. */ + + if (state_offset < 0) + { + if (current_state->data > 0) + { + DPRINTF(("%.*sSkipping this character\n", rlevel*2-2, SP)); + ADD_NEW_DATA(state_offset, current_state->count, + current_state->data - 1); + if (could_continue) reset_could_continue = TRUE; + continue; + } + else + { + current_state->offset = state_offset = -state_offset; + } + } + + /* Check for a duplicate state with the same count, and skip if found. + See the note at the head of this module about the possibility of improving + performance here. */ + + for (j = 0; j < i; j++) + { + if (active_states[j].offset == state_offset && + active_states[j].count == current_state->count) + { + DPRINTF(("%.*sDuplicate state: skipped\n", rlevel*2-2, SP)); + goto NEXT_ACTIVE_STATE; + } + } + + /* The state offset is the offset to the opcode */ + + code = start_code + state_offset; + codevalue = *code; + + /* If this opcode inspects a character, but we are at the end of the + subject, remember the fact for use when testing for a partial match. */ + + if (clen == 0 && poptable[codevalue] != 0) + could_continue = TRUE; + + /* If this opcode is followed by an inline character, load it. It is + tempting to test for the presence of a subject character here, but that + is wrong, because sometimes zero repetitions of the subject are + permitted. + + We also use this mechanism for opcodes such as OP_TYPEPLUS that take an + argument that is not a data character - but is always one byte long because + the values are small. We have to take special action to deal with \P, \p, + \H, \h, \V, \v and \X in this case. To keep the other cases fast, convert + these ones to new opcodes. */ + + if (coptable[codevalue] > 0) + { + dlen = 1; +#ifdef SUPPORT_UTF + if (utf) { GETCHARLEN(d, (code + coptable[codevalue]), dlen); } else +#endif /* SUPPORT_UTF */ + d = code[coptable[codevalue]]; + if (codevalue >= OP_TYPESTAR) + { + switch(d) + { + case OP_ANYBYTE: return PCRE_ERROR_DFA_UITEM; + case OP_NOTPROP: + case OP_PROP: codevalue += OP_PROP_EXTRA; break; + case OP_ANYNL: codevalue += OP_ANYNL_EXTRA; break; + case OP_EXTUNI: codevalue += OP_EXTUNI_EXTRA; break; + case OP_NOT_HSPACE: + case OP_HSPACE: codevalue += OP_HSPACE_EXTRA; break; + case OP_NOT_VSPACE: + case OP_VSPACE: codevalue += OP_VSPACE_EXTRA; break; + default: break; + } + } + } + else + { + dlen = 0; /* Not strictly necessary, but compilers moan */ + d = NOTACHAR; /* if these variables are not set. */ + } + + + /* Now process the individual opcodes */ + + switch (codevalue) + { +/* ========================================================================== */ + /* These cases are never obeyed. This is a fudge that causes a compile- + time error if the vectors coptable or poptable, which are indexed by + opcode, are not the correct length. It seems to be the only way to do + such a check at compile time, as the sizeof() operator does not work + in the C preprocessor. */ + + case OP_TABLE_LENGTH: + case OP_TABLE_LENGTH + + ((sizeof(coptable) == OP_TABLE_LENGTH) && + (sizeof(poptable) == OP_TABLE_LENGTH)): + break; + +/* ========================================================================== */ + /* Reached a closing bracket. If not at the end of the pattern, carry + on with the next opcode. For repeating opcodes, also add the repeat + state. Note that KETRPOS will always be encountered at the end of the + subpattern, because the possessive subpattern repeats are always handled + using recursive calls. Thus, it never adds any new states. + + At the end of the (sub)pattern, unless we have an empty string and + PCRE_NOTEMPTY is set, or PCRE_NOTEMPTY_ATSTART is set and we are at the + start of the subject, save the match data, shifting up all previous + matches so we always have the longest first. */ + + case OP_KET: + case OP_KETRMIN: + case OP_KETRMAX: + case OP_KETRPOS: + if (code != end_code) + { + ADD_ACTIVE(state_offset + 1 + LINK_SIZE, 0); + if (codevalue != OP_KET) + { + ADD_ACTIVE(state_offset - GET(code, 1), 0); + } + } + else + { + if (ptr > current_subject || + ((md->moptions & PCRE_NOTEMPTY) == 0 && + ((md->moptions & PCRE_NOTEMPTY_ATSTART) == 0 || + current_subject > start_subject + md->start_offset))) + { + if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0; + else if (match_count > 0 && ++match_count * 2 > offsetcount) + match_count = 0; + count = ((match_count == 0)? offsetcount : match_count * 2) - 2; + if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int)); + if (offsetcount >= 2) + { + offsets[0] = (int)(current_subject - start_subject); + offsets[1] = (int)(ptr - start_subject); + DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP, + offsets[1] - offsets[0], (char *)current_subject)); + } + if ((md->moptions & PCRE_DFA_SHORTEST) != 0) + { + DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n" + "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel, + match_count, rlevel*2-2, SP)); + return match_count; + } + } + } + break; + +/* ========================================================================== */ + /* These opcodes add to the current list of states without looking + at the current character. */ + + /*-----------------------------------------------------------------*/ + case OP_ALT: + do { code += GET(code, 1); } while (*code == OP_ALT); + ADD_ACTIVE((int)(code - start_code), 0); + break; + + /*-----------------------------------------------------------------*/ + case OP_BRA: + case OP_SBRA: + do + { + ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0); + code += GET(code, 1); + } + while (*code == OP_ALT); + break; + + /*-----------------------------------------------------------------*/ + case OP_CBRA: + case OP_SCBRA: + ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE + IMM2_SIZE), 0); + code += GET(code, 1); + while (*code == OP_ALT) + { + ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0); + code += GET(code, 1); + } + break; + + /*-----------------------------------------------------------------*/ + case OP_BRAZERO: + case OP_BRAMINZERO: + ADD_ACTIVE(state_offset + 1, 0); + code += 1 + GET(code, 2); + while (*code == OP_ALT) code += GET(code, 1); + ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0); + break; + + /*-----------------------------------------------------------------*/ + case OP_SKIPZERO: + code += 1 + GET(code, 2); + while (*code == OP_ALT) code += GET(code, 1); + ADD_ACTIVE((int)(code - start_code + 1 + LINK_SIZE), 0); + break; + + /*-----------------------------------------------------------------*/ + case OP_CIRC: + if (ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) + { ADD_ACTIVE(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_CIRCM: + if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) || + (ptr != end_subject && WAS_NEWLINE(ptr))) + { ADD_ACTIVE(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_EOD: + if (ptr >= end_subject) + { + if ((md->moptions & PCRE_PARTIAL_HARD) != 0) + could_continue = TRUE; + else { ADD_ACTIVE(state_offset + 1, 0); } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_SOD: + if (ptr == start_subject) { ADD_ACTIVE(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_SOM: + if (ptr == start_subject + start_offset) { ADD_ACTIVE(state_offset + 1, 0); } + break; + + +/* ========================================================================== */ + /* These opcodes inspect the next subject character, and sometimes + the previous one as well, but do not have an argument. The variable + clen contains the length of the current character and is zero if we are + at the end of the subject. */ + + /*-----------------------------------------------------------------*/ + case OP_ANY: + if (clen > 0 && !IS_NEWLINE(ptr)) + { + if (ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + could_continue = partial_newline = TRUE; + } + else + { + ADD_NEW(state_offset + 1, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_ALLANY: + if (clen > 0) + { ADD_NEW(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_EODN: + if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0) + could_continue = TRUE; + else if (clen == 0 || (IS_NEWLINE(ptr) && ptr == end_subject - md->nllen)) + { ADD_ACTIVE(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_DOLL: + if ((md->moptions & PCRE_NOTEOL) == 0) + { + if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0) + could_continue = TRUE; + else if (clen == 0 || + ((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr) && + (ptr == end_subject - md->nllen) + )) + { ADD_ACTIVE(state_offset + 1, 0); } + else if (ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + if ((md->moptions & PCRE_PARTIAL_HARD) != 0) + { + reset_could_continue = TRUE; + ADD_NEW_DATA(-(state_offset + 1), 0, 1); + } + else could_continue = partial_newline = TRUE; + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_DOLLM: + if ((md->moptions & PCRE_NOTEOL) == 0) + { + if (clen == 0 && (md->moptions & PCRE_PARTIAL_HARD) != 0) + could_continue = TRUE; + else if (clen == 0 || + ((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr))) + { ADD_ACTIVE(state_offset + 1, 0); } + else if (ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + if ((md->moptions & PCRE_PARTIAL_HARD) != 0) + { + reset_could_continue = TRUE; + ADD_NEW_DATA(-(state_offset + 1), 0, 1); + } + else could_continue = partial_newline = TRUE; + } + } + else if (IS_NEWLINE(ptr)) + { ADD_ACTIVE(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + + case OP_DIGIT: + case OP_WHITESPACE: + case OP_WORDCHAR: + if (clen > 0 && c < 256 && + ((ctypes[c] & toptable1[codevalue]) ^ toptable2[codevalue]) != 0) + { ADD_NEW(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_NOT_DIGIT: + case OP_NOT_WHITESPACE: + case OP_NOT_WORDCHAR: + if (clen > 0 && (c >= 256 || + ((ctypes[c] & toptable1[codevalue]) ^ toptable2[codevalue]) != 0)) + { ADD_NEW(state_offset + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_WORD_BOUNDARY: + case OP_NOT_WORD_BOUNDARY: + { + int left_word, right_word; + + if (ptr > start_subject) + { + const pcre_uchar *temp = ptr - 1; + if (temp < md->start_used_ptr) md->start_used_ptr = temp; +#ifdef SUPPORT_UTF + if (utf) { BACKCHAR(temp); } +#endif + GETCHARTEST(d, temp); +#ifdef SUPPORT_UCP + if ((md->poptions & PCRE_UCP) != 0) + { + if (d == '_') left_word = TRUE; else + { + int cat = UCD_CATEGORY(d); + left_word = (cat == ucp_L || cat == ucp_N); + } + } + else +#endif + left_word = d < 256 && (ctypes[d] & ctype_word) != 0; + } + else left_word = FALSE; + + if (clen > 0) + { +#ifdef SUPPORT_UCP + if ((md->poptions & PCRE_UCP) != 0) + { + if (c == '_') right_word = TRUE; else + { + int cat = UCD_CATEGORY(c); + right_word = (cat == ucp_L || cat == ucp_N); + } + } + else +#endif + right_word = c < 256 && (ctypes[c] & ctype_word) != 0; + } + else right_word = FALSE; + + if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY)) + { ADD_ACTIVE(state_offset + 1, 0); } + } + break; + + + /*-----------------------------------------------------------------*/ + /* Check the next character by Unicode property. We will get here only + if the support is in the binary; otherwise a compile-time error occurs. + */ + +#ifdef SUPPORT_UCP + case OP_PROP: + case OP_NOTPROP: + if (clen > 0) + { + BOOL OK; + const ucd_record * prop = GET_UCD(c); + switch(code[1]) + { + case PT_ANY: + OK = TRUE; + break; + + case PT_LAMP: + OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt; + break; + + case PT_GC: + OK = PRIV(ucp_gentype)[prop->chartype] == code[2]; + break; + + case PT_PC: + OK = prop->chartype == code[2]; + break; + + case PT_SC: + OK = prop->script == code[2]; + break; + + /* These are specials for combination cases. */ + + case PT_ALNUM: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N; + break; + + case PT_SPACE: /* Perl space */ + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR; + break; + + case PT_PXSPACE: /* POSIX space */ + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_VT || + c == CHAR_FF || c == CHAR_CR; + break; + + case PT_WORD: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || + c == CHAR_UNDERSCORE; + break; + + /* Should never occur, but keep compilers from grumbling. */ + + default: + OK = codevalue != OP_PROP; + break; + } + + if (OK == (codevalue == OP_PROP)) { ADD_NEW(state_offset + 3, 0); } + } + break; +#endif + + + +/* ========================================================================== */ + /* These opcodes likewise inspect the subject character, but have an + argument that is not a data character. It is one of these opcodes: + OP_ANY, OP_ALLANY, OP_DIGIT, OP_NOT_DIGIT, OP_WHITESPACE, OP_NOT_SPACE, + OP_WORDCHAR, OP_NOT_WORDCHAR. The value is loaded into d. */ + + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } + if (clen > 0) + { + if (d == OP_ANY && ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + could_continue = partial_newline = TRUE; + } + else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || + (c < 256 && + (d != OP_ANY || !IS_NEWLINE(ptr)) && + ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) + { + if (count > 0 && codevalue == OP_TYPEPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW(state_offset, count); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSQUERY: + ADD_ACTIVE(state_offset + 2, 0); + if (clen > 0) + { + if (d == OP_ANY && ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + could_continue = partial_newline = TRUE; + } + else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || + (c < 256 && + (d != OP_ANY || !IS_NEWLINE(ptr)) && + ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) + { + if (codevalue == OP_TYPEPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW(state_offset + 2, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPOSSTAR: + ADD_ACTIVE(state_offset + 2, 0); + if (clen > 0) + { + if (d == OP_ANY && ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + could_continue = partial_newline = TRUE; + } + else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || + (c < 256 && + (d != OP_ANY || !IS_NEWLINE(ptr)) && + ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) + { + if (codevalue == OP_TYPEPOSSTAR) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW(state_offset, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_TYPEEXACT: + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + if (d == OP_ANY && ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + could_continue = partial_newline = TRUE; + } + else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || + (c < 256 && + (d != OP_ANY || !IS_NEWLINE(ptr)) && + ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) + { + if (++count >= GET2(code, 1)) + { ADD_NEW(state_offset + 1 + IMM2_SIZE + 1, 0); } + else + { ADD_NEW(state_offset, count); } + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEPOSUPTO: + ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + if (d == OP_ANY && ptr + 1 >= md->end_subject && + (md->moptions & (PCRE_PARTIAL_HARD)) != 0 && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + could_continue = partial_newline = TRUE; + } + else if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || + (c < 256 && + (d != OP_ANY || !IS_NEWLINE(ptr)) && + ((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) + { + if (codevalue == OP_TYPEPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= GET2(code, 1)) + { ADD_NEW(state_offset + 2 + IMM2_SIZE, 0); } + else + { ADD_NEW(state_offset, count); } + } + } + break; + +/* ========================================================================== */ + /* These are virtual opcodes that are used when something like + OP_TYPEPLUS has OP_PROP, OP_NOTPROP, OP_ANYNL, or OP_EXTUNI as its + argument. It keeps the code above fast for the other cases. The argument + is in the d variable. */ + +#ifdef SUPPORT_UCP + case OP_PROP_EXTRA + OP_TYPEPLUS: + case OP_PROP_EXTRA + OP_TYPEMINPLUS: + case OP_PROP_EXTRA + OP_TYPEPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + 4, 0); } + if (clen > 0) + { + BOOL OK; + const ucd_record * prop = GET_UCD(c); + switch(code[2]) + { + case PT_ANY: + OK = TRUE; + break; + + case PT_LAMP: + OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt; + break; + + case PT_GC: + OK = PRIV(ucp_gentype)[prop->chartype] == code[3]; + break; + + case PT_PC: + OK = prop->chartype == code[3]; + break; + + case PT_SC: + OK = prop->script == code[3]; + break; + + /* These are specials for combination cases. */ + + case PT_ALNUM: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N; + break; + + case PT_SPACE: /* Perl space */ + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR; + break; + + case PT_PXSPACE: /* POSIX space */ + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_VT || + c == CHAR_FF || c == CHAR_CR; + break; + + case PT_WORD: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || + c == CHAR_UNDERSCORE; + break; + + /* Should never occur, but keep compilers from grumbling. */ + + default: + OK = codevalue != OP_PROP; + break; + } + + if (OK == (d == OP_PROP)) + { + if (count > 0 && codevalue == OP_PROP_EXTRA + OP_TYPEPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW(state_offset, count); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_EXTUNI_EXTRA + OP_TYPEPLUS: + case OP_EXTUNI_EXTRA + OP_TYPEMINPLUS: + case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } + if (clen > 0 && UCD_CATEGORY(c) != ucp_M) + { + const pcre_uchar *nptr = ptr + clen; + int ncount = 0; + if (count > 0 && codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + while (nptr < end_subject) + { + int nd; + int ndlen = 1; + GETCHARLEN(nd, nptr, ndlen); + if (UCD_CATEGORY(nd) != ucp_M) break; + ncount++; + nptr += ndlen; + } + count++; + ADD_NEW_DATA(-state_offset, count, ncount); + } + break; +#endif + + /*-----------------------------------------------------------------*/ + case OP_ANYNL_EXTRA + OP_TYPEPLUS: + case OP_ANYNL_EXTRA + OP_TYPEMINPLUS: + case OP_ANYNL_EXTRA + OP_TYPEPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } + if (clen > 0) + { + int ncount = 0; + switch (c) + { + case 0x000b: + case 0x000c: + case 0x0085: + case 0x2028: + case 0x2029: + if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; + goto ANYNL01; + + case 0x000d: + if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; + /* Fall through */ + + ANYNL01: + case 0x000a: + if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW_DATA(-state_offset, count, ncount); + break; + + default: + break; + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_VSPACE_EXTRA + OP_TYPEPLUS: + case OP_VSPACE_EXTRA + OP_TYPEMINPLUS: + case OP_VSPACE_EXTRA + OP_TYPEPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } + if (clen > 0) + { + BOOL OK; + switch (c) + { + case 0x000a: + case 0x000b: + case 0x000c: + case 0x000d: + case 0x0085: + case 0x2028: + case 0x2029: + OK = TRUE; + break; + + default: + OK = FALSE; + break; + } + + if (OK == (d == OP_VSPACE)) + { + if (count > 0 && codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW_DATA(-state_offset, count, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_HSPACE_EXTRA + OP_TYPEPLUS: + case OP_HSPACE_EXTRA + OP_TYPEMINPLUS: + case OP_HSPACE_EXTRA + OP_TYPEPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } + if (clen > 0) + { + BOOL OK; + switch (c) + { + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ + OK = TRUE; + break; + + default: + OK = FALSE; + break; + } + + if (OK == (d == OP_HSPACE)) + { + if (count > 0 && codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSPLUS) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW_DATA(-state_offset, count, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ +#ifdef SUPPORT_UCP + case OP_PROP_EXTRA + OP_TYPEQUERY: + case OP_PROP_EXTRA + OP_TYPEMINQUERY: + case OP_PROP_EXTRA + OP_TYPEPOSQUERY: + count = 4; + goto QS1; + + case OP_PROP_EXTRA + OP_TYPESTAR: + case OP_PROP_EXTRA + OP_TYPEMINSTAR: + case OP_PROP_EXTRA + OP_TYPEPOSSTAR: + count = 0; + + QS1: + + ADD_ACTIVE(state_offset + 4, 0); + if (clen > 0) + { + BOOL OK; + const ucd_record * prop = GET_UCD(c); + switch(code[2]) + { + case PT_ANY: + OK = TRUE; + break; + + case PT_LAMP: + OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt; + break; + + case PT_GC: + OK = PRIV(ucp_gentype)[prop->chartype] == code[3]; + break; + + case PT_PC: + OK = prop->chartype == code[3]; + break; + + case PT_SC: + OK = prop->script == code[3]; + break; + + /* These are specials for combination cases. */ + + case PT_ALNUM: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N; + break; + + case PT_SPACE: /* Perl space */ + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR; + break; + + case PT_PXSPACE: /* POSIX space */ + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_VT || + c == CHAR_FF || c == CHAR_CR; + break; + + case PT_WORD: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || + c == CHAR_UNDERSCORE; + break; + + /* Should never occur, but keep compilers from grumbling. */ + + default: + OK = codevalue != OP_PROP; + break; + } + + if (OK == (d == OP_PROP)) + { + if (codevalue == OP_PROP_EXTRA + OP_TYPEPOSSTAR || + codevalue == OP_PROP_EXTRA + OP_TYPEPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW(state_offset + count, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_EXTUNI_EXTRA + OP_TYPEQUERY: + case OP_EXTUNI_EXTRA + OP_TYPEMINQUERY: + case OP_EXTUNI_EXTRA + OP_TYPEPOSQUERY: + count = 2; + goto QS2; + + case OP_EXTUNI_EXTRA + OP_TYPESTAR: + case OP_EXTUNI_EXTRA + OP_TYPEMINSTAR: + case OP_EXTUNI_EXTRA + OP_TYPEPOSSTAR: + count = 0; + + QS2: + + ADD_ACTIVE(state_offset + 2, 0); + if (clen > 0 && UCD_CATEGORY(c) != ucp_M) + { + const pcre_uchar *nptr = ptr + clen; + int ncount = 0; + if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSSTAR || + codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + while (nptr < end_subject) + { + int nd; + int ndlen = 1; + GETCHARLEN(nd, nptr, ndlen); + if (UCD_CATEGORY(nd) != ucp_M) break; + ncount++; + nptr += ndlen; + } + ADD_NEW_DATA(-(state_offset + count), 0, ncount); + } + break; +#endif + + /*-----------------------------------------------------------------*/ + case OP_ANYNL_EXTRA + OP_TYPEQUERY: + case OP_ANYNL_EXTRA + OP_TYPEMINQUERY: + case OP_ANYNL_EXTRA + OP_TYPEPOSQUERY: + count = 2; + goto QS3; + + case OP_ANYNL_EXTRA + OP_TYPESTAR: + case OP_ANYNL_EXTRA + OP_TYPEMINSTAR: + case OP_ANYNL_EXTRA + OP_TYPEPOSSTAR: + count = 0; + + QS3: + ADD_ACTIVE(state_offset + 2, 0); + if (clen > 0) + { + int ncount = 0; + switch (c) + { + case 0x000b: + case 0x000c: + case 0x0085: + case 0x2028: + case 0x2029: + if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; + goto ANYNL02; + + case 0x000d: + if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; + /* Fall through */ + + ANYNL02: + case 0x000a: + if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR || + codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW_DATA(-(state_offset + count), 0, ncount); + break; + + default: + break; + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_VSPACE_EXTRA + OP_TYPEQUERY: + case OP_VSPACE_EXTRA + OP_TYPEMINQUERY: + case OP_VSPACE_EXTRA + OP_TYPEPOSQUERY: + count = 2; + goto QS4; + + case OP_VSPACE_EXTRA + OP_TYPESTAR: + case OP_VSPACE_EXTRA + OP_TYPEMINSTAR: + case OP_VSPACE_EXTRA + OP_TYPEPOSSTAR: + count = 0; + + QS4: + ADD_ACTIVE(state_offset + 2, 0); + if (clen > 0) + { + BOOL OK; + switch (c) + { + case 0x000a: + case 0x000b: + case 0x000c: + case 0x000d: + case 0x0085: + case 0x2028: + case 0x2029: + OK = TRUE; + break; + + default: + OK = FALSE; + break; + } + if (OK == (d == OP_VSPACE)) + { + if (codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSSTAR || + codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW_DATA(-(state_offset + count), 0, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_HSPACE_EXTRA + OP_TYPEQUERY: + case OP_HSPACE_EXTRA + OP_TYPEMINQUERY: + case OP_HSPACE_EXTRA + OP_TYPEPOSQUERY: + count = 2; + goto QS5; + + case OP_HSPACE_EXTRA + OP_TYPESTAR: + case OP_HSPACE_EXTRA + OP_TYPEMINSTAR: + case OP_HSPACE_EXTRA + OP_TYPEPOSSTAR: + count = 0; + + QS5: + ADD_ACTIVE(state_offset + 2, 0); + if (clen > 0) + { + BOOL OK; + switch (c) + { + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ + OK = TRUE; + break; + + default: + OK = FALSE; + break; + } + + if (OK == (d == OP_HSPACE)) + { + if (codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSSTAR || + codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW_DATA(-(state_offset + count), 0, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ +#ifdef SUPPORT_UCP + case OP_PROP_EXTRA + OP_TYPEEXACT: + case OP_PROP_EXTRA + OP_TYPEUPTO: + case OP_PROP_EXTRA + OP_TYPEMINUPTO: + case OP_PROP_EXTRA + OP_TYPEPOSUPTO: + if (codevalue != OP_PROP_EXTRA + OP_TYPEEXACT) + { ADD_ACTIVE(state_offset + 1 + IMM2_SIZE + 3, 0); } + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + BOOL OK; + const ucd_record * prop = GET_UCD(c); + switch(code[1 + IMM2_SIZE + 1]) + { + case PT_ANY: + OK = TRUE; + break; + + case PT_LAMP: + OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt; + break; + + case PT_GC: + OK = PRIV(ucp_gentype)[prop->chartype] == code[1 + IMM2_SIZE + 2]; + break; + + case PT_PC: + OK = prop->chartype == code[1 + IMM2_SIZE + 2]; + break; + + case PT_SC: + OK = prop->script == code[1 + IMM2_SIZE + 2]; + break; + + /* These are specials for combination cases. */ + + case PT_ALNUM: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N; + break; + + case PT_SPACE: /* Perl space */ + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR; + break; + + case PT_PXSPACE: /* POSIX space */ + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_VT || + c == CHAR_FF || c == CHAR_CR; + break; + + case PT_WORD: + OK = PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || + c == CHAR_UNDERSCORE; + break; + + /* Should never occur, but keep compilers from grumbling. */ + + default: + OK = codevalue != OP_PROP; + break; + } + + if (OK == (d == OP_PROP)) + { + if (codevalue == OP_PROP_EXTRA + OP_TYPEPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= GET2(code, 1)) + { ADD_NEW(state_offset + 1 + IMM2_SIZE + 3, 0); } + else + { ADD_NEW(state_offset, count); } + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_EXTUNI_EXTRA + OP_TYPEEXACT: + case OP_EXTUNI_EXTRA + OP_TYPEUPTO: + case OP_EXTUNI_EXTRA + OP_TYPEMINUPTO: + case OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO: + if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT) + { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); } + count = current_state->count; /* Number already matched */ + if (clen > 0 && UCD_CATEGORY(c) != ucp_M) + { + const pcre_uchar *nptr = ptr + clen; + int ncount = 0; + if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + while (nptr < end_subject) + { + int nd; + int ndlen = 1; + GETCHARLEN(nd, nptr, ndlen); + if (UCD_CATEGORY(nd) != ucp_M) break; + ncount++; + nptr += ndlen; + } + if (nptr >= end_subject && (md->moptions & PCRE_PARTIAL_HARD) != 0) + reset_could_continue = TRUE; + if (++count >= GET2(code, 1)) + { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, ncount); } + else + { ADD_NEW_DATA(-state_offset, count, ncount); } + } + break; +#endif + + /*-----------------------------------------------------------------*/ + case OP_ANYNL_EXTRA + OP_TYPEEXACT: + case OP_ANYNL_EXTRA + OP_TYPEUPTO: + case OP_ANYNL_EXTRA + OP_TYPEMINUPTO: + case OP_ANYNL_EXTRA + OP_TYPEPOSUPTO: + if (codevalue != OP_ANYNL_EXTRA + OP_TYPEEXACT) + { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); } + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + int ncount = 0; + switch (c) + { + case 0x000b: + case 0x000c: + case 0x0085: + case 0x2028: + case 0x2029: + if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; + goto ANYNL03; + + case 0x000d: + if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; + /* Fall through */ + + ANYNL03: + case 0x000a: + if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= GET2(code, 1)) + { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, ncount); } + else + { ADD_NEW_DATA(-state_offset, count, ncount); } + break; + + default: + break; + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_VSPACE_EXTRA + OP_TYPEEXACT: + case OP_VSPACE_EXTRA + OP_TYPEUPTO: + case OP_VSPACE_EXTRA + OP_TYPEMINUPTO: + case OP_VSPACE_EXTRA + OP_TYPEPOSUPTO: + if (codevalue != OP_VSPACE_EXTRA + OP_TYPEEXACT) + { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); } + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + BOOL OK; + switch (c) + { + case 0x000a: + case 0x000b: + case 0x000c: + case 0x000d: + case 0x0085: + case 0x2028: + case 0x2029: + OK = TRUE; + break; + + default: + OK = FALSE; + } + + if (OK == (d == OP_VSPACE)) + { + if (codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= GET2(code, 1)) + { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, 0); } + else + { ADD_NEW_DATA(-state_offset, count, 0); } + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_HSPACE_EXTRA + OP_TYPEEXACT: + case OP_HSPACE_EXTRA + OP_TYPEUPTO: + case OP_HSPACE_EXTRA + OP_TYPEMINUPTO: + case OP_HSPACE_EXTRA + OP_TYPEPOSUPTO: + if (codevalue != OP_HSPACE_EXTRA + OP_TYPEEXACT) + { ADD_ACTIVE(state_offset + 2 + IMM2_SIZE, 0); } + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + BOOL OK; + switch (c) + { + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ + OK = TRUE; + break; + + default: + OK = FALSE; + break; + } + + if (OK == (d == OP_HSPACE)) + { + if (codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= GET2(code, 1)) + { ADD_NEW_DATA(-(state_offset + 2 + IMM2_SIZE), 0, 0); } + else + { ADD_NEW_DATA(-state_offset, count, 0); } + } + } + break; + +/* ========================================================================== */ + /* These opcodes are followed by a character that is usually compared + to the current subject character; it is loaded into d. We still get + here even if there is no subject character, because in some cases zero + repetitions are permitted. */ + + /*-----------------------------------------------------------------*/ + case OP_CHAR: + if (clen > 0 && c == d) { ADD_NEW(state_offset + dlen + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + case OP_CHARI: + if (clen == 0) break; + +#ifdef SUPPORT_UTF + if (utf) + { + if (c == d) { ADD_NEW(state_offset + dlen + 1, 0); } else + { + unsigned int othercase; + if (c < 128) + othercase = fcc[c]; + else + /* If we have Unicode property support, we can use it to test the + other case of the character. */ +#ifdef SUPPORT_UCP + othercase = UCD_OTHERCASE(c); +#else + othercase = NOTACHAR; +#endif + + if (d == othercase) { ADD_NEW(state_offset + dlen + 1, 0); } + } + } + else +#endif /* SUPPORT_UTF */ + /* Not UTF mode */ + { + if (TABLE_GET(c, lcc, c) == TABLE_GET(d, lcc, d)) + { ADD_NEW(state_offset + 2, 0); } + } + break; + + +#ifdef SUPPORT_UCP + /*-----------------------------------------------------------------*/ + /* This is a tricky one because it can match more than one character. + Find out how many characters to skip, and then set up a negative state + to wait for them to pass before continuing. */ + + case OP_EXTUNI: + if (clen > 0 && UCD_CATEGORY(c) != ucp_M) + { + const pcre_uchar *nptr = ptr + clen; + int ncount = 0; + while (nptr < end_subject) + { + int nclen = 1; + GETCHARLEN(c, nptr, nclen); + if (UCD_CATEGORY(c) != ucp_M) break; + ncount++; + nptr += nclen; + } + if (nptr >= end_subject && (md->moptions & PCRE_PARTIAL_HARD) != 0) + reset_could_continue = TRUE; + ADD_NEW_DATA(-(state_offset + 1), 0, ncount); + } + break; +#endif + + /*-----------------------------------------------------------------*/ + /* This is a tricky like EXTUNI because it too can match more than one + character (when CR is followed by LF). In this case, set up a negative + state to wait for one character to pass before continuing. */ + + case OP_ANYNL: + if (clen > 0) switch(c) + { + case 0x000b: + case 0x000c: + case 0x0085: + case 0x2028: + case 0x2029: + if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; + + case 0x000a: + ADD_NEW(state_offset + 1, 0); + break; + + case 0x000d: + if (ptr + 1 >= end_subject) + { + ADD_NEW(state_offset + 1, 0); + if ((md->moptions & PCRE_PARTIAL_HARD) != 0) + reset_could_continue = TRUE; + } + else if (ptr[1] == 0x0a) + { + ADD_NEW_DATA(-(state_offset + 1), 0, 1); + } + else + { + ADD_NEW(state_offset + 1, 0); + } + break; + } + break; + + /*-----------------------------------------------------------------*/ + case OP_NOT_VSPACE: + if (clen > 0) switch(c) + { + case 0x000a: + case 0x000b: + case 0x000c: + case 0x000d: + case 0x0085: + case 0x2028: + case 0x2029: + break; + + default: + ADD_NEW(state_offset + 1, 0); + break; + } + break; + + /*-----------------------------------------------------------------*/ + case OP_VSPACE: + if (clen > 0) switch(c) + { + case 0x000a: + case 0x000b: + case 0x000c: + case 0x000d: + case 0x0085: + case 0x2028: + case 0x2029: + ADD_NEW(state_offset + 1, 0); + break; + + default: break; + } + break; + + /*-----------------------------------------------------------------*/ + case OP_NOT_HSPACE: + if (clen > 0) switch(c) + { + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ + break; + + default: + ADD_NEW(state_offset + 1, 0); + break; + } + break; + + /*-----------------------------------------------------------------*/ + case OP_HSPACE: + if (clen > 0) switch(c) + { + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ + ADD_NEW(state_offset + 1, 0); + break; + } + break; + + /*-----------------------------------------------------------------*/ + /* Match a negated single character casefully. */ + + case OP_NOT: + if (clen > 0 && c != d) { ADD_NEW(state_offset + dlen + 1, 0); } + break; + + /*-----------------------------------------------------------------*/ + /* Match a negated single character caselessly. */ + + case OP_NOTI: + if (clen > 0) + { + unsigned int otherd; +#ifdef SUPPORT_UTF + if (utf && d >= 128) + { +#ifdef SUPPORT_UCP + otherd = UCD_OTHERCASE(d); +#endif /* SUPPORT_UCP */ + } + else +#endif /* SUPPORT_UTF */ + otherd = TABLE_GET(d, fcc, d); + if (c != d && c != otherd) + { ADD_NEW(state_offset + dlen + 1, 0); } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_PLUSI: + case OP_MINPLUSI: + case OP_POSPLUSI: + case OP_NOTPLUSI: + case OP_NOTMINPLUSI: + case OP_NOTPOSPLUSI: + caseless = TRUE; + codevalue -= OP_STARI - OP_STAR; + + /* Fall through */ + case OP_PLUS: + case OP_MINPLUS: + case OP_POSPLUS: + case OP_NOTPLUS: + case OP_NOTMINPLUS: + case OP_NOTPOSPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(state_offset + dlen + 1, 0); } + if (clen > 0) + { + unsigned int otherd = NOTACHAR; + if (caseless) + { +#ifdef SUPPORT_UTF + if (utf && d >= 128) + { +#ifdef SUPPORT_UCP + otherd = UCD_OTHERCASE(d); +#endif /* SUPPORT_UCP */ + } + else +#endif /* SUPPORT_UTF */ + otherd = TABLE_GET(d, fcc, d); + } + if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) + { + if (count > 0 && + (codevalue == OP_POSPLUS || codevalue == OP_NOTPOSPLUS)) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + count++; + ADD_NEW(state_offset, count); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_QUERYI: + case OP_MINQUERYI: + case OP_POSQUERYI: + case OP_NOTQUERYI: + case OP_NOTMINQUERYI: + case OP_NOTPOSQUERYI: + caseless = TRUE; + codevalue -= OP_STARI - OP_STAR; + /* Fall through */ + case OP_QUERY: + case OP_MINQUERY: + case OP_POSQUERY: + case OP_NOTQUERY: + case OP_NOTMINQUERY: + case OP_NOTPOSQUERY: + ADD_ACTIVE(state_offset + dlen + 1, 0); + if (clen > 0) + { + unsigned int otherd = NOTACHAR; + if (caseless) + { +#ifdef SUPPORT_UTF + if (utf && d >= 128) + { +#ifdef SUPPORT_UCP + otherd = UCD_OTHERCASE(d); +#endif /* SUPPORT_UCP */ + } + else +#endif /* SUPPORT_UTF */ + otherd = TABLE_GET(d, fcc, d); + } + if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) + { + if (codevalue == OP_POSQUERY || codevalue == OP_NOTPOSQUERY) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW(state_offset + dlen + 1, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_STARI: + case OP_MINSTARI: + case OP_POSSTARI: + case OP_NOTSTARI: + case OP_NOTMINSTARI: + case OP_NOTPOSSTARI: + caseless = TRUE; + codevalue -= OP_STARI - OP_STAR; + /* Fall through */ + case OP_STAR: + case OP_MINSTAR: + case OP_POSSTAR: + case OP_NOTSTAR: + case OP_NOTMINSTAR: + case OP_NOTPOSSTAR: + ADD_ACTIVE(state_offset + dlen + 1, 0); + if (clen > 0) + { + unsigned int otherd = NOTACHAR; + if (caseless) + { +#ifdef SUPPORT_UTF + if (utf && d >= 128) + { +#ifdef SUPPORT_UCP + otherd = UCD_OTHERCASE(d); +#endif /* SUPPORT_UCP */ + } + else +#endif /* SUPPORT_UTF */ + otherd = TABLE_GET(d, fcc, d); + } + if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) + { + if (codevalue == OP_POSSTAR || codevalue == OP_NOTPOSSTAR) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + ADD_NEW(state_offset, 0); + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_EXACTI: + case OP_NOTEXACTI: + caseless = TRUE; + codevalue -= OP_STARI - OP_STAR; + /* Fall through */ + case OP_EXACT: + case OP_NOTEXACT: + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + unsigned int otherd = NOTACHAR; + if (caseless) + { +#ifdef SUPPORT_UTF + if (utf && d >= 128) + { +#ifdef SUPPORT_UCP + otherd = UCD_OTHERCASE(d); +#endif /* SUPPORT_UCP */ + } + else +#endif /* SUPPORT_UTF */ + otherd = TABLE_GET(d, fcc, d); + } + if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) + { + if (++count >= GET2(code, 1)) + { ADD_NEW(state_offset + dlen + 1 + IMM2_SIZE, 0); } + else + { ADD_NEW(state_offset, count); } + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_UPTOI: + case OP_MINUPTOI: + case OP_POSUPTOI: + case OP_NOTUPTOI: + case OP_NOTMINUPTOI: + case OP_NOTPOSUPTOI: + caseless = TRUE; + codevalue -= OP_STARI - OP_STAR; + /* Fall through */ + case OP_UPTO: + case OP_MINUPTO: + case OP_POSUPTO: + case OP_NOTUPTO: + case OP_NOTMINUPTO: + case OP_NOTPOSUPTO: + ADD_ACTIVE(state_offset + dlen + 1 + IMM2_SIZE, 0); + count = current_state->count; /* Number already matched */ + if (clen > 0) + { + unsigned int otherd = NOTACHAR; + if (caseless) + { +#ifdef SUPPORT_UTF + if (utf && d >= 128) + { +#ifdef SUPPORT_UCP + otherd = UCD_OTHERCASE(d); +#endif /* SUPPORT_UCP */ + } + else +#endif /* SUPPORT_UTF */ + otherd = TABLE_GET(d, fcc, d); + } + if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) + { + if (codevalue == OP_POSUPTO || codevalue == OP_NOTPOSUPTO) + { + active_count--; /* Remove non-match possibility */ + next_active_state--; + } + if (++count >= GET2(code, 1)) + { ADD_NEW(state_offset + dlen + 1 + IMM2_SIZE, 0); } + else + { ADD_NEW(state_offset, count); } + } + } + break; + + +/* ========================================================================== */ + /* These are the class-handling opcodes */ + + case OP_CLASS: + case OP_NCLASS: + case OP_XCLASS: + { + BOOL isinclass = FALSE; + int next_state_offset; + const pcre_uchar *ecode; + + /* For a simple class, there is always just a 32-byte table, and we + can set isinclass from it. */ + + if (codevalue != OP_XCLASS) + { + ecode = code + 1 + (32 / sizeof(pcre_uchar)); + if (clen > 0) + { + isinclass = (c > 255)? (codevalue == OP_NCLASS) : + ((((pcre_uint8 *)(code + 1))[c/8] & (1 << (c&7))) != 0); + } + } + + /* An extended class may have a table or a list of single characters, + ranges, or both, and it may be positive or negative. There's a + function that sorts all this out. */ + + else + { + ecode = code + GET(code, 1); + if (clen > 0) isinclass = PRIV(xclass)(c, code + 1 + LINK_SIZE, utf); + } + + /* At this point, isinclass is set for all kinds of class, and ecode + points to the byte after the end of the class. If there is a + quantifier, this is where it will be. */ + + next_state_offset = (int)(ecode - start_code); + + switch (*ecode) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + ADD_ACTIVE(next_state_offset + 1, 0); + if (isinclass) { ADD_NEW(state_offset, 0); } + break; + + case OP_CRPLUS: + case OP_CRMINPLUS: + count = current_state->count; /* Already matched */ + if (count > 0) { ADD_ACTIVE(next_state_offset + 1, 0); } + if (isinclass) { count++; ADD_NEW(state_offset, count); } + break; + + case OP_CRQUERY: + case OP_CRMINQUERY: + ADD_ACTIVE(next_state_offset + 1, 0); + if (isinclass) { ADD_NEW(next_state_offset + 1, 0); } + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + count = current_state->count; /* Already matched */ + if (count >= GET2(ecode, 1)) + { ADD_ACTIVE(next_state_offset + 1 + 2 * IMM2_SIZE, 0); } + if (isinclass) + { + int max = GET2(ecode, 1 + IMM2_SIZE); + if (++count >= max && max != 0) /* Max 0 => no limit */ + { ADD_NEW(next_state_offset + 1 + 2 * IMM2_SIZE, 0); } + else + { ADD_NEW(state_offset, count); } + } + break; + + default: + if (isinclass) { ADD_NEW(next_state_offset, 0); } + break; + } + } + break; + +/* ========================================================================== */ + /* These are the opcodes for fancy brackets of various kinds. We have + to use recursion in order to handle them. The "always failing" assertion + (?!) is optimised to OP_FAIL when compiling, so we have to support that, + though the other "backtracking verbs" are not supported. */ + + case OP_FAIL: + forced_fail++; /* Count FAILs for multiple states */ + break; + + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + { + int rc; + int local_offsets[2]; + int local_workspace[1000]; + const pcre_uchar *endasscode = code + GET(code, 1); + + while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1); + + rc = internal_dfa_exec( + md, /* static match data */ + code, /* this subexpression's code */ + ptr, /* where we currently are */ + (int)(ptr - start_subject), /* start offset */ + local_offsets, /* offset vector */ + sizeof(local_offsets)/sizeof(int), /* size of same */ + local_workspace, /* workspace vector */ + sizeof(local_workspace)/sizeof(int), /* size of same */ + rlevel); /* function recursion level */ + + if (rc == PCRE_ERROR_DFA_UITEM) return rc; + if ((rc >= 0) == (codevalue == OP_ASSERT || codevalue == OP_ASSERTBACK)) + { ADD_ACTIVE((int)(endasscode + LINK_SIZE + 1 - start_code), 0); } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_COND: + case OP_SCOND: + { + int local_offsets[1000]; + int local_workspace[1000]; + int codelink = GET(code, 1); + int condcode; + + /* Because of the way auto-callout works during compile, a callout item + is inserted between OP_COND and an assertion condition. This does not + happen for the other conditions. */ + + if (code[LINK_SIZE+1] == OP_CALLOUT) + { + rrc = 0; + if (PUBL(callout) != NULL) + { + PUBL(callout_block) cb; + cb.version = 1; /* Version 1 of the callout block */ + cb.callout_number = code[LINK_SIZE+2]; + cb.offset_vector = offsets; +#ifdef COMPILE_PCRE8 + cb.subject = (PCRE_SPTR)start_subject; +#else + cb.subject = (PCRE_SPTR16)start_subject; +#endif + cb.subject_length = (int)(end_subject - start_subject); + cb.start_match = (int)(current_subject - start_subject); + cb.current_position = (int)(ptr - start_subject); + cb.pattern_position = GET(code, LINK_SIZE + 3); + cb.next_item_length = GET(code, 3 + 2*LINK_SIZE); + cb.capture_top = 1; + cb.capture_last = -1; + cb.callout_data = md->callout_data; + cb.mark = NULL; /* No (*MARK) support */ + if ((rrc = (*PUBL(callout))(&cb)) < 0) return rrc; /* Abandon */ + } + if (rrc > 0) break; /* Fail this thread */ + code += PRIV(OP_lengths)[OP_CALLOUT]; /* Skip callout data */ + } + + condcode = code[LINK_SIZE+1]; + + /* Back reference conditions are not supported */ + + if (condcode == OP_CREF || condcode == OP_NCREF) + return PCRE_ERROR_DFA_UCOND; + + /* The DEFINE condition is always false */ + + if (condcode == OP_DEF) + { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); } + + /* The only supported version of OP_RREF is for the value RREF_ANY, + which means "test if in any recursion". We can't test for specifically + recursed groups. */ + + else if (condcode == OP_RREF || condcode == OP_NRREF) + { + int value = GET2(code, LINK_SIZE + 2); + if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND; + if (md->recursive != NULL) + { ADD_ACTIVE(state_offset + LINK_SIZE + 2 + IMM2_SIZE, 0); } + else { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); } + } + + /* Otherwise, the condition is an assertion */ + + else + { + int rc; + const pcre_uchar *asscode = code + LINK_SIZE + 1; + const pcre_uchar *endasscode = asscode + GET(asscode, 1); + + while (*endasscode == OP_ALT) endasscode += GET(endasscode, 1); + + rc = internal_dfa_exec( + md, /* fixed match data */ + asscode, /* this subexpression's code */ + ptr, /* where we currently are */ + (int)(ptr - start_subject), /* start offset */ + local_offsets, /* offset vector */ + sizeof(local_offsets)/sizeof(int), /* size of same */ + local_workspace, /* workspace vector */ + sizeof(local_workspace)/sizeof(int), /* size of same */ + rlevel); /* function recursion level */ + + if (rc == PCRE_ERROR_DFA_UITEM) return rc; + if ((rc >= 0) == + (condcode == OP_ASSERT || condcode == OP_ASSERTBACK)) + { ADD_ACTIVE((int)(endasscode + LINK_SIZE + 1 - start_code), 0); } + else + { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); } + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_RECURSE: + { + dfa_recursion_info *ri; + int local_offsets[1000]; + int local_workspace[1000]; + const pcre_uchar *callpat = start_code + GET(code, 1); + int recno = (callpat == md->start_code)? 0 : + GET2(callpat, 1 + LINK_SIZE); + int rc; + + DPRINTF(("%.*sStarting regex recursion\n", rlevel*2-2, SP)); + + /* Check for repeating a recursion without advancing the subject + pointer. This should catch convoluted mutual recursions. (Some simple + cases are caught at compile time.) */ + + for (ri = md->recursive; ri != NULL; ri = ri->prevrec) + if (recno == ri->group_num && ptr == ri->subject_position) + return PCRE_ERROR_RECURSELOOP; + + /* Remember this recursion and where we started it so as to + catch infinite loops. */ + + new_recursive.group_num = recno; + new_recursive.subject_position = ptr; + new_recursive.prevrec = md->recursive; + md->recursive = &new_recursive; + + rc = internal_dfa_exec( + md, /* fixed match data */ + callpat, /* this subexpression's code */ + ptr, /* where we currently are */ + (int)(ptr - start_subject), /* start offset */ + local_offsets, /* offset vector */ + sizeof(local_offsets)/sizeof(int), /* size of same */ + local_workspace, /* workspace vector */ + sizeof(local_workspace)/sizeof(int), /* size of same */ + rlevel); /* function recursion level */ + + md->recursive = new_recursive.prevrec; /* Done this recursion */ + + DPRINTF(("%.*sReturn from regex recursion: rc=%d\n", rlevel*2-2, SP, + rc)); + + /* Ran out of internal offsets */ + + if (rc == 0) return PCRE_ERROR_DFA_RECURSE; + + /* For each successful matched substring, set up the next state with a + count of characters to skip before trying it. Note that the count is in + characters, not bytes. */ + + if (rc > 0) + { + for (rc = rc*2 - 2; rc >= 0; rc -= 2) + { + int charcount = local_offsets[rc+1] - local_offsets[rc]; +#ifdef SUPPORT_UTF + if (utf) + { + const pcre_uchar *p = start_subject + local_offsets[rc]; + const pcre_uchar *pp = start_subject + local_offsets[rc+1]; + while (p < pp) if (NOT_FIRSTCHAR(*p++)) charcount--; + } +#endif + if (charcount > 0) + { + ADD_NEW_DATA(-(state_offset + LINK_SIZE + 1), 0, (charcount - 1)); + } + else + { + ADD_ACTIVE(state_offset + LINK_SIZE + 1, 0); + } + } + } + else if (rc != PCRE_ERROR_NOMATCH) return rc; + } + break; + + /*-----------------------------------------------------------------*/ + case OP_BRAPOS: + case OP_SBRAPOS: + case OP_CBRAPOS: + case OP_SCBRAPOS: + case OP_BRAPOSZERO: + { + int charcount, matched_count; + const pcre_uchar *local_ptr = ptr; + BOOL allow_zero; + + if (codevalue == OP_BRAPOSZERO) + { + allow_zero = TRUE; + codevalue = *(++code); /* Codevalue will be one of above BRAs */ + } + else allow_zero = FALSE; + + /* Loop to match the subpattern as many times as possible as if it were + a complete pattern. */ + + for (matched_count = 0;; matched_count++) + { + int local_offsets[2]; + int local_workspace[1000]; + + int rc = internal_dfa_exec( + md, /* fixed match data */ + code, /* this subexpression's code */ + local_ptr, /* where we currently are */ + (int)(ptr - start_subject), /* start offset */ + local_offsets, /* offset vector */ + sizeof(local_offsets)/sizeof(int), /* size of same */ + local_workspace, /* workspace vector */ + sizeof(local_workspace)/sizeof(int), /* size of same */ + rlevel); /* function recursion level */ + + /* Failed to match */ + + if (rc < 0) + { + if (rc != PCRE_ERROR_NOMATCH) return rc; + break; + } + + /* Matched: break the loop if zero characters matched. */ + + charcount = local_offsets[1] - local_offsets[0]; + if (charcount == 0) break; + local_ptr += charcount; /* Advance temporary position ptr */ + } + + /* At this point we have matched the subpattern matched_count + times, and local_ptr is pointing to the character after the end of the + last match. */ + + if (matched_count > 0 || allow_zero) + { + const pcre_uchar *end_subpattern = code; + int next_state_offset; + + do { end_subpattern += GET(end_subpattern, 1); } + while (*end_subpattern == OP_ALT); + next_state_offset = + (int)(end_subpattern - start_code + LINK_SIZE + 1); + + /* Optimization: if there are no more active states, and there + are no new states yet set up, then skip over the subject string + right here, to save looping. Otherwise, set up the new state to swing + into action when the end of the matched substring is reached. */ + + if (i + 1 >= active_count && new_count == 0) + { + ptr = local_ptr; + clen = 0; + ADD_NEW(next_state_offset, 0); + } + else + { + const pcre_uchar *p = ptr; + const pcre_uchar *pp = local_ptr; + charcount = (int)(pp - p); +#ifdef SUPPORT_UTF + if (utf) while (p < pp) if (NOT_FIRSTCHAR(*p++)) charcount--; +#endif + ADD_NEW_DATA(-next_state_offset, 0, (charcount - 1)); + } + } + } + break; + + /*-----------------------------------------------------------------*/ + case OP_ONCE: + case OP_ONCE_NC: + { + int local_offsets[2]; + int local_workspace[1000]; + + int rc = internal_dfa_exec( + md, /* fixed match data */ + code, /* this subexpression's code */ + ptr, /* where we currently are */ + (int)(ptr - start_subject), /* start offset */ + local_offsets, /* offset vector */ + sizeof(local_offsets)/sizeof(int), /* size of same */ + local_workspace, /* workspace vector */ + sizeof(local_workspace)/sizeof(int), /* size of same */ + rlevel); /* function recursion level */ + + if (rc >= 0) + { + const pcre_uchar *end_subpattern = code; + int charcount = local_offsets[1] - local_offsets[0]; + int next_state_offset, repeat_state_offset; + + do { end_subpattern += GET(end_subpattern, 1); } + while (*end_subpattern == OP_ALT); + next_state_offset = + (int)(end_subpattern - start_code + LINK_SIZE + 1); + + /* If the end of this subpattern is KETRMAX or KETRMIN, we must + arrange for the repeat state also to be added to the relevant list. + Calculate the offset, or set -1 for no repeat. */ + + repeat_state_offset = (*end_subpattern == OP_KETRMAX || + *end_subpattern == OP_KETRMIN)? + (int)(end_subpattern - start_code - GET(end_subpattern, 1)) : -1; + + /* If we have matched an empty string, add the next state at the + current character pointer. This is important so that the duplicate + checking kicks in, which is what breaks infinite loops that match an + empty string. */ + + if (charcount == 0) + { + ADD_ACTIVE(next_state_offset, 0); + } + + /* Optimization: if there are no more active states, and there + are no new states yet set up, then skip over the subject string + right here, to save looping. Otherwise, set up the new state to swing + into action when the end of the matched substring is reached. */ + + else if (i + 1 >= active_count && new_count == 0) + { + ptr += charcount; + clen = 0; + ADD_NEW(next_state_offset, 0); + + /* If we are adding a repeat state at the new character position, + we must fudge things so that it is the only current state. + Otherwise, it might be a duplicate of one we processed before, and + that would cause it to be skipped. */ + + if (repeat_state_offset >= 0) + { + next_active_state = active_states; + active_count = 0; + i = -1; + ADD_ACTIVE(repeat_state_offset, 0); + } + } + else + { +#ifdef SUPPORT_UTF + if (utf) + { + const pcre_uchar *p = start_subject + local_offsets[0]; + const pcre_uchar *pp = start_subject + local_offsets[1]; + while (p < pp) if (NOT_FIRSTCHAR(*p++)) charcount--; + } +#endif + ADD_NEW_DATA(-next_state_offset, 0, (charcount - 1)); + if (repeat_state_offset >= 0) + { ADD_NEW_DATA(-repeat_state_offset, 0, (charcount - 1)); } + } + } + else if (rc != PCRE_ERROR_NOMATCH) return rc; + } + break; + + +/* ========================================================================== */ + /* Handle callouts */ + + case OP_CALLOUT: + rrc = 0; + if (PUBL(callout) != NULL) + { + PUBL(callout_block) cb; + cb.version = 1; /* Version 1 of the callout block */ + cb.callout_number = code[1]; + cb.offset_vector = offsets; +#ifdef COMPILE_PCRE8 + cb.subject = (PCRE_SPTR)start_subject; +#else + cb.subject = (PCRE_SPTR16)start_subject; +#endif + cb.subject_length = (int)(end_subject - start_subject); + cb.start_match = (int)(current_subject - start_subject); + cb.current_position = (int)(ptr - start_subject); + cb.pattern_position = GET(code, 2); + cb.next_item_length = GET(code, 2 + LINK_SIZE); + cb.capture_top = 1; + cb.capture_last = -1; + cb.callout_data = md->callout_data; + cb.mark = NULL; /* No (*MARK) support */ + if ((rrc = (*PUBL(callout))(&cb)) < 0) return rrc; /* Abandon */ + } + if (rrc == 0) + { ADD_ACTIVE(state_offset + PRIV(OP_lengths)[OP_CALLOUT], 0); } + break; + + +/* ========================================================================== */ + default: /* Unsupported opcode */ + return PCRE_ERROR_DFA_UITEM; + } + + NEXT_ACTIVE_STATE: continue; + + } /* End of loop scanning active states */ + + /* We have finished the processing at the current subject character. If no + new states have been set for the next character, we have found all the + matches that we are going to find. If we are at the top level and partial + matching has been requested, check for appropriate conditions. + + The "forced_ fail" variable counts the number of (*F) encountered for the + character. If it is equal to the original active_count (saved in + workspace[1]) it means that (*F) was found on every active state. In this + case we don't want to give a partial match. + + The "could_continue" variable is true if a state could have continued but + for the fact that the end of the subject was reached. */ + + if (new_count <= 0) + { + if (rlevel == 1 && /* Top level, and */ + could_continue && /* Some could go on, and */ + forced_fail != workspace[1] && /* Not all forced fail & */ + ( /* either... */ + (md->moptions & PCRE_PARTIAL_HARD) != 0 /* Hard partial */ + || /* or... */ + ((md->moptions & PCRE_PARTIAL_SOFT) != 0 && /* Soft partial and */ + match_count < 0) /* no matches */ + ) && /* And... */ + ( + partial_newline || /* Either partial NL */ + ( /* or ... */ + ptr >= end_subject && /* End of subject and */ + ptr > md->start_used_ptr) /* Inspected non-empty string */ + ) + ) + { + if (offsetcount >= 2) + { + offsets[0] = (int)(md->start_used_ptr - start_subject); + offsets[1] = (int)(end_subject - start_subject); + } + match_count = PCRE_ERROR_PARTIAL; + } + + DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n" + "%.*s---------------------\n\n", rlevel*2-2, SP, rlevel, match_count, + rlevel*2-2, SP)); + break; /* In effect, "return", but see the comment below */ + } + + /* One or more states are active for the next character. */ + + ptr += clen; /* Advance to next subject character */ + } /* Loop to move along the subject string */ + +/* Control gets here from "break" a few lines above. We do it this way because +if we use "return" above, we have compiler trouble. Some compilers warn if +there's nothing here because they think the function doesn't return a value. On +the other hand, if we put a dummy statement here, some more clever compilers +complain that it can't be reached. Sigh. */ + +return match_count; +} + + + + +/************************************************* +* Execute a Regular Expression - DFA engine * +*************************************************/ + +/* This external function applies a compiled re to a subject string using a DFA +engine. This function calls the internal function multiple times if the pattern +is not anchored. + +Arguments: + argument_re points to the compiled expression + extra_data points to extra data or is NULL + subject points to the subject string + length length of subject string (may contain binary zeros) + start_offset where to start in the subject string + options option bits + offsets vector of match offsets + offsetcount size of same + workspace workspace vector + wscount size of same + +Returns: > 0 => number of match offset pairs placed in offsets + = 0 => offsets overflowed; longest matches are present + -1 => failed to match + < -1 => some kind of unexpected problem +*/ + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data, + const char *subject, int length, int start_offset, int options, int *offsets, + int offsetcount, int *workspace, int wscount) +#else +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_dfa_exec(const pcre16 *argument_re, const pcre16_extra *extra_data, + PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets, + int offsetcount, int *workspace, int wscount) +#endif +{ +REAL_PCRE *re = (REAL_PCRE *)argument_re; +dfa_match_data match_block; +dfa_match_data *md = &match_block; +BOOL utf, anchored, startline, firstline; +const pcre_uchar *current_subject, *end_subject; +const pcre_study_data *study = NULL; + +const pcre_uchar *req_char_ptr; +const pcre_uint8 *start_bits = NULL; +BOOL has_first_char = FALSE; +BOOL has_req_char = FALSE; +pcre_uchar first_char = 0; +pcre_uchar first_char2 = 0; +pcre_uchar req_char = 0; +pcre_uchar req_char2 = 0; +int newline; + +/* Plausibility checks */ + +if ((options & ~PUBLIC_DFA_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION; +if (re == NULL || subject == NULL || workspace == NULL || + (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL; +if (offsetcount < 0) return PCRE_ERROR_BADCOUNT; +if (wscount < 20) return PCRE_ERROR_DFA_WSSIZE; +if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET; + +/* Check that the first field in the block is the magic number. If it is not, +return with PCRE_ERROR_BADMAGIC. However, if the magic number is equal to +REVERSED_MAGIC_NUMBER we return with PCRE_ERROR_BADENDIANNESS, which +means that the pattern is likely compiled with different endianness. */ + +if (re->magic_number != MAGIC_NUMBER) + return re->magic_number == REVERSED_MAGIC_NUMBER? + PCRE_ERROR_BADENDIANNESS:PCRE_ERROR_BADMAGIC; +if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE; + +/* If restarting after a partial match, do some sanity checks on the contents +of the workspace. */ + +if ((options & PCRE_DFA_RESTART) != 0) + { + if ((workspace[0] & (-2)) != 0 || workspace[1] < 1 || + workspace[1] > (wscount - 2)/INTS_PER_STATEBLOCK) + return PCRE_ERROR_DFA_BADRESTART; + } + +/* Set up study, callout, and table data */ + +md->tables = re->tables; +md->callout_data = NULL; + +if (extra_data != NULL) + { + unsigned int flags = extra_data->flags; + if ((flags & PCRE_EXTRA_STUDY_DATA) != 0) + study = (const pcre_study_data *)extra_data->study_data; + if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0) return PCRE_ERROR_DFA_UMLIMIT; + if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0) + return PCRE_ERROR_DFA_UMLIMIT; + if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0) + md->callout_data = extra_data->callout_data; + if ((flags & PCRE_EXTRA_TABLES) != 0) + md->tables = extra_data->tables; + } + +/* Set some local values */ + +current_subject = (const pcre_uchar *)subject + start_offset; +end_subject = (const pcre_uchar *)subject + length; +req_char_ptr = current_subject - 1; + +#ifdef SUPPORT_UTF +/* PCRE_UTF16 has the same value as PCRE_UTF8. */ +utf = (re->options & PCRE_UTF8) != 0; +#else +utf = FALSE; +#endif + +anchored = (options & (PCRE_ANCHORED|PCRE_DFA_RESTART)) != 0 || + (re->options & PCRE_ANCHORED) != 0; + +/* The remaining fixed data for passing around. */ + +md->start_code = (const pcre_uchar *)argument_re + + re->name_table_offset + re->name_count * re->name_entry_size; +md->start_subject = (const pcre_uchar *)subject; +md->end_subject = end_subject; +md->start_offset = start_offset; +md->moptions = options; +md->poptions = re->options; + +/* If the BSR option is not set at match time, copy what was set +at compile time. */ + +if ((md->moptions & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) == 0) + { + if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0) + md->moptions |= re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE); +#ifdef BSR_ANYCRLF + else md->moptions |= PCRE_BSR_ANYCRLF; +#endif + } + +/* Handle different types of newline. The three bits give eight cases. If +nothing is set at run time, whatever was used at compile time applies. */ + +switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : (pcre_uint32)options) & + PCRE_NEWLINE_BITS) + { + case 0: newline = NEWLINE; break; /* Compile-time default */ + case PCRE_NEWLINE_CR: newline = CHAR_CR; break; + case PCRE_NEWLINE_LF: newline = CHAR_NL; break; + case PCRE_NEWLINE_CR+ + PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break; + case PCRE_NEWLINE_ANY: newline = -1; break; + case PCRE_NEWLINE_ANYCRLF: newline = -2; break; + default: return PCRE_ERROR_BADNEWLINE; + } + +if (newline == -2) + { + md->nltype = NLTYPE_ANYCRLF; + } +else if (newline < 0) + { + md->nltype = NLTYPE_ANY; + } +else + { + md->nltype = NLTYPE_FIXED; + if (newline > 255) + { + md->nllen = 2; + md->nl[0] = (newline >> 8) & 255; + md->nl[1] = newline & 255; + } + else + { + md->nllen = 1; + md->nl[0] = newline; + } + } + +/* Check a UTF-8 string if required. Unfortunately there's no way of passing +back the character offset. */ + +#ifdef SUPPORT_UTF +if (utf && (options & PCRE_NO_UTF8_CHECK) == 0) + { + int erroroffset; + int errorcode = PRIV(valid_utf)((pcre_uchar *)subject, length, &erroroffset); + if (errorcode != 0) + { + if (offsetcount >= 2) + { + offsets[0] = erroroffset; + offsets[1] = errorcode; + } + return (errorcode <= PCRE_UTF8_ERR5 && (options & PCRE_PARTIAL_HARD) != 0)? + PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8; + } + if (start_offset > 0 && start_offset < length && + NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset])) + return PCRE_ERROR_BADUTF8_OFFSET; + } +#endif + +/* If the exec call supplied NULL for tables, use the inbuilt ones. This +is a feature that makes it possible to save compiled regex and re-use them +in other programs later. */ + +if (md->tables == NULL) md->tables = PRIV(default_tables); + +/* The "must be at the start of a line" flags are used in a loop when finding +where to start. */ + +startline = (re->flags & PCRE_STARTLINE) != 0; +firstline = (re->options & PCRE_FIRSTLINE) != 0; + +/* Set up the first character to match, if available. The first_byte value is +never set for an anchored regular expression, but the anchoring may be forced +at run time, so we have to test for anchoring. The first char may be unset for +an unanchored pattern, of course. If there's no first char and the pattern was +studied, there may be a bitmap of possible first characters. */ + +if (!anchored) + { + if ((re->flags & PCRE_FIRSTSET) != 0) + { + has_first_char = TRUE; + first_char = first_char2 = (pcre_uchar)(re->first_char); + if ((re->flags & PCRE_FCH_CASELESS) != 0) + { + first_char2 = TABLE_GET(first_char, md->tables + fcc_offset, first_char); +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + if (utf && first_char > 127) + first_char2 = UCD_OTHERCASE(first_char); +#endif + } + } + else + { + if (!startline && study != NULL && + (study->flags & PCRE_STUDY_MAPPED) != 0) + start_bits = study->start_bits; + } + } + +/* For anchored or unanchored matches, there may be a "last known required +character" set. */ + +if ((re->flags & PCRE_REQCHSET) != 0) + { + has_req_char = TRUE; + req_char = req_char2 = (pcre_uchar)(re->req_char); + if ((re->flags & PCRE_RCH_CASELESS) != 0) + { + req_char2 = TABLE_GET(req_char, md->tables + fcc_offset, req_char); +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + if (utf && req_char > 127) + req_char2 = UCD_OTHERCASE(req_char); +#endif + } + } + +/* Call the main matching function, looping for a non-anchored regex after a +failed match. If not restarting, perform certain optimizations at the start of +a match. */ + +for (;;) + { + int rc; + + if ((options & PCRE_DFA_RESTART) == 0) + { + const pcre_uchar *save_end_subject = end_subject; + + /* If firstline is TRUE, the start of the match is constrained to the first + line of a multiline string. Implement this by temporarily adjusting + end_subject so that we stop scanning at a newline. If the match fails at + the newline, later code breaks this loop. */ + + if (firstline) + { + PCRE_PUCHAR t = current_subject; +#ifdef SUPPORT_UTF + if (utf) + { + while (t < md->end_subject && !IS_NEWLINE(t)) + { + t++; + ACROSSCHAR(t < end_subject, *t, t++); + } + } + else +#endif + while (t < md->end_subject && !IS_NEWLINE(t)) t++; + end_subject = t; + } + + /* There are some optimizations that avoid running the match if a known + starting point is not found. However, there is an option that disables + these, for testing and for ensuring that all callouts do actually occur. + The option can be set in the regex by (*NO_START_OPT) or passed in + match-time options. */ + + if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0) + { + /* Advance to a known first char. */ + + if (has_first_char) + { + if (first_char != first_char2) + while (current_subject < end_subject && + *current_subject != first_char && *current_subject != first_char2) + current_subject++; + else + while (current_subject < end_subject && + *current_subject != first_char) + current_subject++; + } + + /* Or to just after a linebreak for a multiline match if possible */ + + else if (startline) + { + if (current_subject > md->start_subject + start_offset) + { +#ifdef SUPPORT_UTF + if (utf) + { + while (current_subject < end_subject && + !WAS_NEWLINE(current_subject)) + { + current_subject++; + ACROSSCHAR(current_subject < end_subject, *current_subject, + current_subject++); + } + } + else +#endif + while (current_subject < end_subject && !WAS_NEWLINE(current_subject)) + current_subject++; + + /* If we have just passed a CR and the newline option is ANY or + ANYCRLF, and we are now at a LF, advance the match position by one + more character. */ + + if (current_subject[-1] == CHAR_CR && + (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) && + current_subject < end_subject && + *current_subject == CHAR_NL) + current_subject++; + } + } + + /* Or to a non-unique first char after study */ + + else if (start_bits != NULL) + { + while (current_subject < end_subject) + { + register unsigned int c = *current_subject; +#ifndef COMPILE_PCRE8 + if (c > 255) c = 255; +#endif + if ((start_bits[c/8] & (1 << (c&7))) == 0) + { + current_subject++; +#if defined SUPPORT_UTF && defined COMPILE_PCRE8 + /* In non 8-bit mode, the iteration will stop for + characters > 255 at the beginning or not stop at all. */ + if (utf) + ACROSSCHAR(current_subject < end_subject, *current_subject, + current_subject++); +#endif + } + else break; + } + } + } + + /* Restore fudged end_subject */ + + end_subject = save_end_subject; + + /* The following two optimizations are disabled for partial matching or if + disabling is explicitly requested (and of course, by the test above, this + code is not obeyed when restarting after a partial match). */ + + if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0 && + (options & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT)) == 0) + { + /* If the pattern was studied, a minimum subject length may be set. This + is a lower bound; no actual string of that length may actually match the + pattern. Although the value is, strictly, in characters, we treat it as + bytes to avoid spending too much time in this optimization. */ + + if (study != NULL && (study->flags & PCRE_STUDY_MINLEN) != 0 && + (pcre_uint32)(end_subject - current_subject) < study->minlength) + return PCRE_ERROR_NOMATCH; + + /* If req_char is set, we know that that character must appear in the + subject for the match to succeed. If the first character is set, req_char + must be later in the subject; otherwise the test starts at the match + point. This optimization can save a huge amount of work in patterns with + nested unlimited repeats that aren't going to match. Writing separate + code for cased/caseless versions makes it go faster, as does using an + autoincrement and backing off on a match. + + HOWEVER: when the subject string is very, very long, searching to its end + can take a long time, and give bad performance on quite ordinary + patterns. This showed up when somebody was matching /^C/ on a 32-megabyte + string... so we don't do this when the string is sufficiently long. */ + + if (has_req_char && end_subject - current_subject < REQ_BYTE_MAX) + { + register PCRE_PUCHAR p = current_subject + (has_first_char? 1:0); + + /* We don't need to repeat the search if we haven't yet reached the + place we found it at last time. */ + + if (p > req_char_ptr) + { + if (req_char != req_char2) + { + while (p < end_subject) + { + register int pp = *p++; + if (pp == req_char || pp == req_char2) { p--; break; } + } + } + else + { + while (p < end_subject) + { + if (*p++ == req_char) { p--; break; } + } + } + + /* If we can't find the required character, break the matching loop, + which will cause a return or PCRE_ERROR_NOMATCH. */ + + if (p >= end_subject) break; + + /* If we have found the required character, save the point where we + found it, so that we don't search again next time round the loop if + the start hasn't passed this character yet. */ + + req_char_ptr = p; + } + } + } + } /* End of optimizations that are done when not restarting */ + + /* OK, now we can do the business */ + + md->start_used_ptr = current_subject; + md->recursive = NULL; + + rc = internal_dfa_exec( + md, /* fixed match data */ + md->start_code, /* this subexpression's code */ + current_subject, /* where we currently are */ + start_offset, /* start offset in subject */ + offsets, /* offset vector */ + offsetcount, /* size of same */ + workspace, /* workspace vector */ + wscount, /* size of same */ + 0); /* function recurse level */ + + /* Anything other than "no match" means we are done, always; otherwise, carry + on only if not anchored. */ + + if (rc != PCRE_ERROR_NOMATCH || anchored) return rc; + + /* Advance to the next subject character unless we are at the end of a line + and firstline is set. */ + + if (firstline && IS_NEWLINE(current_subject)) break; + current_subject++; +#ifdef SUPPORT_UTF + if (utf) + { + ACROSSCHAR(current_subject < end_subject, *current_subject, + current_subject++); + } +#endif + if (current_subject > end_subject) break; + + /* If we have just passed a CR and we are now at a LF, and the pattern does + not contain any explicit matches for \r or \n, and the newline option is CRLF + or ANY or ANYCRLF, advance the match position by one more character. */ + + if (current_subject[-1] == CHAR_CR && + current_subject < end_subject && + *current_subject == CHAR_NL && (re->flags & PCRE_HASCRORLF) == 0 && (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF || diff --git a/pcre/pcre_exec.c b/pcre/pcre_exec.c --- a/pcre/pcre_exec.c +++ b/pcre/pcre_exec.c @@ -6,4948 +6,7144 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* This module contains pcre_exec(), the externally visible function that does -pattern matching using an NFA algorithm, trying to mimic Perl as closely as -possible. There are also some static supporting functions. */ - -#ifdef HAVE_CONFIG_H -#include "config.h" -#endif - -#define NLBLOCK md /* Block containing newline information */ -#define PSSTART start_subject /* Field containing processed string start */ -#define PSEND end_subject /* Field containing processed string end */ - -#include "pcre_internal.h" - -/* Undefine some potentially clashing cpp symbols */ - -#undef min -#undef max - -/* Flag bits for the match() function */ - -#define match_condassert 0x01 /* Called to check a condition assertion */ -#define match_cbegroup 0x02 /* Could-be-empty unlimited repeat group */ - -/* Non-error returns from the match() function. Error returns are externally -defined PCRE_ERROR_xxx codes, which are all negative. */ - -#define MATCH_MATCH 1 -#define MATCH_NOMATCH 0 - -/* Special internal returns from the match() function. Make them sufficiently -negative to avoid the external error codes. */ - -#define MATCH_COMMIT (-999) -#define MATCH_PRUNE (-998) -#define MATCH_SKIP (-997) -#define MATCH_THEN (-996) - -/* Maximum number of ints of offset to save on the stack for recursive calls. -If the offset vector is bigger, malloc is used. This should be a multiple of 3, -because the offset vector is always a multiple of 3 long. */ - -#define REC_STACK_SAVE_MAX 30 - -/* Min and max values for the common repeats; for the maxima, 0 => infinity */ - -static const char rep_min[] = { 0, 0, 1, 1, 0, 0 }; -static const char rep_max[] = { 0, 0, 0, 0, 1, 1 }; - - - -#ifdef DEBUG -/************************************************* -* Debugging function to print chars * -*************************************************/ - -/* Print a sequence of chars in printable format, stopping at the end of the -subject if the requested. - -Arguments: - p points to characters - length number to print - is_subject TRUE if printing from within md->start_subject - md pointer to matching data block, if is_subject is TRUE - -Returns: nothing -*/ - -static void -pchars(const uschar *p, int length, BOOL is_subject, match_data *md) -{ -unsigned int c; -if (is_subject && length > md->end_subject - p) length = md->end_subject - p; -while (length-- > 0) - if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c); -} -#endif - - - -/************************************************* -* Match a back-reference * -*************************************************/ - -/* If a back reference hasn't been set, the length that is passed is greater -than the number of characters left in the string, so the match fails. - -Arguments: - offset index into the offset vector - eptr points into the subject - length length to be matched - md points to match data block - ims the ims flags - -Returns: TRUE if matched -*/ - -static BOOL -match_ref(int offset, register USPTR eptr, int length, match_data *md, - unsigned long int ims) -{ -USPTR p = md->start_subject + md->offset_vector[offset]; - -#ifdef DEBUG -if (eptr >= md->end_subject) - printf("matching subject "); -else - { - printf("matching subject "); - pchars(eptr, length, TRUE, md); - } -printf(" against backref "); -pchars(p, length, FALSE, md); -printf("\n"); -#endif - -/* Always fail if not enough characters left */ - -if (length > md->end_subject - eptr) return FALSE; - -/* Separate the caselesss case for speed */ - -if ((ims & PCRE_CASELESS) != 0) - { - while (length-- > 0) - if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE; - } -else - { while (length-- > 0) if (*p++ != *eptr++) return FALSE; } - -return TRUE; -} - - - -/*************************************************************************** -**************************************************************************** - RECURSION IN THE match() FUNCTION - -The match() function is highly recursive, though not every recursive call -increases the recursive depth. Nevertheless, some regular expressions can cause -it to recurse to a great depth. I was writing for Unix, so I just let it call -itself recursively. This uses the stack for saving everything that has to be -saved for a recursive call. On Unix, the stack can be large, and this works -fine. - -It turns out that on some non-Unix-like systems there are problems with -programs that use a lot of stack. (This despite the fact that every last chip -has oodles of memory these days, and techniques for extending the stack have -been known for decades.) So.... - -There is a fudge, triggered by defining NO_RECURSE, which avoids recursive -calls by keeping local variables that need to be preserved in blocks of memory -obtained from malloc() instead instead of on the stack. Macros are used to -achieve this so that the actual code doesn't look very different to what it -always used to. - -The original heap-recursive code used longjmp(). However, it seems that this -can be very slow on some operating systems. Following a suggestion from Stan -Switzer, the use of longjmp() has been abolished, at the cost of having to -provide a unique number for each call to RMATCH. There is no way of generating -a sequence of numbers at compile time in C. I have given them names, to make -them stand out more clearly. - -Crude tests on x86 Linux show a small speedup of around 5-8%. However, on -FreeBSD, avoiding longjmp() more than halves the time taken to run the standard -tests. Furthermore, not using longjmp() means that local dynamic variables -don't have indeterminate values; this has meant that the frame size can be -reduced because the result can be "passed back" by straight setting of the -variable instead of being passed in the frame. -**************************************************************************** -***************************************************************************/ - -/* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN -below must be updated in sync. */ - -enum { RM1=1, RM2, RM3, RM4, RM5, RM6, RM7, RM8, RM9, RM10, - RM11, RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20, - RM21, RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30, - RM31, RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40, - RM41, RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50, - RM51, RM52, RM53, RM54 }; - -/* These versions of the macros use the stack, as normal. There are debugging -versions and production versions. Note that the "rw" argument of RMATCH isn't -actuall used in this definition. */ - -#ifndef NO_RECURSE -#define REGISTER register - -#ifdef DEBUG -#define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \ - { \ - printf("match() called in line %d\n", __LINE__); \ - rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1); \ - printf("to line %d\n", __LINE__); \ - } -#define RRETURN(ra) \ - { \ - printf("match() returned %d from line %d ", ra, __LINE__); \ - return ra; \ - } -#else -#define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \ - rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1) -#define RRETURN(ra) return ra -#endif - -#else - - -/* These versions of the macros manage a private stack on the heap. Note that -the "rd" argument of RMATCH isn't actually used in this definition. It's the md -argument of match(), which never changes. */ - -#define REGISTER - -#define RMATCH(ra,rb,rc,rd,re,rf,rg,rw)\ - {\ - heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\ - frame->Xwhere = rw; \ - newframe->Xeptr = ra;\ - newframe->Xecode = rb;\ - newframe->Xmstart = mstart;\ - newframe->Xoffset_top = rc;\ - newframe->Xims = re;\ - newframe->Xeptrb = rf;\ - newframe->Xflags = rg;\ - newframe->Xrdepth = frame->Xrdepth + 1;\ - newframe->Xprevframe = frame;\ - frame = newframe;\ - DPRINTF(("restarting from line %d\n", __LINE__));\ - goto HEAP_RECURSE;\ - L_##rw:\ - DPRINTF(("jumped back to line %d\n", __LINE__));\ - } - -#define RRETURN(ra)\ - {\ - heapframe *newframe = frame;\ - frame = newframe->Xprevframe;\ - (pcre_stack_free)(newframe);\ - if (frame != NULL)\ - {\ - rrc = ra;\ - goto HEAP_RETURN;\ - }\ - return ra;\ - } - - -/* Structure for remembering the local variables in a private frame */ - -typedef struct heapframe { - struct heapframe *Xprevframe; - - /* Function arguments that may change */ - - const uschar *Xeptr; - const uschar *Xecode; - const uschar *Xmstart; - int Xoffset_top; - long int Xims; - eptrblock *Xeptrb; - int Xflags; - unsigned int Xrdepth; - - /* Function local variables */ - - const uschar *Xcallpat; - const uschar *Xcharptr; - const uschar *Xdata; - const uschar *Xnext; - const uschar *Xpp; - const uschar *Xprev; - const uschar *Xsaved_eptr; - - recursion_info Xnew_recursive; - - BOOL Xcur_is_word; - BOOL Xcondition; - BOOL Xprev_is_word; - - unsigned long int Xoriginal_ims; - -#ifdef SUPPORT_UCP - int Xprop_type; - int Xprop_value; - int Xprop_fail_result; - int Xprop_category; - int Xprop_chartype; - int Xprop_script; - int Xoclength; - uschar Xocchars[8]; -#endif - - int Xctype; - unsigned int Xfc; - int Xfi; - int Xlength; - int Xmax; - int Xmin; - int Xnumber; - int Xoffset; - int Xop; - int Xsave_capture_last; - int Xsave_offset1, Xsave_offset2, Xsave_offset3; - int Xstacksave[REC_STACK_SAVE_MAX]; - - eptrblock Xnewptrb; - - /* Where to jump back to */ - - int Xwhere; - -} heapframe; - -#endif - - -/*************************************************************************** -***************************************************************************/ - - - -/************************************************* -* Match from current position * -*************************************************/ - -/* This function is called recursively in many circumstances. Whenever it -returns a negative (error) response, the outer incarnation must also return the -same response. - -Performance note: It might be tempting to extract commonly used fields from the -md structure (e.g. utf8, end_subject) into individual variables to improve -performance. Tests using gcc on a SPARC disproved this; in the first case, it -made performance worse. - -Arguments: - eptr pointer to current character in subject - ecode pointer to current position in compiled code - mstart pointer to the current match start position (can be modified - by encountering \K) - offset_top current top pointer - md pointer to "static" info for the match - ims current /i, /m, and /s options - eptrb pointer to chain of blocks containing eptr at start of - brackets - for testing for empty matches - flags can contain - match_condassert - this is an assertion condition - match_cbegroup - this is the start of an unlimited repeat - group that can match an empty string - rdepth the recursion depth - -Returns: MATCH_MATCH if matched ) these values are >= 0 - MATCH_NOMATCH if failed to match ) - a negative PCRE_ERROR_xxx value if aborted by an error condition - (e.g. stopped by repeated call or recursion limit) -*/ - -static int -match(REGISTER USPTR eptr, REGISTER const uschar *ecode, const uschar *mstart, - int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb, - int flags, unsigned int rdepth) -{ -/* These variables do not need to be preserved over recursion in this function, -so they can be ordinary variables in all cases. Mark some of them with -"register" because they are used a lot in loops. */ - -register int rrc; /* Returns from recursive calls */ -register int i; /* Used for loops not involving calls to RMATCH() */ -register unsigned int c; /* Character values not kept over RMATCH() calls */ -register BOOL utf8; /* Local copy of UTF-8 flag for speed */ - -BOOL minimize, possessive; /* Quantifier options */ - -/* When recursion is not being used, all "local" variables that have to be -preserved over calls to RMATCH() are part of a "frame" which is obtained from -heap storage. Set up the top-level frame here; others are obtained from the -heap whenever RMATCH() does a "recursion". See the macro definitions above. */ - -#ifdef NO_RECURSE -heapframe *frame = (pcre_stack_malloc)(sizeof(heapframe)); -frame->Xprevframe = NULL; /* Marks the top level */ - -/* Copy in the original argument variables */ - -frame->Xeptr = eptr; -frame->Xecode = ecode; -frame->Xmstart = mstart; -frame->Xoffset_top = offset_top; -frame->Xims = ims; -frame->Xeptrb = eptrb; -frame->Xflags = flags; -frame->Xrdepth = rdepth; - -/* This is where control jumps back to to effect "recursion" */ - -HEAP_RECURSE: - -/* Macros make the argument variables come from the current frame */ - -#define eptr frame->Xeptr -#define ecode frame->Xecode -#define mstart frame->Xmstart -#define offset_top frame->Xoffset_top -#define ims frame->Xims -#define eptrb frame->Xeptrb -#define flags frame->Xflags -#define rdepth frame->Xrdepth - -/* Ditto for the local variables */ - -#ifdef SUPPORT_UTF8 -#define charptr frame->Xcharptr -#endif -#define callpat frame->Xcallpat -#define data frame->Xdata -#define next frame->Xnext -#define pp frame->Xpp -#define prev frame->Xprev -#define saved_eptr frame->Xsaved_eptr - -#define new_recursive frame->Xnew_recursive - -#define cur_is_word frame->Xcur_is_word -#define condition frame->Xcondition -#define prev_is_word frame->Xprev_is_word - -#define original_ims frame->Xoriginal_ims - -#ifdef SUPPORT_UCP -#define prop_type frame->Xprop_type -#define prop_value frame->Xprop_value -#define prop_fail_result frame->Xprop_fail_result -#define prop_category frame->Xprop_category -#define prop_chartype frame->Xprop_chartype -#define prop_script frame->Xprop_script -#define oclength frame->Xoclength -#define occhars frame->Xocchars -#endif - -#define ctype frame->Xctype -#define fc frame->Xfc -#define fi frame->Xfi -#define length frame->Xlength -#define max frame->Xmax -#define min frame->Xmin -#define number frame->Xnumber -#define offset frame->Xoffset -#define op frame->Xop -#define save_capture_last frame->Xsave_capture_last -#define save_offset1 frame->Xsave_offset1 -#define save_offset2 frame->Xsave_offset2 -#define save_offset3 frame->Xsave_offset3 -#define stacksave frame->Xstacksave - -#define newptrb frame->Xnewptrb - -/* When recursion is being used, local variables are allocated on the stack and -get preserved during recursion in the normal way. In this environment, fi and -i, and fc and c, can be the same variables. */ - -#else /* NO_RECURSE not defined */ -#define fi i -#define fc c - - -#ifdef SUPPORT_UTF8 /* Many of these variables are used only */ -const uschar *charptr; /* in small blocks of the code. My normal */ -#endif /* style of coding would have declared */ -const uschar *callpat; /* them within each of those blocks. */ -const uschar *data; /* However, in order to accommodate the */ -const uschar *next; /* version of this code that uses an */ -USPTR pp; /* external "stack" implemented on the */ -const uschar *prev; /* heap, it is easier to declare them all */ -USPTR saved_eptr; /* here, so the declarations can be cut */ - /* out in a block. The only declarations */ -recursion_info new_recursive; /* within blocks below are for variables */ - /* that do not have to be preserved over */ -BOOL cur_is_word; /* a recursive call to RMATCH(). */ -BOOL condition; -BOOL prev_is_word; - -unsigned long int original_ims; - -#ifdef SUPPORT_UCP -int prop_type; -int prop_value; -int prop_fail_result; -int prop_category; -int prop_chartype; -int prop_script; -int oclength; -uschar occhars[8]; -#endif - -int ctype; -int length; -int max; -int min; -int number; -int offset; -int op; -int save_capture_last; -int save_offset1, save_offset2, save_offset3; -int stacksave[REC_STACK_SAVE_MAX]; - -eptrblock newptrb; -#endif /* NO_RECURSE */ - -/* These statements are here to stop the compiler complaining about unitialized -variables. */ - -#ifdef SUPPORT_UCP -prop_value = 0; -prop_fail_result = 0; -#endif - - -/* This label is used for tail recursion, which is used in a few cases even -when NO_RECURSE is not defined, in order to reduce the amount of stack that is -used. Thanks to Ian Taylor for noticing this possibility and sending the -original patch. */ - -TAIL_RECURSE: - -/* OK, now we can get on with the real code of the function. Recursive calls -are specified by the macro RMATCH and RRETURN is used to return. When -NO_RECURSE is *not* defined, these just turn into a recursive call to match() -and a "return", respectively (possibly with some debugging if DEBUG is -defined). However, RMATCH isn't like a function call because it's quite a -complicated macro. It has to be used in one particular way. This shouldn't, -however, impact performance when true recursion is being used. */ - -#ifdef SUPPORT_UTF8 -utf8 = md->utf8; /* Local copy of the flag */ -#else -utf8 = FALSE; -#endif - -/* First check that we haven't called match() too many times, or that we -haven't exceeded the recursive call limit. */ - -if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT); -if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT); - -original_ims = ims; /* Save for resetting on ')' */ - -/* At the start of a group with an unlimited repeat that may match an empty -string, the match_cbegroup flag is set. When this is the case, add the current -subject pointer to the chain of such remembered pointers, to be checked when we -hit the closing ket, in order to break infinite loops that match no characters. -When match() is called in other circumstances, don't add to the chain. The -match_cbegroup flag must NOT be used with tail recursion, because the memory -block that is used is on the stack, so a new one may be required for each -match(). */ - -if ((flags & match_cbegroup) != 0) - { - newptrb.epb_saved_eptr = eptr; - newptrb.epb_prev = eptrb; - eptrb = &newptrb; - } - -/* Now start processing the opcodes. */ - -for (;;) - { - minimize = possessive = FALSE; - op = *ecode; - - /* For partial matching, remember if we ever hit the end of the subject after - matching at least one subject character. */ - - if (md->partial && - eptr >= md->end_subject && - eptr > mstart) - md->hitend = TRUE; - - switch(op) - { - case OP_FAIL: - RRETURN(MATCH_NOMATCH); - - case OP_PRUNE: - RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, - ims, eptrb, flags, RM51); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - RRETURN(MATCH_PRUNE); - - case OP_COMMIT: - RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, - ims, eptrb, flags, RM52); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - RRETURN(MATCH_COMMIT); - - case OP_SKIP: - RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, - ims, eptrb, flags, RM53); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - md->start_match_ptr = eptr; /* Pass back current position */ - RRETURN(MATCH_SKIP); - - case OP_THEN: - RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, - ims, eptrb, flags, RM54); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - RRETURN(MATCH_THEN); - - /* Handle a capturing bracket. If there is space in the offset vector, save - the current subject position in the working slot at the top of the vector. - We mustn't change the current values of the data slot, because they may be - set from a previous iteration of this group, and be referred to by a - reference inside the group. - - If the bracket fails to match, we need to restore this value and also the - values of the final offsets, in case they were set by a previous iteration - of the same bracket. - - If there isn't enough space in the offset vector, treat this as if it were - a non-capturing bracket. Don't worry about setting the flag for the error - case here; that is handled in the code for KET. */ - - case OP_CBRA: - case OP_SCBRA: - number = GET2(ecode, 1+LINK_SIZE); - offset = number << 1; - -#ifdef DEBUG - printf("start bracket %d\n", number); - printf("subject="); - pchars(eptr, 16, TRUE, md); - printf("\n"); -#endif - - if (offset < md->offset_max) - { - save_offset1 = md->offset_vector[offset]; - save_offset2 = md->offset_vector[offset+1]; - save_offset3 = md->offset_vector[md->offset_end - number]; - save_capture_last = md->capture_last; - - DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3)); - md->offset_vector[md->offset_end - number] = eptr - md->start_subject; - - flags = (op == OP_SCBRA)? match_cbegroup : 0; - do - { - RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, - ims, eptrb, flags, RM1); - if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); - md->capture_last = save_capture_last; - ecode += GET(ecode, 1); - } - while (*ecode == OP_ALT); - - DPRINTF(("bracket %d failed\n", number)); - - md->offset_vector[offset] = save_offset1; - md->offset_vector[offset+1] = save_offset2; - md->offset_vector[md->offset_end - number] = save_offset3; - - RRETURN(MATCH_NOMATCH); - } - - /* FALL THROUGH ... Insufficient room for saving captured contents. Treat - as a non-capturing bracket. */ - - /* VVVVVVVVVVVVVVVVVVVVVVVVV */ - /* VVVVVVVVVVVVVVVVVVVVVVVVV */ - - DPRINTF(("insufficient capture room: treat as non-capturing\n")); - - /* VVVVVVVVVVVVVVVVVVVVVVVVV */ - /* VVVVVVVVVVVVVVVVVVVVVVVVV */ - - /* Non-capturing bracket. Loop for all the alternatives. When we get to the - final alternative within the brackets, we would return the result of a - recursive call to match() whatever happened. We can reduce stack usage by - turning this into a tail recursion, except in the case when match_cbegroup - is set.*/ - - case OP_BRA: - case OP_SBRA: - DPRINTF(("start non-capturing bracket\n")); - flags = (op >= OP_SBRA)? match_cbegroup : 0; - for (;;) - { - if (ecode[GET(ecode, 1)] != OP_ALT) /* Final alternative */ - { - if (flags == 0) /* Not a possibly empty group */ - { - ecode += _pcre_OP_lengths[*ecode]; - DPRINTF(("bracket 0 tail recursion\n")); - goto TAIL_RECURSE; - } - - /* Possibly empty group; can't use tail recursion. */ - - RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims, - eptrb, flags, RM48); - RRETURN(rrc); - } - - /* For non-final alternatives, continue the loop for a NOMATCH result; - otherwise return. */ - - RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims, - eptrb, flags, RM2); - if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); - ecode += GET(ecode, 1); - } - /* Control never reaches here. */ - - /* Conditional group: compilation checked that there are no more than - two branches. If the condition is false, skipping the first branch takes us - past the end if there is only one branch, but that's OK because that is - exactly what going to the ket would do. As there is only one branch to be - obeyed, we can use tail recursion to avoid using another stack frame. */ - - case OP_COND: - case OP_SCOND: - if (ecode[LINK_SIZE+1] == OP_RREF) /* Recursion test */ - { - offset = GET2(ecode, LINK_SIZE + 2); /* Recursion group number*/ - condition = md->recursive != NULL && - (offset == RREF_ANY || offset == md->recursive->group_num); - ecode += condition? 3 : GET(ecode, 1); - } - - else if (ecode[LINK_SIZE+1] == OP_CREF) /* Group used test */ - { - offset = GET2(ecode, LINK_SIZE+2) << 1; /* Doubled ref number */ - condition = offset < offset_top && md->offset_vector[offset] >= 0; - ecode += condition? 3 : GET(ecode, 1); - } - - else if (ecode[LINK_SIZE+1] == OP_DEF) /* DEFINE - always false */ - { - condition = FALSE; - ecode += GET(ecode, 1); - } - - /* The condition is an assertion. Call match() to evaluate it - setting - the final argument match_condassert causes it to stop at the end of an - assertion. */ - - else - { - RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, - match_condassert, RM3); - if (rrc == MATCH_MATCH) - { - condition = TRUE; - ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2); - while (*ecode == OP_ALT) ecode += GET(ecode, 1); - } - else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) - { - RRETURN(rrc); /* Need braces because of following else */ - } - else - { - condition = FALSE; - ecode += GET(ecode, 1); - } - } - - /* We are now at the branch that is to be obeyed. As there is only one, - we can use tail recursion to avoid using another stack frame, except when - match_cbegroup is required for an unlimited repeat of a possibly empty - group. If the second alternative doesn't exist, we can just plough on. */ - - if (condition || *ecode == OP_ALT) - { - ecode += 1 + LINK_SIZE; - if (op == OP_SCOND) /* Possibly empty group */ - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49); - RRETURN(rrc); - } - else /* Group must match something */ - { - flags = 0; - goto TAIL_RECURSE; - } - } - else /* Condition false & no 2nd alternative */ - { - ecode += 1 + LINK_SIZE; - } - break; - - - /* End of the pattern, either real or forced. If we are in a top-level - recursion, we should restore the offsets appropriately and continue from - after the call. */ - - case OP_ACCEPT: - case OP_END: - if (md->recursive != NULL && md->recursive->group_num == 0) - { - recursion_info *rec = md->recursive; - DPRINTF(("End of pattern in a (?0) recursion\n")); - md->recursive = rec->prevrec; - memmove(md->offset_vector, rec->offset_save, - rec->saved_max * sizeof(int)); - mstart = rec->save_start; - ims = original_ims; - ecode = rec->after_call; - break; - } - - /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty - string - backtracking will then try other alternatives, if any. */ - - if (md->notempty && eptr == mstart) RRETURN(MATCH_NOMATCH); - md->end_match_ptr = eptr; /* Record where we ended */ - md->end_offset_top = offset_top; /* and how many extracts were taken */ - md->start_match_ptr = mstart; /* and the start (\K can modify) */ - RRETURN(MATCH_MATCH); - - /* Change option settings */ - - case OP_OPT: - ims = ecode[1]; - ecode += 2; - DPRINTF(("ims set to %02lx\n", ims)); - break; - - /* Assertion brackets. Check the alternative branches in turn - the - matching won't pass the KET for an assertion. If any one branch matches, - the assertion is true. Lookbehind assertions have an OP_REVERSE item at the - start of each branch to move the current point backwards, so the code at - this level is identical to the lookahead case. */ - - case OP_ASSERT: - case OP_ASSERTBACK: - do - { - RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0, - RM4); - if (rrc == MATCH_MATCH) break; - if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); - ecode += GET(ecode, 1); - } - while (*ecode == OP_ALT); - if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH); - - /* If checking an assertion for a condition, return MATCH_MATCH. */ - - if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH); - - /* Continue from after the assertion, updating the offsets high water - mark, since extracts may have been taken during the assertion. */ - - do ecode += GET(ecode,1); while (*ecode == OP_ALT); - ecode += 1 + LINK_SIZE; - offset_top = md->end_offset_top; - continue; - - /* Negative assertion: all branches must fail to match */ - - case OP_ASSERT_NOT: - case OP_ASSERTBACK_NOT: - do - { - RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0, - RM5); - if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH); - if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); - ecode += GET(ecode,1); - } - while (*ecode == OP_ALT); - - if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH); - - ecode += 1 + LINK_SIZE; - continue; - - /* Move the subject pointer back. This occurs only at the start of - each branch of a lookbehind assertion. If we are too close to the start to - move back, this match function fails. When working with UTF-8 we move - back a number of characters, not bytes. */ - - case OP_REVERSE: -#ifdef SUPPORT_UTF8 - if (utf8) - { - i = GET(ecode, 1); - while (i-- > 0) - { - eptr--; - if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH); - BACKCHAR(eptr); - } - } - else -#endif - - /* No UTF-8 support, or not in UTF-8 mode: count is byte count */ - - { - eptr -= GET(ecode, 1); - if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH); - } - - /* Skip to next op code */ - - ecode += 1 + LINK_SIZE; - break; - - /* The callout item calls an external function, if one is provided, passing - details of the match so far. This is mainly for debugging, though the - function is able to force a failure. */ - - case OP_CALLOUT: - if (pcre_callout != NULL) - { - pcre_callout_block cb; - cb.version = 1; /* Version 1 of the callout block */ - cb.callout_number = ecode[1]; - cb.offset_vector = md->offset_vector; - cb.subject = (PCRE_SPTR)md->start_subject; - cb.subject_length = md->end_subject - md->start_subject; - cb.start_match = mstart - md->start_subject; - cb.current_position = eptr - md->start_subject; - cb.pattern_position = GET(ecode, 2); - cb.next_item_length = GET(ecode, 2 + LINK_SIZE); - cb.capture_top = offset_top/2; - cb.capture_last = md->capture_last; - cb.callout_data = md->callout_data; - if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH); - if (rrc < 0) RRETURN(rrc); - } - ecode += 2 + 2*LINK_SIZE; - break; - - /* Recursion either matches the current regex, or some subexpression. The - offset data is the offset to the starting bracket from the start of the - whole pattern. (This is so that it works from duplicated subpatterns.) - - If there are any capturing brackets started but not finished, we have to - save their starting points and reinstate them after the recursion. However, - we don't know how many such there are (offset_top records the completed - total) so we just have to save all the potential data. There may be up to - 65535 such values, which is too large to put on the stack, but using malloc - for small numbers seems expensive. As a compromise, the stack is used when - there are no more than REC_STACK_SAVE_MAX values to store; otherwise malloc - is used. A problem is what to do if the malloc fails ... there is no way of - returning to the top level with an error. Save the top REC_STACK_SAVE_MAX - values on the stack, and accept that the rest may be wrong. - - There are also other values that have to be saved. We use a chained - sequence of blocks that actually live on the stack. Thanks to Robin Houston - for the original version of this logic. */ - - case OP_RECURSE: - { - callpat = md->start_code + GET(ecode, 1); - new_recursive.group_num = (callpat == md->start_code)? 0 : - GET2(callpat, 1 + LINK_SIZE); - - /* Add to "recursing stack" */ - - new_recursive.prevrec = md->recursive; - md->recursive = &new_recursive; - - /* Find where to continue from afterwards */ - - ecode += 1 + LINK_SIZE; - new_recursive.after_call = ecode; - - /* Now save the offset data. */ - - new_recursive.saved_max = md->offset_end; - if (new_recursive.saved_max <= REC_STACK_SAVE_MAX) - new_recursive.offset_save = stacksave; - else - { - new_recursive.offset_save = - (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int)); - if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY); - } - - memcpy(new_recursive.offset_save, md->offset_vector, - new_recursive.saved_max * sizeof(int)); - new_recursive.save_start = mstart; - mstart = eptr; - - /* OK, now we can do the recursion. For each top-level alternative we - restore the offset and recursion data. */ - - DPRINTF(("Recursing into group %d\n", new_recursive.group_num)); - flags = (*callpat >= OP_SBRA)? match_cbegroup : 0; - do - { - RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top, - md, ims, eptrb, flags, RM6); - if (rrc == MATCH_MATCH) - { - DPRINTF(("Recursion matched\n")); - md->recursive = new_recursive.prevrec; - if (new_recursive.offset_save != stacksave) - (pcre_free)(new_recursive.offset_save); - RRETURN(MATCH_MATCH); - } - else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) - { - DPRINTF(("Recursion gave error %d\n", rrc)); - RRETURN(rrc); - } - - md->recursive = &new_recursive; - memcpy(md->offset_vector, new_recursive.offset_save, - new_recursive.saved_max * sizeof(int)); - callpat += GET(callpat, 1); - } - while (*callpat == OP_ALT); - - DPRINTF(("Recursion didn't match\n")); - md->recursive = new_recursive.prevrec; - if (new_recursive.offset_save != stacksave) - (pcre_free)(new_recursive.offset_save); - RRETURN(MATCH_NOMATCH); - } - /* Control never reaches here */ - - /* "Once" brackets are like assertion brackets except that after a match, - the point in the subject string is not moved back. Thus there can never be - a move back into the brackets. Friedl calls these "atomic" subpatterns. - Check the alternative branches in turn - the matching won't pass the KET - for this kind of subpattern. If any one branch matches, we carry on as at - the end of a normal bracket, leaving the subject pointer. */ - - case OP_ONCE: - prev = ecode; - saved_eptr = eptr; - - do - { - RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7); - if (rrc == MATCH_MATCH) break; - if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); - ecode += GET(ecode,1); - } - while (*ecode == OP_ALT); - - /* If hit the end of the group (which could be repeated), fail */ - - if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH); - - /* Continue as from after the assertion, updating the offsets high water - mark, since extracts may have been taken. */ - - do ecode += GET(ecode, 1); while (*ecode == OP_ALT); - - offset_top = md->end_offset_top; - eptr = md->end_match_ptr; - - /* For a non-repeating ket, just continue at this level. This also - happens for a repeating ket if no characters were matched in the group. - This is the forcible breaking of infinite loops as implemented in Perl - 5.005. If there is an options reset, it will get obeyed in the normal - course of events. */ - - if (*ecode == OP_KET || eptr == saved_eptr) - { - ecode += 1+LINK_SIZE; - break; - } - - /* The repeating kets try the rest of the pattern or restart from the - preceding bracket, in the appropriate order. The second "call" of match() - uses tail recursion, to avoid using another stack frame. We need to reset - any options that changed within the bracket before re-running it, so - check the next opcode. */ - - if (ecode[1+LINK_SIZE] == OP_OPT) - { - ims = (ims & ~PCRE_IMS) | ecode[4]; - DPRINTF(("ims set to %02lx at group repeat\n", ims)); - } - - if (*ecode == OP_KETRMIN) - { - RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - ecode = prev; - flags = 0; - goto TAIL_RECURSE; - } - else /* OP_KETRMAX */ - { - RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - ecode += 1 + LINK_SIZE; - flags = 0; - goto TAIL_RECURSE; - } - /* Control never gets here */ - - /* An alternation is the end of a branch; scan along to find the end of the - bracketed group and go to there. */ - - case OP_ALT: - do ecode += GET(ecode,1); while (*ecode == OP_ALT); - break; - - /* BRAZERO, BRAMINZERO and SKIPZERO occur just before a bracket group, - indicating that it may occur zero times. It may repeat infinitely, or not - at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets - with fixed upper repeat limits are compiled as a number of copies, with the - optional ones preceded by BRAZERO or BRAMINZERO. */ - - case OP_BRAZERO: - { - next = ecode+1; - RMATCH(eptr, next, offset_top, md, ims, eptrb, 0, RM10); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - do next += GET(next,1); while (*next == OP_ALT); - ecode = next + 1 + LINK_SIZE; - } - break; - - case OP_BRAMINZERO: - { - next = ecode+1; - do next += GET(next, 1); while (*next == OP_ALT); - RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0, RM11); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - ecode++; - } - break; - - case OP_SKIPZERO: - { - next = ecode+1; - do next += GET(next,1); while (*next == OP_ALT); - ecode = next + 1 + LINK_SIZE; - } - break; - - /* End of a group, repeated or non-repeating. */ - - case OP_KET: - case OP_KETRMIN: - case OP_KETRMAX: - prev = ecode - GET(ecode, 1); - - /* If this was a group that remembered the subject start, in order to break - infinite repeats of empty string matches, retrieve the subject start from - the chain. Otherwise, set it NULL. */ - - if (*prev >= OP_SBRA) - { - saved_eptr = eptrb->epb_saved_eptr; /* Value at start of group */ - eptrb = eptrb->epb_prev; /* Backup to previous group */ - } - else saved_eptr = NULL; - - /* If we are at the end of an assertion group, stop matching and return - MATCH_MATCH, but record the current high water mark for use by positive - assertions. Do this also for the "once" (atomic) groups. */ - - if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT || - *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT || - *prev == OP_ONCE) - { - md->end_match_ptr = eptr; /* For ONCE */ - md->end_offset_top = offset_top; - RRETURN(MATCH_MATCH); - } - - /* For capturing groups we have to check the group number back at the start - and if necessary complete handling an extraction by setting the offsets and - bumping the high water mark. Note that whole-pattern recursion is coded as - a recurse into group 0, so it won't be picked up here. Instead, we catch it - when the OP_END is reached. Other recursion is handled here. */ - - if (*prev == OP_CBRA || *prev == OP_SCBRA) - { - number = GET2(prev, 1+LINK_SIZE); - offset = number << 1; - -#ifdef DEBUG - printf("end bracket %d", number); - printf("\n"); -#endif - - md->capture_last = number; - if (offset >= md->offset_max) md->offset_overflow = TRUE; else - { - md->offset_vector[offset] = - md->offset_vector[md->offset_end - number]; - md->offset_vector[offset+1] = eptr - md->start_subject; - if (offset_top <= offset) offset_top = offset + 2; - } - - /* Handle a recursively called group. Restore the offsets - appropriately and continue from after the call. */ - - if (md->recursive != NULL && md->recursive->group_num == number) - { - recursion_info *rec = md->recursive; - DPRINTF(("Recursion (%d) succeeded - continuing\n", number)); - md->recursive = rec->prevrec; - mstart = rec->save_start; - memcpy(md->offset_vector, rec->offset_save, - rec->saved_max * sizeof(int)); - ecode = rec->after_call; - ims = original_ims; - break; - } - } - - /* For both capturing and non-capturing groups, reset the value of the ims - flags, in case they got changed during the group. */ - - ims = original_ims; - DPRINTF(("ims reset to %02lx\n", ims)); - - /* For a non-repeating ket, just continue at this level. This also - happens for a repeating ket if no characters were matched in the group. - This is the forcible breaking of infinite loops as implemented in Perl - 5.005. If there is an options reset, it will get obeyed in the normal - course of events. */ - - if (*ecode == OP_KET || eptr == saved_eptr) - { - ecode += 1 + LINK_SIZE; - break; - } - - /* The repeating kets try the rest of the pattern or restart from the - preceding bracket, in the appropriate order. In the second case, we can use - tail recursion to avoid using another stack frame, unless we have an - unlimited repeat of a group that can match an empty string. */ - - flags = (*prev >= OP_SBRA)? match_cbegroup : 0; - - if (*ecode == OP_KETRMIN) - { - RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (flags != 0) /* Could match an empty string */ - { - RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50); - RRETURN(rrc); - } - ecode = prev; - goto TAIL_RECURSE; - } - else /* OP_KETRMAX */ - { - RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - ecode += 1 + LINK_SIZE; - flags = 0; - goto TAIL_RECURSE; - } - /* Control never gets here */ - - /* Start of subject unless notbol, or after internal newline if multiline */ - - case OP_CIRC: - if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH); - if ((ims & PCRE_MULTILINE) != 0) - { - if (eptr != md->start_subject && - (eptr == md->end_subject || !WAS_NEWLINE(eptr))) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - } - /* ... else fall through */ - - /* Start of subject assertion */ - - case OP_SOD: - if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH); - ecode++; - break; - - /* Start of match assertion */ - - case OP_SOM: - if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH); - ecode++; - break; - - /* Reset the start of match point */ - - case OP_SET_SOM: - mstart = eptr; - ecode++; - break; - - /* Assert before internal newline if multiline, or before a terminating - newline unless endonly is set, else end of subject unless noteol is set. */ - - case OP_DOLL: - if ((ims & PCRE_MULTILINE) != 0) - { - if (eptr < md->end_subject) - { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); } - else - { if (md->noteol) RRETURN(MATCH_NOMATCH); } - ecode++; - break; - } - else - { - if (md->noteol) RRETURN(MATCH_NOMATCH); - if (!md->endonly) - { - if (eptr != md->end_subject && - (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen)) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - } - } - /* ... else fall through for endonly */ - - /* End of subject assertion (\z) */ - - case OP_EOD: - if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH); - ecode++; - break; - - /* End of subject or ending \n assertion (\Z) */ - - case OP_EODN: - if (eptr != md->end_subject && - (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen)) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - - /* Word boundary assertions */ - - case OP_NOT_WORD_BOUNDARY: - case OP_WORD_BOUNDARY: - { - - /* Find out if the previous and current characters are "word" characters. - It takes a bit more work in UTF-8 mode. Characters > 255 are assumed to - be "non-word" characters. */ - -#ifdef SUPPORT_UTF8 - if (utf8) - { - if (eptr == md->start_subject) prev_is_word = FALSE; else - { - const uschar *lastptr = eptr - 1; - while((*lastptr & 0xc0) == 0x80) lastptr--; - GETCHAR(c, lastptr); - prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0; - } - if (eptr >= md->end_subject) cur_is_word = FALSE; else - { - GETCHAR(c, eptr); - cur_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0; - } - } - else -#endif - - /* More streamlined when not in UTF-8 mode */ - - { - prev_is_word = (eptr != md->start_subject) && - ((md->ctypes[eptr[-1]] & ctype_word) != 0); - cur_is_word = (eptr < md->end_subject) && - ((md->ctypes[*eptr] & ctype_word) != 0); - } - - /* Now see if the situation is what we want */ - - if ((*ecode++ == OP_WORD_BOUNDARY)? - cur_is_word == prev_is_word : cur_is_word != prev_is_word) - RRETURN(MATCH_NOMATCH); - } - break; - - /* Match a single character type; inline for speed */ - - case OP_ANY: - if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); - /* Fall through */ - - case OP_ALLANY: - if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH); - if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; - ecode++; - break; - - /* Match a single byte, even in UTF-8 mode. This opcode really does match - any byte, even newline, independent of the setting of PCRE_DOTALL. */ - - case OP_ANYBYTE: - if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_NOT_DIGIT: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - if ( -#ifdef SUPPORT_UTF8 - c < 256 && -#endif - (md->ctypes[c] & ctype_digit) != 0 - ) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_DIGIT: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - if ( -#ifdef SUPPORT_UTF8 - c >= 256 || -#endif - (md->ctypes[c] & ctype_digit) == 0 - ) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_NOT_WHITESPACE: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - if ( -#ifdef SUPPORT_UTF8 - c < 256 && -#endif - (md->ctypes[c] & ctype_space) != 0 - ) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_WHITESPACE: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - if ( -#ifdef SUPPORT_UTF8 - c >= 256 || -#endif - (md->ctypes[c] & ctype_space) == 0 - ) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_NOT_WORDCHAR: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - if ( -#ifdef SUPPORT_UTF8 - c < 256 && -#endif - (md->ctypes[c] & ctype_word) != 0 - ) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_WORDCHAR: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - if ( -#ifdef SUPPORT_UTF8 - c >= 256 || -#endif - (md->ctypes[c] & ctype_word) == 0 - ) - RRETURN(MATCH_NOMATCH); - ecode++; - break; - - case OP_ANYNL: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - switch(c) - { - default: RRETURN(MATCH_NOMATCH); - case 0x000d: - if (eptr < md->end_subject && *eptr == 0x0a) eptr++; - break; - - case 0x000a: - break; - - case 0x000b: - case 0x000c: - case 0x0085: - case 0x2028: - case 0x2029: - if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); - break; - } - ecode++; - break; - - case OP_NOT_HSPACE: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - switch(c) - { - default: break; - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - case 0x1680: /* OGHAM SPACE MARK */ - case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ - case 0x2000: /* EN QUAD */ - case 0x2001: /* EM QUAD */ - case 0x2002: /* EN SPACE */ - case 0x2003: /* EM SPACE */ - case 0x2004: /* THREE-PER-EM SPACE */ - case 0x2005: /* FOUR-PER-EM SPACE */ - case 0x2006: /* SIX-PER-EM SPACE */ - case 0x2007: /* FIGURE SPACE */ - case 0x2008: /* PUNCTUATION SPACE */ - case 0x2009: /* THIN SPACE */ - case 0x200A: /* HAIR SPACE */ - case 0x202f: /* NARROW NO-BREAK SPACE */ - case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ - case 0x3000: /* IDEOGRAPHIC SPACE */ - RRETURN(MATCH_NOMATCH); - } - ecode++; - break; - - case OP_HSPACE: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - switch(c) - { - default: RRETURN(MATCH_NOMATCH); - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - case 0x1680: /* OGHAM SPACE MARK */ - case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ - case 0x2000: /* EN QUAD */ - case 0x2001: /* EM QUAD */ - case 0x2002: /* EN SPACE */ - case 0x2003: /* EM SPACE */ - case 0x2004: /* THREE-PER-EM SPACE */ - case 0x2005: /* FOUR-PER-EM SPACE */ - case 0x2006: /* SIX-PER-EM SPACE */ - case 0x2007: /* FIGURE SPACE */ - case 0x2008: /* PUNCTUATION SPACE */ - case 0x2009: /* THIN SPACE */ - case 0x200A: /* HAIR SPACE */ - case 0x202f: /* NARROW NO-BREAK SPACE */ - case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ - case 0x3000: /* IDEOGRAPHIC SPACE */ - break; - } - ecode++; - break; - - case OP_NOT_VSPACE: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - switch(c) - { - default: break; - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - case 0x0d: /* CR */ - case 0x85: /* NEL */ - case 0x2028: /* LINE SEPARATOR */ - case 0x2029: /* PARAGRAPH SEPARATOR */ - RRETURN(MATCH_NOMATCH); - } - ecode++; - break; - - case OP_VSPACE: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - switch(c) - { - default: RRETURN(MATCH_NOMATCH); - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - case 0x0d: /* CR */ - case 0x85: /* NEL */ - case 0x2028: /* LINE SEPARATOR */ - case 0x2029: /* PARAGRAPH SEPARATOR */ - break; - } - ecode++; - break; - -#ifdef SUPPORT_UCP - /* Check the next character by Unicode property. We will get here only - if the support is in the binary; otherwise a compile-time error occurs. */ - - case OP_PROP: - case OP_NOTPROP: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - { - int chartype, script; - int category = _pcre_ucp_findprop(c, &chartype, &script); - - switch(ecode[1]) - { - case PT_ANY: - if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH); - break; - - case PT_LAMP: - if ((chartype == ucp_Lu || - chartype == ucp_Ll || - chartype == ucp_Lt) == (op == OP_NOTPROP)) - RRETURN(MATCH_NOMATCH); - break; - - case PT_GC: - if ((ecode[2] != category) == (op == OP_PROP)) - RRETURN(MATCH_NOMATCH); - break; - - case PT_PC: - if ((ecode[2] != chartype) == (op == OP_PROP)) - RRETURN(MATCH_NOMATCH); - break; - - case PT_SC: - if ((ecode[2] != script) == (op == OP_PROP)) - RRETURN(MATCH_NOMATCH); - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - - ecode += 3; - } - break; - - /* Match an extended Unicode sequence. We will get here only if the support - is in the binary; otherwise a compile-time error occurs. */ - - case OP_EXTUNI: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - { - int chartype, script; - int category = _pcre_ucp_findprop(c, &chartype, &script); - if (category == ucp_M) RRETURN(MATCH_NOMATCH); - while (eptr < md->end_subject) - { - int len = 1; - if (!utf8) c = *eptr; else - { - GETCHARLEN(c, eptr, len); - } - category = _pcre_ucp_findprop(c, &chartype, &script); - if (category != ucp_M) break; - eptr += len; - } - } - ecode++; - break; -#endif - - - /* Match a back reference, possibly repeatedly. Look past the end of the - item to see if there is repeat information following. The code is similar - to that for character classes, but repeated for efficiency. Then obey - similar code to character type repeats - written out again for speed. - However, if the referenced string is the empty string, always treat - it as matched, any number of times (otherwise there could be infinite - loops). */ - - case OP_REF: - { - offset = GET2(ecode, 1) << 1; /* Doubled ref number */ - ecode += 3; - - /* If the reference is unset, there are two possibilities: - - (a) In the default, Perl-compatible state, set the length to be longer - than the amount of subject left; this ensures that every attempt at a - match fails. We can't just fail here, because of the possibility of - quantifiers with zero minima. - - (b) If the JavaScript compatibility flag is set, set the length to zero - so that the back reference matches an empty string. - - Otherwise, set the length to the length of what was matched by the - referenced subpattern. */ - - if (offset >= offset_top || md->offset_vector[offset] < 0) - length = (md->jscript_compat)? 0 : md->end_subject - eptr + 1; - else - length = md->offset_vector[offset+1] - md->offset_vector[offset]; - - /* Set up for repetition, or handle the non-repeated case */ - - switch (*ecode) - { - case OP_CRSTAR: - case OP_CRMINSTAR: - case OP_CRPLUS: - case OP_CRMINPLUS: - case OP_CRQUERY: - case OP_CRMINQUERY: - c = *ecode++ - OP_CRSTAR; - minimize = (c & 1) != 0; - min = rep_min[c]; /* Pick up values from tables; */ - max = rep_max[c]; /* zero for max => infinity */ - if (max == 0) max = INT_MAX; - break; - - case OP_CRRANGE: - case OP_CRMINRANGE: - minimize = (*ecode == OP_CRMINRANGE); - min = GET2(ecode, 1); - max = GET2(ecode, 3); - if (max == 0) max = INT_MAX; - ecode += 5; - break; - - default: /* No repeat follows */ - if (!match_ref(offset, eptr, length, md, ims)) RRETURN(MATCH_NOMATCH); - eptr += length; - continue; /* With the main loop */ - } - - /* If the length of the reference is zero, just continue with the - main loop. */ - - if (length == 0) continue; - - /* First, ensure the minimum number of matches are present. We get back - the length of the reference string explicitly rather than passing the - address of eptr, so that eptr can be a register variable. */ - - for (i = 1; i <= min; i++) - { - if (!match_ref(offset, eptr, length, md, ims)) RRETURN(MATCH_NOMATCH); - eptr += length; - } - - /* If min = max, continue at the same level without recursion. - They are not both allowed to be zero. */ - - if (min == max) continue; - - /* If minimizing, keep trying and advancing the pointer */ - - if (minimize) - { - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || !match_ref(offset, eptr, length, md, ims)) - RRETURN(MATCH_NOMATCH); - eptr += length; - } - /* Control never gets here */ - } - - /* If maximizing, find the longest string and work backwards */ - - else - { - pp = eptr; - for (i = min; i < max; i++) - { - if (!match_ref(offset, eptr, length, md, ims)) break; - eptr += length; - } - while (eptr >= pp) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - eptr -= length; - } - RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - - - - /* Match a bit-mapped character class, possibly repeatedly. This op code is - used when all the characters in the class have values in the range 0-255, - and either the matching is caseful, or the characters are in the range - 0-127 when UTF-8 processing is enabled. The only difference between - OP_CLASS and OP_NCLASS occurs when a data character outside the range is - encountered. - - First, look past the end of the item to see if there is repeat information - following. Then obey similar code to character type repeats - written out - again for speed. */ - - case OP_NCLASS: - case OP_CLASS: - { - data = ecode + 1; /* Save for matching */ - ecode += 33; /* Advance past the item */ - - switch (*ecode) - { - case OP_CRSTAR: - case OP_CRMINSTAR: - case OP_CRPLUS: - case OP_CRMINPLUS: - case OP_CRQUERY: - case OP_CRMINQUERY: - c = *ecode++ - OP_CRSTAR; - minimize = (c & 1) != 0; - min = rep_min[c]; /* Pick up values from tables; */ - max = rep_max[c]; /* zero for max => infinity */ - if (max == 0) max = INT_MAX; - break; - - case OP_CRRANGE: - case OP_CRMINRANGE: - minimize = (*ecode == OP_CRMINRANGE); - min = GET2(ecode, 1); - max = GET2(ecode, 3); - if (max == 0) max = INT_MAX; - ecode += 5; - break; - - default: /* No repeat follows */ - min = max = 1; - break; - } - - /* First, ensure the minimum number of matches are present. */ - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (utf8) - { - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - if (c > 255) - { - if (op == OP_CLASS) RRETURN(MATCH_NOMATCH); - } - else - { - if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); - } - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - c = *eptr++; - if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); - } - } - - /* If max == min we can continue with the main loop without the - need to recurse. */ - - if (min == max) continue; - - /* If minimizing, keep testing the rest of the expression and advancing - the pointer while it matches the class. */ - - if (minimize) - { -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (utf8) - { - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - if (c > 255) - { - if (op == OP_CLASS) RRETURN(MATCH_NOMATCH); - } - else - { - if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); - } - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - c = *eptr++; - if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - } - - /* If maximizing, find the longest possible run, then work backwards. */ - - else - { - pp = eptr; - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (utf8) - { - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c > 255) - { - if (op == OP_CLASS) break; - } - else - { - if ((data[c/8] & (1 << (c&7))) == 0) break; - } - eptr += len; - } - for (;;) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - BACKCHAR(eptr); - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject) break; - c = *eptr; - if ((data[c/8] & (1 << (c&7))) == 0) break; - eptr++; - } - while (eptr >= pp) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - eptr--; - } - } - - RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - - - /* Match an extended character class. This opcode is encountered only - in UTF-8 mode, because that's the only time it is compiled. */ - -#ifdef SUPPORT_UTF8 - case OP_XCLASS: - { - data = ecode + 1 + LINK_SIZE; /* Save for matching */ - ecode += GET(ecode, 1); /* Advance past the item */ - - switch (*ecode) - { - case OP_CRSTAR: - case OP_CRMINSTAR: - case OP_CRPLUS: - case OP_CRMINPLUS: - case OP_CRQUERY: - case OP_CRMINQUERY: - c = *ecode++ - OP_CRSTAR; - minimize = (c & 1) != 0; - min = rep_min[c]; /* Pick up values from tables; */ - max = rep_max[c]; /* zero for max => infinity */ - if (max == 0) max = INT_MAX; - break; - - case OP_CRRANGE: - case OP_CRMINRANGE: - minimize = (*ecode == OP_CRMINRANGE); - min = GET2(ecode, 1); - max = GET2(ecode, 3); - if (max == 0) max = INT_MAX; - ecode += 5; - break; - - default: /* No repeat follows */ - min = max = 1; - break; - } - - /* First, ensure the minimum number of matches are present. */ - - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH); - } - - /* If max == min we can continue with the main loop without the - need to recurse. */ - - if (min == max) continue; - - /* If minimizing, keep testing the rest of the expression and advancing - the pointer while it matches the class. */ - - if (minimize) - { - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - } - - /* If maximizing, find the longest possible run, then work backwards. */ - - else - { - pp = eptr; - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (!_pcre_xclass(c, data)) break; - eptr += len; - } - for(;;) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - if (utf8) BACKCHAR(eptr); - } - RRETURN(MATCH_NOMATCH); - } - - /* Control never gets here */ - } -#endif /* End of XCLASS */ - - /* Match a single character, casefully */ - - case OP_CHAR: -#ifdef SUPPORT_UTF8 - if (utf8) - { - length = 1; - ecode++; - GETCHARLEN(fc, ecode, length); - if (length > md->end_subject - eptr) RRETURN(MATCH_NOMATCH); - while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH); - } - else -#endif - - /* Non-UTF-8 mode */ - { - if (md->end_subject - eptr < 1) RRETURN(MATCH_NOMATCH); - if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH); - ecode += 2; - } - break; - - /* Match a single character, caselessly */ - - case OP_CHARNC: -#ifdef SUPPORT_UTF8 - if (utf8) - { - length = 1; - ecode++; - GETCHARLEN(fc, ecode, length); - - if (length > md->end_subject - eptr) RRETURN(MATCH_NOMATCH); - - /* If the pattern character's value is < 128, we have only one byte, and - can use the fast lookup table. */ - - if (fc < 128) - { - if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH); - } - - /* Otherwise we must pick up the subject character */ - - else - { - unsigned int dc; - GETCHARINC(dc, eptr); - ecode += length; - - /* If we have Unicode property support, we can use it to test the other - case of the character, if there is one. */ - - if (fc != dc) - { -#ifdef SUPPORT_UCP - if (dc != _pcre_ucp_othercase(fc)) -#endif - RRETURN(MATCH_NOMATCH); - } - } - } - else -#endif /* SUPPORT_UTF8 */ - - /* Non-UTF-8 mode */ - { - if (md->end_subject - eptr < 1) RRETURN(MATCH_NOMATCH); - if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH); - ecode += 2; - } - break; - - /* Match a single character repeatedly. */ - - case OP_EXACT: - min = max = GET2(ecode, 1); - ecode += 3; - goto REPEATCHAR; - - case OP_POSUPTO: - possessive = TRUE; - /* Fall through */ - - case OP_UPTO: - case OP_MINUPTO: - min = 0; - max = GET2(ecode, 1); - minimize = *ecode == OP_MINUPTO; - ecode += 3; - goto REPEATCHAR; - - case OP_POSSTAR: - possessive = TRUE; - min = 0; - max = INT_MAX; - ecode++; - goto REPEATCHAR; - - case OP_POSPLUS: - possessive = TRUE; - min = 1; - max = INT_MAX; - ecode++; - goto REPEATCHAR; - - case OP_POSQUERY: - possessive = TRUE; - min = 0; - max = 1; - ecode++; - goto REPEATCHAR; - - case OP_STAR: - case OP_MINSTAR: - case OP_PLUS: - case OP_MINPLUS: - case OP_QUERY: - case OP_MINQUERY: - c = *ecode++ - OP_STAR; - minimize = (c & 1) != 0; - min = rep_min[c]; /* Pick up values from tables; */ - max = rep_max[c]; /* zero for max => infinity */ - if (max == 0) max = INT_MAX; - - /* Common code for all repeated single-character matches. We can give - up quickly if there are fewer than the minimum number of characters left in - the subject. */ - - REPEATCHAR: -#ifdef SUPPORT_UTF8 - if (utf8) - { - length = 1; - charptr = ecode; - GETCHARLEN(fc, ecode, length); - if (min * length > md->end_subject - eptr) RRETURN(MATCH_NOMATCH); - ecode += length; - - /* Handle multibyte character matching specially here. There is - support for caseless matching if UCP support is present. */ - - if (length > 1) - { -#ifdef SUPPORT_UCP - unsigned int othercase; - if ((ims & PCRE_CASELESS) != 0 && - (othercase = _pcre_ucp_othercase(fc)) != NOTACHAR) - oclength = _pcre_ord2utf8(othercase, occhars); - else oclength = 0; -#endif /* SUPPORT_UCP */ - - for (i = 1; i <= min; i++) - { - if (memcmp(eptr, charptr, length) == 0) eptr += length; -#ifdef SUPPORT_UCP - /* Need braces because of following else */ - else if (oclength == 0) { RRETURN(MATCH_NOMATCH); } - else - { - if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH); - eptr += oclength; - } -#else /* without SUPPORT_UCP */ - else { RRETURN(MATCH_NOMATCH); } -#endif /* SUPPORT_UCP */ - } - - if (min == max) continue; - - if (minimize) - { - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - if (memcmp(eptr, charptr, length) == 0) eptr += length; -#ifdef SUPPORT_UCP - /* Need braces because of following else */ - else if (oclength == 0) { RRETURN(MATCH_NOMATCH); } - else - { - if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH); - eptr += oclength; - } -#else /* without SUPPORT_UCP */ - else { RRETURN (MATCH_NOMATCH); } -#endif /* SUPPORT_UCP */ - } - /* Control never gets here */ - } - - else /* Maximize */ - { - pp = eptr; - for (i = min; i < max; i++) - { - if (eptr > md->end_subject - length) break; - if (memcmp(eptr, charptr, length) == 0) eptr += length; -#ifdef SUPPORT_UCP - else if (oclength == 0) break; - else - { - if (memcmp(eptr, occhars, oclength) != 0) break; - eptr += oclength; - } -#else /* without SUPPORT_UCP */ - else break; -#endif /* SUPPORT_UCP */ - } - - if (possessive) continue; - for(;;) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr == pp) RRETURN(MATCH_NOMATCH); -#ifdef SUPPORT_UCP - eptr--; - BACKCHAR(eptr); -#else /* without SUPPORT_UCP */ - eptr -= length; -#endif /* SUPPORT_UCP */ - } - } - /* Control never gets here */ - } - - /* If the length of a UTF-8 character is 1, we fall through here, and - obey the code as for non-UTF-8 characters below, though in this case the - value of fc will always be < 128. */ - } - else -#endif /* SUPPORT_UTF8 */ - - /* When not in UTF-8 mode, load a single-byte character. */ - { - if (min > md->end_subject - eptr) RRETURN(MATCH_NOMATCH); - fc = *ecode++; - } - - /* The value of fc at this point is always less than 256, though we may or - may not be in UTF-8 mode. The code is duplicated for the caseless and - caseful cases, for speed, since matching characters is likely to be quite - common. First, ensure the minimum number of matches are present. If min = - max, continue at the same level without recursing. Otherwise, if - minimizing, keep trying the rest of the expression and advancing one - matching character if failing, up to the maximum. Alternatively, if - maximizing, find the maximum number of characters and work backwards. */ - - DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max, - max, eptr)); - - if ((ims & PCRE_CASELESS) != 0) - { - fc = md->lcc[fc]; - for (i = 1; i <= min; i++) - if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH); - if (min == max) continue; - if (minimize) - { - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject || - fc != md->lcc[*eptr++]) - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - } - else /* Maximize */ - { - pp = eptr; - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break; - eptr++; - } - if (possessive) continue; - while (eptr >= pp) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25); - eptr--; - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - } - - /* Caseful comparisons (includes all multi-byte characters) */ - - else - { - for (i = 1; i <= min; i++) if (fc != *eptr++) RRETURN(MATCH_NOMATCH); - if (min == max) continue; - if (minimize) - { - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject || fc != *eptr++) - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - } - else /* Maximize */ - { - pp = eptr; - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || fc != *eptr) break; - eptr++; - } - if (possessive) continue; - while (eptr >= pp) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27); - eptr--; - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - - /* Match a negated single one-byte character. The character we are - checking can be multibyte. */ - - case OP_NOT: - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - ecode++; - GETCHARINCTEST(c, eptr); - if ((ims & PCRE_CASELESS) != 0) - { -#ifdef SUPPORT_UTF8 - if (c < 256) -#endif - c = md->lcc[c]; - if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH); - } - else - { - if (*ecode++ == c) RRETURN(MATCH_NOMATCH); - } - break; - - /* Match a negated single one-byte character repeatedly. This is almost a - repeat of the code for a repeated single character, but I haven't found a - nice way of commoning these up that doesn't require a test of the - positive/negative option for each character match. Maybe that wouldn't add - very much to the time taken, but character matching *is* what this is all - about... */ - - case OP_NOTEXACT: - min = max = GET2(ecode, 1); - ecode += 3; - goto REPEATNOTCHAR; - - case OP_NOTUPTO: - case OP_NOTMINUPTO: - min = 0; - max = GET2(ecode, 1); - minimize = *ecode == OP_NOTMINUPTO; - ecode += 3; - goto REPEATNOTCHAR; - - case OP_NOTPOSSTAR: - possessive = TRUE; - min = 0; - max = INT_MAX; - ecode++; - goto REPEATNOTCHAR; - - case OP_NOTPOSPLUS: - possessive = TRUE; - min = 1; - max = INT_MAX; - ecode++; - goto REPEATNOTCHAR; - - case OP_NOTPOSQUERY: - possessive = TRUE; - min = 0; - max = 1; - ecode++; - goto REPEATNOTCHAR; - - case OP_NOTPOSUPTO: - possessive = TRUE; - min = 0; - max = GET2(ecode, 1); - ecode += 3; - goto REPEATNOTCHAR; - - case OP_NOTSTAR: - case OP_NOTMINSTAR: - case OP_NOTPLUS: - case OP_NOTMINPLUS: - case OP_NOTQUERY: - case OP_NOTMINQUERY: - c = *ecode++ - OP_NOTSTAR; - minimize = (c & 1) != 0; - min = rep_min[c]; /* Pick up values from tables; */ - max = rep_max[c]; /* zero for max => infinity */ - if (max == 0) max = INT_MAX; - - /* Common code for all repeated single-byte matches. We can give up quickly - if there are fewer than the minimum number of bytes left in the - subject. */ - - REPEATNOTCHAR: - if (min > md->end_subject - eptr) RRETURN(MATCH_NOMATCH); - fc = *ecode++; - - /* The code is duplicated for the caseless and caseful cases, for speed, - since matching characters is likely to be quite common. First, ensure the - minimum number of matches are present. If min = max, continue at the same - level without recursing. Otherwise, if minimizing, keep trying the rest of - the expression and advancing one matching character if failing, up to the - maximum. Alternatively, if maximizing, find the maximum number of - characters and work backwards. */ - - DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max, - max, eptr)); - - if ((ims & PCRE_CASELESS) != 0) - { - fc = md->lcc[fc]; - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (utf8) - { - register unsigned int d; - for (i = 1; i <= min; i++) - { - GETCHARINC(d, eptr); - if (d < 256) d = md->lcc[d]; - if (fc == d) RRETURN(MATCH_NOMATCH); - } - } - else -#endif - - /* Not UTF-8 mode */ - { - for (i = 1; i <= min; i++) - if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH); - } - - if (min == max) continue; - - if (minimize) - { -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (utf8) - { - register unsigned int d; - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - GETCHARINC(d, eptr); - if (d < 256) d = md->lcc[d]; - if (fi >= max || eptr >= md->end_subject || fc == d) - RRETURN(MATCH_NOMATCH); - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++]) - RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - } - - /* Maximize case */ - - else - { - pp = eptr; - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (utf8) - { - register unsigned int d; - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(d, eptr, len); - if (d < 256) d = md->lcc[d]; - if (fc == d) break; - eptr += len; - } - if (possessive) continue; - for(;;) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - BACKCHAR(eptr); - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break; - eptr++; - } - if (possessive) continue; - while (eptr >= pp) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - eptr--; - } - } - - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - } - - /* Caseful comparisons */ - - else - { -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (utf8) - { - register unsigned int d; - for (i = 1; i <= min; i++) - { - GETCHARINC(d, eptr); - if (fc == d) RRETURN(MATCH_NOMATCH); - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (i = 1; i <= min; i++) - if (fc == *eptr++) RRETURN(MATCH_NOMATCH); - } - - if (min == max) continue; - - if (minimize) - { -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (utf8) - { - register unsigned int d; - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - GETCHARINC(d, eptr); - if (fi >= max || eptr >= md->end_subject || fc == d) - RRETURN(MATCH_NOMATCH); - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject || fc == *eptr++) - RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - } - - /* Maximize case */ - - else - { - pp = eptr; - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (utf8) - { - register unsigned int d; - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(d, eptr, len); - if (fc == d) break; - eptr += len; - } - if (possessive) continue; - for(;;) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - BACKCHAR(eptr); - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || fc == *eptr) break; - eptr++; - } - if (possessive) continue; - while (eptr >= pp) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - eptr--; - } - } - - RRETURN(MATCH_NOMATCH); - } - } - /* Control never gets here */ - - /* Match a single character type repeatedly; several different opcodes - share code. This is very similar to the code for single characters, but we - repeat it in the interests of efficiency. */ - - case OP_TYPEEXACT: - min = max = GET2(ecode, 1); - minimize = TRUE; - ecode += 3; - goto REPEATTYPE; - - case OP_TYPEUPTO: - case OP_TYPEMINUPTO: - min = 0; - max = GET2(ecode, 1); - minimize = *ecode == OP_TYPEMINUPTO; - ecode += 3; - goto REPEATTYPE; - - case OP_TYPEPOSSTAR: - possessive = TRUE; - min = 0; - max = INT_MAX; - ecode++; - goto REPEATTYPE; - - case OP_TYPEPOSPLUS: - possessive = TRUE; - min = 1; - max = INT_MAX; - ecode++; - goto REPEATTYPE; - - case OP_TYPEPOSQUERY: - possessive = TRUE; - min = 0; - max = 1; - ecode++; - goto REPEATTYPE; - - case OP_TYPEPOSUPTO: - possessive = TRUE; - min = 0; - max = GET2(ecode, 1); - ecode += 3; - goto REPEATTYPE; - - case OP_TYPESTAR: - case OP_TYPEMINSTAR: - case OP_TYPEPLUS: - case OP_TYPEMINPLUS: - case OP_TYPEQUERY: - case OP_TYPEMINQUERY: - c = *ecode++ - OP_TYPESTAR; - minimize = (c & 1) != 0; - min = rep_min[c]; /* Pick up values from tables; */ - max = rep_max[c]; /* zero for max => infinity */ - if (max == 0) max = INT_MAX; - - /* Common code for all repeated single character type matches. Note that - in UTF-8 mode, '.' matches a character of any length, but for the other - character types, the valid characters are all one-byte long. */ - - REPEATTYPE: - ctype = *ecode++; /* Code for the character type */ - -#ifdef SUPPORT_UCP - if (ctype == OP_PROP || ctype == OP_NOTPROP) - { - prop_fail_result = ctype == OP_NOTPROP; - prop_type = *ecode++; - prop_value = *ecode++; - } - else prop_type = -1; -#endif - - /* First, ensure the minimum number of matches are present. Use inline - code for maximizing the speed, and do the type test once at the start - (i.e. keep it out of the loop). Also we can test that there are at least - the minimum number of bytes before we start. This isn't as effective in - UTF-8 mode, but it does no harm. Separate the UTF-8 code completely as that - is tidier. Also separate the UCP code, which can be the same for both UTF-8 - and single-bytes. */ - - if (min > md->end_subject - eptr) RRETURN(MATCH_NOMATCH); - if (min > 0) - { -#ifdef SUPPORT_UCP - if (prop_type >= 0) - { - switch(prop_type) - { - case PT_ANY: - if (prop_fail_result) RRETURN(MATCH_NOMATCH); - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - } - break; - - case PT_LAMP: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if ((prop_chartype == ucp_Lu || - prop_chartype == ucp_Ll || - prop_chartype == ucp_Lt) == prop_fail_result) - RRETURN(MATCH_NOMATCH); - } - break; - - case PT_GC: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if ((prop_category == prop_value) == prop_fail_result) - RRETURN(MATCH_NOMATCH); - } - break; - - case PT_PC: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if ((prop_chartype == prop_value) == prop_fail_result) - RRETURN(MATCH_NOMATCH); - } - break; - - case PT_SC: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if ((prop_script == prop_value) == prop_fail_result) - RRETURN(MATCH_NOMATCH); - } - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - } - - /* Match extended Unicode sequences. We will get here only if the - support is in the binary; otherwise a compile-time error occurs. */ - - else if (ctype == OP_EXTUNI) - { - for (i = 1; i <= min; i++) - { - GETCHARINCTEST(c, eptr); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH); - while (eptr < md->end_subject) - { - int len = 1; - if (!utf8) c = *eptr; else - { - GETCHARLEN(c, eptr, len); - } - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if (prop_category != ucp_M) break; - eptr += len; - } - } - } - - else -#endif /* SUPPORT_UCP */ - -/* Handle all other cases when the coding is UTF-8 */ - -#ifdef SUPPORT_UTF8 - if (utf8) switch(ctype) - { - case OP_ANY: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject || IS_NEWLINE(eptr)) - RRETURN(MATCH_NOMATCH); - eptr++; - while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; - } - break; - - case OP_ALLANY: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - eptr++; - while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; - } - break; - - case OP_ANYBYTE: - eptr += min; - break; - - case OP_ANYNL: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - switch(c) - { - default: RRETURN(MATCH_NOMATCH); - case 0x000d: - if (eptr < md->end_subject && *eptr == 0x0a) eptr++; - break; - - case 0x000a: - break; - - case 0x000b: - case 0x000c: - case 0x0085: - case 0x2028: - case 0x2029: - if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); - break; - } - } - break; - - case OP_NOT_HSPACE: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - switch(c) - { - default: break; - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - case 0x1680: /* OGHAM SPACE MARK */ - case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ - case 0x2000: /* EN QUAD */ - case 0x2001: /* EM QUAD */ - case 0x2002: /* EN SPACE */ - case 0x2003: /* EM SPACE */ - case 0x2004: /* THREE-PER-EM SPACE */ - case 0x2005: /* FOUR-PER-EM SPACE */ - case 0x2006: /* SIX-PER-EM SPACE */ - case 0x2007: /* FIGURE SPACE */ - case 0x2008: /* PUNCTUATION SPACE */ - case 0x2009: /* THIN SPACE */ - case 0x200A: /* HAIR SPACE */ - case 0x202f: /* NARROW NO-BREAK SPACE */ - case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ - case 0x3000: /* IDEOGRAPHIC SPACE */ - RRETURN(MATCH_NOMATCH); - } - } - break; - - case OP_HSPACE: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - switch(c) - { - default: RRETURN(MATCH_NOMATCH); - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - case 0x1680: /* OGHAM SPACE MARK */ - case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ - case 0x2000: /* EN QUAD */ - case 0x2001: /* EM QUAD */ - case 0x2002: /* EN SPACE */ - case 0x2003: /* EM SPACE */ - case 0x2004: /* THREE-PER-EM SPACE */ - case 0x2005: /* FOUR-PER-EM SPACE */ - case 0x2006: /* SIX-PER-EM SPACE */ - case 0x2007: /* FIGURE SPACE */ - case 0x2008: /* PUNCTUATION SPACE */ - case 0x2009: /* THIN SPACE */ - case 0x200A: /* HAIR SPACE */ - case 0x202f: /* NARROW NO-BREAK SPACE */ - case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ - case 0x3000: /* IDEOGRAPHIC SPACE */ - break; - } - } - break; - - case OP_NOT_VSPACE: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - switch(c) - { - default: break; - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - case 0x0d: /* CR */ - case 0x85: /* NEL */ - case 0x2028: /* LINE SEPARATOR */ - case 0x2029: /* PARAGRAPH SEPARATOR */ - RRETURN(MATCH_NOMATCH); - } - } - break; - - case OP_VSPACE: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - switch(c) - { - default: RRETURN(MATCH_NOMATCH); - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - case 0x0d: /* CR */ - case 0x85: /* NEL */ - case 0x2028: /* LINE SEPARATOR */ - case 0x2029: /* PARAGRAPH SEPARATOR */ - break; - } - } - break; - - case OP_NOT_DIGIT: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - if (c < 128 && (md->ctypes[c] & ctype_digit) != 0) - RRETURN(MATCH_NOMATCH); - } - break; - - case OP_DIGIT: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject || - *eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0) - RRETURN(MATCH_NOMATCH); - /* No need to skip more bytes - we know it's a 1-byte character */ - } - break; - - case OP_NOT_WHITESPACE: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject || - (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)) - RRETURN(MATCH_NOMATCH); - while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80); - } - break; - - case OP_WHITESPACE: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject || - *eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0) - RRETURN(MATCH_NOMATCH); - /* No need to skip more bytes - we know it's a 1-byte character */ - } - break; - - case OP_NOT_WORDCHAR: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject || - (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)) - RRETURN(MATCH_NOMATCH); - while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80); - } - break; - - case OP_WORDCHAR: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject || - *eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0) - RRETURN(MATCH_NOMATCH); - /* No need to skip more bytes - we know it's a 1-byte character */ - } - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } /* End switch(ctype) */ - - else -#endif /* SUPPORT_UTF8 */ - - /* Code for the non-UTF-8 case for minimum matching of operators other - than OP_PROP and OP_NOTPROP. We can assume that there are the minimum - number of bytes present, as this was tested above. */ - - switch(ctype) - { - case OP_ANY: - for (i = 1; i <= min; i++) - { - if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); - eptr++; - } - break; - - case OP_ALLANY: - eptr += min; - break; - - case OP_ANYBYTE: - eptr += min; - break; - - /* Because of the CRLF case, we can't assume the minimum number of - bytes are present in this case. */ - - case OP_ANYNL: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - switch(*eptr++) - { - default: RRETURN(MATCH_NOMATCH); - case 0x000d: - if (eptr < md->end_subject && *eptr == 0x0a) eptr++; - break; - case 0x000a: - break; - - case 0x000b: - case 0x000c: - case 0x0085: - if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); - break; - } - } - break; - - case OP_NOT_HSPACE: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - switch(*eptr++) - { - default: break; - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - RRETURN(MATCH_NOMATCH); - } - } - break; - - case OP_HSPACE: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - switch(*eptr++) - { - default: RRETURN(MATCH_NOMATCH); - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - break; - } - } - break; - - case OP_NOT_VSPACE: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - switch(*eptr++) - { - default: break; - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - case 0x0d: /* CR */ - case 0x85: /* NEL */ - RRETURN(MATCH_NOMATCH); - } - } - break; - - case OP_VSPACE: - for (i = 1; i <= min; i++) - { - if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - switch(*eptr++) - { - default: RRETURN(MATCH_NOMATCH); - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - case 0x0d: /* CR */ - case 0x85: /* NEL */ - break; - } - } - break; - - case OP_NOT_DIGIT: - for (i = 1; i <= min; i++) - if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_DIGIT: - for (i = 1; i <= min; i++) - if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_NOT_WHITESPACE: - for (i = 1; i <= min; i++) - if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_WHITESPACE: - for (i = 1; i <= min; i++) - if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_NOT_WORDCHAR: - for (i = 1; i <= min; i++) - if ((md->ctypes[*eptr++] & ctype_word) != 0) - RRETURN(MATCH_NOMATCH); - break; - - case OP_WORDCHAR: - for (i = 1; i <= min; i++) - if ((md->ctypes[*eptr++] & ctype_word) == 0) - RRETURN(MATCH_NOMATCH); - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - } - - /* If min = max, continue at the same level without recursing */ - - if (min == max) continue; - - /* If minimizing, we have to test the rest of the pattern before each - subsequent match. Again, separate the UTF-8 case for speed, and also - separate the UCP cases. */ - - if (minimize) - { -#ifdef SUPPORT_UCP - if (prop_type >= 0) - { - switch(prop_type) - { - case PT_ANY: - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - if (prop_fail_result) RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - - case PT_LAMP: - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if ((prop_chartype == ucp_Lu || - prop_chartype == ucp_Ll || - prop_chartype == ucp_Lt) == prop_fail_result) - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - - case PT_GC: - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if ((prop_category == prop_value) == prop_fail_result) - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - - case PT_PC: - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if ((prop_chartype == prop_value) == prop_fail_result) - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - - case PT_SC: - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINC(c, eptr); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if ((prop_script == prop_value) == prop_fail_result) - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - } - - /* Match extended Unicode sequences. We will get here only if the - support is in the binary; otherwise a compile-time error occurs. */ - - else if (ctype == OP_EXTUNI) - { - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH); - GETCHARINCTEST(c, eptr); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH); - while (eptr < md->end_subject) - { - int len = 1; - if (!utf8) c = *eptr; else - { - GETCHARLEN(c, eptr, len); - } - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if (prop_category != ucp_M) break; - eptr += len; - } - } - } - - else -#endif /* SUPPORT_UCP */ - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - if (utf8) - { - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject || - (ctype == OP_ANY && IS_NEWLINE(eptr))) - RRETURN(MATCH_NOMATCH); - - GETCHARINC(c, eptr); - switch(ctype) - { - case OP_ANY: /* This is the non-NL case */ - case OP_ALLANY: - case OP_ANYBYTE: - break; - - case OP_ANYNL: - switch(c) - { - default: RRETURN(MATCH_NOMATCH); - case 0x000d: - if (eptr < md->end_subject && *eptr == 0x0a) eptr++; - break; - case 0x000a: - break; - - case 0x000b: - case 0x000c: - case 0x0085: - case 0x2028: - case 0x2029: - if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); - break; - } - break; - - case OP_NOT_HSPACE: - switch(c) - { - default: break; - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - case 0x1680: /* OGHAM SPACE MARK */ - case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ - case 0x2000: /* EN QUAD */ - case 0x2001: /* EM QUAD */ - case 0x2002: /* EN SPACE */ - case 0x2003: /* EM SPACE */ - case 0x2004: /* THREE-PER-EM SPACE */ - case 0x2005: /* FOUR-PER-EM SPACE */ - case 0x2006: /* SIX-PER-EM SPACE */ - case 0x2007: /* FIGURE SPACE */ - case 0x2008: /* PUNCTUATION SPACE */ - case 0x2009: /* THIN SPACE */ - case 0x200A: /* HAIR SPACE */ - case 0x202f: /* NARROW NO-BREAK SPACE */ - case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ - case 0x3000: /* IDEOGRAPHIC SPACE */ - RRETURN(MATCH_NOMATCH); - } - break; - - case OP_HSPACE: - switch(c) - { - default: RRETURN(MATCH_NOMATCH); - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - case 0x1680: /* OGHAM SPACE MARK */ - case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ - case 0x2000: /* EN QUAD */ - case 0x2001: /* EM QUAD */ - case 0x2002: /* EN SPACE */ - case 0x2003: /* EM SPACE */ - case 0x2004: /* THREE-PER-EM SPACE */ - case 0x2005: /* FOUR-PER-EM SPACE */ - case 0x2006: /* SIX-PER-EM SPACE */ - case 0x2007: /* FIGURE SPACE */ - case 0x2008: /* PUNCTUATION SPACE */ - case 0x2009: /* THIN SPACE */ - case 0x200A: /* HAIR SPACE */ - case 0x202f: /* NARROW NO-BREAK SPACE */ - case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ - case 0x3000: /* IDEOGRAPHIC SPACE */ - break; - } - break; - - case OP_NOT_VSPACE: - switch(c) - { - default: break; - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - case 0x0d: /* CR */ - case 0x85: /* NEL */ - case 0x2028: /* LINE SEPARATOR */ - case 0x2029: /* PARAGRAPH SEPARATOR */ - RRETURN(MATCH_NOMATCH); - } - break; - - case OP_VSPACE: - switch(c) - { - default: RRETURN(MATCH_NOMATCH); - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - case 0x0d: /* CR */ - case 0x85: /* NEL */ - case 0x2028: /* LINE SEPARATOR */ - case 0x2029: /* PARAGRAPH SEPARATOR */ - break; - } - break; - - case OP_NOT_DIGIT: - if (c < 256 && (md->ctypes[c] & ctype_digit) != 0) - RRETURN(MATCH_NOMATCH); - break; - - case OP_DIGIT: - if (c >= 256 || (md->ctypes[c] & ctype_digit) == 0) - RRETURN(MATCH_NOMATCH); - break; - - case OP_NOT_WHITESPACE: - if (c < 256 && (md->ctypes[c] & ctype_space) != 0) - RRETURN(MATCH_NOMATCH); - break; - - case OP_WHITESPACE: - if (c >= 256 || (md->ctypes[c] & ctype_space) == 0) - RRETURN(MATCH_NOMATCH); - break; - - case OP_NOT_WORDCHAR: - if (c < 256 && (md->ctypes[c] & ctype_word) != 0) - RRETURN(MATCH_NOMATCH); - break; - - case OP_WORDCHAR: - if (c >= 256 || (md->ctypes[c] & ctype_word) == 0) - RRETURN(MATCH_NOMATCH); - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - } - } - else -#endif - /* Not UTF-8 mode */ - { - for (fi = min;; fi++) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (fi >= max || eptr >= md->end_subject || - (ctype == OP_ANY && IS_NEWLINE(eptr))) - RRETURN(MATCH_NOMATCH); - - c = *eptr++; - switch(ctype) - { - case OP_ANY: /* This is the non-NL case */ - case OP_ALLANY: - case OP_ANYBYTE: - break; - - case OP_ANYNL: - switch(c) - { - default: RRETURN(MATCH_NOMATCH); - case 0x000d: - if (eptr < md->end_subject && *eptr == 0x0a) eptr++; - break; - - case 0x000a: - break; - - case 0x000b: - case 0x000c: - case 0x0085: - if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); - break; - } - break; - - case OP_NOT_HSPACE: - switch(c) - { - default: break; - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - RRETURN(MATCH_NOMATCH); - } - break; - - case OP_HSPACE: - switch(c) - { - default: RRETURN(MATCH_NOMATCH); - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - break; - } - break; - - case OP_NOT_VSPACE: - switch(c) - { - default: break; - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - case 0x0d: /* CR */ - case 0x85: /* NEL */ - RRETURN(MATCH_NOMATCH); - } - break; - - case OP_VSPACE: - switch(c) - { - default: RRETURN(MATCH_NOMATCH); - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - case 0x0d: /* CR */ - case 0x85: /* NEL */ - break; - } - break; - - case OP_NOT_DIGIT: - if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_DIGIT: - if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_NOT_WHITESPACE: - if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_WHITESPACE: - if ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_NOT_WORDCHAR: - if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH); - break; - - case OP_WORDCHAR: - if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH); - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - } - } - /* Control never gets here */ - } - - /* If maximizing, it is worth using inline code for speed, doing the type - test once at the start (i.e. keep it out of the loop). Again, keep the - UTF-8 and UCP stuff separate. */ - - else - { - pp = eptr; /* Remember where we started */ - -#ifdef SUPPORT_UCP - if (prop_type >= 0) - { - switch(prop_type) - { - case PT_ANY: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (prop_fail_result) break; - eptr+= len; - } - break; - - case PT_LAMP: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if ((prop_chartype == ucp_Lu || - prop_chartype == ucp_Ll || - prop_chartype == ucp_Lt) == prop_fail_result) - break; - eptr+= len; - } - break; - - case PT_GC: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if ((prop_category == prop_value) == prop_fail_result) - break; - eptr+= len; - } - break; - - case PT_PC: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if ((prop_chartype == prop_value) == prop_fail_result) - break; - eptr+= len; - } - break; - - case PT_SC: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if ((prop_script == prop_value) == prop_fail_result) - break; - eptr+= len; - } - break; - } - - /* eptr is now past the end of the maximum run */ - - if (possessive) continue; - for(;;) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - if (utf8) BACKCHAR(eptr); - } - } - - /* Match extended Unicode sequences. We will get here only if the - support is in the binary; otherwise a compile-time error occurs. */ - - else if (ctype == OP_EXTUNI) - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject) break; - GETCHARINCTEST(c, eptr); - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if (prop_category == ucp_M) break; - while (eptr < md->end_subject) - { - int len = 1; - if (!utf8) c = *eptr; else - { - GETCHARLEN(c, eptr, len); - } - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if (prop_category != ucp_M) break; - eptr += len; - } - } - - /* eptr is now past the end of the maximum run */ - - if (possessive) continue; - for(;;) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM45); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - for (;;) /* Move back over one extended */ - { - int len = 1; - if (!utf8) c = *eptr; else - { - BACKCHAR(eptr); - GETCHARLEN(c, eptr, len); - } - prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script); - if (prop_category != ucp_M) break; - eptr--; - } - } - } - - else -#endif /* SUPPORT_UCP */ - -#ifdef SUPPORT_UTF8 - /* UTF-8 mode */ - - if (utf8) - { - switch(ctype) - { - case OP_ANY: - if (max < INT_MAX) - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break; - eptr++; - while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; - } - } - - /* Handle unlimited UTF-8 repeat */ - - else - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break; - eptr++; - while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; - } - } - break; - - case OP_ALLANY: - if (max < INT_MAX) - { - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject) break; - eptr++; - while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++; - } - } - else eptr = md->end_subject; /* Unlimited UTF-8 repeat */ - break; - - /* The byte case is the same as non-UTF8 */ - - case OP_ANYBYTE: - c = max - min; - if (c > (unsigned int)(md->end_subject - eptr)) - c = md->end_subject - eptr; - eptr += c; - break; - - case OP_ANYNL: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c == 0x000d) - { - if (++eptr >= md->end_subject) break; - if (*eptr == 0x000a) eptr++; - } - else - { - if (c != 0x000a && - (md->bsr_anycrlf || - (c != 0x000b && c != 0x000c && - c != 0x0085 && c != 0x2028 && c != 0x2029))) - break; - eptr += len; - } - } - break; - - case OP_NOT_HSPACE: - case OP_HSPACE: - for (i = min; i < max; i++) - { - BOOL gotspace; - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - switch(c) - { - default: gotspace = FALSE; break; - case 0x09: /* HT */ - case 0x20: /* SPACE */ - case 0xa0: /* NBSP */ - case 0x1680: /* OGHAM SPACE MARK */ - case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ - case 0x2000: /* EN QUAD */ - case 0x2001: /* EM QUAD */ - case 0x2002: /* EN SPACE */ - case 0x2003: /* EM SPACE */ - case 0x2004: /* THREE-PER-EM SPACE */ - case 0x2005: /* FOUR-PER-EM SPACE */ - case 0x2006: /* SIX-PER-EM SPACE */ - case 0x2007: /* FIGURE SPACE */ - case 0x2008: /* PUNCTUATION SPACE */ - case 0x2009: /* THIN SPACE */ - case 0x200A: /* HAIR SPACE */ - case 0x202f: /* NARROW NO-BREAK SPACE */ - case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ - case 0x3000: /* IDEOGRAPHIC SPACE */ - gotspace = TRUE; - break; - } - if (gotspace == (ctype == OP_NOT_HSPACE)) break; - eptr += len; - } - break; - - case OP_NOT_VSPACE: - case OP_VSPACE: - for (i = min; i < max; i++) - { - BOOL gotspace; - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - switch(c) - { - default: gotspace = FALSE; break; - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - case 0x0d: /* CR */ - case 0x85: /* NEL */ - case 0x2028: /* LINE SEPARATOR */ - case 0x2029: /* PARAGRAPH SEPARATOR */ - gotspace = TRUE; - break; - } - if (gotspace == (ctype == OP_NOT_VSPACE)) break; - eptr += len; - } - break; - - case OP_NOT_DIGIT: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c < 256 && (md->ctypes[c] & ctype_digit) != 0) break; - eptr+= len; - } - break; - - case OP_DIGIT: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c >= 256 ||(md->ctypes[c] & ctype_digit) == 0) break; - eptr+= len; - } - break; - - case OP_NOT_WHITESPACE: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c < 256 && (md->ctypes[c] & ctype_space) != 0) break; - eptr+= len; - } - break; - - case OP_WHITESPACE: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c >= 256 ||(md->ctypes[c] & ctype_space) == 0) break; - eptr+= len; - } - break; - - case OP_NOT_WORDCHAR: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c < 256 && (md->ctypes[c] & ctype_word) != 0) break; - eptr+= len; - } - break; - - case OP_WORDCHAR: - for (i = min; i < max; i++) - { - int len = 1; - if (eptr >= md->end_subject) break; - GETCHARLEN(c, eptr, len); - if (c >= 256 || (md->ctypes[c] & ctype_word) == 0) break; - eptr+= len; - } - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - - /* eptr is now past the end of the maximum run */ - - if (possessive) continue; - for(;;) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM46); - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - if (eptr-- == pp) break; /* Stop if tried at original pos */ - BACKCHAR(eptr); - } - } - else -#endif /* SUPPORT_UTF8 */ - - /* Not UTF-8 mode */ - { - switch(ctype) - { - case OP_ANY: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break; - eptr++; - } - break; - - case OP_ALLANY: - case OP_ANYBYTE: - c = max - min; - if (c > (unsigned int)(md->end_subject - eptr)) - c = md->end_subject - eptr; - eptr += c; - break; - - case OP_ANYNL: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject) break; - c = *eptr; - if (c == 0x000d) - { - if (++eptr >= md->end_subject) break; - if (*eptr == 0x000a) eptr++; - } - else - { - if (c != 0x000a && - (md->bsr_anycrlf || - (c != 0x000b && c != 0x000c && c != 0x0085))) - break; - eptr++; - } - } - break; - - case OP_NOT_HSPACE: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject) break; - c = *eptr; - if (c == 0x09 || c == 0x20 || c == 0xa0) break; - eptr++; - } - break; - - case OP_HSPACE: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject) break; - c = *eptr; - if (c != 0x09 && c != 0x20 && c != 0xa0) break; - eptr++; - } - break; - - case OP_NOT_VSPACE: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject) break; - c = *eptr; - if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85) - break; - eptr++; - } - break; - - case OP_VSPACE: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject) break; - c = *eptr; - if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85) - break; - eptr++; - } - break; - - case OP_NOT_DIGIT: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || (md->ctypes[*eptr] & ctype_digit) != 0) - break; - eptr++; - } - break; - - case OP_DIGIT: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || (md->ctypes[*eptr] & ctype_digit) == 0) - break; - eptr++; - } - break; - - case OP_NOT_WHITESPACE: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || (md->ctypes[*eptr] & ctype_space) != 0) - break; - eptr++; - } - break; - - case OP_WHITESPACE: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || (md->ctypes[*eptr] & ctype_space) == 0) - break; - eptr++; - } - break; - - case OP_NOT_WORDCHAR: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || (md->ctypes[*eptr] & ctype_word) != 0) - break; - eptr++; - } - break; - - case OP_WORDCHAR: - for (i = min; i < max; i++) - { - if (eptr >= md->end_subject || (md->ctypes[*eptr] & ctype_word) == 0) - break; - eptr++; - } - break; - - default: - RRETURN(PCRE_ERROR_INTERNAL); - } - - /* eptr is now past the end of the maximum run */ - - if (possessive) continue; - while (eptr >= pp) - { - RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM47); - eptr--; - if (rrc != MATCH_NOMATCH) RRETURN(rrc); - } - } - - /* Get here if we can't make it match with any permitted repetitions */ - - RRETURN(MATCH_NOMATCH); - } - /* Control never gets here */ - - /* There's been some horrible disaster. Arrival here can only mean there is - something seriously wrong in the code above or the OP_xxx definitions. */ - - default: - DPRINTF(("Unknown opcode %d\n", *ecode)); - RRETURN(PCRE_ERROR_UNKNOWN_OPCODE); - } - - /* Do not stick any code in here without much thought; it is assumed - that "continue" in the code above comes out to here to repeat the main - loop. */ - - } /* End of main loop */ -/* Control never reaches here */ - - -/* When compiling to use the heap rather than the stack for recursive calls to -match(), the RRETURN() macro jumps here. The number that is saved in -frame->Xwhere indicates which label we actually want to return to. */ - -#ifdef NO_RECURSE -#define LBL(val) case val: goto L_RM##val; -HEAP_RETURN: -switch (frame->Xwhere) - { - LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8) - LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17) - LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33) - LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52) - LBL(53) LBL(54) -#ifdef SUPPORT_UTF8 - LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30) - LBL(32) LBL(34) LBL(42) LBL(46) -#ifdef SUPPORT_UCP - LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45) -#endif /* SUPPORT_UCP */ -#endif /* SUPPORT_UTF8 */ - default: - DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere)); - return PCRE_ERROR_INTERNAL; - } -#undef LBL -#endif /* NO_RECURSE */ -} - - -/*************************************************************************** -**************************************************************************** - RECURSION IN THE match() FUNCTION - -Undefine all the macros that were defined above to handle this. */ - -#ifdef NO_RECURSE -#undef eptr -#undef ecode -#undef mstart -#undef offset_top -#undef ims -#undef eptrb -#undef flags - -#undef callpat -#undef charptr -#undef data -#undef next -#undef pp -#undef prev -#undef saved_eptr - -#undef new_recursive - -#undef cur_is_word -#undef condition -#undef prev_is_word - -#undef original_ims - -#undef ctype -#undef length -#undef max -#undef min -#undef number -#undef offset -#undef op -#undef save_capture_last -#undef save_offset1 -#undef save_offset2 -#undef save_offset3 -#undef stacksave - -#undef newptrb - -#endif - -/* These two are defined as macros in both cases */ - -#undef fc -#undef fi - -/*************************************************************************** -***************************************************************************/ - - - -/************************************************* -* Execute a Regular Expression * -*************************************************/ - -/* This function applies a compiled re to a subject string and picks out -portions of the string if it matches. Two elements in the vector are set for -each substring: the offsets to the start and end of the substring. - -Arguments: - argument_re points to the compiled expression - extra_data points to extra data or is NULL - subject points to the subject string - length length of subject string (may contain binary zeros) - start_offset where to start in the subject string - options option bits - offsets points to a vector of ints to be filled in with offsets - offsetcount the number of elements in the vector - -Returns: > 0 => success; value is the number of elements filled in - = 0 => success, but offsets is not big enough - -1 => failed to match - < -1 => some kind of unexpected problem -*/ - -PCRE_EXP_DEFN int -pcre_exec(const pcre *argument_re, const pcre_extra *extra_data, - PCRE_SPTR subject, int length, int start_offset, int options, int *offsets, - int offsetcount) -{ -int rc, resetcount, ocount; -int first_byte = -1; -int req_byte = -1; -int req_byte2 = -1; -int newline; -unsigned long int ims; -BOOL using_temporary_offsets = FALSE; -BOOL anchored; -BOOL startline; -BOOL firstline; -BOOL first_byte_caseless = FALSE; -BOOL req_byte_caseless = FALSE; -BOOL utf8; -match_data match_block; -match_data *md = &match_block; -const uschar *tables; -const uschar *start_bits = NULL; -USPTR start_match = (USPTR)subject + start_offset; -USPTR end_subject; -USPTR req_byte_ptr = start_match - 1; - -pcre_study_data internal_study; -const pcre_study_data *study; - -real_pcre internal_re; -const real_pcre *external_re = (const real_pcre *)argument_re; -const real_pcre *re = external_re; - -/* Plausibility checks */ - -if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION; -if (re == NULL || subject == NULL || - (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL; -if (offsetcount < 0) return PCRE_ERROR_BADCOUNT; - -/* Fish out the optional data from the extra_data structure, first setting -the default values. */ - -study = NULL; -md->match_limit = MATCH_LIMIT; -md->match_limit_recursion = MATCH_LIMIT_RECURSION; -md->callout_data = NULL; - -/* The table pointer is always in native byte order. */ - -tables = external_re->tables; - -if (extra_data != NULL) - { - register unsigned int flags = extra_data->flags; - if ((flags & PCRE_EXTRA_STUDY_DATA) != 0) - study = (const pcre_study_data *)extra_data->study_data; - if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0) - md->match_limit = extra_data->match_limit; - if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0) - md->match_limit_recursion = extra_data->match_limit_recursion; - if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0) - md->callout_data = extra_data->callout_data; - if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables; - } - -/* If the exec call supplied NULL for tables, use the inbuilt ones. This -is a feature that makes it possible to save compiled regex and re-use them -in other programs later. */ - -if (tables == NULL) tables = _pcre_default_tables; - -/* Check that the first field in the block is the magic number. If it is not, -test for a regex that was compiled on a host of opposite endianness. If this is -the case, flipped values are put in internal_re and internal_study if there was -study data too. */ - -if (re->magic_number != MAGIC_NUMBER) - { - re = _pcre_try_flipped(re, &internal_re, study, &internal_study); - if (re == NULL) return PCRE_ERROR_BADMAGIC; - if (study != NULL) study = &internal_study; - } - -/* Set up other data */ - -anchored = ((re->options | options) & PCRE_ANCHORED) != 0; -startline = (re->flags & PCRE_STARTLINE) != 0; -firstline = (re->options & PCRE_FIRSTLINE) != 0; - -/* The code starts after the real_pcre block and the capture name table. */ - -md->start_code = (const uschar *)external_re + re->name_table_offset + - re->name_count * re->name_entry_size; - -md->start_subject = (USPTR)subject; -md->start_offset = start_offset; -md->end_subject = md->start_subject + length; -end_subject = md->end_subject; - -md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0; -utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0; -md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0; - -md->notbol = (options & PCRE_NOTBOL) != 0; -md->noteol = (options & PCRE_NOTEOL) != 0; -md->notempty = (options & PCRE_NOTEMPTY) != 0; -md->partial = (options & PCRE_PARTIAL) != 0; -md->hitend = FALSE; - -md->recursive = NULL; /* No recursion at top level */ - -md->lcc = tables + lcc_offset; -md->ctypes = tables + ctypes_offset; - -/* Handle different \R options. */ - -switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) - { - case 0: - if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0) - md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0; - else -#ifdef BSR_ANYCRLF - md->bsr_anycrlf = TRUE; -#else - md->bsr_anycrlf = FALSE; -#endif - break; - - case PCRE_BSR_ANYCRLF: - md->bsr_anycrlf = TRUE; - break; - - case PCRE_BSR_UNICODE: - md->bsr_anycrlf = FALSE; - break; - - default: return PCRE_ERROR_BADNEWLINE; - } - -/* Handle different types of newline. The three bits give eight cases. If -nothing is set at run time, whatever was used at compile time applies. */ - -switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : - (pcre_uint32)options) & PCRE_NEWLINE_BITS) - { - case 0: newline = NEWLINE; break; /* Compile-time default */ - case PCRE_NEWLINE_CR: newline = '\r'; break; - case PCRE_NEWLINE_LF: newline = '\n'; break; - case PCRE_NEWLINE_CR+ - PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break; - case PCRE_NEWLINE_ANY: newline = -1; break; - case PCRE_NEWLINE_ANYCRLF: newline = -2; break; - default: return PCRE_ERROR_BADNEWLINE; - } - -if (newline == -2) - { - md->nltype = NLTYPE_ANYCRLF; - } -else if (newline < 0) - { - md->nltype = NLTYPE_ANY; - } -else - { - md->nltype = NLTYPE_FIXED; - if (newline > 255) - { - md->nllen = 2; - md->nl[0] = (newline >> 8) & 255; - md->nl[1] = newline & 255; - } - else - { - md->nllen = 1; - md->nl[0] = newline; - } - } - -/* Partial matching is supported only for a restricted set of regexes at the -moment. */ - -if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0) - return PCRE_ERROR_BADPARTIAL; - -/* Check a UTF-8 string if required. Unfortunately there's no way of passing -back the character offset. */ - -#ifdef SUPPORT_UTF8 -if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0) - { - if (_pcre_valid_utf8((uschar *)subject, length) >= 0) - return PCRE_ERROR_BADUTF8; - if (start_offset > 0 && start_offset < length) - { - int tb = ((uschar *)subject)[start_offset]; - if (tb > 127) - { - tb &= 0xc0; - if (tb != 0 && tb != 0xc0) return PCRE_ERROR_BADUTF8_OFFSET; - } - } - } -#endif - -/* The ims options can vary during the matching as a result of the presence -of (?ims) items in the pattern. They are kept in a local variable so that -restoring at the exit of a group is easy. */ - -ims = re->options & (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL); - -/* If the expression has got more back references than the offsets supplied can -hold, we get a temporary chunk of working store to use during the matching. -Otherwise, we can use the vector supplied, rounding down its size to a multiple -of 3. */ - -ocount = offsetcount - (offsetcount % 3); - -if (re->top_backref > 0 && re->top_backref >= ocount/3) - { - ocount = re->top_backref * 3 + 3; - md->offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int)); - if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY; - using_temporary_offsets = TRUE; - DPRINTF(("Got memory to hold back references\n")); - } -else md->offset_vector = offsets; - -md->offset_end = ocount; -md->offset_max = (2*ocount)/3; -md->offset_overflow = FALSE; -md->capture_last = -1; - -/* Compute the minimum number of offsets that we need to reset each time. Doing -this makes a huge difference to execution time when there aren't many brackets -in the pattern. */ - -resetcount = 2 + re->top_bracket * 2; -if (resetcount > offsetcount) resetcount = ocount; - -/* Reset the working variable associated with each extraction. These should -never be used unless previously set, but they get saved and restored, and so we -initialize them to avoid reading uninitialized locations. */ - -if (md->offset_vector != NULL) - { - register int *iptr = md->offset_vector + ocount; - register int *iend = iptr - resetcount/2 + 1; - while (--iptr >= iend) *iptr = -1; - } - -/* Set up the first character to match, if available. The first_byte value is -never set for an anchored regular expression, but the anchoring may be forced -at run time, so we have to test for anchoring. The first char may be unset for -an unanchored pattern, of course. If there's no first char and the pattern was -studied, there may be a bitmap of possible first characters. */ - -if (!anchored) - { - if ((re->flags & PCRE_FIRSTSET) != 0) - { - first_byte = re->first_byte & 255; - if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE) - first_byte = md->lcc[first_byte]; - } - else - if (!startline && study != NULL && - (study->options & PCRE_STUDY_MAPPED) != 0) - start_bits = study->start_bits; - } - -/* For anchored or unanchored matches, there may be a "last known required -character" set. */ - -if ((re->flags & PCRE_REQCHSET) != 0) - { - req_byte = re->req_byte & 255; - req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0; - req_byte2 = (tables + fcc_offset)[req_byte]; /* case flipped */ - } - - -/* ==========================================================================*/ - -/* Loop for handling unanchored repeated matching attempts; for anchored regexs -the loop runs just once. */ - -for(;;) - { - USPTR save_end_subject = end_subject; - USPTR new_start_match; - - /* Reset the maximum number of extractions we might see. */ - - if (md->offset_vector != NULL) - { - register int *iptr = md->offset_vector; - register int *iend = iptr + resetcount; - while (iptr < iend) *iptr++ = -1; - } - - /* Advance to a unique first char if possible. If firstline is TRUE, the - start of the match is constrained to the first line of a multiline string. - That is, the match must be before or at the first newline. Implement this by - temporarily adjusting end_subject so that we stop scanning at a newline. If - the match fails at the newline, later code breaks this loop. */ - - if (firstline) - { - USPTR t = start_match; - while (t < md->end_subject && !IS_NEWLINE(t)) t++; - end_subject = t; - } - - /* Now test for a unique first byte */ - - if (first_byte >= 0) - { - if (first_byte_caseless) - while (start_match < end_subject && - md->lcc[*start_match] != first_byte) - { NEXTCHAR(start_match); } - else - while (start_match < end_subject && *start_match != first_byte) - { NEXTCHAR(start_match); } - } - - /* Or to just after a linebreak for a multiline match if possible */ - - else if (startline) - { - if (start_match > md->start_subject + start_offset) - { - while (start_match <= end_subject && !WAS_NEWLINE(start_match)) - { NEXTCHAR(start_match); } - - /* If we have just passed a CR and the newline option is ANY or ANYCRLF, - and we are now at a LF, advance the match position by one more character. - */ - - if (start_match[-1] == '\r' && - (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) && - start_match < end_subject && - *start_match == '\n') - start_match++; - } - } - - /* Or to a non-unique first char after study */ - - else if (start_bits != NULL) - { - while (start_match < end_subject) - { - register unsigned int c = *start_match; - if ((start_bits[c/8] & (1 << (c&7))) == 0) - { NEXTCHAR(start_match); } - else break; - } - } - - /* Restore fudged end_subject */ - - end_subject = save_end_subject; - -#ifdef DEBUG /* Sigh. Some compilers never learn. */ - printf(">>>> Match against: "); - pchars(start_match, end_subject - start_match, TRUE, md); - printf("\n"); -#endif - - /* If req_byte is set, we know that that character must appear in the subject - for the match to succeed. If the first character is set, req_byte must be - later in the subject; otherwise the test starts at the match point. This - optimization can save a huge amount of backtracking in patterns with nested - unlimited repeats that aren't going to match. Writing separate code for - cased/caseless versions makes it go faster, as does using an autoincrement - and backing off on a match. - - HOWEVER: when the subject string is very, very long, searching to its end can - take a long time, and give bad performance on quite ordinary patterns. This - showed up when somebody was matching something like /^\d+C/ on a 32-megabyte - string... so we don't do this when the string is sufficiently long. - - ALSO: this processing is disabled when partial matching is requested. - */ - - if (req_byte >= 0 && - end_subject - start_match < REQ_BYTE_MAX && - !md->partial) - { - register USPTR p = start_match + ((first_byte >= 0)? 1 : 0); - - /* We don't need to repeat the search if we haven't yet reached the - place we found it at last time. */ - - if (p > req_byte_ptr) - { - if (req_byte_caseless) - { - while (p < end_subject) - { - register int pp = *p++; - if (pp == req_byte || pp == req_byte2) { p--; break; } - } - } - else - { - while (p < end_subject) - { - if (*p++ == req_byte) { p--; break; } - } - } - - /* If we can't find the required character, break the matching loop, - forcing a match failure. */ - - if (p >= end_subject) - { - rc = MATCH_NOMATCH; - break; - } - - /* If we have found the required character, save the point where we - found it, so that we don't search again next time round the loop if - the start hasn't passed this character yet. */ - - req_byte_ptr = p; - } - } - - /* OK, we can now run the match. */ - - md->start_match_ptr = start_match; - md->match_call_count = 0; - rc = match(start_match, md->start_code, start_match, 2, md, ims, NULL, 0, 0); - - switch(rc) - { - /* NOMATCH and PRUNE advance by one character. THEN at this level acts - exactly like PRUNE. */ - - case MATCH_NOMATCH: - case MATCH_PRUNE: - case MATCH_THEN: - new_start_match = start_match + 1; -#ifdef SUPPORT_UTF8 - if (utf8) - while(new_start_match < end_subject && (*new_start_match & 0xc0) == 0x80) - new_start_match++; -#endif - break; - - /* SKIP passes back the next starting point explicitly. */ - - case MATCH_SKIP: - new_start_match = md->start_match_ptr; - break; - - /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */ - - case MATCH_COMMIT: - rc = MATCH_NOMATCH; - goto ENDLOOP; - - /* Any other return is some kind of error. */ - - default: - goto ENDLOOP; - } - - /* Control reaches here for the various types of "no match at this point" - result. Reset the code to MATCH_NOMATCH for subsequent checking. */ - - rc = MATCH_NOMATCH; - - /* If PCRE_FIRSTLINE is set, the match must happen before or at the first - newline in the subject (though it may continue over the newline). Therefore, - if we have just failed to match, starting at a newline, do not continue. */ - - if (firstline && IS_NEWLINE(start_match)) break; - - /* Advance to new matching position */ - - start_match = new_start_match; - - /* Break the loop if the pattern is anchored or if we have passed the end of - the subject. */ - - if (anchored || start_match > end_subject) break; - - /* If we have just passed a CR and we are now at a LF, and the pattern does - not contain any explicit matches for \r or \n, and the newline option is CRLF - or ANY or ANYCRLF, advance the match position by one more character. */ - - if (start_match[-1] == '\r' && - start_match < end_subject && - *start_match == '\n' && - (re->flags & PCRE_HASCRORLF) == 0 && - (md->nltype == NLTYPE_ANY || - md->nltype == NLTYPE_ANYCRLF || - md->nllen == 2)) - start_match++; - - } /* End of for(;;) "bumpalong" loop */ - -/* ==========================================================================*/ - -/* We reach here when rc is not MATCH_NOMATCH, or if one of the stopping -conditions is true: - -(1) The pattern is anchored or the match was failed by (*COMMIT); - -(2) We are past the end of the subject; - -(3) PCRE_FIRSTLINE is set and we have failed to match at a newline, because - this option requests that a match occur at or before the first newline in - the subject. - -When we have a match and the offset vector is big enough to deal with any -backreferences, captured substring offsets will already be set up. In the case -where we had to get some local store to hold offsets for backreference -processing, copy those that we can. In this case there need not be overflow if -certain parts of the pattern were not used, even though there are more -capturing parentheses than vector slots. */ - -ENDLOOP: - -if (rc == MATCH_MATCH) - { - if (using_temporary_offsets) - { - if (offsetcount >= 4) - { - memcpy(offsets + 2, md->offset_vector + 2, - (offsetcount - 2) * sizeof(int)); - DPRINTF(("Copied offsets from temporary memory\n")); - } - if (md->end_offset_top > offsetcount) md->offset_overflow = TRUE; - DPRINTF(("Freeing temporary memory\n")); - (pcre_free)(md->offset_vector); - } - - /* Set the return code to the number of captured strings, or 0 if there are - too many to fit into the vector. */ - - rc = md->offset_overflow? 0 : md->end_offset_top/2; - - /* If there is space, set up the whole thing as substring 0. The value of - md->start_match_ptr might be modified if \K was encountered on the success - matching path. */ - - if (offsetcount < 2) rc = 0; else - { - offsets[0] = md->start_match_ptr - md->start_subject; - offsets[1] = md->end_match_ptr - md->start_subject; - } - - DPRINTF((">>>> returning %d\n", rc)); - return rc; - } - -/* Control gets here if there has been an error, or if the overall match -attempt has failed at all permitted starting positions. */ - -if (using_temporary_offsets) - { - DPRINTF(("Freeing temporary memory\n")); - (pcre_free)(md->offset_vector); - } - -if (rc != MATCH_NOMATCH) - { - DPRINTF((">>>> error: returning %d\n", rc)); - return rc; - } -else if (md->partial && md->hitend) - { - DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n")); - return PCRE_ERROR_PARTIAL; - } -else - { - DPRINTF((">>>> returning PCRE_ERROR_NOMATCH\n")); - return PCRE_ERROR_NOMATCH; - } + Copyright (c) 1997-2012 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + +/* This module contains pcre_exec(), the externally visible function that does +pattern matching using an NFA algorithm, trying to mimic Perl as closely as +possible. There are also some static supporting functions. */ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#define NLBLOCK md /* Block containing newline information */ +#define PSSTART start_subject /* Field containing processed string start */ +#define PSEND end_subject /* Field containing processed string end */ + +#include "pcre_internal.h" + +/* Undefine some potentially clashing cpp symbols */ + +#undef min +#undef max + +/* Values for setting in md->match_function_type to indicate two special types +of call to match(). We do it this way to save on using another stack variable, +as stack usage is to be discouraged. */ + +#define MATCH_CONDASSERT 1 /* Called to check a condition assertion */ +#define MATCH_CBEGROUP 2 /* Could-be-empty unlimited repeat group */ + +/* Non-error returns from the match() function. Error returns are externally +defined PCRE_ERROR_xxx codes, which are all negative. */ + +#define MATCH_MATCH 1 +#define MATCH_NOMATCH 0 + +/* Special internal returns from the match() function. Make them sufficiently +negative to avoid the external error codes. */ + +#define MATCH_ACCEPT (-999) +#define MATCH_COMMIT (-998) +#define MATCH_KETRPOS (-997) +#define MATCH_ONCE (-996) +#define MATCH_PRUNE (-995) +#define MATCH_SKIP (-994) +#define MATCH_SKIP_ARG (-993) +#define MATCH_THEN (-992) + +/* Maximum number of ints of offset to save on the stack for recursive calls. +If the offset vector is bigger, malloc is used. This should be a multiple of 3, +because the offset vector is always a multiple of 3 long. */ + +#define REC_STACK_SAVE_MAX 30 + +/* Min and max values for the common repeats; for the maxima, 0 => infinity */ + +static const char rep_min[] = { 0, 0, 1, 1, 0, 0 }; +static const char rep_max[] = { 0, 0, 0, 0, 1, 1 }; + + + +#ifdef PCRE_DEBUG +/************************************************* +* Debugging function to print chars * +*************************************************/ + +/* Print a sequence of chars in printable format, stopping at the end of the +subject if the requested. + +Arguments: + p points to characters + length number to print + is_subject TRUE if printing from within md->start_subject + md pointer to matching data block, if is_subject is TRUE + +Returns: nothing +*/ + +static void +pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md) +{ +unsigned int c; +if (is_subject && length > md->end_subject - p) length = md->end_subject - p; +while (length-- > 0) + if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c); +} +#endif + + + +/************************************************* +* Match a back-reference * +*************************************************/ + +/* Normally, if a back reference hasn't been set, the length that is passed is +negative, so the match always fails. However, in JavaScript compatibility mode, +the length passed is zero. Note that in caseless UTF-8 mode, the number of +subject bytes matched may be different to the number of reference bytes. + +Arguments: + offset index into the offset vector + eptr pointer into the subject + length length of reference to be matched (number of bytes) + md points to match data block + caseless TRUE if caseless + +Returns: >= 0 the number of subject bytes matched + -1 no match + -2 partial match; always given if at end subject +*/ + +static int +match_ref(int offset, register PCRE_PUCHAR eptr, int length, match_data *md, + BOOL caseless) +{ +PCRE_PUCHAR eptr_start = eptr; +register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset]; + +#ifdef PCRE_DEBUG +if (eptr >= md->end_subject) + printf("matching subject "); +else + { + printf("matching subject "); + pchars(eptr, length, TRUE, md); + } +printf(" against backref "); +pchars(p, length, FALSE, md); +printf("\n"); +#endif + +/* Always fail if reference not set (and not JavaScript compatible - in that +case the length is passed as zero). */ + +if (length < 0) return -1; + +/* Separate the caseless case for speed. In UTF-8 mode we can only do this +properly if Unicode properties are supported. Otherwise, we can check only +ASCII characters. */ + +if (caseless) + { +#ifdef SUPPORT_UTF +#ifdef SUPPORT_UCP + if (md->utf) + { + /* Match characters up to the end of the reference. NOTE: the number of + bytes matched may differ, because there are some characters whose upper and + lower case versions code as different numbers of bytes. For example, U+023A + (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8); + a sequence of 3 of the former uses 6 bytes, as does a sequence of two of + the latter. It is important, therefore, to check the length along the + reference, not along the subject (earlier code did this wrong). */ + + PCRE_PUCHAR endptr = p + length; + while (p < endptr) + { + int c, d; + if (eptr >= md->end_subject) return -2; /* Partial match */ + GETCHARINC(c, eptr); + GETCHARINC(d, p); + if (c != d && c != UCD_OTHERCASE(d)) return -1; + } + } + else +#endif +#endif + + /* The same code works when not in UTF-8 mode and in UTF-8 mode when there + is no UCP support. */ + { + while (length-- > 0) + { + if (eptr >= md->end_subject) return -2; /* Partial match */ + if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1; + p++; + eptr++; + } + } + } + +/* In the caseful case, we can just compare the bytes, whether or not we +are in UTF-8 mode. */ + +else + { + while (length-- > 0) + { + if (eptr >= md->end_subject) return -2; /* Partial match */ + if (*p++ != *eptr++) return -1; + } + } + +return (int)(eptr - eptr_start); +} + + + +/*************************************************************************** +**************************************************************************** + RECURSION IN THE match() FUNCTION + +The match() function is highly recursive, though not every recursive call +increases the recursive depth. Nevertheless, some regular expressions can cause +it to recurse to a great depth. I was writing for Unix, so I just let it call +itself recursively. This uses the stack for saving everything that has to be +saved for a recursive call. On Unix, the stack can be large, and this works +fine. + +It turns out that on some non-Unix-like systems there are problems with +programs that use a lot of stack. (This despite the fact that every last chip +has oodles of memory these days, and techniques for extending the stack have +been known for decades.) So.... + +There is a fudge, triggered by defining NO_RECURSE, which avoids recursive +calls by keeping local variables that need to be preserved in blocks of memory +obtained from malloc() instead instead of on the stack. Macros are used to +achieve this so that the actual code doesn't look very different to what it +always used to. + +The original heap-recursive code used longjmp(). However, it seems that this +can be very slow on some operating systems. Following a suggestion from Stan +Switzer, the use of longjmp() has been abolished, at the cost of having to +provide a unique number for each call to RMATCH. There is no way of generating +a sequence of numbers at compile time in C. I have given them names, to make +them stand out more clearly. + +Crude tests on x86 Linux show a small speedup of around 5-8%. However, on +FreeBSD, avoiding longjmp() more than halves the time taken to run the standard +tests. Furthermore, not using longjmp() means that local dynamic variables +don't have indeterminate values; this has meant that the frame size can be +reduced because the result can be "passed back" by straight setting of the +variable instead of being passed in the frame. +**************************************************************************** +***************************************************************************/ + +/* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN +below must be updated in sync. */ + +enum { RM1=1, RM2, RM3, RM4, RM5, RM6, RM7, RM8, RM9, RM10, + RM11, RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20, + RM21, RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30, + RM31, RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40, + RM41, RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50, + RM51, RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60, + RM61, RM62, RM63, RM64, RM65, RM66 }; + +/* These versions of the macros use the stack, as normal. There are debugging +versions and production versions. Note that the "rw" argument of RMATCH isn't +actually used in this definition. */ + +#ifndef NO_RECURSE +#define REGISTER register + +#ifdef PCRE_DEBUG +#define RMATCH(ra,rb,rc,rd,re,rw) \ + { \ + printf("match() called in line %d\n", __LINE__); \ + rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1); \ + printf("to line %d\n", __LINE__); \ + } +#define RRETURN(ra) \ + { \ + printf("match() returned %d from line %d ", ra, __LINE__); \ + return ra; \ + } +#else +#define RMATCH(ra,rb,rc,rd,re,rw) \ + rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1) +#define RRETURN(ra) return ra +#endif + +#else + + +/* These versions of the macros manage a private stack on the heap. Note that +the "rd" argument of RMATCH isn't actually used in this definition. It's the md +argument of match(), which never changes. */ + +#define REGISTER + +#define RMATCH(ra,rb,rc,rd,re,rw)\ + {\ + heapframe *newframe = frame->Xnextframe;\ + if (newframe == NULL)\ + {\ + newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\ + if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\ + newframe->Xnextframe = NULL;\ + frame->Xnextframe = newframe;\ + }\ + frame->Xwhere = rw;\ + newframe->Xeptr = ra;\ + newframe->Xecode = rb;\ + newframe->Xmstart = mstart;\ + newframe->Xoffset_top = rc;\ + newframe->Xeptrb = re;\ + newframe->Xrdepth = frame->Xrdepth + 1;\ + newframe->Xprevframe = frame;\ + frame = newframe;\ + DPRINTF(("restarting from line %d\n", __LINE__));\ + goto HEAP_RECURSE;\ + L_##rw:\ + DPRINTF(("jumped back to line %d\n", __LINE__));\ + } + +#define RRETURN(ra)\ + {\ + heapframe *oldframe = frame;\ + frame = oldframe->Xprevframe;\ + if (frame != NULL)\ + {\ + rrc = ra;\ + goto HEAP_RETURN;\ + }\ + return ra;\ + } + + +/* Structure for remembering the local variables in a private frame */ + +typedef struct heapframe { + struct heapframe *Xprevframe; + struct heapframe *Xnextframe; + + /* Function arguments that may change */ + + PCRE_PUCHAR Xeptr; + const pcre_uchar *Xecode; + PCRE_PUCHAR Xmstart; + int Xoffset_top; + eptrblock *Xeptrb; + unsigned int Xrdepth; + + /* Function local variables */ + + PCRE_PUCHAR Xcallpat; +#ifdef SUPPORT_UTF + PCRE_PUCHAR Xcharptr; +#endif + PCRE_PUCHAR Xdata; + PCRE_PUCHAR Xnext; + PCRE_PUCHAR Xpp; + PCRE_PUCHAR Xprev; + PCRE_PUCHAR Xsaved_eptr; + + recursion_info Xnew_recursive; + + BOOL Xcur_is_word; + BOOL Xcondition; + BOOL Xprev_is_word; + +#ifdef SUPPORT_UCP + int Xprop_type; + int Xprop_value; + int Xprop_fail_result; + int Xoclength; + pcre_uchar Xocchars[6]; +#endif + + int Xcodelink; + int Xctype; + unsigned int Xfc; + int Xfi; + int Xlength; + int Xmax; + int Xmin; + int Xnumber; + int Xoffset; + int Xop; + int Xsave_capture_last; + int Xsave_offset1, Xsave_offset2, Xsave_offset3; + int Xstacksave[REC_STACK_SAVE_MAX]; + + eptrblock Xnewptrb; + + /* Where to jump back to */ + + int Xwhere; + +} heapframe; + +#endif + + +/*************************************************************************** +***************************************************************************/ + + + +/************************************************* +* Match from current position * +*************************************************/ + +/* This function is called recursively in many circumstances. Whenever it +returns a negative (error) response, the outer incarnation must also return the +same response. */ + +/* These macros pack up tests that are used for partial matching, and which +appear several times in the code. We set the "hit end" flag if the pointer is +at the end of the subject and also past the start of the subject (i.e. +something has been matched). For hard partial matching, we then return +immediately. The second one is used when we already know we are past the end of +the subject. */ + +#define CHECK_PARTIAL()\ + if (md->partial != 0 && eptr >= md->end_subject && \ + eptr > md->start_used_ptr) \ + { \ + md->hitend = TRUE; \ + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \ + } + +#define SCHECK_PARTIAL()\ + if (md->partial != 0 && eptr > md->start_used_ptr) \ + { \ + md->hitend = TRUE; \ + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \ + } + + +/* Performance note: It might be tempting to extract commonly used fields from +the md structure (e.g. utf, end_subject) into individual variables to improve +performance. Tests using gcc on a SPARC disproved this; in the first case, it +made performance worse. + +Arguments: + eptr pointer to current character in subject + ecode pointer to current position in compiled code + mstart pointer to the current match start position (can be modified + by encountering \K) + offset_top current top pointer + md pointer to "static" info for the match + eptrb pointer to chain of blocks containing eptr at start of + brackets - for testing for empty matches + rdepth the recursion depth + +Returns: MATCH_MATCH if matched ) these values are >= 0 + MATCH_NOMATCH if failed to match ) + a negative MATCH_xxx value for PRUNE, SKIP, etc + a negative PCRE_ERROR_xxx value if aborted by an error condition + (e.g. stopped by repeated call or recursion limit) +*/ + +static int +match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode, + PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb, + unsigned int rdepth) +{ +/* These variables do not need to be preserved over recursion in this function, +so they can be ordinary variables in all cases. Mark some of them with +"register" because they are used a lot in loops. */ + +register int rrc; /* Returns from recursive calls */ +register int i; /* Used for loops not involving calls to RMATCH() */ +register unsigned int c; /* Character values not kept over RMATCH() calls */ +register BOOL utf; /* Local copy of UTF flag for speed */ + +BOOL minimize, possessive; /* Quantifier options */ +BOOL caseless; +int condcode; + +/* When recursion is not being used, all "local" variables that have to be +preserved over calls to RMATCH() are part of a "frame". We set up the top-level +frame on the stack here; subsequent instantiations are obtained from the heap +whenever RMATCH() does a "recursion". See the macro definitions above. Putting +the top-level on the stack rather than malloc-ing them all gives a performance +boost in many cases where there is not much "recursion". */ + +#ifdef NO_RECURSE +heapframe *frame = (heapframe *)md->match_frames_base; + +/* Copy in the original argument variables */ + +frame->Xeptr = eptr; +frame->Xecode = ecode; +frame->Xmstart = mstart; +frame->Xoffset_top = offset_top; +frame->Xeptrb = eptrb; +frame->Xrdepth = rdepth; + +/* This is where control jumps back to to effect "recursion" */ + +HEAP_RECURSE: + +/* Macros make the argument variables come from the current frame */ + +#define eptr frame->Xeptr +#define ecode frame->Xecode +#define mstart frame->Xmstart +#define offset_top frame->Xoffset_top +#define eptrb frame->Xeptrb +#define rdepth frame->Xrdepth + +/* Ditto for the local variables */ + +#ifdef SUPPORT_UTF +#define charptr frame->Xcharptr +#endif +#define callpat frame->Xcallpat +#define codelink frame->Xcodelink +#define data frame->Xdata +#define next frame->Xnext +#define pp frame->Xpp +#define prev frame->Xprev +#define saved_eptr frame->Xsaved_eptr + +#define new_recursive frame->Xnew_recursive + +#define cur_is_word frame->Xcur_is_word +#define condition frame->Xcondition +#define prev_is_word frame->Xprev_is_word + +#ifdef SUPPORT_UCP +#define prop_type frame->Xprop_type +#define prop_value frame->Xprop_value +#define prop_fail_result frame->Xprop_fail_result +#define oclength frame->Xoclength +#define occhars frame->Xocchars +#endif + +#define ctype frame->Xctype +#define fc frame->Xfc +#define fi frame->Xfi +#define length frame->Xlength +#define max frame->Xmax +#define min frame->Xmin +#define number frame->Xnumber +#define offset frame->Xoffset +#define op frame->Xop +#define save_capture_last frame->Xsave_capture_last +#define save_offset1 frame->Xsave_offset1 +#define save_offset2 frame->Xsave_offset2 +#define save_offset3 frame->Xsave_offset3 +#define stacksave frame->Xstacksave + +#define newptrb frame->Xnewptrb + +/* When recursion is being used, local variables are allocated on the stack and +get preserved during recursion in the normal way. In this environment, fi and +i, and fc and c, can be the same variables. */ + +#else /* NO_RECURSE not defined */ +#define fi i +#define fc c + +/* Many of the following variables are used only in small blocks of the code. +My normal style of coding would have declared them within each of those blocks. +However, in order to accommodate the version of this code that uses an external +"stack" implemented on the heap, it is easier to declare them all here, so the +declarations can be cut out in a block. The only declarations within blocks +below are for variables that do not have to be preserved over a recursive call +to RMATCH(). */ + +#ifdef SUPPORT_UTF +const pcre_uchar *charptr; +#endif +const pcre_uchar *callpat; +const pcre_uchar *data; +const pcre_uchar *next; +PCRE_PUCHAR pp; +const pcre_uchar *prev; +PCRE_PUCHAR saved_eptr; + +recursion_info new_recursive; + +BOOL cur_is_word; +BOOL condition; +BOOL prev_is_word; + +#ifdef SUPPORT_UCP +int prop_type; +int prop_value; +int prop_fail_result; +int oclength; +pcre_uchar occhars[6]; +#endif + +int codelink; +int ctype; +int length; +int max; +int min; +int number; +int offset; +int op; +int save_capture_last; +int save_offset1, save_offset2, save_offset3; +int stacksave[REC_STACK_SAVE_MAX]; + +eptrblock newptrb; + +/* There is a special fudge for calling match() in a way that causes it to +measure the size of its basic stack frame when the stack is being used for +recursion. The second argument (ecode) being NULL triggers this behaviour. It +cannot normally ever be NULL. The return is the negated value of the frame +size. */ + +if (ecode == NULL) + { + if (rdepth == 0) + return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1); + else + { + int len = (char *)&rdepth - (char *)eptr; + return (len > 0)? -len : len; + } + } +#endif /* NO_RECURSE */ + +/* To save space on the stack and in the heap frame, I have doubled up on some +of the local variables that are used only in localised parts of the code, but +still need to be preserved over recursive calls of match(). These macros define +the alternative names that are used. */ + +#define allow_zero cur_is_word +#define cbegroup condition +#define code_offset codelink +#define condassert condition +#define matched_once prev_is_word +#define foc number +#define save_mark data + +/* These statements are here to stop the compiler complaining about unitialized +variables. */ + +#ifdef SUPPORT_UCP +prop_value = 0; +prop_fail_result = 0; +#endif + + +/* This label is used for tail recursion, which is used in a few cases even +when NO_RECURSE is not defined, in order to reduce the amount of stack that is +used. Thanks to Ian Taylor for noticing this possibility and sending the +original patch. */ + +TAIL_RECURSE: + +/* OK, now we can get on with the real code of the function. Recursive calls +are specified by the macro RMATCH and RRETURN is used to return. When +NO_RECURSE is *not* defined, these just turn into a recursive call to match() +and a "return", respectively (possibly with some debugging if PCRE_DEBUG is +defined). However, RMATCH isn't like a function call because it's quite a +complicated macro. It has to be used in one particular way. This shouldn't, +however, impact performance when true recursion is being used. */ + +#ifdef SUPPORT_UTF +utf = md->utf; /* Local copy of the flag */ +#else +utf = FALSE; +#endif + +/* First check that we haven't called match() too many times, or that we +haven't exceeded the recursive call limit. */ + +if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT); +if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT); + +/* At the start of a group with an unlimited repeat that may match an empty +string, the variable md->match_function_type is set to MATCH_CBEGROUP. It is +done this way to save having to use another function argument, which would take +up space on the stack. See also MATCH_CONDASSERT below. + +When MATCH_CBEGROUP is set, add the current subject pointer to the chain of +such remembered pointers, to be checked when we hit the closing ket, in order +to break infinite loops that match no characters. When match() is called in +other circumstances, don't add to the chain. The MATCH_CBEGROUP feature must +NOT be used with tail recursion, because the memory block that is used is on +the stack, so a new one may be required for each match(). */ + +if (md->match_function_type == MATCH_CBEGROUP) + { + newptrb.epb_saved_eptr = eptr; + newptrb.epb_prev = eptrb; + eptrb = &newptrb; + md->match_function_type = 0; + } + +/* Now start processing the opcodes. */ + +for (;;) + { + minimize = possessive = FALSE; + op = *ecode; + + switch(op) + { + case OP_MARK: + md->nomatch_mark = ecode + 2; + md->mark = NULL; /* In case previously set by assertion */ + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md, + eptrb, RM55); + if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) && + md->mark == NULL) md->mark = ecode + 2; + + /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an + argument, and we must check whether that argument matches this MARK's + argument. It is passed back in md->start_match_ptr (an overloading of that + variable). If it does match, we reset that variable to the current subject + position and return MATCH_SKIP. Otherwise, pass back the return code + unaltered. */ + + else if (rrc == MATCH_SKIP_ARG && + STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0) + { + md->start_match_ptr = eptr; + RRETURN(MATCH_SKIP); + } + RRETURN(rrc); + + case OP_FAIL: + RRETURN(MATCH_NOMATCH); + + /* COMMIT overrides PRUNE, SKIP, and THEN */ + + case OP_COMMIT: + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM52); + if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && + rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG && + rrc != MATCH_THEN) + RRETURN(rrc); + RRETURN(MATCH_COMMIT); + + /* PRUNE overrides THEN */ + + case OP_PRUNE: + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM51); + if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); + RRETURN(MATCH_PRUNE); + + case OP_PRUNE_ARG: + md->nomatch_mark = ecode + 2; + md->mark = NULL; /* In case previously set by assertion */ + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md, + eptrb, RM56); + if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) && + md->mark == NULL) md->mark = ecode + 2; + if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); + RRETURN(MATCH_PRUNE); + + /* SKIP overrides PRUNE and THEN */ + + case OP_SKIP: + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM53); + if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN) + RRETURN(rrc); + md->start_match_ptr = eptr; /* Pass back current position */ + RRETURN(MATCH_SKIP); + + /* Note that, for Perl compatibility, SKIP with an argument does NOT set + nomatch_mark. There is a flag that disables this opcode when re-matching a + pattern that ended with a SKIP for which there was not a matching MARK. */ + + case OP_SKIP_ARG: + if (md->ignore_skip_arg) + { + ecode += PRIV(OP_lengths)[*ecode] + ecode[1]; + break; + } + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md, + eptrb, RM57); + if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN) + RRETURN(rrc); + + /* Pass back the current skip name by overloading md->start_match_ptr and + returning the special MATCH_SKIP_ARG return code. This will either be + caught by a matching MARK, or get to the top, where it causes a rematch + with the md->ignore_skip_arg flag set. */ + + md->start_match_ptr = ecode + 2; + RRETURN(MATCH_SKIP_ARG); + + /* For THEN (and THEN_ARG) we pass back the address of the opcode, so that + the branch in which it occurs can be determined. Overload the start of + match pointer to do this. */ + + case OP_THEN: + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM54); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->start_match_ptr = ecode; + RRETURN(MATCH_THEN); + + case OP_THEN_ARG: + md->nomatch_mark = ecode + 2; + md->mark = NULL; /* In case previously set by assertion */ + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, + md, eptrb, RM58); + if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) && + md->mark == NULL) md->mark = ecode + 2; + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->start_match_ptr = ecode; + RRETURN(MATCH_THEN); + + /* Handle an atomic group that does not contain any capturing parentheses. + This can be handled like an assertion. Prior to 8.13, all atomic groups + were handled this way. In 8.13, the code was changed as below for ONCE, so + that backups pass through the group and thereby reset captured values. + However, this uses a lot more stack, so in 8.20, atomic groups that do not + contain any captures generate OP_ONCE_NC, which can be handled in the old, + less stack intensive way. + + Check the alternative branches in turn - the matching won't pass the KET + for this kind of subpattern. If any one branch matches, we carry on as at + the end of a normal bracket, leaving the subject pointer, but resetting + the start-of-match value in case it was changed by \K. */ + + case OP_ONCE_NC: + prev = ecode; + saved_eptr = eptr; + save_mark = md->mark; + do + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64); + if (rrc == MATCH_MATCH) /* Note: _not_ MATCH_ACCEPT */ + { + mstart = md->start_match_ptr; + break; + } + if (rrc == MATCH_THEN) + { + next = ecode + GET(ecode,1); + if (md->start_match_ptr < next && + (*ecode == OP_ALT || *next == OP_ALT)) + rrc = MATCH_NOMATCH; + } + + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode += GET(ecode,1); + md->mark = save_mark; + } + while (*ecode == OP_ALT); + + /* If hit the end of the group (which could be repeated), fail */ + + if (*ecode != OP_ONCE_NC && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH); + + /* Continue as from after the group, updating the offsets high water + mark, since extracts may have been taken. */ + + do ecode += GET(ecode, 1); while (*ecode == OP_ALT); + + offset_top = md->end_offset_top; + eptr = md->end_match_ptr; + + /* For a non-repeating ket, just continue at this level. This also + happens for a repeating ket if no characters were matched in the group. + This is the forcible breaking of infinite loops as implemented in Perl + 5.005. */ + + if (*ecode == OP_KET || eptr == saved_eptr) + { + ecode += 1+LINK_SIZE; + break; + } + + /* The repeating kets try the rest of the pattern or restart from the + preceding bracket, in the appropriate order. The second "call" of match() + uses tail recursion, to avoid using another stack frame. */ + + if (*ecode == OP_KETRMIN) + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM65); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode = prev; + goto TAIL_RECURSE; + } + else /* OP_KETRMAX */ + { + RMATCH(eptr, prev, offset_top, md, eptrb, RM66); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode += 1 + LINK_SIZE; + goto TAIL_RECURSE; + } + /* Control never gets here */ + + /* Handle a capturing bracket, other than those that are possessive with an + unlimited repeat. If there is space in the offset vector, save the current + subject position in the working slot at the top of the vector. We mustn't + change the current values of the data slot, because they may be set from a + previous iteration of this group, and be referred to by a reference inside + the group. A failure to match might occur after the group has succeeded, + if something later on doesn't match. For this reason, we need to restore + the working value and also the values of the final offsets, in case they + were set by a previous iteration of the same bracket. + + If there isn't enough space in the offset vector, treat this as if it were + a non-capturing bracket. Don't worry about setting the flag for the error + case here; that is handled in the code for KET. */ + + case OP_CBRA: + case OP_SCBRA: + number = GET2(ecode, 1+LINK_SIZE); + offset = number << 1; + +#ifdef PCRE_DEBUG + printf("start bracket %d\n", number); + printf("subject="); + pchars(eptr, 16, TRUE, md); + printf("\n"); +#endif + + if (offset < md->offset_max) + { + save_offset1 = md->offset_vector[offset]; + save_offset2 = md->offset_vector[offset+1]; + save_offset3 = md->offset_vector[md->offset_end - number]; + save_capture_last = md->capture_last; + save_mark = md->mark; + + DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3)); + md->offset_vector[md->offset_end - number] = + (int)(eptr - md->start_subject); + + for (;;) + { + if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP; + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM1); + if (rrc == MATCH_ONCE) break; /* Backing up through an atomic group */ + + /* If we backed up to a THEN, check whether it is within the current + branch by comparing the address of the THEN that is passed back with + the end of the branch. If it is within the current branch, and the + branch is one of two or more alternatives (it either starts or ends + with OP_ALT), we have reached the limit of THEN's action, so convert + the return code to NOMATCH, which will cause normal backtracking to + happen from now on. Otherwise, THEN is passed back to an outer + alternative. This implements Perl's treatment of parenthesized groups, + where a group not containing | does not affect the current alternative, + that is, (X) is NOT the same as (X|(*F)). */ + + if (rrc == MATCH_THEN) + { + next = ecode + GET(ecode,1); + if (md->start_match_ptr < next && + (*ecode == OP_ALT || *next == OP_ALT)) + rrc = MATCH_NOMATCH; + } + + /* Anything other than NOMATCH is passed back. */ + + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->capture_last = save_capture_last; + ecode += GET(ecode, 1); + md->mark = save_mark; + if (*ecode != OP_ALT) break; + } + + DPRINTF(("bracket %d failed\n", number)); + md->offset_vector[offset] = save_offset1; + md->offset_vector[offset+1] = save_offset2; + md->offset_vector[md->offset_end - number] = save_offset3; + + /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */ + + RRETURN(rrc); + } + + /* FALL THROUGH ... Insufficient room for saving captured contents. Treat + as a non-capturing bracket. */ + + /* VVVVVVVVVVVVVVVVVVVVVVVVV */ + /* VVVVVVVVVVVVVVVVVVVVVVVVV */ + + DPRINTF(("insufficient capture room: treat as non-capturing\n")); + + /* VVVVVVVVVVVVVVVVVVVVVVVVV */ + /* VVVVVVVVVVVVVVVVVVVVVVVVV */ + + /* Non-capturing or atomic group, except for possessive with unlimited + repeat and ONCE group with no captures. Loop for all the alternatives. + + When we get to the final alternative within the brackets, we used to return + the result of a recursive call to match() whatever happened so it was + possible to reduce stack usage by turning this into a tail recursion, + except in the case of a possibly empty group. However, now that there is + the possiblity of (*THEN) occurring in the final alternative, this + optimization is no longer always possible. + + We can optimize if we know there are no (*THEN)s in the pattern; at present + this is the best that can be done. + + MATCH_ONCE is returned when the end of an atomic group is successfully + reached, but subsequent matching fails. It passes back up the tree (causing + captured values to be reset) until the original atomic group level is + reached. This is tested by comparing md->once_target with the start of the + group. At this point, the return is converted into MATCH_NOMATCH so that + previous backup points can be taken. */ + + case OP_ONCE: + case OP_BRA: + case OP_SBRA: + DPRINTF(("start non-capturing bracket\n")); + + for (;;) + { + if (op >= OP_SBRA || op == OP_ONCE) + md->match_function_type = MATCH_CBEGROUP; + + /* If this is not a possibly empty group, and there are no (*THEN)s in + the pattern, and this is the final alternative, optimize as described + above. */ + + else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT) + { + ecode += PRIV(OP_lengths)[*ecode]; + goto TAIL_RECURSE; + } + + /* In all other cases, we have to make another call to match(). */ + + save_mark = md->mark; + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb, + RM2); + + /* See comment in the code for capturing groups above about handling + THEN. */ + + if (rrc == MATCH_THEN) + { + next = ecode + GET(ecode,1); + if (md->start_match_ptr < next && + (*ecode == OP_ALT || *next == OP_ALT)) + rrc = MATCH_NOMATCH; + } + + if (rrc != MATCH_NOMATCH) + { + if (rrc == MATCH_ONCE) + { + const pcre_uchar *scode = ecode; + if (*scode != OP_ONCE) /* If not at start, find it */ + { + while (*scode == OP_ALT) scode += GET(scode, 1); + scode -= GET(scode, 1); + } + if (md->once_target == scode) rrc = MATCH_NOMATCH; + } + RRETURN(rrc); + } + ecode += GET(ecode, 1); + md->mark = save_mark; + if (*ecode != OP_ALT) break; + } + + RRETURN(MATCH_NOMATCH); + + /* Handle possessive capturing brackets with an unlimited repeat. We come + here from BRAZERO with allow_zero set TRUE. The offset_vector values are + handled similarly to the normal case above. However, the matching is + different. The end of these brackets will always be OP_KETRPOS, which + returns MATCH_KETRPOS without going further in the pattern. By this means + we can handle the group by iteration rather than recursion, thereby + reducing the amount of stack needed. */ + + case OP_CBRAPOS: + case OP_SCBRAPOS: + allow_zero = FALSE; + + POSSESSIVE_CAPTURE: + number = GET2(ecode, 1+LINK_SIZE); + offset = number << 1; + +#ifdef PCRE_DEBUG + printf("start possessive bracket %d\n", number); + printf("subject="); + pchars(eptr, 16, TRUE, md); + printf("\n"); +#endif + + if (offset < md->offset_max) + { + matched_once = FALSE; + code_offset = (int)(ecode - md->start_code); + + save_offset1 = md->offset_vector[offset]; + save_offset2 = md->offset_vector[offset+1]; + save_offset3 = md->offset_vector[md->offset_end - number]; + save_capture_last = md->capture_last; + + DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3)); + + /* Each time round the loop, save the current subject position for use + when the group matches. For MATCH_MATCH, the group has matched, so we + restart it with a new subject starting position, remembering that we had + at least one match. For MATCH_NOMATCH, carry on with the alternatives, as + usual. If we haven't matched any alternatives in any iteration, check to + see if a previous iteration matched. If so, the group has matched; + continue from afterwards. Otherwise it has failed; restore the previous + capture values before returning NOMATCH. */ + + for (;;) + { + md->offset_vector[md->offset_end - number] = + (int)(eptr - md->start_subject); + if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP; + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM63); + if (rrc == MATCH_KETRPOS) + { + offset_top = md->end_offset_top; + eptr = md->end_match_ptr; + ecode = md->start_code + code_offset; + save_capture_last = md->capture_last; + matched_once = TRUE; + continue; + } + + /* See comment in the code for capturing groups above about handling + THEN. */ + + if (rrc == MATCH_THEN) + { + next = ecode + GET(ecode,1); + if (md->start_match_ptr < next && + (*ecode == OP_ALT || *next == OP_ALT)) + rrc = MATCH_NOMATCH; + } + + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->capture_last = save_capture_last; + ecode += GET(ecode, 1); + if (*ecode != OP_ALT) break; + } + + if (!matched_once) + { + md->offset_vector[offset] = save_offset1; + md->offset_vector[offset+1] = save_offset2; + md->offset_vector[md->offset_end - number] = save_offset3; + } + + if (allow_zero || matched_once) + { + ecode += 1 + LINK_SIZE; + break; + } + + RRETURN(MATCH_NOMATCH); + } + + /* FALL THROUGH ... Insufficient room for saving captured contents. Treat + as a non-capturing bracket. */ + + /* VVVVVVVVVVVVVVVVVVVVVVVVV */ + /* VVVVVVVVVVVVVVVVVVVVVVVVV */ + + DPRINTF(("insufficient capture room: treat as non-capturing\n")); + + /* VVVVVVVVVVVVVVVVVVVVVVVVV */ + /* VVVVVVVVVVVVVVVVVVVVVVVVV */ + + /* Non-capturing possessive bracket with unlimited repeat. We come here + from BRAZERO with allow_zero = TRUE. The code is similar to the above, + without the capturing complication. It is written out separately for speed + and cleanliness. */ + + case OP_BRAPOS: + case OP_SBRAPOS: + allow_zero = FALSE; + + POSSESSIVE_NON_CAPTURE: + matched_once = FALSE; + code_offset = (int)(ecode - md->start_code); + + for (;;) + { + if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP; + RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, + eptrb, RM48); + if (rrc == MATCH_KETRPOS) + { + offset_top = md->end_offset_top; + eptr = md->end_match_ptr; + ecode = md->start_code + code_offset; + matched_once = TRUE; + continue; + } + + /* See comment in the code for capturing groups above about handling + THEN. */ + + if (rrc == MATCH_THEN) + { + next = ecode + GET(ecode,1); + if (md->start_match_ptr < next && + (*ecode == OP_ALT || *next == OP_ALT)) + rrc = MATCH_NOMATCH; + } + + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode += GET(ecode, 1); + if (*ecode != OP_ALT) break; + } + + if (matched_once || allow_zero) + { + ecode += 1 + LINK_SIZE; + break; + } + RRETURN(MATCH_NOMATCH); + + /* Control never reaches here. */ + + /* Conditional group: compilation checked that there are no more than + two branches. If the condition is false, skipping the first branch takes us + past the end if there is only one branch, but that's OK because that is + exactly what going to the ket would do. */ + + case OP_COND: + case OP_SCOND: + codelink = GET(ecode, 1); + + /* Because of the way auto-callout works during compile, a callout item is + inserted between OP_COND and an assertion condition. */ + + if (ecode[LINK_SIZE+1] == OP_CALLOUT) + { + if (PUBL(callout) != NULL) + { + PUBL(callout_block) cb; + cb.version = 2; /* Version 1 of the callout block */ + cb.callout_number = ecode[LINK_SIZE+2]; + cb.offset_vector = md->offset_vector; +#ifdef COMPILE_PCRE8 + cb.subject = (PCRE_SPTR)md->start_subject; +#else + cb.subject = (PCRE_SPTR16)md->start_subject; +#endif + cb.subject_length = (int)(md->end_subject - md->start_subject); + cb.start_match = (int)(mstart - md->start_subject); + cb.current_position = (int)(eptr - md->start_subject); + cb.pattern_position = GET(ecode, LINK_SIZE + 3); + cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE); + cb.capture_top = offset_top/2; + cb.capture_last = md->capture_last; + cb.callout_data = md->callout_data; + cb.mark = md->nomatch_mark; + if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH); + if (rrc < 0) RRETURN(rrc); + } + ecode += PRIV(OP_lengths)[OP_CALLOUT]; + } + + condcode = ecode[LINK_SIZE+1]; + + /* Now see what the actual condition is */ + + if (condcode == OP_RREF || condcode == OP_NRREF) /* Recursion test */ + { + if (md->recursive == NULL) /* Not recursing => FALSE */ + { + condition = FALSE; + ecode += GET(ecode, 1); + } + else + { + int recno = GET2(ecode, LINK_SIZE + 2); /* Recursion group number*/ + condition = (recno == RREF_ANY || recno == md->recursive->group_num); + + /* If the test is for recursion into a specific subpattern, and it is + false, but the test was set up by name, scan the table to see if the + name refers to any other numbers, and test them. The condition is true + if any one is set. */ + + if (!condition && condcode == OP_NRREF) + { + pcre_uchar *slotA = md->name_table; + for (i = 0; i < md->name_count; i++) + { + if (GET2(slotA, 0) == recno) break; + slotA += md->name_entry_size; + } + + /* Found a name for the number - there can be only one; duplicate + names for different numbers are allowed, but not vice versa. First + scan down for duplicates. */ + + if (i < md->name_count) + { + pcre_uchar *slotB = slotA; + while (slotB > md->name_table) + { + slotB -= md->name_entry_size; + if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0) + { + condition = GET2(slotB, 0) == md->recursive->group_num; + if (condition) break; + } + else break; + } + + /* Scan up for duplicates */ + + if (!condition) + { + slotB = slotA; + for (i++; i < md->name_count; i++) + { + slotB += md->name_entry_size; + if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0) + { + condition = GET2(slotB, 0) == md->recursive->group_num; + if (condition) break; + } + else break; + } + } + } + } + + /* Chose branch according to the condition */ + + ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1); + } + } + + else if (condcode == OP_CREF || condcode == OP_NCREF) /* Group used test */ + { + offset = GET2(ecode, LINK_SIZE+2) << 1; /* Doubled ref number */ + condition = offset < offset_top && md->offset_vector[offset] >= 0; + + /* If the numbered capture is unset, but the reference was by name, + scan the table to see if the name refers to any other numbers, and test + them. The condition is true if any one is set. This is tediously similar + to the code above, but not close enough to try to amalgamate. */ + + if (!condition && condcode == OP_NCREF) + { + int refno = offset >> 1; + pcre_uchar *slotA = md->name_table; + + for (i = 0; i < md->name_count; i++) + { + if (GET2(slotA, 0) == refno) break; + slotA += md->name_entry_size; + } + + /* Found a name for the number - there can be only one; duplicate names + for different numbers are allowed, but not vice versa. First scan down + for duplicates. */ + + if (i < md->name_count) + { + pcre_uchar *slotB = slotA; + while (slotB > md->name_table) + { + slotB -= md->name_entry_size; + if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0) + { + offset = GET2(slotB, 0) << 1; + condition = offset < offset_top && + md->offset_vector[offset] >= 0; + if (condition) break; + } + else break; + } + + /* Scan up for duplicates */ + + if (!condition) + { + slotB = slotA; + for (i++; i < md->name_count; i++) + { + slotB += md->name_entry_size; + if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0) + { + offset = GET2(slotB, 0) << 1; + condition = offset < offset_top && + md->offset_vector[offset] >= 0; + if (condition) break; + } + else break; + } + } + } + } + + /* Chose branch according to the condition */ + + ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1); + } + + else if (condcode == OP_DEF) /* DEFINE - always false */ + { + condition = FALSE; + ecode += GET(ecode, 1); + } + + /* The condition is an assertion. Call match() to evaluate it - setting + md->match_function_type to MATCH_CONDASSERT causes it to stop at the end of + an assertion. */ + + else + { + md->match_function_type = MATCH_CONDASSERT; + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3); + if (rrc == MATCH_MATCH) + { + if (md->end_offset_top > offset_top) + offset_top = md->end_offset_top; /* Captures may have happened */ + condition = TRUE; + ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2); + while (*ecode == OP_ALT) ecode += GET(ecode, 1); + } + + /* PCRE doesn't allow the effect of (*THEN) to escape beyond an + assertion; it is therefore treated as NOMATCH. */ + + else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) + { + RRETURN(rrc); /* Need braces because of following else */ + } + else + { + condition = FALSE; + ecode += codelink; + } + } + + /* We are now at the branch that is to be obeyed. As there is only one, can + use tail recursion to avoid using another stack frame, except when there is + unlimited repeat of a possibly empty group. In the latter case, a recursive + call to match() is always required, unless the second alternative doesn't + exist, in which case we can just plough on. Note that, for compatibility + with Perl, the | in a conditional group is NOT treated as creating two + alternatives. If a THEN is encountered in the branch, it propagates out to + the enclosing alternative (unless nested in a deeper set of alternatives, + of course). */ + + if (condition || *ecode == OP_ALT) + { + if (op != OP_SCOND) + { + ecode += 1 + LINK_SIZE; + goto TAIL_RECURSE; + } + + md->match_function_type = MATCH_CBEGROUP; + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49); + RRETURN(rrc); + } + + /* Condition false & no alternative; continue after the group. */ + + else + { + ecode += 1 + LINK_SIZE; + } + break; + + + /* Before OP_ACCEPT there may be any number of OP_CLOSE opcodes, + to close any currently open capturing brackets. */ + + case OP_CLOSE: + number = GET2(ecode, 1); + offset = number << 1; + +#ifdef PCRE_DEBUG + printf("end bracket %d at *ACCEPT", number); + printf("\n"); +#endif + + md->capture_last = number; + if (offset >= md->offset_max) md->offset_overflow = TRUE; else + { + md->offset_vector[offset] = + md->offset_vector[md->offset_end - number]; + md->offset_vector[offset+1] = (int)(eptr - md->start_subject); + if (offset_top <= offset) offset_top = offset + 2; + } + ecode += 1 + IMM2_SIZE; + break; + + + /* End of the pattern, either real or forced. */ + + case OP_END: + case OP_ACCEPT: + case OP_ASSERT_ACCEPT: + + /* If we have matched an empty string, fail if not in an assertion and not + in a recursion if either PCRE_NOTEMPTY is set, or if PCRE_NOTEMPTY_ATSTART + is set and we have matched at the start of the subject. In both cases, + backtracking will then try other alternatives, if any. */ + + if (eptr == mstart && op != OP_ASSERT_ACCEPT && + md->recursive == NULL && + (md->notempty || + (md->notempty_atstart && + mstart == md->start_subject + md->start_offset))) + RRETURN(MATCH_NOMATCH); + + /* Otherwise, we have a match. */ + + md->end_match_ptr = eptr; /* Record where we ended */ + md->end_offset_top = offset_top; /* and how many extracts were taken */ + md->start_match_ptr = mstart; /* and the start (\K can modify) */ + + /* For some reason, the macros don't work properly if an expression is + given as the argument to RRETURN when the heap is in use. */ + + rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT; + RRETURN(rrc); + + /* Assertion brackets. Check the alternative branches in turn - the + matching won't pass the KET for an assertion. If any one branch matches, + the assertion is true. Lookbehind assertions have an OP_REVERSE item at the + start of each branch to move the current point backwards, so the code at + this level is identical to the lookahead case. When the assertion is part + of a condition, we want to return immediately afterwards. The caller of + this incarnation of the match() function will have set MATCH_CONDASSERT in + md->match_function type, and one of these opcodes will be the first opcode + that is processed. We use a local variable that is preserved over calls to + match() to remember this case. */ + + case OP_ASSERT: + case OP_ASSERTBACK: + save_mark = md->mark; + if (md->match_function_type == MATCH_CONDASSERT) + { + condassert = TRUE; + md->match_function_type = 0; + } + else condassert = FALSE; + + do + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4); + if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) + { + mstart = md->start_match_ptr; /* In case \K reset it */ + break; + } + md->mark = save_mark; + + /* A COMMIT failure must fail the entire assertion, without trying any + subsequent branches. */ + + if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH); + + /* PCRE does not allow THEN to escape beyond an assertion; it + is treated as NOMATCH. */ + + if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); + ecode += GET(ecode, 1); + } + while (*ecode == OP_ALT); + + if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH); + + /* If checking an assertion for a condition, return MATCH_MATCH. */ + + if (condassert) RRETURN(MATCH_MATCH); + + /* Continue from after the assertion, updating the offsets high water + mark, since extracts may have been taken during the assertion. */ + + do ecode += GET(ecode,1); while (*ecode == OP_ALT); + ecode += 1 + LINK_SIZE; + offset_top = md->end_offset_top; + continue; + + /* Negative assertion: all branches must fail to match. Encountering SKIP, + PRUNE, or COMMIT means we must assume failure without checking subsequent + branches. */ + + case OP_ASSERT_NOT: + case OP_ASSERTBACK_NOT: + save_mark = md->mark; + if (md->match_function_type == MATCH_CONDASSERT) + { + condassert = TRUE; + md->match_function_type = 0; + } + else condassert = FALSE; + + do + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5); + md->mark = save_mark; + if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH); + if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT) + { + do ecode += GET(ecode,1); while (*ecode == OP_ALT); + break; + } + + /* PCRE does not allow THEN to escape beyond an assertion; it is treated + as NOMATCH. */ + + if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); + ecode += GET(ecode,1); + } + while (*ecode == OP_ALT); + + if (condassert) RRETURN(MATCH_MATCH); /* Condition assertion */ + + ecode += 1 + LINK_SIZE; + continue; + + /* Move the subject pointer back. This occurs only at the start of + each branch of a lookbehind assertion. If we are too close to the start to + move back, this match function fails. When working with UTF-8 we move + back a number of characters, not bytes. */ + + case OP_REVERSE: +#ifdef SUPPORT_UTF + if (utf) + { + i = GET(ecode, 1); + while (i-- > 0) + { + eptr--; + if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH); + BACKCHAR(eptr); + } + } + else +#endif + + /* No UTF-8 support, or not in UTF-8 mode: count is byte count */ + + { + eptr -= GET(ecode, 1); + if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH); + } + + /* Save the earliest consulted character, then skip to next op code */ + + if (eptr < md->start_used_ptr) md->start_used_ptr = eptr; + ecode += 1 + LINK_SIZE; + break; + + /* The callout item calls an external function, if one is provided, passing + details of the match so far. This is mainly for debugging, though the + function is able to force a failure. */ + + case OP_CALLOUT: + if (PUBL(callout) != NULL) + { + PUBL(callout_block) cb; + cb.version = 2; /* Version 1 of the callout block */ + cb.callout_number = ecode[1]; + cb.offset_vector = md->offset_vector; +#ifdef COMPILE_PCRE8 + cb.subject = (PCRE_SPTR)md->start_subject; +#else + cb.subject = (PCRE_SPTR16)md->start_subject; +#endif + cb.subject_length = (int)(md->end_subject - md->start_subject); + cb.start_match = (int)(mstart - md->start_subject); + cb.current_position = (int)(eptr - md->start_subject); + cb.pattern_position = GET(ecode, 2); + cb.next_item_length = GET(ecode, 2 + LINK_SIZE); + cb.capture_top = offset_top/2; + cb.capture_last = md->capture_last; + cb.callout_data = md->callout_data; + cb.mark = md->nomatch_mark; + if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH); + if (rrc < 0) RRETURN(rrc); + } + ecode += 2 + 2*LINK_SIZE; + break; + + /* Recursion either matches the current regex, or some subexpression. The + offset data is the offset to the starting bracket from the start of the + whole pattern. (This is so that it works from duplicated subpatterns.) + + The state of the capturing groups is preserved over recursion, and + re-instated afterwards. We don't know how many are started and not yet + finished (offset_top records the completed total) so we just have to save + all the potential data. There may be up to 65535 such values, which is too + large to put on the stack, but using malloc for small numbers seems + expensive. As a compromise, the stack is used when there are no more than + REC_STACK_SAVE_MAX values to store; otherwise malloc is used. + + There are also other values that have to be saved. We use a chained + sequence of blocks that actually live on the stack. Thanks to Robin Houston + for the original version of this logic. It has, however, been hacked around + a lot, so he is not to blame for the current way it works. */ + + case OP_RECURSE: + { + recursion_info *ri; + int recno; + + callpat = md->start_code + GET(ecode, 1); + recno = (callpat == md->start_code)? 0 : + GET2(callpat, 1 + LINK_SIZE); + + /* Check for repeating a recursion without advancing the subject pointer. + This should catch convoluted mutual recursions. (Some simple cases are + caught at compile time.) */ + + for (ri = md->recursive; ri != NULL; ri = ri->prevrec) + if (recno == ri->group_num && eptr == ri->subject_position) + RRETURN(PCRE_ERROR_RECURSELOOP); + + /* Add to "recursing stack" */ + + new_recursive.group_num = recno; + new_recursive.subject_position = eptr; + new_recursive.prevrec = md->recursive; + md->recursive = &new_recursive; + + /* Where to continue from afterwards */ + + ecode += 1 + LINK_SIZE; + + /* Now save the offset data */ + + new_recursive.saved_max = md->offset_end; + if (new_recursive.saved_max <= REC_STACK_SAVE_MAX) + new_recursive.offset_save = stacksave; + else + { + new_recursive.offset_save = + (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int)); + if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY); + } + memcpy(new_recursive.offset_save, md->offset_vector, + new_recursive.saved_max * sizeof(int)); + + /* OK, now we can do the recursion. After processing each alternative, + restore the offset data. If there were nested recursions, md->recursive + might be changed, so reset it before looping. */ + + DPRINTF(("Recursing into group %d\n", new_recursive.group_num)); + cbegroup = (*callpat >= OP_SBRA); + do + { + if (cbegroup) md->match_function_type = MATCH_CBEGROUP; + RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top, + md, eptrb, RM6); + memcpy(md->offset_vector, new_recursive.offset_save, + new_recursive.saved_max * sizeof(int)); + md->recursive = new_recursive.prevrec; + if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) + { + DPRINTF(("Recursion matched\n")); + if (new_recursive.offset_save != stacksave) + (PUBL(free))(new_recursive.offset_save); + + /* Set where we got to in the subject, and reset the start in case + it was changed by \K. This *is* propagated back out of a recursion, + for Perl compatibility. */ + + eptr = md->end_match_ptr; + mstart = md->start_match_ptr; + goto RECURSION_MATCHED; /* Exit loop; end processing */ + } + + /* PCRE does not allow THEN or COMMIT to escape beyond a recursion; it + is treated as NOMATCH. */ + + else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN && + rrc != MATCH_COMMIT) + { + DPRINTF(("Recursion gave error %d\n", rrc)); + if (new_recursive.offset_save != stacksave) + (PUBL(free))(new_recursive.offset_save); + RRETURN(rrc); + } + + md->recursive = &new_recursive; + callpat += GET(callpat, 1); + } + while (*callpat == OP_ALT); + + DPRINTF(("Recursion didn't match\n")); + md->recursive = new_recursive.prevrec; + if (new_recursive.offset_save != stacksave) + (PUBL(free))(new_recursive.offset_save); + RRETURN(MATCH_NOMATCH); + } + + RECURSION_MATCHED: + break; + + /* An alternation is the end of a branch; scan along to find the end of the + bracketed group and go to there. */ + + case OP_ALT: + do ecode += GET(ecode,1); while (*ecode == OP_ALT); + break; + + /* BRAZERO, BRAMINZERO and SKIPZERO occur just before a bracket group, + indicating that it may occur zero times. It may repeat infinitely, or not + at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets + with fixed upper repeat limits are compiled as a number of copies, with the + optional ones preceded by BRAZERO or BRAMINZERO. */ + + case OP_BRAZERO: + next = ecode + 1; + RMATCH(eptr, next, offset_top, md, eptrb, RM10); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + do next += GET(next, 1); while (*next == OP_ALT); + ecode = next + 1 + LINK_SIZE; + break; + + case OP_BRAMINZERO: + next = ecode + 1; + do next += GET(next, 1); while (*next == OP_ALT); + RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, eptrb, RM11); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + ecode++; + break; + + case OP_SKIPZERO: + next = ecode+1; + do next += GET(next,1); while (*next == OP_ALT); + ecode = next + 1 + LINK_SIZE; + break; + + /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything + here; just jump to the group, with allow_zero set TRUE. */ + + case OP_BRAPOSZERO: + op = *(++ecode); + allow_zero = TRUE; + if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE; + goto POSSESSIVE_NON_CAPTURE; + + /* End of a group, repeated or non-repeating. */ + + case OP_KET: + case OP_KETRMIN: + case OP_KETRMAX: + case OP_KETRPOS: + prev = ecode - GET(ecode, 1); + + /* If this was a group that remembered the subject start, in order to break + infinite repeats of empty string matches, retrieve the subject start from + the chain. Otherwise, set it NULL. */ + + if (*prev >= OP_SBRA || *prev == OP_ONCE) + { + saved_eptr = eptrb->epb_saved_eptr; /* Value at start of group */ + eptrb = eptrb->epb_prev; /* Backup to previous group */ + } + else saved_eptr = NULL; + + /* If we are at the end of an assertion group or a non-capturing atomic + group, stop matching and return MATCH_MATCH, but record the current high + water mark for use by positive assertions. We also need to record the match + start in case it was changed by \K. */ + + if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) || + *prev == OP_ONCE_NC) + { + md->end_match_ptr = eptr; /* For ONCE_NC */ + md->end_offset_top = offset_top; + md->start_match_ptr = mstart; + RRETURN(MATCH_MATCH); /* Sets md->mark */ + } + + /* For capturing groups we have to check the group number back at the start + and if necessary complete handling an extraction by setting the offsets and + bumping the high water mark. Whole-pattern recursion is coded as a recurse + into group 0, so it won't be picked up here. Instead, we catch it when the + OP_END is reached. Other recursion is handled here. We just have to record + the current subject position and start match pointer and give a MATCH + return. */ + + if (*prev == OP_CBRA || *prev == OP_SCBRA || + *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS) + { + number = GET2(prev, 1+LINK_SIZE); + offset = number << 1; + +#ifdef PCRE_DEBUG + printf("end bracket %d", number); + printf("\n"); +#endif + + /* Handle a recursively called group. */ + + if (md->recursive != NULL && md->recursive->group_num == number) + { + md->end_match_ptr = eptr; + md->start_match_ptr = mstart; + RRETURN(MATCH_MATCH); + } + + /* Deal with capturing */ + + md->capture_last = number; + if (offset >= md->offset_max) md->offset_overflow = TRUE; else + { + /* If offset is greater than offset_top, it means that we are + "skipping" a capturing group, and that group's offsets must be marked + unset. In earlier versions of PCRE, all the offsets were unset at the + start of matching, but this doesn't work because atomic groups and + assertions can cause a value to be set that should later be unset. + Example: matching /(?>(a))b|(a)c/ against "ac". This sets group 1 as + part of the atomic group, but this is not on the final matching path, + so must be unset when 2 is set. (If there is no group 2, there is no + problem, because offset_top will then be 2, indicating no capture.) */ + + if (offset > offset_top) + { + register int *iptr = md->offset_vector + offset_top; + register int *iend = md->offset_vector + offset; + while (iptr < iend) *iptr++ = -1; + } + + /* Now make the extraction */ + + md->offset_vector[offset] = + md->offset_vector[md->offset_end - number]; + md->offset_vector[offset+1] = (int)(eptr - md->start_subject); + if (offset_top <= offset) offset_top = offset + 2; + } + } + + /* For an ordinary non-repeating ket, just continue at this level. This + also happens for a repeating ket if no characters were matched in the + group. This is the forcible breaking of infinite loops as implemented in + Perl 5.005. For a non-repeating atomic group that includes captures, + establish a backup point by processing the rest of the pattern at a lower + level. If this results in a NOMATCH return, pass MATCH_ONCE back to the + original OP_ONCE level, thereby bypassing intermediate backup points, but + resetting any captures that happened along the way. */ + + if (*ecode == OP_KET || eptr == saved_eptr) + { + if (*prev == OP_ONCE) + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->once_target = prev; /* Level at which to change to MATCH_NOMATCH */ + RRETURN(MATCH_ONCE); + } + ecode += 1 + LINK_SIZE; /* Carry on at this level */ + break; + } + + /* OP_KETRPOS is a possessive repeating ket. Remember the current position, + and return the MATCH_KETRPOS. This makes it possible to do the repeats one + at a time from the outer level, thus saving stack. */ + + if (*ecode == OP_KETRPOS) + { + md->end_match_ptr = eptr; + md->end_offset_top = offset_top; + RRETURN(MATCH_KETRPOS); + } + + /* The normal repeating kets try the rest of the pattern or restart from + the preceding bracket, in the appropriate order. In the second case, we can + use tail recursion to avoid using another stack frame, unless we have an + an atomic group or an unlimited repeat of a group that can match an empty + string. */ + + if (*ecode == OP_KETRMIN) + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM7); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (*prev == OP_ONCE) + { + RMATCH(eptr, prev, offset_top, md, eptrb, RM8); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->once_target = prev; /* Level at which to change to MATCH_NOMATCH */ + RRETURN(MATCH_ONCE); + } + if (*prev >= OP_SBRA) /* Could match an empty string */ + { + RMATCH(eptr, prev, offset_top, md, eptrb, RM50); + RRETURN(rrc); + } + ecode = prev; + goto TAIL_RECURSE; + } + else /* OP_KETRMAX */ + { + RMATCH(eptr, prev, offset_top, md, eptrb, RM13); + if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH; + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (*prev == OP_ONCE) + { + RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM9); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + md->once_target = prev; + RRETURN(MATCH_ONCE); + } + ecode += 1 + LINK_SIZE; + goto TAIL_RECURSE; + } + /* Control never gets here */ + + /* Not multiline mode: start of subject assertion, unless notbol. */ + + case OP_CIRC: + if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH); + + /* Start of subject assertion */ + + case OP_SOD: + if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH); + ecode++; + break; + + /* Multiline mode: start of subject unless notbol, or after any newline. */ + + case OP_CIRCM: + if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH); + if (eptr != md->start_subject && + (eptr == md->end_subject || !WAS_NEWLINE(eptr))) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + /* Start of match assertion */ + + case OP_SOM: + if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH); + ecode++; + break; + + /* Reset the start of match point */ + + case OP_SET_SOM: + mstart = eptr; + ecode++; + break; + + /* Multiline mode: assert before any newline, or before end of subject + unless noteol is set. */ + + case OP_DOLLM: + if (eptr < md->end_subject) + { + if (!IS_NEWLINE(eptr)) + { + if (md->partial != 0 && + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + *eptr == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + RRETURN(MATCH_NOMATCH); + } + } + else + { + if (md->noteol) RRETURN(MATCH_NOMATCH); + SCHECK_PARTIAL(); + } + ecode++; + break; + + /* Not multiline mode: assert before a terminating newline or before end of + subject unless noteol is set. */ + + case OP_DOLL: + if (md->noteol) RRETURN(MATCH_NOMATCH); + if (!md->endonly) goto ASSERT_NL_OR_EOS; + + /* ... else fall through for endonly */ + + /* End of subject assertion (\z) */ + + case OP_EOD: + if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH); + SCHECK_PARTIAL(); + ecode++; + break; + + /* End of subject or ending \n assertion (\Z) */ + + case OP_EODN: + ASSERT_NL_OR_EOS: + if (eptr < md->end_subject && + (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen)) + { + if (md->partial != 0 && + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + *eptr == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + RRETURN(MATCH_NOMATCH); + } + + /* Either at end of string or \n before end. */ + + SCHECK_PARTIAL(); + ecode++; + break; + + /* Word boundary assertions */ + + case OP_NOT_WORD_BOUNDARY: + case OP_WORD_BOUNDARY: + { + + /* Find out if the previous and current characters are "word" characters. + It takes a bit more work in UTF-8 mode. Characters > 255 are assumed to + be "non-word" characters. Remember the earliest consulted character for + partial matching. */ + +#ifdef SUPPORT_UTF + if (utf) + { + /* Get status of previous character */ + + if (eptr == md->start_subject) prev_is_word = FALSE; else + { + PCRE_PUCHAR lastptr = eptr - 1; + BACKCHAR(lastptr); + if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr; + GETCHAR(c, lastptr); +#ifdef SUPPORT_UCP + if (md->use_ucp) + { + if (c == '_') prev_is_word = TRUE; else + { + int cat = UCD_CATEGORY(c); + prev_is_word = (cat == ucp_L || cat == ucp_N); + } + } + else +#endif + prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0; + } + + /* Get status of next character */ + + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + cur_is_word = FALSE; + } + else + { + GETCHAR(c, eptr); +#ifdef SUPPORT_UCP + if (md->use_ucp) + { + if (c == '_') cur_is_word = TRUE; else + { + int cat = UCD_CATEGORY(c); + cur_is_word = (cat == ucp_L || cat == ucp_N); + } + } + else +#endif + cur_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0; + } + } + else +#endif + + /* Not in UTF-8 mode, but we may still have PCRE_UCP set, and for + consistency with the behaviour of \w we do use it in this case. */ + + { + /* Get status of previous character */ + + if (eptr == md->start_subject) prev_is_word = FALSE; else + { + if (eptr <= md->start_used_ptr) md->start_used_ptr = eptr - 1; +#ifdef SUPPORT_UCP + if (md->use_ucp) + { + c = eptr[-1]; + if (c == '_') prev_is_word = TRUE; else + { + int cat = UCD_CATEGORY(c); + prev_is_word = (cat == ucp_L || cat == ucp_N); + } + } + else +#endif + prev_is_word = MAX_255(eptr[-1]) + && ((md->ctypes[eptr[-1]] & ctype_word) != 0); + } + + /* Get status of next character */ + + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + cur_is_word = FALSE; + } + else +#ifdef SUPPORT_UCP + if (md->use_ucp) + { + c = *eptr; + if (c == '_') cur_is_word = TRUE; else + { + int cat = UCD_CATEGORY(c); + cur_is_word = (cat == ucp_L || cat == ucp_N); + } + } + else +#endif + cur_is_word = MAX_255(*eptr) + && ((md->ctypes[*eptr] & ctype_word) != 0); + } + + /* Now see if the situation is what we want */ + + if ((*ecode++ == OP_WORD_BOUNDARY)? + cur_is_word == prev_is_word : cur_is_word != prev_is_word) + RRETURN(MATCH_NOMATCH); + } + break; + + /* Match any single character type except newline; have to take care with + CRLF newlines and partial matching. */ + + case OP_ANY: + if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); + if (md->partial != 0 && + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + *eptr == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + + /* Fall through */ + + /* Match any single character whatsoever. */ + + case OP_ALLANY: + if (eptr >= md->end_subject) /* DO NOT merge the eptr++ here; it must */ + { /* not be updated before SCHECK_PARTIAL. */ + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr++; +#ifdef SUPPORT_UTF + if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); +#endif + ecode++; + break; + + /* Match a single byte, even in UTF-8 mode. This opcode really does match + any byte, even newline, independent of the setting of PCRE_DOTALL. */ + + case OP_ANYBYTE: + if (eptr >= md->end_subject) /* DO NOT merge the eptr++ here; it must */ + { /* not be updated before SCHECK_PARTIAL. */ + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr++; + ecode++; + break; + + case OP_NOT_DIGIT: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ( +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + c < 256 && +#endif + (md->ctypes[c] & ctype_digit) != 0 + ) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + case OP_DIGIT: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ( +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + c > 255 || +#endif + (md->ctypes[c] & ctype_digit) == 0 + ) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + case OP_NOT_WHITESPACE: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ( +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + c < 256 && +#endif + (md->ctypes[c] & ctype_space) != 0 + ) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + case OP_WHITESPACE: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ( +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + c > 255 || +#endif + (md->ctypes[c] & ctype_space) == 0 + ) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + case OP_NOT_WORDCHAR: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ( +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + c < 256 && +#endif + (md->ctypes[c] & ctype_word) != 0 + ) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + case OP_WORDCHAR: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ( +#if defined SUPPORT_UTF || !(defined COMPILE_PCRE8) + c > 255 || +#endif + (md->ctypes[c] & ctype_word) == 0 + ) + RRETURN(MATCH_NOMATCH); + ecode++; + break; + + case OP_ANYNL: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + + case 0x000d: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + } + else if (*eptr == 0x0a) eptr++; + break; + + case 0x000a: + break; + + case 0x000b: + case 0x000c: + case 0x0085: + case 0x2028: + case 0x2029: + if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); + break; + } + ecode++; + break; + + case OP_NOT_HSPACE: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + switch(c) + { + default: break; + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ + RRETURN(MATCH_NOMATCH); + } + ecode++; + break; + + case OP_HSPACE: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ + break; + } + ecode++; + break; + + case OP_NOT_VSPACE: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + switch(c) + { + default: break; + case 0x0a: /* LF */ + case 0x0b: /* VT */ + case 0x0c: /* FF */ + case 0x0d: /* CR */ + case 0x85: /* NEL */ + case 0x2028: /* LINE SEPARATOR */ + case 0x2029: /* PARAGRAPH SEPARATOR */ + RRETURN(MATCH_NOMATCH); + } + ecode++; + break; + + case OP_VSPACE: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + case 0x0a: /* LF */ + case 0x0b: /* VT */ + case 0x0c: /* FF */ + case 0x0d: /* CR */ + case 0x85: /* NEL */ + case 0x2028: /* LINE SEPARATOR */ + case 0x2029: /* PARAGRAPH SEPARATOR */ + break; + } + ecode++; + break; + +#ifdef SUPPORT_UCP + /* Check the next character by Unicode property. We will get here only + if the support is in the binary; otherwise a compile-time error occurs. */ + + case OP_PROP: + case OP_NOTPROP: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + { + const ucd_record *prop = GET_UCD(c); + + switch(ecode[1]) + { + case PT_ANY: + if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH); + break; + + case PT_LAMP: + if ((prop->chartype == ucp_Lu || + prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt) == (op == OP_NOTPROP)) + RRETURN(MATCH_NOMATCH); + break; + + case PT_GC: + if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP)) + RRETURN(MATCH_NOMATCH); + break; + + case PT_PC: + if ((ecode[2] != prop->chartype) == (op == OP_PROP)) + RRETURN(MATCH_NOMATCH); + break; + + case PT_SC: + if ((ecode[2] != prop->script) == (op == OP_PROP)) + RRETURN(MATCH_NOMATCH); + break; + + /* These are specials */ + + case PT_ALNUM: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP)) + RRETURN(MATCH_NOMATCH); + break; + + case PT_SPACE: /* Perl space */ + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR) + == (op == OP_NOTPROP)) + RRETURN(MATCH_NOMATCH); + break; + + case PT_PXSPACE: /* POSIX space */ + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_VT || + c == CHAR_FF || c == CHAR_CR) + == (op == OP_NOTPROP)) + RRETURN(MATCH_NOMATCH); + break; + + case PT_WORD: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || + c == CHAR_UNDERSCORE) == (op == OP_NOTPROP)) + RRETURN(MATCH_NOMATCH); + break; + + /* This should never occur */ + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + + ecode += 3; + } + break; + + /* Match an extended Unicode sequence. We will get here only if the support + is in the binary; otherwise a compile-time error occurs. */ + + case OP_EXTUNI: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH); + while (eptr < md->end_subject) + { + int len = 1; + if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); } + if (UCD_CATEGORY(c) != ucp_M) break; + eptr += len; + } + CHECK_PARTIAL(); + ecode++; + break; +#endif + + + /* Match a back reference, possibly repeatedly. Look past the end of the + item to see if there is repeat information following. The code is similar + to that for character classes, but repeated for efficiency. Then obey + similar code to character type repeats - written out again for speed. + However, if the referenced string is the empty string, always treat + it as matched, any number of times (otherwise there could be infinite + loops). */ + + case OP_REF: + case OP_REFI: + caseless = op == OP_REFI; + offset = GET2(ecode, 1) << 1; /* Doubled ref number */ + ecode += 1 + IMM2_SIZE; + + /* If the reference is unset, there are two possibilities: + + (a) In the default, Perl-compatible state, set the length negative; + this ensures that every attempt at a match fails. We can't just fail + here, because of the possibility of quantifiers with zero minima. + + (b) If the JavaScript compatibility flag is set, set the length to zero + so that the back reference matches an empty string. + + Otherwise, set the length to the length of what was matched by the + referenced subpattern. */ + + if (offset >= offset_top || md->offset_vector[offset] < 0) + length = (md->jscript_compat)? 0 : -1; + else + length = md->offset_vector[offset+1] - md->offset_vector[offset]; + + /* Set up for repetition, or handle the non-repeated case */ + + switch (*ecode) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRPLUS: + case OP_CRMINPLUS: + case OP_CRQUERY: + case OP_CRMINQUERY: + c = *ecode++ - OP_CRSTAR; + minimize = (c & 1) != 0; + min = rep_min[c]; /* Pick up values from tables; */ + max = rep_max[c]; /* zero for max => infinity */ + if (max == 0) max = INT_MAX; + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + minimize = (*ecode == OP_CRMINRANGE); + min = GET2(ecode, 1); + max = GET2(ecode, 1 + IMM2_SIZE); + if (max == 0) max = INT_MAX; + ecode += 1 + 2 * IMM2_SIZE; + break; + + default: /* No repeat follows */ + if ((length = match_ref(offset, eptr, length, md, caseless)) < 0) + { + if (length == -2) eptr = md->end_subject; /* Partial match */ + CHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr += length; + continue; /* With the main loop */ + } + + /* Handle repeated back references. If the length of the reference is + zero, just continue with the main loop. If the length is negative, it + means the reference is unset in non-Java-compatible mode. If the minimum is + zero, we can continue at the same level without recursion. For any other + minimum, carrying on will result in NOMATCH. */ + + if (length == 0) continue; + if (length < 0 && min == 0) continue; + + /* First, ensure the minimum number of matches are present. We get back + the length of the reference string explicitly rather than passing the + address of eptr, so that eptr can be a register variable. */ + + for (i = 1; i <= min; i++) + { + int slength; + if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0) + { + if (slength == -2) eptr = md->end_subject; /* Partial match */ + CHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr += slength; + } + + /* If min = max, continue at the same level without recursion. + They are not both allowed to be zero. */ + + if (min == max) continue; + + /* If minimizing, keep trying and advancing the pointer */ + + if (minimize) + { + for (fi = min;; fi++) + { + int slength; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM14); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0) + { + if (slength == -2) eptr = md->end_subject; /* Partial match */ + CHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr += slength; + } + /* Control never gets here */ + } + + /* If maximizing, find the longest string and work backwards */ + + else + { + pp = eptr; + for (i = min; i < max; i++) + { + int slength; + if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0) + { + /* Can't use CHECK_PARTIAL because we don't want to update eptr in + the soft partial matching case. */ + + if (slength == -2 && md->partial != 0 && + md->end_subject > md->start_used_ptr) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + break; + } + eptr += slength; + } + + while (eptr >= pp) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM15); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr -= length; + } + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + /* Match a bit-mapped character class, possibly repeatedly. This op code is + used when all the characters in the class have values in the range 0-255, + and either the matching is caseful, or the characters are in the range + 0-127 when UTF-8 processing is enabled. The only difference between + OP_CLASS and OP_NCLASS occurs when a data character outside the range is + encountered. + + First, look past the end of the item to see if there is repeat information + following. Then obey similar code to character type repeats - written out + again for speed. */ + + case OP_NCLASS: + case OP_CLASS: + { + /* The data variable is saved across frames, so the byte map needs to + be stored there. */ +#define BYTE_MAP ((pcre_uint8 *)data) + data = ecode + 1; /* Save for matching */ + ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */ + + switch (*ecode) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRPLUS: + case OP_CRMINPLUS: + case OP_CRQUERY: + case OP_CRMINQUERY: + c = *ecode++ - OP_CRSTAR; + minimize = (c & 1) != 0; + min = rep_min[c]; /* Pick up values from tables; */ + max = rep_max[c]; /* zero for max => infinity */ + if (max == 0) max = INT_MAX; + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + minimize = (*ecode == OP_CRMINRANGE); + min = GET2(ecode, 1); + max = GET2(ecode, 1 + IMM2_SIZE); + if (max == 0) max = INT_MAX; + ecode += 1 + 2 * IMM2_SIZE; + break; + + default: /* No repeat follows */ + min = max = 1; + break; + } + + /* First, ensure the minimum number of matches are present. */ + +#ifdef SUPPORT_UTF + if (utf) + { + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + if (c > 255) + { + if (op == OP_CLASS) RRETURN(MATCH_NOMATCH); + } + else + if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); + } + } + else +#endif + /* Not UTF mode */ + { + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + c = *eptr++; +#ifndef COMPILE_PCRE8 + if (c > 255) + { + if (op == OP_CLASS) RRETURN(MATCH_NOMATCH); + } + else +#endif + if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); + } + } + + /* If max == min we can continue with the main loop without the + need to recurse. */ + + if (min == max) continue; + + /* If minimizing, keep testing the rest of the expression and advancing + the pointer while it matches the class. */ + + if (minimize) + { +#ifdef SUPPORT_UTF + if (utf) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM16); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + if (c > 255) + { + if (op == OP_CLASS) RRETURN(MATCH_NOMATCH); + } + else + if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); + } + } + else +#endif + /* Not UTF mode */ + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM17); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + c = *eptr++; +#ifndef COMPILE_PCRE8 + if (c > 255) + { + if (op == OP_CLASS) RRETURN(MATCH_NOMATCH); + } + else +#endif + if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH); + } + } + /* Control never gets here */ + } + + /* If maximizing, find the longest possible run, then work backwards. */ + + else + { + pp = eptr; + +#ifdef SUPPORT_UTF + if (utf) + { + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c > 255) + { + if (op == OP_CLASS) break; + } + else + if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break; + eptr += len; + } + for (;;) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM18); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (eptr-- == pp) break; /* Stop if tried at original pos */ + BACKCHAR(eptr); + } + } + else +#endif + /* Not UTF mode */ + { + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + c = *eptr; +#ifndef COMPILE_PCRE8 + if (c > 255) + { + if (op == OP_CLASS) break; + } + else +#endif + if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break; + eptr++; + } + while (eptr >= pp) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM19); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; + } + } + + RRETURN(MATCH_NOMATCH); + } +#undef BYTE_MAP + } + /* Control never gets here */ + + + /* Match an extended character class. This opcode is encountered only + when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8 + mode, because Unicode properties are supported in non-UTF-8 mode. */ + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: + { + data = ecode + 1 + LINK_SIZE; /* Save for matching */ + ecode += GET(ecode, 1); /* Advance past the item */ + + switch (*ecode) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRPLUS: + case OP_CRMINPLUS: + case OP_CRQUERY: + case OP_CRMINQUERY: + c = *ecode++ - OP_CRSTAR; + minimize = (c & 1) != 0; + min = rep_min[c]; /* Pick up values from tables; */ + max = rep_max[c]; /* zero for max => infinity */ + if (max == 0) max = INT_MAX; + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + minimize = (*ecode == OP_CRMINRANGE); + min = GET2(ecode, 1); + max = GET2(ecode, 1 + IMM2_SIZE); + if (max == 0) max = INT_MAX; + ecode += 1 + 2 * IMM2_SIZE; + break; + + default: /* No repeat follows */ + min = max = 1; + break; + } + + /* First, ensure the minimum number of matches are present. */ + + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH); + } + + /* If max == min we can continue with the main loop without the + need to recurse. */ + + if (min == max) continue; + + /* If minimizing, keep testing the rest of the expression and advancing + the pointer while it matches the class. */ + + if (minimize) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM20); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + } + + /* If maximizing, find the longest possible run, then work backwards. */ + + else + { + pp = eptr; + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } +#ifdef SUPPORT_UTF + GETCHARLENTEST(c, eptr, len); +#else + c = *eptr; +#endif + if (!PRIV(xclass)(c, data, utf)) break; + eptr += len; + } + for(;;) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM21); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (eptr-- == pp) break; /* Stop if tried at original pos */ +#ifdef SUPPORT_UTF + if (utf) BACKCHAR(eptr); +#endif + } + RRETURN(MATCH_NOMATCH); + } + + /* Control never gets here */ + } +#endif /* End of XCLASS */ + + /* Match a single character, casefully */ + + case OP_CHAR: +#ifdef SUPPORT_UTF + if (utf) + { + length = 1; + ecode++; + GETCHARLEN(fc, ecode, length); + if (length > md->end_subject - eptr) + { + CHECK_PARTIAL(); /* Not SCHECK_PARTIAL() */ + RRETURN(MATCH_NOMATCH); + } + while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH); + } + else +#endif + /* Not UTF mode */ + { + if (md->end_subject - eptr < 1) + { + SCHECK_PARTIAL(); /* This one can use SCHECK_PARTIAL() */ + RRETURN(MATCH_NOMATCH); + } + if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH); + ecode += 2; + } + break; + + /* Match a single character, caselessly. If we are at the end of the + subject, give up immediately. */ + + case OP_CHARI: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + +#ifdef SUPPORT_UTF + if (utf) + { + length = 1; + ecode++; + GETCHARLEN(fc, ecode, length); + + /* If the pattern character's value is < 128, we have only one byte, and + we know that its other case must also be one byte long, so we can use the + fast lookup table. We know that there is at least one byte left in the + subject. */ + + if (fc < 128) + { + if (md->lcc[fc] + != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH); + ecode++; + eptr++; + } + + /* Otherwise we must pick up the subject character. Note that we cannot + use the value of "length" to check for sufficient bytes left, because the + other case of the character may have more or fewer bytes. */ + + else + { + unsigned int dc; + GETCHARINC(dc, eptr); + ecode += length; + + /* If we have Unicode property support, we can use it to test the other + case of the character, if there is one. */ + + if (fc != dc) + { +#ifdef SUPPORT_UCP + if (dc != UCD_OTHERCASE(fc)) +#endif + RRETURN(MATCH_NOMATCH); + } + } + } + else +#endif /* SUPPORT_UTF */ + + /* Not UTF mode */ + { + if (TABLE_GET(ecode[1], md->lcc, ecode[1]) + != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH); + eptr++; + ecode += 2; + } + break; + + /* Match a single character repeatedly. */ + + case OP_EXACT: + case OP_EXACTI: + min = max = GET2(ecode, 1); + ecode += 1 + IMM2_SIZE; + goto REPEATCHAR; + + case OP_POSUPTO: + case OP_POSUPTOI: + possessive = TRUE; + /* Fall through */ + + case OP_UPTO: + case OP_UPTOI: + case OP_MINUPTO: + case OP_MINUPTOI: + min = 0; + max = GET2(ecode, 1); + minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI; + ecode += 1 + IMM2_SIZE; + goto REPEATCHAR; + + case OP_POSSTAR: + case OP_POSSTARI: + possessive = TRUE; + min = 0; + max = INT_MAX; + ecode++; + goto REPEATCHAR; + + case OP_POSPLUS: + case OP_POSPLUSI: + possessive = TRUE; + min = 1; + max = INT_MAX; + ecode++; + goto REPEATCHAR; + + case OP_POSQUERY: + case OP_POSQUERYI: + possessive = TRUE; + min = 0; + max = 1; + ecode++; + goto REPEATCHAR; + + case OP_STAR: + case OP_STARI: + case OP_MINSTAR: + case OP_MINSTARI: + case OP_PLUS: + case OP_PLUSI: + case OP_MINPLUS: + case OP_MINPLUSI: + case OP_QUERY: + case OP_QUERYI: + case OP_MINQUERY: + case OP_MINQUERYI: + c = *ecode++ - ((op < OP_STARI)? OP_STAR : OP_STARI); + minimize = (c & 1) != 0; + min = rep_min[c]; /* Pick up values from tables; */ + max = rep_max[c]; /* zero for max => infinity */ + if (max == 0) max = INT_MAX; + + /* Common code for all repeated single-character matches. */ + + REPEATCHAR: +#ifdef SUPPORT_UTF + if (utf) + { + length = 1; + charptr = ecode; + GETCHARLEN(fc, ecode, length); + ecode += length; + + /* Handle multibyte character matching specially here. There is + support for caseless matching if UCP support is present. */ + + if (length > 1) + { +#ifdef SUPPORT_UCP + unsigned int othercase; + if (op >= OP_STARI && /* Caseless */ + (othercase = UCD_OTHERCASE(fc)) != fc) + oclength = PRIV(ord2utf)(othercase, occhars); + else oclength = 0; +#endif /* SUPPORT_UCP */ + + for (i = 1; i <= min; i++) + { + if (eptr <= md->end_subject - length && + memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length; +#ifdef SUPPORT_UCP + else if (oclength > 0 && + eptr <= md->end_subject - oclength && + memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength; +#endif /* SUPPORT_UCP */ + else + { + CHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + } + + if (min == max) continue; + + if (minimize) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM22); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr <= md->end_subject - length && + memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length; +#ifdef SUPPORT_UCP + else if (oclength > 0 && + eptr <= md->end_subject - oclength && + memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength; +#endif /* SUPPORT_UCP */ + else + { + CHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + } + /* Control never gets here */ + } + + else /* Maximize */ + { + pp = eptr; + for (i = min; i < max; i++) + { + if (eptr <= md->end_subject - length && + memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length; +#ifdef SUPPORT_UCP + else if (oclength > 0 && + eptr <= md->end_subject - oclength && + memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength; +#endif /* SUPPORT_UCP */ + else + { + CHECK_PARTIAL(); + break; + } + } + + if (possessive) continue; + + for(;;) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM23); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (eptr == pp) { RRETURN(MATCH_NOMATCH); } +#ifdef SUPPORT_UCP + eptr--; + BACKCHAR(eptr); +#else /* without SUPPORT_UCP */ + eptr -= length; +#endif /* SUPPORT_UCP */ + } + } + /* Control never gets here */ + } + + /* If the length of a UTF-8 character is 1, we fall through here, and + obey the code as for non-UTF-8 characters below, though in this case the + value of fc will always be < 128. */ + } + else +#endif /* SUPPORT_UTF */ + /* When not in UTF-8 mode, load a single-byte character. */ + fc = *ecode++; + + /* The value of fc at this point is always one character, though we may + or may not be in UTF mode. The code is duplicated for the caseless and + caseful cases, for speed, since matching characters is likely to be quite + common. First, ensure the minimum number of matches are present. If min = + max, continue at the same level without recursing. Otherwise, if + minimizing, keep trying the rest of the expression and advancing one + matching character if failing, up to the maximum. Alternatively, if + maximizing, find the maximum number of characters and work backwards. */ + + DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max, + max, (char *)eptr)); + + if (op >= OP_STARI) /* Caseless */ + { +#ifdef COMPILE_PCRE8 + /* fc must be < 128 if UTF is enabled. */ + foc = md->fcc[fc]; +#else +#ifdef SUPPORT_UTF +#ifdef SUPPORT_UCP + if (utf && fc > 127) + foc = UCD_OTHERCASE(fc); +#else + if (utf && fc > 127) + foc = fc; +#endif /* SUPPORT_UCP */ + else +#endif /* SUPPORT_UTF */ + foc = TABLE_GET(fc, md->fcc, fc); +#endif /* COMPILE_PCRE8 */ + + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH); + eptr++; + } + if (min == max) continue; + if (minimize) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM24); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH); + eptr++; + } + /* Control never gets here */ + } + else /* Maximize */ + { + pp = eptr; + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (fc != *eptr && foc != *eptr) break; + eptr++; + } + + if (possessive) continue; + + while (eptr >= pp) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM25); + eptr--; + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + } + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + } + + /* Caseful comparisons (includes all multi-byte characters) */ + + else + { + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc != *eptr++) RRETURN(MATCH_NOMATCH); + } + + if (min == max) continue; + + if (minimize) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM26); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc != *eptr++) RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + } + else /* Maximize */ + { + pp = eptr; + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (fc != *eptr) break; + eptr++; + } + if (possessive) continue; + + while (eptr >= pp) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM27); + eptr--; + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + } + RRETURN(MATCH_NOMATCH); + } + } + /* Control never gets here */ + + /* Match a negated single one-byte character. The character we are + checking can be multibyte. */ + + case OP_NOT: + case OP_NOTI: + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } +#ifdef SUPPORT_UTF + if (utf) + { + register unsigned int ch, och; + + ecode++; + GETCHARINC(ch, ecode); + GETCHARINC(c, eptr); + + if (op == OP_NOT) + { + if (ch == c) RRETURN(MATCH_NOMATCH); + } + else + { +#ifdef SUPPORT_UCP + if (ch > 127) + och = UCD_OTHERCASE(ch); +#else + if (ch > 127) + och = ch; +#endif /* SUPPORT_UCP */ + else + och = TABLE_GET(ch, md->fcc, ch); + if (ch == c || och == c) RRETURN(MATCH_NOMATCH); + } + } + else +#endif + { + register unsigned int ch = ecode[1]; + c = *eptr++; + if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c)) + RRETURN(MATCH_NOMATCH); + ecode += 2; + } + break; + + /* Match a negated single one-byte character repeatedly. This is almost a + repeat of the code for a repeated single character, but I haven't found a + nice way of commoning these up that doesn't require a test of the + positive/negative option for each character match. Maybe that wouldn't add + very much to the time taken, but character matching *is* what this is all + about... */ + + case OP_NOTEXACT: + case OP_NOTEXACTI: + min = max = GET2(ecode, 1); + ecode += 1 + IMM2_SIZE; + goto REPEATNOTCHAR; + + case OP_NOTUPTO: + case OP_NOTUPTOI: + case OP_NOTMINUPTO: + case OP_NOTMINUPTOI: + min = 0; + max = GET2(ecode, 1); + minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI; + ecode += 1 + IMM2_SIZE; + goto REPEATNOTCHAR; + + case OP_NOTPOSSTAR: + case OP_NOTPOSSTARI: + possessive = TRUE; + min = 0; + max = INT_MAX; + ecode++; + goto REPEATNOTCHAR; + + case OP_NOTPOSPLUS: + case OP_NOTPOSPLUSI: + possessive = TRUE; + min = 1; + max = INT_MAX; + ecode++; + goto REPEATNOTCHAR; + + case OP_NOTPOSQUERY: + case OP_NOTPOSQUERYI: + possessive = TRUE; + min = 0; + max = 1; + ecode++; + goto REPEATNOTCHAR; + + case OP_NOTPOSUPTO: + case OP_NOTPOSUPTOI: + possessive = TRUE; + min = 0; + max = GET2(ecode, 1); + ecode += 1 + IMM2_SIZE; + goto REPEATNOTCHAR; + + case OP_NOTSTAR: + case OP_NOTSTARI: + case OP_NOTMINSTAR: + case OP_NOTMINSTARI: + case OP_NOTPLUS: + case OP_NOTPLUSI: + case OP_NOTMINPLUS: + case OP_NOTMINPLUSI: + case OP_NOTQUERY: + case OP_NOTQUERYI: + case OP_NOTMINQUERY: + case OP_NOTMINQUERYI: + c = *ecode++ - ((op >= OP_NOTSTARI)? OP_NOTSTARI: OP_NOTSTAR); + minimize = (c & 1) != 0; + min = rep_min[c]; /* Pick up values from tables; */ + max = rep_max[c]; /* zero for max => infinity */ + if (max == 0) max = INT_MAX; + + /* Common code for all repeated single-byte matches. */ + + REPEATNOTCHAR: + GETCHARINCTEST(fc, ecode); + + /* The code is duplicated for the caseless and caseful cases, for speed, + since matching characters is likely to be quite common. First, ensure the + minimum number of matches are present. If min = max, continue at the same + level without recursing. Otherwise, if minimizing, keep trying the rest of + the expression and advancing one matching character if failing, up to the + maximum. Alternatively, if maximizing, find the maximum number of + characters and work backwards. */ + + DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max, + max, (char *)eptr)); + + if (op >= OP_NOTSTARI) /* Caseless */ + { +#ifdef SUPPORT_UTF +#ifdef SUPPORT_UCP + if (utf && fc > 127) + foc = UCD_OTHERCASE(fc); +#else + if (utf && fc > 127) + foc = fc; +#endif /* SUPPORT_UCP */ + else +#endif /* SUPPORT_UTF */ + foc = TABLE_GET(fc, md->fcc, fc); + +#ifdef SUPPORT_UTF + if (utf) + { + register unsigned int d; + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(d, eptr); + if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH); + } + } + else +#endif + /* Not UTF mode */ + { + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH); + eptr++; + } + } + + if (min == max) continue; + + if (minimize) + { +#ifdef SUPPORT_UTF + if (utf) + { + register unsigned int d; + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM28); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(d, eptr); + if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH); + } + } + else +#endif + /* Not UTF mode */ + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM29); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH); + eptr++; + } + } + /* Control never gets here */ + } + + /* Maximize case */ + + else + { + pp = eptr; + +#ifdef SUPPORT_UTF + if (utf) + { + register unsigned int d; + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(d, eptr, len); + if (fc == d || (unsigned int)foc == d) break; + eptr += len; + } + if (possessive) continue; + for(;;) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM30); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (eptr-- == pp) break; /* Stop if tried at original pos */ + BACKCHAR(eptr); + } + } + else +#endif + /* Not UTF mode */ + { + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (fc == *eptr || foc == *eptr) break; + eptr++; + } + if (possessive) continue; + while (eptr >= pp) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM31); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; + } + } + + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + } + + /* Caseful comparisons */ + + else + { +#ifdef SUPPORT_UTF + if (utf) + { + register unsigned int d; + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(d, eptr); + if (fc == d) RRETURN(MATCH_NOMATCH); + } + } + else +#endif + /* Not UTF mode */ + { + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc == *eptr++) RRETURN(MATCH_NOMATCH); + } + } + + if (min == max) continue; + + if (minimize) + { +#ifdef SUPPORT_UTF + if (utf) + { + register unsigned int d; + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM32); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(d, eptr); + if (fc == d) RRETURN(MATCH_NOMATCH); + } + } + else +#endif + /* Not UTF mode */ + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM33); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (fc == *eptr++) RRETURN(MATCH_NOMATCH); + } + } + /* Control never gets here */ + } + + /* Maximize case */ + + else + { + pp = eptr; + +#ifdef SUPPORT_UTF + if (utf) + { + register unsigned int d; + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(d, eptr, len); + if (fc == d) break; + eptr += len; + } + if (possessive) continue; + for(;;) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM34); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (eptr-- == pp) break; /* Stop if tried at original pos */ + BACKCHAR(eptr); + } + } + else +#endif + /* Not UTF mode */ + { + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (fc == *eptr) break; + eptr++; + } + if (possessive) continue; + while (eptr >= pp) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM35); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; + } + } + + RRETURN(MATCH_NOMATCH); + } + } + /* Control never gets here */ + + /* Match a single character type repeatedly; several different opcodes + share code. This is very similar to the code for single characters, but we + repeat it in the interests of efficiency. */ + + case OP_TYPEEXACT: + min = max = GET2(ecode, 1); + minimize = TRUE; + ecode += 1 + IMM2_SIZE; + goto REPEATTYPE; + + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + min = 0; + max = GET2(ecode, 1); + minimize = *ecode == OP_TYPEMINUPTO; + ecode += 1 + IMM2_SIZE; + goto REPEATTYPE; + + case OP_TYPEPOSSTAR: + possessive = TRUE; + min = 0; + max = INT_MAX; + ecode++; + goto REPEATTYPE; + + case OP_TYPEPOSPLUS: + possessive = TRUE; + min = 1; + max = INT_MAX; + ecode++; + goto REPEATTYPE; + + case OP_TYPEPOSQUERY: + possessive = TRUE; + min = 0; + max = 1; + ecode++; + goto REPEATTYPE; + + case OP_TYPEPOSUPTO: + possessive = TRUE; + min = 0; + max = GET2(ecode, 1); + ecode += 1 + IMM2_SIZE; + goto REPEATTYPE; + + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + c = *ecode++ - OP_TYPESTAR; + minimize = (c & 1) != 0; + min = rep_min[c]; /* Pick up values from tables; */ + max = rep_max[c]; /* zero for max => infinity */ + if (max == 0) max = INT_MAX; + + /* Common code for all repeated single character type matches. Note that + in UTF-8 mode, '.' matches a character of any length, but for the other + character types, the valid characters are all one-byte long. */ + + REPEATTYPE: + ctype = *ecode++; /* Code for the character type */ + +#ifdef SUPPORT_UCP + if (ctype == OP_PROP || ctype == OP_NOTPROP) + { + prop_fail_result = ctype == OP_NOTPROP; + prop_type = *ecode++; + prop_value = *ecode++; + } + else prop_type = -1; +#endif + + /* First, ensure the minimum number of matches are present. Use inline + code for maximizing the speed, and do the type test once at the start + (i.e. keep it out of the loop). Separate the UTF-8 code completely as that + is tidier. Also separate the UCP code, which can be the same for both UTF-8 + and single-bytes. */ + + if (min > 0) + { +#ifdef SUPPORT_UCP + if (prop_type >= 0) + { + switch(prop_type) + { + case PT_ANY: + if (prop_fail_result) RRETURN(MATCH_NOMATCH); + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + } + break; + + case PT_LAMP: + for (i = 1; i <= min; i++) + { + int chartype; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + chartype = UCD_CHARTYPE(c); + if ((chartype == ucp_Lu || + chartype == ucp_Ll || + chartype == ucp_Lt) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_GC: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_PC: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_SC: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_ALNUM: + for (i = 1; i <= min; i++) + { + int category; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + category = UCD_CATEGORY(c); + if ((category == ucp_L || category == ucp_N) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_SPACE: /* Perl space */ + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL || + c == CHAR_FF || c == CHAR_CR) + == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_PXSPACE: /* POSIX space */ + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL || + c == CHAR_VT || c == CHAR_FF || c == CHAR_CR) + == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + case PT_WORD: + for (i = 1; i <= min; i++) + { + int category; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + category = UCD_CATEGORY(c); + if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE) + == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + break; + + /* This should not occur */ + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + } + + /* Match extended Unicode sequences. We will get here only if the + support is in the binary; otherwise a compile-time error occurs. */ + + else if (ctype == OP_EXTUNI) + { + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH); + while (eptr < md->end_subject) + { + int len = 1; + if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); } + if (UCD_CATEGORY(c) != ucp_M) break; + eptr += len; + } + CHECK_PARTIAL(); + } + } + + else +#endif /* SUPPORT_UCP */ + +/* Handle all other cases when the coding is UTF-8 */ + +#ifdef SUPPORT_UTF + if (utf) switch(ctype) + { + case OP_ANY: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); + if (md->partial != 0 && + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + *eptr == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + break; + + case OP_ALLANY: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + break; + + case OP_ANYBYTE: + if (eptr > md->end_subject - min) RRETURN(MATCH_NOMATCH); + eptr += min; + break; + + case OP_ANYNL: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + + case 0x000d: + if (eptr < md->end_subject && *eptr == 0x0a) eptr++; + break; + + case 0x000a: + break; + + case 0x000b: + case 0x000c: + case 0x0085: + case 0x2028: + case 0x2029: + if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); + break; + } + } + break; + + case OP_NOT_HSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + switch(c) + { + default: break; + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ + RRETURN(MATCH_NOMATCH); + } + } + break; + + case OP_HSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ + break; + } + } + break; + + case OP_NOT_VSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + switch(c) + { + default: break; + case 0x0a: /* LF */ + case 0x0b: /* VT */ + case 0x0c: /* FF */ + case 0x0d: /* CR */ + case 0x85: /* NEL */ + case 0x2028: /* LINE SEPARATOR */ + case 0x2029: /* PARAGRAPH SEPARATOR */ + RRETURN(MATCH_NOMATCH); + } + } + break; + + case OP_VSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + case 0x0a: /* LF */ + case 0x0b: /* VT */ + case 0x0c: /* FF */ + case 0x0d: /* CR */ + case 0x85: /* NEL */ + case 0x2028: /* LINE SEPARATOR */ + case 0x2029: /* PARAGRAPH SEPARATOR */ + break; + } + } + break; + + case OP_NOT_DIGIT: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINC(c, eptr); + if (c < 128 && (md->ctypes[c] & ctype_digit) != 0) + RRETURN(MATCH_NOMATCH); + } + break; + + case OP_DIGIT: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0) + RRETURN(MATCH_NOMATCH); + eptr++; + /* No need to skip more bytes - we know it's a 1-byte character */ + } + break; + + case OP_NOT_WHITESPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0) + RRETURN(MATCH_NOMATCH); + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + break; + + case OP_WHITESPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0) + RRETURN(MATCH_NOMATCH); + eptr++; + /* No need to skip more bytes - we know it's a 1-byte character */ + } + break; + + case OP_NOT_WORDCHAR: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0) + RRETURN(MATCH_NOMATCH); + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + break; + + case OP_WORDCHAR: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0) + RRETURN(MATCH_NOMATCH); + eptr++; + /* No need to skip more bytes - we know it's a 1-byte character */ + } + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } /* End switch(ctype) */ + + else +#endif /* SUPPORT_UTF */ + + /* Code for the non-UTF-8 case for minimum matching of operators other + than OP_PROP and OP_NOTPROP. */ + + switch(ctype) + { + case OP_ANY: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); + if (md->partial != 0 && + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + *eptr == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + eptr++; + } + break; + + case OP_ALLANY: + if (eptr > md->end_subject - min) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr += min; + break; + + case OP_ANYBYTE: + if (eptr > md->end_subject - min) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + eptr += min; + break; + + case OP_ANYNL: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + switch(*eptr++) + { + default: RRETURN(MATCH_NOMATCH); + + case 0x000d: + if (eptr < md->end_subject && *eptr == 0x0a) eptr++; + break; + + case 0x000a: + break; + + case 0x000b: + case 0x000c: + case 0x0085: +#ifdef COMPILE_PCRE16 + case 0x2028: + case 0x2029: +#endif + if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); + break; + } + } + break; + + case OP_NOT_HSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + switch(*eptr++) + { + default: break; + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ +#ifdef COMPILE_PCRE16 + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ +#endif + RRETURN(MATCH_NOMATCH); + } + } + break; + + case OP_HSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + switch(*eptr++) + { + default: RRETURN(MATCH_NOMATCH); + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ +#ifdef COMPILE_PCRE16 + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ +#endif + break; + } + } + break; + + case OP_NOT_VSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + switch(*eptr++) + { + default: break; + case 0x0a: /* LF */ + case 0x0b: /* VT */ + case 0x0c: /* FF */ + case 0x0d: /* CR */ + case 0x85: /* NEL */ +#ifdef COMPILE_PCRE16 + case 0x2028: /* LINE SEPARATOR */ + case 0x2029: /* PARAGRAPH SEPARATOR */ +#endif + RRETURN(MATCH_NOMATCH); + } + } + break; + + case OP_VSPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + switch(*eptr++) + { + default: RRETURN(MATCH_NOMATCH); + case 0x0a: /* LF */ + case 0x0b: /* VT */ + case 0x0c: /* FF */ + case 0x0d: /* CR */ + case 0x85: /* NEL */ +#ifdef COMPILE_PCRE16 + case 0x2028: /* LINE SEPARATOR */ + case 0x2029: /* PARAGRAPH SEPARATOR */ +#endif + break; + } + } + break; + + case OP_NOT_DIGIT: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0) + RRETURN(MATCH_NOMATCH); + eptr++; + } + break; + + case OP_DIGIT: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0) + RRETURN(MATCH_NOMATCH); + eptr++; + } + break; + + case OP_NOT_WHITESPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0) + RRETURN(MATCH_NOMATCH); + eptr++; + } + break; + + case OP_WHITESPACE: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0) + RRETURN(MATCH_NOMATCH); + eptr++; + } + break; + + case OP_NOT_WORDCHAR: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0) + RRETURN(MATCH_NOMATCH); + eptr++; + } + break; + + case OP_WORDCHAR: + for (i = 1; i <= min; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0) + RRETURN(MATCH_NOMATCH); + eptr++; + } + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + } + + /* If min = max, continue at the same level without recursing */ + + if (min == max) continue; + + /* If minimizing, we have to test the rest of the pattern before each + subsequent match. Again, separate the UTF-8 case for speed, and also + separate the UCP cases. */ + + if (minimize) + { +#ifdef SUPPORT_UCP + if (prop_type >= 0) + { + switch(prop_type) + { + case PT_ANY: + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM36); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if (prop_fail_result) RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_LAMP: + for (fi = min;; fi++) + { + int chartype; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM37); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + chartype = UCD_CHARTYPE(c); + if ((chartype == ucp_Lu || + chartype == ucp_Ll || + chartype == ucp_Lt) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_GC: + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM38); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_PC: + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM39); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_SC: + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM40); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_ALNUM: + for (fi = min;; fi++) + { + int category; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM59); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + category = UCD_CATEGORY(c); + if ((category == ucp_L || category == ucp_N) == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_SPACE: /* Perl space */ + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM60); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL || + c == CHAR_FF || c == CHAR_CR) + == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_PXSPACE: /* POSIX space */ + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM61); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL || + c == CHAR_VT || c == CHAR_FF || c == CHAR_CR) + == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + case PT_WORD: + for (fi = min;; fi++) + { + int category; + RMATCH(eptr, ecode, offset_top, md, eptrb, RM62); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + category = UCD_CATEGORY(c); + if ((category == ucp_L || + category == ucp_N || + c == CHAR_UNDERSCORE) + == prop_fail_result) + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + /* This should never occur */ + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + } + + /* Match extended Unicode sequences. We will get here only if the + support is in the binary; otherwise a compile-time error occurs. */ + + else if (ctype == OP_EXTUNI) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM41); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + GETCHARINCTEST(c, eptr); + if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH); + while (eptr < md->end_subject) + { + int len = 1; + if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); } + if (UCD_CATEGORY(c) != ucp_M) break; + eptr += len; + } + CHECK_PARTIAL(); + } + } + else +#endif /* SUPPORT_UCP */ + +#ifdef SUPPORT_UTF + if (utf) + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM42); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (ctype == OP_ANY && IS_NEWLINE(eptr)) + RRETURN(MATCH_NOMATCH); + GETCHARINC(c, eptr); + switch(ctype) + { + case OP_ANY: /* This is the non-NL case */ + if (md->partial != 0 && /* Take care with CRLF partial */ + eptr >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + break; + + case OP_ALLANY: + case OP_ANYBYTE: + break; + + case OP_ANYNL: + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + case 0x000d: + if (eptr < md->end_subject && *eptr == 0x0a) eptr++; + break; + case 0x000a: + break; + + case 0x000b: + case 0x000c: + case 0x0085: + case 0x2028: + case 0x2029: + if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); + break; + } + break; + + case OP_NOT_HSPACE: + switch(c) + { + default: break; + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ + RRETURN(MATCH_NOMATCH); + } + break; + + case OP_HSPACE: + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ + break; + } + break; + + case OP_NOT_VSPACE: + switch(c) + { + default: break; + case 0x0a: /* LF */ + case 0x0b: /* VT */ + case 0x0c: /* FF */ + case 0x0d: /* CR */ + case 0x85: /* NEL */ + case 0x2028: /* LINE SEPARATOR */ + case 0x2029: /* PARAGRAPH SEPARATOR */ + RRETURN(MATCH_NOMATCH); + } + break; + + case OP_VSPACE: + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + case 0x0a: /* LF */ + case 0x0b: /* VT */ + case 0x0c: /* FF */ + case 0x0d: /* CR */ + case 0x85: /* NEL */ + case 0x2028: /* LINE SEPARATOR */ + case 0x2029: /* PARAGRAPH SEPARATOR */ + break; + } + break; + + case OP_NOT_DIGIT: + if (c < 256 && (md->ctypes[c] & ctype_digit) != 0) + RRETURN(MATCH_NOMATCH); + break; + + case OP_DIGIT: + if (c >= 256 || (md->ctypes[c] & ctype_digit) == 0) + RRETURN(MATCH_NOMATCH); + break; + + case OP_NOT_WHITESPACE: + if (c < 256 && (md->ctypes[c] & ctype_space) != 0) + RRETURN(MATCH_NOMATCH); + break; + + case OP_WHITESPACE: + if (c >= 256 || (md->ctypes[c] & ctype_space) == 0) + RRETURN(MATCH_NOMATCH); + break; + + case OP_NOT_WORDCHAR: + if (c < 256 && (md->ctypes[c] & ctype_word) != 0) + RRETURN(MATCH_NOMATCH); + break; + + case OP_WORDCHAR: + if (c >= 256 || (md->ctypes[c] & ctype_word) == 0) + RRETURN(MATCH_NOMATCH); + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + } + } + else +#endif + /* Not UTF mode */ + { + for (fi = min;; fi++) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM43); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (fi >= max) RRETURN(MATCH_NOMATCH); + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + RRETURN(MATCH_NOMATCH); + } + if (ctype == OP_ANY && IS_NEWLINE(eptr)) + RRETURN(MATCH_NOMATCH); + c = *eptr++; + switch(ctype) + { + case OP_ANY: /* This is the non-NL case */ + if (md->partial != 0 && /* Take care with CRLF partial */ + eptr >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + c == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + break; + + case OP_ALLANY: + case OP_ANYBYTE: + break; + + case OP_ANYNL: + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + case 0x000d: + if (eptr < md->end_subject && *eptr == 0x0a) eptr++; + break; + + case 0x000a: + break; + + case 0x000b: + case 0x000c: + case 0x0085: +#ifdef COMPILE_PCRE16 + case 0x2028: + case 0x2029: +#endif + if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH); + break; + } + break; + + case OP_NOT_HSPACE: + switch(c) + { + default: break; + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ +#ifdef COMPILE_PCRE16 + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ +#endif + RRETURN(MATCH_NOMATCH); + } + break; + + case OP_HSPACE: + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ +#ifdef COMPILE_PCRE16 + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ +#endif + break; + } + break; + + case OP_NOT_VSPACE: + switch(c) + { + default: break; + case 0x0a: /* LF */ + case 0x0b: /* VT */ + case 0x0c: /* FF */ + case 0x0d: /* CR */ + case 0x85: /* NEL */ +#ifdef COMPILE_PCRE16 + case 0x2028: /* LINE SEPARATOR */ + case 0x2029: /* PARAGRAPH SEPARATOR */ +#endif + RRETURN(MATCH_NOMATCH); + } + break; + + case OP_VSPACE: + switch(c) + { + default: RRETURN(MATCH_NOMATCH); + case 0x0a: /* LF */ + case 0x0b: /* VT */ + case 0x0c: /* FF */ + case 0x0d: /* CR */ + case 0x85: /* NEL */ +#ifdef COMPILE_PCRE16 + case 0x2028: /* LINE SEPARATOR */ + case 0x2029: /* PARAGRAPH SEPARATOR */ +#endif + break; + } + break; + + case OP_NOT_DIGIT: + if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH); + break; + + case OP_DIGIT: + if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH); + break; + + case OP_NOT_WHITESPACE: + if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH); + break; + + case OP_WHITESPACE: + if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH); + break; + + case OP_NOT_WORDCHAR: + if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH); + break; + + case OP_WORDCHAR: + if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH); + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + } + } + /* Control never gets here */ + } + + /* If maximizing, it is worth using inline code for speed, doing the type + test once at the start (i.e. keep it out of the loop). Again, keep the + UTF-8 and UCP stuff separate. */ + + else + { + pp = eptr; /* Remember where we started */ + +#ifdef SUPPORT_UCP + if (prop_type >= 0) + { + switch(prop_type) + { + case PT_ANY: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + if (prop_fail_result) break; + eptr+= len; + } + break; + + case PT_LAMP: + for (i = min; i < max; i++) + { + int chartype; + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + chartype = UCD_CHARTYPE(c); + if ((chartype == ucp_Lu || + chartype == ucp_Ll || + chartype == ucp_Lt) == prop_fail_result) + break; + eptr+= len; + } + break; + + case PT_GC: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result) break; + eptr+= len; + } + break; + + case PT_PC: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result) break; + eptr+= len; + } + break; + + case PT_SC: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result) break; + eptr+= len; + } + break; + + case PT_ALNUM: + for (i = min; i < max; i++) + { + int category; + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + category = UCD_CATEGORY(c); + if ((category == ucp_L || category == ucp_N) == prop_fail_result) + break; + eptr+= len; + } + break; + + case PT_SPACE: /* Perl space */ + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL || + c == CHAR_FF || c == CHAR_CR) + == prop_fail_result) + break; + eptr+= len; + } + break; + + case PT_PXSPACE: /* POSIX space */ + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL || + c == CHAR_VT || c == CHAR_FF || c == CHAR_CR) + == prop_fail_result) + break; + eptr+= len; + } + break; + + case PT_WORD: + for (i = min; i < max; i++) + { + int category; + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLENTEST(c, eptr, len); + category = UCD_CATEGORY(c); + if ((category == ucp_L || category == ucp_N || + c == CHAR_UNDERSCORE) == prop_fail_result) + break; + eptr+= len; + } + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + + /* eptr is now past the end of the maximum run */ + + if (possessive) continue; + for(;;) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM44); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (eptr-- == pp) break; /* Stop if tried at original pos */ + if (utf) BACKCHAR(eptr); + } + } + + /* Match extended Unicode sequences. We will get here only if the + support is in the binary; otherwise a compile-time error occurs. */ + + else if (ctype == OP_EXTUNI) + { + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); } + if (UCD_CATEGORY(c) == ucp_M) break; + eptr += len; + while (eptr < md->end_subject) + { + len = 1; + if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); } + if (UCD_CATEGORY(c) != ucp_M) break; + eptr += len; + } + CHECK_PARTIAL(); + } + + /* eptr is now past the end of the maximum run */ + + if (possessive) continue; + + for(;;) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM45); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (eptr-- == pp) break; /* Stop if tried at original pos */ + for (;;) /* Move back over one extended */ + { + if (!utf) c = *eptr; else + { + BACKCHAR(eptr); + GETCHAR(c, eptr); + } + if (UCD_CATEGORY(c) != ucp_M) break; + eptr--; + } + } + } + + else +#endif /* SUPPORT_UCP */ + +#ifdef SUPPORT_UTF + if (utf) + { + switch(ctype) + { + case OP_ANY: + if (max < INT_MAX) + { + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (IS_NEWLINE(eptr)) break; + if (md->partial != 0 && /* Take care with CRLF partial */ + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + *eptr == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + } + + /* Handle unlimited UTF-8 repeat */ + + else + { + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (IS_NEWLINE(eptr)) break; + if (md->partial != 0 && /* Take care with CRLF partial */ + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + *eptr == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + } + break; + + case OP_ALLANY: + if (max < INT_MAX) + { + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + eptr++; + ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++); + } + } + else + { + eptr = md->end_subject; /* Unlimited UTF-8 repeat */ + SCHECK_PARTIAL(); + } + break; + + /* The byte case is the same as non-UTF8 */ + + case OP_ANYBYTE: + c = max - min; + if (c > (unsigned int)(md->end_subject - eptr)) + { + eptr = md->end_subject; + SCHECK_PARTIAL(); + } + else eptr += c; + break; + + case OP_ANYNL: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c == 0x000d) + { + if (++eptr >= md->end_subject) break; + if (*eptr == 0x000a) eptr++; + } + else + { + if (c != 0x000a && + (md->bsr_anycrlf || + (c != 0x000b && c != 0x000c && + c != 0x0085 && c != 0x2028 && c != 0x2029))) + break; + eptr += len; + } + } + break; + + case OP_NOT_HSPACE: + case OP_HSPACE: + for (i = min; i < max; i++) + { + BOOL gotspace; + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + switch(c) + { + default: gotspace = FALSE; break; + case 0x09: /* HT */ + case 0x20: /* SPACE */ + case 0xa0: /* NBSP */ + case 0x1680: /* OGHAM SPACE MARK */ + case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ + case 0x2000: /* EN QUAD */ + case 0x2001: /* EM QUAD */ + case 0x2002: /* EN SPACE */ + case 0x2003: /* EM SPACE */ + case 0x2004: /* THREE-PER-EM SPACE */ + case 0x2005: /* FOUR-PER-EM SPACE */ + case 0x2006: /* SIX-PER-EM SPACE */ + case 0x2007: /* FIGURE SPACE */ + case 0x2008: /* PUNCTUATION SPACE */ + case 0x2009: /* THIN SPACE */ + case 0x200A: /* HAIR SPACE */ + case 0x202f: /* NARROW NO-BREAK SPACE */ + case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ + case 0x3000: /* IDEOGRAPHIC SPACE */ + gotspace = TRUE; + break; + } + if (gotspace == (ctype == OP_NOT_HSPACE)) break; + eptr += len; + } + break; + + case OP_NOT_VSPACE: + case OP_VSPACE: + for (i = min; i < max; i++) + { + BOOL gotspace; + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + switch(c) + { + default: gotspace = FALSE; break; + case 0x0a: /* LF */ + case 0x0b: /* VT */ + case 0x0c: /* FF */ + case 0x0d: /* CR */ + case 0x85: /* NEL */ + case 0x2028: /* LINE SEPARATOR */ + case 0x2029: /* PARAGRAPH SEPARATOR */ + gotspace = TRUE; + break; + } + if (gotspace == (ctype == OP_NOT_VSPACE)) break; + eptr += len; + } + break; + + case OP_NOT_DIGIT: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c < 256 && (md->ctypes[c] & ctype_digit) != 0) break; + eptr+= len; + } + break; + + case OP_DIGIT: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c >= 256 ||(md->ctypes[c] & ctype_digit) == 0) break; + eptr+= len; + } + break; + + case OP_NOT_WHITESPACE: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c < 256 && (md->ctypes[c] & ctype_space) != 0) break; + eptr+= len; + } + break; + + case OP_WHITESPACE: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c >= 256 ||(md->ctypes[c] & ctype_space) == 0) break; + eptr+= len; + } + break; + + case OP_NOT_WORDCHAR: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c < 256 && (md->ctypes[c] & ctype_word) != 0) break; + eptr+= len; + } + break; + + case OP_WORDCHAR: + for (i = min; i < max; i++) + { + int len = 1; + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + GETCHARLEN(c, eptr, len); + if (c >= 256 || (md->ctypes[c] & ctype_word) == 0) break; + eptr+= len; + } + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + + /* eptr is now past the end of the maximum run. If possessive, we are + done (no backing up). Otherwise, match at this position; anything other + than no match is immediately returned. For nomatch, back up one + character, unless we are matching \R and the last thing matched was + \r\n, in which case, back up two bytes. */ + + if (possessive) continue; + for(;;) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM46); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + if (eptr-- == pp) break; /* Stop if tried at original pos */ + BACKCHAR(eptr); + if (ctype == OP_ANYNL && eptr > pp && *eptr == '\n' && + eptr[-1] == '\r') eptr--; + } + } + else +#endif /* SUPPORT_UTF */ + /* Not UTF mode */ + { + switch(ctype) + { + case OP_ANY: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (IS_NEWLINE(eptr)) break; + if (md->partial != 0 && /* Take care with CRLF partial */ + eptr + 1 >= md->end_subject && + NLBLOCK->nltype == NLTYPE_FIXED && + NLBLOCK->nllen == 2 && + *eptr == NLBLOCK->nl[0]) + { + md->hitend = TRUE; + if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); + } + eptr++; + } + break; + + case OP_ALLANY: + case OP_ANYBYTE: + c = max - min; + if (c > (unsigned int)(md->end_subject - eptr)) + { + eptr = md->end_subject; + SCHECK_PARTIAL(); + } + else eptr += c; + break; + + case OP_ANYNL: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + c = *eptr; + if (c == 0x000d) + { + if (++eptr >= md->end_subject) break; + if (*eptr == 0x000a) eptr++; + } + else + { + if (c != 0x000a && (md->bsr_anycrlf || + (c != 0x000b && c != 0x000c && c != 0x0085 +#ifdef COMPILE_PCRE16 + && c != 0x2028 && c != 0x2029 +#endif + ))) break; + eptr++; + } + } + break; + + case OP_NOT_HSPACE: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + c = *eptr; + if (c == 0x09 || c == 0x20 || c == 0xa0 +#ifdef COMPILE_PCRE16 + || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A) + || c == 0x202f || c == 0x205f || c == 0x3000 +#endif + ) break; + eptr++; + } + break; + + case OP_HSPACE: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + c = *eptr; + if (c != 0x09 && c != 0x20 && c != 0xa0 +#ifdef COMPILE_PCRE16 + && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A) + && c != 0x202f && c != 0x205f && c != 0x3000 +#endif + ) break; + eptr++; + } + break; + + case OP_NOT_VSPACE: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + c = *eptr; + if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85 +#ifdef COMPILE_PCRE16 + || c == 0x2028 || c == 0x2029 +#endif + ) break; + eptr++; + } + break; + + case OP_VSPACE: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + c = *eptr; + if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85 +#ifdef COMPILE_PCRE16 + && c != 0x2028 && c != 0x2029 +#endif + ) break; + eptr++; + } + break; + + case OP_NOT_DIGIT: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0) break; + eptr++; + } + break; + + case OP_DIGIT: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0) break; + eptr++; + } + break; + + case OP_NOT_WHITESPACE: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0) break; + eptr++; + } + break; + + case OP_WHITESPACE: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0) break; + eptr++; + } + break; + + case OP_NOT_WORDCHAR: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0) break; + eptr++; + } + break; + + case OP_WORDCHAR: + for (i = min; i < max; i++) + { + if (eptr >= md->end_subject) + { + SCHECK_PARTIAL(); + break; + } + if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0) break; + eptr++; + } + break; + + default: + RRETURN(PCRE_ERROR_INTERNAL); + } + + /* eptr is now past the end of the maximum run. If possessive, we are + done (no backing up). Otherwise, match at this position; anything other + than no match is immediately returned. For nomatch, back up one + character (byte), unless we are matching \R and the last thing matched + was \r\n, in which case, back up two bytes. */ + + if (possessive) continue; + while (eptr >= pp) + { + RMATCH(eptr, ecode, offset_top, md, eptrb, RM47); + if (rrc != MATCH_NOMATCH) RRETURN(rrc); + eptr--; + if (ctype == OP_ANYNL && eptr > pp && *eptr == '\n' && + eptr[-1] == '\r') eptr--; + } + } + + /* Get here if we can't make it match with any permitted repetitions */ + + RRETURN(MATCH_NOMATCH); + } + /* Control never gets here */ + + /* There's been some horrible disaster. Arrival here can only mean there is + something seriously wrong in the code above or the OP_xxx definitions. */ + + default: + DPRINTF(("Unknown opcode %d\n", *ecode)); + RRETURN(PCRE_ERROR_UNKNOWN_OPCODE); + } + + /* Do not stick any code in here without much thought; it is assumed + that "continue" in the code above comes out to here to repeat the main + loop. */ + + } /* End of main loop */ +/* Control never reaches here */ + + +/* When compiling to use the heap rather than the stack for recursive calls to +match(), the RRETURN() macro jumps here. The number that is saved in +frame->Xwhere indicates which label we actually want to return to. */ + +#ifdef NO_RECURSE +#define LBL(val) case val: goto L_RM##val; +HEAP_RETURN: +switch (frame->Xwhere) + { + LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8) + LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17) + LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33) + LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52) + LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58) LBL(63) LBL(64) + LBL(65) LBL(66) +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + LBL(21) +#endif +#ifdef SUPPORT_UTF + LBL(16) LBL(18) LBL(20) + LBL(22) LBL(23) LBL(28) LBL(30) + LBL(32) LBL(34) LBL(42) LBL(46) +#ifdef SUPPORT_UCP + LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45) + LBL(59) LBL(60) LBL(61) LBL(62) +#endif /* SUPPORT_UCP */ +#endif /* SUPPORT_UTF */ + default: + DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere)); + +printf("+++jump error in pcre match: label %d non-existent\n", frame->Xwhere); + + return PCRE_ERROR_INTERNAL; + } +#undef LBL +#endif /* NO_RECURSE */ +} + + +/*************************************************************************** +**************************************************************************** + RECURSION IN THE match() FUNCTION + +Undefine all the macros that were defined above to handle this. */ + +#ifdef NO_RECURSE +#undef eptr +#undef ecode +#undef mstart +#undef offset_top +#undef eptrb +#undef flags + +#undef callpat +#undef charptr +#undef data +#undef next +#undef pp +#undef prev +#undef saved_eptr + +#undef new_recursive + +#undef cur_is_word +#undef condition +#undef prev_is_word + +#undef ctype +#undef length +#undef max +#undef min +#undef number +#undef offset +#undef op +#undef save_capture_last +#undef save_offset1 +#undef save_offset2 +#undef save_offset3 +#undef stacksave + +#undef newptrb + +#endif + +/* These two are defined as macros in both cases */ + +#undef fc +#undef fi + +/*************************************************************************** +***************************************************************************/ + + +#ifdef NO_RECURSE +/************************************************* +* Release allocated heap frames * +*************************************************/ + +/* This function releases all the allocated frames. The base frame is on the +machine stack, and so must not be freed. + +Argument: the address of the base frame +Returns: nothing +*/ + +static void +release_match_heapframes (heapframe *frame_base) +{ +heapframe *nextframe = frame_base->Xnextframe; +while (nextframe != NULL) + { + heapframe *oldframe = nextframe; + nextframe = nextframe->Xnextframe; + (PUBL(stack_free))(oldframe); + } +} +#endif + + +/************************************************* +* Execute a Regular Expression * +*************************************************/ + +/* This function applies a compiled re to a subject string and picks out +portions of the string if it matches. Two elements in the vector are set for +each substring: the offsets to the start and end of the substring. + +Arguments: + argument_re points to the compiled expression + extra_data points to extra data or is NULL + subject points to the subject string + length length of subject string (may contain binary zeros) + start_offset where to start in the subject string + options option bits + offsets points to a vector of ints to be filled in with offsets + offsetcount the number of elements in the vector + +Returns: > 0 => success; value is the number of elements filled in + = 0 => success, but offsets is not big enough + -1 => failed to match + < -1 => some kind of unexpected problem +*/ + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre_exec(const pcre *argument_re, const pcre_extra *extra_data, + PCRE_SPTR subject, int length, int start_offset, int options, int *offsets, + int offsetcount) +#else +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data, + PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets, + int offsetcount) +#endif +{ +int rc, ocount, arg_offset_max; +int newline; +BOOL using_temporary_offsets = FALSE; +BOOL anchored; +BOOL startline; +BOOL firstline; +BOOL utf; +BOOL has_first_char = FALSE; +BOOL has_req_char = FALSE; +pcre_uchar first_char = 0; +pcre_uchar first_char2 = 0; +pcre_uchar req_char = 0; +pcre_uchar req_char2 = 0; +match_data match_block; +match_data *md = &match_block; +const pcre_uint8 *tables; +const pcre_uint8 *start_bits = NULL; +PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset; +PCRE_PUCHAR end_subject; +PCRE_PUCHAR start_partial = NULL; +PCRE_PUCHAR req_char_ptr = start_match - 1; + +const pcre_study_data *study; +const REAL_PCRE *re = (const REAL_PCRE *)argument_re; + +#ifdef NO_RECURSE +heapframe frame_zero; +frame_zero.Xprevframe = NULL; /* Marks the top level */ +frame_zero.Xnextframe = NULL; /* None are allocated yet */ +md->match_frames_base = &frame_zero; +#endif + +/* Check for the special magic call that measures the size of the stack used +per recursive call of match(). Without the funny casting for sizeof, a Windows +compiler gave this error: "unary minus operator applied to unsigned type, +result still unsigned". Hopefully the cast fixes that. */ + +if (re == NULL && extra_data == NULL && subject == NULL && length == -999 && + start_offset == -999) +#ifdef NO_RECURSE + return -((int)sizeof(heapframe)); +#else + return match(NULL, NULL, NULL, 0, NULL, NULL, 0); +#endif + +/* Plausibility checks */ + +if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION; +if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0)) + return PCRE_ERROR_NULL; +if (offsetcount < 0) return PCRE_ERROR_BADCOUNT; +if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET; + +/* Check that the first field in the block is the magic number. If it is not, +return with PCRE_ERROR_BADMAGIC. However, if the magic number is equal to +REVERSED_MAGIC_NUMBER we return with PCRE_ERROR_BADENDIANNESS, which +means that the pattern is likely compiled with different endianness. */ + +if (re->magic_number != MAGIC_NUMBER) + return re->magic_number == REVERSED_MAGIC_NUMBER? + PCRE_ERROR_BADENDIANNESS:PCRE_ERROR_BADMAGIC; +if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE; + +/* These two settings are used in the code for checking a UTF-8 string that +follows immediately afterwards. Other values in the md block are used only +during "normal" pcre_exec() processing, not when the JIT support is in use, +so they are set up later. */ + +/* PCRE_UTF16 has the same value as PCRE_UTF8. */ +utf = md->utf = (re->options & PCRE_UTF8) != 0; +md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 : + ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0; + +/* Check a UTF-8 string if required. Pass back the character offset and error +code for an invalid string if a results vector is available. */ + +#ifdef SUPPORT_UTF +if (utf && (options & PCRE_NO_UTF8_CHECK) == 0) + { + int erroroffset; + int errorcode = PRIV(valid_utf)((PCRE_PUCHAR)subject, length, &erroroffset); + if (errorcode != 0) + { + if (offsetcount >= 2) + { + offsets[0] = erroroffset; + offsets[1] = errorcode; + } +#ifdef COMPILE_PCRE16 + return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)? + PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16; +#else + return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)? + PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8; +#endif + } + + /* Check that a start_offset points to the start of a UTF character. */ + if (start_offset > 0 && start_offset < length && + NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset])) + return PCRE_ERROR_BADUTF8_OFFSET; + } +#endif + +/* If the pattern was successfully studied with JIT support, run the JIT +executable instead of the rest of this function. Most options must be set at +compile time for the JIT code to be usable. Fallback to the normal code path if +an unsupported flag is set. */ + +#ifdef SUPPORT_JIT +if (extra_data != NULL + && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT | + PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT + && extra_data->executable_jit != NULL + && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL | + PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART | + PCRE_PARTIAL_SOFT | PCRE_PARTIAL_HARD)) == 0) + { + rc = PRIV(jit_exec)(re, extra_data, (const pcre_uchar *)subject, length, + start_offset, options, offsets, offsetcount); + + /* PCRE_ERROR_NULL means that the selected normal or partial matching + mode is not compiled. In this case we simply fallback to interpreter. */ + + if (rc != PCRE_ERROR_NULL) return rc; + } +#endif + +/* Carry on with non-JIT matching. This information is for finding all the +numbers associated with a given name, for condition testing. */ + +md->name_table = (pcre_uchar *)re + re->name_table_offset; +md->name_count = re->name_count; +md->name_entry_size = re->name_entry_size; + +/* Fish out the optional data from the extra_data structure, first setting +the default values. */ + +study = NULL; +md->match_limit = MATCH_LIMIT; +md->match_limit_recursion = MATCH_LIMIT_RECURSION; +md->callout_data = NULL; + +/* The table pointer is always in native byte order. */ + +tables = re->tables; + +if (extra_data != NULL) + { + register unsigned int flags = extra_data->flags; + if ((flags & PCRE_EXTRA_STUDY_DATA) != 0) + study = (const pcre_study_data *)extra_data->study_data; + if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0) + md->match_limit = extra_data->match_limit; + if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0) + md->match_limit_recursion = extra_data->match_limit_recursion; + if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0) + md->callout_data = extra_data->callout_data; + if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables; + } + +/* If the exec call supplied NULL for tables, use the inbuilt ones. This +is a feature that makes it possible to save compiled regex and re-use them +in other programs later. */ + +if (tables == NULL) tables = PRIV(default_tables); + +/* Set up other data */ + +anchored = ((re->options | options) & PCRE_ANCHORED) != 0; +startline = (re->flags & PCRE_STARTLINE) != 0; +firstline = (re->options & PCRE_FIRSTLINE) != 0; + +/* The code starts after the real_pcre block and the capture name table. */ + +md->start_code = (const pcre_uchar *)re + re->name_table_offset + + re->name_count * re->name_entry_size; + +md->start_subject = (PCRE_PUCHAR)subject; +md->start_offset = start_offset; +md->end_subject = md->start_subject + length; +end_subject = md->end_subject; + +md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0; +md->use_ucp = (re->options & PCRE_UCP) != 0; +md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0; +md->ignore_skip_arg = FALSE; + +/* Some options are unpacked into BOOL variables in the hope that testing +them will be faster than individual option bits. */ + +md->notbol = (options & PCRE_NOTBOL) != 0; +md->noteol = (options & PCRE_NOTEOL) != 0; +md->notempty = (options & PCRE_NOTEMPTY) != 0; +md->notempty_atstart = (options & PCRE_NOTEMPTY_ATSTART) != 0; + +md->hitend = FALSE; +md->mark = md->nomatch_mark = NULL; /* In case never set */ + +md->recursive = NULL; /* No recursion at top level */ +md->hasthen = (re->flags & PCRE_HASTHEN) != 0; + +md->lcc = tables + lcc_offset; +md->fcc = tables + fcc_offset; +md->ctypes = tables + ctypes_offset; + +/* Handle different \R options. */ + +switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) + { + case 0: + if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0) + md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0; + else +#ifdef BSR_ANYCRLF + md->bsr_anycrlf = TRUE; +#else + md->bsr_anycrlf = FALSE; +#endif + break; + + case PCRE_BSR_ANYCRLF: + md->bsr_anycrlf = TRUE; + break; + + case PCRE_BSR_UNICODE: + md->bsr_anycrlf = FALSE; + break; + + default: return PCRE_ERROR_BADNEWLINE; + } + +/* Handle different types of newline. The three bits give eight cases. If +nothing is set at run time, whatever was used at compile time applies. */ + +switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : + (pcre_uint32)options) & PCRE_NEWLINE_BITS) + { + case 0: newline = NEWLINE; break; /* Compile-time default */ + case PCRE_NEWLINE_CR: newline = CHAR_CR; break; + case PCRE_NEWLINE_LF: newline = CHAR_NL; break; + case PCRE_NEWLINE_CR+ + PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break; + case PCRE_NEWLINE_ANY: newline = -1; break; + case PCRE_NEWLINE_ANYCRLF: newline = -2; break; + default: return PCRE_ERROR_BADNEWLINE; + } + +if (newline == -2) + { + md->nltype = NLTYPE_ANYCRLF; + } +else if (newline < 0) + { + md->nltype = NLTYPE_ANY; + } +else + { + md->nltype = NLTYPE_FIXED; + if (newline > 255) + { + md->nllen = 2; + md->nl[0] = (newline >> 8) & 255; + md->nl[1] = newline & 255; + } + else + { + md->nllen = 1; + md->nl[0] = newline; + } + } + +/* Partial matching was originally supported only for a restricted set of +regexes; from release 8.00 there are no restrictions, but the bits are still +defined (though never set). So there's no harm in leaving this code. */ + +if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0) + return PCRE_ERROR_BADPARTIAL; + +/* If the expression has got more back references than the offsets supplied can +hold, we get a temporary chunk of working store to use during the matching. +Otherwise, we can use the vector supplied, rounding down its size to a multiple +of 3. */ + +ocount = offsetcount - (offsetcount % 3); +arg_offset_max = (2*ocount)/3; + +if (re->top_backref > 0 && re->top_backref >= ocount/3) + { + ocount = re->top_backref * 3 + 3; + md->offset_vector = (int *)(PUBL(malloc))(ocount * sizeof(int)); + if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY; + using_temporary_offsets = TRUE; + DPRINTF(("Got memory to hold back references\n")); + } +else md->offset_vector = offsets; + +md->offset_end = ocount; +md->offset_max = (2*ocount)/3; +md->offset_overflow = FALSE; +md->capture_last = -1; + +/* Reset the working variable associated with each extraction. These should +never be used unless previously set, but they get saved and restored, and so we +initialize them to avoid reading uninitialized locations. Also, unset the +offsets for the matched string. This is really just for tidiness with callouts, +in case they inspect these fields. */ + +if (md->offset_vector != NULL) + { + register int *iptr = md->offset_vector + ocount; + register int *iend = iptr - re->top_bracket; + if (iend < md->offset_vector + 2) iend = md->offset_vector + 2; + while (--iptr >= iend) *iptr = -1; + md->offset_vector[0] = md->offset_vector[1] = -1; + } + +/* Set up the first character to match, if available. The first_char value is +never set for an anchored regular expression, but the anchoring may be forced +at run time, so we have to test for anchoring. The first char may be unset for +an unanchored pattern, of course. If there's no first char and the pattern was +studied, there may be a bitmap of possible first characters. */ + +if (!anchored) + { + if ((re->flags & PCRE_FIRSTSET) != 0) + { + has_first_char = TRUE; + first_char = first_char2 = (pcre_uchar)(re->first_char); + if ((re->flags & PCRE_FCH_CASELESS) != 0) + { + first_char2 = TABLE_GET(first_char, md->fcc, first_char); +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + if (utf && first_char > 127) + first_char2 = UCD_OTHERCASE(first_char); +#endif + } + } + else + if (!startline && study != NULL && + (study->flags & PCRE_STUDY_MAPPED) != 0) + start_bits = study->start_bits; + } + +/* For anchored or unanchored matches, there may be a "last known required +character" set. */ + +if ((re->flags & PCRE_REQCHSET) != 0) + { + has_req_char = TRUE; + req_char = req_char2 = (pcre_uchar)(re->req_char); + if ((re->flags & PCRE_RCH_CASELESS) != 0) + { + req_char2 = TABLE_GET(req_char, md->fcc, req_char); +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + if (utf && req_char > 127) + req_char2 = UCD_OTHERCASE(req_char); +#endif + } + } + + +/* ==========================================================================*/ + +/* Loop for handling unanchored repeated matching attempts; for anchored regexs +the loop runs just once. */ + +for(;;) + { + PCRE_PUCHAR save_end_subject = end_subject; + PCRE_PUCHAR new_start_match; + + /* If firstline is TRUE, the start of the match is constrained to the first + line of a multiline string. That is, the match must be before or at the first + newline. Implement this by temporarily adjusting end_subject so that we stop + scanning at a newline. If the match fails at the newline, later code breaks + this loop. */ + + if (firstline) + { + PCRE_PUCHAR t = start_match; +#ifdef SUPPORT_UTF + if (utf) + { + while (t < md->end_subject && !IS_NEWLINE(t)) + { + t++; + ACROSSCHAR(t < end_subject, *t, t++); + } + } + else +#endif + while (t < md->end_subject && !IS_NEWLINE(t)) t++; + end_subject = t; + } + + /* There are some optimizations that avoid running the match if a known + starting point is not found, or if a known later character is not present. + However, there is an option that disables these, for testing and for ensuring + that all callouts do actually occur. The option can be set in the regex by + (*NO_START_OPT) or passed in match-time options. */ + + if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0) + { + /* Advance to a unique first char if there is one. */ + + if (has_first_char) + { + if (first_char != first_char2) + while (start_match < end_subject && + *start_match != first_char && *start_match != first_char2) + start_match++; + else + while (start_match < end_subject && *start_match != first_char) + start_match++; + } + + /* Or to just after a linebreak for a multiline match */ + + else if (startline) + { + if (start_match > md->start_subject + start_offset) + { +#ifdef SUPPORT_UTF + if (utf) + { + while (start_match < end_subject && !WAS_NEWLINE(start_match)) + { + start_match++; + ACROSSCHAR(start_match < end_subject, *start_match, + start_match++); + } + } + else +#endif + while (start_match < end_subject && !WAS_NEWLINE(start_match)) + start_match++; + + /* If we have just passed a CR and the newline option is ANY or ANYCRLF, + and we are now at a LF, advance the match position by one more character. + */ + + if (start_match[-1] == CHAR_CR && + (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) && + start_match < end_subject && + *start_match == CHAR_NL) + start_match++; + } + } + + /* Or to a non-unique first byte after study */ + + else if (start_bits != NULL) + { + while (start_match < end_subject) + { + register unsigned int c = *start_match; +#ifndef COMPILE_PCRE8 + if (c > 255) c = 255; +#endif + if ((start_bits[c/8] & (1 << (c&7))) == 0) + { + start_match++; +#if defined SUPPORT_UTF && defined COMPILE_PCRE8 + /* In non 8-bit mode, the iteration will stop for + characters > 255 at the beginning or not stop at all. */ + if (utf) + ACROSSCHAR(start_match < end_subject, *start_match, + start_match++); +#endif + } + else break; + } + } + } /* Starting optimizations */ + + /* Restore fudged end_subject */ + + end_subject = save_end_subject; + + /* The following two optimizations are disabled for partial matching or if + disabling is explicitly requested. */ + + if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0 && !md->partial) + { + /* If the pattern was studied, a minimum subject length may be set. This is + a lower bound; no actual string of that length may actually match the + pattern. Although the value is, strictly, in characters, we treat it as + bytes to avoid spending too much time in this optimization. */ + + if (study != NULL && (study->flags & PCRE_STUDY_MINLEN) != 0 && + (pcre_uint32)(end_subject - start_match) < study->minlength) + { + rc = MATCH_NOMATCH; + break; + } + + /* If req_char is set, we know that that character must appear in the + subject for the match to succeed. If the first character is set, req_char + must be later in the subject; otherwise the test starts at the match point. + This optimization can save a huge amount of backtracking in patterns with + nested unlimited repeats that aren't going to match. Writing separate code + for cased/caseless versions makes it go faster, as does using an + autoincrement and backing off on a match. + + HOWEVER: when the subject string is very, very long, searching to its end + can take a long time, and give bad performance on quite ordinary patterns. + This showed up when somebody was matching something like /^\d+C/ on a + 32-megabyte string... so we don't do this when the string is sufficiently + long. */ + + if (has_req_char && end_subject - start_match < REQ_BYTE_MAX) + { + register PCRE_PUCHAR p = start_match + (has_first_char? 1:0); + + /* We don't need to repeat the search if we haven't yet reached the + place we found it at last time. */ + + if (p > req_char_ptr) + { + if (req_char != req_char2) + { + while (p < end_subject) + { + register int pp = *p++; + if (pp == req_char || pp == req_char2) { p--; break; } + } + } + else + { + while (p < end_subject) + { + if (*p++ == req_char) { p--; break; } + } + } + + /* If we can't find the required character, break the matching loop, + forcing a match failure. */ + + if (p >= end_subject) + { + rc = MATCH_NOMATCH; + break; + } + + /* If we have found the required character, save the point where we + found it, so that we don't search again next time round the loop if + the start hasn't passed this character yet. */ + + req_char_ptr = p; + } + } + } + +#ifdef PCRE_DEBUG /* Sigh. Some compilers never learn. */ + printf(">>>> Match against: "); + pchars(start_match, end_subject - start_match, TRUE, md); + printf("\n"); +#endif + + /* OK, we can now run the match. If "hitend" is set afterwards, remember the + first starting point for which a partial match was found. */ + + md->start_match_ptr = start_match; + md->start_used_ptr = start_match; + md->match_call_count = 0; + md->match_function_type = 0; + md->end_offset_top = 0; + rc = match(start_match, md->start_code, start_match, 2, md, NULL, 0); + if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr; + + switch(rc) + { + /* If MATCH_SKIP_ARG reaches this level it means that a MARK that matched + the SKIP's arg was not found. In this circumstance, Perl ignores the SKIP + entirely. The only way we can do that is to re-do the match at the same + point, with a flag to force SKIP with an argument to be ignored. Just + treating this case as NOMATCH does not work because it does not check other + alternatives in patterns such as A(*SKIP:A)B|AC when the subject is AC. */ + + case MATCH_SKIP_ARG: + new_start_match = start_match; + md->ignore_skip_arg = TRUE; + break; + + /* SKIP passes back the next starting point explicitly, but if it is the + same as the match we have just done, treat it as NOMATCH. */ + + case MATCH_SKIP: + if (md->start_match_ptr != start_match) + { + new_start_match = md->start_match_ptr; + break; + } + /* Fall through */ + + /* NOMATCH and PRUNE advance by one character. THEN at this level acts + exactly like PRUNE. Unset the ignore SKIP-with-argument flag. */ + + case MATCH_NOMATCH: + case MATCH_PRUNE: + case MATCH_THEN: + md->ignore_skip_arg = FALSE; + new_start_match = start_match + 1; +#ifdef SUPPORT_UTF + if (utf) + ACROSSCHAR(new_start_match < end_subject, *new_start_match, + new_start_match++); +#endif + break; + + /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */ + + case MATCH_COMMIT: + rc = MATCH_NOMATCH; + goto ENDLOOP; + + /* Any other return is either a match, or some kind of error. */ + + default: + goto ENDLOOP; + } + + /* Control reaches here for the various types of "no match at this point" + result. Reset the code to MATCH_NOMATCH for subsequent checking. */ + + rc = MATCH_NOMATCH; + + /* If PCRE_FIRSTLINE is set, the match must happen before or at the first + newline in the subject (though it may continue over the newline). Therefore, + if we have just failed to match, starting at a newline, do not continue. */ + + if (firstline && IS_NEWLINE(start_match)) break; + + /* Advance to new matching position */ + + start_match = new_start_match; + + /* Break the loop if the pattern is anchored or if we have passed the end of + the subject. */ + + if (anchored || start_match > end_subject) break; + + /* If we have just passed a CR and we are now at a LF, and the pattern does + not contain any explicit matches for \r or \n, and the newline option is CRLF + or ANY or ANYCRLF, advance the match position by one more character. In + normal matching start_match will aways be greater than the first position at + this stage, but a failed *SKIP can cause a return at the same point, which is + why the first test exists. */ + + if (start_match > (PCRE_PUCHAR)subject + start_offset && + start_match[-1] == CHAR_CR && + start_match < end_subject && + *start_match == CHAR_NL && + (re->flags & PCRE_HASCRORLF) == 0 && + (md->nltype == NLTYPE_ANY || + md->nltype == NLTYPE_ANYCRLF || + md->nllen == 2)) + start_match++; + + md->mark = NULL; /* Reset for start of next match attempt */ + } /* End of for(;;) "bumpalong" loop */ + +/* ==========================================================================*/ + +/* We reach here when rc is not MATCH_NOMATCH, or if one of the stopping +conditions is true: + +(1) The pattern is anchored or the match was failed by (*COMMIT); + +(2) We are past the end of the subject; + +(3) PCRE_FIRSTLINE is set and we have failed to match at a newline, because + this option requests that a match occur at or before the first newline in + the subject. + +When we have a match and the offset vector is big enough to deal with any +backreferences, captured substring offsets will already be set up. In the case +where we had to get some local store to hold offsets for backreference +processing, copy those that we can. In this case there need not be overflow if +certain parts of the pattern were not used, even though there are more +capturing parentheses than vector slots. */ + +ENDLOOP: + +if (rc == MATCH_MATCH || rc == MATCH_ACCEPT) + { + if (using_temporary_offsets) + { + if (arg_offset_max >= 4) + { + memcpy(offsets + 2, md->offset_vector + 2, + (arg_offset_max - 2) * sizeof(int)); + DPRINTF(("Copied offsets from temporary memory\n")); + } + if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE; + DPRINTF(("Freeing temporary memory\n")); + (PUBL(free))(md->offset_vector); + } + + /* Set the return code to the number of captured strings, or 0 if there were + too many to fit into the vector. */ + + rc = (md->offset_overflow && md->end_offset_top >= arg_offset_max)? + 0 : md->end_offset_top/2; + + /* If there is space in the offset vector, set any unused pairs at the end of + the pattern to -1 for backwards compatibility. It is documented that this + happens. In earlier versions, the whole set of potential capturing offsets + was set to -1 each time round the loop, but this is handled differently now. + "Gaps" are set to -1 dynamically instead (this fixes a bug). Thus, it is only + those at the end that need unsetting here. We can't just unset them all at + the start of the whole thing because they may get set in one branch that is + not the final matching branch. */ + + if (md->end_offset_top/2 <= re->top_bracket && offsets != NULL) + { + register int *iptr, *iend; + int resetcount = 2 + re->top_bracket * 2; + if (resetcount > offsetcount) resetcount = offsetcount; + iptr = offsets + md->end_offset_top; + iend = offsets + resetcount; + while (iptr < iend) *iptr++ = -1; + } + + /* If there is space, set up the whole thing as substring 0. The value of + md->start_match_ptr might be modified if \K was encountered on the success + matching path. */ + + if (offsetcount < 2) rc = 0; else + { + offsets[0] = (int)(md->start_match_ptr - md->start_subject); + offsets[1] = (int)(md->end_match_ptr - md->start_subject); + } + + /* Return MARK data if requested */ + + if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0) + *(extra_data->mark) = (pcre_uchar *)md->mark; + DPRINTF((">>>> returning %d\n", rc)); +#ifdef NO_RECURSE + release_match_heapframes(&frame_zero); +#endif + return rc; + } + +/* Control gets here if there has been an error, or if the overall match +attempt has failed at all permitted starting positions. */ + +if (using_temporary_offsets) + { + DPRINTF(("Freeing temporary memory\n")); + (PUBL(free))(md->offset_vector); + } + +/* For anything other than nomatch or partial match, just return the code. */ + +if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL) + { + DPRINTF((">>>> error: returning %d\n", rc)); +#ifdef NO_RECURSE + release_match_heapframes(&frame_zero); +#endif + return rc; + } + +/* Handle partial matches - disable any mark data */ + +if (start_partial != NULL) + { + DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n")); + md->mark = NULL; + if (offsetcount > 1) + { + offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject); + offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject); + } + rc = PCRE_ERROR_PARTIAL; + } + +/* This is the classic nomatch case */ + +else + { + DPRINTF((">>>> returning PCRE_ERROR_NOMATCH\n")); + rc = PCRE_ERROR_NOMATCH; + } + +/* Return the MARK data if it has been requested. */ + +if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0) + *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark; +#ifdef NO_RECURSE + release_match_heapframes(&frame_zero); +#endif +return rc; } /* End of pcre_exec.c */ diff --git a/pcre/pcre_fullinfo.c b/pcre/pcre_fullinfo.c --- a/pcre/pcre_fullinfo.c +++ b/pcre/pcre_fullinfo.c @@ -6,7 +6,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -65,13 +65,17 @@ Returns: 0 if data returned, negative on error */ -PCRE_EXP_DEFN int -pcre_fullinfo(const pcre *argument_re, const pcre_extra *extra_data, int what, - void *where) +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre_fullinfo(const pcre *argument_re, const pcre_extra *extra_data, + int what, void *where) +#else +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_fullinfo(const pcre16 *argument_re, const pcre16_extra *extra_data, + int what, void *where) +#endif { -real_pcre internal_re; -pcre_study_data internal_study; -const real_pcre *re = (const real_pcre *)argument_re; +const REAL_PCRE *re = (const REAL_PCRE *)argument_re; const pcre_study_data *study = NULL; if (re == NULL || where == NULL) return PCRE_ERROR_NULL; @@ -79,17 +83,23 @@ if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_STUDY_DATA) != 0) study = (const pcre_study_data *)extra_data->study_data; +/* Check that the first field in the block is the magic number. If it is not, +return with PCRE_ERROR_BADMAGIC. However, if the magic number is equal to +REVERSED_MAGIC_NUMBER we return with PCRE_ERROR_BADENDIANNESS, which +means that the pattern is likely compiled with different endianness. */ + if (re->magic_number != MAGIC_NUMBER) - { - re = _pcre_try_flipped(re, &internal_re, study, &internal_study); - if (re == NULL) return PCRE_ERROR_BADMAGIC; - if (study != NULL) study = &internal_study; - } + return re->magic_number == REVERSED_MAGIC_NUMBER? + PCRE_ERROR_BADENDIANNESS:PCRE_ERROR_BADMAGIC; +/* Check that this pattern was compiled in the correct bit mode */ + +if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE; + switch (what) { case PCRE_INFO_OPTIONS: - *((unsigned long int *)where) = re->options & PUBLIC_OPTIONS; + *((unsigned long int *)where) = re->options & PUBLIC_COMPILE_OPTIONS; break; case PCRE_INFO_SIZE: @@ -100,6 +110,18 @@ *((size_t *)where) = (study == NULL)? 0 : study->size; break; + case PCRE_INFO_JITSIZE: +#ifdef SUPPORT_JIT + *((size_t *)where) = + (extra_data != NULL && + (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0 && + extra_data->executable_jit != NULL)? + PRIV(jit_get_size)(extra_data->executable_jit) : 0; +#else + *((size_t *)where) = 0; +#endif + break; + case PCRE_INFO_CAPTURECOUNT: *((int *)where) = re->top_bracket; break; @@ -110,7 +132,7 @@ case PCRE_INFO_FIRSTBYTE: *((int *)where) = - ((re->flags & PCRE_FIRSTSET) != 0)? re->first_byte : + ((re->flags & PCRE_FIRSTSET) != 0)? re->first_char : ((re->flags & PCRE_STARTLINE) != 0)? -1 : -2; break; @@ -118,14 +140,26 @@ block, not the internal copy (with flipped integer fields). */ case PCRE_INFO_FIRSTTABLE: - *((const uschar **)where) = - (study != NULL && (study->options & PCRE_STUDY_MAPPED) != 0)? + *((const pcre_uint8 **)where) = + (study != NULL && (study->flags & PCRE_STUDY_MAPPED) != 0)? ((const pcre_study_data *)extra_data->study_data)->start_bits : NULL; break; + case PCRE_INFO_MINLENGTH: + *((int *)where) = + (study != NULL && (study->flags & PCRE_STUDY_MINLEN) != 0)? + (int)(study->minlength) : -1; + break; + + case PCRE_INFO_JIT: + *((int *)where) = extra_data != NULL && + (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0 && + extra_data->executable_jit != NULL; + break; + case PCRE_INFO_LASTLITERAL: *((int *)where) = - ((re->flags & PCRE_REQCHSET) != 0)? re->req_byte : -1; + ((re->flags & PCRE_REQCHSET) != 0)? re->req_char : -1; break; case PCRE_INFO_NAMEENTRYSIZE: @@ -137,13 +171,16 @@ break; case PCRE_INFO_NAMETABLE: - *((const uschar **)where) = (const uschar *)re + re->name_table_offset; + *((const pcre_uchar **)where) = (const pcre_uchar *)re + re->name_table_offset; break; case PCRE_INFO_DEFAULT_TABLES: - *((const uschar **)where) = (const uschar *)(_pcre_default_tables); + *((const pcre_uint8 **)where) = (const pcre_uint8 *)(PRIV(default_tables)); break; + /* From release 8.00 this will always return TRUE because NOPARTIAL is + no longer ever set (the restrictions have been removed). */ + case PCRE_INFO_OKPARTIAL: *((int *)where) = (re->flags & PCRE_NOPARTIAL) == 0; break; @@ -154,6 +191,10 @@ case PCRE_INFO_HASCRORLF: *((int *)where) = (re->flags & PCRE_HASCRORLF) != 0; + break; + + case PCRE_INFO_MAXLOOKBEHIND: + *((int *)where) = re->max_lookbehind; break; default: return PCRE_ERROR_BADOPTION; diff --git a/pcre/pcre_get.c b/pcre/pcre_get.c --- a/pcre/pcre_get.c +++ b/pcre/pcre_get.c @@ -6,7 +6,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -65,14 +65,20 @@ (PCRE_ERROR_NOSUBSTRING) if not found */ -int +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION pcre_get_stringnumber(const pcre *code, const char *stringname) +#else +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_get_stringnumber(const pcre16 *code, PCRE_SPTR16 stringname) +#endif { int rc; int entrysize; int top, bot; -uschar *nametable; +pcre_uchar *nametable; +#ifdef COMPILE_PCRE8 if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMECOUNT, &top)) != 0) return rc; if (top <= 0) return PCRE_ERROR_NOSUBSTRING; @@ -81,14 +87,26 @@ return rc; if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMETABLE, &nametable)) != 0) return rc; +#endif +#ifdef COMPILE_PCRE16 +if ((rc = pcre16_fullinfo(code, NULL, PCRE_INFO_NAMECOUNT, &top)) != 0) + return rc; +if (top <= 0) return PCRE_ERROR_NOSUBSTRING; +if ((rc = pcre16_fullinfo(code, NULL, PCRE_INFO_NAMEENTRYSIZE, &entrysize)) != 0) + return rc; +if ((rc = pcre16_fullinfo(code, NULL, PCRE_INFO_NAMETABLE, &nametable)) != 0) + return rc; +#endif + bot = 0; while (top > bot) { int mid = (top + bot) / 2; - uschar *entry = nametable + entrysize*mid; - int c = strcmp(stringname, (char *)(entry + 2)); - if (c == 0) return (entry[0] << 8) + entry[1]; + pcre_uchar *entry = nametable + entrysize*mid; + int c = STRCMP_UC_UC((pcre_uchar *)stringname, + (pcre_uchar *)(entry + IMM2_SIZE)); + if (c == 0) return GET2(entry, 0); if (c > 0) bot = mid + 1; else top = mid; } @@ -114,15 +132,22 @@ (PCRE_ERROR_NOSUBSTRING) if not found */ -int +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION pcre_get_stringtable_entries(const pcre *code, const char *stringname, char **firstptr, char **lastptr) +#else +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_get_stringtable_entries(const pcre16 *code, PCRE_SPTR16 stringname, + PCRE_UCHAR16 **firstptr, PCRE_UCHAR16 **lastptr) +#endif { int rc; int entrysize; int top, bot; -uschar *nametable, *lastentry; +pcre_uchar *nametable, *lastentry; +#ifdef COMPILE_PCRE8 if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMECOUNT, &top)) != 0) return rc; if (top <= 0) return PCRE_ERROR_NOSUBSTRING; @@ -131,30 +156,49 @@ return rc; if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMETABLE, &nametable)) != 0) return rc; +#endif +#ifdef COMPILE_PCRE16 +if ((rc = pcre16_fullinfo(code, NULL, PCRE_INFO_NAMECOUNT, &top)) != 0) + return rc; +if (top <= 0) return PCRE_ERROR_NOSUBSTRING; +if ((rc = pcre16_fullinfo(code, NULL, PCRE_INFO_NAMEENTRYSIZE, &entrysize)) != 0) + return rc; +if ((rc = pcre16_fullinfo(code, NULL, PCRE_INFO_NAMETABLE, &nametable)) != 0) + return rc; +#endif + lastentry = nametable + entrysize * (top - 1); bot = 0; while (top > bot) { int mid = (top + bot) / 2; - uschar *entry = nametable + entrysize*mid; - int c = strcmp(stringname, (char *)(entry + 2)); + pcre_uchar *entry = nametable + entrysize*mid; + int c = STRCMP_UC_UC((pcre_uchar *)stringname, + (pcre_uchar *)(entry + IMM2_SIZE)); if (c == 0) { - uschar *first = entry; - uschar *last = entry; + pcre_uchar *first = entry; + pcre_uchar *last = entry; while (first > nametable) { - if (strcmp(stringname, (char *)(first - entrysize + 2)) != 0) break; + if (STRCMP_UC_UC((pcre_uchar *)stringname, + (pcre_uchar *)(first - entrysize + IMM2_SIZE)) != 0) break; first -= entrysize; } while (last < lastentry) { - if (strcmp(stringname, (char *)(last + entrysize + 2)) != 0) break; + if (STRCMP_UC_UC((pcre_uchar *)stringname, + (pcre_uchar *)(last + entrysize + IMM2_SIZE)) != 0) break; last += entrysize; } +#ifdef COMPILE_PCRE8 *firstptr = (char *)first; *lastptr = (char *)last; +#else + *firstptr = (PCRE_UCHAR16 *)first; + *lastptr = (PCRE_UCHAR16 *)last; +#endif return entrysize; } if (c > 0) bot = mid + 1; else top = mid; @@ -182,23 +226,39 @@ or a negative number on error */ +#ifdef COMPILE_PCRE8 static int get_first_set(const pcre *code, const char *stringname, int *ovector) +#else +static int +get_first_set(const pcre16 *code, PCRE_SPTR16 stringname, int *ovector) +#endif { -const real_pcre *re = (const real_pcre *)code; +const REAL_PCRE *re = (const REAL_PCRE *)code; int entrysize; +pcre_uchar *entry; +#ifdef COMPILE_PCRE8 char *first, *last; -uschar *entry; +#else +PCRE_UCHAR16 *first, *last; +#endif + +#ifdef COMPILE_PCRE8 if ((re->options & PCRE_DUPNAMES) == 0 && (re->flags & PCRE_JCHANGED) == 0) return pcre_get_stringnumber(code, stringname); entrysize = pcre_get_stringtable_entries(code, stringname, &first, &last); +#else +if ((re->options & PCRE_DUPNAMES) == 0 && (re->flags & PCRE_JCHANGED) == 0) + return pcre16_get_stringnumber(code, stringname); +entrysize = pcre16_get_stringtable_entries(code, stringname, &first, &last); +#endif if (entrysize <= 0) return entrysize; -for (entry = (uschar *)first; entry <= (uschar *)last; entry += entrysize) +for (entry = (pcre_uchar *)first; entry <= (pcre_uchar *)last; entry += entrysize) { - int n = (entry[0] << 8) + entry[1]; + int n = GET2(entry, 0); if (ovector[n*2] >= 0) return n; } -return (first[0] << 8) + first[1]; +return GET2(entry, 0); } @@ -231,9 +291,15 @@ PCRE_ERROR_NOSUBSTRING (-7) no such captured substring */ -int +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION pcre_copy_substring(const char *subject, int *ovector, int stringcount, int stringnumber, char *buffer, int size) +#else +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_copy_substring(PCRE_SPTR16 subject, int *ovector, int stringcount, + int stringnumber, PCRE_UCHAR16 *buffer, int size) +#endif { int yield; if (stringnumber < 0 || stringnumber >= stringcount) @@ -241,7 +307,7 @@ stringnumber *= 2; yield = ovector[stringnumber+1] - ovector[stringnumber]; if (size < yield + 1) return PCRE_ERROR_NOMEMORY; -memcpy(buffer, subject + ovector[stringnumber], yield); +memcpy(buffer, subject + ovector[stringnumber], IN_UCHARS(yield)); buffer[yield] = 0; return yield; } @@ -276,13 +342,25 @@ PCRE_ERROR_NOSUBSTRING (-7) no such captured substring */ -int -pcre_copy_named_substring(const pcre *code, const char *subject, int *ovector, - int stringcount, const char *stringname, char *buffer, int size) +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre_copy_named_substring(const pcre *code, const char *subject, + int *ovector, int stringcount, const char *stringname, + char *buffer, int size) +#else +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_copy_named_substring(const pcre16 *code, PCRE_SPTR16 subject, + int *ovector, int stringcount, PCRE_SPTR16 stringname, + PCRE_UCHAR16 *buffer, int size) +#endif { int n = get_first_set(code, stringname, ovector); if (n <= 0) return n; +#ifdef COMPILE_PCRE8 return pcre_copy_substring(subject, ovector, stringcount, n, buffer, size); +#else +return pcre16_copy_substring(subject, ovector, stringcount, n, buffer, size); +#endif } @@ -308,29 +386,39 @@ PCRE_ERROR_NOMEMORY (-6) failed to get store */ -int +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION pcre_get_substring_list(const char *subject, int *ovector, int stringcount, const char ***listptr) +#else +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_get_substring_list(PCRE_SPTR16 subject, int *ovector, int stringcount, + PCRE_SPTR16 **listptr) +#endif { int i; -int size = sizeof(char *); +int size = sizeof(pcre_uchar *); int double_count = stringcount * 2; -char **stringlist; -char *p; +pcre_uchar **stringlist; +pcre_uchar *p; for (i = 0; i < double_count; i += 2) - size += sizeof(char *) + ovector[i+1] - ovector[i] + 1; + size += sizeof(pcre_uchar *) + IN_UCHARS(ovector[i+1] - ovector[i] + 1); -stringlist = (char **)(pcre_malloc)(size); +stringlist = (pcre_uchar **)(PUBL(malloc))(size); if (stringlist == NULL) return PCRE_ERROR_NOMEMORY; +#ifdef COMPILE_PCRE8 *listptr = (const char **)stringlist; -p = (char *)(stringlist + stringcount + 1); +#else +*listptr = (PCRE_SPTR16 *)stringlist; +#endif +p = (pcre_uchar *)(stringlist + stringcount + 1); for (i = 0; i < double_count; i += 2) { int len = ovector[i+1] - ovector[i]; - memcpy(p, subject + ovector[i], len); + memcpy(p, subject + ovector[i], IN_UCHARS(len)); *stringlist++ = p; p += len; *p++ = 0; @@ -347,16 +435,22 @@ *************************************************/ /* This function exists for the benefit of people calling PCRE from non-C -programs that can call its functions, but not free() or (pcre_free)() directly. +programs that can call its functions, but not free() or (PUBL(free))() +directly. Argument: the result of a previous pcre_get_substring_list() Returns: nothing */ -void +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN void PCRE_CALL_CONVENTION pcre_free_substring_list(const char **pointer) +#else +PCRE_EXP_DEFN void PCRE_CALL_CONVENTION +pcre16_free_substring_list(PCRE_SPTR16 *pointer) +#endif { -(pcre_free)((void *)pointer); +(PUBL(free))((void *)pointer); } @@ -386,21 +480,31 @@ PCRE_ERROR_NOSUBSTRING (-7) substring not present */ -int +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION pcre_get_substring(const char *subject, int *ovector, int stringcount, int stringnumber, const char **stringptr) +#else +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_get_substring(PCRE_SPTR16 subject, int *ovector, int stringcount, + int stringnumber, PCRE_SPTR16 *stringptr) +#endif { int yield; -char *substring; +pcre_uchar *substring; if (stringnumber < 0 || stringnumber >= stringcount) return PCRE_ERROR_NOSUBSTRING; stringnumber *= 2; yield = ovector[stringnumber+1] - ovector[stringnumber]; -substring = (char *)(pcre_malloc)(yield + 1); +substring = (pcre_uchar *)(PUBL(malloc))(IN_UCHARS(yield + 1)); if (substring == NULL) return PCRE_ERROR_NOMEMORY; -memcpy(substring, subject + ovector[stringnumber], yield); +memcpy(substring, subject + ovector[stringnumber], IN_UCHARS(yield)); substring[yield] = 0; -*stringptr = substring; +#ifdef COMPILE_PCRE8 +*stringptr = (const char *)substring; +#else +*stringptr = (PCRE_SPTR16)substring; +#endif return yield; } @@ -433,13 +537,25 @@ PCRE_ERROR_NOSUBSTRING (-7) no such captured substring */ -int -pcre_get_named_substring(const pcre *code, const char *subject, int *ovector, - int stringcount, const char *stringname, const char **stringptr) +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre_get_named_substring(const pcre *code, const char *subject, + int *ovector, int stringcount, const char *stringname, + const char **stringptr) +#else +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_get_named_substring(const pcre16 *code, PCRE_SPTR16 subject, + int *ovector, int stringcount, PCRE_SPTR16 stringname, + PCRE_SPTR16 *stringptr) +#endif { int n = get_first_set(code, stringname, ovector); if (n <= 0) return n; +#ifdef COMPILE_PCRE8 return pcre_get_substring(subject, ovector, stringcount, n, stringptr); +#else +return pcre16_get_substring(subject, ovector, stringcount, n, stringptr); +#endif } @@ -450,16 +566,22 @@ *************************************************/ /* This function exists for the benefit of people calling PCRE from non-C -programs that can call its functions, but not free() or (pcre_free)() directly. +programs that can call its functions, but not free() or (PUBL(free))() +directly. Argument: the result of a previous pcre_get_substring() Returns: nothing */ -void +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN void PCRE_CALL_CONVENTION pcre_free_substring(const char *pointer) +#else +PCRE_EXP_DEFN void PCRE_CALL_CONVENTION +pcre16_free_substring(PCRE_SPTR16 pointer) +#endif { -(pcre_free)((void *)pointer); +(PUBL(free))((void *)pointer); } /* End of pcre_get.c */ diff --git a/pcre/pcre_globals.c b/pcre/pcre_globals.c --- a/pcre/pcre_globals.c +++ b/pcre/pcre_globals.c @@ -6,7 +6,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -43,21 +43,42 @@ However, it calls memory allocation and freeing functions via the four indirections below, and it can optionally do callouts, using the fifth indirection. These values can be changed by the caller, but are shared between -all threads. However, when compiling for Virtual Pascal, things are done -differently, and global variables are not used (see pcre.in). */ +all threads. +For MS Visual Studio and Symbian OS, there are problems in initializing these +variables to non-local functions. In these cases, therefore, an indirection via +a local function is used. + +Also, when compiling for Virtual Pascal, things are done differently, and +global variables are not used. */ + #ifdef HAVE_CONFIG_H #include "config.h" #endif #include "pcre_internal.h" -#ifndef VPCOMPAT -PCRE_EXP_DATA_DEFN void *(*pcre_malloc)(size_t) = malloc; -PCRE_EXP_DATA_DEFN void (*pcre_free)(void *) = free; -PCRE_EXP_DATA_DEFN void *(*pcre_stack_malloc)(size_t) = malloc; -PCRE_EXP_DATA_DEFN void (*pcre_stack_free)(void *) = free; -PCRE_EXP_DATA_DEFN int (*pcre_callout)(pcre_callout_block *) = NULL; +#if defined _MSC_VER || defined __SYMBIAN32__ +static void* LocalPcreMalloc(size_t aSize) + { + return malloc(aSize); + } +static void LocalPcreFree(void* aPtr) + { + free(aPtr); + } +PCRE_EXP_DATA_DEFN void *(*PUBL(malloc))(size_t) = LocalPcreMalloc; +PCRE_EXP_DATA_DEFN void (*PUBL(free))(void *) = LocalPcreFree; +PCRE_EXP_DATA_DEFN void *(*PUBL(stack_malloc))(size_t) = LocalPcreMalloc; +PCRE_EXP_DATA_DEFN void (*PUBL(stack_free))(void *) = LocalPcreFree; +PCRE_EXP_DATA_DEFN int (*PUBL(callout))(PUBL(callout_block) *) = NULL; + +#elif !defined VPCOMPAT +PCRE_EXP_DATA_DEFN void *(*PUBL(malloc))(size_t) = malloc; +PCRE_EXP_DATA_DEFN void (*PUBL(free))(void *) = free; +PCRE_EXP_DATA_DEFN void *(*PUBL(stack_malloc))(size_t) = malloc; +PCRE_EXP_DATA_DEFN void (*PUBL(stack_free))(void *) = free; +PCRE_EXP_DATA_DEFN int (*PUBL(callout))(PUBL(callout_block) *) = NULL; #endif /* End of pcre_globals.c */ diff --git a/pcre/pcre_info.c b/pcre/pcre_info.c deleted file mode 100644 --- a/pcre/pcre_info.c +++ /dev/null @@ -1,93 +0,0 @@ -/************************************************* -* Perl-Compatible Regular Expressions * -*************************************************/ - -/* PCRE is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. - - Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* This module contains the external function pcre_info(), which gives some -information about a compiled pattern. However, use of this function is now -deprecated, as it has been superseded by pcre_fullinfo(). */ - - -#ifdef HAVE_CONFIG_H -#include "config.h" -#endif - -#include "pcre_internal.h" - - -/************************************************* -* (Obsolete) Return info about compiled pattern * -*************************************************/ - -/* This is the original "info" function. It picks potentially useful data out -of the private structure, but its interface was too rigid. It remains for -backwards compatibility. The public options are passed back in an int - though -the re->options field has been expanded to a long int, all the public options -at the low end of it, and so even on 16-bit systems this will still be OK. -Therefore, I haven't changed the API for pcre_info(). - -Arguments: - argument_re points to compiled code - optptr where to pass back the options - first_byte where to pass back the first character, - or -1 if multiline and all branches start ^, - or -2 otherwise - -Returns: number of capturing subpatterns - or negative values on error -*/ - -PCRE_EXP_DEFN int -pcre_info(const pcre *argument_re, int *optptr, int *first_byte) -{ -real_pcre internal_re; -const real_pcre *re = (const real_pcre *)argument_re; -if (re == NULL) return PCRE_ERROR_NULL; -if (re->magic_number != MAGIC_NUMBER) - { - re = _pcre_try_flipped(re, &internal_re, NULL, NULL); - if (re == NULL) return PCRE_ERROR_BADMAGIC; - } -if (optptr != NULL) *optptr = (int)(re->options & PUBLIC_OPTIONS); -if (first_byte != NULL) - *first_byte = ((re->flags & PCRE_FIRSTSET) != 0)? re->first_byte : - ((re->flags & PCRE_STARTLINE) != 0)? -1 : -2; -return re->top_bracket; -} - -/* End of pcre_info.c */ diff --git a/pcre/pcre_internal.h b/pcre/pcre_internal.h --- a/pcre/pcre_internal.h +++ b/pcre/pcre_internal.h @@ -7,1133 +7,2339 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - -/* This header contains definitions that are shared between the different -modules, but which are not relevant to the exported API. This includes some -functions whose names all begin with "_pcre_". */ - -#ifndef PCRE_INTERNAL_H -#define PCRE_INTERNAL_H - -#include "config.h" - -/* Define DEBUG to get debugging output on stdout. */ - -#if 0 -#define DEBUG -#endif - -/* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef -inline, and there are *still* stupid compilers about that don't like indented -pre-processor statements, or at least there were when I first wrote this. After -all, it had only been about 10 years then... - -It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so -be absolutely sure we get our version. */ - -#undef DPRINTF -#ifdef DEBUG -#define DPRINTF(p) printf p -#else -#define DPRINTF(p) /* Nothing */ -#endif - - -/* Standard C headers plus the external interface definition. The only time -setjmp and stdarg are used is when NO_RECURSE is set. */ - -#include -#include -#include -#include -#include -#include -#include -#include - -/* When compiling a DLL for Windows, the exported symbols have to be declared -using some MS magic. I found some useful information on this web page: -http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the -information there, using __declspec(dllexport) without "extern" we have a -definition; with "extern" we have a declaration. The settings here override the -setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL, -which is all that is needed for applications (they just import the symbols). We -use: - - PCRE_EXP_DECL for declarations - PCRE_EXP_DEFN for definitions of exported functions - PCRE_EXP_DATA_DEFN for definitions of exported variables - -The reason for the two DEFN macros is that in non-Windows environments, one -does not want to have "extern" before variable definitions because it leads to -compiler warnings. So we distinguish between functions and variables. In -Windows, the two should always be the same. - -The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest, -which is an application, but needs to import this file in order to "peek" at -internals, can #include pcre.h first to get an application's-eye view. - -In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon, -special-purpose environments) might want to stick other stuff in front of -exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and -PCRE_EXP_DATA_DEFN only if they are not already set. */ - -#ifndef PCRE_EXP_DECL -# ifdef _WIN32 -# ifndef PCRE_STATIC -# define PCRE_EXP_DECL extern __declspec(dllexport) -# define PCRE_EXP_DEFN __declspec(dllexport) -# define PCRE_EXP_DATA_DEFN __declspec(dllexport) -# else -# define PCRE_EXP_DECL extern -# define PCRE_EXP_DEFN -# define PCRE_EXP_DATA_DEFN -# endif -# else -# ifdef __cplusplus -# define PCRE_EXP_DECL extern "C" -# else -# define PCRE_EXP_DECL extern -# endif -# ifndef PCRE_EXP_DEFN -# define PCRE_EXP_DEFN PCRE_EXP_DECL -# endif -# ifndef PCRE_EXP_DATA_DEFN -# define PCRE_EXP_DATA_DEFN -# endif -# endif -#endif - -/* We need to have types that specify unsigned 16-bit and 32-bit integers. We -cannot determine these outside the compilation (e.g. by running a program as -part of "configure") because PCRE is often cross-compiled for use on other -systems. Instead we make use of the maximum sizes that are available at -preprocessor time in standard C environments. */ - -#if USHRT_MAX == 65535 - typedef unsigned short pcre_uint16; -#elif UINT_MAX == 65535 - typedef unsigned int pcre_uint16; -#else - #error Cannot determine a type for 16-bit unsigned integers -#endif - -#if UINT_MAX == 4294967295 - typedef unsigned int pcre_uint32; -#elif ULONG_MAX == 4294967295 - typedef unsigned long int pcre_uint32; -#else - #error Cannot determine a type for 32-bit unsigned integers -#endif - -/* All character handling must be done as unsigned characters. Otherwise there -are problems with top-bit-set characters and functions such as isspace(). -However, we leave the interface to the outside world as char *, because that -should make things easier for callers. We define a short type for unsigned char -to save lots of typing. I tried "uchar", but it causes problems on Digital -Unix, where it is defined in sys/types, so use "uschar" instead. */ - -typedef unsigned char uschar; - -/* This is an unsigned int value that no character can ever have. UTF-8 -characters only go up to 0x7fffffff (though Unicode doesn't go beyond -0x0010ffff). */ - -#define NOTACHAR 0xffffffff - -/* PCRE is able to support several different kinds of newline (CR, LF, CRLF, -"any" and "anycrlf" at present). The following macros are used to package up -testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various -modules to indicate in which datablock the parameters exist, and what the -start/end of string field names are. */ - -#define NLTYPE_FIXED 0 /* Newline is a fixed length string */ -#define NLTYPE_ANY 1 /* Newline is any Unicode line ending */ -#define NLTYPE_ANYCRLF 2 /* Newline is CR, LF, or CRLF */ - -/* This macro checks for a newline at the given position */ - -#define IS_NEWLINE(p) \ - ((NLBLOCK->nltype != NLTYPE_FIXED)? \ - ((p) < NLBLOCK->PSEND && \ - _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\ - utf8)) \ - : \ - ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \ - (p)[0] == NLBLOCK->nl[0] && \ - (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \ - ) \ - ) - -/* This macro checks for a newline immediately preceding the given position */ - -#define WAS_NEWLINE(p) \ - ((NLBLOCK->nltype != NLTYPE_FIXED)? \ - ((p) > NLBLOCK->PSSTART && \ - _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \ - &(NLBLOCK->nllen), utf8)) \ - : \ - ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \ - (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \ - (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \ - ) \ - ) - -/* When PCRE is compiled as a C++ library, the subject pointer can be replaced -with a custom type. This makes it possible, for example, to allow pcre_exec() -to process subject strings that are discontinuous by using a smart pointer -class. It must always be possible to inspect all of the subject string in -pcre_exec() because of the way it backtracks. Two macros are required in the -normal case, for sign-unspecified and unsigned char pointers. The former is -used for the external interface and appears in pcre.h, which is why its name -must begin with PCRE_. */ - -#ifdef CUSTOM_SUBJECT_PTR -#define PCRE_SPTR CUSTOM_SUBJECT_PTR -#define USPTR CUSTOM_SUBJECT_PTR -#else -#define PCRE_SPTR const char * -#define USPTR const unsigned char * -#endif - - - -/* Include the public PCRE header and the definitions of UCP character property -values. */ - -#include "pcre.h" -#include "ucp.h" - -/* When compiling for use with the Virtual Pascal compiler, these functions -need to have their names changed. PCRE must be compiled with the -DVPCOMPAT -option on the command line. */ - -#ifdef VPCOMPAT -#define strlen(s) _strlen(s) -#define strncmp(s1,s2,m) _strncmp(s1,s2,m) -#define memcmp(s,c,n) _memcmp(s,c,n) -#define memcpy(d,s,n) _memcpy(d,s,n) -#define memmove(d,s,n) _memmove(d,s,n) -#define memset(s,c,n) _memset(s,c,n) -#else /* VPCOMPAT */ - -/* To cope with SunOS4 and other systems that lack memmove() but have bcopy(), -define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY -is set. Otherwise, include an emulating function for those systems that have -neither (there some non-Unix environments where this is the case). */ - -#ifndef HAVE_MEMMOVE -#undef memmove /* some systems may have a macro */ -#ifdef HAVE_BCOPY -#define memmove(a, b, c) bcopy(b, a, c) -#else /* HAVE_BCOPY */ -static void * -pcre_memmove(void *d, const void *s, size_t n) -{ -size_t i; -unsigned char *dest = (unsigned char *)d; -const unsigned char *src = (const unsigned char *)s; -if (dest > src) - { - dest += n; - src += n; - for (i = 0; i < n; ++i) *(--dest) = *(--src); - return (void *)dest; - } -else - { - for (i = 0; i < n; ++i) *dest++ = *src++; - return (void *)(dest - n); - } -} -#define memmove(a, b, c) pcre_memmove(a, b, c) -#endif /* not HAVE_BCOPY */ -#endif /* not HAVE_MEMMOVE */ -#endif /* not VPCOMPAT */ - - -/* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored -in big-endian order) by default. These are used, for example, to link from the -start of a subpattern to its alternatives and its end. The use of 2 bytes per -offset limits the size of the compiled regex to around 64K, which is big enough -for almost everybody. However, I received a request for an even bigger limit. -For this reason, and also to make the code easier to maintain, the storing and -loading of offsets from the byte string is now handled by the macros that are -defined here. - -The macros are controlled by the value of LINK_SIZE. This defaults to 2 in -the config.h file, but can be overridden by using -D on the command line. This -is automated on Unix systems via the "configure" command. */ - -#if LINK_SIZE == 2 - -#define PUT(a,n,d) \ - (a[n] = (d) >> 8), \ - (a[(n)+1] = (d) & 255) - -#define GET(a,n) \ - (((a)[n] << 8) | (a)[(n)+1]) - -#define MAX_PATTERN_SIZE (1 << 16) - - -#elif LINK_SIZE == 3 - -#define PUT(a,n,d) \ - (a[n] = (d) >> 16), \ - (a[(n)+1] = (d) >> 8), \ - (a[(n)+2] = (d) & 255) - -#define GET(a,n) \ - (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2]) - -#define MAX_PATTERN_SIZE (1 << 24) - - -#elif LINK_SIZE == 4 - -#define PUT(a,n,d) \ - (a[n] = (d) >> 24), \ - (a[(n)+1] = (d) >> 16), \ - (a[(n)+2] = (d) >> 8), \ - (a[(n)+3] = (d) & 255) - -#define GET(a,n) \ - (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3]) - -#define MAX_PATTERN_SIZE (1 << 30) /* Keep it positive */ - - -#else -#error LINK_SIZE must be either 2, 3, or 4 -#endif - - -/* Convenience macro defined in terms of the others */ - -#define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE - - -/* PCRE uses some other 2-byte quantities that do not change when the size of -offsets changes. There are used for repeat counts and for other things such as -capturing parenthesis numbers in back references. */ - -#define PUT2(a,n,d) \ - a[n] = (d) >> 8; \ - a[(n)+1] = (d) & 255 - -#define GET2(a,n) \ - (((a)[n] << 8) | (a)[(n)+1]) - -#define PUT2INC(a,n,d) PUT2(a,n,d), a += 2 - - -/* When UTF-8 encoding is being used, a character is no longer just a single -byte. The macros for character handling generate simple sequences when used in -byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should -never be called in byte mode. To make sure it can never even appear when UTF-8 -support is omitted, we don't even define it. */ - -#ifndef SUPPORT_UTF8 -#define NEXTCHAR(p) p++; -#define GETCHAR(c, eptr) c = *eptr; -#define GETCHARTEST(c, eptr) c = *eptr; -#define GETCHARINC(c, eptr) c = *eptr++; -#define GETCHARINCTEST(c, eptr) c = *eptr++; -#define GETCHARLEN(c, eptr, len) c = *eptr; -/* #define BACKCHAR(eptr) */ - -#else /* SUPPORT_UTF8 */ - -/* Advance a character pointer one byte in non-UTF-8 mode and by one character -in UTF-8 mode. */ - -#define NEXTCHAR(p) \ - p++; \ - if (utf8) { while((*p & 0xc0) == 0x80) p++; } - -/* Get the next UTF-8 character, not advancing the pointer. This is called when -we know we are in UTF-8 mode. */ - -#define GETCHAR(c, eptr) \ - c = *eptr; \ - if (c >= 0xc0) \ - { \ - int gcii; \ - int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \ - int gcss = 6*gcaa; \ - c = (c & _pcre_utf8_table3[gcaa]) << gcss; \ - for (gcii = 1; gcii <= gcaa; gcii++) \ - { \ - gcss -= 6; \ - c |= (eptr[gcii] & 0x3f) << gcss; \ - } \ - } - -/* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the -pointer. */ - -#define GETCHARTEST(c, eptr) \ - c = *eptr; \ - if (utf8 && c >= 0xc0) \ - { \ - int gcii; \ - int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \ - int gcss = 6*gcaa; \ - c = (c & _pcre_utf8_table3[gcaa]) << gcss; \ - for (gcii = 1; gcii <= gcaa; gcii++) \ - { \ - gcss -= 6; \ - c |= (eptr[gcii] & 0x3f) << gcss; \ - } \ - } - -/* Get the next UTF-8 character, advancing the pointer. This is called when we -know we are in UTF-8 mode. */ - -#define GETCHARINC(c, eptr) \ - c = *eptr++; \ - if (c >= 0xc0) \ - { \ - int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \ - int gcss = 6*gcaa; \ - c = (c & _pcre_utf8_table3[gcaa]) << gcss; \ - while (gcaa-- > 0) \ - { \ - gcss -= 6; \ - c |= (*eptr++ & 0x3f) << gcss; \ - } \ - } - -/* Get the next character, testing for UTF-8 mode, and advancing the pointer */ - -#define GETCHARINCTEST(c, eptr) \ - c = *eptr++; \ - if (utf8 && c >= 0xc0) \ - { \ - int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \ - int gcss = 6*gcaa; \ - c = (c & _pcre_utf8_table3[gcaa]) << gcss; \ - while (gcaa-- > 0) \ - { \ - gcss -= 6; \ - c |= (*eptr++ & 0x3f) << gcss; \ - } \ - } - -/* Get the next UTF-8 character, not advancing the pointer, incrementing length -if there are extra bytes. This is called when we know we are in UTF-8 mode. */ - -#define GETCHARLEN(c, eptr, len) \ - c = *eptr; \ - if (c >= 0xc0) \ - { \ - int gcii; \ - int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \ - int gcss = 6*gcaa; \ - c = (c & _pcre_utf8_table3[gcaa]) << gcss; \ - for (gcii = 1; gcii <= gcaa; gcii++) \ - { \ - gcss -= 6; \ - c |= (eptr[gcii] & 0x3f) << gcss; \ - } \ - len += gcaa; \ - } - -/* If the pointer is not at the start of a character, move it back until -it is. This is called only in UTF-8 mode - we don't put a test within the macro -because almost all calls are already within a block of UTF-8 only code. */ - -#define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr-- - -#endif - - -/* In case there is no definition of offsetof() provided - though any proper -Standard C system should have one. */ - -#ifndef offsetof -#define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field)) -#endif - - -/* These are the public options that can change during matching. */ - -#define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL) - -/* Private flags containing information about the compiled regex. They used to -live at the top end of the options word, but that got almost full, so now they -are in a 16-bit flags word. */ - -#define PCRE_NOPARTIAL 0x0001 /* can't use partial with this regex */ -#define PCRE_FIRSTSET 0x0002 /* first_byte is set */ -#define PCRE_REQCHSET 0x0004 /* req_byte is set */ -#define PCRE_STARTLINE 0x0008 /* start after \n for multiline */ -#define PCRE_JCHANGED 0x0010 /* j option used in regex */ -#define PCRE_HASCRORLF 0x0020 /* explicit \r or \n in pattern */ - -/* Options for the "extra" block produced by pcre_study(). */ - -#define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */ - -/* Masks for identifying the public options that are permitted at compile -time, run time, or study time, respectively. */ - -#define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \ - PCRE_NEWLINE_ANYCRLF) - -#define PUBLIC_OPTIONS \ - (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \ - PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \ - PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \ - PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \ - PCRE_JAVASCRIPT_COMPAT) - -#define PUBLIC_EXEC_OPTIONS \ - (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \ - PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE) - -#define PUBLIC_DFA_EXEC_OPTIONS \ - (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \ - PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \ - PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE) - -#define PUBLIC_STUDY_OPTIONS 0 /* None defined */ - -/* Magic number to provide a small check against being handed junk. Also used -to detect whether a pattern was compiled on a host of different endianness. */ - -#define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */ - -/* Negative values for the firstchar and reqchar variables */ - -#define REQ_UNSET (-2) -#define REQ_NONE (-1) - -/* The maximum remaining length of subject we are prepared to search for a -req_byte match. */ - -#define REQ_BYTE_MAX 1000 - -/* Flags added to firstbyte or reqbyte; a "non-literal" item is either a -variable-length repeat, or a anything other than literal characters. */ - -#define REQ_CASELESS 0x0100 /* indicates caselessness */ -#define REQ_VARY 0x0200 /* reqbyte followed non-literal item */ - -/* Miscellaneous definitions */ - -typedef int BOOL; - -#define FALSE 0 -#define TRUE 1 - -/* Escape items that are just an encoding of a particular data value. */ - -#ifndef ESC_e -#define ESC_e 27 -#endif - -#ifndef ESC_f -#define ESC_f '\f' -#endif - -#ifndef ESC_n -#define ESC_n '\n' -#endif - -#ifndef ESC_r -#define ESC_r '\r' -#endif - -/* We can't officially use ESC_t because it is a POSIX reserved identifier -(presumably because of all the others like size_t). */ - -#ifndef ESC_tee -#define ESC_tee '\t' -#endif - -/* Codes for different types of Unicode property */ - -#define PT_ANY 0 /* Any property - matches all chars */ -#define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */ -#define PT_GC 2 /* General characteristic (e.g. L) */ -#define PT_PC 3 /* Particular characteristic (e.g. Lu) */ -#define PT_SC 4 /* Script (e.g. Han) */ - -/* Flag bits and data types for the extended class (OP_XCLASS) for classes that -contain UTF-8 characters with values greater than 255. */ - -#define XCL_NOT 0x01 /* Flag: this is a negative class */ -#define XCL_MAP 0x02 /* Flag: a 32-byte map is present */ - -#define XCL_END 0 /* Marks end of individual items */ -#define XCL_SINGLE 1 /* Single item (one multibyte char) follows */ -#define XCL_RANGE 2 /* A range (two multibyte chars) follows */ -#define XCL_PROP 3 /* Unicode property (2-byte property code follows) */ -#define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */ - -/* These are escaped items that aren't just an encoding of a particular data -value such as \n. They must have non-zero values, as check_escape() returns -their negation. Also, they must appear in the same order as in the opcode -definitions below, up to ESC_z. There's a dummy for OP_ANY because it -corresponds to "." rather than an escape sequence, and another for OP_ALLANY -(which is used for [^] in JavaScript compatibility mode). - -The final escape must be ESC_REF as subsequent values are used for -backreferences (\1, \2, \3, etc). There are two tests in the code for an escape -greater than ESC_b and less than ESC_Z to detect the types that may be -repeated. These are the types that consume characters. If any new escapes are -put in between that don't consume a character, that code will have to change. -*/ - -enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, - ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, - ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k, - ESC_REF }; - - -/* Opcode table: Starting from 1 (i.e. after OP_END), the values up to -OP_EOD must correspond in order to the list of escapes immediately above. - -*** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions -that follow must also be updated to match. There is also a table called -"coptable" in pcre_dfa_exec.c that must be updated. */ - -enum { - OP_END, /* 0 End of pattern */ - - /* Values corresponding to backslashed metacharacters */ - - OP_SOD, /* 1 Start of data: \A */ - OP_SOM, /* 2 Start of match (subject + offset): \G */ - OP_SET_SOM, /* 3 Set start of match (\K) */ - OP_NOT_WORD_BOUNDARY, /* 4 \B */ - OP_WORD_BOUNDARY, /* 5 \b */ - OP_NOT_DIGIT, /* 6 \D */ - OP_DIGIT, /* 7 \d */ - OP_NOT_WHITESPACE, /* 8 \S */ - OP_WHITESPACE, /* 9 \s */ - OP_NOT_WORDCHAR, /* 10 \W */ - OP_WORDCHAR, /* 11 \w */ - OP_ANY, /* 12 Match any character (subject to DOTALL) */ - OP_ALLANY, /* 13 Match any character (not subject to DOTALL) */ - OP_ANYBYTE, /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */ - OP_NOTPROP, /* 15 \P (not Unicode property) */ - OP_PROP, /* 16 \p (Unicode property) */ - OP_ANYNL, /* 17 \R (any newline sequence) */ - OP_NOT_HSPACE, /* 18 \H (not horizontal whitespace) */ - OP_HSPACE, /* 19 \h (horizontal whitespace) */ - OP_NOT_VSPACE, /* 20 \V (not vertical whitespace) */ - OP_VSPACE, /* 21 \v (vertical whitespace) */ - OP_EXTUNI, /* 22 \X (extended Unicode sequence */ - OP_EODN, /* 23 End of data or \n at end of data: \Z. */ - OP_EOD, /* 24 End of data: \z */ - - OP_OPT, /* 25 Set runtime options */ - OP_CIRC, /* 26 Start of line - varies with multiline switch */ - OP_DOLL, /* 27 End of line - varies with multiline switch */ - OP_CHAR, /* 28 Match one character, casefully */ - OP_CHARNC, /* 29 Match one character, caselessly */ - OP_NOT, /* 30 Match one character, not the following one */ - - OP_STAR, /* 31 The maximizing and minimizing versions of */ - OP_MINSTAR, /* 32 these six opcodes must come in pairs, with */ - OP_PLUS, /* 33 the minimizing one second. */ - OP_MINPLUS, /* 34 This first set applies to single characters.*/ - OP_QUERY, /* 35 */ - OP_MINQUERY, /* 36 */ - - OP_UPTO, /* 37 From 0 to n matches */ - OP_MINUPTO, /* 38 */ - OP_EXACT, /* 39 Exactly n matches */ - - OP_POSSTAR, /* 40 Possessified star */ - OP_POSPLUS, /* 41 Possessified plus */ - OP_POSQUERY, /* 42 Posesssified query */ - OP_POSUPTO, /* 43 Possessified upto */ - - OP_NOTSTAR, /* 44 The maximizing and minimizing versions of */ - OP_NOTMINSTAR, /* 45 these six opcodes must come in pairs, with */ - OP_NOTPLUS, /* 46 the minimizing one second. They must be in */ - OP_NOTMINPLUS, /* 47 exactly the same order as those above. */ - OP_NOTQUERY, /* 48 This set applies to "not" single characters. */ - OP_NOTMINQUERY, /* 49 */ - - OP_NOTUPTO, /* 50 From 0 to n matches */ - OP_NOTMINUPTO, /* 51 */ - OP_NOTEXACT, /* 52 Exactly n matches */ - - OP_NOTPOSSTAR, /* 53 Possessified versions */ - OP_NOTPOSPLUS, /* 54 */ - OP_NOTPOSQUERY, /* 55 */ - OP_NOTPOSUPTO, /* 56 */ - - OP_TYPESTAR, /* 57 The maximizing and minimizing versions of */ - OP_TYPEMINSTAR, /* 58 these six opcodes must come in pairs, with */ - OP_TYPEPLUS, /* 59 the minimizing one second. These codes must */ - OP_TYPEMINPLUS, /* 60 be in exactly the same order as those above. */ - OP_TYPEQUERY, /* 61 This set applies to character types such as \d */ - OP_TYPEMINQUERY, /* 62 */ - - OP_TYPEUPTO, /* 63 From 0 to n matches */ - OP_TYPEMINUPTO, /* 64 */ - OP_TYPEEXACT, /* 65 Exactly n matches */ - - OP_TYPEPOSSTAR, /* 66 Possessified versions */ - OP_TYPEPOSPLUS, /* 67 */ - OP_TYPEPOSQUERY, /* 68 */ - OP_TYPEPOSUPTO, /* 69 */ - - OP_CRSTAR, /* 70 The maximizing and minimizing versions of */ - OP_CRMINSTAR, /* 71 all these opcodes must come in pairs, with */ - OP_CRPLUS, /* 72 the minimizing one second. These codes must */ - OP_CRMINPLUS, /* 73 be in exactly the same order as those above. */ - OP_CRQUERY, /* 74 These are for character classes and back refs */ - OP_CRMINQUERY, /* 75 */ - OP_CRRANGE, /* 76 These are different to the three sets above. */ - OP_CRMINRANGE, /* 77 */ - - OP_CLASS, /* 78 Match a character class, chars < 256 only */ - OP_NCLASS, /* 79 Same, but the bitmap was created from a negative - class - the difference is relevant only when a UTF-8 - character > 255 is encountered. */ - - OP_XCLASS, /* 80 Extended class for handling UTF-8 chars within the - class. This does both positive and negative. */ - - OP_REF, /* 81 Match a back reference */ - OP_RECURSE, /* 82 Match a numbered subpattern (possibly recursive) */ - OP_CALLOUT, /* 83 Call out to external function if provided */ - - OP_ALT, /* 84 Start of alternation */ - OP_KET, /* 85 End of group that doesn't have an unbounded repeat */ - OP_KETRMAX, /* 86 These two must remain together and in this */ - OP_KETRMIN, /* 87 order. They are for groups the repeat for ever. */ - - /* The assertions must come before BRA, CBRA, ONCE, and COND.*/ - - OP_ASSERT, /* 88 Positive lookahead */ - OP_ASSERT_NOT, /* 89 Negative lookahead */ - OP_ASSERTBACK, /* 90 Positive lookbehind */ - OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */ - OP_REVERSE, /* 92 Move pointer back - used in lookbehind assertions */ - - /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first, - as there's a test for >= ONCE for a subpattern that isn't an assertion. */ - - OP_ONCE, /* 93 Atomic group */ - OP_BRA, /* 94 Start of non-capturing bracket */ - OP_CBRA, /* 95 Start of capturing bracket */ - OP_COND, /* 96 Conditional group */ - - /* These three must follow the previous three, in the same order. There's a - check for >= SBRA to distinguish the two sets. */ - - OP_SBRA, /* 97 Start of non-capturing bracket, check empty */ - OP_SCBRA, /* 98 Start of capturing bracket, check empty */ - OP_SCOND, /* 99 Conditional group, check empty */ - - OP_CREF, /* 100 Used to hold a capture number as condition */ - OP_RREF, /* 101 Used to hold a recursion number as condition */ - OP_DEF, /* 102 The DEFINE condition */ - - OP_BRAZERO, /* 103 These two must remain together and in this */ - OP_BRAMINZERO, /* 104 order. */ - - /* These are backtracking control verbs */ - - OP_PRUNE, /* 105 */ - OP_SKIP, /* 106 */ - OP_THEN, /* 107 */ - OP_COMMIT, /* 108 */ - - /* These are forced failure and success verbs */ - - OP_FAIL, /* 109 */ - OP_ACCEPT, /* 110 */ - - /* This is used to skip a subpattern with a {0} quantifier */ - - OP_SKIPZERO /* 111 */ -}; - - -/* This macro defines textual names for all the opcodes. These are used only -for debugging. The macro is referenced only in pcre_printint.c. */ - -#define OP_NAME_LIST \ - "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d", \ - "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte", \ - "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v", \ - "extuni", "\\Z", "\\z", \ - "Opt", "^", "$", "char", "charnc", "not", \ - "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \ - "*+","++", "?+", "{", \ - "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \ - "*+","++", "?+", "{", \ - "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \ - "*+","++", "?+", "{", \ - "*", "*?", "+", "+?", "?", "??", "{", "{", \ - "class", "nclass", "xclass", "Ref", "Recurse", "Callout", \ - "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not", \ - "AssertB", "AssertB not", "Reverse", \ - "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond", \ - "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero", \ - "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT", \ - "Skip zero" - - -/* This macro defines the length of fixed length operations in the compiled -regex. The lengths are used when searching for specific things, and also in the -debugging printing of a compiled regex. We use a macro so that it can be -defined close to the definitions of the opcodes themselves. - -As things have been extended, some of these are no longer fixed lenths, but are -minima instead. For example, the length of a single-character repeat may vary -in UTF-8 mode. The code that uses this table must know about such things. */ - -#define OP_LENGTHS \ - 1, /* End */ \ - 1, 1, 1, 1, 1, /* \A, \G, \K, \B, \b */ \ - 1, 1, 1, 1, 1, 1, /* \D, \d, \S, \s, \W, \w */ \ - 1, 1, 1, /* Any, AllAny, Anybyte */ \ - 3, 3, 1, /* NOTPROP, PROP, EXTUNI */ \ - 1, 1, 1, 1, 1, /* \R, \H, \h, \V, \v */ \ - 1, 1, 2, 1, 1, /* \Z, \z, Opt, ^, $ */ \ - 2, /* Char - the minimum length */ \ - 2, /* Charnc - the minimum length */ \ - 2, /* not */ \ - /* Positive single-char repeats ** These are */ \ - 2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \ - 4, 4, 4, /* upto, minupto, exact ** UTF-8 mode */ \ - 2, 2, 2, 4, /* *+, ++, ?+, upto+ */ \ - /* Negative single-char repeats - only for chars < 256 */ \ - 2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \ - 4, 4, 4, /* NOT upto, minupto, exact */ \ - 2, 2, 2, 4, /* Possessive *, +, ?, upto */ \ - /* Positive type repeats */ \ - 2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \ - 4, 4, 4, /* Type upto, minupto, exact */ \ - 2, 2, 2, 4, /* Possessive *+, ++, ?+, upto+ */ \ - /* Character class & ref repeats */ \ - 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \ - 5, 5, /* CRRANGE, CRMINRANGE */ \ - 33, /* CLASS */ \ - 33, /* NCLASS */ \ - 0, /* XCLASS - variable length */ \ - 3, /* REF */ \ - 1+LINK_SIZE, /* RECURSE */ \ - 2+2*LINK_SIZE, /* CALLOUT */ \ - 1+LINK_SIZE, /* Alt */ \ - 1+LINK_SIZE, /* Ket */ \ - 1+LINK_SIZE, /* KetRmax */ \ - 1+LINK_SIZE, /* KetRmin */ \ - 1+LINK_SIZE, /* Assert */ \ - 1+LINK_SIZE, /* Assert not */ \ - 1+LINK_SIZE, /* Assert behind */ \ - 1+LINK_SIZE, /* Assert behind not */ \ - 1+LINK_SIZE, /* Reverse */ \ - 1+LINK_SIZE, /* ONCE */ \ - 1+LINK_SIZE, /* BRA */ \ - 3+LINK_SIZE, /* CBRA */ \ - 1+LINK_SIZE, /* COND */ \ - 1+LINK_SIZE, /* SBRA */ \ - 3+LINK_SIZE, /* SCBRA */ \ - 1+LINK_SIZE, /* SCOND */ \ - 3, /* CREF */ \ - 3, /* RREF */ \ - 1, /* DEF */ \ - 1, 1, /* BRAZERO, BRAMINZERO */ \ - 1, 1, 1, 1, /* PRUNE, SKIP, THEN, COMMIT, */ \ - 1, 1, 1 /* FAIL, ACCEPT, SKIPZERO */ - - -/* A magic value for OP_RREF to indicate the "any recursion" condition. */ - -#define RREF_ANY 0xffff - -/* Error code numbers. They are given names so that they can more easily be -tracked. */ - -enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9, - ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19, - ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29, - ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39, - ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49, - ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59, - ERR60, ERR61, ERR62, ERR63, ERR64 }; - -/* The real format of the start of the pcre block; the index of names and the -code vector run on as long as necessary after the end. We store an explicit -offset to the name table so that if a regex is compiled on one host, saved, and -then run on another where the size of pointers is different, all might still -be well. For the case of compiled-on-4 and run-on-8, we include an extra -pointer that is always NULL. For future-proofing, a few dummy fields were -originally included - even though you can never get this planning right - but -there is only one left now. - -NOTE NOTE NOTE: -Because people can now save and re-use compiled patterns, any additions to this -structure should be made at the end, and something earlier (e.g. a new -flag in the options or one of the dummy fields) should indicate that the new -fields are present. Currently PCRE always sets the dummy fields to zero. -NOTE NOTE NOTE: -*/ - -typedef struct real_pcre { - pcre_uint32 magic_number; - pcre_uint32 size; /* Total that was malloced */ - pcre_uint32 options; /* Public options */ - pcre_uint16 flags; /* Private flags */ - pcre_uint16 dummy1; /* For future use */ - pcre_uint16 top_bracket; - pcre_uint16 top_backref; - pcre_uint16 first_byte; - pcre_uint16 req_byte; - pcre_uint16 name_table_offset; /* Offset to name table that follows */ - pcre_uint16 name_entry_size; /* Size of any name items */ - pcre_uint16 name_count; /* Number of name items */ - pcre_uint16 ref_count; /* Reference count */ - - const unsigned char *tables; /* Pointer to tables or NULL for std */ - const unsigned char *nullpad; /* NULL padding */ -} real_pcre; - -/* The format of the block used to store data from pcre_study(). The same -remark (see NOTE above) about extending this structure applies. */ - -typedef struct pcre_study_data { - pcre_uint32 size; /* Total that was malloced */ - pcre_uint32 options; - uschar start_bits[32]; -} pcre_study_data; - -/* Structure for passing "static" information around between the functions -doing the compiling, so that they are thread-safe. */ - -typedef struct compile_data { - const uschar *lcc; /* Points to lower casing table */ - const uschar *fcc; /* Points to case-flipping table */ - const uschar *cbits; /* Points to character type table */ - const uschar *ctypes; /* Points to table of type maps */ - const uschar *start_workspace;/* The start of working space */ - const uschar *start_code; /* The start of the compiled code */ - const uschar *start_pattern; /* The start of the pattern */ - const uschar *end_pattern; /* The end of the pattern */ - uschar *hwm; /* High watermark of workspace */ - uschar *name_table; /* The name/number table */ - int names_found; /* Number of entries so far */ - int name_entry_size; /* Size of each entry */ - int bracount; /* Count of capturing parens as we compile */ - int final_bracount; /* Saved value after first pass */ - int top_backref; /* Maximum back reference */ - unsigned int backref_map; /* Bitmap of low back refs */ - int external_options; /* External (initial) options */ - int external_flags; /* External flag bits to be set */ - int req_varyopt; /* "After variable item" flag for reqbyte */ - BOOL had_accept; /* (*ACCEPT) encountered */ - int nltype; /* Newline type */ - int nllen; /* Newline string length */ - uschar nl[4]; /* Newline string when fixed length */ -} compile_data; - -/* Structure for maintaining a chain of pointers to the currently incomplete -branches, for testing for left recursion. */ - -typedef struct branch_chain { - struct branch_chain *outer; - uschar *current; -} branch_chain; - -/* Structure for items in a linked list that represents an explicit recursive -call within the pattern. */ - -typedef struct recursion_info { - struct recursion_info *prevrec; /* Previous recursion record (or NULL) */ - int group_num; /* Number of group that was called */ - const uschar *after_call; /* "Return value": points after the call in the expr */ - USPTR save_start; /* Old value of mstart */ - int *offset_save; /* Pointer to start of saved offsets */ - int saved_max; /* Number of saved offsets */ -} recursion_info; - -/* Structure for building a chain of data for holding the values of the subject -pointer at the start of each subpattern, so as to detect when an empty string -has been matched by a subpattern - to break infinite loops. */ - -typedef struct eptrblock { - struct eptrblock *epb_prev; - USPTR epb_saved_eptr; -} eptrblock; - - -/* Structure for passing "static" information around between the functions -doing traditional NFA matching, so that they are thread-safe. */ - -typedef struct match_data { - unsigned long int match_call_count; /* As it says */ - unsigned long int match_limit; /* As it says */ - unsigned long int match_limit_recursion; /* As it says */ - int *offset_vector; /* Offset vector */ - int offset_end; /* One past the end */ - int offset_max; /* The maximum usable for return data */ - int nltype; /* Newline type */ - int nllen; /* Newline string length */ - uschar nl[4]; /* Newline string when fixed */ - const uschar *lcc; /* Points to lower casing table */ - const uschar *ctypes; /* Points to table of type maps */ - BOOL offset_overflow; /* Set if too many extractions */ - BOOL notbol; /* NOTBOL flag */ - BOOL noteol; /* NOTEOL flag */ - BOOL utf8; /* UTF8 flag */ - BOOL jscript_compat; /* JAVASCRIPT_COMPAT flag */ - BOOL endonly; /* Dollar not before final \n */ - BOOL notempty; /* Empty string match not wanted */ - BOOL partial; /* PARTIAL flag */ - BOOL hitend; /* Hit the end of the subject at some point */ - BOOL bsr_anycrlf; /* \R is just any CRLF, not full Unicode */ - const uschar *start_code; /* For use when recursing */ - USPTR start_subject; /* Start of the subject string */ - USPTR end_subject; /* End of the subject string */ - USPTR start_match_ptr; /* Start of matched string */ - USPTR end_match_ptr; /* Subject position at end match */ - int end_offset_top; /* Highwater mark at end of match */ - int capture_last; /* Most recent capture number */ - int start_offset; /* The start offset value */ - eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */ - int eptrn; /* Next free eptrblock */ - recursion_info *recursive; /* Linked list of recursion data */ - void *callout_data; /* To pass back to callouts */ -} match_data; - -/* A similar structure is used for the same purpose by the DFA matching -functions. */ - -typedef struct dfa_match_data { - const uschar *start_code; /* Start of the compiled pattern */ - const uschar *start_subject; /* Start of the subject string */ - const uschar *end_subject; /* End of subject string */ - const uschar *tables; /* Character tables */ - int moptions; /* Match options */ - int poptions; /* Pattern options */ - int nltype; /* Newline type */ - int nllen; /* Newline string length */ - uschar nl[4]; /* Newline string when fixed */ - void *callout_data; /* To pass back to callouts */ -} dfa_match_data; - -/* Bit definitions for entries in the pcre_ctypes table. */ - -#define ctype_space 0x01 -#define ctype_letter 0x02 -#define ctype_digit 0x04 -#define ctype_xdigit 0x08 -#define ctype_word 0x10 /* alphanumeric or '_' */ -#define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */ - -/* Offsets for the bitmap tables in pcre_cbits. Each table contains a set -of bits for a class map. Some classes are built by combining these tables. */ - -#define cbit_space 0 /* [:space:] or \s */ -#define cbit_xdigit 32 /* [:xdigit:] */ -#define cbit_digit 64 /* [:digit:] or \d */ -#define cbit_upper 96 /* [:upper:] */ -#define cbit_lower 128 /* [:lower:] */ -#define cbit_word 160 /* [:word:] or \w */ -#define cbit_graph 192 /* [:graph:] */ -#define cbit_print 224 /* [:print:] */ -#define cbit_punct 256 /* [:punct:] */ -#define cbit_cntrl 288 /* [:cntrl:] */ -#define cbit_length 320 /* Length of the cbits table */ - -/* Offsets of the various tables from the base tables pointer, and -total length. */ - -#define lcc_offset 0 -#define fcc_offset 256 -#define cbits_offset 512 -#define ctypes_offset (cbits_offset + cbit_length) -#define tables_length (ctypes_offset + 256) - -/* Layout of the UCP type table that translates property names into types and -codes. Each entry used to point directly to a name, but to reduce the number of -relocations in shared libraries, it now has an offset into a single string -instead. */ - -typedef struct { - pcre_uint16 name_offset; - pcre_uint16 type; - pcre_uint16 value; -} ucp_type_table; - - -/* Internal shared data tables. These are tables that are used by more than one -of the exported public functions. They have to be "external" in the C sense, -but are not part of the PCRE public API. The data for these tables is in the -pcre_tables.c module. */ - -extern const int _pcre_utf8_table1[]; -extern const int _pcre_utf8_table2[]; -extern const int _pcre_utf8_table3[]; -extern const uschar _pcre_utf8_table4[]; - -extern const int _pcre_utf8_table1_size; - -extern const char _pcre_utt_names[]; -extern const ucp_type_table _pcre_utt[]; -extern const int _pcre_utt_size; - -extern const uschar _pcre_default_tables[]; - -extern const uschar _pcre_OP_lengths[]; - - -/* Internal shared functions. These are functions that are used by more than -one of the exported public functions. They have to be "external" in the C -sense, but are not part of the PCRE public API. */ - -extern BOOL _pcre_is_newline(const uschar *, int, const uschar *, - int *, BOOL); -extern int _pcre_ord2utf8(int, uschar *); -extern real_pcre *_pcre_try_flipped(const real_pcre *, real_pcre *, - const pcre_study_data *, pcre_study_data *); -extern int _pcre_ucp_findprop(const unsigned int, int *, int *); -extern unsigned int _pcre_ucp_othercase(const unsigned int); -extern int _pcre_valid_utf8(const uschar *, int); -extern BOOL _pcre_was_newline(const uschar *, int, const uschar *, - int *, BOOL); -extern BOOL _pcre_xclass(int, const uschar *); + Copyright (c) 1997-2012 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + +/* This header contains definitions that are shared between the different +modules, but which are not relevant to the exported API. This includes some +functions whose names all begin with "_pcre_" or "_pcre16_" depending on +the PRIV macro. */ + +#ifndef PCRE_INTERNAL_H +#define PCRE_INTERNAL_H + +/* Define PCRE_DEBUG to get debugging output on stdout. */ + +#if 0 +#define PCRE_DEBUG +#endif + +/* PCRE is compiled as an 8 bit library if it is not requested otherwise. */ +#ifndef COMPILE_PCRE16 +#define COMPILE_PCRE8 +#endif + +/* If SUPPORT_UCP is defined, SUPPORT_UTF must also be defined. The +"configure" script ensures this, but not everybody uses "configure". */ + +#if defined SUPPORT_UCP && !(defined SUPPORT_UTF) +#define SUPPORT_UTF 1 +#endif + +/* We define SUPPORT_UTF if SUPPORT_UTF8 is enabled for compatibility +reasons with existing code. */ + +#if defined SUPPORT_UTF8 && !(defined SUPPORT_UTF) +#define SUPPORT_UTF 1 +#endif + +/* Fixme: SUPPORT_UTF8 should be eventually disappear from the code. +Until then we define it if SUPPORT_UTF is defined. */ + +#if defined SUPPORT_UTF && !(defined SUPPORT_UTF8) +#define SUPPORT_UTF8 1 +#endif + +/* We do not support both EBCDIC and UTF-8/16 at the same time. The "configure" +script prevents both being selected, but not everybody uses "configure". */ + +#if defined EBCDIC && defined SUPPORT_UTF +#error The use of both EBCDIC and SUPPORT_UTF8/16 is not supported. +#endif + +/* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef +inline, and there are *still* stupid compilers about that don't like indented +pre-processor statements, or at least there were when I first wrote this. After +all, it had only been about 10 years then... + +It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so +be absolutely sure we get our version. */ + +#undef DPRINTF +#ifdef PCRE_DEBUG +#define DPRINTF(p) printf p +#else +#define DPRINTF(p) /* Nothing */ +#endif + + +/* Standard C headers plus the external interface definition. The only time +setjmp and stdarg are used is when NO_RECURSE is set. */ + +#include +#include +#include +#include +#include +#include + +/* When compiling a DLL for Windows, the exported symbols have to be declared +using some MS magic. I found some useful information on this web page: +http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the +information there, using __declspec(dllexport) without "extern" we have a +definition; with "extern" we have a declaration. The settings here override the +setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL, +which is all that is needed for applications (they just import the symbols). We +use: + + PCRE_EXP_DECL for declarations + PCRE_EXP_DEFN for definitions of exported functions + PCRE_EXP_DATA_DEFN for definitions of exported variables + +The reason for the two DEFN macros is that in non-Windows environments, one +does not want to have "extern" before variable definitions because it leads to +compiler warnings. So we distinguish between functions and variables. In +Windows, the two should always be the same. + +The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest, +which is an application, but needs to import this file in order to "peek" at +internals, can #include pcre.h first to get an application's-eye view. + +In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon, +special-purpose environments) might want to stick other stuff in front of +exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and +PCRE_EXP_DATA_DEFN only if they are not already set. */ + +#ifndef PCRE_EXP_DECL +# ifdef _WIN32 +# ifndef PCRE_STATIC +# define PCRE_EXP_DECL extern __declspec(dllexport) +# define PCRE_EXP_DEFN __declspec(dllexport) +# define PCRE_EXP_DATA_DEFN __declspec(dllexport) +# else +# define PCRE_EXP_DECL extern +# define PCRE_EXP_DEFN +# define PCRE_EXP_DATA_DEFN +# endif +# else +# ifdef __cplusplus +# define PCRE_EXP_DECL extern "C" +# else +# define PCRE_EXP_DECL extern +# endif +# ifndef PCRE_EXP_DEFN +# define PCRE_EXP_DEFN PCRE_EXP_DECL +# endif +# ifndef PCRE_EXP_DATA_DEFN +# define PCRE_EXP_DATA_DEFN +# endif +# endif +#endif + +/* When compiling with the MSVC compiler, it is sometimes necessary to include +a "calling convention" before exported function names. (This is secondhand +information; I know nothing about MSVC myself). For example, something like + + void __cdecl function(....) + +might be needed. In order so make this easy, all the exported functions have +PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not +set, we ensure here that it has no effect. */ + +#ifndef PCRE_CALL_CONVENTION +#define PCRE_CALL_CONVENTION +#endif + +/* We need to have types that specify unsigned 8, 16 and 32-bit integers. We +cannot determine these outside the compilation (e.g. by running a program as +part of "configure") because PCRE is often cross-compiled for use on other +systems. Instead we make use of the maximum sizes that are available at +preprocessor time in standard C environments. */ + +typedef unsigned char pcre_uint8; + +#if USHRT_MAX == 65535 + typedef unsigned short pcre_uint16; + typedef short pcre_int16; +#elif UINT_MAX == 65535 + typedef unsigned int pcre_uint16; + typedef int pcre_int16; +#else + #error Cannot determine a type for 16-bit unsigned integers +#endif + +#if UINT_MAX == 4294967295 + typedef unsigned int pcre_uint32; + typedef int pcre_int32; +#elif ULONG_MAX == 4294967295 + typedef unsigned long int pcre_uint32; + typedef long int pcre_int32; +#else + #error Cannot determine a type for 32-bit unsigned integers +#endif + +/* When checking for integer overflow in pcre_compile(), we need to handle +large integers. If a 64-bit integer type is available, we can use that. +Otherwise we have to cast to double, which of course requires floating point +arithmetic. Handle this by defining a macro for the appropriate type. If +stdint.h is available, include it; it may define INT64_MAX. Systems that do not +have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set +by "configure". */ + +#if HAVE_STDINT_H +#include +#elif HAVE_INTTYPES_H +#include +#endif + +#if defined INT64_MAX || defined int64_t +#define INT64_OR_DOUBLE int64_t +#else +#define INT64_OR_DOUBLE double +#endif + +/* All character handling must be done as unsigned characters. Otherwise there +are problems with top-bit-set characters and functions such as isspace(). +However, we leave the interface to the outside world as char * or short *, +because that should make things easier for callers. This character type is +called pcre_uchar. + +The IN_UCHARS macro multiply its argument with the byte size of the current +pcre_uchar type. Useful for memcpy and such operations, whose require the +byte size of their input/output buffers. + +The MAX_255 macro checks whether its pcre_uchar input is less than 256. + +The TABLE_GET macro is designed for accessing elements of tables whose contain +exactly 256 items. When the character is able to contain more than 256 +items, some check is needed before accessing these tables. +*/ + +#ifdef COMPILE_PCRE8 + +typedef unsigned char pcre_uchar; +#define IN_UCHARS(x) (x) +#define MAX_255(c) 1 +#define TABLE_GET(c, table, default) ((table)[c]) + +#else + +#ifdef COMPILE_PCRE16 +#if USHRT_MAX != 65535 +/* This is a warning message. Change PCRE_UCHAR16 to a 16 bit data type in +pcre.h(.in) and disable (comment out) this message. */ +#error Warning: PCRE_UCHAR16 is not a 16 bit data type. +#endif + +typedef pcre_uint16 pcre_uchar; +#define IN_UCHARS(x) ((x) << 1) +#define MAX_255(c) ((c) <= 255u) +#define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default)) + +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE16 */ + +#endif /* COMPILE_PCRE8 */ + +/* This is an unsigned int value that no character can ever have. UTF-8 +characters only go up to 0x7fffffff (though Unicode doesn't go beyond +0x0010ffff). */ + +#define NOTACHAR 0xffffffff + +/* PCRE is able to support several different kinds of newline (CR, LF, CRLF, +"any" and "anycrlf" at present). The following macros are used to package up +testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various +modules to indicate in which datablock the parameters exist, and what the +start/end of string field names are. */ + +#define NLTYPE_FIXED 0 /* Newline is a fixed length string */ +#define NLTYPE_ANY 1 /* Newline is any Unicode line ending */ +#define NLTYPE_ANYCRLF 2 /* Newline is CR, LF, or CRLF */ + +/* This macro checks for a newline at the given position */ + +#define IS_NEWLINE(p) \ + ((NLBLOCK->nltype != NLTYPE_FIXED)? \ + ((p) < NLBLOCK->PSEND && \ + PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \ + &(NLBLOCK->nllen), utf)) \ + : \ + ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \ + (p)[0] == NLBLOCK->nl[0] && \ + (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \ + ) \ + ) + +/* This macro checks for a newline immediately preceding the given position */ + +#define WAS_NEWLINE(p) \ + ((NLBLOCK->nltype != NLTYPE_FIXED)? \ + ((p) > NLBLOCK->PSSTART && \ + PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \ + &(NLBLOCK->nllen), utf)) \ + : \ + ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \ + (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \ + (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \ + ) \ + ) + +/* When PCRE is compiled as a C++ library, the subject pointer can be replaced +with a custom type. This makes it possible, for example, to allow pcre_exec() +to process subject strings that are discontinuous by using a smart pointer +class. It must always be possible to inspect all of the subject string in +pcre_exec() because of the way it backtracks. Two macros are required in the +normal case, for sign-unspecified and unsigned char pointers. The former is +used for the external interface and appears in pcre.h, which is why its name +must begin with PCRE_. */ + +#ifdef CUSTOM_SUBJECT_PTR +#define PCRE_PUCHAR CUSTOM_SUBJECT_PTR +#else +#define PCRE_PUCHAR const pcre_uchar * +#endif + +/* Include the public PCRE header and the definitions of UCP character property +values. */ + +#include "pcre.h" +#include "ucp.h" + +/* When compiling for use with the Virtual Pascal compiler, these functions +need to have their names changed. PCRE must be compiled with the -DVPCOMPAT +option on the command line. */ + +#ifdef VPCOMPAT +#define strlen(s) _strlen(s) +#define strncmp(s1,s2,m) _strncmp(s1,s2,m) +#define memcmp(s,c,n) _memcmp(s,c,n) +#define memcpy(d,s,n) _memcpy(d,s,n) +#define memmove(d,s,n) _memmove(d,s,n) +#define memset(s,c,n) _memset(s,c,n) +#else /* VPCOMPAT */ + +/* To cope with SunOS4 and other systems that lack memmove() but have bcopy(), +define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY +is set. Otherwise, include an emulating function for those systems that have +neither (there some non-Unix environments where this is the case). */ + +#ifndef HAVE_MEMMOVE +#undef memmove /* some systems may have a macro */ +#ifdef HAVE_BCOPY +#define memmove(a, b, c) bcopy(b, a, c) +#else /* HAVE_BCOPY */ +static void * +pcre_memmove(void *d, const void *s, size_t n) +{ +size_t i; +unsigned char *dest = (unsigned char *)d; +const unsigned char *src = (const unsigned char *)s; +if (dest > src) + { + dest += n; + src += n; + for (i = 0; i < n; ++i) *(--dest) = *(--src); + return (void *)dest; + } +else + { + for (i = 0; i < n; ++i) *dest++ = *src++; + return (void *)(dest - n); + } +} +#define memmove(a, b, c) pcre_memmove(a, b, c) +#endif /* not HAVE_BCOPY */ +#endif /* not HAVE_MEMMOVE */ +#endif /* not VPCOMPAT */ + + +/* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored +in big-endian order) by default. These are used, for example, to link from the +start of a subpattern to its alternatives and its end. The use of 2 bytes per +offset limits the size of the compiled regex to around 64K, which is big enough +for almost everybody. However, I received a request for an even bigger limit. +For this reason, and also to make the code easier to maintain, the storing and +loading of offsets from the byte string is now handled by the macros that are +defined here. + +The macros are controlled by the value of LINK_SIZE. This defaults to 2 in +the config.h file, but can be overridden by using -D on the command line. This +is automated on Unix systems via the "configure" command. */ + +#ifdef COMPILE_PCRE8 + +#if LINK_SIZE == 2 + +#define PUT(a,n,d) \ + (a[n] = (d) >> 8), \ + (a[(n)+1] = (d) & 255) + +#define GET(a,n) \ + (((a)[n] << 8) | (a)[(n)+1]) + +#define MAX_PATTERN_SIZE (1 << 16) + + +#elif LINK_SIZE == 3 + +#define PUT(a,n,d) \ + (a[n] = (d) >> 16), \ + (a[(n)+1] = (d) >> 8), \ + (a[(n)+2] = (d) & 255) + +#define GET(a,n) \ + (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2]) + +#define MAX_PATTERN_SIZE (1 << 24) + + +#elif LINK_SIZE == 4 + +#define PUT(a,n,d) \ + (a[n] = (d) >> 24), \ + (a[(n)+1] = (d) >> 16), \ + (a[(n)+2] = (d) >> 8), \ + (a[(n)+3] = (d) & 255) + +#define GET(a,n) \ + (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3]) + +/* Keep it positive */ +#define MAX_PATTERN_SIZE (1 << 30) + +#else +#error LINK_SIZE must be either 2, 3, or 4 +#endif + +#else /* COMPILE_PCRE8 */ + +#ifdef COMPILE_PCRE16 + +#if LINK_SIZE == 2 + +#undef LINK_SIZE +#define LINK_SIZE 1 + +#define PUT(a,n,d) \ + (a[n] = (d)) + +#define GET(a,n) \ + (a[n]) + +#define MAX_PATTERN_SIZE (1 << 16) + +#elif LINK_SIZE == 3 || LINK_SIZE == 4 + +#undef LINK_SIZE +#define LINK_SIZE 2 + +#define PUT(a,n,d) \ + (a[n] = (d) >> 16), \ + (a[(n)+1] = (d) & 65535) + +#define GET(a,n) \ + (((a)[n] << 16) | (a)[(n)+1]) + +/* Keep it positive */ +#define MAX_PATTERN_SIZE (1 << 30) + +#else +#error LINK_SIZE must be either 2, 3, or 4 +#endif + +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE16 */ + +#endif /* COMPILE_PCRE8 */ + +/* Convenience macro defined in terms of the others */ + +#define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE + + +/* PCRE uses some other 2-byte quantities that do not change when the size of +offsets changes. There are used for repeat counts and for other things such as +capturing parenthesis numbers in back references. */ + +#ifdef COMPILE_PCRE8 + +#define IMM2_SIZE 2 + +#define PUT2(a,n,d) \ + a[n] = (d) >> 8; \ + a[(n)+1] = (d) & 255 + +#define GET2(a,n) \ + (((a)[n] << 8) | (a)[(n)+1]) + +#else /* COMPILE_PCRE8 */ + +#ifdef COMPILE_PCRE16 + +#define IMM2_SIZE 1 + +#define PUT2(a,n,d) \ + a[n] = d + +#define GET2(a,n) \ + a[n] + +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE16 */ + +#endif /* COMPILE_PCRE8 */ + +#define PUT2INC(a,n,d) PUT2(a,n,d), a += IMM2_SIZE + +/* The maximum length of a MARK name is currently one data unit; it may be +changed in future to be a fixed number of bytes or to depend on LINK_SIZE. */ + +#define MAX_MARK ((1 << (sizeof(pcre_uchar)*8)) - 1) + +/* When UTF encoding is being used, a character is no longer just a single +character. The macros for character handling generate simple sequences when +used in character-mode, and more complicated ones for UTF characters. +GETCHARLENTEST and other macros are not used when UTF is not supported, +so they are not defined. To make sure they can never even appear when +UTF support is omitted, we don't even define them. */ + +#ifndef SUPPORT_UTF + +/* #define MAX_VALUE_FOR_SINGLE_CHAR */ +/* #define HAS_EXTRALEN(c) */ +/* #define GET_EXTRALEN(c) */ +/* #define NOT_FIRSTCHAR(c) */ +#define GETCHAR(c, eptr) c = *eptr; +#define GETCHARTEST(c, eptr) c = *eptr; +#define GETCHARINC(c, eptr) c = *eptr++; +#define GETCHARINCTEST(c, eptr) c = *eptr++; +#define GETCHARLEN(c, eptr, len) c = *eptr; +/* #define GETCHARLENTEST(c, eptr, len) */ +/* #define BACKCHAR(eptr) */ +/* #define FORWARDCHAR(eptr) */ +/* #define ACROSSCHAR(condition, eptr, action) */ + +#else /* SUPPORT_UTF */ + +#ifdef COMPILE_PCRE8 + +/* These macros were originally written in the form of loops that used data +from the tables whose names start with PRIV(utf8_table). They were rewritten by +a user so as not to use loops, because in some environments this gives a +significant performance advantage, and it seems never to do any harm. */ + +/* Tells the biggest code point which can be encoded as a single character. */ + +#define MAX_VALUE_FOR_SINGLE_CHAR 127 + +/* Tests whether the code point needs extra characters to decode. */ + +#define HAS_EXTRALEN(c) ((c) >= 0xc0) + +/* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE. +Otherwise it has an undefined behaviour. */ + +#define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f]) + +/* Returns TRUE, if the given character is not the first character +of a UTF sequence. */ + +#define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80) + +/* Base macro to pick up the remaining bytes of a UTF-8 character, not +advancing the pointer. */ + +#define GETUTF8(c, eptr) \ + { \ + if ((c & 0x20) == 0) \ + c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \ + else if ((c & 0x10) == 0) \ + c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \ + else if ((c & 0x08) == 0) \ + c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \ + ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \ + else if ((c & 0x04) == 0) \ + c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \ + ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \ + (eptr[4] & 0x3f); \ + else \ + c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \ + ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \ + ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \ + } + +/* Get the next UTF-8 character, not advancing the pointer. This is called when +we know we are in UTF-8 mode. */ + +#define GETCHAR(c, eptr) \ + c = *eptr; \ + if (c >= 0xc0) GETUTF8(c, eptr); + +/* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the +pointer. */ + +#define GETCHARTEST(c, eptr) \ + c = *eptr; \ + if (utf && c >= 0xc0) GETUTF8(c, eptr); + +/* Base macro to pick up the remaining bytes of a UTF-8 character, advancing +the pointer. */ + +#define GETUTF8INC(c, eptr) \ + { \ + if ((c & 0x20) == 0) \ + c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \ + else if ((c & 0x10) == 0) \ + { \ + c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \ + eptr += 2; \ + } \ + else if ((c & 0x08) == 0) \ + { \ + c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \ + ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \ + eptr += 3; \ + } \ + else if ((c & 0x04) == 0) \ + { \ + c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \ + ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \ + (eptr[3] & 0x3f); \ + eptr += 4; \ + } \ + else \ + { \ + c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \ + ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \ + ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \ + eptr += 5; \ + } \ + } + +/* Get the next UTF-8 character, advancing the pointer. This is called when we +know we are in UTF-8 mode. */ + +#define GETCHARINC(c, eptr) \ + c = *eptr++; \ + if (c >= 0xc0) GETUTF8INC(c, eptr); + +/* Get the next character, testing for UTF-8 mode, and advancing the pointer. +This is called when we don't know if we are in UTF-8 mode. */ + +#define GETCHARINCTEST(c, eptr) \ + c = *eptr++; \ + if (utf && c >= 0xc0) GETUTF8INC(c, eptr); + +/* Base macro to pick up the remaining bytes of a UTF-8 character, not +advancing the pointer, incrementing the length. */ + +#define GETUTF8LEN(c, eptr, len) \ + { \ + if ((c & 0x20) == 0) \ + { \ + c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \ + len++; \ + } \ + else if ((c & 0x10) == 0) \ + { \ + c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \ + len += 2; \ + } \ + else if ((c & 0x08) == 0) \ + {\ + c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \ + ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \ + len += 3; \ + } \ + else if ((c & 0x04) == 0) \ + { \ + c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \ + ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \ + (eptr[4] & 0x3f); \ + len += 4; \ + } \ + else \ + {\ + c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \ + ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \ + ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \ + len += 5; \ + } \ + } + +/* Get the next UTF-8 character, not advancing the pointer, incrementing length +if there are extra bytes. This is called when we know we are in UTF-8 mode. */ + +#define GETCHARLEN(c, eptr, len) \ + c = *eptr; \ + if (c >= 0xc0) GETUTF8LEN(c, eptr, len); + +/* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the +pointer, incrementing length if there are extra bytes. This is called when we +do not know if we are in UTF-8 mode. */ + +#define GETCHARLENTEST(c, eptr, len) \ + c = *eptr; \ + if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len); + +/* If the pointer is not at the start of a character, move it back until +it is. This is called only in UTF-8 mode - we don't put a test within the macro +because almost all calls are already within a block of UTF-8 only code. */ + +#define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr-- + +/* Same as above, just in the other direction. */ +#define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++ + +/* Same as above, but it allows a fully customizable form. */ +#define ACROSSCHAR(condition, eptr, action) \ + while((condition) && ((eptr) & 0xc0) == 0x80) action + +#else /* COMPILE_PCRE8 */ + +#ifdef COMPILE_PCRE16 + +/* Tells the biggest code point which can be encoded as a single character. */ + +#define MAX_VALUE_FOR_SINGLE_CHAR 65535 + +/* Tests whether the code point needs extra characters to decode. */ + +#define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800) + +/* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE. +Otherwise it has an undefined behaviour. */ + +#define GET_EXTRALEN(c) 1 + +/* Returns TRUE, if the given character is not the first character +of a UTF sequence. */ + +#define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00) + +/* Base macro to pick up the low surrogate of a UTF-16 character, not +advancing the pointer. */ + +#define GETUTF16(c, eptr) \ + { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; } + +/* Get the next UTF-16 character, not advancing the pointer. This is called when +we know we are in UTF-16 mode. */ + +#define GETCHAR(c, eptr) \ + c = *eptr; \ + if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr); + +/* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the +pointer. */ + +#define GETCHARTEST(c, eptr) \ + c = *eptr; \ + if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr); + +/* Base macro to pick up the low surrogate of a UTF-16 character, advancing +the pointer. */ + +#define GETUTF16INC(c, eptr) \ + { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; } + +/* Get the next UTF-16 character, advancing the pointer. This is called when we +know we are in UTF-16 mode. */ + +#define GETCHARINC(c, eptr) \ + c = *eptr++; \ + if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr); + +/* Get the next character, testing for UTF-16 mode, and advancing the pointer. +This is called when we don't know if we are in UTF-16 mode. */ + +#define GETCHARINCTEST(c, eptr) \ + c = *eptr++; \ + if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr); + +/* Base macro to pick up the low surrogate of a UTF-16 character, not +advancing the pointer, incrementing the length. */ + +#define GETUTF16LEN(c, eptr, len) \ + { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; } + +/* Get the next UTF-16 character, not advancing the pointer, incrementing +length if there is a low surrogate. This is called when we know we are in +UTF-16 mode. */ + +#define GETCHARLEN(c, eptr, len) \ + c = *eptr; \ + if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len); + +/* Get the next UTF-816character, testing for UTF-16 mode, not advancing the +pointer, incrementing length if there is a low surrogate. This is called when +we do not know if we are in UTF-16 mode. */ + +#define GETCHARLENTEST(c, eptr, len) \ + c = *eptr; \ + if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len); + +/* If the pointer is not at the start of a character, move it back until +it is. This is called only in UTF-16 mode - we don't put a test within the +macro because almost all calls are already within a block of UTF-16 only +code. */ + +#define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr-- + +/* Same as above, just in the other direction. */ +#define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++ + +/* Same as above, but it allows a fully customizable form. */ +#define ACROSSCHAR(condition, eptr, action) \ + if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action + +#endif + +#endif /* COMPILE_PCRE8 */ + +#endif /* SUPPORT_UTF */ + + +/* In case there is no definition of offsetof() provided - though any proper +Standard C system should have one. */ + +#ifndef offsetof +#define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field)) +#endif + + +/* Private flags containing information about the compiled regex. They used to +live at the top end of the options word, but that got almost full, so now they +are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as +the restrictions on partial matching have been lifted. It remains for backwards +compatibility. */ + +#ifdef COMPILE_PCRE8 +#define PCRE_MODE 0x0001 /* compiled in 8 bit mode */ +#endif +#ifdef COMPILE_PCRE16 +#define PCRE_MODE 0x0002 /* compiled in 16 bit mode */ +#endif +#define PCRE_FIRSTSET 0x0010 /* first_char is set */ +#define PCRE_FCH_CASELESS 0x0020 /* caseless first char */ +#define PCRE_REQCHSET 0x0040 /* req_byte is set */ +#define PCRE_RCH_CASELESS 0x0080 /* caseless requested char */ +#define PCRE_STARTLINE 0x0100 /* start after \n for multiline */ +#define PCRE_NOPARTIAL 0x0200 /* can't use partial with this regex */ +#define PCRE_JCHANGED 0x0400 /* j option used in regex */ +#define PCRE_HASCRORLF 0x0800 /* explicit \r or \n in pattern */ +#define PCRE_HASTHEN 0x1000 /* pattern contains (*THEN) */ + +/* Flags for the "extra" block produced by pcre_study(). */ + +#define PCRE_STUDY_MAPPED 0x0001 /* a map of starting chars exists */ +#define PCRE_STUDY_MINLEN 0x0002 /* a minimum length field exists */ + +/* Masks for identifying the public options that are permitted at compile +time, run time, or study time, respectively. */ + +#define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \ + PCRE_NEWLINE_ANYCRLF) + +#define PUBLIC_COMPILE_OPTIONS \ + (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \ + PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \ + PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \ + PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \ + PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE) + +#define PUBLIC_EXEC_OPTIONS \ + (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \ + PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \ + PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE) + +#define PUBLIC_DFA_EXEC_OPTIONS \ + (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \ + PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \ + PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \ + PCRE_NO_START_OPTIMIZE) + +#define PUBLIC_STUDY_OPTIONS \ + (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \ + PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE) + +/* Magic number to provide a small check against being handed junk. */ + +#define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */ + +/* This variable is used to detect a loaded regular expression +in different endianness. */ + +#define REVERSED_MAGIC_NUMBER 0x45524350UL /* 'ERCP' */ + +/* Negative values for the firstchar and reqchar variables */ + +#define REQ_UNSET (-2) +#define REQ_NONE (-1) + +/* The maximum remaining length of subject we are prepared to search for a +req_byte match. */ + +#define REQ_BYTE_MAX 1000 + +/* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in +environments where these macros are defined elsewhere. Unfortunately, there +is no way to do the same for the typedef. */ + +typedef int BOOL; + +#ifndef FALSE +#define FALSE 0 +#define TRUE 1 +#endif + +/* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal +character constants like '*' because the compiler would emit their EBCDIC code, +which is different from their ASCII/UTF-8 code. Instead we define macros for +the characters so that they always use the ASCII/UTF-8 code when UTF-8 support +is enabled. When UTF-8 support is not enabled, the definitions use character +literals. Both character and string versions of each character are needed, and +there are some longer strings as well. + +This means that, on EBCDIC platforms, the PCRE library can handle either +EBCDIC, or UTF-8, but not both. To support both in the same compiled library +would need different lookups depending on whether PCRE_UTF8 was set or not. +This would make it impossible to use characters in switch/case statements, +which would reduce performance. For a theoretical use (which nobody has asked +for) in a minority area (EBCDIC platforms), this is not sensible. Any +application that did need both could compile two versions of the library, using +macros to give the functions distinct names. */ + +#ifndef SUPPORT_UTF + +/* UTF-8 support is not enabled; use the platform-dependent character literals +so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */ + +#define CHAR_HT '\t' +#define CHAR_VT '\v' +#define CHAR_FF '\f' +#define CHAR_CR '\r' +#define CHAR_NL '\n' +#define CHAR_BS '\b' +#define CHAR_BEL '\a' +#ifdef EBCDIC +#define CHAR_ESC '\047' +#define CHAR_DEL '\007' +#else +#define CHAR_ESC '\033' +#define CHAR_DEL '\177' +#endif + +#define CHAR_SPACE ' ' +#define CHAR_EXCLAMATION_MARK '!' +#define CHAR_QUOTATION_MARK '"' +#define CHAR_NUMBER_SIGN '#' +#define CHAR_DOLLAR_SIGN '$' +#define CHAR_PERCENT_SIGN '%' +#define CHAR_AMPERSAND '&' +#define CHAR_APOSTROPHE '\'' +#define CHAR_LEFT_PARENTHESIS '(' +#define CHAR_RIGHT_PARENTHESIS ')' +#define CHAR_ASTERISK '*' +#define CHAR_PLUS '+' +#define CHAR_COMMA ',' +#define CHAR_MINUS '-' +#define CHAR_DOT '.' +#define CHAR_SLASH '/' +#define CHAR_0 '0' +#define CHAR_1 '1' +#define CHAR_2 '2' +#define CHAR_3 '3' +#define CHAR_4 '4' +#define CHAR_5 '5' +#define CHAR_6 '6' +#define CHAR_7 '7' +#define CHAR_8 '8' +#define CHAR_9 '9' +#define CHAR_COLON ':' +#define CHAR_SEMICOLON ';' +#define CHAR_LESS_THAN_SIGN '<' +#define CHAR_EQUALS_SIGN '=' +#define CHAR_GREATER_THAN_SIGN '>' +#define CHAR_QUESTION_MARK '?' +#define CHAR_COMMERCIAL_AT '@' +#define CHAR_A 'A' +#define CHAR_B 'B' +#define CHAR_C 'C' +#define CHAR_D 'D' +#define CHAR_E 'E' +#define CHAR_F 'F' +#define CHAR_G 'G' +#define CHAR_H 'H' +#define CHAR_I 'I' +#define CHAR_J 'J' +#define CHAR_K 'K' +#define CHAR_L 'L' +#define CHAR_M 'M' +#define CHAR_N 'N' +#define CHAR_O 'O' +#define CHAR_P 'P' +#define CHAR_Q 'Q' +#define CHAR_R 'R' +#define CHAR_S 'S' +#define CHAR_T 'T' +#define CHAR_U 'U' +#define CHAR_V 'V' +#define CHAR_W 'W' +#define CHAR_X 'X' +#define CHAR_Y 'Y' +#define CHAR_Z 'Z' +#define CHAR_LEFT_SQUARE_BRACKET '[' +#define CHAR_BACKSLASH '\\' +#define CHAR_RIGHT_SQUARE_BRACKET ']' +#define CHAR_CIRCUMFLEX_ACCENT '^' +#define CHAR_UNDERSCORE '_' +#define CHAR_GRAVE_ACCENT '`' +#define CHAR_a 'a' +#define CHAR_b 'b' +#define CHAR_c 'c' +#define CHAR_d 'd' +#define CHAR_e 'e' +#define CHAR_f 'f' +#define CHAR_g 'g' +#define CHAR_h 'h' +#define CHAR_i 'i' +#define CHAR_j 'j' +#define CHAR_k 'k' +#define CHAR_l 'l' +#define CHAR_m 'm' +#define CHAR_n 'n' +#define CHAR_o 'o' +#define CHAR_p 'p' +#define CHAR_q 'q' +#define CHAR_r 'r' +#define CHAR_s 's' +#define CHAR_t 't' +#define CHAR_u 'u' +#define CHAR_v 'v' +#define CHAR_w 'w' +#define CHAR_x 'x' +#define CHAR_y 'y' +#define CHAR_z 'z' +#define CHAR_LEFT_CURLY_BRACKET '{' +#define CHAR_VERTICAL_LINE '|' +#define CHAR_RIGHT_CURLY_BRACKET '}' +#define CHAR_TILDE '~' + +#define STR_HT "\t" +#define STR_VT "\v" +#define STR_FF "\f" +#define STR_CR "\r" +#define STR_NL "\n" +#define STR_BS "\b" +#define STR_BEL "\a" +#ifdef EBCDIC +#define STR_ESC "\047" +#define STR_DEL "\007" +#else +#define STR_ESC "\033" +#define STR_DEL "\177" +#endif + +#define STR_SPACE " " +#define STR_EXCLAMATION_MARK "!" +#define STR_QUOTATION_MARK "\"" +#define STR_NUMBER_SIGN "#" +#define STR_DOLLAR_SIGN "$" +#define STR_PERCENT_SIGN "%" +#define STR_AMPERSAND "&" +#define STR_APOSTROPHE "'" +#define STR_LEFT_PARENTHESIS "(" +#define STR_RIGHT_PARENTHESIS ")" +#define STR_ASTERISK "*" +#define STR_PLUS "+" +#define STR_COMMA "," +#define STR_MINUS "-" +#define STR_DOT "." +#define STR_SLASH "/" +#define STR_0 "0" +#define STR_1 "1" +#define STR_2 "2" +#define STR_3 "3" +#define STR_4 "4" +#define STR_5 "5" +#define STR_6 "6" +#define STR_7 "7" +#define STR_8 "8" +#define STR_9 "9" +#define STR_COLON ":" +#define STR_SEMICOLON ";" +#define STR_LESS_THAN_SIGN "<" +#define STR_EQUALS_SIGN "=" +#define STR_GREATER_THAN_SIGN ">" +#define STR_QUESTION_MARK "?" +#define STR_COMMERCIAL_AT "@" +#define STR_A "A" +#define STR_B "B" +#define STR_C "C" +#define STR_D "D" +#define STR_E "E" +#define STR_F "F" +#define STR_G "G" +#define STR_H "H" +#define STR_I "I" +#define STR_J "J" +#define STR_K "K" +#define STR_L "L" +#define STR_M "M" +#define STR_N "N" +#define STR_O "O" +#define STR_P "P" +#define STR_Q "Q" +#define STR_R "R" +#define STR_S "S" +#define STR_T "T" +#define STR_U "U" +#define STR_V "V" +#define STR_W "W" +#define STR_X "X" +#define STR_Y "Y" +#define STR_Z "Z" +#define STR_LEFT_SQUARE_BRACKET "[" +#define STR_BACKSLASH "\\" +#define STR_RIGHT_SQUARE_BRACKET "]" +#define STR_CIRCUMFLEX_ACCENT "^" +#define STR_UNDERSCORE "_" +#define STR_GRAVE_ACCENT "`" +#define STR_a "a" +#define STR_b "b" +#define STR_c "c" +#define STR_d "d" +#define STR_e "e" +#define STR_f "f" +#define STR_g "g" +#define STR_h "h" +#define STR_i "i" +#define STR_j "j" +#define STR_k "k" +#define STR_l "l" +#define STR_m "m" +#define STR_n "n" +#define STR_o "o" +#define STR_p "p" +#define STR_q "q" +#define STR_r "r" +#define STR_s "s" +#define STR_t "t" +#define STR_u "u" +#define STR_v "v" +#define STR_w "w" +#define STR_x "x" +#define STR_y "y" +#define STR_z "z" +#define STR_LEFT_CURLY_BRACKET "{" +#define STR_VERTICAL_LINE "|" +#define STR_RIGHT_CURLY_BRACKET "}" +#define STR_TILDE "~" + +#define STRING_ACCEPT0 "ACCEPT\0" +#define STRING_COMMIT0 "COMMIT\0" +#define STRING_F0 "F\0" +#define STRING_FAIL0 "FAIL\0" +#define STRING_MARK0 "MARK\0" +#define STRING_PRUNE0 "PRUNE\0" +#define STRING_SKIP0 "SKIP\0" +#define STRING_THEN "THEN" + +#define STRING_alpha0 "alpha\0" +#define STRING_lower0 "lower\0" +#define STRING_upper0 "upper\0" +#define STRING_alnum0 "alnum\0" +#define STRING_ascii0 "ascii\0" +#define STRING_blank0 "blank\0" +#define STRING_cntrl0 "cntrl\0" +#define STRING_digit0 "digit\0" +#define STRING_graph0 "graph\0" +#define STRING_print0 "print\0" +#define STRING_punct0 "punct\0" +#define STRING_space0 "space\0" +#define STRING_word0 "word\0" +#define STRING_xdigit "xdigit" + +#define STRING_DEFINE "DEFINE" + +#define STRING_CR_RIGHTPAR "CR)" +#define STRING_LF_RIGHTPAR "LF)" +#define STRING_CRLF_RIGHTPAR "CRLF)" +#define STRING_ANY_RIGHTPAR "ANY)" +#define STRING_ANYCRLF_RIGHTPAR "ANYCRLF)" +#define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)" +#define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)" +#ifdef COMPILE_PCRE8 +#define STRING_UTF_RIGHTPAR "UTF8)" +#endif +#ifdef COMPILE_PCRE16 +#define STRING_UTF_RIGHTPAR "UTF16)" +#endif +#define STRING_UCP_RIGHTPAR "UCP)" +#define STRING_NO_START_OPT_RIGHTPAR "NO_START_OPT)" + +#else /* SUPPORT_UTF */ + +/* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This +works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode +only. */ + +#define CHAR_HT '\011' +#define CHAR_VT '\013' +#define CHAR_FF '\014' +#define CHAR_CR '\015' +#define CHAR_NL '\012' +#define CHAR_BS '\010' +#define CHAR_BEL '\007' +#define CHAR_ESC '\033' +#define CHAR_DEL '\177' + +#define CHAR_SPACE '\040' +#define CHAR_EXCLAMATION_MARK '\041' +#define CHAR_QUOTATION_MARK '\042' +#define CHAR_NUMBER_SIGN '\043' +#define CHAR_DOLLAR_SIGN '\044' +#define CHAR_PERCENT_SIGN '\045' +#define CHAR_AMPERSAND '\046' +#define CHAR_APOSTROPHE '\047' +#define CHAR_LEFT_PARENTHESIS '\050' +#define CHAR_RIGHT_PARENTHESIS '\051' +#define CHAR_ASTERISK '\052' +#define CHAR_PLUS '\053' +#define CHAR_COMMA '\054' +#define CHAR_MINUS '\055' +#define CHAR_DOT '\056' +#define CHAR_SLASH '\057' +#define CHAR_0 '\060' +#define CHAR_1 '\061' +#define CHAR_2 '\062' +#define CHAR_3 '\063' +#define CHAR_4 '\064' +#define CHAR_5 '\065' +#define CHAR_6 '\066' +#define CHAR_7 '\067' +#define CHAR_8 '\070' +#define CHAR_9 '\071' +#define CHAR_COLON '\072' +#define CHAR_SEMICOLON '\073' +#define CHAR_LESS_THAN_SIGN '\074' +#define CHAR_EQUALS_SIGN '\075' +#define CHAR_GREATER_THAN_SIGN '\076' +#define CHAR_QUESTION_MARK '\077' +#define CHAR_COMMERCIAL_AT '\100' +#define CHAR_A '\101' +#define CHAR_B '\102' +#define CHAR_C '\103' +#define CHAR_D '\104' +#define CHAR_E '\105' +#define CHAR_F '\106' +#define CHAR_G '\107' +#define CHAR_H '\110' +#define CHAR_I '\111' +#define CHAR_J '\112' +#define CHAR_K '\113' +#define CHAR_L '\114' +#define CHAR_M '\115' +#define CHAR_N '\116' +#define CHAR_O '\117' +#define CHAR_P '\120' +#define CHAR_Q '\121' +#define CHAR_R '\122' +#define CHAR_S '\123' +#define CHAR_T '\124' +#define CHAR_U '\125' +#define CHAR_V '\126' +#define CHAR_W '\127' +#define CHAR_X '\130' +#define CHAR_Y '\131' +#define CHAR_Z '\132' +#define CHAR_LEFT_SQUARE_BRACKET '\133' +#define CHAR_BACKSLASH '\134' +#define CHAR_RIGHT_SQUARE_BRACKET '\135' +#define CHAR_CIRCUMFLEX_ACCENT '\136' +#define CHAR_UNDERSCORE '\137' +#define CHAR_GRAVE_ACCENT '\140' +#define CHAR_a '\141' +#define CHAR_b '\142' +#define CHAR_c '\143' +#define CHAR_d '\144' +#define CHAR_e '\145' +#define CHAR_f '\146' +#define CHAR_g '\147' +#define CHAR_h '\150' +#define CHAR_i '\151' +#define CHAR_j '\152' +#define CHAR_k '\153' +#define CHAR_l '\154' +#define CHAR_m '\155' +#define CHAR_n '\156' +#define CHAR_o '\157' +#define CHAR_p '\160' +#define CHAR_q '\161' +#define CHAR_r '\162' +#define CHAR_s '\163' +#define CHAR_t '\164' +#define CHAR_u '\165' +#define CHAR_v '\166' +#define CHAR_w '\167' +#define CHAR_x '\170' +#define CHAR_y '\171' +#define CHAR_z '\172' +#define CHAR_LEFT_CURLY_BRACKET '\173' +#define CHAR_VERTICAL_LINE '\174' +#define CHAR_RIGHT_CURLY_BRACKET '\175' +#define CHAR_TILDE '\176' + +#define STR_HT "\011" +#define STR_VT "\013" +#define STR_FF "\014" +#define STR_CR "\015" +#define STR_NL "\012" +#define STR_BS "\010" +#define STR_BEL "\007" +#define STR_ESC "\033" +#define STR_DEL "\177" + +#define STR_SPACE "\040" +#define STR_EXCLAMATION_MARK "\041" +#define STR_QUOTATION_MARK "\042" +#define STR_NUMBER_SIGN "\043" +#define STR_DOLLAR_SIGN "\044" +#define STR_PERCENT_SIGN "\045" +#define STR_AMPERSAND "\046" +#define STR_APOSTROPHE "\047" +#define STR_LEFT_PARENTHESIS "\050" +#define STR_RIGHT_PARENTHESIS "\051" +#define STR_ASTERISK "\052" +#define STR_PLUS "\053" +#define STR_COMMA "\054" +#define STR_MINUS "\055" +#define STR_DOT "\056" +#define STR_SLASH "\057" +#define STR_0 "\060" +#define STR_1 "\061" +#define STR_2 "\062" +#define STR_3 "\063" +#define STR_4 "\064" +#define STR_5 "\065" +#define STR_6 "\066" +#define STR_7 "\067" +#define STR_8 "\070" +#define STR_9 "\071" +#define STR_COLON "\072" +#define STR_SEMICOLON "\073" +#define STR_LESS_THAN_SIGN "\074" +#define STR_EQUALS_SIGN "\075" +#define STR_GREATER_THAN_SIGN "\076" +#define STR_QUESTION_MARK "\077" +#define STR_COMMERCIAL_AT "\100" +#define STR_A "\101" +#define STR_B "\102" +#define STR_C "\103" +#define STR_D "\104" +#define STR_E "\105" +#define STR_F "\106" +#define STR_G "\107" +#define STR_H "\110" +#define STR_I "\111" +#define STR_J "\112" +#define STR_K "\113" +#define STR_L "\114" +#define STR_M "\115" +#define STR_N "\116" +#define STR_O "\117" +#define STR_P "\120" +#define STR_Q "\121" +#define STR_R "\122" +#define STR_S "\123" +#define STR_T "\124" +#define STR_U "\125" +#define STR_V "\126" +#define STR_W "\127" +#define STR_X "\130" +#define STR_Y "\131" +#define STR_Z "\132" +#define STR_LEFT_SQUARE_BRACKET "\133" +#define STR_BACKSLASH "\134" +#define STR_RIGHT_SQUARE_BRACKET "\135" +#define STR_CIRCUMFLEX_ACCENT "\136" +#define STR_UNDERSCORE "\137" +#define STR_GRAVE_ACCENT "\140" +#define STR_a "\141" +#define STR_b "\142" +#define STR_c "\143" +#define STR_d "\144" +#define STR_e "\145" +#define STR_f "\146" +#define STR_g "\147" +#define STR_h "\150" +#define STR_i "\151" +#define STR_j "\152" +#define STR_k "\153" +#define STR_l "\154" +#define STR_m "\155" +#define STR_n "\156" +#define STR_o "\157" +#define STR_p "\160" +#define STR_q "\161" +#define STR_r "\162" +#define STR_s "\163" +#define STR_t "\164" +#define STR_u "\165" +#define STR_v "\166" +#define STR_w "\167" +#define STR_x "\170" +#define STR_y "\171" +#define STR_z "\172" +#define STR_LEFT_CURLY_BRACKET "\173" +#define STR_VERTICAL_LINE "\174" +#define STR_RIGHT_CURLY_BRACKET "\175" +#define STR_TILDE "\176" + +#define STRING_ACCEPT0 STR_A STR_C STR_C STR_E STR_P STR_T "\0" +#define STRING_COMMIT0 STR_C STR_O STR_M STR_M STR_I STR_T "\0" +#define STRING_F0 STR_F "\0" +#define STRING_FAIL0 STR_F STR_A STR_I STR_L "\0" +#define STRING_MARK0 STR_M STR_A STR_R STR_K "\0" +#define STRING_PRUNE0 STR_P STR_R STR_U STR_N STR_E "\0" +#define STRING_SKIP0 STR_S STR_K STR_I STR_P "\0" +#define STRING_THEN STR_T STR_H STR_E STR_N + +#define STRING_alpha0 STR_a STR_l STR_p STR_h STR_a "\0" +#define STRING_lower0 STR_l STR_o STR_w STR_e STR_r "\0" +#define STRING_upper0 STR_u STR_p STR_p STR_e STR_r "\0" +#define STRING_alnum0 STR_a STR_l STR_n STR_u STR_m "\0" +#define STRING_ascii0 STR_a STR_s STR_c STR_i STR_i "\0" +#define STRING_blank0 STR_b STR_l STR_a STR_n STR_k "\0" +#define STRING_cntrl0 STR_c STR_n STR_t STR_r STR_l "\0" +#define STRING_digit0 STR_d STR_i STR_g STR_i STR_t "\0" +#define STRING_graph0 STR_g STR_r STR_a STR_p STR_h "\0" +#define STRING_print0 STR_p STR_r STR_i STR_n STR_t "\0" +#define STRING_punct0 STR_p STR_u STR_n STR_c STR_t "\0" +#define STRING_space0 STR_s STR_p STR_a STR_c STR_e "\0" +#define STRING_word0 STR_w STR_o STR_r STR_d "\0" +#define STRING_xdigit STR_x STR_d STR_i STR_g STR_i STR_t + +#define STRING_DEFINE STR_D STR_E STR_F STR_I STR_N STR_E + +#define STRING_CR_RIGHTPAR STR_C STR_R STR_RIGHT_PARENTHESIS +#define STRING_LF_RIGHTPAR STR_L STR_F STR_RIGHT_PARENTHESIS +#define STRING_CRLF_RIGHTPAR STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS +#define STRING_ANY_RIGHTPAR STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS +#define STRING_ANYCRLF_RIGHTPAR STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS +#define STRING_BSR_ANYCRLF_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS +#define STRING_BSR_UNICODE_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS +#ifdef COMPILE_PCRE8 +#define STRING_UTF_RIGHTPAR STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS +#endif +#ifdef COMPILE_PCRE16 +#define STRING_UTF_RIGHTPAR STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS +#endif +#define STRING_UCP_RIGHTPAR STR_U STR_C STR_P STR_RIGHT_PARENTHESIS +#define STRING_NO_START_OPT_RIGHTPAR STR_N STR_O STR_UNDERSCORE STR_S STR_T STR_A STR_R STR_T STR_UNDERSCORE STR_O STR_P STR_T STR_RIGHT_PARENTHESIS + +#endif /* SUPPORT_UTF */ + +/* Escape items that are just an encoding of a particular data value. */ + +#ifndef ESC_e +#define ESC_e CHAR_ESC +#endif + +#ifndef ESC_f +#define ESC_f CHAR_FF +#endif + +#ifndef ESC_n +#define ESC_n CHAR_NL +#endif + +#ifndef ESC_r +#define ESC_r CHAR_CR +#endif + +/* We can't officially use ESC_t because it is a POSIX reserved identifier +(presumably because of all the others like size_t). */ + +#ifndef ESC_tee +#define ESC_tee CHAR_HT +#endif + +/* Codes for different types of Unicode property */ + +#define PT_ANY 0 /* Any property - matches all chars */ +#define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */ +#define PT_GC 2 /* Specified general characteristic (e.g. L) */ +#define PT_PC 3 /* Specified particular characteristic (e.g. Lu) */ +#define PT_SC 4 /* Script (e.g. Han) */ +#define PT_ALNUM 5 /* Alphanumeric - the union of L and N */ +#define PT_SPACE 6 /* Perl space - Z plus 9,10,12,13 */ +#define PT_PXSPACE 7 /* POSIX space - Z plus 9,10,11,12,13 */ +#define PT_WORD 8 /* Word - L plus N plus underscore */ + +/* Flag bits and data types for the extended class (OP_XCLASS) for classes that +contain characters with values greater than 255. */ + +#define XCL_NOT 0x01 /* Flag: this is a negative class */ +#define XCL_MAP 0x02 /* Flag: a 32-byte map is present */ + +#define XCL_END 0 /* Marks end of individual items */ +#define XCL_SINGLE 1 /* Single item (one multibyte char) follows */ +#define XCL_RANGE 2 /* A range (two multibyte chars) follows */ +#define XCL_PROP 3 /* Unicode property (2-byte property code follows) */ +#define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */ + +/* These are escaped items that aren't just an encoding of a particular data +value such as \n. They must have non-zero values, as check_escape() returns +their negation. Also, they must appear in the same order as in the opcode +definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it +corresponds to "." in DOTALL mode rather than an escape sequence. It is also +used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In +non-DOTALL mode, "." behaves like \N. + +The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc. +when PCRE_UCP is set, when replacement of \d etc by \p sequences is required. +They must be contiguous, and remain in order so that the replacements can be +looked up from a table. + +The final escape must be ESC_REF as subsequent values are used for +backreferences (\1, \2, \3, etc). There are two tests in the code for an escape +greater than ESC_b and less than ESC_Z to detect the types that may be +repeated. These are the types that consume characters. If any new escapes are +put in between that don't consume a character, that code will have to change. +*/ + +enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, + ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, + ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, + ESC_E, ESC_Q, ESC_g, ESC_k, + ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu, + ESC_REF }; + +/* Opcode table: Starting from 1 (i.e. after OP_END), the values up to +OP_EOD must correspond in order to the list of escapes immediately above. + +*** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions +that follow must also be updated to match. There are also tables called +"coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */ + +enum { + OP_END, /* 0 End of pattern */ + + /* Values corresponding to backslashed metacharacters */ + + OP_SOD, /* 1 Start of data: \A */ + OP_SOM, /* 2 Start of match (subject + offset): \G */ + OP_SET_SOM, /* 3 Set start of match (\K) */ + OP_NOT_WORD_BOUNDARY, /* 4 \B */ + OP_WORD_BOUNDARY, /* 5 \b */ + OP_NOT_DIGIT, /* 6 \D */ + OP_DIGIT, /* 7 \d */ + OP_NOT_WHITESPACE, /* 8 \S */ + OP_WHITESPACE, /* 9 \s */ + OP_NOT_WORDCHAR, /* 10 \W */ + OP_WORDCHAR, /* 11 \w */ + + OP_ANY, /* 12 Match any character except newline */ + OP_ALLANY, /* 13 Match any character */ + OP_ANYBYTE, /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */ + OP_NOTPROP, /* 15 \P (not Unicode property) */ + OP_PROP, /* 16 \p (Unicode property) */ + OP_ANYNL, /* 17 \R (any newline sequence) */ + OP_NOT_HSPACE, /* 18 \H (not horizontal whitespace) */ + OP_HSPACE, /* 19 \h (horizontal whitespace) */ + OP_NOT_VSPACE, /* 20 \V (not vertical whitespace) */ + OP_VSPACE, /* 21 \v (vertical whitespace) */ + OP_EXTUNI, /* 22 \X (extended Unicode sequence */ + OP_EODN, /* 23 End of data or \n at end of data: \Z. */ + OP_EOD, /* 24 End of data: \z */ + + OP_CIRC, /* 25 Start of line - not multiline */ + OP_CIRCM, /* 26 Start of line - multiline */ + OP_DOLL, /* 27 End of line - not multiline */ + OP_DOLLM, /* 28 End of line - multiline */ + OP_CHAR, /* 29 Match one character, casefully */ + OP_CHARI, /* 30 Match one character, caselessly */ + OP_NOT, /* 31 Match one character, not the given one, casefully */ + OP_NOTI, /* 32 Match one character, not the given one, caselessly */ + + /* The following sets of 13 opcodes must always be kept in step because + the offset from the first one is used to generate the others. */ + + /**** Single characters, caseful, must precede the caseless ones ****/ + + OP_STAR, /* 33 The maximizing and minimizing versions of */ + OP_MINSTAR, /* 34 these six opcodes must come in pairs, with */ + OP_PLUS, /* 35 the minimizing one second. */ + OP_MINPLUS, /* 36 */ + OP_QUERY, /* 37 */ + OP_MINQUERY, /* 38 */ + + OP_UPTO, /* 39 From 0 to n matches of one character, caseful*/ + OP_MINUPTO, /* 40 */ + OP_EXACT, /* 41 Exactly n matches */ + + OP_POSSTAR, /* 42 Possessified star, caseful */ + OP_POSPLUS, /* 43 Possessified plus, caseful */ + OP_POSQUERY, /* 44 Posesssified query, caseful */ + OP_POSUPTO, /* 45 Possessified upto, caseful */ + + /**** Single characters, caseless, must follow the caseful ones */ + + OP_STARI, /* 46 */ + OP_MINSTARI, /* 47 */ + OP_PLUSI, /* 48 */ + OP_MINPLUSI, /* 49 */ + OP_QUERYI, /* 50 */ + OP_MINQUERYI, /* 51 */ + + OP_UPTOI, /* 52 From 0 to n matches of one character, caseless */ + OP_MINUPTOI, /* 53 */ + OP_EXACTI, /* 54 */ + + OP_POSSTARI, /* 55 Possessified star, caseless */ + OP_POSPLUSI, /* 56 Possessified plus, caseless */ + OP_POSQUERYI, /* 57 Posesssified query, caseless */ + OP_POSUPTOI, /* 58 Possessified upto, caseless */ + + /**** The negated ones must follow the non-negated ones, and match them ****/ + /**** Negated single character, caseful; must precede the caseless ones ****/ + + OP_NOTSTAR, /* 59 The maximizing and minimizing versions of */ + OP_NOTMINSTAR, /* 60 these six opcodes must come in pairs, with */ + OP_NOTPLUS, /* 61 the minimizing one second. They must be in */ + OP_NOTMINPLUS, /* 62 exactly the same order as those above. */ + OP_NOTQUERY, /* 63 */ + OP_NOTMINQUERY, /* 64 */ + + OP_NOTUPTO, /* 65 From 0 to n matches, caseful */ + OP_NOTMINUPTO, /* 66 */ + OP_NOTEXACT, /* 67 Exactly n matches */ + + OP_NOTPOSSTAR, /* 68 Possessified versions, caseful */ + OP_NOTPOSPLUS, /* 69 */ + OP_NOTPOSQUERY, /* 70 */ + OP_NOTPOSUPTO, /* 71 */ + + /**** Negated single character, caseless; must follow the caseful ones ****/ + + OP_NOTSTARI, /* 72 */ + OP_NOTMINSTARI, /* 73 */ + OP_NOTPLUSI, /* 74 */ + OP_NOTMINPLUSI, /* 75 */ + OP_NOTQUERYI, /* 76 */ + OP_NOTMINQUERYI, /* 77 */ + + OP_NOTUPTOI, /* 78 From 0 to n matches, caseless */ + OP_NOTMINUPTOI, /* 79 */ + OP_NOTEXACTI, /* 80 Exactly n matches */ + + OP_NOTPOSSTARI, /* 81 Possessified versions, caseless */ + OP_NOTPOSPLUSI, /* 82 */ + OP_NOTPOSQUERYI, /* 83 */ + OP_NOTPOSUPTOI, /* 84 */ + + /**** Character types ****/ + + OP_TYPESTAR, /* 85 The maximizing and minimizing versions of */ + OP_TYPEMINSTAR, /* 86 these six opcodes must come in pairs, with */ + OP_TYPEPLUS, /* 87 the minimizing one second. These codes must */ + OP_TYPEMINPLUS, /* 88 be in exactly the same order as those above. */ + OP_TYPEQUERY, /* 89 */ + OP_TYPEMINQUERY, /* 90 */ + + OP_TYPEUPTO, /* 91 From 0 to n matches */ + OP_TYPEMINUPTO, /* 92 */ + OP_TYPEEXACT, /* 93 Exactly n matches */ + + OP_TYPEPOSSTAR, /* 94 Possessified versions */ + OP_TYPEPOSPLUS, /* 95 */ + OP_TYPEPOSQUERY, /* 96 */ + OP_TYPEPOSUPTO, /* 97 */ + + /* These are used for character classes and back references; only the + first six are the same as the sets above. */ + + OP_CRSTAR, /* 98 The maximizing and minimizing versions of */ + OP_CRMINSTAR, /* 99 all these opcodes must come in pairs, with */ + OP_CRPLUS, /* 100 the minimizing one second. These codes must */ + OP_CRMINPLUS, /* 101 be in exactly the same order as those above. */ + OP_CRQUERY, /* 102 */ + OP_CRMINQUERY, /* 103 */ + + OP_CRRANGE, /* 104 These are different to the three sets above. */ + OP_CRMINRANGE, /* 105 */ + + /* End of quantifier opcodes */ + + OP_CLASS, /* 106 Match a character class, chars < 256 only */ + OP_NCLASS, /* 107 Same, but the bitmap was created from a negative + class - the difference is relevant only when a + character > 255 is encountered. */ + OP_XCLASS, /* 108 Extended class for handling > 255 chars within the + class. This does both positive and negative. */ + OP_REF, /* 109 Match a back reference, casefully */ + OP_REFI, /* 110 Match a back reference, caselessly */ + OP_RECURSE, /* 111 Match a numbered subpattern (possibly recursive) */ + OP_CALLOUT, /* 112 Call out to external function if provided */ + + OP_ALT, /* 113 Start of alternation */ + OP_KET, /* 114 End of group that doesn't have an unbounded repeat */ + OP_KETRMAX, /* 115 These two must remain together and in this */ + OP_KETRMIN, /* 116 order. They are for groups the repeat for ever. */ + OP_KETRPOS, /* 117 Possessive unlimited repeat. */ + + /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four + asserts must remain in order. */ + + OP_REVERSE, /* 118 Move pointer back - used in lookbehind assertions */ + OP_ASSERT, /* 119 Positive lookahead */ + OP_ASSERT_NOT, /* 120 Negative lookahead */ + OP_ASSERTBACK, /* 121 Positive lookbehind */ + OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */ + + /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately + after the assertions, with ONCE first, as there's a test for >= ONCE for a + subpattern that isn't an assertion. The POS versions must immediately follow + the non-POS versions in each case. */ + + OP_ONCE, /* 123 Atomic group, contains captures */ + OP_ONCE_NC, /* 124 Atomic group containing no captures */ + OP_BRA, /* 125 Start of non-capturing bracket */ + OP_BRAPOS, /* 126 Ditto, with unlimited, possessive repeat */ + OP_CBRA, /* 127 Start of capturing bracket */ + OP_CBRAPOS, /* 128 Ditto, with unlimited, possessive repeat */ + OP_COND, /* 129 Conditional group */ + + /* These five must follow the previous five, in the same order. There's a + check for >= SBRA to distinguish the two sets. */ + + OP_SBRA, /* 130 Start of non-capturing bracket, check empty */ + OP_SBRAPOS, /* 131 Ditto, with unlimited, possessive repeat */ + OP_SCBRA, /* 132 Start of capturing bracket, check empty */ + OP_SCBRAPOS, /* 133 Ditto, with unlimited, possessive repeat */ + OP_SCOND, /* 134 Conditional group, check empty */ + + /* The next two pairs must (respectively) be kept together. */ + + OP_CREF, /* 135 Used to hold a capture number as condition */ + OP_NCREF, /* 136 Same, but generated by a name reference*/ + OP_RREF, /* 137 Used to hold a recursion number as condition */ + OP_NRREF, /* 138 Same, but generated by a name reference*/ + OP_DEF, /* 139 The DEFINE condition */ + + OP_BRAZERO, /* 140 These two must remain together and in this */ + OP_BRAMINZERO, /* 141 order. */ + OP_BRAPOSZERO, /* 142 */ + + /* These are backtracking control verbs */ + + OP_MARK, /* 143 always has an argument */ + OP_PRUNE, /* 144 */ + OP_PRUNE_ARG, /* 145 same, but with argument */ + OP_SKIP, /* 146 */ + OP_SKIP_ARG, /* 147 same, but with argument */ + OP_THEN, /* 148 */ + OP_THEN_ARG, /* 149 same, but with argument */ + OP_COMMIT, /* 150 */ + + /* These are forced failure and success verbs */ + + OP_FAIL, /* 151 */ + OP_ACCEPT, /* 152 */ + OP_ASSERT_ACCEPT, /* 153 Used inside assertions */ + OP_CLOSE, /* 154 Used before OP_ACCEPT to close open captures */ + + /* This is used to skip a subpattern with a {0} quantifier */ + + OP_SKIPZERO, /* 155 */ + + /* This is not an opcode, but is used to check that tables indexed by opcode + are the correct length, in order to catch updating errors - there have been + some in the past. */ + + OP_TABLE_LENGTH +}; + +/* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro +definitions that follow must also be updated to match. There are also tables +called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */ + + +/* This macro defines textual names for all the opcodes. These are used only +for debugging, and some of them are only partial names. The macro is referenced +only in pcre_printint.c, which fills out the full names in many cases (and in +some cases doesn't actually use these names at all). */ + +#define OP_NAME_LIST \ + "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d", \ + "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte", \ + "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v", \ + "extuni", "\\Z", "\\z", \ + "^", "^", "$", "$", "char", "chari", "not", "noti", \ + "*", "*?", "+", "+?", "?", "??", \ + "{", "{", "{", \ + "*+","++", "?+", "{", \ + "*", "*?", "+", "+?", "?", "??", \ + "{", "{", "{", \ + "*+","++", "?+", "{", \ + "*", "*?", "+", "+?", "?", "??", \ + "{", "{", "{", \ + "*+","++", "?+", "{", \ + "*", "*?", "+", "+?", "?", "??", \ + "{", "{", "{", \ + "*+","++", "?+", "{", \ + "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \ + "*+","++", "?+", "{", \ + "*", "*?", "+", "+?", "?", "??", "{", "{", \ + "class", "nclass", "xclass", "Ref", "Refi", \ + "Recurse", "Callout", \ + "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos", \ + "Reverse", "Assert", "Assert not", "AssertB", "AssertB not", \ + "Once", "Once_NC", \ + "Bra", "BraPos", "CBra", "CBraPos", \ + "Cond", \ + "SBra", "SBraPos", "SCBra", "SCBraPos", \ + "SCond", \ + "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def", \ + "Brazero", "Braminzero", "Braposzero", \ + "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP", \ + "*THEN", "*THEN", "*COMMIT", "*FAIL", \ + "*ACCEPT", "*ASSERT_ACCEPT", \ + "Close", "Skip zero" + + +/* This macro defines the length of fixed length operations in the compiled +regex. The lengths are used when searching for specific things, and also in the +debugging printing of a compiled regex. We use a macro so that it can be +defined close to the definitions of the opcodes themselves. + +As things have been extended, some of these are no longer fixed lenths, but are +minima instead. For example, the length of a single-character repeat may vary +in UTF-8 mode. The code that uses this table must know about such things. */ + +#define OP_LENGTHS \ + 1, /* End */ \ + 1, 1, 1, 1, 1, /* \A, \G, \K, \B, \b */ \ + 1, 1, 1, 1, 1, 1, /* \D, \d, \S, \s, \W, \w */ \ + 1, 1, 1, /* Any, AllAny, Anybyte */ \ + 3, 3, /* \P, \p */ \ + 1, 1, 1, 1, 1, /* \R, \H, \h, \V, \v */ \ + 1, /* \X */ \ + 1, 1, 1, 1, 1, 1, /* \Z, \z, ^, ^M, $, $M */ \ + 2, /* Char - the minimum length */ \ + 2, /* Chari - the minimum length */ \ + 2, /* not */ \ + 2, /* noti */ \ + /* Positive single-char repeats ** These are */ \ + 2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* upto, minupto ** mode */ \ + 2+IMM2_SIZE, /* exact */ \ + 2, 2, 2, 2+IMM2_SIZE, /* *+, ++, ?+, upto+ */ \ + 2, 2, 2, 2, 2, 2, /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8 */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* upto I, minupto I */ \ + 2+IMM2_SIZE, /* exact I */ \ + 2, 2, 2, 2+IMM2_SIZE, /* *+I, ++I, ?+I, upto+I */ \ + /* Negative single-char repeats - only for chars < 256 */ \ + 2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* NOT upto, minupto */ \ + 2+IMM2_SIZE, /* NOT exact */ \ + 2, 2, 2, 2+IMM2_SIZE, /* Possessive NOT *, +, ?, upto */ \ + 2, 2, 2, 2, 2, 2, /* NOT *I, *?I, +I, +?I, ?I, ??I */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* NOT upto I, minupto I */ \ + 2+IMM2_SIZE, /* NOT exact I */ \ + 2, 2, 2, 2+IMM2_SIZE, /* Possessive NOT *I, +I, ?I, upto I */ \ + /* Positive type repeats */ \ + 2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \ + 2+IMM2_SIZE, 2+IMM2_SIZE, /* Type upto, minupto */ \ + 2+IMM2_SIZE, /* Type exact */ \ + 2, 2, 2, 2+IMM2_SIZE, /* Possessive *+, ++, ?+, upto+ */ \ + /* Character class & ref repeats */ \ + 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \ + 1+2*IMM2_SIZE, 1+2*IMM2_SIZE, /* CRRANGE, CRMINRANGE */ \ + 1+(32/sizeof(pcre_uchar)), /* CLASS */ \ + 1+(32/sizeof(pcre_uchar)), /* NCLASS */ \ + 0, /* XCLASS - variable length */ \ + 1+IMM2_SIZE, /* REF */ \ + 1+IMM2_SIZE, /* REFI */ \ + 1+LINK_SIZE, /* RECURSE */ \ + 2+2*LINK_SIZE, /* CALLOUT */ \ + 1+LINK_SIZE, /* Alt */ \ + 1+LINK_SIZE, /* Ket */ \ + 1+LINK_SIZE, /* KetRmax */ \ + 1+LINK_SIZE, /* KetRmin */ \ + 1+LINK_SIZE, /* KetRpos */ \ + 1+LINK_SIZE, /* Reverse */ \ + 1+LINK_SIZE, /* Assert */ \ + 1+LINK_SIZE, /* Assert not */ \ + 1+LINK_SIZE, /* Assert behind */ \ + 1+LINK_SIZE, /* Assert behind not */ \ + 1+LINK_SIZE, /* ONCE */ \ + 1+LINK_SIZE, /* ONCE_NC */ \ + 1+LINK_SIZE, /* BRA */ \ + 1+LINK_SIZE, /* BRAPOS */ \ + 1+LINK_SIZE+IMM2_SIZE, /* CBRA */ \ + 1+LINK_SIZE+IMM2_SIZE, /* CBRAPOS */ \ + 1+LINK_SIZE, /* COND */ \ + 1+LINK_SIZE, /* SBRA */ \ + 1+LINK_SIZE, /* SBRAPOS */ \ + 1+LINK_SIZE+IMM2_SIZE, /* SCBRA */ \ + 1+LINK_SIZE+IMM2_SIZE, /* SCBRAPOS */ \ + 1+LINK_SIZE, /* SCOND */ \ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* CREF, NCREF */ \ + 1+IMM2_SIZE, 1+IMM2_SIZE, /* RREF, NRREF */ \ + 1, /* DEF */ \ + 1, 1, 1, /* BRAZERO, BRAMINZERO, BRAPOSZERO */ \ + 3, 1, 3, /* MARK, PRUNE, PRUNE_ARG */ \ + 1, 3, /* SKIP, SKIP_ARG */ \ + 1, 3, /* THEN, THEN_ARG */ \ + 1, 1, 1, 1, /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT */ \ + 1+IMM2_SIZE, 1 /* CLOSE, SKIPZERO */ + +/* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion" +condition. */ + +#define RREF_ANY 0xffff + +/* Compile time error code numbers. They are given names so that they can more +easily be tracked. When a new number is added, the table called eint in +pcreposix.c must be updated. */ + +enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9, + ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19, + ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29, + ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39, + ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49, + ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59, + ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69, + ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERRCOUNT }; + +/* JIT compiling modes. The function list is indexed by them. */ +enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE, + JIT_NUMBER_OF_COMPILE_MODES }; + +/* The real format of the start of the pcre block; the index of names and the +code vector run on as long as necessary after the end. We store an explicit +offset to the name table so that if a regex is compiled on one host, saved, and +then run on another where the size of pointers is different, all might still +be well. For the case of compiled-on-4 and run-on-8, we include an extra +pointer that is always NULL. For future-proofing, a few dummy fields were +originally included - even though you can never get this planning right - but +there is only one left now. + +NOTE NOTE NOTE: +Because people can now save and re-use compiled patterns, any additions to this +structure should be made at the end, and something earlier (e.g. a new +flag in the options or one of the dummy fields) should indicate that the new +fields are present. Currently PCRE always sets the dummy fields to zero. +NOTE NOTE NOTE +*/ + +#ifdef COMPILE_PCRE8 +#define REAL_PCRE real_pcre +#else +#define REAL_PCRE real_pcre16 +#endif + +typedef struct REAL_PCRE { + pcre_uint32 magic_number; + pcre_uint32 size; /* Total that was malloced */ + pcre_uint32 options; /* Public options */ + pcre_uint16 flags; /* Private flags */ + pcre_uint16 max_lookbehind; /* Longest lookbehind (characters) */ + pcre_uint16 top_bracket; /* Highest numbered group */ + pcre_uint16 top_backref; /* Highest numbered back reference */ + pcre_uint16 first_char; /* Starting character */ + pcre_uint16 req_char; /* This character must be seen */ + pcre_uint16 name_table_offset; /* Offset to name table that follows */ + pcre_uint16 name_entry_size; /* Size of any name items */ + pcre_uint16 name_count; /* Number of name items */ + pcre_uint16 ref_count; /* Reference count */ + const pcre_uint8 *tables; /* Pointer to tables or NULL for std */ + const pcre_uint8 *nullpad; /* NULL padding */ +} REAL_PCRE; + +/* The format of the block used to store data from pcre_study(). The same +remark (see NOTE above) about extending this structure applies. */ + +typedef struct pcre_study_data { + pcre_uint32 size; /* Total that was malloced */ + pcre_uint32 flags; /* Private flags */ + pcre_uint8 start_bits[32]; /* Starting char bits */ + pcre_uint32 minlength; /* Minimum subject length */ +} pcre_study_data; + +/* Structure for building a chain of open capturing subpatterns during +compiling, so that instructions to close them can be compiled when (*ACCEPT) is +encountered. This is also used to identify subpatterns that contain recursive +back references to themselves, so that they can be made atomic. */ + +typedef struct open_capitem { + struct open_capitem *next; /* Chain link */ + pcre_uint16 number; /* Capture number */ + pcre_uint16 flag; /* Set TRUE if recursive back ref */ +} open_capitem; + +/* Structure for passing "static" information around between the functions +doing the compiling, so that they are thread-safe. */ + +typedef struct compile_data { + const pcre_uint8 *lcc; /* Points to lower casing table */ + const pcre_uint8 *fcc; /* Points to case-flipping table */ + const pcre_uint8 *cbits; /* Points to character type table */ + const pcre_uint8 *ctypes; /* Points to table of type maps */ + const pcre_uchar *start_workspace;/* The start of working space */ + const pcre_uchar *start_code; /* The start of the compiled code */ + const pcre_uchar *start_pattern; /* The start of the pattern */ + const pcre_uchar *end_pattern; /* The end of the pattern */ + open_capitem *open_caps; /* Chain of open capture items */ + pcre_uchar *hwm; /* High watermark of workspace */ + pcre_uchar *name_table; /* The name/number table */ + int names_found; /* Number of entries so far */ + int name_entry_size; /* Size of each entry */ + int workspace_size; /* Size of workspace */ + int bracount; /* Count of capturing parens as we compile */ + int final_bracount; /* Saved value after first pass */ + int max_lookbehind; /* Maximum lookbehind (characters) */ + int top_backref; /* Maximum back reference */ + unsigned int backref_map; /* Bitmap of low back refs */ + int assert_depth; /* Depth of nested assertions */ + int external_options; /* External (initial) options */ + int external_flags; /* External flag bits to be set */ + int req_varyopt; /* "After variable item" flag for reqbyte */ + BOOL had_accept; /* (*ACCEPT) encountered */ + BOOL check_lookbehind; /* Lookbehinds need later checking */ + int nltype; /* Newline type */ + int nllen; /* Newline string length */ + pcre_uchar nl[4]; /* Newline string when fixed length */ +} compile_data; + +/* Structure for maintaining a chain of pointers to the currently incomplete +branches, for testing for left recursion while compiling. */ + +typedef struct branch_chain { + struct branch_chain *outer; + pcre_uchar *current_branch; +} branch_chain; + +/* Structure for items in a linked list that represents an explicit recursive +call within the pattern; used by pcre_exec(). */ + +typedef struct recursion_info { + struct recursion_info *prevrec; /* Previous recursion record (or NULL) */ + int group_num; /* Number of group that was called */ + int *offset_save; /* Pointer to start of saved offsets */ + int saved_max; /* Number of saved offsets */ + PCRE_PUCHAR subject_position; /* Position at start of recursion */ +} recursion_info; + +/* A similar structure for pcre_dfa_exec(). */ + +typedef struct dfa_recursion_info { + struct dfa_recursion_info *prevrec; + int group_num; + PCRE_PUCHAR subject_position; +} dfa_recursion_info; + +/* Structure for building a chain of data for holding the values of the subject +pointer at the start of each subpattern, so as to detect when an empty string +has been matched by a subpattern - to break infinite loops; used by +pcre_exec(). */ + +typedef struct eptrblock { + struct eptrblock *epb_prev; + PCRE_PUCHAR epb_saved_eptr; +} eptrblock; + + +/* Structure for passing "static" information around between the functions +doing traditional NFA matching, so that they are thread-safe. */ + +typedef struct match_data { + unsigned long int match_call_count; /* As it says */ + unsigned long int match_limit; /* As it says */ + unsigned long int match_limit_recursion; /* As it says */ + int *offset_vector; /* Offset vector */ + int offset_end; /* One past the end */ + int offset_max; /* The maximum usable for return data */ + int nltype; /* Newline type */ + int nllen; /* Newline string length */ + int name_count; /* Number of names in name table */ + int name_entry_size; /* Size of entry in names table */ + pcre_uchar *name_table; /* Table of names */ + pcre_uchar nl[4]; /* Newline string when fixed */ + const pcre_uint8 *lcc; /* Points to lower casing table */ + const pcre_uint8 *fcc; /* Points to case-flipping table */ + const pcre_uint8 *ctypes; /* Points to table of type maps */ + BOOL offset_overflow; /* Set if too many extractions */ + BOOL notbol; /* NOTBOL flag */ + BOOL noteol; /* NOTEOL flag */ + BOOL utf; /* UTF-8 / UTF-16 flag */ + BOOL jscript_compat; /* JAVASCRIPT_COMPAT flag */ + BOOL use_ucp; /* PCRE_UCP flag */ + BOOL endonly; /* Dollar not before final \n */ + BOOL notempty; /* Empty string match not wanted */ + BOOL notempty_atstart; /* Empty string match at start not wanted */ + BOOL hitend; /* Hit the end of the subject at some point */ + BOOL bsr_anycrlf; /* \R is just any CRLF, not full Unicode */ + BOOL hasthen; /* Pattern contains (*THEN) */ + BOOL ignore_skip_arg; /* For re-run when SKIP name not found */ + const pcre_uchar *start_code; /* For use when recursing */ + PCRE_PUCHAR start_subject; /* Start of the subject string */ + PCRE_PUCHAR end_subject; /* End of the subject string */ + PCRE_PUCHAR start_match_ptr; /* Start of matched string */ + PCRE_PUCHAR end_match_ptr; /* Subject position at end match */ + PCRE_PUCHAR start_used_ptr; /* Earliest consulted character */ + int partial; /* PARTIAL options */ + int end_offset_top; /* Highwater mark at end of match */ + int capture_last; /* Most recent capture number */ + int start_offset; /* The start offset value */ + int match_function_type; /* Set for certain special calls of MATCH() */ + eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */ + int eptrn; /* Next free eptrblock */ + recursion_info *recursive; /* Linked list of recursion data */ + void *callout_data; /* To pass back to callouts */ + const pcre_uchar *mark; /* Mark pointer to pass back on success */ + const pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */ + const pcre_uchar *once_target; /* Where to back up to for atomic groups */ +#ifdef NO_RECURSE + void *match_frames_base; /* For remembering malloc'd frames */ +#endif +} match_data; + +/* A similar structure is used for the same purpose by the DFA matching +functions. */ + +typedef struct dfa_match_data { + const pcre_uchar *start_code; /* Start of the compiled pattern */ + const pcre_uchar *start_subject ; /* Start of the subject string */ + const pcre_uchar *end_subject; /* End of subject string */ + const pcre_uchar *start_used_ptr; /* Earliest consulted character */ + const pcre_uint8 *tables; /* Character tables */ + int start_offset; /* The start offset value */ + int moptions; /* Match options */ + int poptions; /* Pattern options */ + int nltype; /* Newline type */ + int nllen; /* Newline string length */ + pcre_uchar nl[4]; /* Newline string when fixed */ + void *callout_data; /* To pass back to callouts */ + dfa_recursion_info *recursive; /* Linked list of recursion data */ +} dfa_match_data; + +/* Bit definitions for entries in the pcre_ctypes table. */ + +#define ctype_space 0x01 +#define ctype_letter 0x02 +#define ctype_digit 0x04 +#define ctype_xdigit 0x08 +#define ctype_word 0x10 /* alphanumeric or '_' */ +#define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */ + +/* Offsets for the bitmap tables in pcre_cbits. Each table contains a set +of bits for a class map. Some classes are built by combining these tables. */ + +#define cbit_space 0 /* [:space:] or \s */ +#define cbit_xdigit 32 /* [:xdigit:] */ +#define cbit_digit 64 /* [:digit:] or \d */ +#define cbit_upper 96 /* [:upper:] */ +#define cbit_lower 128 /* [:lower:] */ +#define cbit_word 160 /* [:word:] or \w */ +#define cbit_graph 192 /* [:graph:] */ +#define cbit_print 224 /* [:print:] */ +#define cbit_punct 256 /* [:punct:] */ +#define cbit_cntrl 288 /* [:cntrl:] */ +#define cbit_length 320 /* Length of the cbits table */ + +/* Offsets of the various tables from the base tables pointer, and +total length. */ + +#define lcc_offset 0 +#define fcc_offset 256 +#define cbits_offset 512 +#define ctypes_offset (cbits_offset + cbit_length) +#define tables_length (ctypes_offset + 256) + +/* Internal function and data prefixes. */ + +#ifdef COMPILE_PCRE8 +#ifndef PUBL +#define PUBL(name) pcre_##name +#endif +#ifndef PRIV +#define PRIV(name) _pcre_##name +#endif +#else /* COMPILE_PCRE8 */ +#ifdef COMPILE_PCRE16 +#ifndef PUBL +#define PUBL(name) pcre16_##name +#endif +#ifndef PRIV +#define PRIV(name) _pcre16_##name +#endif +#else +#error Unsupported compiling mode +#endif /* COMPILE_PCRE16 */ +#endif /* COMPILE_PCRE8 */ + +/* Layout of the UCP type table that translates property names into types and +codes. Each entry used to point directly to a name, but to reduce the number of +relocations in shared libraries, it now has an offset into a single string +instead. */ + +typedef struct { + pcre_uint16 name_offset; + pcre_uint16 type; + pcre_uint16 value; +} ucp_type_table; + + +/* Internal shared data tables. These are tables that are used by more than one +of the exported public functions. They have to be "external" in the C sense, +but are not part of the PCRE public API. The data for these tables is in the +pcre_tables.c module. */ + +#ifdef COMPILE_PCRE8 + +extern const int PRIV(utf8_table1)[]; +extern const int PRIV(utf8_table1_size); +extern const int PRIV(utf8_table2)[]; +extern const int PRIV(utf8_table3)[]; +extern const pcre_uint8 PRIV(utf8_table4)[]; + +#endif /* COMPILE_PCRE8 */ + +extern const char PRIV(utt_names)[]; +extern const ucp_type_table PRIV(utt)[]; +extern const int PRIV(utt_size); + +extern const pcre_uint8 PRIV(default_tables)[]; + +extern const pcre_uint8 PRIV(OP_lengths)[]; + + +/* Internal shared functions. These are functions that are used by more than +one of the exported public functions. They have to be "external" in the C +sense, but are not part of the PCRE public API. */ + +/* String comparison functions. */ +#ifdef COMPILE_PCRE8 + +#define STRCMP_UC_UC(str1, str2) \ + strcmp((char *)(str1), (char *)(str2)) +#define STRCMP_UC_C8(str1, str2) \ + strcmp((char *)(str1), (str2)) +#define STRNCMP_UC_UC(str1, str2, num) \ + strncmp((char *)(str1), (char *)(str2), (num)) +#define STRNCMP_UC_C8(str1, str2, num) \ + strncmp((char *)(str1), (str2), (num)) +#define STRLEN_UC(str) strlen((const char *)str) + +#else + +extern int PRIV(strcmp_uc_uc)(const pcre_uchar *, + const pcre_uchar *); +extern int PRIV(strcmp_uc_c8)(const pcre_uchar *, + const char *); +extern int PRIV(strncmp_uc_uc)(const pcre_uchar *, + const pcre_uchar *, unsigned int num); +extern int PRIV(strncmp_uc_c8)(const pcre_uchar *, + const char *, unsigned int num); +extern unsigned int PRIV(strlen_uc)(const pcre_uchar *str); + +#define STRCMP_UC_UC(str1, str2) \ + PRIV(strcmp_uc_uc)((str1), (str2)) +#define STRCMP_UC_C8(str1, str2) \ + PRIV(strcmp_uc_c8)((str1), (str2)) +#define STRNCMP_UC_UC(str1, str2, num) \ + PRIV(strncmp_uc_uc)((str1), (str2), (num)) +#define STRNCMP_UC_C8(str1, str2, num) \ + PRIV(strncmp_uc_c8)((str1), (str2), (num)) +#define STRLEN_UC(str) PRIV(strlen_uc)(str) + +#endif /* COMPILE_PCRE8 */ + +extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int); +extern BOOL PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR, + int *, BOOL); +extern int PRIV(ord2utf)(pcre_uint32, pcre_uchar *); +extern int PRIV(valid_utf)(PCRE_PUCHAR, int, int *); +extern BOOL PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR, + int *, BOOL); +extern BOOL PRIV(xclass)(int, const pcre_uchar *, BOOL); + +#ifdef SUPPORT_JIT +extern void PRIV(jit_compile)(const REAL_PCRE *, + PUBL(extra) *, int); +extern int PRIV(jit_exec)(const REAL_PCRE *, const PUBL(extra) *, + const pcre_uchar *, int, int, int, int *, int); +extern void PRIV(jit_free)(void *); +extern int PRIV(jit_get_size)(void *); +extern const char* PRIV(jit_get_target)(void); +#endif + +/* Unicode character database (UCD) */ + +typedef struct { + pcre_uint8 script; + pcre_uint8 chartype; + pcre_int32 other_case; +} ucd_record; + +extern const ucd_record PRIV(ucd_records)[]; +extern const pcre_uint8 PRIV(ucd_stage1)[]; +extern const pcre_uint16 PRIV(ucd_stage2)[]; +extern const int PRIV(ucp_gentype)[]; +#ifdef SUPPORT_JIT +extern const int PRIV(ucp_typerange)[]; +#endif + +#ifdef SUPPORT_UCP +/* UCD access macros */ + +#define UCD_BLOCK_SIZE 128 +#define GET_UCD(ch) (PRIV(ucd_records) + \ + PRIV(ucd_stage2)[PRIV(ucd_stage1)[(ch) / UCD_BLOCK_SIZE] * \ + UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE]) + +#define UCD_CHARTYPE(ch) GET_UCD(ch)->chartype +#define UCD_SCRIPT(ch) GET_UCD(ch)->script +#define UCD_CATEGORY(ch) PRIV(ucp_gentype)[UCD_CHARTYPE(ch)] +#define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case) + +#endif /* SUPPORT_UCP */ #endif diff --git a/pcre/pcre_jit_compile.c b/pcre/pcre_jit_compile.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_jit_compile.c @@ -0,0 +1,7454 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Written by Philip Hazel + Copyright (c) 1997-2012 University of Cambridge + + The machine code generator part (this module) was written by Zoltan Herczeg + Copyright (c) 2010-2012 + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + +#ifdef SUPPORT_JIT + +/* All-in-one: Since we use the JIT compiler only from here, +we just include it. This way we don't need to touch the build +system files. */ + +#define SLJIT_MALLOC(size) (PUBL(malloc))(size) +#define SLJIT_FREE(ptr) (PUBL(free))(ptr) +#define SLJIT_CONFIG_AUTO 1 +#define SLJIT_CONFIG_STATIC 1 +#define SLJIT_VERBOSE 0 +#define SLJIT_DEBUG 0 + +#include "sljit/sljitLir.c" + +#if defined SLJIT_CONFIG_UNSUPPORTED && SLJIT_CONFIG_UNSUPPORTED +#error Unsupported architecture +#endif + +/* Allocate memory on the stack. Fast, but limited size. */ +#define LOCAL_SPACE_SIZE 32768 + +#define STACK_GROWTH_RATE 8192 + +/* Enable to check that the allocation could destroy temporaries. */ +#if defined SLJIT_DEBUG && SLJIT_DEBUG +#define DESTROY_REGISTERS 1 +#endif + +/* +Short summary about the backtracking mechanism empolyed by the jit code generator: + +The code generator follows the recursive nature of the PERL compatible regular +expressions. The basic blocks of regular expressions are condition checkers +whose execute different commands depending on the result of the condition check. +The relationship between the operators can be horizontal (concatenation) and +vertical (sub-expression) (See struct backtrack_common for more details). + + 'ab' - 'a' and 'b' regexps are concatenated + 'a+' - 'a' is the sub-expression of the '+' operator + +The condition checkers are boolean (true/false) checkers. Machine code is generated +for the checker itself and for the actions depending on the result of the checker. +The 'true' case is called as the try path (expected path), and the other is called as +the 'backtrack' path. Branch instructions are expesive for all CPUs, so we avoid taken +branches on the try path. + + Greedy star operator (*) : + Try path: match happens. + Backtrack path: match failed. + Non-greedy star operator (*?) : + Try path: no need to perform a match. + Backtrack path: match is required. + +The following example shows how the code generated for a capturing bracket +with two alternatives. Let A, B, C, D are arbirary regular expressions, and +we have the following regular expression: + + A(B|C)D + +The generated code will be the following: + + A try path + '(' try path (pushing arguments to the stack) + B try path + ')' try path (pushing arguments to the stack) + D try path + return with successful match + + D backtrack path + ')' backtrack path (If we arrived from "C" jump to the backtrack of "C") + B backtrack path + C expected path + jump to D try path + C backtrack path + A backtrack path + + Notice, that the order of backtrack code paths are the opposite of the fast + code paths. In this way the topmost value on the stack is always belong + to the current backtrack code path. The backtrack path must check + whether there is a next alternative. If so, it needs to jump back to + the try path eventually. Otherwise it needs to clear out its own stack + frame and continue the execution on the backtrack code paths. +*/ + +/* +Saved stack frames: + +Atomic blocks and asserts require reloading the values of local variables +when the backtrack mechanism performed. Because of OP_RECURSE, the locals +are not necessarly known in compile time, thus we need a dynamic restore +mechanism. + +The stack frames are stored in a chain list, and have the following format: +([ capturing bracket offset ][ start value ][ end value ])+ ... [ 0 ] [ previous head ] + +Thus we can restore the locals to a particular point in the stack. +*/ + +typedef struct jit_arguments { + /* Pointers first. */ + struct sljit_stack *stack; + const pcre_uchar *str; + const pcre_uchar *begin; + const pcre_uchar *end; + int *offsets; + pcre_uchar *uchar_ptr; + pcre_uchar *mark_ptr; + /* Everything else after. */ + int offsetcount; + int calllimit; + pcre_uint8 notbol; + pcre_uint8 noteol; + pcre_uint8 notempty; + pcre_uint8 notempty_atstart; +} jit_arguments; + +typedef struct executable_functions { + void *executable_funcs[JIT_NUMBER_OF_COMPILE_MODES]; + PUBL(jit_callback) callback; + void *userdata; + sljit_uw executable_sizes[JIT_NUMBER_OF_COMPILE_MODES]; +} executable_functions; + +typedef struct jump_list { + struct sljit_jump *jump; + struct jump_list *next; +} jump_list; + +enum stub_types { stack_alloc }; + +typedef struct stub_list { + enum stub_types type; + int data; + struct sljit_jump *start; + struct sljit_label *leave; + struct stub_list *next; +} stub_list; + +typedef int (SLJIT_CALL *jit_function)(jit_arguments *args); + +/* The following structure is the key data type for the recursive +code generator. It is allocated by compile_trypath, and contains +the aguments for compile_backtrackpath. Must be the first member +of its descendants. */ +typedef struct backtrack_common { + /* Concatenation stack. */ + struct backtrack_common *prev; + jump_list *nextbacktracks; + /* Internal stack (for component operators). */ + struct backtrack_common *top; + jump_list *topbacktracks; + /* Opcode pointer. */ + pcre_uchar *cc; +} backtrack_common; + +typedef struct assert_backtrack { + backtrack_common common; + jump_list *condfailed; + /* Less than 0 (-1) if a frame is not needed. */ + int framesize; + /* Points to our private memory word on the stack. */ + int localptr; + /* For iterators. */ + struct sljit_label *trypath; +} assert_backtrack; + +typedef struct bracket_backtrack { + backtrack_common common; + /* Where to coninue if an alternative is successfully matched. */ + struct sljit_label *alttrypath; + /* For rmin and rmax iterators. */ + struct sljit_label *recursivetrypath; + /* For greedy ? operator. */ + struct sljit_label *zerotrypath; + /* Contains the branches of a failed condition. */ + union { + /* Both for OP_COND, OP_SCOND. */ + jump_list *condfailed; + assert_backtrack *assert; + /* For OP_ONCE. -1 if not needed. */ + int framesize; + } u; + /* Points to our private memory word on the stack. */ + int localptr; +} bracket_backtrack; + +typedef struct bracketpos_backtrack { + backtrack_common common; + /* Points to our private memory word on the stack. */ + int localptr; + /* Reverting stack is needed. */ + int framesize; + /* Allocated stack size. */ + int stacksize; +} bracketpos_backtrack; + +typedef struct braminzero_backtrack { + backtrack_common common; + struct sljit_label *trypath; +} braminzero_backtrack; + +typedef struct iterator_backtrack { + backtrack_common common; + /* Next iteration. */ + struct sljit_label *trypath; +} iterator_backtrack; + +typedef struct recurse_entry { + struct recurse_entry *next; + /* Contains the function entry. */ + struct sljit_label *entry; + /* Collects the calls until the function is not created. */ + jump_list *calls; + /* Points to the starting opcode. */ + int start; +} recurse_entry; + +typedef struct recurse_backtrack { + backtrack_common common; +} recurse_backtrack; + +typedef struct compiler_common { + struct sljit_compiler *compiler; + pcre_uchar *start; + + /* Opcode local area direct map. */ + int *localptrs; + int cbraptr; + /* OVector starting point. Must be divisible by 2. */ + int ovector_start; + /* Last known position of the requested byte. */ + int req_char_ptr; + /* Head of the last recursion. */ + int recursive_head; + /* First inspected character for partial matching. */ + int start_used_ptr; + /* Starting pointer for partial soft matches. */ + int hit_start; + /* End pointer of the first line. */ + int first_line_end; + /* Points to the marked string. */ + int mark_ptr; + + /* Other */ + const pcre_uint8 *fcc; + sljit_w lcc; + int mode; + int nltype; + int newline; + int bsr_nltype; + int endonly; + BOOL has_set_som; + sljit_w ctypes; + sljit_uw name_table; + sljit_w name_count; + sljit_w name_entry_size; + + /* Labels and jump lists. */ + struct sljit_label *partialmatchlabel; + struct sljit_label *leavelabel; + struct sljit_label *acceptlabel; + stub_list *stubs; + recurse_entry *entries; + recurse_entry *currententry; + jump_list *partialmatch; + jump_list *leave; + jump_list *accept; + jump_list *calllimit; + jump_list *stackalloc; + jump_list *revertframes; + jump_list *wordboundary; + jump_list *anynewline; + jump_list *hspace; + jump_list *vspace; + jump_list *casefulcmp; + jump_list *caselesscmp; + BOOL jscript_compat; +#ifdef SUPPORT_UTF + BOOL utf; +#ifdef SUPPORT_UCP + BOOL use_ucp; +#endif + jump_list *utfreadchar; +#ifdef COMPILE_PCRE8 + jump_list *utfreadtype8; +#endif +#endif /* SUPPORT_UTF */ +#ifdef SUPPORT_UCP + jump_list *getucd; +#endif +} compiler_common; + +/* For byte_sequence_compare. */ + +typedef struct compare_context { + int length; + int sourcereg; +#if defined SLJIT_UNALIGNED && SLJIT_UNALIGNED + int ucharptr; + union { + sljit_i asint; + sljit_uh asushort; +#ifdef COMPILE_PCRE8 + sljit_ub asbyte; + sljit_ub asuchars[4]; +#else +#ifdef COMPILE_PCRE16 + sljit_uh asuchars[2]; +#endif +#endif + } c; + union { + sljit_i asint; + sljit_uh asushort; +#ifdef COMPILE_PCRE8 + sljit_ub asbyte; + sljit_ub asuchars[4]; +#else +#ifdef COMPILE_PCRE16 + sljit_uh asuchars[2]; +#endif +#endif + } oc; +#endif +} compare_context; + +enum { + frame_end = 0, + frame_setstrbegin = -1, + frame_setmark = -2 +}; + +/* Undefine sljit macros. */ +#undef CMP + +/* Used for accessing the elements of the stack. */ +#define STACK(i) ((-(i) - 1) * (int)sizeof(sljit_w)) + +#define TMP1 SLJIT_TEMPORARY_REG1 +#define TMP2 SLJIT_TEMPORARY_REG3 +#define TMP3 SLJIT_TEMPORARY_EREG2 +#define STR_PTR SLJIT_SAVED_REG1 +#define STR_END SLJIT_SAVED_REG2 +#define STACK_TOP SLJIT_TEMPORARY_REG2 +#define STACK_LIMIT SLJIT_SAVED_REG3 +#define ARGUMENTS SLJIT_SAVED_EREG1 +#define CALL_COUNT SLJIT_SAVED_EREG2 +#define RETURN_ADDR SLJIT_TEMPORARY_EREG1 + +/* Locals layout. */ +/* These two locals can be used by the current opcode. */ +#define LOCALS0 (0 * sizeof(sljit_w)) +#define LOCALS1 (1 * sizeof(sljit_w)) +/* Two local variables for possessive quantifiers (char1 cannot use them). */ +#define POSSESSIVE0 (2 * sizeof(sljit_w)) +#define POSSESSIVE1 (3 * sizeof(sljit_w)) +/* Max limit of recursions. */ +#define CALL_LIMIT (4 * sizeof(sljit_w)) +/* The output vector is stored on the stack, and contains pointers +to characters. The vector data is divided into two groups: the first +group contains the start / end character pointers, and the second is +the start pointers when the end of the capturing group has not yet reached. */ +#define OVECTOR_START (common->ovector_start) +#define OVECTOR(i) (OVECTOR_START + (i) * sizeof(sljit_w)) +#define OVECTOR_PRIV(i) (common->cbraptr + (i) * sizeof(sljit_w)) +#define PRIV_DATA(cc) (common->localptrs[(cc) - common->start]) + +#ifdef COMPILE_PCRE8 +#define MOV_UCHAR SLJIT_MOV_UB +#define MOVU_UCHAR SLJIT_MOVU_UB +#else +#ifdef COMPILE_PCRE16 +#define MOV_UCHAR SLJIT_MOV_UH +#define MOVU_UCHAR SLJIT_MOVU_UH +#else +#error Unsupported compiling mode +#endif +#endif + +/* Shortcuts. */ +#define DEFINE_COMPILER \ + struct sljit_compiler *compiler = common->compiler +#define OP1(op, dst, dstw, src, srcw) \ + sljit_emit_op1(compiler, (op), (dst), (dstw), (src), (srcw)) +#define OP2(op, dst, dstw, src1, src1w, src2, src2w) \ + sljit_emit_op2(compiler, (op), (dst), (dstw), (src1), (src1w), (src2), (src2w)) +#define LABEL() \ + sljit_emit_label(compiler) +#define JUMP(type) \ + sljit_emit_jump(compiler, (type)) +#define JUMPTO(type, label) \ + sljit_set_label(sljit_emit_jump(compiler, (type)), (label)) +#define JUMPHERE(jump) \ + sljit_set_label((jump), sljit_emit_label(compiler)) +#define CMP(type, src1, src1w, src2, src2w) \ + sljit_emit_cmp(compiler, (type), (src1), (src1w), (src2), (src2w)) +#define CMPTO(type, src1, src1w, src2, src2w, label) \ + sljit_set_label(sljit_emit_cmp(compiler, (type), (src1), (src1w), (src2), (src2w)), (label)) +#define COND_VALUE(op, dst, dstw, type) \ + sljit_emit_cond_value(compiler, (op), (dst), (dstw), (type)) +#define GET_LOCAL_BASE(dst, dstw, offset) \ + sljit_get_local_base(compiler, (dst), (dstw), (offset)) + +static pcre_uchar* bracketend(pcre_uchar* cc) +{ +SLJIT_ASSERT((*cc >= OP_ASSERT && *cc <= OP_ASSERTBACK_NOT) || (*cc >= OP_ONCE && *cc <= OP_SCOND)); +do cc += GET(cc, 1); while (*cc == OP_ALT); +SLJIT_ASSERT(*cc >= OP_KET && *cc <= OP_KETRPOS); +cc += 1 + LINK_SIZE; +return cc; +} + +/* Functions whose might need modification for all new supported opcodes: + next_opcode + get_localspace + set_localptrs + get_framesize + init_frame + get_localsize + copy_locals + compile_trypath + compile_backtrackpath +*/ + +static pcre_uchar *next_opcode(compiler_common *common, pcre_uchar *cc) +{ +SLJIT_UNUSED_ARG(common); +switch(*cc) + { + case OP_SOD: + case OP_SOM: + case OP_SET_SOM: + case OP_NOT_WORD_BOUNDARY: + case OP_WORD_BOUNDARY: + case OP_NOT_DIGIT: + case OP_DIGIT: + case OP_NOT_WHITESPACE: + case OP_WHITESPACE: + case OP_NOT_WORDCHAR: + case OP_WORDCHAR: + case OP_ANY: + case OP_ALLANY: + case OP_ANYNL: + case OP_NOT_HSPACE: + case OP_HSPACE: + case OP_NOT_VSPACE: + case OP_VSPACE: + case OP_EXTUNI: + case OP_EODN: + case OP_EOD: + case OP_CIRC: + case OP_CIRCM: + case OP_DOLL: + case OP_DOLLM: + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSSTAR: + case OP_TYPEPOSPLUS: + case OP_TYPEPOSQUERY: + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRPLUS: + case OP_CRMINPLUS: + case OP_CRQUERY: + case OP_CRMINQUERY: + case OP_DEF: + case OP_BRAZERO: + case OP_BRAMINZERO: + case OP_BRAPOSZERO: + case OP_COMMIT: + case OP_FAIL: + case OP_ACCEPT: + case OP_ASSERT_ACCEPT: + case OP_SKIPZERO: + return cc + 1; + + case OP_ANYBYTE: +#ifdef SUPPORT_UTF + if (common->utf) return NULL; +#endif + return cc + 1; + + case OP_CHAR: + case OP_CHARI: + case OP_NOT: + case OP_NOTI: + case OP_STAR: + case OP_MINSTAR: + case OP_PLUS: + case OP_MINPLUS: + case OP_QUERY: + case OP_MINQUERY: + case OP_POSSTAR: + case OP_POSPLUS: + case OP_POSQUERY: + case OP_STARI: + case OP_MINSTARI: + case OP_PLUSI: + case OP_MINPLUSI: + case OP_QUERYI: + case OP_MINQUERYI: + case OP_POSSTARI: + case OP_POSPLUSI: + case OP_POSQUERYI: + case OP_NOTSTAR: + case OP_NOTMINSTAR: + case OP_NOTPLUS: + case OP_NOTMINPLUS: + case OP_NOTQUERY: + case OP_NOTMINQUERY: + case OP_NOTPOSSTAR: + case OP_NOTPOSPLUS: + case OP_NOTPOSQUERY: + case OP_NOTSTARI: + case OP_NOTMINSTARI: + case OP_NOTPLUSI: + case OP_NOTMINPLUSI: + case OP_NOTQUERYI: + case OP_NOTMINQUERYI: + case OP_NOTPOSSTARI: + case OP_NOTPOSPLUSI: + case OP_NOTPOSQUERYI: + cc += 2; +#ifdef SUPPORT_UTF + if (common->utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]); +#endif + return cc; + + case OP_UPTO: + case OP_MINUPTO: + case OP_EXACT: + case OP_POSUPTO: + case OP_UPTOI: + case OP_MINUPTOI: + case OP_EXACTI: + case OP_POSUPTOI: + case OP_NOTUPTO: + case OP_NOTMINUPTO: + case OP_NOTEXACT: + case OP_NOTPOSUPTO: + case OP_NOTUPTOI: + case OP_NOTMINUPTOI: + case OP_NOTEXACTI: + case OP_NOTPOSUPTOI: + cc += 2 + IMM2_SIZE; +#ifdef SUPPORT_UTF + if (common->utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]); +#endif + return cc; + + case OP_NOTPROP: + case OP_PROP: + return cc + 1 + 2; + + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEEXACT: + case OP_TYPEPOSUPTO: + case OP_REF: + case OP_REFI: + case OP_CREF: + case OP_NCREF: + case OP_RREF: + case OP_NRREF: + case OP_CLOSE: + cc += 1 + IMM2_SIZE; + return cc; + + case OP_CRRANGE: + case OP_CRMINRANGE: + return cc + 1 + 2 * IMM2_SIZE; + + case OP_CLASS: + case OP_NCLASS: + return cc + 1 + 32 / sizeof(pcre_uchar); + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: + return cc + GET(cc, 1); +#endif + + case OP_RECURSE: + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + case OP_REVERSE: + case OP_ONCE: + case OP_ONCE_NC: + case OP_BRA: + case OP_BRAPOS: + case OP_COND: + case OP_SBRA: + case OP_SBRAPOS: + case OP_SCOND: + case OP_ALT: + case OP_KET: + case OP_KETRMAX: + case OP_KETRMIN: + case OP_KETRPOS: + return cc + 1 + LINK_SIZE; + + case OP_CBRA: + case OP_CBRAPOS: + case OP_SCBRA: + case OP_SCBRAPOS: + return cc + 1 + LINK_SIZE + IMM2_SIZE; + + case OP_MARK: + return cc + 1 + 2 + cc[1]; + + default: + return NULL; + } +} + +static int get_localspace(compiler_common *common, pcre_uchar *cc, pcre_uchar *ccend) +{ +int localspace = 0; +pcre_uchar *alternative; +/* Calculate important variables (like stack size) and checks whether all opcodes are supported. */ +while (cc < ccend) + { + switch(*cc) + { + case OP_SET_SOM: + common->has_set_som = TRUE; + cc += 1; + break; + + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + case OP_ONCE: + case OP_ONCE_NC: + case OP_BRAPOS: + case OP_SBRA: + case OP_SBRAPOS: + case OP_SCOND: + localspace += sizeof(sljit_w); + cc += 1 + LINK_SIZE; + break; + + case OP_CBRAPOS: + case OP_SCBRAPOS: + localspace += sizeof(sljit_w); + cc += 1 + LINK_SIZE + IMM2_SIZE; + break; + + case OP_COND: + /* Might be a hidden SCOND. */ + alternative = cc + GET(cc, 1); + if (*alternative == OP_KETRMAX || *alternative == OP_KETRMIN) + localspace += sizeof(sljit_w); + cc += 1 + LINK_SIZE; + break; + + case OP_RECURSE: + /* Set its value only once. */ + if (common->recursive_head == 0) + { + common->recursive_head = common->ovector_start; + common->ovector_start += sizeof(sljit_w); + } + cc += 1 + LINK_SIZE; + break; + + case OP_MARK: + if (common->mark_ptr == 0) + { + common->mark_ptr = common->ovector_start; + common->ovector_start += sizeof(sljit_w); + } + cc += 1 + 2 + cc[1]; + break; + + default: + cc = next_opcode(common, cc); + if (cc == NULL) + return -1; + break; + } + } +return localspace; +} + +static void set_localptrs(compiler_common *common, int localptr, pcre_uchar *ccend) +{ +pcre_uchar *cc = common->start; +pcre_uchar *alternative; +while (cc < ccend) + { + switch(*cc) + { + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + case OP_ONCE: + case OP_ONCE_NC: + case OP_BRAPOS: + case OP_SBRA: + case OP_SBRAPOS: + case OP_SCOND: + common->localptrs[cc - common->start] = localptr; + localptr += sizeof(sljit_w); + cc += 1 + LINK_SIZE; + break; + + case OP_CBRAPOS: + case OP_SCBRAPOS: + common->localptrs[cc - common->start] = localptr; + localptr += sizeof(sljit_w); + cc += 1 + LINK_SIZE + IMM2_SIZE; + break; + + case OP_COND: + /* Might be a hidden SCOND. */ + alternative = cc + GET(cc, 1); + if (*alternative == OP_KETRMAX || *alternative == OP_KETRMIN) + { + common->localptrs[cc - common->start] = localptr; + localptr += sizeof(sljit_w); + } + cc += 1 + LINK_SIZE; + break; + + default: + cc = next_opcode(common, cc); + SLJIT_ASSERT(cc != NULL); + break; + } + } +} + +/* Returns with -1 if no need for frame. */ +static int get_framesize(compiler_common *common, pcre_uchar *cc, BOOL recursive) +{ +pcre_uchar *ccend = bracketend(cc); +int length = 0; +BOOL possessive = FALSE; +BOOL setsom_found = recursive; +BOOL setmark_found = recursive; + +if (!recursive && (*cc == OP_CBRAPOS || *cc == OP_SCBRAPOS)) + { + length = 3; + possessive = TRUE; + } + +cc = next_opcode(common, cc); +SLJIT_ASSERT(cc != NULL); +while (cc < ccend) + switch(*cc) + { + case OP_SET_SOM: + SLJIT_ASSERT(common->has_set_som); + if (!setsom_found) + { + length += 2; + setsom_found = TRUE; + } + cc += 1; + break; + + case OP_MARK: + SLJIT_ASSERT(common->mark_ptr != 0); + if (!setmark_found) + { + length += 2; + setmark_found = TRUE; + } + cc += 1 + 2 + cc[1]; + break; + + case OP_RECURSE: + if (common->has_set_som && !setsom_found) + { + length += 2; + setsom_found = TRUE; + } + if (common->mark_ptr != 0 && !setmark_found) + { + length += 2; + setmark_found = TRUE; + } + cc += 1 + LINK_SIZE; + break; + + case OP_CBRA: + case OP_CBRAPOS: + case OP_SCBRA: + case OP_SCBRAPOS: + length += 3; + cc += 1 + LINK_SIZE + IMM2_SIZE; + break; + + default: + cc = next_opcode(common, cc); + SLJIT_ASSERT(cc != NULL); + break; + } + +/* Possessive quantifiers can use a special case. */ +if (SLJIT_UNLIKELY(possessive) && length == 3) + return -1; + +if (length > 0) + return length + 1; +return -1; +} + +static void init_frame(compiler_common *common, pcre_uchar *cc, int stackpos, int stacktop, BOOL recursive) +{ +DEFINE_COMPILER; +pcre_uchar *ccend = bracketend(cc); +BOOL setsom_found = recursive; +BOOL setmark_found = recursive; +int offset; + +/* >= 1 + shortest item size (2) */ +SLJIT_UNUSED_ARG(stacktop); +SLJIT_ASSERT(stackpos >= stacktop + 2); + +stackpos = STACK(stackpos); +if (recursive || (*cc != OP_CBRAPOS && *cc != OP_SCBRAPOS)) + cc = next_opcode(common, cc); +SLJIT_ASSERT(cc != NULL); +while (cc < ccend) + switch(*cc) + { + case OP_SET_SOM: + SLJIT_ASSERT(common->has_set_som); + if (!setsom_found) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackpos, SLJIT_IMM, frame_setstrbegin); + stackpos += (int)sizeof(sljit_w); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackpos, TMP1, 0); + stackpos += (int)sizeof(sljit_w); + setsom_found = TRUE; + } + cc += 1; + break; + + case OP_MARK: + SLJIT_ASSERT(common->mark_ptr != 0); + if (!setmark_found) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->mark_ptr); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackpos, SLJIT_IMM, frame_setmark); + stackpos += (int)sizeof(sljit_w); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackpos, TMP1, 0); + stackpos += (int)sizeof(sljit_w); + setmark_found = TRUE; + } + cc += 1 + 2 + cc[1]; + break; + + case OP_RECURSE: + if (common->has_set_som && !setsom_found) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackpos, SLJIT_IMM, frame_setstrbegin); + stackpos += (int)sizeof(sljit_w); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackpos, TMP1, 0); + stackpos += (int)sizeof(sljit_w); + setsom_found = TRUE; + } + if (common->mark_ptr != 0 && !setmark_found) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->mark_ptr); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackpos, SLJIT_IMM, frame_setmark); + stackpos += (int)sizeof(sljit_w); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackpos, TMP1, 0); + stackpos += (int)sizeof(sljit_w); + setmark_found = TRUE; + } + cc += 1 + LINK_SIZE; + break; + + case OP_CBRA: + case OP_CBRAPOS: + case OP_SCBRA: + case OP_SCBRAPOS: + offset = (GET2(cc, 1 + LINK_SIZE)) << 1; + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackpos, SLJIT_IMM, OVECTOR(offset)); + stackpos += (int)sizeof(sljit_w); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset)); + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackpos, TMP1, 0); + stackpos += (int)sizeof(sljit_w); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackpos, TMP2, 0); + stackpos += (int)sizeof(sljit_w); + + cc += 1 + LINK_SIZE + IMM2_SIZE; + break; + + default: + cc = next_opcode(common, cc); + SLJIT_ASSERT(cc != NULL); + break; + } + +OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackpos, SLJIT_IMM, frame_end); +SLJIT_ASSERT(stackpos == STACK(stacktop)); +} + +static SLJIT_INLINE int get_localsize(compiler_common *common, pcre_uchar *cc, pcre_uchar *ccend) +{ +int localsize = 2; +pcre_uchar *alternative; +/* Calculate the sum of the local variables. */ +while (cc < ccend) + { + switch(*cc) + { + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + case OP_ONCE: + case OP_ONCE_NC: + case OP_BRAPOS: + case OP_SBRA: + case OP_SBRAPOS: + case OP_SCOND: + localsize++; + cc += 1 + LINK_SIZE; + break; + + case OP_CBRA: + case OP_SCBRA: + localsize++; + cc += 1 + LINK_SIZE + IMM2_SIZE; + break; + + case OP_CBRAPOS: + case OP_SCBRAPOS: + localsize += 2; + cc += 1 + LINK_SIZE + IMM2_SIZE; + break; + + case OP_COND: + /* Might be a hidden SCOND. */ + alternative = cc + GET(cc, 1); + if (*alternative == OP_KETRMAX || *alternative == OP_KETRMIN) + localsize++; + cc += 1 + LINK_SIZE; + break; + + default: + cc = next_opcode(common, cc); + SLJIT_ASSERT(cc != NULL); + break; + } + } +SLJIT_ASSERT(cc == ccend); +return localsize; +} + +static void copy_locals(compiler_common *common, pcre_uchar *cc, pcre_uchar *ccend, + BOOL save, int stackptr, int stacktop) +{ +DEFINE_COMPILER; +int srcw[2]; +int count; +BOOL tmp1next = TRUE; +BOOL tmp1empty = TRUE; +BOOL tmp2empty = TRUE; +pcre_uchar *alternative; +enum { + start, + loop, + end +} status; + +status = save ? start : loop; +stackptr = STACK(stackptr - 2); +stacktop = STACK(stacktop - 1); + +if (!save) + { + stackptr += sizeof(sljit_w); + if (stackptr < stacktop) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), stackptr); + stackptr += sizeof(sljit_w); + tmp1empty = FALSE; + } + if (stackptr < stacktop) + { + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(STACK_TOP), stackptr); + stackptr += sizeof(sljit_w); + tmp2empty = FALSE; + } + /* The tmp1next must be TRUE in either way. */ + } + +while (status != end) + { + count = 0; + switch(status) + { + case start: + SLJIT_ASSERT(save && common->recursive_head != 0); + count = 1; + srcw[0] = common->recursive_head; + status = loop; + break; + + case loop: + if (cc >= ccend) + { + status = end; + break; + } + + switch(*cc) + { + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + case OP_ONCE: + case OP_ONCE_NC: + case OP_BRAPOS: + case OP_SBRA: + case OP_SBRAPOS: + case OP_SCOND: + count = 1; + srcw[0] = PRIV_DATA(cc); + SLJIT_ASSERT(srcw[0] != 0); + cc += 1 + LINK_SIZE; + break; + + case OP_CBRA: + case OP_SCBRA: + count = 1; + srcw[0] = OVECTOR_PRIV(GET2(cc, 1 + LINK_SIZE)); + cc += 1 + LINK_SIZE + IMM2_SIZE; + break; + + case OP_CBRAPOS: + case OP_SCBRAPOS: + count = 2; + srcw[1] = OVECTOR_PRIV(GET2(cc, 1 + LINK_SIZE)); + srcw[0] = PRIV_DATA(cc); + SLJIT_ASSERT(srcw[0] != 0); + cc += 1 + LINK_SIZE + IMM2_SIZE; + break; + + case OP_COND: + /* Might be a hidden SCOND. */ + alternative = cc + GET(cc, 1); + if (*alternative == OP_KETRMAX || *alternative == OP_KETRMIN) + { + count = 1; + srcw[0] = PRIV_DATA(cc); + SLJIT_ASSERT(srcw[0] != 0); + } + cc += 1 + LINK_SIZE; + break; + + default: + cc = next_opcode(common, cc); + SLJIT_ASSERT(cc != NULL); + break; + } + break; + + case end: + SLJIT_ASSERT_STOP(); + break; + } + + while (count > 0) + { + count--; + if (save) + { + if (tmp1next) + { + if (!tmp1empty) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackptr, TMP1, 0); + stackptr += sizeof(sljit_w); + } + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), srcw[count]); + tmp1empty = FALSE; + tmp1next = FALSE; + } + else + { + if (!tmp2empty) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackptr, TMP2, 0); + stackptr += sizeof(sljit_w); + } + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), srcw[count]); + tmp2empty = FALSE; + tmp1next = TRUE; + } + } + else + { + if (tmp1next) + { + SLJIT_ASSERT(!tmp1empty); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), srcw[count], TMP1, 0); + tmp1empty = stackptr >= stacktop; + if (!tmp1empty) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), stackptr); + stackptr += sizeof(sljit_w); + } + tmp1next = FALSE; + } + else + { + SLJIT_ASSERT(!tmp2empty); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), srcw[count], TMP2, 0); + tmp2empty = stackptr >= stacktop; + if (!tmp2empty) + { + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(STACK_TOP), stackptr); + stackptr += sizeof(sljit_w); + } + tmp1next = TRUE; + } + } + } + } + +if (save) + { + if (tmp1next) + { + if (!tmp1empty) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackptr, TMP1, 0); + stackptr += sizeof(sljit_w); + } + if (!tmp2empty) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackptr, TMP2, 0); + stackptr += sizeof(sljit_w); + } + } + else + { + if (!tmp2empty) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackptr, TMP2, 0); + stackptr += sizeof(sljit_w); + } + if (!tmp1empty) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), stackptr, TMP1, 0); + stackptr += sizeof(sljit_w); + } + } + } +SLJIT_ASSERT(cc == ccend && stackptr == stacktop && (save || (tmp1empty && tmp2empty))); +} + +static SLJIT_INLINE BOOL ispowerof2(unsigned int value) +{ +return (value & (value - 1)) == 0; +} + +static SLJIT_INLINE void set_jumps(jump_list *list, struct sljit_label *label) +{ +while (list) + { + /* sljit_set_label is clever enough to do nothing + if either the jump or the label is NULL */ + sljit_set_label(list->jump, label); + list = list->next; + } +} + +static SLJIT_INLINE void add_jump(struct sljit_compiler *compiler, jump_list **list, struct sljit_jump* jump) +{ +jump_list *list_item = sljit_alloc_memory(compiler, sizeof(jump_list)); +if (list_item) + { + list_item->next = *list; + list_item->jump = jump; + *list = list_item; + } +} + +static void add_stub(compiler_common *common, enum stub_types type, int data, struct sljit_jump *start) +{ +DEFINE_COMPILER; +stub_list* list_item = sljit_alloc_memory(compiler, sizeof(stub_list)); + +if (list_item) + { + list_item->type = type; + list_item->data = data; + list_item->start = start; + list_item->leave = LABEL(); + list_item->next = common->stubs; + common->stubs = list_item; + } +} + +static void flush_stubs(compiler_common *common) +{ +DEFINE_COMPILER; +stub_list* list_item = common->stubs; + +while (list_item) + { + JUMPHERE(list_item->start); + switch(list_item->type) + { + case stack_alloc: + add_jump(compiler, &common->stackalloc, JUMP(SLJIT_FAST_CALL)); + break; + } + JUMPTO(SLJIT_JUMP, list_item->leave); + list_item = list_item->next; + } +common->stubs = NULL; +} + +static SLJIT_INLINE void decrease_call_count(compiler_common *common) +{ +DEFINE_COMPILER; + +OP2(SLJIT_SUB | SLJIT_SET_E, CALL_COUNT, 0, CALL_COUNT, 0, SLJIT_IMM, 1); +add_jump(compiler, &common->calllimit, JUMP(SLJIT_C_ZERO)); +} + +static SLJIT_INLINE void allocate_stack(compiler_common *common, int size) +{ +/* May destroy all locals and registers except TMP2. */ +DEFINE_COMPILER; + +OP2(SLJIT_ADD, STACK_TOP, 0, STACK_TOP, 0, SLJIT_IMM, size * sizeof(sljit_w)); +#ifdef DESTROY_REGISTERS +OP1(SLJIT_MOV, TMP1, 0, SLJIT_IMM, 12345); +OP1(SLJIT_MOV, TMP3, 0, TMP1, 0); +OP1(SLJIT_MOV, RETURN_ADDR, 0, TMP1, 0); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0, TMP1, 0); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1, TMP1, 0); +#endif +add_stub(common, stack_alloc, 0, CMP(SLJIT_C_GREATER, STACK_TOP, 0, STACK_LIMIT, 0)); +} + +static SLJIT_INLINE void free_stack(compiler_common *common, int size) +{ +DEFINE_COMPILER; +OP2(SLJIT_SUB, STACK_TOP, 0, STACK_TOP, 0, SLJIT_IMM, size * sizeof(sljit_w)); +} + +static SLJIT_INLINE void reset_ovector(compiler_common *common, int length) +{ +DEFINE_COMPILER; +struct sljit_label *loop; +int i; +/* At this point we can freely use all temporary registers. */ +/* TMP1 returns with begin - 1. */ +OP2(SLJIT_SUB, SLJIT_TEMPORARY_REG1, 0, SLJIT_MEM1(SLJIT_SAVED_REG1), SLJIT_OFFSETOF(jit_arguments, begin), SLJIT_IMM, IN_UCHARS(1)); +if (length < 8) + { + for (i = 0; i < length; i++) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(i), SLJIT_TEMPORARY_REG1, 0); + } +else + { + GET_LOCAL_BASE(SLJIT_TEMPORARY_REG2, 0, OVECTOR_START - sizeof(sljit_w)); + OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG3, 0, SLJIT_IMM, length); + loop = LABEL(); + OP1(SLJIT_MOVU, SLJIT_MEM1(SLJIT_TEMPORARY_REG2), sizeof(sljit_w), SLJIT_TEMPORARY_REG1, 0); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_TEMPORARY_REG3, 0, SLJIT_TEMPORARY_REG3, 0, SLJIT_IMM, 1); + JUMPTO(SLJIT_C_NOT_ZERO, loop); + } +} + +static SLJIT_INLINE void copy_ovector(compiler_common *common, int topbracket) +{ +DEFINE_COMPILER; +struct sljit_label *loop; +struct sljit_jump *earlyexit; + +/* At this point we can freely use all registers. */ +OP1(SLJIT_MOV, SLJIT_SAVED_REG3, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(1)); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(1), STR_PTR, 0); + +OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG1, 0, ARGUMENTS, 0); +if (common->mark_ptr != 0) + OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG3, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->mark_ptr); +OP1(SLJIT_MOV_SI, SLJIT_TEMPORARY_REG2, 0, SLJIT_MEM1(SLJIT_TEMPORARY_REG1), SLJIT_OFFSETOF(jit_arguments, offsetcount)); +if (common->mark_ptr != 0) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_TEMPORARY_REG1), SLJIT_OFFSETOF(jit_arguments, mark_ptr), SLJIT_TEMPORARY_REG3, 0); +OP2(SLJIT_SUB, SLJIT_TEMPORARY_REG3, 0, SLJIT_MEM1(SLJIT_TEMPORARY_REG1), SLJIT_OFFSETOF(jit_arguments, offsets), SLJIT_IMM, sizeof(int)); +OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG1, 0, SLJIT_MEM1(SLJIT_TEMPORARY_REG1), SLJIT_OFFSETOF(jit_arguments, begin)); +GET_LOCAL_BASE(SLJIT_SAVED_REG1, 0, OVECTOR_START); +/* Unlikely, but possible */ +earlyexit = CMP(SLJIT_C_EQUAL, SLJIT_TEMPORARY_REG2, 0, SLJIT_IMM, 0); +loop = LABEL(); +OP2(SLJIT_SUB, SLJIT_SAVED_REG2, 0, SLJIT_MEM1(SLJIT_SAVED_REG1), 0, SLJIT_TEMPORARY_REG1, 0); +OP2(SLJIT_ADD, SLJIT_SAVED_REG1, 0, SLJIT_SAVED_REG1, 0, SLJIT_IMM, sizeof(sljit_w)); +/* Copy the integer value to the output buffer */ +#ifdef COMPILE_PCRE16 +OP2(SLJIT_ASHR, SLJIT_SAVED_REG2, 0, SLJIT_SAVED_REG2, 0, SLJIT_IMM, 1); +#endif +OP1(SLJIT_MOVU_SI, SLJIT_MEM1(SLJIT_TEMPORARY_REG3), sizeof(int), SLJIT_SAVED_REG2, 0); +OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_TEMPORARY_REG2, 0, SLJIT_TEMPORARY_REG2, 0, SLJIT_IMM, 1); +JUMPTO(SLJIT_C_NOT_ZERO, loop); +JUMPHERE(earlyexit); + +/* Calculate the return value, which is the maximum ovector value. */ +if (topbracket > 1) + { + GET_LOCAL_BASE(SLJIT_TEMPORARY_REG1, 0, OVECTOR_START + topbracket * 2 * sizeof(sljit_w)); + OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG2, 0, SLJIT_IMM, topbracket + 1); + + /* OVECTOR(0) is never equal to SLJIT_SAVED_REG3. */ + loop = LABEL(); + OP1(SLJIT_MOVU, SLJIT_TEMPORARY_REG3, 0, SLJIT_MEM1(SLJIT_TEMPORARY_REG1), -(2 * (sljit_w)sizeof(sljit_w))); + OP2(SLJIT_SUB, SLJIT_TEMPORARY_REG2, 0, SLJIT_TEMPORARY_REG2, 0, SLJIT_IMM, 1); + CMPTO(SLJIT_C_EQUAL, SLJIT_TEMPORARY_REG3, 0, SLJIT_SAVED_REG3, 0, loop); + OP1(SLJIT_MOV, SLJIT_RETURN_REG, 0, SLJIT_TEMPORARY_REG2, 0); + } +else + OP1(SLJIT_MOV, SLJIT_RETURN_REG, 0, SLJIT_IMM, 1); +} + +static SLJIT_INLINE void return_with_partial_match(compiler_common *common, struct sljit_label *leave) +{ +DEFINE_COMPILER; + +SLJIT_COMPILE_ASSERT(STR_END == SLJIT_SAVED_REG2, str_end_must_be_saved_reg2); +SLJIT_ASSERT(common->start_used_ptr != 0 && (common->mode == JIT_PARTIAL_SOFT_COMPILE ? common->hit_start != 0 : common->hit_start == 0)); + +OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG2, 0, ARGUMENTS, 0); +OP1(SLJIT_MOV, SLJIT_RETURN_REG, 0, SLJIT_IMM, PCRE_ERROR_PARTIAL); +OP1(SLJIT_MOV_SI, SLJIT_TEMPORARY_REG3, 0, SLJIT_MEM1(SLJIT_TEMPORARY_REG2), SLJIT_OFFSETOF(jit_arguments, offsetcount)); +CMPTO(SLJIT_C_LESS, SLJIT_TEMPORARY_REG3, 0, SLJIT_IMM, 2, leave); + +/* Store match begin and end. */ +OP1(SLJIT_MOV, SLJIT_SAVED_REG1, 0, SLJIT_MEM1(SLJIT_TEMPORARY_REG2), SLJIT_OFFSETOF(jit_arguments, begin)); +OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG2, 0, SLJIT_MEM1(SLJIT_TEMPORARY_REG2), SLJIT_OFFSETOF(jit_arguments, offsets)); +OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG3, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->mode == JIT_PARTIAL_HARD_COMPILE ? common->start_used_ptr : common->hit_start); +OP2(SLJIT_SUB, SLJIT_SAVED_REG2, 0, STR_END, 0, SLJIT_SAVED_REG1, 0); +#ifdef COMPILE_PCRE16 +OP2(SLJIT_ASHR, SLJIT_SAVED_REG2, 0, SLJIT_SAVED_REG2, 0, SLJIT_IMM, 1); +#endif +OP1(SLJIT_MOV_SI, SLJIT_MEM1(SLJIT_TEMPORARY_REG2), sizeof(int), SLJIT_SAVED_REG2, 0); + +OP2(SLJIT_SUB, SLJIT_TEMPORARY_REG3, 0, SLJIT_TEMPORARY_REG3, 0, SLJIT_SAVED_REG1, 0); +#ifdef COMPILE_PCRE16 +OP2(SLJIT_ASHR, SLJIT_TEMPORARY_REG3, 0, SLJIT_TEMPORARY_REG3, 0, SLJIT_IMM, 1); +#endif +OP1(SLJIT_MOV_SI, SLJIT_MEM1(SLJIT_TEMPORARY_REG2), 0, SLJIT_TEMPORARY_REG3, 0); + +JUMPTO(SLJIT_JUMP, leave); +} + +static SLJIT_INLINE void check_start_used_ptr(compiler_common *common) +{ +/* May destroy TMP1. */ +DEFINE_COMPILER; +struct sljit_jump *jump; + +if (common->mode == JIT_PARTIAL_SOFT_COMPILE) + { + /* The value of -1 must be kept for start_used_ptr! */ + OP2(SLJIT_ADD, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr, SLJIT_IMM, 1); + /* Jumps if start_used_ptr < STR_PTR, or start_used_ptr == -1. Although overwriting + is not necessary if start_used_ptr == STR_PTR, it does not hurt as well. */ + jump = CMP(SLJIT_C_LESS_EQUAL, TMP1, 0, STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr, STR_PTR, 0); + JUMPHERE(jump); + } +else if (common->mode == JIT_PARTIAL_HARD_COMPILE) + { + jump = CMP(SLJIT_C_LESS_EQUAL, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr, STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr, STR_PTR, 0); + JUMPHERE(jump); + } +} + +static SLJIT_INLINE BOOL char_has_othercase(compiler_common *common, pcre_uchar* cc) +{ +/* Detects if the character has an othercase. */ +unsigned int c; + +#ifdef SUPPORT_UTF +if (common->utf) + { + GETCHAR(c, cc); + if (c > 127) + { +#ifdef SUPPORT_UCP + return c != UCD_OTHERCASE(c); +#else + return FALSE; +#endif + } +#ifndef COMPILE_PCRE8 + return common->fcc[c] != c; +#endif + } +else +#endif + c = *cc; +return MAX_255(c) ? common->fcc[c] != c : FALSE; +} + +static SLJIT_INLINE unsigned int char_othercase(compiler_common *common, unsigned int c) +{ +/* Returns with the othercase. */ +#ifdef SUPPORT_UTF +if (common->utf && c > 127) + { +#ifdef SUPPORT_UCP + return UCD_OTHERCASE(c); +#else + return c; +#endif + } +#endif +return TABLE_GET(c, common->fcc, c); +} + +static unsigned int char_get_othercase_bit(compiler_common *common, pcre_uchar* cc) +{ +/* Detects if the character and its othercase has only 1 bit difference. */ +unsigned int c, oc, bit; +#if defined SUPPORT_UTF && defined COMPILE_PCRE8 +int n; +#endif + +#ifdef SUPPORT_UTF +if (common->utf) + { + GETCHAR(c, cc); + if (c <= 127) + oc = common->fcc[c]; + else + { +#ifdef SUPPORT_UCP + oc = UCD_OTHERCASE(c); +#else + oc = c; +#endif + } + } +else + { + c = *cc; + oc = TABLE_GET(c, common->fcc, c); + } +#else +c = *cc; +oc = TABLE_GET(c, common->fcc, c); +#endif + +SLJIT_ASSERT(c != oc); + +bit = c ^ oc; +/* Optimized for English alphabet. */ +if (c <= 127 && bit == 0x20) + return (0 << 8) | 0x20; + +/* Since c != oc, they must have at least 1 bit difference. */ +if (!ispowerof2(bit)) + return 0; + +#ifdef COMPILE_PCRE8 + +#ifdef SUPPORT_UTF +if (common->utf && c > 127) + { + n = GET_EXTRALEN(*cc); + while ((bit & 0x3f) == 0) + { + n--; + bit >>= 6; + } + return (n << 8) | bit; + } +#endif /* SUPPORT_UTF */ +return (0 << 8) | bit; + +#else /* COMPILE_PCRE8 */ + +#ifdef COMPILE_PCRE16 +#ifdef SUPPORT_UTF +if (common->utf && c > 65535) + { + if (bit >= (1 << 10)) + bit >>= 10; + else + return (bit < 256) ? ((2 << 8) | bit) : ((3 << 8) | (bit >> 8)); + } +#endif /* SUPPORT_UTF */ +return (bit < 256) ? ((0 << 8) | bit) : ((1 << 8) | (bit >> 8)); +#endif /* COMPILE_PCRE16 */ + +#endif /* COMPILE_PCRE8 */ +} + +static void check_partial(compiler_common *common, BOOL force) +{ +/* Checks whether a partial matching is occured. Does not modify registers. */ +DEFINE_COMPILER; +struct sljit_jump *jump = NULL; + +SLJIT_ASSERT(!force || common->mode != JIT_COMPILE); + +if (common->mode == JIT_COMPILE) + return; + +if (!force) + jump = CMP(SLJIT_C_GREATER_EQUAL, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr, STR_PTR, 0); +else if (common->mode == JIT_PARTIAL_SOFT_COMPILE) + jump = CMP(SLJIT_C_EQUAL, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr, SLJIT_IMM, -1); + +if (common->mode == JIT_PARTIAL_SOFT_COMPILE) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->hit_start, SLJIT_IMM, -1); +else + { + if (common->partialmatchlabel != NULL) + JUMPTO(SLJIT_JUMP, common->partialmatchlabel); + else + add_jump(compiler, &common->partialmatch, JUMP(SLJIT_JUMP)); + } + +if (jump != NULL) + JUMPHERE(jump); +} + +static struct sljit_jump *check_str_end(compiler_common *common) +{ +/* Does not affect registers. Usually used in a tight spot. */ +DEFINE_COMPILER; +struct sljit_jump *jump; +struct sljit_jump *nohit; +struct sljit_jump *return_value; + +if (common->mode == JIT_COMPILE) + return CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); + +jump = CMP(SLJIT_C_LESS, STR_PTR, 0, STR_END, 0); +if (common->mode == JIT_PARTIAL_SOFT_COMPILE) + { + nohit = CMP(SLJIT_C_GREATER_EQUAL, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr, STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->hit_start, SLJIT_IMM, -1); + JUMPHERE(nohit); + return_value = JUMP(SLJIT_JUMP); + } +else + { + return_value = CMP(SLJIT_C_GREATER_EQUAL, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr, STR_PTR, 0); + if (common->partialmatchlabel != NULL) + JUMPTO(SLJIT_JUMP, common->partialmatchlabel); + else + add_jump(compiler, &common->partialmatch, JUMP(SLJIT_JUMP)); + } +JUMPHERE(jump); +return return_value; +} + +static void detect_partial_match(compiler_common *common, jump_list **backtracks) +{ +DEFINE_COMPILER; +struct sljit_jump *jump; + +if (common->mode == JIT_COMPILE) + { + add_jump(compiler, backtracks, CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0)); + return; + } + +/* Partial matching mode. */ +jump = CMP(SLJIT_C_LESS, STR_PTR, 0, STR_END, 0); +add_jump(compiler, backtracks, CMP(SLJIT_C_GREATER_EQUAL, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr, STR_PTR, 0)); +if (common->mode == JIT_PARTIAL_SOFT_COMPILE) + { + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->hit_start, SLJIT_IMM, -1); + add_jump(compiler, backtracks, JUMP(SLJIT_JUMP)); + } +else + { + if (common->partialmatchlabel != NULL) + JUMPTO(SLJIT_JUMP, common->partialmatchlabel); + else + add_jump(compiler, &common->partialmatch, JUMP(SLJIT_JUMP)); + } +JUMPHERE(jump); +} + +static void read_char(compiler_common *common) +{ +/* Reads the character into TMP1, updates STR_PTR. +Does not check STR_END. TMP2 Destroyed. */ +DEFINE_COMPILER; +#ifdef SUPPORT_UTF +struct sljit_jump *jump; +#endif + +OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), 0); +#ifdef SUPPORT_UTF +if (common->utf) + { +#ifdef COMPILE_PCRE8 + jump = CMP(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xc0); +#else +#ifdef COMPILE_PCRE16 + jump = CMP(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xd800); +#endif +#endif /* COMPILE_PCRE8 */ + add_jump(compiler, &common->utfreadchar, JUMP(SLJIT_FAST_CALL)); + JUMPHERE(jump); + } +#endif +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +} + +static void peek_char(compiler_common *common) +{ +/* Reads the character into TMP1, keeps STR_PTR. +Does not check STR_END. TMP2 Destroyed. */ +DEFINE_COMPILER; +#ifdef SUPPORT_UTF +struct sljit_jump *jump; +#endif + +OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), 0); +#ifdef SUPPORT_UTF +if (common->utf) + { +#ifdef COMPILE_PCRE8 + jump = CMP(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xc0); +#else +#ifdef COMPILE_PCRE16 + jump = CMP(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xd800); +#endif +#endif /* COMPILE_PCRE8 */ + add_jump(compiler, &common->utfreadchar, JUMP(SLJIT_FAST_CALL)); + OP2(SLJIT_SUB, STR_PTR, 0, STR_PTR, 0, TMP2, 0); + JUMPHERE(jump); + } +#endif +} + +static void read_char8_type(compiler_common *common) +{ +/* Reads the character type into TMP1, updates STR_PTR. Does not check STR_END. */ +DEFINE_COMPILER; +#if defined SUPPORT_UTF || defined COMPILE_PCRE16 +struct sljit_jump *jump; +#endif + +#ifdef SUPPORT_UTF +if (common->utf) + { + OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), 0); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +#ifdef COMPILE_PCRE8 + /* This can be an extra read in some situations, but hopefully + it is needed in most cases. */ + OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP2), common->ctypes); + jump = CMP(SLJIT_C_LESS, TMP2, 0, SLJIT_IMM, 0xc0); + add_jump(compiler, &common->utfreadtype8, JUMP(SLJIT_FAST_CALL)); + JUMPHERE(jump); +#else +#ifdef COMPILE_PCRE16 + OP1(SLJIT_MOV, TMP1, 0, SLJIT_IMM, 0); + jump = CMP(SLJIT_C_GREATER, TMP2, 0, SLJIT_IMM, 255); + OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP2), common->ctypes); + JUMPHERE(jump); + /* Skip low surrogate if necessary. */ + OP2(SLJIT_AND, TMP2, 0, TMP2, 0, SLJIT_IMM, 0xfc00); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP2, 0, SLJIT_IMM, 0xd800); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SHL, TMP2, 0, TMP2, 0, SLJIT_IMM, 1); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP2, 0); +#endif +#endif /* COMPILE_PCRE8 */ + return; + } +#endif +OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), 0); +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +#ifdef COMPILE_PCRE16 +/* The ctypes array contains only 256 values. */ +OP1(SLJIT_MOV, TMP1, 0, SLJIT_IMM, 0); +jump = CMP(SLJIT_C_GREATER, TMP2, 0, SLJIT_IMM, 255); +#endif +OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP2), common->ctypes); +#ifdef COMPILE_PCRE16 +JUMPHERE(jump); +#endif +} + +static void skip_char_back(compiler_common *common) +{ +/* Goes one character back. Affects STR_PTR and TMP1. Does not check begin. */ +DEFINE_COMPILER; +#if defined SUPPORT_UTF && defined COMPILE_PCRE8 +struct sljit_label *label; + +if (common->utf) + { + label = LABEL(); + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), -IN_UCHARS(1)); + OP2(SLJIT_SUB, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, 0xc0); + CMPTO(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, 0x80, label); + return; + } +#endif +#if defined SUPPORT_UTF && defined COMPILE_PCRE16 +if (common->utf) + { + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), -IN_UCHARS(1)); + OP2(SLJIT_SUB, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + /* Skip low surrogate if necessary. */ + OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, 0xfc00); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0xdc00); + COND_VALUE(SLJIT_MOV, TMP1, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SHL, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + OP2(SLJIT_SUB, STR_PTR, 0, STR_PTR, 0, TMP1, 0); + return; + } +#endif +OP2(SLJIT_SUB, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +} + +static void check_newlinechar(compiler_common *common, int nltype, jump_list **backtracks, BOOL jumpiftrue) +{ +/* Character comes in TMP1. Checks if it is a newline. TMP2 may be destroyed. */ +DEFINE_COMPILER; + +if (nltype == NLTYPE_ANY) + { + add_jump(compiler, &common->anynewline, JUMP(SLJIT_FAST_CALL)); + add_jump(compiler, backtracks, JUMP(jumpiftrue ? SLJIT_C_NOT_ZERO : SLJIT_C_ZERO)); + } +else if (nltype == NLTYPE_ANYCRLF) + { + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, CHAR_CR); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, CHAR_NL); + COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_EQUAL); + add_jump(compiler, backtracks, JUMP(jumpiftrue ? SLJIT_C_NOT_ZERO : SLJIT_C_ZERO)); + } +else + { + SLJIT_ASSERT(nltype == NLTYPE_FIXED && common->newline < 256); + add_jump(compiler, backtracks, CMP(jumpiftrue ? SLJIT_C_EQUAL : SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, common->newline)); + } +} + +#ifdef SUPPORT_UTF + +#ifdef COMPILE_PCRE8 +static void do_utfreadchar(compiler_common *common) +{ +/* Fast decoding a UTF-8 character. TMP1 contains the first byte +of the character (>= 0xc0). Return char value in TMP1, length - 1 in TMP2. */ +DEFINE_COMPILER; +struct sljit_jump *jump; + +sljit_emit_fast_enter(compiler, RETURN_ADDR, 0); +/* Searching for the first zero. */ +OP2(SLJIT_AND | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x20); +jump = JUMP(SLJIT_C_NOT_ZERO); +/* Two byte sequence. */ +OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(1)); +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, 0x1f); +OP2(SLJIT_SHL, TMP1, 0, TMP1, 0, SLJIT_IMM, 6); +OP2(SLJIT_AND, TMP2, 0, TMP2, 0, SLJIT_IMM, 0x3f); +OP2(SLJIT_OR, TMP1, 0, TMP1, 0, TMP2, 0); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, IN_UCHARS(1)); +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); +JUMPHERE(jump); + +OP2(SLJIT_AND | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x10); +jump = JUMP(SLJIT_C_NOT_ZERO); +/* Three byte sequence. */ +OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(1)); +OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, 0x0f); +OP2(SLJIT_SHL, TMP1, 0, TMP1, 0, SLJIT_IMM, 12); +OP2(SLJIT_AND, TMP2, 0, TMP2, 0, SLJIT_IMM, 0x3f); +OP2(SLJIT_SHL, TMP2, 0, TMP2, 0, SLJIT_IMM, 6); +OP2(SLJIT_OR, TMP1, 0, TMP1, 0, TMP2, 0); +OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(2)); +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(2)); +OP2(SLJIT_AND, TMP2, 0, TMP2, 0, SLJIT_IMM, 0x3f); +OP2(SLJIT_OR, TMP1, 0, TMP1, 0, TMP2, 0); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, IN_UCHARS(2)); +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); +JUMPHERE(jump); + +/* Four byte sequence. */ +OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(1)); +OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, 0x07); +OP2(SLJIT_SHL, TMP1, 0, TMP1, 0, SLJIT_IMM, 18); +OP2(SLJIT_AND, TMP2, 0, TMP2, 0, SLJIT_IMM, 0x3f); +OP2(SLJIT_SHL, TMP2, 0, TMP2, 0, SLJIT_IMM, 12); +OP2(SLJIT_OR, TMP1, 0, TMP1, 0, TMP2, 0); +OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(2)); +OP2(SLJIT_AND, TMP2, 0, TMP2, 0, SLJIT_IMM, 0x3f); +OP2(SLJIT_SHL, TMP2, 0, TMP2, 0, SLJIT_IMM, 6); +OP2(SLJIT_OR, TMP1, 0, TMP1, 0, TMP2, 0); +OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(3)); +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(3)); +OP2(SLJIT_AND, TMP2, 0, TMP2, 0, SLJIT_IMM, 0x3f); +OP2(SLJIT_OR, TMP1, 0, TMP1, 0, TMP2, 0); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, IN_UCHARS(3)); +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); +} + +static void do_utfreadtype8(compiler_common *common) +{ +/* Fast decoding a UTF-8 character type. TMP2 contains the first byte +of the character (>= 0xc0). Return value in TMP1. */ +DEFINE_COMPILER; +struct sljit_jump *jump; +struct sljit_jump *compare; + +sljit_emit_fast_enter(compiler, RETURN_ADDR, 0); + +OP2(SLJIT_AND | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP2, 0, SLJIT_IMM, 0x20); +jump = JUMP(SLJIT_C_NOT_ZERO); +/* Two byte sequence. */ +OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(0)); +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +OP2(SLJIT_AND, TMP2, 0, TMP2, 0, SLJIT_IMM, 0x1f); +OP2(SLJIT_SHL, TMP2, 0, TMP2, 0, SLJIT_IMM, 6); +OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, 0x3f); +OP2(SLJIT_OR, TMP2, 0, TMP2, 0, TMP1, 0); +compare = CMP(SLJIT_C_GREATER, TMP2, 0, SLJIT_IMM, 255); +OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP2), common->ctypes); +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); + +JUMPHERE(compare); +OP1(SLJIT_MOV, TMP1, 0, SLJIT_IMM, 0); +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); +JUMPHERE(jump); + +/* We only have types for characters less than 256. */ +OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP2), (sljit_w)PRIV(utf8_table4) - 0xc0); +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP1, 0); +OP1(SLJIT_MOV, TMP1, 0, SLJIT_IMM, 0); +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); +} + +#else /* COMPILE_PCRE8 */ + +#ifdef COMPILE_PCRE16 +static void do_utfreadchar(compiler_common *common) +{ +/* Fast decoding a UTF-16 character. TMP1 contains the first 16 bit char +of the character (>= 0xd800). Return char value in TMP1, length - 1 in TMP2. */ +DEFINE_COMPILER; +struct sljit_jump *jump; + +sljit_emit_fast_enter(compiler, RETURN_ADDR, 0); +jump = CMP(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xdc00); +/* Do nothing, only return. */ +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); + +JUMPHERE(jump); +/* Combine two 16 bit characters. */ +OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(1)); +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, 0x3ff); +OP2(SLJIT_SHL, TMP1, 0, TMP1, 0, SLJIT_IMM, 10); +OP2(SLJIT_AND, TMP2, 0, TMP2, 0, SLJIT_IMM, 0x3ff); +OP2(SLJIT_OR, TMP1, 0, TMP1, 0, TMP2, 0); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, IN_UCHARS(1)); +OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 0x10000); +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); +} +#endif /* COMPILE_PCRE16 */ + +#endif /* COMPILE_PCRE8 */ + +#endif /* SUPPORT_UTF */ + +#ifdef SUPPORT_UCP + +/* UCD_BLOCK_SIZE must be 128 (see the assert below). */ +#define UCD_BLOCK_MASK 127 +#define UCD_BLOCK_SHIFT 7 + +static void do_getucd(compiler_common *common) +{ +/* Search the UCD record for the character comes in TMP1. +Returns chartype in TMP1 and UCD offset in TMP2. */ +DEFINE_COMPILER; + +SLJIT_ASSERT(UCD_BLOCK_SIZE == 128 && sizeof(ucd_record) == 8); + +sljit_emit_fast_enter(compiler, RETURN_ADDR, 0); +OP2(SLJIT_LSHR, TMP2, 0, TMP1, 0, SLJIT_IMM, UCD_BLOCK_SHIFT); +OP1(SLJIT_MOV_UB, TMP2, 0, SLJIT_MEM1(TMP2), (sljit_w)PRIV(ucd_stage1)); +OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, UCD_BLOCK_MASK); +OP2(SLJIT_SHL, TMP2, 0, TMP2, 0, SLJIT_IMM, UCD_BLOCK_SHIFT); +OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, TMP2, 0); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, (sljit_w)PRIV(ucd_stage2)); +OP1(SLJIT_MOV_UH, TMP2, 0, SLJIT_MEM2(TMP2, TMP1), 1); +OP1(SLJIT_MOV, TMP1, 0, SLJIT_IMM, (sljit_w)PRIV(ucd_records) + SLJIT_OFFSETOF(ucd_record, chartype)); +OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM2(TMP1, TMP2), 3); +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); +} +#endif + +static SLJIT_INLINE struct sljit_label *mainloop_entry(compiler_common *common, BOOL hascrorlf, BOOL firstline) +{ +DEFINE_COMPILER; +struct sljit_label *mainloop; +struct sljit_label *newlinelabel = NULL; +struct sljit_jump *start; +struct sljit_jump *end = NULL; +struct sljit_jump *nl = NULL; +#ifdef SUPPORT_UTF +struct sljit_jump *singlechar; +#endif +jump_list *newline = NULL; +BOOL newlinecheck = FALSE; +BOOL readuchar = FALSE; + +if (!(hascrorlf || firstline) && (common->nltype == NLTYPE_ANY || + common->nltype == NLTYPE_ANYCRLF || common->newline > 255)) + newlinecheck = TRUE; + +if (firstline) + { + /* Search for the end of the first line. */ + SLJIT_ASSERT(common->first_line_end != 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0, STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->first_line_end, STR_END, 0); + + if (common->nltype == NLTYPE_FIXED && common->newline > 255) + { + mainloop = LABEL(); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + end = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(-1)); + OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(0)); + CMPTO(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, (common->newline >> 8) & 0xff, mainloop); + CMPTO(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, common->newline & 0xff, mainloop); + OP2(SLJIT_SUB, SLJIT_MEM1(SLJIT_LOCALS_REG), common->first_line_end, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + } + else + { + end = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); + mainloop = LABEL(); + /* Continual stores does not cause data dependency. */ + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->first_line_end, STR_PTR, 0); + read_char(common); + check_newlinechar(common, common->nltype, &newline, TRUE); + CMPTO(SLJIT_C_LESS, STR_PTR, 0, STR_END, 0, mainloop); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->first_line_end, STR_PTR, 0); + set_jumps(newline, LABEL()); + } + + JUMPHERE(end); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0); + } + +start = JUMP(SLJIT_JUMP); + +if (newlinecheck) + { + newlinelabel = LABEL(); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + end = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), 0); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, common->newline & 0xff); + COND_VALUE(SLJIT_MOV, TMP1, 0, SLJIT_C_EQUAL); +#ifdef COMPILE_PCRE16 + OP2(SLJIT_SHL, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); +#endif + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP1, 0); + nl = JUMP(SLJIT_JUMP); + } + +mainloop = LABEL(); + +/* Increasing the STR_PTR here requires one less jump in the most common case. */ +#ifdef SUPPORT_UTF +if (common->utf) readuchar = TRUE; +#endif +if (newlinecheck) readuchar = TRUE; + +if (readuchar) + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), 0); + +if (newlinecheck) + CMPTO(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, (common->newline >> 8) & 0xff, newlinelabel); + +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +#if defined SUPPORT_UTF && defined COMPILE_PCRE8 +if (common->utf) + { + singlechar = CMP(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xc0); + OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP1), (sljit_w)PRIV(utf8_table4) - 0xc0); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP1, 0); + JUMPHERE(singlechar); + } +#endif +#if defined SUPPORT_UTF && defined COMPILE_PCRE16 +if (common->utf) + { + singlechar = CMP(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xd800); + OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, 0xfc00); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0xd800); + COND_VALUE(SLJIT_MOV, TMP1, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SHL, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP1, 0); + JUMPHERE(singlechar); + } +#endif +JUMPHERE(start); + +if (newlinecheck) + { + JUMPHERE(end); + JUMPHERE(nl); + } + +return mainloop; +} + +static SLJIT_INLINE void fast_forward_first_char(compiler_common *common, pcre_uchar first_char, BOOL caseless, BOOL firstline) +{ +DEFINE_COMPILER; +struct sljit_label *start; +struct sljit_jump *leave; +struct sljit_jump *found; +pcre_uchar oc, bit; + +if (firstline) + { + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, STR_END, 0); + OP1(SLJIT_MOV, STR_END, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->first_line_end); + } + +start = LABEL(); +leave = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); +OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), 0); + +oc = first_char; +if (caseless) + { + oc = TABLE_GET(first_char, common->fcc, first_char); +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + if (first_char > 127 && common->utf) + oc = UCD_OTHERCASE(first_char); +#endif + } +if (first_char == oc) + found = CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, first_char); +else + { + bit = first_char ^ oc; + if (ispowerof2(bit)) + { + OP2(SLJIT_OR, TMP2, 0, TMP1, 0, SLJIT_IMM, bit); + found = CMP(SLJIT_C_EQUAL, TMP2, 0, SLJIT_IMM, first_char | bit); + } + else + { + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, first_char); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, oc); + COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_EQUAL); + found = JUMP(SLJIT_C_NOT_ZERO); + } + } + +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +#if defined SUPPORT_UTF && defined COMPILE_PCRE8 +if (common->utf) + { + CMPTO(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xc0, start); + OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP1), (sljit_w)PRIV(utf8_table4) - 0xc0); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP1, 0); + } +#endif +#if defined SUPPORT_UTF && defined COMPILE_PCRE16 +if (common->utf) + { + CMPTO(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xd800, start); + OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, 0xfc00); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0xd800); + COND_VALUE(SLJIT_MOV, TMP1, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SHL, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP1, 0); + } +#endif +JUMPTO(SLJIT_JUMP, start); +JUMPHERE(found); +JUMPHERE(leave); + +if (firstline) + OP1(SLJIT_MOV, STR_END, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0); +} + +static SLJIT_INLINE void fast_forward_newline(compiler_common *common, BOOL firstline) +{ +DEFINE_COMPILER; +struct sljit_label *loop; +struct sljit_jump *lastchar; +struct sljit_jump *firstchar; +struct sljit_jump *leave; +struct sljit_jump *foundcr = NULL; +struct sljit_jump *notfoundnl; +jump_list *newline = NULL; + +if (firstline) + { + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, STR_END, 0); + OP1(SLJIT_MOV, STR_END, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->first_line_end); + } + +if (common->nltype == NLTYPE_FIXED && common->newline > 255) + { + lastchar = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); + OP1(SLJIT_MOV, TMP1, 0, ARGUMENTS, 0); + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, str)); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, begin)); + firstchar = CMP(SLJIT_C_LESS_EQUAL, STR_PTR, 0, TMP2, 0); + + OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, IN_UCHARS(2)); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, STR_PTR, 0, TMP1, 0); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_GREATER_EQUAL); +#ifdef COMPILE_PCRE16 + OP2(SLJIT_SHL, TMP2, 0, TMP2, 0, SLJIT_IMM, 1); +#endif + OP2(SLJIT_SUB, STR_PTR, 0, STR_PTR, 0, TMP2, 0); + + loop = LABEL(); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + leave = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(-2)); + OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(-1)); + CMPTO(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, (common->newline >> 8) & 0xff, loop); + CMPTO(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, common->newline & 0xff, loop); + + JUMPHERE(leave); + JUMPHERE(firstchar); + JUMPHERE(lastchar); + + if (firstline) + OP1(SLJIT_MOV, STR_END, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0); + return; + } + +OP1(SLJIT_MOV, TMP1, 0, ARGUMENTS, 0); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, str)); +firstchar = CMP(SLJIT_C_LESS_EQUAL, STR_PTR, 0, TMP2, 0); +skip_char_back(common); + +loop = LABEL(); +read_char(common); +lastchar = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); +if (common->nltype == NLTYPE_ANY || common->nltype == NLTYPE_ANYCRLF) + foundcr = CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, CHAR_CR); +check_newlinechar(common, common->nltype, &newline, FALSE); +set_jumps(newline, loop); + +if (common->nltype == NLTYPE_ANY || common->nltype == NLTYPE_ANYCRLF) + { + leave = JUMP(SLJIT_JUMP); + JUMPHERE(foundcr); + notfoundnl = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), 0); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, CHAR_NL); + COND_VALUE(SLJIT_MOV, TMP1, 0, SLJIT_C_EQUAL); +#ifdef COMPILE_PCRE16 + OP2(SLJIT_SHL, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); +#endif + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP1, 0); + JUMPHERE(notfoundnl); + JUMPHERE(leave); + } +JUMPHERE(lastchar); +JUMPHERE(firstchar); + +if (firstline) + OP1(SLJIT_MOV, STR_END, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0); +} + +static SLJIT_INLINE void fast_forward_start_bits(compiler_common *common, sljit_uw start_bits, BOOL firstline) +{ +DEFINE_COMPILER; +struct sljit_label *start; +struct sljit_jump *leave; +struct sljit_jump *found; +#ifndef COMPILE_PCRE8 +struct sljit_jump *jump; +#endif + +if (firstline) + { + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, STR_END, 0); + OP1(SLJIT_MOV, STR_END, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->first_line_end); + } + +start = LABEL(); +leave = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); +OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), 0); +#ifdef SUPPORT_UTF +if (common->utf) + OP1(SLJIT_MOV, TMP3, 0, TMP1, 0); +#endif +#ifndef COMPILE_PCRE8 +jump = CMP(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 255); +OP1(SLJIT_MOV, TMP1, 0, SLJIT_IMM, 255); +JUMPHERE(jump); +#endif +OP2(SLJIT_AND, TMP2, 0, TMP1, 0, SLJIT_IMM, 0x7); +OP2(SLJIT_LSHR, TMP1, 0, TMP1, 0, SLJIT_IMM, 3); +OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP1), start_bits); +OP2(SLJIT_SHL, TMP2, 0, SLJIT_IMM, 1, TMP2, 0); +OP2(SLJIT_AND | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, TMP2, 0); +found = JUMP(SLJIT_C_NOT_ZERO); + +#ifdef SUPPORT_UTF +if (common->utf) + OP1(SLJIT_MOV, TMP1, 0, TMP3, 0); +#endif +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +#if defined SUPPORT_UTF && defined COMPILE_PCRE8 +if (common->utf) + { + CMPTO(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xc0, start); + OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP1), (sljit_w)PRIV(utf8_table4) - 0xc0); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP1, 0); + } +#endif +#if defined SUPPORT_UTF && defined COMPILE_PCRE16 +if (common->utf) + { + CMPTO(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xd800, start); + OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, 0xfc00); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0xd800); + COND_VALUE(SLJIT_MOV, TMP1, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SHL, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP1, 0); + } +#endif +JUMPTO(SLJIT_JUMP, start); +JUMPHERE(found); +JUMPHERE(leave); + +if (firstline) + OP1(SLJIT_MOV, STR_END, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0); +} + +static SLJIT_INLINE struct sljit_jump *search_requested_char(compiler_common *common, pcre_uchar req_char, BOOL caseless, BOOL has_firstchar) +{ +DEFINE_COMPILER; +struct sljit_label *loop; +struct sljit_jump *toolong; +struct sljit_jump *alreadyfound; +struct sljit_jump *found; +struct sljit_jump *foundoc = NULL; +struct sljit_jump *notfound; +pcre_uchar oc, bit; + +SLJIT_ASSERT(common->req_char_ptr != 0); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->req_char_ptr); +OP2(SLJIT_ADD, TMP1, 0, STR_PTR, 0, SLJIT_IMM, REQ_BYTE_MAX); +toolong = CMP(SLJIT_C_LESS, TMP1, 0, STR_END, 0); +alreadyfound = CMP(SLJIT_C_LESS, STR_PTR, 0, TMP2, 0); + +if (has_firstchar) + OP2(SLJIT_ADD, TMP1, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +else + OP1(SLJIT_MOV, TMP1, 0, STR_PTR, 0); + +loop = LABEL(); +notfound = CMP(SLJIT_C_GREATER_EQUAL, TMP1, 0, STR_END, 0); + +OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(TMP1), 0); +oc = req_char; +if (caseless) + { + oc = TABLE_GET(req_char, common->fcc, req_char); +#if defined SUPPORT_UCP && !(defined COMPILE_PCRE8) + if (req_char > 127 && common->utf) + oc = UCD_OTHERCASE(req_char); +#endif + } +if (req_char == oc) + found = CMP(SLJIT_C_EQUAL, TMP2, 0, SLJIT_IMM, req_char); +else + { + bit = req_char ^ oc; + if (ispowerof2(bit)) + { + OP2(SLJIT_OR, TMP2, 0, TMP2, 0, SLJIT_IMM, bit); + found = CMP(SLJIT_C_EQUAL, TMP2, 0, SLJIT_IMM, req_char | bit); + } + else + { + found = CMP(SLJIT_C_EQUAL, TMP2, 0, SLJIT_IMM, req_char); + foundoc = CMP(SLJIT_C_EQUAL, TMP2, 0, SLJIT_IMM, oc); + } + } +OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, IN_UCHARS(1)); +JUMPTO(SLJIT_JUMP, loop); + +JUMPHERE(found); +if (foundoc) + JUMPHERE(foundoc); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->req_char_ptr, TMP1, 0); +JUMPHERE(alreadyfound); +JUMPHERE(toolong); +return notfound; +} + +static void do_revertframes(compiler_common *common) +{ +DEFINE_COMPILER; +struct sljit_jump *jump; +struct sljit_label *mainloop; + +sljit_emit_fast_enter(compiler, RETURN_ADDR, 0); +OP1(SLJIT_MOV, TMP1, 0, STACK_TOP, 0); +GET_LOCAL_BASE(TMP3, 0, 0); + +/* Drop frames until we reach STACK_TOP. */ +mainloop = LABEL(); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(TMP1), 0); +jump = CMP(SLJIT_C_SIG_LESS_EQUAL, TMP2, 0, SLJIT_IMM, frame_end); +OP2(SLJIT_ADD, TMP2, 0, TMP2, 0, TMP3, 0); +OP1(SLJIT_MOV, SLJIT_MEM1(TMP2), 0, SLJIT_MEM1(TMP1), sizeof(sljit_w)); +OP1(SLJIT_MOV, SLJIT_MEM1(TMP2), sizeof(sljit_w), SLJIT_MEM1(TMP1), 2 * sizeof(sljit_w)); +OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 3 * sizeof(sljit_w)); +JUMPTO(SLJIT_JUMP, mainloop); + +JUMPHERE(jump); +jump = CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, frame_end); +/* End of dropping frames. */ +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); + +JUMPHERE(jump); +jump = CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, frame_setstrbegin); +/* Set string begin. */ +OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(TMP1), sizeof(sljit_w)); +OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 2 * sizeof(sljit_w)); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0), TMP2, 0); +JUMPTO(SLJIT_JUMP, mainloop); + +JUMPHERE(jump); +if (common->mark_ptr != 0) + { + jump = CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, frame_setmark); + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(TMP1), sizeof(sljit_w)); + OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 2 * sizeof(sljit_w)); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->mark_ptr, TMP2, 0); + JUMPTO(SLJIT_JUMP, mainloop); + + JUMPHERE(jump); + } + +/* Unknown command. */ +OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 2 * sizeof(sljit_w)); +JUMPTO(SLJIT_JUMP, mainloop); +} + +static void check_wordboundary(compiler_common *common) +{ +DEFINE_COMPILER; +struct sljit_jump *skipread; +#if !(defined COMPILE_PCRE8) || defined SUPPORT_UTF +struct sljit_jump *jump; +#endif + +SLJIT_COMPILE_ASSERT(ctype_word == 0x10, ctype_word_must_be_16); + +sljit_emit_fast_enter(compiler, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0); +/* Get type of the previous char, and put it to LOCALS1. */ +OP1(SLJIT_MOV, TMP1, 0, ARGUMENTS, 0); +OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, begin)); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1, SLJIT_IMM, 0); +skipread = CMP(SLJIT_C_LESS_EQUAL, STR_PTR, 0, TMP1, 0); +skip_char_back(common); +check_start_used_ptr(common); +read_char(common); + +/* Testing char type. */ +#ifdef SUPPORT_UCP +if (common->use_ucp) + { + OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, 1); + jump = CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, CHAR_UNDERSCORE); + add_jump(compiler, &common->getucd, JUMP(SLJIT_FAST_CALL)); + OP2(SLJIT_SUB, TMP1, 0, TMP1, 0, SLJIT_IMM, ucp_Ll); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, ucp_Lu - ucp_Ll); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_LESS_EQUAL); + OP2(SLJIT_SUB, TMP1, 0, TMP1, 0, SLJIT_IMM, ucp_Nd - ucp_Ll); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, ucp_No - ucp_Nd); + COND_VALUE(SLJIT_OR, TMP2, 0, SLJIT_C_LESS_EQUAL); + JUMPHERE(jump); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1, TMP2, 0); + } +else +#endif + { +#ifndef COMPILE_PCRE8 + jump = CMP(SLJIT_C_GREATER, TMP1, 0, SLJIT_IMM, 255); +#elif defined SUPPORT_UTF + /* Here LOCALS1 has already been zeroed. */ + jump = NULL; + if (common->utf) + jump = CMP(SLJIT_C_GREATER, TMP1, 0, SLJIT_IMM, 255); +#endif /* COMPILE_PCRE8 */ + OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP1), common->ctypes); + OP2(SLJIT_LSHR, TMP1, 0, TMP1, 0, SLJIT_IMM, 4 /* ctype_word */); + OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1, TMP1, 0); +#ifndef COMPILE_PCRE8 + JUMPHERE(jump); +#elif defined SUPPORT_UTF + if (jump != NULL) + JUMPHERE(jump); +#endif /* COMPILE_PCRE8 */ + } +JUMPHERE(skipread); + +OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, 0); +skipread = check_str_end(common); +peek_char(common); + +/* Testing char type. This is a code duplication. */ +#ifdef SUPPORT_UCP +if (common->use_ucp) + { + OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, 1); + jump = CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, CHAR_UNDERSCORE); + add_jump(compiler, &common->getucd, JUMP(SLJIT_FAST_CALL)); + OP2(SLJIT_SUB, TMP1, 0, TMP1, 0, SLJIT_IMM, ucp_Ll); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, ucp_Lu - ucp_Ll); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_LESS_EQUAL); + OP2(SLJIT_SUB, TMP1, 0, TMP1, 0, SLJIT_IMM, ucp_Nd - ucp_Ll); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, ucp_No - ucp_Nd); + COND_VALUE(SLJIT_OR, TMP2, 0, SLJIT_C_LESS_EQUAL); + JUMPHERE(jump); + } +else +#endif + { +#ifndef COMPILE_PCRE8 + /* TMP2 may be destroyed by peek_char. */ + OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, 0); + jump = CMP(SLJIT_C_GREATER, TMP1, 0, SLJIT_IMM, 255); +#elif defined SUPPORT_UTF + OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, 0); + jump = NULL; + if (common->utf) + jump = CMP(SLJIT_C_GREATER, TMP1, 0, SLJIT_IMM, 255); +#endif + OP1(SLJIT_MOV_UB, TMP2, 0, SLJIT_MEM1(TMP1), common->ctypes); + OP2(SLJIT_LSHR, TMP2, 0, TMP2, 0, SLJIT_IMM, 4 /* ctype_word */); + OP2(SLJIT_AND, TMP2, 0, TMP2, 0, SLJIT_IMM, 1); +#ifndef COMPILE_PCRE8 + JUMPHERE(jump); +#elif defined SUPPORT_UTF + if (jump != NULL) + JUMPHERE(jump); +#endif /* COMPILE_PCRE8 */ + } +JUMPHERE(skipread); + +OP2(SLJIT_XOR | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1); +sljit_emit_fast_return(compiler, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0); +} + +static void check_anynewline(compiler_common *common) +{ +/* Check whether TMP1 contains a newline character. TMP2 destroyed. */ +DEFINE_COMPILER; + +sljit_emit_fast_enter(compiler, RETURN_ADDR, 0); + +OP2(SLJIT_SUB, TMP1, 0, TMP1, 0, SLJIT_IMM, 0x0a); +OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x0d - 0x0a); +COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_LESS_EQUAL); +OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x85 - 0x0a); +#if defined SUPPORT_UTF || defined COMPILE_PCRE16 +#ifdef COMPILE_PCRE8 +if (common->utf) + { +#endif + COND_VALUE(SLJIT_OR, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_OR, TMP1, 0, TMP1, 0, SLJIT_IMM, 0x1); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x2029 - 0x0a); +#ifdef COMPILE_PCRE8 + } +#endif +#endif /* SUPPORT_UTF || COMPILE_PCRE16 */ +COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_EQUAL); +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); +} + +static void check_hspace(compiler_common *common) +{ +/* Check whether TMP1 contains a newline character. TMP2 destroyed. */ +DEFINE_COMPILER; + +sljit_emit_fast_enter(compiler, RETURN_ADDR, 0); + +OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x09); +COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_EQUAL); +OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x20); +COND_VALUE(SLJIT_OR, TMP2, 0, SLJIT_C_EQUAL); +OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0xa0); +#if defined SUPPORT_UTF || defined COMPILE_PCRE16 +#ifdef COMPILE_PCRE8 +if (common->utf) + { +#endif + COND_VALUE(SLJIT_OR, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x1680); + COND_VALUE(SLJIT_OR, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x180e); + COND_VALUE(SLJIT_OR, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SUB, TMP1, 0, TMP1, 0, SLJIT_IMM, 0x2000); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x200A - 0x2000); + COND_VALUE(SLJIT_OR, TMP2, 0, SLJIT_C_LESS_EQUAL); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x202f - 0x2000); + COND_VALUE(SLJIT_OR, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x205f - 0x2000); + COND_VALUE(SLJIT_OR, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x3000 - 0x2000); +#ifdef COMPILE_PCRE8 + } +#endif +#endif /* SUPPORT_UTF || COMPILE_PCRE16 */ +COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_EQUAL); + +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); +} + +static void check_vspace(compiler_common *common) +{ +/* Check whether TMP1 contains a newline character. TMP2 destroyed. */ +DEFINE_COMPILER; + +sljit_emit_fast_enter(compiler, RETURN_ADDR, 0); + +OP2(SLJIT_SUB, TMP1, 0, TMP1, 0, SLJIT_IMM, 0x0a); +OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x0d - 0x0a); +COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_LESS_EQUAL); +OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x85 - 0x0a); +#if defined SUPPORT_UTF || defined COMPILE_PCRE16 +#ifdef COMPILE_PCRE8 +if (common->utf) + { +#endif + COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_OR, TMP1, 0, TMP1, 0, SLJIT_IMM, 0x1); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x2029 - 0x0a); +#ifdef COMPILE_PCRE8 + } +#endif +#endif /* SUPPORT_UTF || COMPILE_PCRE16 */ +COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_EQUAL); + +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); +} + +#define CHAR1 STR_END +#define CHAR2 STACK_TOP + +static void do_casefulcmp(compiler_common *common) +{ +DEFINE_COMPILER; +struct sljit_jump *jump; +struct sljit_label *label; + +sljit_emit_fast_enter(compiler, RETURN_ADDR, 0); +OP2(SLJIT_SUB, STR_PTR, 0, STR_PTR, 0, TMP2, 0); +OP1(SLJIT_MOV, TMP3, 0, CHAR1, 0); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0, CHAR2, 0); +OP2(SLJIT_SUB, TMP1, 0, TMP1, 0, SLJIT_IMM, IN_UCHARS(1)); +OP2(SLJIT_SUB, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + +label = LABEL(); +OP1(MOVU_UCHAR, CHAR1, 0, SLJIT_MEM1(TMP1), IN_UCHARS(1)); +OP1(MOVU_UCHAR, CHAR2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(1)); +jump = CMP(SLJIT_C_NOT_EQUAL, CHAR1, 0, CHAR2, 0); +OP2(SLJIT_SUB | SLJIT_SET_E, TMP2, 0, TMP2, 0, SLJIT_IMM, IN_UCHARS(1)); +JUMPTO(SLJIT_C_NOT_ZERO, label); + +JUMPHERE(jump); +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +OP1(SLJIT_MOV, CHAR1, 0, TMP3, 0); +OP1(SLJIT_MOV, CHAR2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0); +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); +} + +#define LCC_TABLE STACK_LIMIT + +static void do_caselesscmp(compiler_common *common) +{ +DEFINE_COMPILER; +struct sljit_jump *jump; +struct sljit_label *label; + +sljit_emit_fast_enter(compiler, RETURN_ADDR, 0); +OP2(SLJIT_SUB, STR_PTR, 0, STR_PTR, 0, TMP2, 0); + +OP1(SLJIT_MOV, TMP3, 0, LCC_TABLE, 0); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0, CHAR1, 0); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1, CHAR2, 0); +OP1(SLJIT_MOV, LCC_TABLE, 0, SLJIT_IMM, common->lcc); +OP2(SLJIT_SUB, TMP1, 0, TMP1, 0, SLJIT_IMM, IN_UCHARS(1)); +OP2(SLJIT_SUB, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + +label = LABEL(); +OP1(MOVU_UCHAR, CHAR1, 0, SLJIT_MEM1(TMP1), IN_UCHARS(1)); +OP1(MOVU_UCHAR, CHAR2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(1)); +#ifndef COMPILE_PCRE8 +jump = CMP(SLJIT_C_GREATER, CHAR1, 0, SLJIT_IMM, 255); +#endif +OP1(SLJIT_MOV_UB, CHAR1, 0, SLJIT_MEM2(LCC_TABLE, CHAR1), 0); +#ifndef COMPILE_PCRE8 +JUMPHERE(jump); +jump = CMP(SLJIT_C_GREATER, CHAR2, 0, SLJIT_IMM, 255); +#endif +OP1(SLJIT_MOV_UB, CHAR2, 0, SLJIT_MEM2(LCC_TABLE, CHAR2), 0); +#ifndef COMPILE_PCRE8 +JUMPHERE(jump); +#endif +jump = CMP(SLJIT_C_NOT_EQUAL, CHAR1, 0, CHAR2, 0); +OP2(SLJIT_SUB | SLJIT_SET_E, TMP2, 0, TMP2, 0, SLJIT_IMM, IN_UCHARS(1)); +JUMPTO(SLJIT_C_NOT_ZERO, label); + +JUMPHERE(jump); +OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +OP1(SLJIT_MOV, LCC_TABLE, 0, TMP3, 0); +OP1(SLJIT_MOV, CHAR1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0); +OP1(SLJIT_MOV, CHAR2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1); +sljit_emit_fast_return(compiler, RETURN_ADDR, 0); +} + +#undef LCC_TABLE +#undef CHAR1 +#undef CHAR2 + +#if defined SUPPORT_UTF && defined SUPPORT_UCP + +static const pcre_uchar *SLJIT_CALL do_utf_caselesscmp(pcre_uchar *src1, jit_arguments *args, pcre_uchar *end1) +{ +/* This function would be ineffective to do in JIT level. */ +int c1, c2; +const pcre_uchar *src2 = args->uchar_ptr; +const pcre_uchar *end2 = args->end; + +while (src1 < end1) + { + if (src2 >= end2) + return (pcre_uchar*)1; + GETCHARINC(c1, src1); + GETCHARINC(c2, src2); + if (c1 != c2 && c1 != UCD_OTHERCASE(c2)) return NULL; + } +return src2; +} + +#endif /* SUPPORT_UTF && SUPPORT_UCP */ + +static pcre_uchar *byte_sequence_compare(compiler_common *common, BOOL caseless, pcre_uchar *cc, + compare_context* context, jump_list **backtracks) +{ +DEFINE_COMPILER; +unsigned int othercasebit = 0; +pcre_uchar *othercasechar = NULL; +#ifdef SUPPORT_UTF +int utflength; +#endif + +if (caseless && char_has_othercase(common, cc)) + { + othercasebit = char_get_othercase_bit(common, cc); + SLJIT_ASSERT(othercasebit); + /* Extracting bit difference info. */ +#ifdef COMPILE_PCRE8 + othercasechar = cc + (othercasebit >> 8); + othercasebit &= 0xff; +#else +#ifdef COMPILE_PCRE16 + othercasechar = cc + (othercasebit >> 9); + if ((othercasebit & 0x100) != 0) + othercasebit = (othercasebit & 0xff) << 8; + else + othercasebit &= 0xff; +#endif +#endif + } + +if (context->sourcereg == -1) + { +#ifdef COMPILE_PCRE8 +#if defined SLJIT_UNALIGNED && SLJIT_UNALIGNED + if (context->length >= 4) + OP1(SLJIT_MOV_SI, TMP1, 0, SLJIT_MEM1(STR_PTR), -context->length); + else if (context->length >= 2) + OP1(SLJIT_MOV_UH, TMP1, 0, SLJIT_MEM1(STR_PTR), -context->length); + else +#endif + OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(STR_PTR), -context->length); +#else +#ifdef COMPILE_PCRE16 +#if defined SLJIT_UNALIGNED && SLJIT_UNALIGNED + if (context->length >= 4) + OP1(SLJIT_MOV_SI, TMP1, 0, SLJIT_MEM1(STR_PTR), -context->length); + else +#endif + OP1(SLJIT_MOV_UH, TMP1, 0, SLJIT_MEM1(STR_PTR), -context->length); +#endif +#endif /* COMPILE_PCRE8 */ + context->sourcereg = TMP2; + } + +#ifdef SUPPORT_UTF +utflength = 1; +if (common->utf && HAS_EXTRALEN(*cc)) + utflength += GET_EXTRALEN(*cc); + +do + { +#endif + + context->length -= IN_UCHARS(1); +#if defined SLJIT_UNALIGNED && SLJIT_UNALIGNED + + /* Unaligned read is supported. */ + if (othercasebit != 0 && othercasechar == cc) + { + context->c.asuchars[context->ucharptr] = *cc | othercasebit; + context->oc.asuchars[context->ucharptr] = othercasebit; + } + else + { + context->c.asuchars[context->ucharptr] = *cc; + context->oc.asuchars[context->ucharptr] = 0; + } + context->ucharptr++; + +#ifdef COMPILE_PCRE8 + if (context->ucharptr >= 4 || context->length == 0 || (context->ucharptr == 2 && context->length == 1)) +#else + if (context->ucharptr >= 2 || context->length == 0) +#endif + { + if (context->length >= 4) + OP1(SLJIT_MOV_SI, context->sourcereg, 0, SLJIT_MEM1(STR_PTR), -context->length); +#ifdef COMPILE_PCRE8 + else if (context->length >= 2) + OP1(SLJIT_MOV_UH, context->sourcereg, 0, SLJIT_MEM1(STR_PTR), -context->length); + else if (context->length >= 1) + OP1(SLJIT_MOV_UB, context->sourcereg, 0, SLJIT_MEM1(STR_PTR), -context->length); +#else + else if (context->length >= 2) + OP1(SLJIT_MOV_UH, context->sourcereg, 0, SLJIT_MEM1(STR_PTR), -context->length); +#endif + context->sourcereg = context->sourcereg == TMP1 ? TMP2 : TMP1; + + switch(context->ucharptr) + { + case 4 / sizeof(pcre_uchar): + if (context->oc.asint != 0) + OP2(SLJIT_OR, context->sourcereg, 0, context->sourcereg, 0, SLJIT_IMM, context->oc.asint); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, context->sourcereg, 0, SLJIT_IMM, context->c.asint | context->oc.asint)); + break; + + case 2 / sizeof(pcre_uchar): + if (context->oc.asushort != 0) + OP2(SLJIT_OR, context->sourcereg, 0, context->sourcereg, 0, SLJIT_IMM, context->oc.asushort); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, context->sourcereg, 0, SLJIT_IMM, context->c.asushort | context->oc.asushort)); + break; + +#ifdef COMPILE_PCRE8 + case 1: + if (context->oc.asbyte != 0) + OP2(SLJIT_OR, context->sourcereg, 0, context->sourcereg, 0, SLJIT_IMM, context->oc.asbyte); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, context->sourcereg, 0, SLJIT_IMM, context->c.asbyte | context->oc.asbyte)); + break; +#endif + + default: + SLJIT_ASSERT_STOP(); + break; + } + context->ucharptr = 0; + } + +#else + + /* Unaligned read is unsupported. */ +#ifdef COMPILE_PCRE8 + if (context->length > 0) + OP1(SLJIT_MOV_UB, context->sourcereg, 0, SLJIT_MEM1(STR_PTR), -context->length); +#else + if (context->length > 0) + OP1(SLJIT_MOV_UH, context->sourcereg, 0, SLJIT_MEM1(STR_PTR), -context->length); +#endif + context->sourcereg = context->sourcereg == TMP1 ? TMP2 : TMP1; + + if (othercasebit != 0 && othercasechar == cc) + { + OP2(SLJIT_OR, context->sourcereg, 0, context->sourcereg, 0, SLJIT_IMM, othercasebit); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, context->sourcereg, 0, SLJIT_IMM, *cc | othercasebit)); + } + else + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, context->sourcereg, 0, SLJIT_IMM, *cc)); + +#endif + + cc++; +#ifdef SUPPORT_UTF + utflength--; + } +while (utflength > 0); +#endif + +return cc; +} + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + +#define SET_TYPE_OFFSET(value) \ + if ((value) != typeoffset) \ + { \ + if ((value) > typeoffset) \ + OP2(SLJIT_SUB, typereg, 0, typereg, 0, SLJIT_IMM, (value) - typeoffset); \ + else \ + OP2(SLJIT_ADD, typereg, 0, typereg, 0, SLJIT_IMM, typeoffset - (value)); \ + } \ + typeoffset = (value); + +#define SET_CHAR_OFFSET(value) \ + if ((value) != charoffset) \ + { \ + if ((value) > charoffset) \ + OP2(SLJIT_SUB, TMP1, 0, TMP1, 0, SLJIT_IMM, (value) - charoffset); \ + else \ + OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, charoffset - (value)); \ + } \ + charoffset = (value); + +static void compile_xclass_trypath(compiler_common *common, pcre_uchar *cc, jump_list **backtracks) +{ +DEFINE_COMPILER; +jump_list *found = NULL; +jump_list **list = (*cc & XCL_NOT) == 0 ? &found : backtracks; +unsigned int c; +int compares; +struct sljit_jump *jump = NULL; +pcre_uchar *ccbegin; +#ifdef SUPPORT_UCP +BOOL needstype = FALSE, needsscript = FALSE, needschar = FALSE; +BOOL charsaved = FALSE; +int typereg = TMP1, scriptreg = TMP1; +unsigned int typeoffset; +#endif +int invertcmp, numberofcmps; +unsigned int charoffset; + +/* Although SUPPORT_UTF must be defined, we are not necessary in utf mode. */ +detect_partial_match(common, backtracks); +read_char(common); + +if ((*cc++ & XCL_MAP) != 0) + { + OP1(SLJIT_MOV, TMP3, 0, TMP1, 0); +#ifndef COMPILE_PCRE8 + jump = CMP(SLJIT_C_GREATER, TMP1, 0, SLJIT_IMM, 255); +#elif defined SUPPORT_UTF + if (common->utf) + jump = CMP(SLJIT_C_GREATER, TMP1, 0, SLJIT_IMM, 255); +#endif + + OP2(SLJIT_AND, TMP2, 0, TMP1, 0, SLJIT_IMM, 0x7); + OP2(SLJIT_LSHR, TMP1, 0, TMP1, 0, SLJIT_IMM, 3); + OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP1), (sljit_w)cc); + OP2(SLJIT_SHL, TMP2, 0, SLJIT_IMM, 1, TMP2, 0); + OP2(SLJIT_AND | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, TMP2, 0); + add_jump(compiler, list, JUMP(SLJIT_C_NOT_ZERO)); + +#ifndef COMPILE_PCRE8 + JUMPHERE(jump); +#elif defined SUPPORT_UTF + if (common->utf) + JUMPHERE(jump); +#endif + OP1(SLJIT_MOV, TMP1, 0, TMP3, 0); +#ifdef SUPPORT_UCP + charsaved = TRUE; +#endif + cc += 32 / sizeof(pcre_uchar); + } + +/* Scanning the necessary info. */ +ccbegin = cc; +compares = 0; +while (*cc != XCL_END) + { + compares++; + if (*cc == XCL_SINGLE) + { + cc += 2; +#ifdef SUPPORT_UTF + if (common->utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]); +#endif +#ifdef SUPPORT_UCP + needschar = TRUE; +#endif + } + else if (*cc == XCL_RANGE) + { + cc += 2; +#ifdef SUPPORT_UTF + if (common->utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]); +#endif + cc++; +#ifdef SUPPORT_UTF + if (common->utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]); +#endif +#ifdef SUPPORT_UCP + needschar = TRUE; +#endif + } +#ifdef SUPPORT_UCP + else + { + SLJIT_ASSERT(*cc == XCL_PROP || *cc == XCL_NOTPROP); + cc++; + switch(*cc) + { + case PT_ANY: + break; + + case PT_LAMP: + case PT_GC: + case PT_PC: + case PT_ALNUM: + needstype = TRUE; + break; + + case PT_SC: + needsscript = TRUE; + break; + + case PT_SPACE: + case PT_PXSPACE: + case PT_WORD: + needstype = TRUE; + needschar = TRUE; + break; + + default: + SLJIT_ASSERT_STOP(); + break; + } + cc += 2; + } +#endif + } + +#ifdef SUPPORT_UCP +/* Simple register allocation. TMP1 is preferred if possible. */ +if (needstype || needsscript) + { + if (needschar && !charsaved) + OP1(SLJIT_MOV, TMP3, 0, TMP1, 0); + add_jump(compiler, &common->getucd, JUMP(SLJIT_FAST_CALL)); + if (needschar) + { + if (needstype) + { + OP1(SLJIT_MOV, RETURN_ADDR, 0, TMP1, 0); + typereg = RETURN_ADDR; + } + + if (needsscript) + scriptreg = TMP3; + OP1(SLJIT_MOV, TMP1, 0, TMP3, 0); + } + else if (needstype && needsscript) + scriptreg = TMP3; + /* In all other cases only one of them was specified, and that can goes to TMP1. */ + + if (needsscript) + { + if (scriptreg == TMP1) + { + OP1(SLJIT_MOV, scriptreg, 0, SLJIT_IMM, (sljit_w)PRIV(ucd_records) + SLJIT_OFFSETOF(ucd_record, script)); + OP1(SLJIT_MOV_UB, scriptreg, 0, SLJIT_MEM2(scriptreg, TMP2), 3); + } + else + { + OP2(SLJIT_SHL, TMP2, 0, TMP2, 0, SLJIT_IMM, 3); + OP2(SLJIT_ADD, TMP2, 0, TMP2, 0, SLJIT_IMM, (sljit_w)PRIV(ucd_records) + SLJIT_OFFSETOF(ucd_record, script)); + OP1(SLJIT_MOV_UB, scriptreg, 0, SLJIT_MEM1(TMP2), 0); + } + } + } +#endif + +/* Generating code. */ +cc = ccbegin; +charoffset = 0; +numberofcmps = 0; +#ifdef SUPPORT_UCP +typeoffset = 0; +#endif + +while (*cc != XCL_END) + { + compares--; + invertcmp = (compares == 0 && list != backtracks); + jump = NULL; + + if (*cc == XCL_SINGLE) + { + cc ++; +#ifdef SUPPORT_UTF + if (common->utf) + { + GETCHARINC(c, cc); + } + else +#endif + c = *cc++; + + if (numberofcmps < 3 && (*cc == XCL_SINGLE || *cc == XCL_RANGE)) + { + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, c - charoffset); + COND_VALUE(numberofcmps == 0 ? SLJIT_MOV : SLJIT_OR, TMP2, 0, SLJIT_C_EQUAL); + numberofcmps++; + } + else if (numberofcmps > 0) + { + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, c - charoffset); + COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_EQUAL); + jump = JUMP(SLJIT_C_NOT_ZERO ^ invertcmp); + numberofcmps = 0; + } + else + { + jump = CMP(SLJIT_C_EQUAL ^ invertcmp, TMP1, 0, SLJIT_IMM, c - charoffset); + numberofcmps = 0; + } + } + else if (*cc == XCL_RANGE) + { + cc ++; +#ifdef SUPPORT_UTF + if (common->utf) + { + GETCHARINC(c, cc); + } + else +#endif + c = *cc++; + SET_CHAR_OFFSET(c); +#ifdef SUPPORT_UTF + if (common->utf) + { + GETCHARINC(c, cc); + } + else +#endif + c = *cc++; + if (numberofcmps < 3 && (*cc == XCL_SINGLE || *cc == XCL_RANGE)) + { + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, c - charoffset); + COND_VALUE(numberofcmps == 0 ? SLJIT_MOV : SLJIT_OR, TMP2, 0, SLJIT_C_LESS_EQUAL); + numberofcmps++; + } + else if (numberofcmps > 0) + { + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, c - charoffset); + COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_LESS_EQUAL); + jump = JUMP(SLJIT_C_NOT_ZERO ^ invertcmp); + numberofcmps = 0; + } + else + { + jump = CMP(SLJIT_C_LESS_EQUAL ^ invertcmp, TMP1, 0, SLJIT_IMM, c - charoffset); + numberofcmps = 0; + } + } +#ifdef SUPPORT_UCP + else + { + if (*cc == XCL_NOTPROP) + invertcmp ^= 0x1; + cc++; + switch(*cc) + { + case PT_ANY: + if (list != backtracks) + { + if ((cc[-1] == XCL_NOTPROP && compares > 0) || (cc[-1] == XCL_PROP && compares == 0)) + continue; + } + else if (cc[-1] == XCL_NOTPROP) + continue; + jump = JUMP(SLJIT_JUMP); + break; + + case PT_LAMP: + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, typereg, 0, SLJIT_IMM, ucp_Lu - typeoffset); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, typereg, 0, SLJIT_IMM, ucp_Ll - typeoffset); + COND_VALUE(SLJIT_OR, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, typereg, 0, SLJIT_IMM, ucp_Lt - typeoffset); + COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_EQUAL); + jump = JUMP(SLJIT_C_NOT_ZERO ^ invertcmp); + break; + + case PT_GC: + c = PRIV(ucp_typerange)[(int)cc[1] * 2]; + SET_TYPE_OFFSET(c); + jump = CMP(SLJIT_C_LESS_EQUAL ^ invertcmp, typereg, 0, SLJIT_IMM, PRIV(ucp_typerange)[(int)cc[1] * 2 + 1] - c); + break; + + case PT_PC: + jump = CMP(SLJIT_C_EQUAL ^ invertcmp, typereg, 0, SLJIT_IMM, (int)cc[1] - typeoffset); + break; + + case PT_SC: + jump = CMP(SLJIT_C_EQUAL ^ invertcmp, scriptreg, 0, SLJIT_IMM, (int)cc[1]); + break; + + case PT_SPACE: + case PT_PXSPACE: + if (*cc == PT_SPACE) + { + OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, 0); + jump = CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, 11 - charoffset); + } + SET_CHAR_OFFSET(9); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 13 - 9); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_LESS_EQUAL); + if (*cc == PT_SPACE) + JUMPHERE(jump); + + SET_TYPE_OFFSET(ucp_Zl); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, typereg, 0, SLJIT_IMM, ucp_Zs - ucp_Zl); + COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_LESS_EQUAL); + jump = JUMP(SLJIT_C_NOT_ZERO ^ invertcmp); + break; + + case PT_WORD: + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, CHAR_UNDERSCORE - charoffset); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_EQUAL); + /* ... fall through */ + + case PT_ALNUM: + SET_TYPE_OFFSET(ucp_Ll); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, typereg, 0, SLJIT_IMM, ucp_Lu - ucp_Ll); + COND_VALUE((*cc == PT_ALNUM) ? SLJIT_MOV : SLJIT_OR, TMP2, 0, SLJIT_C_LESS_EQUAL); + SET_TYPE_OFFSET(ucp_Nd); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, typereg, 0, SLJIT_IMM, ucp_No - ucp_Nd); + COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_LESS_EQUAL); + jump = JUMP(SLJIT_C_NOT_ZERO ^ invertcmp); + break; + } + cc += 2; + } +#endif + + if (jump != NULL) + add_jump(compiler, compares > 0 ? list : backtracks, jump); + } + +if (found != NULL) + set_jumps(found, LABEL()); +} + +#undef SET_TYPE_OFFSET +#undef SET_CHAR_OFFSET + +#endif + +static pcre_uchar *compile_char1_trypath(compiler_common *common, pcre_uchar type, pcre_uchar *cc, jump_list **backtracks) +{ +DEFINE_COMPILER; +int length; +unsigned int c, oc, bit; +compare_context context; +struct sljit_jump *jump[4]; +#ifdef SUPPORT_UTF +struct sljit_label *label; +#ifdef SUPPORT_UCP +pcre_uchar propdata[5]; +#endif +#endif + +switch(type) + { + case OP_SOD: + OP1(SLJIT_MOV, TMP1, 0, ARGUMENTS, 0); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, begin)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, STR_PTR, 0, TMP1, 0)); + return cc; + + case OP_SOM: + OP1(SLJIT_MOV, TMP1, 0, ARGUMENTS, 0); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, str)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, STR_PTR, 0, TMP1, 0)); + return cc; + + case OP_NOT_WORD_BOUNDARY: + case OP_WORD_BOUNDARY: + add_jump(compiler, &common->wordboundary, JUMP(SLJIT_FAST_CALL)); + add_jump(compiler, backtracks, JUMP(type == OP_NOT_WORD_BOUNDARY ? SLJIT_C_NOT_ZERO : SLJIT_C_ZERO)); + return cc; + + case OP_NOT_DIGIT: + case OP_DIGIT: + detect_partial_match(common, backtracks); + read_char8_type(common); + OP2(SLJIT_AND | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, ctype_digit); + add_jump(compiler, backtracks, JUMP(type == OP_DIGIT ? SLJIT_C_ZERO : SLJIT_C_NOT_ZERO)); + return cc; + + case OP_NOT_WHITESPACE: + case OP_WHITESPACE: + detect_partial_match(common, backtracks); + read_char8_type(common); + OP2(SLJIT_AND | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, ctype_space); + add_jump(compiler, backtracks, JUMP(type == OP_WHITESPACE ? SLJIT_C_ZERO : SLJIT_C_NOT_ZERO)); + return cc; + + case OP_NOT_WORDCHAR: + case OP_WORDCHAR: + detect_partial_match(common, backtracks); + read_char8_type(common); + OP2(SLJIT_AND | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, ctype_word); + add_jump(compiler, backtracks, JUMP(type == OP_WORDCHAR ? SLJIT_C_ZERO : SLJIT_C_NOT_ZERO)); + return cc; + + case OP_ANY: + detect_partial_match(common, backtracks); + read_char(common); + if (common->nltype == NLTYPE_FIXED && common->newline > 255) + { + jump[0] = CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, (common->newline >> 8) & 0xff); + if (common->mode != JIT_PARTIAL_HARD_COMPILE) + jump[1] = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); + else + jump[1] = check_str_end(common); + + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), 0); + add_jump(compiler, backtracks, CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, common->newline & 0xff)); + if (jump[1] != NULL) + JUMPHERE(jump[1]); + JUMPHERE(jump[0]); + } + else + check_newlinechar(common, common->nltype, backtracks, TRUE); + return cc; + + case OP_ALLANY: + detect_partial_match(common, backtracks); +#ifdef SUPPORT_UTF + if (common->utf) + { + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), 0); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); +#ifdef COMPILE_PCRE8 + jump[0] = CMP(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xc0); + OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP1), (sljit_w)PRIV(utf8_table4) - 0xc0); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP1, 0); +#else /* COMPILE_PCRE8 */ +#ifdef COMPILE_PCRE16 + jump[0] = CMP(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xd800); + OP2(SLJIT_AND, TMP1, 0, TMP1, 0, SLJIT_IMM, 0xfc00); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0xd800); + COND_VALUE(SLJIT_MOV, TMP1, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SHL, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP1, 0); +#endif /* COMPILE_PCRE16 */ +#endif /* COMPILE_PCRE8 */ + JUMPHERE(jump[0]); + return cc; + } +#endif + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + return cc; + + case OP_ANYBYTE: + detect_partial_match(common, backtracks); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + return cc; + +#ifdef SUPPORT_UTF +#ifdef SUPPORT_UCP + case OP_NOTPROP: + case OP_PROP: + propdata[0] = 0; + propdata[1] = type == OP_NOTPROP ? XCL_NOTPROP : XCL_PROP; + propdata[2] = cc[0]; + propdata[3] = cc[1]; + propdata[4] = XCL_END; + compile_xclass_trypath(common, propdata, backtracks); + return cc + 2; +#endif +#endif + + case OP_ANYNL: + detect_partial_match(common, backtracks); + read_char(common); + jump[0] = CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, CHAR_CR); + /* We don't need to handle soft partial matching case. */ + if (common->mode != JIT_PARTIAL_HARD_COMPILE) + jump[1] = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); + else + jump[1] = check_str_end(common); + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), 0); + jump[2] = CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, CHAR_NL); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + jump[3] = JUMP(SLJIT_JUMP); + JUMPHERE(jump[0]); + check_newlinechar(common, common->bsr_nltype, backtracks, FALSE); + JUMPHERE(jump[1]); + JUMPHERE(jump[2]); + JUMPHERE(jump[3]); + return cc; + + case OP_NOT_HSPACE: + case OP_HSPACE: + detect_partial_match(common, backtracks); + read_char(common); + add_jump(compiler, &common->hspace, JUMP(SLJIT_FAST_CALL)); + add_jump(compiler, backtracks, JUMP(type == OP_NOT_HSPACE ? SLJIT_C_NOT_ZERO : SLJIT_C_ZERO)); + return cc; + + case OP_NOT_VSPACE: + case OP_VSPACE: + detect_partial_match(common, backtracks); + read_char(common); + add_jump(compiler, &common->vspace, JUMP(SLJIT_FAST_CALL)); + add_jump(compiler, backtracks, JUMP(type == OP_NOT_VSPACE ? SLJIT_C_NOT_ZERO : SLJIT_C_ZERO)); + return cc; + +#ifdef SUPPORT_UCP + case OP_EXTUNI: + detect_partial_match(common, backtracks); + read_char(common); + add_jump(compiler, &common->getucd, JUMP(SLJIT_FAST_CALL)); + OP2(SLJIT_SUB, TMP1, 0, TMP1, 0, SLJIT_IMM, ucp_Mc); + add_jump(compiler, backtracks, CMP(SLJIT_C_LESS_EQUAL, TMP1, 0, SLJIT_IMM, ucp_Mn - ucp_Mc)); + + label = LABEL(); + jump[0] = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); + OP1(SLJIT_MOV, TMP3, 0, STR_PTR, 0); + read_char(common); + add_jump(compiler, &common->getucd, JUMP(SLJIT_FAST_CALL)); + OP2(SLJIT_SUB, TMP1, 0, TMP1, 0, SLJIT_IMM, ucp_Mc); + CMPTO(SLJIT_C_LESS_EQUAL, TMP1, 0, SLJIT_IMM, ucp_Mn - ucp_Mc, label); + + OP1(SLJIT_MOV, STR_PTR, 0, TMP3, 0); + JUMPHERE(jump[0]); + if (common->mode == JIT_PARTIAL_HARD_COMPILE) + { + jump[0] = CMP(SLJIT_C_LESS, STR_PTR, 0, STR_END, 0); + /* Since we successfully read a char above, partial matching must occure. */ + check_partial(common, TRUE); + JUMPHERE(jump[0]); + } + return cc; +#endif + + case OP_EODN: + /* Requires rather complex checks. */ + jump[0] = CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0); + if (common->nltype == NLTYPE_FIXED && common->newline > 255) + { + OP2(SLJIT_ADD, TMP2, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(2)); + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(0)); + if (common->mode == JIT_COMPILE) + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, STR_END, 0)); + else + { + jump[1] = CMP(SLJIT_C_EQUAL, TMP2, 0, STR_END, 0); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP2, 0, STR_END, 0); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_LESS); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, (common->newline >> 8) & 0xff); + COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_NOT_EQUAL); + add_jump(compiler, backtracks, JUMP(SLJIT_C_NOT_EQUAL)); + check_partial(common, TRUE); + add_jump(compiler, backtracks, JUMP(SLJIT_JUMP)); + JUMPHERE(jump[1]); + } + OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(1)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, (common->newline >> 8) & 0xff)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, common->newline & 0xff)); + } + else if (common->nltype == NLTYPE_FIXED) + { + OP2(SLJIT_ADD, TMP2, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(0)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, STR_END, 0)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, common->newline)); + } + else + { + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(0)); + jump[1] = CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, CHAR_CR); + OP2(SLJIT_ADD, TMP2, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(2)); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP2, 0, STR_END, 0); + jump[2] = JUMP(SLJIT_C_GREATER); + add_jump(compiler, backtracks, JUMP(SLJIT_C_LESS)); + /* Equal. */ + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(1)); + jump[3] = CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, CHAR_NL); + add_jump(compiler, backtracks, JUMP(SLJIT_JUMP)); + + JUMPHERE(jump[1]); + if (common->nltype == NLTYPE_ANYCRLF) + { + OP2(SLJIT_ADD, TMP2, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + add_jump(compiler, backtracks, CMP(SLJIT_C_LESS, TMP2, 0, STR_END, 0)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, CHAR_NL)); + } + else + { + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1, STR_PTR, 0); + read_char(common); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, STR_PTR, 0, STR_END, 0)); + add_jump(compiler, &common->anynewline, JUMP(SLJIT_FAST_CALL)); + add_jump(compiler, backtracks, JUMP(SLJIT_C_ZERO)); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1); + } + JUMPHERE(jump[2]); + JUMPHERE(jump[3]); + } + JUMPHERE(jump[0]); + check_partial(common, FALSE); + return cc; + + case OP_EOD: + add_jump(compiler, backtracks, CMP(SLJIT_C_LESS, STR_PTR, 0, STR_END, 0)); + check_partial(common, FALSE); + return cc; + + case OP_CIRC: + OP1(SLJIT_MOV, TMP2, 0, ARGUMENTS, 0); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(TMP2), SLJIT_OFFSETOF(jit_arguments, begin)); + add_jump(compiler, backtracks, CMP(SLJIT_C_GREATER, STR_PTR, 0, TMP1, 0)); + OP1(SLJIT_MOV_UB, TMP2, 0, SLJIT_MEM1(TMP2), SLJIT_OFFSETOF(jit_arguments, notbol)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, 0)); + return cc; + + case OP_CIRCM: + OP1(SLJIT_MOV, TMP2, 0, ARGUMENTS, 0); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(TMP2), SLJIT_OFFSETOF(jit_arguments, begin)); + jump[1] = CMP(SLJIT_C_GREATER, STR_PTR, 0, TMP1, 0); + OP1(SLJIT_MOV_UB, TMP2, 0, SLJIT_MEM1(TMP2), SLJIT_OFFSETOF(jit_arguments, notbol)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, 0)); + jump[0] = JUMP(SLJIT_JUMP); + JUMPHERE(jump[1]); + + add_jump(compiler, backtracks, CMP(SLJIT_C_GREATER_EQUAL, STR_PTR, 0, STR_END, 0)); + if (common->nltype == NLTYPE_FIXED && common->newline > 255) + { + OP2(SLJIT_SUB, TMP2, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(2)); + add_jump(compiler, backtracks, CMP(SLJIT_C_LESS, TMP2, 0, TMP1, 0)); + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(-2)); + OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(-1)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, (common->newline >> 8) & 0xff)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, common->newline & 0xff)); + } + else + { + skip_char_back(common); + read_char(common); + check_newlinechar(common, common->nltype, backtracks, FALSE); + } + JUMPHERE(jump[0]); + return cc; + + case OP_DOLL: + OP1(SLJIT_MOV, TMP2, 0, ARGUMENTS, 0); + OP1(SLJIT_MOV_UB, TMP2, 0, SLJIT_MEM1(TMP2), SLJIT_OFFSETOF(jit_arguments, noteol)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, 0)); + + if (!common->endonly) + compile_char1_trypath(common, OP_EODN, cc, backtracks); + else + { + add_jump(compiler, backtracks, CMP(SLJIT_C_LESS, STR_PTR, 0, STR_END, 0)); + check_partial(common, FALSE); + } + return cc; + + case OP_DOLLM: + jump[1] = CMP(SLJIT_C_LESS, STR_PTR, 0, STR_END, 0); + OP1(SLJIT_MOV, TMP2, 0, ARGUMENTS, 0); + OP1(SLJIT_MOV_UB, TMP2, 0, SLJIT_MEM1(TMP2), SLJIT_OFFSETOF(jit_arguments, noteol)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, 0)); + check_partial(common, FALSE); + jump[0] = JUMP(SLJIT_JUMP); + JUMPHERE(jump[1]); + + if (common->nltype == NLTYPE_FIXED && common->newline > 255) + { + OP2(SLJIT_ADD, TMP2, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(2)); + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(0)); + if (common->mode == JIT_COMPILE) + add_jump(compiler, backtracks, CMP(SLJIT_C_GREATER, TMP2, 0, STR_END, 0)); + else + { + jump[1] = CMP(SLJIT_C_LESS_EQUAL, TMP2, 0, STR_END, 0); + /* STR_PTR = STR_END - IN_UCHARS(1) */ + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, (common->newline >> 8) & 0xff)); + check_partial(common, TRUE); + add_jump(compiler, backtracks, JUMP(SLJIT_JUMP)); + JUMPHERE(jump[1]); + } + + OP1(MOV_UCHAR, TMP2, 0, SLJIT_MEM1(STR_PTR), IN_UCHARS(1)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, (common->newline >> 8) & 0xff)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, common->newline & 0xff)); + } + else + { + peek_char(common); + check_newlinechar(common, common->nltype, backtracks, FALSE); + } + JUMPHERE(jump[0]); + return cc; + + case OP_CHAR: + case OP_CHARI: + length = 1; +#ifdef SUPPORT_UTF + if (common->utf && HAS_EXTRALEN(*cc)) length += GET_EXTRALEN(*cc); +#endif + if (common->mode == JIT_COMPILE && (type == OP_CHAR || !char_has_othercase(common, cc) || char_get_othercase_bit(common, cc) != 0)) + { + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(length)); + add_jump(compiler, backtracks, CMP(SLJIT_C_GREATER, STR_PTR, 0, STR_END, 0)); + + context.length = IN_UCHARS(length); + context.sourcereg = -1; +#if defined SLJIT_UNALIGNED && SLJIT_UNALIGNED + context.ucharptr = 0; +#endif + return byte_sequence_compare(common, type == OP_CHARI, cc, &context, backtracks); + } + detect_partial_match(common, backtracks); + read_char(common); +#ifdef SUPPORT_UTF + if (common->utf) + { + GETCHAR(c, cc); + } + else +#endif + c = *cc; + if (type == OP_CHAR || !char_has_othercase(common, cc)) + { + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, c)); + return cc + length; + } + oc = char_othercase(common, c); + bit = c ^ oc; + if (ispowerof2(bit)) + { + OP2(SLJIT_OR, TMP1, 0, TMP1, 0, SLJIT_IMM, bit); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, c | bit)); + return cc + length; + } + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, c); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, char_othercase(common, c)); + COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_EQUAL); + add_jump(compiler, backtracks, JUMP(SLJIT_C_ZERO)); + return cc + length; + + case OP_NOT: + case OP_NOTI: + detect_partial_match(common, backtracks); + length = 1; +#ifdef SUPPORT_UTF + if (common->utf) + { +#ifdef COMPILE_PCRE8 + c = *cc; + if (c < 128) + { + OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(STR_PTR), 0); + if (type == OP_NOT || !char_has_othercase(common, cc)) + add_jump(compiler, backtracks, CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, c)); + else + { + /* Since UTF8 code page is fixed, we know that c is in [a-z] or [A-Z] range. */ + OP2(SLJIT_OR, TMP2, 0, TMP1, 0, SLJIT_IMM, 0x20); + add_jump(compiler, backtracks, CMP(SLJIT_C_EQUAL, TMP2, 0, SLJIT_IMM, c | 0x20)); + } + /* Skip the variable-length character. */ + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1)); + jump[0] = CMP(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, 0xc0); + OP1(MOV_UCHAR, TMP1, 0, SLJIT_MEM1(TMP1), (sljit_w)PRIV(utf8_table4) - 0xc0); + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP1, 0); + JUMPHERE(jump[0]); + return cc + 1; + } + else +#endif /* COMPILE_PCRE8 */ + { + GETCHARLEN(c, cc, length); + read_char(common); + } + } + else +#endif /* SUPPORT_UTF */ + { + read_char(common); + c = *cc; + } + + if (type == OP_NOT || !char_has_othercase(common, cc)) + add_jump(compiler, backtracks, CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, c)); + else + { + oc = char_othercase(common, c); + bit = c ^ oc; + if (ispowerof2(bit)) + { + OP2(SLJIT_OR, TMP1, 0, TMP1, 0, SLJIT_IMM, bit); + add_jump(compiler, backtracks, CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, c | bit)); + } + else + { + add_jump(compiler, backtracks, CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, c)); + add_jump(compiler, backtracks, CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, oc)); + } + } + return cc + length; + + case OP_CLASS: + case OP_NCLASS: + detect_partial_match(common, backtracks); + read_char(common); +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + jump[0] = NULL; +#ifdef COMPILE_PCRE8 + /* This check only affects 8 bit mode. In other modes, we + always need to compare the value with 255. */ + if (common->utf) +#endif /* COMPILE_PCRE8 */ + { + jump[0] = CMP(SLJIT_C_GREATER, TMP1, 0, SLJIT_IMM, 255); + if (type == OP_CLASS) + { + add_jump(compiler, backtracks, jump[0]); + jump[0] = NULL; + } + } +#endif /* SUPPORT_UTF || !COMPILE_PCRE8 */ + OP2(SLJIT_AND, TMP2, 0, TMP1, 0, SLJIT_IMM, 0x7); + OP2(SLJIT_LSHR, TMP1, 0, TMP1, 0, SLJIT_IMM, 3); + OP1(SLJIT_MOV_UB, TMP1, 0, SLJIT_MEM1(TMP1), (sljit_w)cc); + OP2(SLJIT_SHL, TMP2, 0, SLJIT_IMM, 1, TMP2, 0); + OP2(SLJIT_AND | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, TMP2, 0); + add_jump(compiler, backtracks, JUMP(SLJIT_C_ZERO)); +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + if (jump[0] != NULL) + JUMPHERE(jump[0]); +#endif /* SUPPORT_UTF || !COMPILE_PCRE8 */ + return cc + 32 / sizeof(pcre_uchar); + +#if defined SUPPORT_UTF || defined COMPILE_PCRE16 + case OP_XCLASS: + compile_xclass_trypath(common, cc + LINK_SIZE, backtracks); + return cc + GET(cc, 0) - 1; +#endif + + case OP_REVERSE: + length = GET(cc, 0); + if (length == 0) + return cc + LINK_SIZE; + OP1(SLJIT_MOV, TMP1, 0, ARGUMENTS, 0); +#ifdef SUPPORT_UTF + if (common->utf) + { + OP1(SLJIT_MOV, TMP3, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, begin)); + OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, length); + label = LABEL(); + add_jump(compiler, backtracks, CMP(SLJIT_C_LESS_EQUAL, STR_PTR, 0, TMP3, 0)); + skip_char_back(common); + OP2(SLJIT_SUB | SLJIT_SET_E, TMP2, 0, TMP2, 0, SLJIT_IMM, 1); + JUMPTO(SLJIT_C_NOT_ZERO, label); + } + else +#endif + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, begin)); + OP2(SLJIT_SUB, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(length)); + add_jump(compiler, backtracks, CMP(SLJIT_C_LESS, STR_PTR, 0, TMP1, 0)); + } + check_start_used_ptr(common); + return cc + LINK_SIZE; + } +SLJIT_ASSERT_STOP(); +return cc; +} + +static SLJIT_INLINE pcre_uchar *compile_charn_trypath(compiler_common *common, pcre_uchar *cc, pcre_uchar *ccend, jump_list **backtracks) +{ +/* This function consumes at least one input character. */ +/* To decrease the number of length checks, we try to concatenate the fixed length character sequences. */ +DEFINE_COMPILER; +pcre_uchar *ccbegin = cc; +compare_context context; +int size; + +context.length = 0; +do + { + if (cc >= ccend) + break; + + if (*cc == OP_CHAR) + { + size = 1; +#ifdef SUPPORT_UTF + if (common->utf && HAS_EXTRALEN(cc[1])) + size += GET_EXTRALEN(cc[1]); +#endif + } + else if (*cc == OP_CHARI) + { + size = 1; +#ifdef SUPPORT_UTF + if (common->utf) + { + if (char_has_othercase(common, cc + 1) && char_get_othercase_bit(common, cc + 1) == 0) + size = 0; + else if (HAS_EXTRALEN(cc[1])) + size += GET_EXTRALEN(cc[1]); + } + else +#endif + if (char_has_othercase(common, cc + 1) && char_get_othercase_bit(common, cc + 1) == 0) + size = 0; + } + else + size = 0; + + cc += 1 + size; + context.length += IN_UCHARS(size); + } +while (size > 0 && context.length <= 128); + +cc = ccbegin; +if (context.length > 0) + { + /* We have a fixed-length byte sequence. */ + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, context.length); + add_jump(compiler, backtracks, CMP(SLJIT_C_GREATER, STR_PTR, 0, STR_END, 0)); + + context.sourcereg = -1; +#if defined SLJIT_UNALIGNED && SLJIT_UNALIGNED + context.ucharptr = 0; +#endif + do cc = byte_sequence_compare(common, *cc == OP_CHARI, cc + 1, &context, backtracks); while (context.length > 0); + return cc; + } + +/* A non-fixed length character will be checked if length == 0. */ +return compile_char1_trypath(common, *cc, cc + 1, backtracks); +} + +static struct sljit_jump *compile_ref_checks(compiler_common *common, pcre_uchar *cc, jump_list **backtracks) +{ +DEFINE_COMPILER; +int offset = GET2(cc, 1) << 1; + +OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset)); +if (!common->jscript_compat) + { + if (backtracks == NULL) + { + /* OVECTOR(1) contains the "string begin - 1" constant. */ + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(1)); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_EQUAL); + OP2(SLJIT_SUB | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1)); + COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_EQUAL); + return JUMP(SLJIT_C_NOT_ZERO); + } + add_jump(compiler, backtracks, CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(1))); + } +return CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1)); +} + +/* Forward definitions. */ +static void compile_trypath(compiler_common *, pcre_uchar *, pcre_uchar *, backtrack_common *); +static void compile_backtrackpath(compiler_common *, struct backtrack_common *); + +#define PUSH_BACKTRACK(size, ccstart, error) \ + do \ + { \ + backtrack = sljit_alloc_memory(compiler, (size)); \ + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) \ + return error; \ + memset(backtrack, 0, size); \ + backtrack->prev = parent->top; \ + backtrack->cc = (ccstart); \ + parent->top = backtrack; \ + } \ + while (0) + +#define PUSH_BACKTRACK_NOVALUE(size, ccstart) \ + do \ + { \ + backtrack = sljit_alloc_memory(compiler, (size)); \ + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) \ + return; \ + memset(backtrack, 0, size); \ + backtrack->prev = parent->top; \ + backtrack->cc = (ccstart); \ + parent->top = backtrack; \ + } \ + while (0) + +#define BACKTRACK_AS(type) ((type *)backtrack) + +static pcre_uchar *compile_ref_trypath(compiler_common *common, pcre_uchar *cc, jump_list **backtracks, BOOL withchecks, BOOL emptyfail) +{ +DEFINE_COMPILER; +int offset = GET2(cc, 1) << 1; +struct sljit_jump *jump = NULL; +struct sljit_jump *partial; +struct sljit_jump *nopartial; + +OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset)); +/* OVECTOR(1) contains the "string begin - 1" constant. */ +if (withchecks && !common->jscript_compat) + add_jump(compiler, backtracks, CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(1))); + +#if defined SUPPORT_UTF && defined SUPPORT_UCP +if (common->utf && *cc == OP_REFI) + { + SLJIT_ASSERT(TMP1 == SLJIT_TEMPORARY_REG1 && STACK_TOP == SLJIT_TEMPORARY_REG2 && TMP2 == SLJIT_TEMPORARY_REG3); + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1)); + if (withchecks) + jump = CMP(SLJIT_C_EQUAL, TMP1, 0, TMP2, 0); + + /* Needed to save important temporary registers. */ + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0, STACK_TOP, 0); + OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG2, 0, ARGUMENTS, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_TEMPORARY_REG2), SLJIT_OFFSETOF(jit_arguments, uchar_ptr), STR_PTR, 0); + sljit_emit_ijump(compiler, SLJIT_CALL3, SLJIT_IMM, SLJIT_FUNC_OFFSET(do_utf_caselesscmp)); + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0); + if (common->mode == JIT_COMPILE) + add_jump(compiler, backtracks, CMP(SLJIT_C_LESS_EQUAL, SLJIT_RETURN_REG, 0, SLJIT_IMM, 1)); + else + { + add_jump(compiler, backtracks, CMP(SLJIT_C_EQUAL, SLJIT_RETURN_REG, 0, SLJIT_IMM, 0)); + nopartial = CMP(SLJIT_C_NOT_EQUAL, SLJIT_RETURN_REG, 0, SLJIT_IMM, 1); + check_partial(common, FALSE); + add_jump(compiler, backtracks, JUMP(SLJIT_JUMP)); + JUMPHERE(nopartial); + } + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_RETURN_REG, 0); + } +else +#endif /* SUPPORT_UTF && SUPPORT_UCP */ + { + OP2(SLJIT_SUB | SLJIT_SET_E, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1), TMP1, 0); + if (withchecks) + jump = JUMP(SLJIT_C_ZERO); + + OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, TMP2, 0); + partial = CMP(SLJIT_C_GREATER, STR_PTR, 0, STR_END, 0); + if (common->mode == JIT_COMPILE) + add_jump(compiler, backtracks, partial); + + add_jump(compiler, *cc == OP_REF ? &common->casefulcmp : &common->caselesscmp, JUMP(SLJIT_FAST_CALL)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, 0)); + + if (common->mode != JIT_COMPILE) + { + nopartial = JUMP(SLJIT_JUMP); + JUMPHERE(partial); + /* TMP2 -= STR_END - STR_PTR */ + OP2(SLJIT_SUB, TMP2, 0, TMP2, 0, STR_PTR, 0); + OP2(SLJIT_ADD, TMP2, 0, TMP2, 0, STR_END, 0); + partial = CMP(SLJIT_C_EQUAL, TMP2, 0, SLJIT_IMM, 0); + OP1(SLJIT_MOV, STR_PTR, 0, STR_END, 0); + add_jump(compiler, *cc == OP_REF ? &common->casefulcmp : &common->caselesscmp, JUMP(SLJIT_FAST_CALL)); + add_jump(compiler, backtracks, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, 0)); + JUMPHERE(partial); + check_partial(common, FALSE); + add_jump(compiler, backtracks, JUMP(SLJIT_JUMP)); + JUMPHERE(nopartial); + } + } + +if (jump != NULL) + { + if (emptyfail) + add_jump(compiler, backtracks, jump); + else + JUMPHERE(jump); + } +return cc + 1 + IMM2_SIZE; +} + +static SLJIT_INLINE pcre_uchar *compile_ref_iterator_trypath(compiler_common *common, pcre_uchar *cc, backtrack_common *parent) +{ +DEFINE_COMPILER; +backtrack_common *backtrack; +pcre_uchar type; +struct sljit_label *label; +struct sljit_jump *zerolength; +struct sljit_jump *jump = NULL; +pcre_uchar *ccbegin = cc; +int min = 0, max = 0; +BOOL minimize; + +PUSH_BACKTRACK(sizeof(iterator_backtrack), cc, NULL); + +type = cc[1 + IMM2_SIZE]; +minimize = (type & 0x1) != 0; +switch(type) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + min = 0; + max = 0; + cc += 1 + IMM2_SIZE + 1; + break; + case OP_CRPLUS: + case OP_CRMINPLUS: + min = 1; + max = 0; + cc += 1 + IMM2_SIZE + 1; + break; + case OP_CRQUERY: + case OP_CRMINQUERY: + min = 0; + max = 1; + cc += 1 + IMM2_SIZE + 1; + break; + case OP_CRRANGE: + case OP_CRMINRANGE: + min = GET2(cc, 1 + IMM2_SIZE + 1); + max = GET2(cc, 1 + IMM2_SIZE + 1 + IMM2_SIZE); + cc += 1 + IMM2_SIZE + 1 + 2 * IMM2_SIZE; + break; + default: + SLJIT_ASSERT_STOP(); + break; + } + +if (!minimize) + { + if (min == 0) + { + allocate_stack(common, 2); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), SLJIT_IMM, 0); + /* Temporary release of STR_PTR. */ + OP2(SLJIT_SUB, STACK_TOP, 0, STACK_TOP, 0, SLJIT_IMM, sizeof(sljit_w)); + zerolength = compile_ref_checks(common, ccbegin, NULL); + /* Restore if not zero length. */ + OP2(SLJIT_ADD, STACK_TOP, 0, STACK_TOP, 0, SLJIT_IMM, sizeof(sljit_w)); + } + else + { + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + zerolength = compile_ref_checks(common, ccbegin, &backtrack->topbacktracks); + } + + if (min > 1 || max > 1) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, SLJIT_IMM, 0); + + label = LABEL(); + compile_ref_trypath(common, ccbegin, &backtrack->topbacktracks, FALSE, FALSE); + + if (min > 1 || max > 1) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0); + OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, TMP1, 0); + if (min > 1) + CMPTO(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, min, label); + if (max > 1) + { + jump = CMP(SLJIT_C_GREATER_EQUAL, TMP1, 0, SLJIT_IMM, max); + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + JUMPTO(SLJIT_JUMP, label); + JUMPHERE(jump); + } + } + + if (max == 0) + { + /* Includes min > 1 case as well. */ + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + JUMPTO(SLJIT_JUMP, label); + } + + JUMPHERE(zerolength); + BACKTRACK_AS(iterator_backtrack)->trypath = LABEL(); + + decrease_call_count(common); + return cc; + } + +allocate_stack(common, 2); +OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); +if (type != OP_CRMINSTAR) + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), SLJIT_IMM, 0); + +if (min == 0) + { + zerolength = compile_ref_checks(common, ccbegin, NULL); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + jump = JUMP(SLJIT_JUMP); + } +else + zerolength = compile_ref_checks(common, ccbegin, &backtrack->topbacktracks); + +BACKTRACK_AS(iterator_backtrack)->trypath = LABEL(); +if (max > 0) + add_jump(compiler, &backtrack->topbacktracks, CMP(SLJIT_C_GREATER_EQUAL, SLJIT_MEM1(STACK_TOP), STACK(1), SLJIT_IMM, max)); + +compile_ref_trypath(common, ccbegin, &backtrack->topbacktracks, TRUE, TRUE); +OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + +if (min > 1) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), TMP1, 0); + CMPTO(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, min, BACKTRACK_AS(iterator_backtrack)->trypath); + } +else if (max > 0) + OP2(SLJIT_ADD, SLJIT_MEM1(STACK_TOP), STACK(1), SLJIT_MEM1(STACK_TOP), STACK(1), SLJIT_IMM, 1); + +if (jump != NULL) + JUMPHERE(jump); +JUMPHERE(zerolength); + +decrease_call_count(common); +return cc; +} + +static SLJIT_INLINE pcre_uchar *compile_recurse_trypath(compiler_common *common, pcre_uchar *cc, backtrack_common *parent) +{ +DEFINE_COMPILER; +backtrack_common *backtrack; +recurse_entry *entry = common->entries; +recurse_entry *prev = NULL; +int start = GET(cc, 1); + +PUSH_BACKTRACK(sizeof(recurse_backtrack), cc, NULL); +while (entry != NULL) + { + if (entry->start == start) + break; + prev = entry; + entry = entry->next; + } + +if (entry == NULL) + { + entry = sljit_alloc_memory(compiler, sizeof(recurse_entry)); + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + return NULL; + entry->next = NULL; + entry->entry = NULL; + entry->calls = NULL; + entry->start = start; + + if (prev != NULL) + prev->next = entry; + else + common->entries = entry; + } + +if (common->has_set_som && common->mark_ptr != 0) + { + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0)); + allocate_stack(common, 2); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->mark_ptr); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), TMP2, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), TMP1, 0); + } +else if (common->has_set_som || common->mark_ptr != 0) + { + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->has_set_som ? (int)(OVECTOR(0)) : common->mark_ptr); + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), TMP2, 0); + } + +if (entry->entry == NULL) + add_jump(compiler, &entry->calls, JUMP(SLJIT_FAST_CALL)); +else + JUMPTO(SLJIT_FAST_CALL, entry->entry); +/* Leave if the match is failed. */ +add_jump(compiler, &backtrack->topbacktracks, CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, 0)); +return cc + 1 + LINK_SIZE; +} + +static pcre_uchar *compile_assert_trypath(compiler_common *common, pcre_uchar *cc, assert_backtrack *backtrack, BOOL conditional) +{ +DEFINE_COMPILER; +int framesize; +int localptr; +backtrack_common altbacktrack; +pcre_uchar *ccbegin; +pcre_uchar opcode; +pcre_uchar bra = OP_BRA; +jump_list *tmp = NULL; +jump_list **target = (conditional) ? &backtrack->condfailed : &backtrack->common.topbacktracks; +jump_list **found; +/* Saving previous accept variables. */ +struct sljit_label *save_leavelabel = common->leavelabel; +struct sljit_label *save_acceptlabel = common->acceptlabel; +jump_list *save_leave = common->leave; +jump_list *save_accept = common->accept; +struct sljit_jump *jump; +struct sljit_jump *brajump = NULL; + +if (*cc == OP_BRAZERO || *cc == OP_BRAMINZERO) + { + SLJIT_ASSERT(!conditional); + bra = *cc; + cc++; + } +localptr = PRIV_DATA(cc); +SLJIT_ASSERT(localptr != 0); +framesize = get_framesize(common, cc, FALSE); +backtrack->framesize = framesize; +backtrack->localptr = localptr; +opcode = *cc; +SLJIT_ASSERT(opcode >= OP_ASSERT && opcode <= OP_ASSERTBACK_NOT); +found = (opcode == OP_ASSERT || opcode == OP_ASSERTBACK) ? &tmp : target; +ccbegin = cc; +cc += GET(cc, 1); + +if (bra == OP_BRAMINZERO) + { + /* This is a braminzero backtrack path. */ + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + free_stack(common, 1); + brajump = CMP(SLJIT_C_EQUAL, STR_PTR, 0, SLJIT_IMM, 0); + } + +if (framesize < 0) + { + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, STACK_TOP, 0); + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + } +else + { + allocate_stack(common, framesize + 2); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + OP2(SLJIT_SUB, TMP2, 0, STACK_TOP, 0, SLJIT_IMM, -STACK(framesize + 1)); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, TMP2, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), TMP1, 0); + init_frame(common, ccbegin, framesize + 1, 2, FALSE); + } + +memset(&altbacktrack, 0, sizeof(backtrack_common)); +common->leavelabel = NULL; +common->leave = NULL; +while (1) + { + common->acceptlabel = NULL; + common->accept = NULL; + altbacktrack.top = NULL; + altbacktrack.topbacktracks = NULL; + + if (*ccbegin == OP_ALT) + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + + altbacktrack.cc = ccbegin; + compile_trypath(common, ccbegin + 1 + LINK_SIZE, cc, &altbacktrack); + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + { + common->leavelabel = save_leavelabel; + common->acceptlabel = save_acceptlabel; + common->leave = save_leave; + common->accept = save_accept; + return NULL; + } + common->acceptlabel = LABEL(); + if (common->accept != NULL) + set_jumps(common->accept, common->acceptlabel); + + /* Reset stack. */ + if (framesize < 0) + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + else { + if ((opcode != OP_ASSERT_NOT && opcode != OP_ASSERTBACK_NOT) || conditional) + { + /* We don't need to keep the STR_PTR, only the previous localptr. */ + OP2(SLJIT_ADD, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_IMM, (framesize + 1) * sizeof(sljit_w)); + } + else + { + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + add_jump(compiler, &common->revertframes, JUMP(SLJIT_FAST_CALL)); + } + } + + if (opcode == OP_ASSERT_NOT || opcode == OP_ASSERTBACK_NOT) + { + /* We know that STR_PTR was stored on the top of the stack. */ + if (conditional) + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), 0); + else if (bra == OP_BRAZERO) + { + if (framesize < 0) + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), 0); + else + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), framesize * sizeof(sljit_w)); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), (framesize + 1) * sizeof(sljit_w)); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, TMP1, 0); + } + OP2(SLJIT_ADD, STACK_TOP, 0, STACK_TOP, 0, SLJIT_IMM, sizeof(sljit_w)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + } + else if (framesize >= 0) + { + /* For OP_BRA and OP_BRAMINZERO. */ + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_MEM1(STACK_TOP), framesize * sizeof(sljit_w)); + } + } + add_jump(compiler, found, JUMP(SLJIT_JUMP)); + + compile_backtrackpath(common, altbacktrack.top); + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + { + common->leavelabel = save_leavelabel; + common->acceptlabel = save_acceptlabel; + common->leave = save_leave; + common->accept = save_accept; + return NULL; + } + set_jumps(altbacktrack.topbacktracks, LABEL()); + + if (*cc != OP_ALT) + break; + + ccbegin = cc; + cc += GET(cc, 1); + } +/* None of them matched. */ +if (common->leave != NULL) + set_jumps(common->leave, LABEL()); + +if (opcode == OP_ASSERT || opcode == OP_ASSERTBACK) + { + /* Assert is failed. */ + if (conditional || bra == OP_BRAZERO) + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + + if (framesize < 0) + { + /* The topmost item should be 0. */ + if (bra == OP_BRAZERO) + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + else + free_stack(common, 1); + } + else + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + /* The topmost item should be 0. */ + if (bra == OP_BRAZERO) + { + free_stack(common, framesize + 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + } + else + free_stack(common, framesize + 2); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, TMP1, 0); + } + jump = JUMP(SLJIT_JUMP); + if (bra != OP_BRAZERO) + add_jump(compiler, target, jump); + + /* Assert is successful. */ + set_jumps(tmp, LABEL()); + if (framesize < 0) + { + /* We know that STR_PTR was stored on the top of the stack. */ + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), 0); + /* Keep the STR_PTR on the top of the stack. */ + if (bra == OP_BRAZERO) + OP2(SLJIT_ADD, STACK_TOP, 0, STACK_TOP, 0, SLJIT_IMM, sizeof(sljit_w)); + else if (bra == OP_BRAMINZERO) + { + OP2(SLJIT_ADD, STACK_TOP, 0, STACK_TOP, 0, SLJIT_IMM, sizeof(sljit_w)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + } + } + else + { + if (bra == OP_BRA) + { + /* We don't need to keep the STR_PTR, only the previous localptr. */ + OP2(SLJIT_ADD, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_IMM, (framesize + 1) * sizeof(sljit_w)); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), 0); + } + else + { + /* We don't need to keep the STR_PTR, only the previous localptr. */ + OP2(SLJIT_ADD, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_IMM, (framesize + 2) * sizeof(sljit_w)); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), bra == OP_BRAZERO ? STR_PTR : SLJIT_IMM, 0); + } + } + + if (bra == OP_BRAZERO) + { + backtrack->trypath = LABEL(); + sljit_set_label(jump, backtrack->trypath); + } + else if (bra == OP_BRAMINZERO) + { + JUMPTO(SLJIT_JUMP, backtrack->trypath); + JUMPHERE(brajump); + if (framesize >= 0) + { + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + add_jump(compiler, &common->revertframes, JUMP(SLJIT_FAST_CALL)); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_MEM1(STACK_TOP), framesize * sizeof(sljit_w)); + } + set_jumps(backtrack->common.topbacktracks, LABEL()); + } + } +else + { + /* AssertNot is successful. */ + if (framesize < 0) + { + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + if (bra != OP_BRA) + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + else + free_stack(common, 1); + } + else + { + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + /* The topmost item should be 0. */ + if (bra != OP_BRA) + { + free_stack(common, framesize + 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + } + else + free_stack(common, framesize + 2); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, TMP1, 0); + } + + if (bra == OP_BRAZERO) + backtrack->trypath = LABEL(); + else if (bra == OP_BRAMINZERO) + { + JUMPTO(SLJIT_JUMP, backtrack->trypath); + JUMPHERE(brajump); + } + + if (bra != OP_BRA) + { + SLJIT_ASSERT(found == &backtrack->common.topbacktracks); + set_jumps(backtrack->common.topbacktracks, LABEL()); + backtrack->common.topbacktracks = NULL; + } + } + +common->leavelabel = save_leavelabel; +common->acceptlabel = save_acceptlabel; +common->leave = save_leave; +common->accept = save_accept; +return cc + 1 + LINK_SIZE; +} + +static sljit_w SLJIT_CALL do_searchovector(sljit_w refno, sljit_w* locals, pcre_uchar *name_table) +{ +int condition = FALSE; +pcre_uchar *slotA = name_table; +pcre_uchar *slotB; +sljit_w name_count = locals[LOCALS0 / sizeof(sljit_w)]; +sljit_w name_entry_size = locals[LOCALS1 / sizeof(sljit_w)]; +sljit_w no_capture; +int i; + +locals += refno & 0xff; +refno >>= 8; +no_capture = locals[1]; + +for (i = 0; i < name_count; i++) + { + if (GET2(slotA, 0) == refno) break; + slotA += name_entry_size; + } + +if (i < name_count) + { + /* Found a name for the number - there can be only one; duplicate names + for different numbers are allowed, but not vice versa. First scan down + for duplicates. */ + + slotB = slotA; + while (slotB > name_table) + { + slotB -= name_entry_size; + if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0) + { + condition = locals[GET2(slotB, 0) << 1] != no_capture; + if (condition) break; + } + else break; + } + + /* Scan up for duplicates */ + if (!condition) + { + slotB = slotA; + for (i++; i < name_count; i++) + { + slotB += name_entry_size; + if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0) + { + condition = locals[GET2(slotB, 0) << 1] != no_capture; + if (condition) break; + } + else break; + } + } + } +return condition; +} + +static sljit_w SLJIT_CALL do_searchgroups(sljit_w recno, sljit_w* locals, pcre_uchar *name_table) +{ +int condition = FALSE; +pcre_uchar *slotA = name_table; +pcre_uchar *slotB; +sljit_w name_count = locals[LOCALS0 / sizeof(sljit_w)]; +sljit_w name_entry_size = locals[LOCALS1 / sizeof(sljit_w)]; +sljit_w group_num = locals[POSSESSIVE0 / sizeof(sljit_w)]; +int i; + +for (i = 0; i < name_count; i++) + { + if (GET2(slotA, 0) == recno) break; + slotA += name_entry_size; + } + +if (i < name_count) + { + /* Found a name for the number - there can be only one; duplicate + names for different numbers are allowed, but not vice versa. First + scan down for duplicates. */ + + slotB = slotA; + while (slotB > name_table) + { + slotB -= name_entry_size; + if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0) + { + condition = GET2(slotB, 0) == group_num; + if (condition) break; + } + else break; + } + + /* Scan up for duplicates */ + if (!condition) + { + slotB = slotA; + for (i++; i < name_count; i++) + { + slotB += name_entry_size; + if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0) + { + condition = GET2(slotB, 0) == group_num; + if (condition) break; + } + else break; + } + } + } +return condition; +} + +/* + Handling bracketed expressions is probably the most complex part. + + Stack layout naming characters: + S - Push the current STR_PTR + 0 - Push a 0 (NULL) + A - Push the current STR_PTR. Needed for restoring the STR_PTR + before the next alternative. Not pushed if there are no alternatives. + M - Any values pushed by the current alternative. Can be empty, or anything. + C - Push the previous OVECTOR(i), OVECTOR(i+1) and OVECTOR_PRIV(i) to the stack. + L - Push the previous local (pointed by localptr) to the stack + () - opional values stored on the stack + ()* - optonal, can be stored multiple times + + The following list shows the regular expression templates, their PCRE byte codes + and stack layout supported by pcre-sljit. + + (?:) OP_BRA | OP_KET A M + () OP_CBRA | OP_KET C M + (?:)+ OP_BRA | OP_KETRMAX 0 A M S ( A M S )* + OP_SBRA | OP_KETRMAX 0 L M S ( L M S )* + (?:)+? OP_BRA | OP_KETRMIN 0 A M S ( A M S )* + OP_SBRA | OP_KETRMIN 0 L M S ( L M S )* + ()+ OP_CBRA | OP_KETRMAX 0 C M S ( C M S )* + OP_SCBRA | OP_KETRMAX 0 C M S ( C M S )* + ()+? OP_CBRA | OP_KETRMIN 0 C M S ( C M S )* + OP_SCBRA | OP_KETRMIN 0 C M S ( C M S )* + (?:)? OP_BRAZERO | OP_BRA | OP_KET S ( A M 0 ) + (?:)?? OP_BRAMINZERO | OP_BRA | OP_KET S ( A M 0 ) + ()? OP_BRAZERO | OP_CBRA | OP_KET S ( C M 0 ) + ()?? OP_BRAMINZERO | OP_CBRA | OP_KET S ( C M 0 ) + (?:)* OP_BRAZERO | OP_BRA | OP_KETRMAX S 0 ( A M S )* + OP_BRAZERO | OP_SBRA | OP_KETRMAX S 0 ( L M S )* + (?:)*? OP_BRAMINZERO | OP_BRA | OP_KETRMIN S 0 ( A M S )* + OP_BRAMINZERO | OP_SBRA | OP_KETRMIN S 0 ( L M S )* + ()* OP_BRAZERO | OP_CBRA | OP_KETRMAX S 0 ( C M S )* + OP_BRAZERO | OP_SCBRA | OP_KETRMAX S 0 ( C M S )* + ()*? OP_BRAMINZERO | OP_CBRA | OP_KETRMIN S 0 ( C M S )* + OP_BRAMINZERO | OP_SCBRA | OP_KETRMIN S 0 ( C M S )* + + + Stack layout naming characters: + A - Push the alternative index (starting from 0) on the stack. + Not pushed if there is no alternatives. + M - Any values pushed by the current alternative. Can be empty, or anything. + + The next list shows the possible content of a bracket: + (|) OP_*BRA | OP_ALT ... M A + (?()|) OP_*COND | OP_ALT M A + (?>|) OP_ONCE | OP_ALT ... [stack trace] M A + (?>|) OP_ONCE_NC | OP_ALT ... [stack trace] M A + Or nothing, if trace is unnecessary +*/ + +static pcre_uchar *compile_bracket_trypath(compiler_common *common, pcre_uchar *cc, backtrack_common *parent) +{ +DEFINE_COMPILER; +backtrack_common *backtrack; +pcre_uchar opcode; +int localptr = 0; +int offset = 0; +int stacksize; +pcre_uchar *ccbegin; +pcre_uchar *trypath; +pcre_uchar bra = OP_BRA; +pcre_uchar ket; +assert_backtrack *assert; +BOOL has_alternatives; +struct sljit_jump *jump; +struct sljit_jump *skip; +struct sljit_label *rmaxlabel = NULL; +struct sljit_jump *braminzerojump = NULL; + +PUSH_BACKTRACK(sizeof(bracket_backtrack), cc, NULL); + +if (*cc == OP_BRAZERO || *cc == OP_BRAMINZERO) + { + bra = *cc; + cc++; + opcode = *cc; + } + +opcode = *cc; +ccbegin = cc; +trypath = ccbegin + 1 + LINK_SIZE; + +if ((opcode == OP_COND || opcode == OP_SCOND) && cc[1 + LINK_SIZE] == OP_DEF) + { + /* Drop this bracket_backtrack. */ + parent->top = backtrack->prev; + return bracketend(cc); + } + +ket = *(bracketend(cc) - 1 - LINK_SIZE); +SLJIT_ASSERT(ket == OP_KET || ket == OP_KETRMAX || ket == OP_KETRMIN); +SLJIT_ASSERT(!((bra == OP_BRAZERO && ket == OP_KETRMIN) || (bra == OP_BRAMINZERO && ket == OP_KETRMAX))); +cc += GET(cc, 1); + +has_alternatives = *cc == OP_ALT; +if (SLJIT_UNLIKELY(opcode == OP_COND) || SLJIT_UNLIKELY(opcode == OP_SCOND)) + { + has_alternatives = (*trypath == OP_RREF) ? FALSE : TRUE; + if (*trypath == OP_NRREF) + { + stacksize = GET2(trypath, 1); + if (common->currententry == NULL || stacksize == RREF_ANY) + has_alternatives = FALSE; + else if (common->currententry->start == 0) + has_alternatives = stacksize != 0; + else + has_alternatives = stacksize != GET2(common->start, common->currententry->start + 1 + LINK_SIZE); + } + } + +if (SLJIT_UNLIKELY(opcode == OP_COND) && (*cc == OP_KETRMAX || *cc == OP_KETRMIN)) + opcode = OP_SCOND; +if (SLJIT_UNLIKELY(opcode == OP_ONCE_NC)) + opcode = OP_ONCE; + +if (opcode == OP_CBRA || opcode == OP_SCBRA) + { + /* Capturing brackets has a pre-allocated space. */ + offset = GET2(ccbegin, 1 + LINK_SIZE); + localptr = OVECTOR_PRIV(offset); + offset <<= 1; + BACKTRACK_AS(bracket_backtrack)->localptr = localptr; + trypath += IMM2_SIZE; + } +else if (opcode == OP_ONCE || opcode == OP_SBRA || opcode == OP_SCOND) + { + /* Other brackets simply allocate the next entry. */ + localptr = PRIV_DATA(ccbegin); + SLJIT_ASSERT(localptr != 0); + BACKTRACK_AS(bracket_backtrack)->localptr = localptr; + if (opcode == OP_ONCE) + BACKTRACK_AS(bracket_backtrack)->u.framesize = get_framesize(common, ccbegin, FALSE); + } + +/* Instructions before the first alternative. */ +stacksize = 0; +if ((ket == OP_KETRMAX) || (ket == OP_KETRMIN && bra != OP_BRAMINZERO)) + stacksize++; +if (bra == OP_BRAZERO) + stacksize++; + +if (stacksize > 0) + allocate_stack(common, stacksize); + +stacksize = 0; +if ((ket == OP_KETRMAX) || (ket == OP_KETRMIN && bra != OP_BRAMINZERO)) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stacksize), SLJIT_IMM, 0); + stacksize++; + } + +if (bra == OP_BRAZERO) + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stacksize), STR_PTR, 0); + +if (bra == OP_BRAMINZERO) + { + /* This is a backtrack path! (Since the try-path of OP_BRAMINZERO matches to the empty string) */ + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + if (ket != OP_KETRMIN) + { + free_stack(common, 1); + braminzerojump = CMP(SLJIT_C_EQUAL, STR_PTR, 0, SLJIT_IMM, 0); + } + else + { + if (opcode == OP_ONCE || opcode >= OP_SBRA) + { + jump = CMP(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_IMM, 0); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + /* Nothing stored during the first run. */ + skip = JUMP(SLJIT_JUMP); + JUMPHERE(jump); + /* Checking zero-length iteration. */ + if (opcode != OP_ONCE || BACKTRACK_AS(bracket_backtrack)->u.framesize < 0) + { + /* When we come from outside, localptr contains the previous STR_PTR. */ + braminzerojump = CMP(SLJIT_C_EQUAL, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + } + else + { + /* Except when the whole stack frame must be saved. */ + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + braminzerojump = CMP(SLJIT_C_EQUAL, STR_PTR, 0, SLJIT_MEM1(TMP1), (BACKTRACK_AS(bracket_backtrack)->u.framesize + 1) * sizeof(sljit_w)); + } + JUMPHERE(skip); + } + else + { + jump = CMP(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_IMM, 0); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + JUMPHERE(jump); + } + } + } + +if (ket == OP_KETRMIN) + BACKTRACK_AS(bracket_backtrack)->recursivetrypath = LABEL(); + +if (ket == OP_KETRMAX) + { + rmaxlabel = LABEL(); + if (has_alternatives && opcode != OP_ONCE && opcode < OP_SBRA) + BACKTRACK_AS(bracket_backtrack)->alttrypath = rmaxlabel; + } + +/* Handling capturing brackets and alternatives. */ +if (opcode == OP_ONCE) + { + if (BACKTRACK_AS(bracket_backtrack)->u.framesize < 0) + { + /* Neither capturing brackets nor recursions are not found in the block. */ + if (ket == OP_KETRMIN) + { + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + allocate_stack(common, 2); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), TMP2, 0); + OP2(SLJIT_SUB, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, STACK_TOP, 0, SLJIT_IMM, sizeof(sljit_w)); + } + else if (ket == OP_KETRMAX || has_alternatives) + { + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, STACK_TOP, 0); + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + } + else + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, STACK_TOP, 0); + } + else + { + if (ket == OP_KETRMIN || ket == OP_KETRMAX || has_alternatives) + { + allocate_stack(common, BACKTRACK_AS(bracket_backtrack)->u.framesize + 2); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + OP2(SLJIT_SUB, TMP2, 0, STACK_TOP, 0, SLJIT_IMM, -STACK(BACKTRACK_AS(bracket_backtrack)->u.framesize + 1)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, TMP2, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), TMP1, 0); + init_frame(common, ccbegin, BACKTRACK_AS(bracket_backtrack)->u.framesize + 1, 2, FALSE); + } + else + { + allocate_stack(common, BACKTRACK_AS(bracket_backtrack)->u.framesize + 1); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + OP2(SLJIT_SUB, TMP2, 0, STACK_TOP, 0, SLJIT_IMM, -STACK(BACKTRACK_AS(bracket_backtrack)->u.framesize)); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, TMP2, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), TMP1, 0); + init_frame(common, ccbegin, BACKTRACK_AS(bracket_backtrack)->u.framesize, 1, FALSE); + } + } + } +else if (opcode == OP_CBRA || opcode == OP_SCBRA) + { + /* Saving the previous values. */ + allocate_stack(common, 3); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset)); + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), TMP1, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), TMP2, 0); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(2), TMP1, 0); + } +else if (opcode == OP_SBRA || opcode == OP_SCOND) + { + /* Saving the previous value. */ + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), TMP2, 0); + } +else if (has_alternatives) + { + /* Pushing the starting string pointer. */ + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + } + +/* Generating code for the first alternative. */ +if (opcode == OP_COND || opcode == OP_SCOND) + { + if (*trypath == OP_CREF) + { + SLJIT_ASSERT(has_alternatives); + add_jump(compiler, &(BACKTRACK_AS(bracket_backtrack)->u.condfailed), + CMP(SLJIT_C_EQUAL, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(GET2(trypath, 1) << 1), SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(1))); + trypath += 1 + IMM2_SIZE; + } + else if (*trypath == OP_NCREF) + { + SLJIT_ASSERT(has_alternatives); + stacksize = GET2(trypath, 1); + jump = CMP(SLJIT_C_NOT_EQUAL, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(stacksize << 1), SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(1)); + + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE1, STACK_TOP, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0, SLJIT_IMM, common->name_count); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1, SLJIT_IMM, common->name_entry_size); + OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG1, 0, SLJIT_IMM, (stacksize << 8) | (common->ovector_start / sizeof(sljit_w))); + GET_LOCAL_BASE(SLJIT_TEMPORARY_REG2, 0, 0); + OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG3, 0, SLJIT_IMM, common->name_table); + sljit_emit_ijump(compiler, SLJIT_CALL3, SLJIT_IMM, SLJIT_FUNC_OFFSET(do_searchovector)); + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE1); + add_jump(compiler, &(BACKTRACK_AS(bracket_backtrack)->u.condfailed), CMP(SLJIT_C_EQUAL, SLJIT_TEMPORARY_REG1, 0, SLJIT_IMM, 0)); + + JUMPHERE(jump); + trypath += 1 + IMM2_SIZE; + } + else if (*trypath == OP_RREF || *trypath == OP_NRREF) + { + /* Never has other case. */ + BACKTRACK_AS(bracket_backtrack)->u.condfailed = NULL; + + stacksize = GET2(trypath, 1); + if (common->currententry == NULL) + stacksize = 0; + else if (stacksize == RREF_ANY) + stacksize = 1; + else if (common->currententry->start == 0) + stacksize = stacksize == 0; + else + stacksize = stacksize == GET2(common->start, common->currententry->start + 1 + LINK_SIZE); + + if (*trypath == OP_RREF || stacksize || common->currententry == NULL) + { + SLJIT_ASSERT(!has_alternatives); + if (stacksize != 0) + trypath += 1 + IMM2_SIZE; + else + { + if (*cc == OP_ALT) + { + trypath = cc + 1 + LINK_SIZE; + cc += GET(cc, 1); + } + else + trypath = cc; + } + } + else + { + SLJIT_ASSERT(has_alternatives); + + stacksize = GET2(trypath, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE1, STACK_TOP, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0, SLJIT_IMM, common->name_count); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1, SLJIT_IMM, common->name_entry_size); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, SLJIT_IMM, GET2(common->start, common->currententry->start + 1 + LINK_SIZE)); + OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG1, 0, SLJIT_IMM, stacksize); + GET_LOCAL_BASE(SLJIT_TEMPORARY_REG2, 0, 0); + OP1(SLJIT_MOV, SLJIT_TEMPORARY_REG3, 0, SLJIT_IMM, common->name_table); + sljit_emit_ijump(compiler, SLJIT_CALL3, SLJIT_IMM, SLJIT_FUNC_OFFSET(do_searchgroups)); + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE1); + add_jump(compiler, &(BACKTRACK_AS(bracket_backtrack)->u.condfailed), CMP(SLJIT_C_EQUAL, SLJIT_TEMPORARY_REG1, 0, SLJIT_IMM, 0)); + trypath += 1 + IMM2_SIZE; + } + } + else + { + SLJIT_ASSERT(has_alternatives && *trypath >= OP_ASSERT && *trypath <= OP_ASSERTBACK_NOT); + /* Similar code as PUSH_BACKTRACK macro. */ + assert = sljit_alloc_memory(compiler, sizeof(assert_backtrack)); + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + return NULL; + memset(assert, 0, sizeof(assert_backtrack)); + assert->common.cc = trypath; + BACKTRACK_AS(bracket_backtrack)->u.assert = assert; + trypath = compile_assert_trypath(common, trypath, assert, TRUE); + } + } + +compile_trypath(common, trypath, cc, backtrack); +if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + return NULL; + +if (opcode == OP_ONCE) + { + if (BACKTRACK_AS(bracket_backtrack)->u.framesize < 0) + { + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + /* TMP2 which is set here used by OP_KETRMAX below. */ + if (ket == OP_KETRMAX) + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(STACK_TOP), 0); + else if (ket == OP_KETRMIN) + { + /* Move the STR_PTR to the localptr. */ + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_MEM1(STACK_TOP), 0); + } + } + else + { + stacksize = (ket == OP_KETRMIN || ket == OP_KETRMAX || has_alternatives) ? 2 : 1; + OP2(SLJIT_ADD, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_IMM, (BACKTRACK_AS(bracket_backtrack)->u.framesize + stacksize) * sizeof(sljit_w)); + if (ket == OP_KETRMAX) + { + /* TMP2 which is set here used by OP_KETRMAX below. */ + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + } + } + } + +stacksize = 0; +if (ket != OP_KET || bra != OP_BRA) + stacksize++; +if (has_alternatives && opcode != OP_ONCE) + stacksize++; + +if (stacksize > 0) + allocate_stack(common, stacksize); + +stacksize = 0; +if (ket != OP_KET) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stacksize), STR_PTR, 0); + stacksize++; + } +else if (bra != OP_BRA) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stacksize), SLJIT_IMM, 0); + stacksize++; + } + +if (has_alternatives) + { + if (opcode != OP_ONCE) + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stacksize), SLJIT_IMM, 0); + if (ket != OP_KETRMAX) + BACKTRACK_AS(bracket_backtrack)->alttrypath = LABEL(); + } + +/* Must be after the trypath label. */ +if (offset != 0) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1), STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 0), TMP1, 0); + } + +if (ket == OP_KETRMAX) + { + if (opcode == OP_ONCE || opcode >= OP_SBRA) + { + if (has_alternatives) + BACKTRACK_AS(bracket_backtrack)->alttrypath = LABEL(); + /* Checking zero-length iteration. */ + if (opcode != OP_ONCE) + { + CMPTO(SLJIT_C_NOT_EQUAL, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, STR_PTR, 0, rmaxlabel); + /* Drop STR_PTR for greedy plus quantifier. */ + if (bra != OP_BRAZERO) + free_stack(common, 1); + } + else + /* TMP2 must contain the starting STR_PTR. */ + CMPTO(SLJIT_C_NOT_EQUAL, TMP2, 0, STR_PTR, 0, rmaxlabel); + } + else + JUMPTO(SLJIT_JUMP, rmaxlabel); + BACKTRACK_AS(bracket_backtrack)->recursivetrypath = LABEL(); + } + +if (bra == OP_BRAZERO) + BACKTRACK_AS(bracket_backtrack)->zerotrypath = LABEL(); + +if (bra == OP_BRAMINZERO) + { + /* This is a backtrack path! (From the viewpoint of OP_BRAMINZERO) */ + JUMPTO(SLJIT_JUMP, ((braminzero_backtrack *)parent)->trypath); + if (braminzerojump != NULL) + { + JUMPHERE(braminzerojump); + /* We need to release the end pointer to perform the + backtrack for the zero-length iteration. When + framesize is < 0, OP_ONCE will do the release itself. */ + if (opcode == OP_ONCE && BACKTRACK_AS(bracket_backtrack)->u.framesize >= 0) + { + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + add_jump(compiler, &common->revertframes, JUMP(SLJIT_FAST_CALL)); + } + else if (ket == OP_KETRMIN && opcode != OP_ONCE) + free_stack(common, 1); + } + /* Continue to the normal backtrack. */ + } + +if ((ket != OP_KET && bra != OP_BRAMINZERO) || bra == OP_BRAZERO) + decrease_call_count(common); + +/* Skip the other alternatives. */ +while (*cc == OP_ALT) + cc += GET(cc, 1); +cc += 1 + LINK_SIZE; +return cc; +} + +static pcre_uchar *compile_bracketpos_trypath(compiler_common *common, pcre_uchar *cc, backtrack_common *parent) +{ +DEFINE_COMPILER; +backtrack_common *backtrack; +pcre_uchar opcode; +int localptr; +int cbraprivptr = 0; +int framesize; +int stacksize; +int offset = 0; +BOOL zero = FALSE; +pcre_uchar *ccbegin = NULL; +int stack; +struct sljit_label *loop = NULL; +struct jump_list *emptymatch = NULL; + +PUSH_BACKTRACK(sizeof(bracketpos_backtrack), cc, NULL); +if (*cc == OP_BRAPOSZERO) + { + zero = TRUE; + cc++; + } + +opcode = *cc; +localptr = PRIV_DATA(cc); +SLJIT_ASSERT(localptr != 0); +BACKTRACK_AS(bracketpos_backtrack)->localptr = localptr; +switch(opcode) + { + case OP_BRAPOS: + case OP_SBRAPOS: + ccbegin = cc + 1 + LINK_SIZE; + break; + + case OP_CBRAPOS: + case OP_SCBRAPOS: + offset = GET2(cc, 1 + LINK_SIZE); + cbraprivptr = OVECTOR_PRIV(offset); + offset <<= 1; + ccbegin = cc + 1 + LINK_SIZE + IMM2_SIZE; + break; + + default: + SLJIT_ASSERT_STOP(); + break; + } + +framesize = get_framesize(common, cc, FALSE); +BACKTRACK_AS(bracketpos_backtrack)->framesize = framesize; +if (framesize < 0) + { + stacksize = (opcode == OP_CBRAPOS || opcode == OP_SCBRAPOS) ? 2 : 1; + if (!zero) + stacksize++; + BACKTRACK_AS(bracketpos_backtrack)->stacksize = stacksize; + allocate_stack(common, stacksize); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, STACK_TOP, 0); + + if (opcode == OP_CBRAPOS || opcode == OP_SCBRAPOS) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset)); + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), TMP1, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), TMP2, 0); + } + else + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + + if (!zero) + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stacksize - 1), SLJIT_IMM, 1); + } +else + { + stacksize = framesize + 1; + if (!zero) + stacksize++; + if (opcode == OP_BRAPOS || opcode == OP_SBRAPOS) + stacksize++; + BACKTRACK_AS(bracketpos_backtrack)->stacksize = stacksize; + allocate_stack(common, stacksize); + + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + OP2(SLJIT_SUB, TMP2, 0, STACK_TOP, 0, SLJIT_IMM, -STACK(stacksize - 1)); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, TMP2, 0); + stack = 0; + if (!zero) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 1); + stack++; + } + if (opcode == OP_BRAPOS || opcode == OP_SBRAPOS) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stack), STR_PTR, 0); + stack++; + } + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stack), TMP1, 0); + init_frame(common, cc, stacksize - 1, stacksize - framesize, FALSE); + } + +if (opcode == OP_CBRAPOS || opcode == OP_SCBRAPOS) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), cbraprivptr, STR_PTR, 0); + +loop = LABEL(); +while (*cc != OP_KETRPOS) + { + backtrack->top = NULL; + backtrack->topbacktracks = NULL; + cc += GET(cc, 1); + + compile_trypath(common, ccbegin, cc, backtrack); + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + return NULL; + + if (framesize < 0) + { + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + + if (opcode == OP_CBRAPOS || opcode == OP_SCBRAPOS) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), cbraprivptr); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1), STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), cbraprivptr, STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset), TMP1, 0); + } + else + { + if (opcode == OP_SBRAPOS) + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + } + + if (opcode == OP_SBRAPOS || opcode == OP_SCBRAPOS) + add_jump(compiler, &emptymatch, CMP(SLJIT_C_EQUAL, TMP1, 0, STR_PTR, 0)); + + if (!zero) + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stacksize - 1), SLJIT_IMM, 0); + } + else + { + if (opcode == OP_CBRAPOS || opcode == OP_SCBRAPOS) + { + OP2(SLJIT_ADD, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_IMM, stacksize * sizeof(sljit_w)); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), cbraprivptr); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1), STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), cbraprivptr, STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset), TMP1, 0); + } + else + { + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + OP2(SLJIT_ADD, STACK_TOP, 0, TMP2, 0, SLJIT_IMM, stacksize * sizeof(sljit_w)); + if (opcode == OP_SBRAPOS) + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(TMP2), (framesize + 1) * sizeof(sljit_w)); + OP1(SLJIT_MOV, SLJIT_MEM1(TMP2), (framesize + 1) * sizeof(sljit_w), STR_PTR, 0); + } + + if (opcode == OP_SBRAPOS || opcode == OP_SCBRAPOS) + add_jump(compiler, &emptymatch, CMP(SLJIT_C_EQUAL, TMP1, 0, STR_PTR, 0)); + + if (!zero) + { + if (framesize < 0) + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stacksize - 1), SLJIT_IMM, 0); + else + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + } + } + JUMPTO(SLJIT_JUMP, loop); + flush_stubs(common); + + compile_backtrackpath(common, backtrack->top); + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + return NULL; + set_jumps(backtrack->topbacktracks, LABEL()); + + if (framesize < 0) + { + if (opcode == OP_CBRAPOS || opcode == OP_SCBRAPOS) + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), cbraprivptr); + else + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + } + else + { + if (opcode == OP_CBRAPOS || opcode == OP_SCBRAPOS) + { + /* Last alternative. */ + if (*cc == OP_KETRPOS) + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), cbraprivptr); + } + else + { + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(TMP2), (framesize + 1) * sizeof(sljit_w)); + } + } + + if (*cc == OP_KETRPOS) + break; + ccbegin = cc + 1 + LINK_SIZE; + } + +backtrack->topbacktracks = NULL; +if (!zero) + { + if (framesize < 0) + add_jump(compiler, &backtrack->topbacktracks, CMP(SLJIT_C_NOT_EQUAL, SLJIT_MEM1(STACK_TOP), STACK(stacksize - 1), SLJIT_IMM, 0)); + else /* TMP2 is set to [localptr] above. */ + add_jump(compiler, &backtrack->topbacktracks, CMP(SLJIT_C_NOT_EQUAL, SLJIT_MEM1(TMP2), (stacksize - 1) * sizeof(sljit_w), SLJIT_IMM, 0)); + } + +/* None of them matched. */ +set_jumps(emptymatch, LABEL()); +decrease_call_count(common); +return cc + 1 + LINK_SIZE; +} + +static SLJIT_INLINE pcre_uchar *get_iterator_parameters(compiler_common *common, pcre_uchar *cc, pcre_uchar *opcode, pcre_uchar *type, int *arg1, int *arg2, pcre_uchar **end) +{ +int class_len; + +*opcode = *cc; +if (*opcode >= OP_STAR && *opcode <= OP_POSUPTO) + { + cc++; + *type = OP_CHAR; + } +else if (*opcode >= OP_STARI && *opcode <= OP_POSUPTOI) + { + cc++; + *type = OP_CHARI; + *opcode -= OP_STARI - OP_STAR; + } +else if (*opcode >= OP_NOTSTAR && *opcode <= OP_NOTPOSUPTO) + { + cc++; + *type = OP_NOT; + *opcode -= OP_NOTSTAR - OP_STAR; + } +else if (*opcode >= OP_NOTSTARI && *opcode <= OP_NOTPOSUPTOI) + { + cc++; + *type = OP_NOTI; + *opcode -= OP_NOTSTARI - OP_STAR; + } +else if (*opcode >= OP_TYPESTAR && *opcode <= OP_TYPEPOSUPTO) + { + cc++; + *opcode -= OP_TYPESTAR - OP_STAR; + *type = 0; + } +else + { + SLJIT_ASSERT(*opcode >= OP_CLASS || *opcode <= OP_XCLASS); + *type = *opcode; + cc++; + class_len = (*type < OP_XCLASS) ? (int)(1 + (32 / sizeof(pcre_uchar))) : GET(cc, 0); + *opcode = cc[class_len - 1]; + if (*opcode >= OP_CRSTAR && *opcode <= OP_CRMINQUERY) + { + *opcode -= OP_CRSTAR - OP_STAR; + if (end != NULL) + *end = cc + class_len; + } + else + { + SLJIT_ASSERT(*opcode == OP_CRRANGE || *opcode == OP_CRMINRANGE); + *arg1 = GET2(cc, (class_len + IMM2_SIZE)); + *arg2 = GET2(cc, class_len); + + if (*arg2 == 0) + { + SLJIT_ASSERT(*arg1 != 0); + *opcode = (*opcode == OP_CRRANGE) ? OP_UPTO : OP_MINUPTO; + } + if (*arg1 == *arg2) + *opcode = OP_EXACT; + + if (end != NULL) + *end = cc + class_len + 2 * IMM2_SIZE; + } + return cc; + } + +if (*opcode == OP_UPTO || *opcode == OP_MINUPTO || *opcode == OP_EXACT || *opcode == OP_POSUPTO) + { + *arg1 = GET2(cc, 0); + cc += IMM2_SIZE; + } + +if (*type == 0) + { + *type = *cc; + if (end != NULL) + *end = next_opcode(common, cc); + cc++; + return cc; + } + +if (end != NULL) + { + *end = cc + 1; +#ifdef SUPPORT_UTF + if (common->utf && HAS_EXTRALEN(*cc)) *end += GET_EXTRALEN(*cc); +#endif + } +return cc; +} + +static pcre_uchar *compile_iterator_trypath(compiler_common *common, pcre_uchar *cc, backtrack_common *parent) +{ +DEFINE_COMPILER; +backtrack_common *backtrack; +pcre_uchar opcode; +pcre_uchar type; +int arg1 = -1, arg2 = -1; +pcre_uchar* end; +jump_list *nomatch = NULL; +struct sljit_jump *jump = NULL; +struct sljit_label *label; + +PUSH_BACKTRACK(sizeof(iterator_backtrack), cc, NULL); + +cc = get_iterator_parameters(common, cc, &opcode, &type, &arg1, &arg2, &end); + +switch(opcode) + { + case OP_STAR: + case OP_PLUS: + case OP_UPTO: + case OP_CRRANGE: + if (type == OP_ANYNL || type == OP_EXTUNI) + { + if (opcode == OP_STAR || opcode == OP_UPTO) + { + allocate_stack(common, 2); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), SLJIT_IMM, 0); + } + else + { + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + } + if (opcode == OP_UPTO || opcode == OP_CRRANGE) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, SLJIT_IMM, 0); + + label = LABEL(); + compile_char1_trypath(common, type, cc, &backtrack->topbacktracks); + if (opcode == OP_UPTO || opcode == OP_CRRANGE) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0); + OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + if (opcode == OP_CRRANGE && arg2 > 0) + CMPTO(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, arg2, label); + if (opcode == OP_UPTO || (opcode == OP_CRRANGE && arg1 > 0)) + jump = CMP(SLJIT_C_GREATER_EQUAL, TMP1, 0, SLJIT_IMM, arg1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, TMP1, 0); + } + + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + JUMPTO(SLJIT_JUMP, label); + if (jump != NULL) + JUMPHERE(jump); + } + else + { + if (opcode == OP_PLUS) + compile_char1_trypath(common, type, cc, &backtrack->topbacktracks); + allocate_stack(common, 2); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), SLJIT_IMM, 1); + label = LABEL(); + compile_char1_trypath(common, type, cc, &nomatch); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + if (opcode <= OP_PLUS || (opcode == OP_CRRANGE && arg1 == 0)) + { + OP2(SLJIT_ADD, SLJIT_MEM1(STACK_TOP), STACK(1), SLJIT_MEM1(STACK_TOP), STACK(1), SLJIT_IMM, 1); + JUMPTO(SLJIT_JUMP, label); + } + else + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), TMP1, 0); + CMPTO(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, arg1 + 1, label); + } + set_jumps(nomatch, LABEL()); + if (opcode == OP_CRRANGE) + add_jump(compiler, &backtrack->topbacktracks, CMP(SLJIT_C_LESS, SLJIT_MEM1(STACK_TOP), STACK(1), SLJIT_IMM, arg2 + 1)); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + } + BACKTRACK_AS(iterator_backtrack)->trypath = LABEL(); + break; + + case OP_MINSTAR: + case OP_MINPLUS: + if (opcode == OP_MINPLUS) + compile_char1_trypath(common, type, cc, &backtrack->topbacktracks); + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + BACKTRACK_AS(iterator_backtrack)->trypath = LABEL(); + break; + + case OP_MINUPTO: + case OP_CRMINRANGE: + allocate_stack(common, 2); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), SLJIT_IMM, 1); + if (opcode == OP_CRMINRANGE) + add_jump(compiler, &backtrack->topbacktracks, JUMP(SLJIT_JUMP)); + BACKTRACK_AS(iterator_backtrack)->trypath = LABEL(); + break; + + case OP_QUERY: + case OP_MINQUERY: + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + if (opcode == OP_QUERY) + compile_char1_trypath(common, type, cc, &backtrack->topbacktracks); + BACKTRACK_AS(iterator_backtrack)->trypath = LABEL(); + break; + + case OP_EXACT: + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, SLJIT_IMM, 1); + label = LABEL(); + compile_char1_trypath(common, type, cc, &backtrack->topbacktracks); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0); + OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, TMP1, 0); + CMPTO(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, arg1 + 1, label); + break; + + case OP_POSSTAR: + case OP_POSPLUS: + case OP_POSUPTO: + if (opcode != OP_POSSTAR) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, SLJIT_IMM, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE1, STR_PTR, 0); + label = LABEL(); + compile_char1_trypath(common, type, cc, &nomatch); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE1, STR_PTR, 0); + if (opcode != OP_POSUPTO) + { + if (opcode == OP_POSPLUS) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, SLJIT_IMM, 2); + JUMPTO(SLJIT_JUMP, label); + } + else + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0); + OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, TMP1, 0); + CMPTO(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, arg1 + 1, label); + } + set_jumps(nomatch, LABEL()); + if (opcode == OP_POSPLUS) + add_jump(compiler, &backtrack->topbacktracks, CMP(SLJIT_C_LESS, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE0, SLJIT_IMM, 2)); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE1); + break; + + case OP_POSQUERY: + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE1, STR_PTR, 0); + compile_char1_trypath(common, type, cc, &nomatch); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE1, STR_PTR, 0); + set_jumps(nomatch, LABEL()); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), POSSESSIVE1); + break; + + default: + SLJIT_ASSERT_STOP(); + break; + } + +decrease_call_count(common); +return end; +} + +static SLJIT_INLINE pcre_uchar *compile_fail_accept_trypath(compiler_common *common, pcre_uchar *cc, backtrack_common *parent) +{ +DEFINE_COMPILER; +backtrack_common *backtrack; + +PUSH_BACKTRACK(sizeof(bracket_backtrack), cc, NULL); + +if (*cc == OP_FAIL) + { + add_jump(compiler, &backtrack->topbacktracks, JUMP(SLJIT_JUMP)); + return cc + 1; + } + +if (*cc == OP_ASSERT_ACCEPT || common->currententry != NULL) + { + /* No need to check notempty conditions. */ + if (common->acceptlabel == NULL) + add_jump(compiler, &common->accept, JUMP(SLJIT_JUMP)); + else + JUMPTO(SLJIT_JUMP, common->acceptlabel); + return cc + 1; + } + +if (common->acceptlabel == NULL) + add_jump(compiler, &common->accept, CMP(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0))); +else + CMPTO(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0), common->acceptlabel); +OP1(SLJIT_MOV, TMP1, 0, ARGUMENTS, 0); +OP1(SLJIT_MOV_UB, TMP2, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, notempty)); +add_jump(compiler, &backtrack->topbacktracks, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, 0)); +OP1(SLJIT_MOV_UB, TMP2, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, notempty_atstart)); +if (common->acceptlabel == NULL) + add_jump(compiler, &common->accept, CMP(SLJIT_C_EQUAL, TMP2, 0, SLJIT_IMM, 0)); +else + CMPTO(SLJIT_C_EQUAL, TMP2, 0, SLJIT_IMM, 0, common->acceptlabel); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, str)); +if (common->acceptlabel == NULL) + add_jump(compiler, &common->accept, CMP(SLJIT_C_NOT_EQUAL, TMP2, 0, STR_PTR, 0)); +else + CMPTO(SLJIT_C_NOT_EQUAL, TMP2, 0, STR_PTR, 0, common->acceptlabel); +add_jump(compiler, &backtrack->topbacktracks, JUMP(SLJIT_JUMP)); +return cc + 1; +} + +static SLJIT_INLINE pcre_uchar *compile_close_trypath(compiler_common *common, pcre_uchar *cc) +{ +DEFINE_COMPILER; +int offset = GET2(cc, 1); + +/* Data will be discarded anyway... */ +if (common->currententry != NULL) + return cc + 1 + IMM2_SIZE; + +OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR_PRIV(offset)); +offset <<= 1; +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1), STR_PTR, 0); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset), TMP1, 0); +return cc + 1 + IMM2_SIZE; +} + +static void compile_trypath(compiler_common *common, pcre_uchar *cc, pcre_uchar *ccend, backtrack_common *parent) +{ +DEFINE_COMPILER; +backtrack_common *backtrack; + +while (cc < ccend) + { + switch(*cc) + { + case OP_SOD: + case OP_SOM: + case OP_NOT_WORD_BOUNDARY: + case OP_WORD_BOUNDARY: + case OP_NOT_DIGIT: + case OP_DIGIT: + case OP_NOT_WHITESPACE: + case OP_WHITESPACE: + case OP_NOT_WORDCHAR: + case OP_WORDCHAR: + case OP_ANY: + case OP_ALLANY: + case OP_ANYBYTE: + case OP_NOTPROP: + case OP_PROP: + case OP_ANYNL: + case OP_NOT_HSPACE: + case OP_HSPACE: + case OP_NOT_VSPACE: + case OP_VSPACE: + case OP_EXTUNI: + case OP_EODN: + case OP_EOD: + case OP_CIRC: + case OP_CIRCM: + case OP_DOLL: + case OP_DOLLM: + case OP_NOT: + case OP_NOTI: + case OP_REVERSE: + cc = compile_char1_trypath(common, *cc, cc + 1, parent->top != NULL ? &parent->top->nextbacktracks : &parent->topbacktracks); + break; + + case OP_SET_SOM: + PUSH_BACKTRACK_NOVALUE(sizeof(backtrack_common), cc); + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0)); + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0), STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), TMP2, 0); + cc++; + break; + + case OP_CHAR: + case OP_CHARI: + if (common->mode == JIT_COMPILE) + cc = compile_charn_trypath(common, cc, ccend, parent->top != NULL ? &parent->top->nextbacktracks : &parent->topbacktracks); + else + cc = compile_char1_trypath(common, *cc, cc + 1, parent->top != NULL ? &parent->top->nextbacktracks : &parent->topbacktracks); + break; + + case OP_STAR: + case OP_MINSTAR: + case OP_PLUS: + case OP_MINPLUS: + case OP_QUERY: + case OP_MINQUERY: + case OP_UPTO: + case OP_MINUPTO: + case OP_EXACT: + case OP_POSSTAR: + case OP_POSPLUS: + case OP_POSQUERY: + case OP_POSUPTO: + case OP_STARI: + case OP_MINSTARI: + case OP_PLUSI: + case OP_MINPLUSI: + case OP_QUERYI: + case OP_MINQUERYI: + case OP_UPTOI: + case OP_MINUPTOI: + case OP_EXACTI: + case OP_POSSTARI: + case OP_POSPLUSI: + case OP_POSQUERYI: + case OP_POSUPTOI: + case OP_NOTSTAR: + case OP_NOTMINSTAR: + case OP_NOTPLUS: + case OP_NOTMINPLUS: + case OP_NOTQUERY: + case OP_NOTMINQUERY: + case OP_NOTUPTO: + case OP_NOTMINUPTO: + case OP_NOTEXACT: + case OP_NOTPOSSTAR: + case OP_NOTPOSPLUS: + case OP_NOTPOSQUERY: + case OP_NOTPOSUPTO: + case OP_NOTSTARI: + case OP_NOTMINSTARI: + case OP_NOTPLUSI: + case OP_NOTMINPLUSI: + case OP_NOTQUERYI: + case OP_NOTMINQUERYI: + case OP_NOTUPTOI: + case OP_NOTMINUPTOI: + case OP_NOTEXACTI: + case OP_NOTPOSSTARI: + case OP_NOTPOSPLUSI: + case OP_NOTPOSQUERYI: + case OP_NOTPOSUPTOI: + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEEXACT: + case OP_TYPEPOSSTAR: + case OP_TYPEPOSPLUS: + case OP_TYPEPOSQUERY: + case OP_TYPEPOSUPTO: + cc = compile_iterator_trypath(common, cc, parent); + break; + + case OP_CLASS: + case OP_NCLASS: + if (cc[1 + (32 / sizeof(pcre_uchar))] >= OP_CRSTAR && cc[1 + (32 / sizeof(pcre_uchar))] <= OP_CRMINRANGE) + cc = compile_iterator_trypath(common, cc, parent); + else + cc = compile_char1_trypath(common, *cc, cc + 1, parent->top != NULL ? &parent->top->nextbacktracks : &parent->topbacktracks); + break; + +#if defined SUPPORT_UTF || defined COMPILE_PCRE16 + case OP_XCLASS: + if (*(cc + GET(cc, 1)) >= OP_CRSTAR && *(cc + GET(cc, 1)) <= OP_CRMINRANGE) + cc = compile_iterator_trypath(common, cc, parent); + else + cc = compile_char1_trypath(common, *cc, cc + 1, parent->top != NULL ? &parent->top->nextbacktracks : &parent->topbacktracks); + break; +#endif + + case OP_REF: + case OP_REFI: + if (cc[1 + IMM2_SIZE] >= OP_CRSTAR && cc[1 + IMM2_SIZE] <= OP_CRMINRANGE) + cc = compile_ref_iterator_trypath(common, cc, parent); + else + cc = compile_ref_trypath(common, cc, parent->top != NULL ? &parent->top->nextbacktracks : &parent->topbacktracks, TRUE, FALSE); + break; + + case OP_RECURSE: + cc = compile_recurse_trypath(common, cc, parent); + break; + + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + PUSH_BACKTRACK_NOVALUE(sizeof(assert_backtrack), cc); + cc = compile_assert_trypath(common, cc, BACKTRACK_AS(assert_backtrack), FALSE); + break; + + case OP_BRAMINZERO: + PUSH_BACKTRACK_NOVALUE(sizeof(braminzero_backtrack), cc); + cc = bracketend(cc + 1); + if (*(cc - 1 - LINK_SIZE) != OP_KETRMIN) + { + allocate_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + } + else + { + allocate_stack(common, 2); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), STR_PTR, 0); + } + BACKTRACK_AS(braminzero_backtrack)->trypath = LABEL(); + if (cc[1] > OP_ASSERTBACK_NOT) + decrease_call_count(common); + break; + + case OP_ONCE: + case OP_ONCE_NC: + case OP_BRA: + case OP_CBRA: + case OP_COND: + case OP_SBRA: + case OP_SCBRA: + case OP_SCOND: + cc = compile_bracket_trypath(common, cc, parent); + break; + + case OP_BRAZERO: + if (cc[1] > OP_ASSERTBACK_NOT) + cc = compile_bracket_trypath(common, cc, parent); + else + { + PUSH_BACKTRACK_NOVALUE(sizeof(assert_backtrack), cc); + cc = compile_assert_trypath(common, cc, BACKTRACK_AS(assert_backtrack), FALSE); + } + break; + + case OP_BRAPOS: + case OP_CBRAPOS: + case OP_SBRAPOS: + case OP_SCBRAPOS: + case OP_BRAPOSZERO: + cc = compile_bracketpos_trypath(common, cc, parent); + break; + + case OP_MARK: + PUSH_BACKTRACK_NOVALUE(sizeof(backtrack_common), cc); + SLJIT_ASSERT(common->mark_ptr != 0); + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->mark_ptr); + allocate_stack(common, 1); + OP1(SLJIT_MOV, TMP1, 0, ARGUMENTS, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), TMP2, 0); + OP1(SLJIT_MOV, TMP2, 0, SLJIT_IMM, (sljit_w)(cc + 2)); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->mark_ptr, TMP2, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, mark_ptr), TMP2, 0); + cc += 1 + 2 + cc[1]; + break; + + case OP_COMMIT: + PUSH_BACKTRACK_NOVALUE(sizeof(backtrack_common), cc); + cc += 1; + break; + + case OP_FAIL: + case OP_ACCEPT: + case OP_ASSERT_ACCEPT: + cc = compile_fail_accept_trypath(common, cc, parent); + break; + + case OP_CLOSE: + cc = compile_close_trypath(common, cc); + break; + + case OP_SKIPZERO: + cc = bracketend(cc + 1); + break; + + default: + SLJIT_ASSERT_STOP(); + return; + } + if (cc == NULL) + return; + } +SLJIT_ASSERT(cc == ccend); +} + +#undef PUSH_BACKTRACK +#undef PUSH_BACKTRACK_NOVALUE +#undef BACKTRACK_AS + +#define COMPILE_BACKTRACKPATH(current) \ + do \ + { \ + compile_backtrackpath(common, (current)); \ + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) \ + return; \ + } \ + while (0) + +#define CURRENT_AS(type) ((type *)current) + +static void compile_iterator_backtrackpath(compiler_common *common, struct backtrack_common *current) +{ +DEFINE_COMPILER; +pcre_uchar *cc = current->cc; +pcre_uchar opcode; +pcre_uchar type; +int arg1 = -1, arg2 = -1; +struct sljit_label *label = NULL; +struct sljit_jump *jump = NULL; +jump_list *jumplist = NULL; + +cc = get_iterator_parameters(common, cc, &opcode, &type, &arg1, &arg2, NULL); + +switch(opcode) + { + case OP_STAR: + case OP_PLUS: + case OP_UPTO: + case OP_CRRANGE: + if (type == OP_ANYNL || type == OP_EXTUNI) + { + set_jumps(current->topbacktracks, LABEL()); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + free_stack(common, 1); + CMPTO(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_IMM, 0, CURRENT_AS(iterator_backtrack)->trypath); + } + else + { + if (opcode <= OP_PLUS || opcode == OP_UPTO) + arg2 = 0; + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + jump = CMP(SLJIT_C_LESS_EQUAL, TMP1, 0, SLJIT_IMM, arg2 + 1); + OP2(SLJIT_SUB, SLJIT_MEM1(STACK_TOP), STACK(1), TMP1, 0, SLJIT_IMM, 1); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + skip_char_back(common); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + JUMPTO(SLJIT_JUMP, CURRENT_AS(iterator_backtrack)->trypath); + if (opcode == OP_CRRANGE) + set_jumps(current->topbacktracks, LABEL()); + JUMPHERE(jump); + free_stack(common, 2); + if (opcode == OP_PLUS) + set_jumps(current->topbacktracks, LABEL()); + } + break; + + case OP_MINSTAR: + case OP_MINPLUS: + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + compile_char1_trypath(common, type, cc, &jumplist); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + JUMPTO(SLJIT_JUMP, CURRENT_AS(iterator_backtrack)->trypath); + set_jumps(jumplist, LABEL()); + free_stack(common, 1); + if (opcode == OP_MINPLUS) + set_jumps(current->topbacktracks, LABEL()); + break; + + case OP_MINUPTO: + case OP_CRMINRANGE: + if (opcode == OP_CRMINRANGE) + { + label = LABEL(); + set_jumps(current->topbacktracks, label); + } + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + compile_char1_trypath(common, type, cc, &jumplist); + + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), TMP1, 0); + + if (opcode == OP_CRMINRANGE) + CMPTO(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, arg2 + 1, label); + + if (opcode == OP_CRMINRANGE && arg1 == 0) + JUMPTO(SLJIT_JUMP, CURRENT_AS(iterator_backtrack)->trypath); + else + CMPTO(SLJIT_C_LESS, TMP1, 0, SLJIT_IMM, arg1 + 2, CURRENT_AS(iterator_backtrack)->trypath); + + set_jumps(jumplist, LABEL()); + free_stack(common, 2); + break; + + case OP_QUERY: + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + CMPTO(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_IMM, 0, CURRENT_AS(iterator_backtrack)->trypath); + jump = JUMP(SLJIT_JUMP); + set_jumps(current->topbacktracks, LABEL()); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + JUMPTO(SLJIT_JUMP, CURRENT_AS(iterator_backtrack)->trypath); + JUMPHERE(jump); + free_stack(common, 1); + break; + + case OP_MINQUERY: + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + jump = CMP(SLJIT_C_EQUAL, STR_PTR, 0, SLJIT_IMM, 0); + compile_char1_trypath(common, type, cc, &jumplist); + JUMPTO(SLJIT_JUMP, CURRENT_AS(iterator_backtrack)->trypath); + set_jumps(jumplist, LABEL()); + JUMPHERE(jump); + free_stack(common, 1); + break; + + case OP_EXACT: + case OP_POSPLUS: + set_jumps(current->topbacktracks, LABEL()); + break; + + case OP_POSSTAR: + case OP_POSQUERY: + case OP_POSUPTO: + break; + + default: + SLJIT_ASSERT_STOP(); + break; + } +} + +static void compile_ref_iterator_backtrackpath(compiler_common *common, struct backtrack_common *current) +{ +DEFINE_COMPILER; +pcre_uchar *cc = current->cc; +pcre_uchar type; + +type = cc[1 + IMM2_SIZE]; +if ((type & 0x1) == 0) + { + set_jumps(current->topbacktracks, LABEL()); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + free_stack(common, 1); + CMPTO(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_IMM, 0, CURRENT_AS(iterator_backtrack)->trypath); + return; + } + +OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); +CMPTO(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_IMM, 0, CURRENT_AS(iterator_backtrack)->trypath); +set_jumps(current->topbacktracks, LABEL()); +free_stack(common, 2); +} + +static void compile_recurse_backtrackpath(compiler_common *common, struct backtrack_common *current) +{ +DEFINE_COMPILER; + +set_jumps(current->topbacktracks, LABEL()); + +if (common->has_set_som && common->mark_ptr != 0) + { + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + free_stack(common, 2); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0), TMP2, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->mark_ptr, TMP1, 0); + } +else if (common->has_set_som || common->mark_ptr != 0) + { + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + free_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->has_set_som ? (int)(OVECTOR(0)) : common->mark_ptr, TMP2, 0); + } +} + +static void compile_assert_backtrackpath(compiler_common *common, struct backtrack_common *current) +{ +DEFINE_COMPILER; +pcre_uchar *cc = current->cc; +pcre_uchar bra = OP_BRA; +struct sljit_jump *brajump = NULL; + +SLJIT_ASSERT(*cc != OP_BRAMINZERO); +if (*cc == OP_BRAZERO) + { + bra = *cc; + cc++; + } + +if (bra == OP_BRAZERO) + { + SLJIT_ASSERT(current->topbacktracks == NULL); + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + } + +if (CURRENT_AS(assert_backtrack)->framesize < 0) + { + set_jumps(current->topbacktracks, LABEL()); + + if (bra == OP_BRAZERO) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + CMPTO(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_IMM, 0, CURRENT_AS(assert_backtrack)->trypath); + free_stack(common, 1); + } + return; + } + +if (bra == OP_BRAZERO) + { + if (*cc == OP_ASSERT_NOT || *cc == OP_ASSERTBACK_NOT) + { + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + CMPTO(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_IMM, 0, CURRENT_AS(assert_backtrack)->trypath); + free_stack(common, 1); + return; + } + free_stack(common, 1); + brajump = CMP(SLJIT_C_EQUAL, STR_PTR, 0, SLJIT_IMM, 0); + } + +if (*cc == OP_ASSERT || *cc == OP_ASSERTBACK) + { + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), CURRENT_AS(assert_backtrack)->localptr); + add_jump(compiler, &common->revertframes, JUMP(SLJIT_FAST_CALL)); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), CURRENT_AS(assert_backtrack)->localptr, SLJIT_MEM1(STACK_TOP), CURRENT_AS(assert_backtrack)->framesize * sizeof(sljit_w)); + + set_jumps(current->topbacktracks, LABEL()); + } +else + set_jumps(current->topbacktracks, LABEL()); + +if (bra == OP_BRAZERO) + { + /* We know there is enough place on the stack. */ + OP2(SLJIT_ADD, STACK_TOP, 0, STACK_TOP, 0, SLJIT_IMM, sizeof(sljit_w)); + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), SLJIT_IMM, 0); + JUMPTO(SLJIT_JUMP, CURRENT_AS(assert_backtrack)->trypath); + JUMPHERE(brajump); + } +} + +static void compile_bracket_backtrackpath(compiler_common *common, struct backtrack_common *current) +{ +DEFINE_COMPILER; +int opcode; +int offset = 0; +int localptr = CURRENT_AS(bracket_backtrack)->localptr; +int stacksize; +int count; +pcre_uchar *cc = current->cc; +pcre_uchar *ccbegin; +pcre_uchar *ccprev; +jump_list *jumplist = NULL; +jump_list *jumplistitem = NULL; +pcre_uchar bra = OP_BRA; +pcre_uchar ket; +assert_backtrack *assert; +BOOL has_alternatives; +struct sljit_jump *brazero = NULL; +struct sljit_jump *once = NULL; +struct sljit_jump *cond = NULL; +struct sljit_label *rminlabel = NULL; + +if (*cc == OP_BRAZERO || *cc == OP_BRAMINZERO) + { + bra = *cc; + cc++; + } + +opcode = *cc; +ccbegin = cc; +ket = *(bracketend(ccbegin) - 1 - LINK_SIZE); +cc += GET(cc, 1); +has_alternatives = *cc == OP_ALT; +if (SLJIT_UNLIKELY(opcode == OP_COND) || SLJIT_UNLIKELY(opcode == OP_SCOND)) + has_alternatives = (ccbegin[1 + LINK_SIZE] >= OP_ASSERT && ccbegin[1 + LINK_SIZE] <= OP_ASSERTBACK_NOT) || CURRENT_AS(bracket_backtrack)->u.condfailed != NULL; +if (opcode == OP_CBRA || opcode == OP_SCBRA) + offset = (GET2(ccbegin, 1 + LINK_SIZE)) << 1; +if (SLJIT_UNLIKELY(opcode == OP_COND) && (*cc == OP_KETRMAX || *cc == OP_KETRMIN)) + opcode = OP_SCOND; +if (SLJIT_UNLIKELY(opcode == OP_ONCE_NC)) + opcode = OP_ONCE; + +if (ket == OP_KETRMAX) + { + if (bra == OP_BRAZERO) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + free_stack(common, 1); + brazero = CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, 0); + } + } +else if (ket == OP_KETRMIN) + { + if (bra != OP_BRAMINZERO) + { + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + if (opcode >= OP_SBRA || opcode == OP_ONCE) + { + /* Checking zero-length iteration. */ + if (opcode != OP_ONCE || CURRENT_AS(bracket_backtrack)->u.framesize < 0) + CMPTO(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, CURRENT_AS(bracket_backtrack)->recursivetrypath); + else + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + CMPTO(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_MEM1(TMP1), (CURRENT_AS(bracket_backtrack)->u.framesize + 1) * sizeof(sljit_w), CURRENT_AS(bracket_backtrack)->recursivetrypath); + } + if (opcode != OP_ONCE) + free_stack(common, 1); + } + else + JUMPTO(SLJIT_JUMP, CURRENT_AS(bracket_backtrack)->recursivetrypath); + } + rminlabel = LABEL(); + } +else if (bra == OP_BRAZERO) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + free_stack(common, 1); + brazero = CMP(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, 0); + } + +if (SLJIT_UNLIKELY(opcode == OP_ONCE)) + { + if (CURRENT_AS(bracket_backtrack)->u.framesize >= 0) + { + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + add_jump(compiler, &common->revertframes, JUMP(SLJIT_FAST_CALL)); + } + once = JUMP(SLJIT_JUMP); + } +else if (SLJIT_UNLIKELY(opcode == OP_COND) || SLJIT_UNLIKELY(opcode == OP_SCOND)) + { + if (has_alternatives) + { + /* Always exactly one alternative. */ + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + free_stack(common, 1); + + jumplistitem = sljit_alloc_memory(compiler, sizeof(jump_list)); + if (SLJIT_UNLIKELY(!jumplistitem)) + return; + jumplist = jumplistitem; + jumplistitem->next = NULL; + jumplistitem->jump = CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, 1); + } + } +else if (*cc == OP_ALT) + { + /* Build a jump list. Get the last successfully matched branch index. */ + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + free_stack(common, 1); + count = 1; + do + { + /* Append as the last item. */ + if (jumplist != NULL) + { + jumplistitem->next = sljit_alloc_memory(compiler, sizeof(jump_list)); + jumplistitem = jumplistitem->next; + } + else + { + jumplistitem = sljit_alloc_memory(compiler, sizeof(jump_list)); + jumplist = jumplistitem; + } + + if (SLJIT_UNLIKELY(!jumplistitem)) + return; + + jumplistitem->next = NULL; + jumplistitem->jump = CMP(SLJIT_C_EQUAL, TMP1, 0, SLJIT_IMM, count++); + cc += GET(cc, 1); + } + while (*cc == OP_ALT); + + cc = ccbegin + GET(ccbegin, 1); + } + +COMPILE_BACKTRACKPATH(current->top); +if (current->topbacktracks) + set_jumps(current->topbacktracks, LABEL()); + +if (SLJIT_UNLIKELY(opcode == OP_COND) || SLJIT_UNLIKELY(opcode == OP_SCOND)) + { + /* Conditional block always has at most one alternative. */ + if (ccbegin[1 + LINK_SIZE] >= OP_ASSERT && ccbegin[1 + LINK_SIZE] <= OP_ASSERTBACK_NOT) + { + SLJIT_ASSERT(has_alternatives); + assert = CURRENT_AS(bracket_backtrack)->u.assert; + if (assert->framesize >= 0 && (ccbegin[1 + LINK_SIZE] == OP_ASSERT || ccbegin[1 + LINK_SIZE] == OP_ASSERTBACK)) + { + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), assert->localptr); + add_jump(compiler, &common->revertframes, JUMP(SLJIT_FAST_CALL)); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), assert->localptr, SLJIT_MEM1(STACK_TOP), assert->framesize * sizeof(sljit_w)); + } + cond = JUMP(SLJIT_JUMP); + set_jumps(CURRENT_AS(bracket_backtrack)->u.assert->condfailed, LABEL()); + } + else if (CURRENT_AS(bracket_backtrack)->u.condfailed != NULL) + { + SLJIT_ASSERT(has_alternatives); + cond = JUMP(SLJIT_JUMP); + set_jumps(CURRENT_AS(bracket_backtrack)->u.condfailed, LABEL()); + } + else + SLJIT_ASSERT(!has_alternatives); + } + +if (has_alternatives) + { + count = 1; + do + { + current->top = NULL; + current->topbacktracks = NULL; + current->nextbacktracks = NULL; + if (*cc == OP_ALT) + { + ccprev = cc + 1 + LINK_SIZE; + cc += GET(cc, 1); + if (opcode != OP_COND && opcode != OP_SCOND) + { + if (localptr != 0 && opcode != OP_ONCE) + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + else + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + } + compile_trypath(common, ccprev, cc, current); + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + return; + } + + /* Instructions after the current alternative is succesfully matched. */ + /* There is a similar code in compile_bracket_trypath. */ + if (opcode == OP_ONCE) + { + if (CURRENT_AS(bracket_backtrack)->u.framesize < 0) + { + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + /* TMP2 which is set here used by OP_KETRMAX below. */ + if (ket == OP_KETRMAX) + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(STACK_TOP), 0); + else if (ket == OP_KETRMIN) + { + /* Move the STR_PTR to the localptr. */ + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_MEM1(STACK_TOP), 0); + } + } + else + { + OP2(SLJIT_ADD, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_IMM, (CURRENT_AS(bracket_backtrack)->u.framesize + 2) * sizeof(sljit_w)); + if (ket == OP_KETRMAX) + { + /* TMP2 which is set here used by OP_KETRMAX below. */ + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + } + } + } + + stacksize = 0; + if (opcode != OP_ONCE) + stacksize++; + if (ket != OP_KET || bra != OP_BRA) + stacksize++; + + if (stacksize > 0) { + if (opcode != OP_ONCE || CURRENT_AS(bracket_backtrack)->u.framesize >= 0) + allocate_stack(common, stacksize); + else + { + /* We know we have place at least for one item on the top of the stack. */ + SLJIT_ASSERT(stacksize == 1); + OP2(SLJIT_ADD, STACK_TOP, 0, STACK_TOP, 0, SLJIT_IMM, sizeof(sljit_w)); + } + } + + stacksize = 0; + if (ket != OP_KET || bra != OP_BRA) + { + if (ket != OP_KET) + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stacksize), STR_PTR, 0); + else + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stacksize), SLJIT_IMM, 0); + stacksize++; + } + + if (opcode != OP_ONCE) + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(stacksize), SLJIT_IMM, count++); + + if (offset != 0) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1), STR_PTR, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 0), TMP1, 0); + } + + JUMPTO(SLJIT_JUMP, CURRENT_AS(bracket_backtrack)->alttrypath); + + if (opcode != OP_ONCE) + { + SLJIT_ASSERT(jumplist); + JUMPHERE(jumplist->jump); + jumplist = jumplist->next; + } + + COMPILE_BACKTRACKPATH(current->top); + if (current->topbacktracks) + set_jumps(current->topbacktracks, LABEL()); + SLJIT_ASSERT(!current->nextbacktracks); + } + while (*cc == OP_ALT); + SLJIT_ASSERT(!jumplist); + + if (cond != NULL) + { + SLJIT_ASSERT(opcode == OP_COND || opcode == OP_SCOND); + assert = CURRENT_AS(bracket_backtrack)->u.assert; + if ((ccbegin[1 + LINK_SIZE] == OP_ASSERT_NOT || ccbegin[1 + LINK_SIZE] == OP_ASSERTBACK_NOT) && assert->framesize >= 0) + + { + OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), assert->localptr); + add_jump(compiler, &common->revertframes, JUMP(SLJIT_FAST_CALL)); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), assert->localptr, SLJIT_MEM1(STACK_TOP), assert->framesize * sizeof(sljit_w)); + } + JUMPHERE(cond); + } + + /* Free the STR_PTR. */ + if (localptr == 0) + free_stack(common, 1); + } + +if (offset != 0) + { + /* Using both tmp register is better for instruction scheduling. */ + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset), TMP1, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1), TMP2, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_MEM1(STACK_TOP), STACK(2)); + free_stack(common, 3); + } +else if (opcode == OP_SBRA || opcode == OP_SCOND) + { + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_MEM1(STACK_TOP), STACK(0)); + free_stack(common, 1); + } +else if (opcode == OP_ONCE) + { + cc = ccbegin + GET(ccbegin, 1); + if (CURRENT_AS(bracket_backtrack)->u.framesize >= 0) + { + /* Reset head and drop saved frame. */ + stacksize = (ket == OP_KETRMAX || ket == OP_KETRMIN || *cc == OP_ALT) ? 2 : 1; + free_stack(common, CURRENT_AS(bracket_backtrack)->u.framesize + stacksize); + } + else if (ket == OP_KETRMAX || (*cc == OP_ALT && ket != OP_KETRMIN)) + { + /* The STR_PTR must be released. */ + free_stack(common, 1); + } + + JUMPHERE(once); + /* Restore previous localptr */ + if (CURRENT_AS(bracket_backtrack)->u.framesize >= 0) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, SLJIT_MEM1(STACK_TOP), CURRENT_AS(bracket_backtrack)->u.framesize * sizeof(sljit_w)); + else if (ket == OP_KETRMIN) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + /* See the comment below. */ + free_stack(common, 2); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), localptr, TMP1, 0); + } + } + +if (ket == OP_KETRMAX) + { + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + if (bra != OP_BRAZERO) + free_stack(common, 1); + CMPTO(SLJIT_C_NOT_EQUAL, STR_PTR, 0, SLJIT_IMM, 0, CURRENT_AS(bracket_backtrack)->recursivetrypath); + if (bra == OP_BRAZERO) + { + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + JUMPTO(SLJIT_JUMP, CURRENT_AS(bracket_backtrack)->zerotrypath); + JUMPHERE(brazero); + free_stack(common, 1); + } + } +else if (ket == OP_KETRMIN) + { + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + + /* OP_ONCE removes everything in case of a backtrack, so we don't + need to explicitly release the STR_PTR. The extra release would + affect badly the free_stack(2) above. */ + if (opcode != OP_ONCE) + free_stack(common, 1); + CMPTO(SLJIT_C_NOT_EQUAL, TMP1, 0, SLJIT_IMM, 0, rminlabel); + if (opcode == OP_ONCE) + free_stack(common, bra == OP_BRAMINZERO ? 2 : 1); + else if (bra == OP_BRAMINZERO) + free_stack(common, 1); + } +else if (bra == OP_BRAZERO) + { + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + JUMPTO(SLJIT_JUMP, CURRENT_AS(bracket_backtrack)->zerotrypath); + JUMPHERE(brazero); + } +} + +static void compile_bracketpos_backtrackpath(compiler_common *common, struct backtrack_common *current) +{ +DEFINE_COMPILER; +int offset; +struct sljit_jump *jump; + +if (CURRENT_AS(bracketpos_backtrack)->framesize < 0) + { + if (*current->cc == OP_CBRAPOS || *current->cc == OP_SCBRAPOS) + { + offset = (GET2(current->cc, 1 + LINK_SIZE)) << 1; + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(STACK_TOP), STACK(1)); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset), TMP1, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(offset + 1), TMP2, 0); + } + set_jumps(current->topbacktracks, LABEL()); + free_stack(common, CURRENT_AS(bracketpos_backtrack)->stacksize); + return; + } + +OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), CURRENT_AS(bracketpos_backtrack)->localptr); +add_jump(compiler, &common->revertframes, JUMP(SLJIT_FAST_CALL)); + +if (current->topbacktracks) + { + jump = JUMP(SLJIT_JUMP); + set_jumps(current->topbacktracks, LABEL()); + /* Drop the stack frame. */ + free_stack(common, CURRENT_AS(bracketpos_backtrack)->stacksize); + JUMPHERE(jump); + } +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), CURRENT_AS(bracketpos_backtrack)->localptr, SLJIT_MEM1(STACK_TOP), CURRENT_AS(bracketpos_backtrack)->framesize * sizeof(sljit_w)); +} + +static void compile_braminzero_backtrackpath(compiler_common *common, struct backtrack_common *current) +{ +assert_backtrack backtrack; + +current->top = NULL; +current->topbacktracks = NULL; +current->nextbacktracks = NULL; +if (current->cc[1] > OP_ASSERTBACK_NOT) + { + /* Manual call of compile_bracket_trypath and compile_bracket_backtrackpath. */ + compile_bracket_trypath(common, current->cc, current); + compile_bracket_backtrackpath(common, current->top); + } +else + { + memset(&backtrack, 0, sizeof(backtrack)); + backtrack.common.cc = current->cc; + backtrack.trypath = CURRENT_AS(braminzero_backtrack)->trypath; + /* Manual call of compile_assert_trypath. */ + compile_assert_trypath(common, current->cc, &backtrack, FALSE); + } +SLJIT_ASSERT(!current->nextbacktracks && !current->topbacktracks); +} + +static void compile_backtrackpath(compiler_common *common, struct backtrack_common *current) +{ +DEFINE_COMPILER; + +while (current) + { + if (current->nextbacktracks != NULL) + set_jumps(current->nextbacktracks, LABEL()); + switch(*current->cc) + { + case OP_SET_SOM: + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + free_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0), TMP1, 0); + break; + + case OP_STAR: + case OP_MINSTAR: + case OP_PLUS: + case OP_MINPLUS: + case OP_QUERY: + case OP_MINQUERY: + case OP_UPTO: + case OP_MINUPTO: + case OP_EXACT: + case OP_POSSTAR: + case OP_POSPLUS: + case OP_POSQUERY: + case OP_POSUPTO: + case OP_STARI: + case OP_MINSTARI: + case OP_PLUSI: + case OP_MINPLUSI: + case OP_QUERYI: + case OP_MINQUERYI: + case OP_UPTOI: + case OP_MINUPTOI: + case OP_EXACTI: + case OP_POSSTARI: + case OP_POSPLUSI: + case OP_POSQUERYI: + case OP_POSUPTOI: + case OP_NOTSTAR: + case OP_NOTMINSTAR: + case OP_NOTPLUS: + case OP_NOTMINPLUS: + case OP_NOTQUERY: + case OP_NOTMINQUERY: + case OP_NOTUPTO: + case OP_NOTMINUPTO: + case OP_NOTEXACT: + case OP_NOTPOSSTAR: + case OP_NOTPOSPLUS: + case OP_NOTPOSQUERY: + case OP_NOTPOSUPTO: + case OP_NOTSTARI: + case OP_NOTMINSTARI: + case OP_NOTPLUSI: + case OP_NOTMINPLUSI: + case OP_NOTQUERYI: + case OP_NOTMINQUERYI: + case OP_NOTUPTOI: + case OP_NOTMINUPTOI: + case OP_NOTEXACTI: + case OP_NOTPOSSTARI: + case OP_NOTPOSPLUSI: + case OP_NOTPOSQUERYI: + case OP_NOTPOSUPTOI: + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEEXACT: + case OP_TYPEPOSSTAR: + case OP_TYPEPOSPLUS: + case OP_TYPEPOSQUERY: + case OP_TYPEPOSUPTO: + case OP_CLASS: + case OP_NCLASS: +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: +#endif + compile_iterator_backtrackpath(common, current); + break; + + case OP_REF: + case OP_REFI: + compile_ref_iterator_backtrackpath(common, current); + break; + + case OP_RECURSE: + compile_recurse_backtrackpath(common, current); + break; + + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + compile_assert_backtrackpath(common, current); + break; + + case OP_ONCE: + case OP_ONCE_NC: + case OP_BRA: + case OP_CBRA: + case OP_COND: + case OP_SBRA: + case OP_SCBRA: + case OP_SCOND: + compile_bracket_backtrackpath(common, current); + break; + + case OP_BRAZERO: + if (current->cc[1] > OP_ASSERTBACK_NOT) + compile_bracket_backtrackpath(common, current); + else + compile_assert_backtrackpath(common, current); + break; + + case OP_BRAPOS: + case OP_CBRAPOS: + case OP_SBRAPOS: + case OP_SCBRAPOS: + case OP_BRAPOSZERO: + compile_bracketpos_backtrackpath(common, current); + break; + + case OP_BRAMINZERO: + compile_braminzero_backtrackpath(common, current); + break; + + case OP_MARK: + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + free_stack(common, 1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->mark_ptr, TMP1, 0); + break; + + case OP_COMMIT: + OP1(SLJIT_MOV, SLJIT_RETURN_REG, 0, SLJIT_IMM, PCRE_ERROR_NOMATCH); + if (common->leavelabel == NULL) + add_jump(compiler, &common->leave, JUMP(SLJIT_JUMP)); + else + JUMPTO(SLJIT_JUMP, common->leavelabel); + break; + + case OP_FAIL: + case OP_ACCEPT: + case OP_ASSERT_ACCEPT: + set_jumps(current->topbacktracks, LABEL()); + break; + + default: + SLJIT_ASSERT_STOP(); + break; + } + current = current->prev; + } +} + +static SLJIT_INLINE void compile_recurse(compiler_common *common) +{ +DEFINE_COMPILER; +pcre_uchar *cc = common->start + common->currententry->start; +pcre_uchar *ccbegin = cc + 1 + LINK_SIZE + (*cc == OP_BRA ? 0 : IMM2_SIZE); +pcre_uchar *ccend = bracketend(cc); +int localsize = get_localsize(common, ccbegin, ccend); +int framesize = get_framesize(common, cc, TRUE); +int alternativesize; +BOOL needsframe; +backtrack_common altbacktrack; +struct sljit_label *save_leavelabel = common->leavelabel; +jump_list *save_leave = common->leave; +struct sljit_jump *jump; + +SLJIT_ASSERT(*cc == OP_BRA || *cc == OP_CBRA || *cc == OP_CBRAPOS || *cc == OP_SCBRA || *cc == OP_SCBRAPOS); +needsframe = framesize >= 0; +if (!needsframe) + framesize = 0; +alternativesize = *(cc + GET(cc, 1)) == OP_ALT ? 1 : 0; + +SLJIT_ASSERT(common->currententry->entry == NULL && common->recursive_head != 0); +common->currententry->entry = LABEL(); +set_jumps(common->currententry->calls, common->currententry->entry); + +sljit_emit_fast_enter(compiler, TMP2, 0); +allocate_stack(common, localsize + framesize + alternativesize); +OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(localsize + framesize + alternativesize - 1), TMP2, 0); +copy_locals(common, ccbegin, ccend, TRUE, localsize + framesize + alternativesize, framesize + alternativesize); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->recursive_head, STACK_TOP, 0); +if (needsframe) + init_frame(common, cc, framesize + alternativesize - 1, alternativesize, TRUE); + +if (alternativesize > 0) + OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); + +memset(&altbacktrack, 0, sizeof(backtrack_common)); +common->leavelabel = NULL; +common->acceptlabel = NULL; +common->leave = NULL; +common->accept = NULL; +altbacktrack.cc = ccbegin; +cc += GET(cc, 1); +while (1) + { + altbacktrack.top = NULL; + altbacktrack.topbacktracks = NULL; + + if (altbacktrack.cc != ccbegin) + OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(STACK_TOP), STACK(0)); + + compile_trypath(common, altbacktrack.cc, cc, &altbacktrack); + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + { + common->leavelabel = save_leavelabel; + common->leave = save_leave; + return; + } + + add_jump(compiler, &common->accept, JUMP(SLJIT_JUMP)); + + compile_backtrackpath(common, altbacktrack.top); + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + { + common->leavelabel = save_leavelabel; + common->leave = save_leave; + return; + } + set_jumps(altbacktrack.topbacktracks, LABEL()); + + if (*cc != OP_ALT) + break; + + altbacktrack.cc = cc + 1 + LINK_SIZE; + cc += GET(cc, 1); + } +/* None of them matched. */ +if (common->leave != NULL) + set_jumps(common->leave, LABEL()); + +OP1(SLJIT_MOV, TMP3, 0, SLJIT_IMM, 0); +jump = JUMP(SLJIT_JUMP); + +set_jumps(common->accept, LABEL()); +OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->recursive_head); +if (needsframe) + { + OP2(SLJIT_SUB, STACK_TOP, 0, STACK_TOP, 0, SLJIT_IMM, (framesize + alternativesize) * sizeof(sljit_w)); + add_jump(compiler, &common->revertframes, JUMP(SLJIT_FAST_CALL)); + OP2(SLJIT_ADD, STACK_TOP, 0, STACK_TOP, 0, SLJIT_IMM, (framesize + alternativesize) * sizeof(sljit_w)); + } +OP1(SLJIT_MOV, TMP3, 0, SLJIT_IMM, 1); + +JUMPHERE(jump); +copy_locals(common, ccbegin, ccend, FALSE, localsize + framesize + alternativesize, framesize + alternativesize); +free_stack(common, localsize + framesize + alternativesize); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(STACK_TOP), sizeof(sljit_w)); +OP1(SLJIT_MOV, TMP1, 0, TMP3, 0); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->recursive_head, TMP2, 0); +sljit_emit_fast_return(compiler, SLJIT_MEM1(STACK_TOP), 0); + +common->leavelabel = save_leavelabel; +common->leave = save_leave; +} + +#undef COMPILE_BACKTRACKPATH +#undef CURRENT_AS + +void +PRIV(jit_compile)(const REAL_PCRE *re, PUBL(extra) *extra, int mode) +{ +struct sljit_compiler *compiler; +backtrack_common rootbacktrack; +compiler_common common_data; +compiler_common *common = &common_data; +const pcre_uint8 *tables = re->tables; +pcre_study_data *study; +int localsize; +pcre_uchar *ccend; +executable_functions *functions; +void *executable_func; +sljit_uw executable_size; +struct sljit_label *mainloop = NULL; +struct sljit_label *empty_match_found; +struct sljit_label *empty_match_backtrack; +struct sljit_jump *jump; +struct sljit_jump *reqbyte_notfound = NULL; +struct sljit_jump *empty_match; + +SLJIT_ASSERT((extra->flags & PCRE_EXTRA_STUDY_DATA) != 0); +study = extra->study_data; + +if (!tables) + tables = PRIV(default_tables); + +memset(&rootbacktrack, 0, sizeof(backtrack_common)); +memset(common, 0, sizeof(compiler_common)); +rootbacktrack.cc = (pcre_uchar *)re + re->name_table_offset + re->name_count * re->name_entry_size; + +common->start = rootbacktrack.cc; +common->fcc = tables + fcc_offset; +common->lcc = (sljit_w)(tables + lcc_offset); +common->mode = mode; +common->nltype = NLTYPE_FIXED; +switch(re->options & PCRE_NEWLINE_BITS) + { + case 0: + /* Compile-time default */ + switch (NEWLINE) + { + case -1: common->newline = (CHAR_CR << 8) | CHAR_NL; common->nltype = NLTYPE_ANY; break; + case -2: common->newline = (CHAR_CR << 8) | CHAR_NL; common->nltype = NLTYPE_ANYCRLF; break; + default: common->newline = NEWLINE; break; + } + break; + case PCRE_NEWLINE_CR: common->newline = CHAR_CR; break; + case PCRE_NEWLINE_LF: common->newline = CHAR_NL; break; + case PCRE_NEWLINE_CR+ + PCRE_NEWLINE_LF: common->newline = (CHAR_CR << 8) | CHAR_NL; break; + case PCRE_NEWLINE_ANY: common->newline = (CHAR_CR << 8) | CHAR_NL; common->nltype = NLTYPE_ANY; break; + case PCRE_NEWLINE_ANYCRLF: common->newline = (CHAR_CR << 8) | CHAR_NL; common->nltype = NLTYPE_ANYCRLF; break; + default: return; + } +if ((re->options & PCRE_BSR_ANYCRLF) != 0) + common->bsr_nltype = NLTYPE_ANYCRLF; +else if ((re->options & PCRE_BSR_UNICODE) != 0) + common->bsr_nltype = NLTYPE_ANY; +else + { +#ifdef BSR_ANYCRLF + common->bsr_nltype = NLTYPE_ANYCRLF; +#else + common->bsr_nltype = NLTYPE_ANY; +#endif + } +common->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0; +common->ctypes = (sljit_w)(tables + ctypes_offset); +common->name_table = (sljit_w)((pcre_uchar *)re + re->name_table_offset); +common->name_count = re->name_count; +common->name_entry_size = re->name_entry_size; +common->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0; +#ifdef SUPPORT_UTF +/* PCRE_UTF16 has the same value as PCRE_UTF8. */ +common->utf = (re->options & PCRE_UTF8) != 0; +#ifdef SUPPORT_UCP +common->use_ucp = (re->options & PCRE_UCP) != 0; +#endif +#endif /* SUPPORT_UTF */ +ccend = bracketend(rootbacktrack.cc); + +/* Calculate the local space size on the stack. */ +common->ovector_start = CALL_LIMIT + sizeof(sljit_w); + +SLJIT_ASSERT(*rootbacktrack.cc == OP_BRA && ccend[-(1 + LINK_SIZE)] == OP_KET); +localsize = get_localspace(common, rootbacktrack.cc, ccend); +if (localsize < 0) + return; + +/* Checking flags and updating ovector_start. */ +if (mode == JIT_COMPILE && (re->flags & PCRE_REQCHSET) != 0 && (re->options & PCRE_NO_START_OPTIMIZE) == 0) + { + common->req_char_ptr = common->ovector_start; + common->ovector_start += sizeof(sljit_w); + } +if (mode != JIT_COMPILE) + { + common->start_used_ptr = common->ovector_start; + common->ovector_start += sizeof(sljit_w); + if (mode == JIT_PARTIAL_SOFT_COMPILE) + { + common->hit_start = common->ovector_start; + common->ovector_start += sizeof(sljit_w); + } + } +if ((re->options & PCRE_FIRSTLINE) != 0) + { + common->first_line_end = common->ovector_start; + common->ovector_start += sizeof(sljit_w); + } + +/* Aligning ovector to even number of sljit words. */ +if ((common->ovector_start & sizeof(sljit_w)) != 0) + common->ovector_start += sizeof(sljit_w); + +SLJIT_ASSERT(!(common->req_char_ptr != 0 && common->start_used_ptr != 0)); +common->cbraptr = OVECTOR_START + (re->top_bracket + 1) * 2 * sizeof(sljit_w); +localsize += common->cbraptr + (re->top_bracket + 1) * sizeof(sljit_w); +if (localsize > SLJIT_MAX_LOCAL_SIZE) + return; +common->localptrs = (int *)SLJIT_MALLOC((ccend - rootbacktrack.cc) * sizeof(int)); +if (!common->localptrs) + return; +memset(common->localptrs, 0, (ccend - rootbacktrack.cc) * sizeof(int)); +set_localptrs(common, common->cbraptr + (re->top_bracket + 1) * sizeof(sljit_w), ccend); + +compiler = sljit_create_compiler(); +if (!compiler) + { + SLJIT_FREE(common->localptrs); + return; + } +common->compiler = compiler; + +/* Main pcre_jit_exec entry. */ +sljit_emit_enter(compiler, 1, 5, 5, localsize); + +/* Register init. */ +reset_ovector(common, (re->top_bracket + 1) * 2); +if (common->req_char_ptr != 0) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->req_char_ptr, SLJIT_TEMPORARY_REG1, 0); + +OP1(SLJIT_MOV, ARGUMENTS, 0, SLJIT_SAVED_REG1, 0); +OP1(SLJIT_MOV, TMP1, 0, SLJIT_SAVED_REG1, 0); +OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, str)); +OP1(SLJIT_MOV, STR_END, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, end)); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, stack)); +OP1(SLJIT_MOV_SI, TMP1, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, calllimit)); +OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(TMP2), SLJIT_OFFSETOF(struct sljit_stack, base)); +OP1(SLJIT_MOV, STACK_LIMIT, 0, SLJIT_MEM1(TMP2), SLJIT_OFFSETOF(struct sljit_stack, limit)); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), CALL_LIMIT, TMP1, 0); + +if (mode == JIT_PARTIAL_SOFT_COMPILE) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->hit_start, SLJIT_IMM, 0); + +/* Main part of the matching */ +if ((re->options & PCRE_ANCHORED) == 0) + { + mainloop = mainloop_entry(common, (re->flags & PCRE_HASCRORLF) != 0, (re->options & PCRE_FIRSTLINE) != 0); + /* Forward search if possible. */ + if ((re->options & PCRE_NO_START_OPTIMIZE) == 0) + { + if ((re->flags & PCRE_FIRSTSET) != 0) + fast_forward_first_char(common, (pcre_uchar)re->first_char, (re->flags & PCRE_FCH_CASELESS) != 0, (re->options & PCRE_FIRSTLINE) != 0); + else if ((re->flags & PCRE_STARTLINE) != 0) + fast_forward_newline(common, (re->options & PCRE_FIRSTLINE) != 0); + else if ((re->flags & PCRE_STARTLINE) == 0 && study != NULL && (study->flags & PCRE_STUDY_MAPPED) != 0) + fast_forward_start_bits(common, (sljit_uw)study->start_bits, (re->options & PCRE_FIRSTLINE) != 0); + } + } +if (common->req_char_ptr != 0) + reqbyte_notfound = search_requested_char(common, (pcre_uchar)re->req_char, (re->flags & PCRE_RCH_CASELESS) != 0, (re->flags & PCRE_FIRSTSET) != 0); + +/* Store the current STR_PTR in OVECTOR(0). */ +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0), STR_PTR, 0); +/* Copy the limit of allowed recursions. */ +OP1(SLJIT_MOV, CALL_COUNT, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), CALL_LIMIT); +if (common->mark_ptr != 0) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->mark_ptr, SLJIT_IMM, 0); +/* Copy the beginning of the string. */ +if (mode == JIT_PARTIAL_SOFT_COMPILE) + { + jump = CMP(SLJIT_C_NOT_EQUAL, SLJIT_MEM1(SLJIT_LOCALS_REG), common->hit_start, SLJIT_IMM, 0); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr, STR_PTR, 0); + JUMPHERE(jump); + } +else if (mode == JIT_PARTIAL_HARD_COMPILE) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr, STR_PTR, 0); + +compile_trypath(common, rootbacktrack.cc, ccend, &rootbacktrack); +if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + { + sljit_free_compiler(compiler); + SLJIT_FREE(common->localptrs); + return; + } + +empty_match = CMP(SLJIT_C_EQUAL, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0)); +empty_match_found = LABEL(); + +common->acceptlabel = LABEL(); +if (common->accept != NULL) + set_jumps(common->accept, common->acceptlabel); + +/* This means we have a match. Update the ovector. */ +copy_ovector(common, re->top_bracket + 1); +common->leavelabel = LABEL(); +if (common->leave != NULL) + set_jumps(common->leave, common->leavelabel); +sljit_emit_return(compiler, SLJIT_MOV, SLJIT_RETURN_REG, 0); + +if (mode != JIT_COMPILE) + { + common->partialmatchlabel = LABEL(); + set_jumps(common->partialmatch, common->partialmatchlabel); + return_with_partial_match(common, common->leavelabel); + } + +empty_match_backtrack = LABEL(); +compile_backtrackpath(common, rootbacktrack.top); +if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + { + sljit_free_compiler(compiler); + SLJIT_FREE(common->localptrs); + return; + } + +SLJIT_ASSERT(rootbacktrack.prev == NULL); + +if (mode == JIT_PARTIAL_SOFT_COMPILE) + { + /* Update hit_start only in the first time. */ + jump = CMP(SLJIT_C_NOT_EQUAL, SLJIT_MEM1(SLJIT_LOCALS_REG), common->hit_start, SLJIT_IMM, -1); + OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->start_used_ptr, SLJIT_IMM, -1); + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), common->hit_start, TMP1, 0); + JUMPHERE(jump); + } + +/* Check we have remaining characters. */ +OP1(SLJIT_MOV, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), OVECTOR(0)); + +if ((re->options & PCRE_ANCHORED) == 0) + { + if ((re->options & PCRE_FIRSTLINE) == 0) + { + if (mode == JIT_COMPILE && study != NULL && study->minlength > 1 && (re->options & PCRE_NO_START_OPTIMIZE) == 0) + { + OP2(SLJIT_ADD, TMP1, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(study->minlength + 1)); + CMPTO(SLJIT_C_LESS_EQUAL, TMP1, 0, STR_END, 0, mainloop); + } + else + CMPTO(SLJIT_C_LESS, STR_PTR, 0, STR_END, 0, mainloop); + } + else + { + SLJIT_ASSERT(common->first_line_end != 0); + if (mode == JIT_COMPILE && study != NULL && study->minlength > 1 && (re->options & PCRE_NO_START_OPTIMIZE) == 0) + { + OP2(SLJIT_ADD, TMP1, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(study->minlength + 1)); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, STR_END, 0); + COND_VALUE(SLJIT_MOV, TMP2, 0, SLJIT_C_GREATER); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->first_line_end); + COND_VALUE(SLJIT_OR | SLJIT_SET_E, TMP2, 0, SLJIT_C_GREATER_EQUAL); + JUMPTO(SLJIT_C_ZERO, mainloop); + } + else + CMPTO(SLJIT_C_LESS, STR_PTR, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), common->first_line_end, mainloop); + } + } + +/* No more remaining characters. */ +if (reqbyte_notfound != NULL) + JUMPHERE(reqbyte_notfound); + +if (mode == JIT_PARTIAL_SOFT_COMPILE) + CMPTO(SLJIT_C_NOT_EQUAL, SLJIT_MEM1(SLJIT_LOCALS_REG), common->hit_start, SLJIT_IMM, 0, common->partialmatchlabel); + +OP1(SLJIT_MOV, SLJIT_RETURN_REG, 0, SLJIT_IMM, PCRE_ERROR_NOMATCH); +JUMPTO(SLJIT_JUMP, common->leavelabel); + +flush_stubs(common); + +JUMPHERE(empty_match); +OP1(SLJIT_MOV, TMP1, 0, ARGUMENTS, 0); +OP1(SLJIT_MOV_UB, TMP2, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, notempty)); +CMPTO(SLJIT_C_NOT_EQUAL, TMP2, 0, SLJIT_IMM, 0, empty_match_backtrack); +OP1(SLJIT_MOV_UB, TMP2, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, notempty_atstart)); +CMPTO(SLJIT_C_EQUAL, TMP2, 0, SLJIT_IMM, 0, empty_match_found); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, str)); +CMPTO(SLJIT_C_NOT_EQUAL, TMP2, 0, STR_PTR, 0, empty_match_found); +JUMPTO(SLJIT_JUMP, empty_match_backtrack); + +common->currententry = common->entries; +while (common->currententry != NULL) + { + /* Might add new entries. */ + compile_recurse(common); + if (SLJIT_UNLIKELY(sljit_get_compiler_error(compiler))) + { + sljit_free_compiler(compiler); + SLJIT_FREE(common->localptrs); + return; + } + flush_stubs(common); + common->currententry = common->currententry->next; + } + +/* Allocating stack, returns with PCRE_ERROR_JIT_STACKLIMIT if fails. */ +/* This is a (really) rare case. */ +set_jumps(common->stackalloc, LABEL()); +/* RETURN_ADDR is not a saved register. */ +sljit_emit_fast_enter(compiler, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0); +OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1, TMP2, 0); +OP1(SLJIT_MOV, TMP1, 0, ARGUMENTS, 0); +OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, stack)); +OP1(SLJIT_MOV, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(struct sljit_stack, top), STACK_TOP, 0); +OP2(SLJIT_ADD, TMP2, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(struct sljit_stack, limit), SLJIT_IMM, STACK_GROWTH_RATE); + +sljit_emit_ijump(compiler, SLJIT_CALL2, SLJIT_IMM, SLJIT_FUNC_OFFSET(sljit_stack_resize)); +jump = CMP(SLJIT_C_NOT_EQUAL, SLJIT_RETURN_REG, 0, SLJIT_IMM, 0); +OP1(SLJIT_MOV, TMP1, 0, ARGUMENTS, 0); +OP1(SLJIT_MOV, TMP1, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, stack)); +OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(struct sljit_stack, top)); +OP1(SLJIT_MOV, STACK_LIMIT, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(struct sljit_stack, limit)); +OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS1); +sljit_emit_fast_return(compiler, SLJIT_MEM1(SLJIT_LOCALS_REG), LOCALS0); + +/* Allocation failed. */ +JUMPHERE(jump); +/* We break the return address cache here, but this is a really rare case. */ +OP1(SLJIT_MOV, SLJIT_RETURN_REG, 0, SLJIT_IMM, PCRE_ERROR_JIT_STACKLIMIT); +JUMPTO(SLJIT_JUMP, common->leavelabel); + +/* Call limit reached. */ +set_jumps(common->calllimit, LABEL()); +OP1(SLJIT_MOV, SLJIT_RETURN_REG, 0, SLJIT_IMM, PCRE_ERROR_MATCHLIMIT); +JUMPTO(SLJIT_JUMP, common->leavelabel); + +if (common->revertframes != NULL) + { + set_jumps(common->revertframes, LABEL()); + do_revertframes(common); + } +if (common->wordboundary != NULL) + { + set_jumps(common->wordboundary, LABEL()); + check_wordboundary(common); + } +if (common->anynewline != NULL) + { + set_jumps(common->anynewline, LABEL()); + check_anynewline(common); + } +if (common->hspace != NULL) + { + set_jumps(common->hspace, LABEL()); + check_hspace(common); + } +if (common->vspace != NULL) + { + set_jumps(common->vspace, LABEL()); + check_vspace(common); + } +if (common->casefulcmp != NULL) + { + set_jumps(common->casefulcmp, LABEL()); + do_casefulcmp(common); + } +if (common->caselesscmp != NULL) + { + set_jumps(common->caselesscmp, LABEL()); + do_caselesscmp(common); + } +#ifdef SUPPORT_UTF +if (common->utfreadchar != NULL) + { + set_jumps(common->utfreadchar, LABEL()); + do_utfreadchar(common); + } +#ifdef COMPILE_PCRE8 +if (common->utfreadtype8 != NULL) + { + set_jumps(common->utfreadtype8, LABEL()); + do_utfreadtype8(common); + } +#endif +#endif /* COMPILE_PCRE8 */ +#ifdef SUPPORT_UCP +if (common->getucd != NULL) + { + set_jumps(common->getucd, LABEL()); + do_getucd(common); + } +#endif + +SLJIT_FREE(common->localptrs); +executable_func = sljit_generate_code(compiler); +executable_size = sljit_get_generated_code_size(compiler); +sljit_free_compiler(compiler); +if (executable_func == NULL) + return; + +/* Reuse the function descriptor if possible. */ +if ((extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0 && extra->executable_jit != NULL) + functions = (executable_functions *)extra->executable_jit; +else + { + functions = SLJIT_MALLOC(sizeof(executable_functions)); + if (functions == NULL) + { + /* This case is highly unlikely since we just recently + freed a lot of memory. Although not impossible. */ + sljit_free_code(executable_func); + return; + } + memset(functions, 0, sizeof(executable_functions)); + extra->executable_jit = functions; + extra->flags |= PCRE_EXTRA_EXECUTABLE_JIT; + } + +functions->executable_funcs[mode] = executable_func; +functions->executable_sizes[mode] = executable_size; +} + +static int jit_machine_stack_exec(jit_arguments *arguments, void* executable_func) +{ +union { + void* executable_func; + jit_function call_executable_func; +} convert_executable_func; +pcre_uint8 local_area[LOCAL_SPACE_SIZE]; +struct sljit_stack local_stack; + +local_stack.top = (sljit_w)&local_area; +local_stack.base = local_stack.top; +local_stack.limit = local_stack.base + LOCAL_SPACE_SIZE; +local_stack.max_limit = local_stack.limit; +arguments->stack = &local_stack; +convert_executable_func.executable_func = executable_func; +return convert_executable_func.call_executable_func(arguments); +} + +int +PRIV(jit_exec)(const REAL_PCRE *re, const PUBL(extra) *extra_data, const pcre_uchar *subject, + int length, int start_offset, int options, int *offsets, int offsetcount) +{ +executable_functions *functions = (executable_functions *)extra_data->executable_jit; +union { + void* executable_func; + jit_function call_executable_func; +} convert_executable_func; +jit_arguments arguments; +int maxoffsetcount; +int retval; +int mode = JIT_COMPILE; + +if ((options & PCRE_PARTIAL_HARD) != 0) + mode = JIT_PARTIAL_HARD_COMPILE; +else if ((options & PCRE_PARTIAL_SOFT) != 0) + mode = JIT_PARTIAL_SOFT_COMPILE; + +if (functions->executable_funcs[mode] == NULL) + return PCRE_ERROR_NULL; + +/* Sanity checks should be handled by pcre_exec. */ +arguments.stack = NULL; +arguments.str = subject + start_offset; +arguments.begin = subject; +arguments.end = subject + length; +arguments.mark_ptr = NULL; +/* JIT decreases this value less frequently than the interpreter. */ +arguments.calllimit = ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0) ? MATCH_LIMIT : extra_data->match_limit; +arguments.notbol = (options & PCRE_NOTBOL) != 0; +arguments.noteol = (options & PCRE_NOTEOL) != 0; +arguments.notempty = (options & PCRE_NOTEMPTY) != 0; +arguments.notempty_atstart = (options & PCRE_NOTEMPTY_ATSTART) != 0; +arguments.offsets = offsets; + +/* pcre_exec() rounds offsetcount to a multiple of 3, and then uses only 2/3 of +the output vector for storing captured strings, with the remainder used as +workspace. We don't need the workspace here. For compatibility, we limit the +number of captured strings in the same way as pcre_exec(), so that the user +gets the same result with and without JIT. */ + +if (offsetcount != 2) + offsetcount = ((offsetcount - (offsetcount % 3)) * 2) / 3; +maxoffsetcount = (re->top_bracket + 1) * 2; +if (offsetcount > maxoffsetcount) + offsetcount = maxoffsetcount; +arguments.offsetcount = offsetcount; + +if (functions->callback) + arguments.stack = (struct sljit_stack *)functions->callback(functions->userdata); +else + arguments.stack = (struct sljit_stack *)functions->userdata; + +if (arguments.stack == NULL) + retval = jit_machine_stack_exec(&arguments, functions->executable_funcs[mode]); +else + { + convert_executable_func.executable_func = functions->executable_funcs[mode]; + retval = convert_executable_func.call_executable_func(&arguments); + } + +if (retval * 2 > offsetcount) + retval = 0; +if ((extra_data->flags & PCRE_EXTRA_MARK) != 0) + *(extra_data->mark) = arguments.mark_ptr; + +return retval; +} + +void +PRIV(jit_free)(void *executable_funcs) +{ +int i; +executable_functions *functions = (executable_functions *)executable_funcs; +for (i = 0; i < JIT_NUMBER_OF_COMPILE_MODES; i++) + { + if (functions->executable_funcs[i] != NULL) + sljit_free_code(functions->executable_funcs[i]); + } +SLJIT_FREE(functions); +} + +int +PRIV(jit_get_size)(void *executable_funcs) +{ +int i; +sljit_uw size = 0; +sljit_uw *executable_sizes = ((executable_functions *)executable_funcs)->executable_sizes; +for (i = 0; i < JIT_NUMBER_OF_COMPILE_MODES; i++) + size += executable_sizes[i]; +return (int)size; +} + +const char* +PRIV(jit_get_target)(void) +{ +return sljit_get_platform_name(); +} + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DECL pcre_jit_stack * +pcre_jit_stack_alloc(int startsize, int maxsize) +#else +PCRE_EXP_DECL pcre16_jit_stack * +pcre16_jit_stack_alloc(int startsize, int maxsize) +#endif +{ +if (startsize < 1 || maxsize < 1) + return NULL; +if (startsize > maxsize) + startsize = maxsize; +startsize = (startsize + STACK_GROWTH_RATE - 1) & ~(STACK_GROWTH_RATE - 1); +maxsize = (maxsize + STACK_GROWTH_RATE - 1) & ~(STACK_GROWTH_RATE - 1); +return (PUBL(jit_stack)*)sljit_allocate_stack(startsize, maxsize); +} + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DECL void +pcre_jit_stack_free(pcre_jit_stack *stack) +#else +PCRE_EXP_DECL void +pcre16_jit_stack_free(pcre16_jit_stack *stack) +#endif +{ +sljit_free_stack((struct sljit_stack *)stack); +} + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DECL void +pcre_assign_jit_stack(pcre_extra *extra, pcre_jit_callback callback, void *userdata) +#else +PCRE_EXP_DECL void +pcre16_assign_jit_stack(pcre16_extra *extra, pcre16_jit_callback callback, void *userdata) +#endif +{ +executable_functions *functions; +if (extra != NULL && + (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0 && + extra->executable_jit != NULL) + { + functions = (executable_functions *)extra->executable_jit; + functions->callback = callback; + functions->userdata = userdata; + } +} + +#else /* SUPPORT_JIT */ + +/* These are dummy functions to avoid linking errors when JIT support is not +being compiled. */ + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DECL pcre_jit_stack * +pcre_jit_stack_alloc(int startsize, int maxsize) +#else +PCRE_EXP_DECL pcre16_jit_stack * +pcre16_jit_stack_alloc(int startsize, int maxsize) +#endif +{ +(void)startsize; +(void)maxsize; +return NULL; +} + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DECL void +pcre_jit_stack_free(pcre_jit_stack *stack) +#else +PCRE_EXP_DECL void +pcre16_jit_stack_free(pcre16_jit_stack *stack) +#endif +{ +(void)stack; +} + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DECL void +pcre_assign_jit_stack(pcre_extra *extra, pcre_jit_callback callback, void *userdata) +#else +PCRE_EXP_DECL void +pcre16_assign_jit_stack(pcre16_extra *extra, pcre16_jit_callback callback, void *userdata) +#endif +{ +(void)extra; +(void)callback; +(void)userdata; +} + +#endif + +/* End of pcre_jit_compile.c */ diff --git a/pcre/pcre_jit_test.c b/pcre/pcre_jit_test.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_jit_test.c @@ -0,0 +1,1279 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Main Library written by Philip Hazel + Copyright (c) 1997-2012 University of Cambridge + + This JIT compiler regression test program was written by Zoltan Herczeg + Copyright (c) 2010-2012 + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include +#include +#include "pcre.h" + +#define PCRE_BUG 0x80000000 + +/* + Letter characters: + \xe6\x92\xad = 0x64ad = 25773 (kanji) + Non-letter characters: + \xc2\xa1 = 0xa1 = (Inverted Exclamation Mark) + \xf3\xa9\xb7\x80 = 0xe9dc0 = 957888 + \xed\xa0\x80 = 55296 = 0xd800 (Invalid UTF character) + \xed\xb0\x80 = 56320 = 0xdc00 (Invalid UTF character) + Newlines: + \xc2\x85 = 0x85 = 133 (NExt Line = NEL) + \xe2\x80\xa8 = 0x2028 = 8232 (Line Separator) + Othercase pairs: + \xc3\xa9 = 0xe9 = 233 (e') + \xc3\x89 = 0xc9 = 201 (E') + \xc3\xa1 = 0xe1 = 225 (a') + \xc3\x81 = 0xc1 = 193 (A') + \xc8\xba = 0x23a = 570 + \xe2\xb1\xa5 = 0x2c65 = 11365 + \xe1\xbd\xb8 = 0x1f78 = 8056 + \xe1\xbf\xb8 = 0x1ff8 = 8184 + \xf0\x90\x90\x80 = 0x10400 = 66560 + \xf0\x90\x90\xa8 = 0x10428 = 66600 + Mark property: + \xcc\x8d = 0x30d = 781 + Special: + \xdf\xbf = 0x7ff = 2047 (highest 2 byte character) + \xe0\xa0\x80 = 0x800 = 2048 (lowest 2 byte character) + \xef\xbf\xbf = 0xffff = 65535 (highest 3 byte character) + \xf0\x90\x80\x80 = 0x10000 = 65536 (lowest 4 byte character) + \xf4\x8f\xbf\xbf = 0x10ffff = 1114111 (highest allowed utf character) +*/ + +static int regression_tests(void); + +int main(void) +{ + int jit = 0; +#ifdef SUPPORT_PCRE8 + pcre_config(PCRE_CONFIG_JIT, &jit); +#else + pcre16_config(PCRE_CONFIG_JIT, &jit); +#endif + if (!jit) { + printf("JIT must be enabled to run pcre_jit_test\n"); + return 1; + } + return regression_tests(); +} + +/* --------------------------------------------------------------------------------------- */ + +#if !(defined SUPPORT_PCRE8) && !(defined SUPPORT_PCRE16) +#error SUPPORT_PCRE8 or SUPPORT_PCRE16 must be defined +#endif + +#define MUA (PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_ANYCRLF) +#define MUAP (PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_ANYCRLF | PCRE_UCP) +#define CMUA (PCRE_CASELESS | PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_ANYCRLF) +#define CMUAP (PCRE_CASELESS | PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_ANYCRLF | PCRE_UCP) +#define MA (PCRE_MULTILINE | PCRE_NEWLINE_ANYCRLF) +#define MAP (PCRE_MULTILINE | PCRE_NEWLINE_ANYCRLF | PCRE_UCP) +#define CMA (PCRE_CASELESS | PCRE_MULTILINE | PCRE_NEWLINE_ANYCRLF) + +#define OFFSET_MASK 0x00ffff +#define F_NO8 0x010000 +#define F_NO16 0x020000 +#define F_NOMATCH 0x040000 +#define F_DIFF 0x080000 +#define F_FORCECONV 0x100000 +#define F_PROPERTY 0x200000 + +struct regression_test_case { + int flags; + int start_offset; + const char *pattern; + const char *input; +}; + +static struct regression_test_case regression_test_cases[] = { + /* Constant strings. */ + { MUA, 0, "AbC", "AbAbC" }, + { MUA, 0, "ACCEPT", "AACACCACCEACCEPACCEPTACCEPTT" }, + { CMUA, 0, "aA#\xc3\xa9\xc3\x81", "aA#Aa#\xc3\x89\xc3\xa1" }, + { MA, 0, "[^a]", "aAbB" }, + { CMA, 0, "[^m]", "mMnN" }, + { MA, 0, "a[^b][^#]", "abacd" }, + { CMA, 0, "A[^B][^E]", "abacd" }, + { CMUA, 0, "[^x][^#]", "XxBll" }, + { MUA, 0, "[^a]", "aaa\xc3\xa1#Ab" }, + { CMUA, 0, "[^A]", "aA\xe6\x92\xad" }, + { MUA, 0, "\\W(\\W)?\\w", "\r\n+bc" }, + { MUA, 0, "\\W(\\W)?\\w", "\n\r+bc" }, + { MUA, 0, "\\W(\\W)?\\w", "\r\r+bc" }, + { MUA, 0, "\\W(\\W)?\\w", "\n\n+bc" }, + { MUA, 0, "[axd]", "sAXd" }, + { CMUA, 0, "[axd]", "sAXd" }, + { CMUA, 0 | F_NOMATCH, "[^axd]", "DxA" }, + { MUA, 0, "[a-dA-C]", "\xe6\x92\xad\xc3\xa9.B" }, + { MUA, 0, "[^a-dA-C]", "\xe6\x92\xad\xc3\xa9" }, + { CMUA, 0, "[^\xc3\xa9]", "\xc3\xa9\xc3\x89." }, + { MUA, 0, "[^\xc3\xa9]", "\xc3\xa9\xc3\x89." }, + { MUA, 0, "[^a]", "\xc2\x80[]" }, + { CMUA, 0, "\xf0\x90\x90\xa7", "\xf0\x90\x91\x8f" }, + { CMA, 0, "1a2b3c4", "1a2B3c51A2B3C4" }, + { PCRE_CASELESS, 0, "\xff#a", "\xff#\xff\xfe##\xff#A" }, + { PCRE_CASELESS, 0, "\xfe", "\xff\xfc#\xfe\xfe" }, + { PCRE_CASELESS, 0, "a1", "Aa1" }, + { MA, 0, "\\Ca", "cda" }, + { CMA, 0, "\\Ca", "CDA" }, + { MA, 0 | F_NOMATCH, "\\Cx", "cda" }, + { CMA, 0 | F_NOMATCH, "\\Cx", "CDA" }, + { CMUAP, 0, "\xf0\x90\x90\x80\xf0\x90\x90\xa8", "\xf0\x90\x90\xa8\xf0\x90\x90\x80" }, + { CMUAP, 0, "\xf0\x90\x90\x80{2}", "\xf0\x90\x90\x80#\xf0\x90\x90\xa8\xf0\x90\x90\x80" }, + { CMUAP, 0, "\xf0\x90\x90\xa8{2}", "\xf0\x90\x90\x80#\xf0\x90\x90\xa8\xf0\x90\x90\x80" }, + { CMUAP, 0, "\xe1\xbd\xb8\xe1\xbf\xb8", "\xe1\xbf\xb8\xe1\xbd\xb8" }, + + /* Assertions. */ + { MUA, 0, "\\b[^A]", "A_B#" }, + { MA, 0 | F_NOMATCH, "\\b\\W", "\n*" }, + { MUA, 0, "\\B[^,]\\b[^s]\\b", "#X" }, + { MAP, 0, "\\B", "_\xa1" }, + { MAP, 0, "\\b_\\b[,A]\\B", "_," }, + { MUAP, 0, "\\b", "\xe6\x92\xad!" }, + { MUAP, 0, "\\B", "_\xc2\xa1\xc3\xa1\xc2\x85" }, + { MUAP, 0, "\\b[^A]\\B[^c]\\b[^_]\\B", "_\xc3\xa1\xe2\x80\xa8" }, + { MUAP, 0, "\\b\\w+\\B", "\xc3\x89\xc2\xa1\xe6\x92\xad\xc3\x81\xc3\xa1" }, + { MUA, 0 | F_NOMATCH, "\\b.", "\xcd\xbe" }, + { CMUAP, 0, "\\By", "\xf0\x90\x90\xa8y" }, + { MA, 0 | F_NOMATCH, "\\R^", "\n" }, + { MA, 1 | F_NOMATCH, "^", "\n" }, + { 0, 0, "^ab", "ab" }, + { 0, 0 | F_NOMATCH, "^ab", "aab" }, + { PCRE_MULTILINE | PCRE_NEWLINE_CRLF, 0, "^a", "\r\raa\n\naa\r\naa" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_ANYCRLF, 0, "^-", "\xe2\x80\xa8--\xc2\x85-\r\n-" }, + { PCRE_MULTILINE | PCRE_NEWLINE_ANY, 0, "^-", "a--b--\x85--" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_ANY, 0, "^-", "a--\xe2\x80\xa8--" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_ANY, 0, "^-", "a--\xc2\x85--" }, + { 0, 0, "ab$", "ab" }, + { 0, 0 | F_NOMATCH, "ab$", "abab\n\n" }, + { PCRE_DOLLAR_ENDONLY, 0 | F_NOMATCH, "ab$", "abab\r\n" }, + { PCRE_MULTILINE | PCRE_NEWLINE_CRLF, 0, "a$", "\r\raa\n\naa\r\naa" }, + { PCRE_MULTILINE | PCRE_NEWLINE_ANY, 0, "a$", "aaa" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_ANYCRLF, 0, "#$", "#\xc2\x85###\r#" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_ANY, 0, "#$", "#\xe2\x80\xa9" }, + { PCRE_NOTBOL | PCRE_NEWLINE_ANY, 0 | F_NOMATCH, "^a", "aa\naa" }, + { PCRE_NOTBOL | PCRE_MULTILINE | PCRE_NEWLINE_ANY, 0, "^a", "aa\naa" }, + { PCRE_NOTEOL | PCRE_NEWLINE_ANY, 0 | F_NOMATCH, "a$", "aa\naa" }, + { PCRE_NOTEOL | PCRE_NEWLINE_ANY, 0 | F_NOMATCH, "a$", "aa\r\n" }, + { PCRE_UTF8 | PCRE_DOLLAR_ENDONLY | PCRE_NEWLINE_ANY, 0 | F_PROPERTY, "\\p{Any}{2,}$", "aa\r\n" }, + { PCRE_NOTEOL | PCRE_MULTILINE | PCRE_NEWLINE_ANY, 0, "a$", "aa\naa" }, + { PCRE_NEWLINE_CR, 0, ".\\Z", "aaa" }, + { PCRE_NEWLINE_CR | PCRE_UTF8, 0, "a\\Z", "aaa\r" }, + { PCRE_NEWLINE_CR, 0, ".\\Z", "aaa\n" }, + { PCRE_NEWLINE_CRLF, 0, ".\\Z", "aaa\r" }, + { PCRE_NEWLINE_CRLF | PCRE_UTF8, 0, ".\\Z", "aaa\n" }, + { PCRE_NEWLINE_CRLF, 0, ".\\Z", "aaa\r\n" }, + { PCRE_NEWLINE_ANYCRLF | PCRE_UTF8, 0, ".\\Z", "aaa" }, + { PCRE_NEWLINE_ANYCRLF | PCRE_UTF8, 0, ".\\Z", "aaa\r" }, + { PCRE_NEWLINE_ANYCRLF | PCRE_UTF8, 0, ".\\Z", "aaa\n" }, + { PCRE_NEWLINE_ANYCRLF | PCRE_UTF8, 0, ".\\Z", "aaa\r\n" }, + { PCRE_NEWLINE_ANYCRLF | PCRE_UTF8, 0, ".\\Z", "aaa\xe2\x80\xa8" }, + { PCRE_NEWLINE_ANYCRLF | PCRE_UTF8, 0, ".\\Z", "aaa" }, + { PCRE_NEWLINE_ANYCRLF | PCRE_UTF8, 0, ".\\Z", "aaa\r" }, + { PCRE_NEWLINE_ANYCRLF | PCRE_UTF8, 0, ".\\Z", "aaa\n" }, + { PCRE_NEWLINE_ANYCRLF | PCRE_UTF8, 0, ".\\Z", "aaa\r\n" }, + { PCRE_NEWLINE_ANY | PCRE_UTF8, 0, ".\\Z", "aaa\xc2\x85" }, + { PCRE_NEWLINE_ANY | PCRE_UTF8, 0, ".\\Z", "aaa\xe2\x80\xa8" }, + { MA, 0, "\\Aa", "aaa" }, + { MA, 1 | F_NOMATCH, "\\Aa", "aaa" }, + { MA, 1, "\\Ga", "aaa" }, + { MA, 1 | F_NOMATCH, "\\Ga", "aba" }, + { MA, 0, "a\\z", "aaa" }, + { MA, 0 | F_NOMATCH, "a\\z", "aab" }, + + /* Brackets. */ + { MUA, 0, "(ab|bb|cd)", "bacde" }, + { MUA, 0, "(?:ab|a)(bc|c)", "ababc" }, + { MUA, 0, "((ab|(cc))|(bb)|(?:cd|efg))", "abac" }, + { CMUA, 0, "((aB|(Cc))|(bB)|(?:cd|EFg))", "AcCe" }, + { MUA, 0, "((ab|(cc))|(bb)|(?:cd|ebg))", "acebebg" }, + { MUA, 0, "(?:(a)|(?:b))(cc|(?:d|e))(a|b)k", "accabdbbccbk" }, + + /* Greedy and non-greedy ? operators. */ + { MUA, 0, "(?:a)?a", "laab" }, + { CMUA, 0, "(A)?A", "llaab" }, + { MUA, 0, "(a)?\?a", "aab" }, /* ?? is the prefix of trygraphs in GCC. */ + { MUA, 0, "(a)?a", "manm" }, + { CMUA, 0, "(a|b)?\?d((?:e)?)", "ABABdx" }, + { MUA, 0, "(a|b)?\?d((?:e)?)", "abcde" }, + { MUA, 0, "((?:ab)?\?g|b(?:g(nn|d)?\?)?)?\?(?:n)?m", "abgnbgnnbgdnmm" }, + + /* Greedy and non-greedy + operators */ + { MUA, 0, "(aa)+aa", "aaaaaaa" }, + { MUA, 0, "(aa)+?aa", "aaaaaaa" }, + { MUA, 0, "(?:aba|ab|a)+l", "ababamababal" }, + { MUA, 0, "(?:aba|ab|a)+?l", "ababamababal" }, + { MUA, 0, "(a(?:bc|cb|b|c)+?|ss)+e", "accssabccbcacbccbbXaccssabccbcacbccbbe" }, + { MUA, 0, "(a(?:bc|cb|b|c)+|ss)+?e", "accssabccbcacbccbbXaccssabccbcacbccbbe" }, + { MUA, 0, "(?:(b(c)+?)+)?\?(?:(bc)+|(cb)+)+(?:m)+", "bccbcccbcbccbcbPbccbcccbcbccbcbmmn" }, + + /* Greedy and non-greedy * operators */ + { CMUA, 0, "(?:AA)*AB", "aaaaaaamaaaaaaab" }, + { MUA, 0, "(?:aa)*?ab", "aaaaaaamaaaaaaab" }, + { MUA, 0, "(aa|ab)*ab", "aaabaaab" }, + { CMUA, 0, "(aa|Ab)*?aB", "aaabaaab" }, + { MUA, 0, "(a|b)*(?:a)*(?:b)*m", "abbbaaababanabbbaaababamm" }, + { MUA, 0, "(a|b)*?(?:a)*?(?:b)*?m", "abbbaaababanabbbaaababamm" }, + { MA, 0, "a(a(\\1*)a|(b)b+){0}a", "aa" }, + { MA, 0, "((?:a|)*){0}a", "a" }, + + /* Combining ? + * operators */ + { MUA, 0, "((bm)+)?\?(?:a)*(bm)+n|((am)+?)?(?:a)+(am)*n", "bmbmabmamaaamambmaman" }, + { MUA, 0, "(((ab)?cd)*ef)+g", "abcdcdefcdefefmabcdcdefcdefefgg" }, + { MUA, 0, "(((ab)?\?cd)*?ef)+?g", "abcdcdefcdefefmabcdcdefcdefefgg" }, + { MUA, 0, "(?:(ab)?c|(?:ab)+?d)*g", "ababcdccababddg" }, + { MUA, 0, "(?:(?:ab)?\?c|(ab)+d)*?g", "ababcdccababddg" }, + + /* Single character iterators. */ + { MUA, 0, "(a+aab)+aaaab", "aaaabcaaaabaabcaabcaaabaaaab" }, + { MUA, 0, "(a*a*aab)+x", "aaaaabaabaaabmaabx" }, + { MUA, 0, "(a*?(b|ab)a*?)+x", "aaaabcxbbaabaacbaaabaabax" }, + { MUA, 0, "(a+(ab|ad)a+)+x", "aaabaaaadaabaaabaaaadaaax" }, + { MUA, 0, "(a?(a)a?)+(aaa)", "abaaabaaaaaaaa" }, + { MUA, 0, "(a?\?(a)a?\?)+(b)", "aaaacaaacaacacbaaab" }, + { MUA, 0, "(a{0,4}(b))+d", "aaaaaabaabcaaaaabaaaaabd" }, + { MUA, 0, "(a{0,4}?[^b])+d+(a{0,4}[^b])d+", "aaaaadaaaacaadddaaddd" }, + { MUA, 0, "(ba{2})+c", "baabaaabacbaabaac" }, + { MUA, 0, "(a*+bc++)+", "aaabbcaaabcccab" }, + { MUA, 0, "(a?+[^b])+", "babaacacb" }, + { MUA, 0, "(a{0,3}+b)(a{0,3}+b)(a{0,3}+)[^c]", "abaabaaacbaabaaaac" }, + { CMUA, 0, "([a-c]+[d-f]+?)+?g", "aBdacdehAbDaFgA" }, + { CMUA, 0, "[c-f]+k", "DemmFke" }, + { MUA, 0, "([DGH]{0,4}M)+", "GGDGHDGMMHMDHHGHM" }, + { MUA, 0, "([a-c]{4,}s)+", "abasabbasbbaabsbba" }, + { CMUA, 0, "[ace]{3,7}", "AcbDAcEEcEd" }, + { CMUA, 0, "[ace]{3,7}?", "AcbDAcEEcEd" }, + { CMUA, 0, "[ace]{3,}", "AcbDAcEEcEd" }, + { CMUA, 0, "[ace]{3,}?", "AcbDAcEEcEd" }, + { MUA, 0, "[ckl]{2,}?g", "cdkkmlglglkcg" }, + { CMUA, 0, "[ace]{5}?", "AcCebDAcEEcEd" }, + { MUA, 0, "([AbC]{3,5}?d)+", "BACaAbbAEAACCbdCCbdCCAAbb" }, + { MUA, 0, "([^ab]{0,}s){2}", "abaabcdsABamsDDs" }, + { MUA, 0, "\\b\\w+\\B", "x,a_cd" }, + { MUAP, 0, "\\b[^\xc2\xa1]+\\B", "\xc3\x89\xc2\xa1\xe6\x92\xad\xc3\x81\xc3\xa1" }, + { CMUA, 0, "[^b]+(a*)([^c]?d{3})", "aaaaddd" }, + { CMUAP, 0, "\xe1\xbd\xb8{2}", "\xe1\xbf\xb8#\xe1\xbf\xb8\xe1\xbd\xb8" }, + { CMUA, 0, "[^\xf0\x90\x90\x80]{2,4}@", "\xf0\x90\x90\xa8\xf0\x90\x90\x80###\xf0\x90\x90\x80@@@" }, + { CMUA, 0, "[^\xe1\xbd\xb8][^\xc3\xa9]", "\xe1\xbd\xb8\xe1\xbf\xb8\xc3\xa9\xc3\x89#" }, + { MUA, 0, "[^\xe1\xbd\xb8][^\xc3\xa9]", "\xe1\xbd\xb8\xe1\xbf\xb8\xc3\xa9\xc3\x89#" }, + { MUA, 0, "[^\xe1\xbd\xb8]{3,}?", "##\xe1\xbd\xb8#\xe1\xbd\xb8#\xc3\x89#\xe1\xbd\xb8" }, + + /* Basic character sets. */ + { MUA, 0, "(?:\\s)+(?:\\S)+", "ab \t\xc3\xa9\xe6\x92\xad " }, + { MUA, 0, "(\\w)*(k)(\\W)?\?", "abcdef abck11" }, + { MUA, 0, "\\((\\d)+\\)\\D", "a() (83 (8)2 (9)ab" }, + { MUA, 0, "\\w(\\s|(?:\\d)*,)+\\w\\wb", "a 5, 4,, bb 5, 4,, aab" }, + { MUA, 0, "(\\v+)(\\V+)", "\x0e\xc2\x85\xe2\x80\xa8\x0b\x09\xe2\x80\xa9" }, + { MUA, 0, "(\\h+)(\\H+)", "\xe2\x80\xa8\xe2\x80\x80\x20\xe2\x80\x8a\xe2\x81\x9f\xe3\x80\x80\x09\x20\xc2\xa0\x0a" }, + + /* Unicode properties. */ + { MUAP, 0, "[1-5\xc3\xa9\\w]", "\xc3\xa1_" }, + { MUAP, 0 | F_PROPERTY, "[\xc3\x81\\p{Ll}]", "A_\xc3\x89\xc3\xa1" }, + { MUAP, 0, "[\\Wd-h_x-z]+", "a\xc2\xa1#_yhzdxi" }, + { MUAP, 0 | F_NOMATCH | F_PROPERTY, "[\\P{Any}]", "abc" }, + { MUAP, 0 | F_NOMATCH | F_PROPERTY, "[^\\p{Any}]", "abc" }, + { MUAP, 0 | F_NOMATCH | F_PROPERTY, "[\\P{Any}\xc3\xa1-\xc3\xa8]", "abc" }, + { MUAP, 0 | F_NOMATCH | F_PROPERTY, "[^\\p{Any}\xc3\xa1-\xc3\xa8]", "abc" }, + { MUAP, 0 | F_NOMATCH | F_PROPERTY, "[\xc3\xa1-\xc3\xa8\\P{Any}]", "abc" }, + { MUAP, 0 | F_NOMATCH | F_PROPERTY, "[^\xc3\xa1-\xc3\xa8\\p{Any}]", "abc" }, + { MUAP, 0 | F_PROPERTY, "[\xc3\xa1-\xc3\xa8\\p{Any}]", "abc" }, + { MUAP, 0 | F_PROPERTY, "[^\xc3\xa1-\xc3\xa8\\P{Any}]", "abc" }, + { MUAP, 0, "[b-\xc3\xa9\\s]", "a\xc\xe6\x92\xad" }, + { CMUAP, 0, "[\xc2\x85-\xc2\x89\xc3\x89]", "\xc2\x84\xc3\xa9" }, + { MUAP, 0, "[^b-d^&\\s]{3,}", "db^ !a\xe2\x80\xa8_ae" }, + { MUAP, 0 | F_PROPERTY, "[^\\S\\P{Any}][\\sN]{1,3}[\\P{N}]{4}", "\xe2\x80\xaa\xa N\x9\xc3\xa9_0" }, + { MUA, 0 | F_PROPERTY, "[^\\P{L}\x9!D-F\xa]{2,3}", "\x9,.DF\xa.CG\xc3\x81" }, + { CMUAP, 0, "[\xc3\xa1-\xc3\xa9_\xe2\x80\xa0-\xe2\x80\xaf]{1,5}[^\xe2\x80\xa0-\xe2\x80\xaf]", "\xc2\xa1\xc3\x89\xc3\x89\xe2\x80\xaf_\xe2\x80\xa0" }, + { MUAP, 0 | F_PROPERTY, "[\xc3\xa2-\xc3\xa6\xc3\x81-\xc3\x84\xe2\x80\xa8-\xe2\x80\xa9\xe6\x92\xad\\p{Zs}]{2,}", "\xe2\x80\xa7\xe2\x80\xa9\xe6\x92\xad \xe6\x92\xae" }, + { MUAP, 0 | F_PROPERTY, "[\\P{L&}]{2}[^\xc2\x85-\xc2\x89\\p{Ll}\\p{Lu}]{2}", "\xc3\xa9\xe6\x92\xad.a\xe6\x92\xad|\xc2\x8a#" }, + { PCRE_UCP, 0, "[a-b\\s]{2,5}[^a]", "AB baaa" }, + + /* Possible empty brackets. */ + { MUA, 0, "(?:|ab||bc|a)+d", "abcxabcabd" }, + { MUA, 0, "(|ab||bc|a)+d", "abcxabcabd" }, + { MUA, 0, "(?:|ab||bc|a)*d", "abcxabcabd" }, + { MUA, 0, "(|ab||bc|a)*d", "abcxabcabd" }, + { MUA, 0, "(?:|ab||bc|a)+?d", "abcxabcabd" }, + { MUA, 0, "(|ab||bc|a)+?d", "abcxabcabd" }, + { MUA, 0, "(?:|ab||bc|a)*?d", "abcxabcabd" }, + { MUA, 0, "(|ab||bc|a)*?d", "abcxabcabd" }, + { MUA, 0, "(((a)*?|(?:ba)+)+?|(?:|c|ca)*)*m", "abaacaccabacabalabaacaccabacabamm" }, + { MUA, 0, "(?:((?:a)*|(ba)+?)+|(|c|ca)*?)*?m", "abaacaccabacabalabaacaccabacabamm" }, + + /* Start offset. */ + { MUA, 3, "(\\d|(?:\\w)*\\w)+", "0ac01Hb" }, + { MUA, 4 | F_NOMATCH, "(\\w\\W\\w)+", "ab#d" }, + { MUA, 2 | F_NOMATCH, "(\\w\\W\\w)+", "ab#d" }, + { MUA, 1, "(\\w\\W\\w)+", "ab#d" }, + + /* Newline. */ + { PCRE_MULTILINE | PCRE_NEWLINE_CRLF, 0, "\\W{0,2}[^#]{3}", "\r\n#....." }, + { PCRE_MULTILINE | PCRE_NEWLINE_CR, 0, "\\W{0,2}[^#]{3}", "\r\n#....." }, + { PCRE_MULTILINE | PCRE_NEWLINE_CRLF, 0, "\\W{1,3}[^#]", "\r\n##...." }, + + /* Any character except newline or any newline. */ + { PCRE_NEWLINE_CRLF, 0, ".", "\r" }, + { PCRE_NEWLINE_CRLF | PCRE_UTF8, 0, ".(.).", "a\xc3\xa1\r\n\n\r\r" }, + { PCRE_NEWLINE_ANYCRLF, 0, ".(.)", "a\rb\nc\r\n\xc2\x85\xe2\x80\xa8" }, + { PCRE_NEWLINE_ANYCRLF | PCRE_UTF8, 0, ".(.)", "a\rb\nc\r\n\xc2\x85\xe2\x80\xa8" }, + { PCRE_NEWLINE_ANY | PCRE_UTF8, 0, "(.).", "a\rb\nc\r\n\xc2\x85\xe2\x80\xa9$de" }, + { PCRE_NEWLINE_ANYCRLF | PCRE_UTF8, 0 | F_NOMATCH, ".(.).", "\xe2\x80\xa8\nb\r" }, + { PCRE_NEWLINE_ANY, 0, "(.)(.)", "#\x85#\r#\n#\r\n#\x84" }, + { PCRE_NEWLINE_ANY | PCRE_UTF8, 0, "(.+)#", "#\rMn\xc2\x85#\n###" }, + { PCRE_BSR_ANYCRLF, 0, "\\R", "\r" }, + { PCRE_BSR_ANYCRLF, 0, "\\R", "\x85#\r\n#" }, + { PCRE_BSR_UNICODE | PCRE_UTF8, 0, "\\R", "ab\xe2\x80\xa8#c" }, + { PCRE_BSR_UNICODE | PCRE_UTF8, 0, "\\R", "ab\r\nc" }, + { PCRE_NEWLINE_CRLF | PCRE_BSR_UNICODE | PCRE_UTF8, 0, "(\\R.)+", "\xc2\x85\r\n#\xe2\x80\xa8\n\r\n\r" }, + { MUA, 0 | F_NOMATCH, "\\R+", "ab" }, + { MUA, 0, "\\R+", "ab\r\n\r" }, + { MUA, 0, "\\R*", "ab\r\n\r" }, + { MUA, 0, "\\R*", "\r\n\r" }, + { MUA, 0, "\\R{2,4}", "\r\nab\r\r" }, + { MUA, 0, "\\R{2,4}", "\r\nab\n\n\n\r\r\r" }, + { MUA, 0, "\\R{2,}", "\r\nab\n\n\n\r\r\r" }, + { MUA, 0, "\\R{0,3}", "\r\n\r\n\r\n\r\n\r\n" }, + { MUA, 0 | F_NOMATCH, "\\R+\\R\\R", "\r\n\r\n" }, + { MUA, 0, "\\R+\\R\\R", "\r\r\r" }, + { MUA, 0, "\\R*\\R\\R", "\n\r" }, + { MUA, 0 | F_NOMATCH, "\\R{2,4}\\R\\R", "\r\r\r" }, + { MUA, 0, "\\R{2,4}\\R\\R", "\r\r\r\r" }, + + /* Atomic groups (no fallback from "next" direction). */ + { MUA, 0 | F_NOMATCH, "(?>ab)ab", "bab" }, + { MUA, 0 | F_NOMATCH, "(?>(ab))ab", "bab" }, + { MUA, 0, "(?>ab)+abc(?>de)*def(?>gh)?ghe(?>ij)+?k(?>lm)*?n(?>op)?\?op", + "bababcdedefgheijijklmlmnop" }, + { MUA, 0, "(?>a(b)+a|(ab)?\?(b))an", "abban" }, + { MUA, 0, "(?>ab+a|(?:ab)?\?b)an", "abban" }, + { MUA, 0, "((?>ab|ad|)*?)(?>|c)*abad", "abababcababad" }, + { MUA, 0, "(?>(aa|b|)*+(?>(##)|###)*d|(aa)(?>(baa)?)m)", "aabaa#####da" }, + { MUA, 0, "((?>a|)+?)b", "aaacaaab" }, + { MUA, 0, "(?>x|)*$", "aaa" }, + { MUA, 0, "(?>(x)|)*$", "aaa" }, + { MUA, 0, "(?>x|())*$", "aaa" }, + { MUA, 0, "((?>[cxy]a|[a-d])*?)b", "aaa+ aaab" }, + { MUA, 0, "((?>[cxy](a)|[a-d])*?)b", "aaa+ aaab" }, + { MUA, 0, "(?>((?>(a+))))bab|(?>((?>(a+))))bb", "aaaabaaabaabab" }, + { MUA, 0, "(?>(?>a+))bab|(?>(?>a+))bb", "aaaabaaabaabab" }, + { MUA, 0, "(?>(a)c|(?>(c)|(a))a)b*?bab", "aaaabaaabaabab" }, + { MUA, 0, "(?>ac|(?>c|a)a)b*?bab", "aaaabaaabaabab" }, + { MUA, 0, "(?>(b)b|(a))*b(?>(c)|d)?x", "ababcaaabdbx" }, + { MUA, 0, "(?>bb|a)*b(?>c|d)?x", "ababcaaabdbx" }, + { MUA, 0, "(?>(bb)|a)*b(?>c|(d))?x", "ababcaaabdbx" }, + { MUA, 0, "(?>(a))*?(?>(a))+?(?>(a))??x", "aaaaaacccaaaaabax" }, + { MUA, 0, "(?>a)*?(?>a)+?(?>a)??x", "aaaaaacccaaaaabax" }, + { MUA, 0, "(?>(a)|)*?(?>(a)|)+?(?>(a)|)??x", "aaaaaacccaaaaabax" }, + { MUA, 0, "(?>a|)*?(?>a|)+?(?>a|)??x", "aaaaaacccaaaaabax" }, + { MUA, 0, "(?>a(?>(a{0,2}))*?b|aac)+b", "aaaaaaacaaaabaaaaacaaaabaacaaabb" }, + { CMA, 0, "(?>((?>a{32}|b+|(a*))?(?>c+|d*)?\?)+e)+?f", "aaccebbdde bbdaaaccebbdee bbdaaaccebbdeef" }, + { MUA, 0, "(?>(?:(?>aa|a||x)+?b|(?>aa|a||(x))+?c)?(?>[ad]{0,2})*?d)+d", "aaacdbaabdcabdbaaacd aacaabdbdcdcaaaadaabcbaadd" }, + { MUA, 0, "(?>(?:(?>aa|a||(x))+?b|(?>aa|a||x)+?c)?(?>[ad]{0,2})*?d)+d", "aaacdbaabdcabdbaaacd aacaabdbdcdcaaaadaabcbaadd" }, + { MUA, 0 | F_NOMATCH | F_PROPERTY, "\\X", "\xcc\x8d\xcc\x8d" }, + { MUA, 0 | F_PROPERTY, "\\X", "\xcc\x8d\xcc\x8d#\xcc\x8d\xcc\x8d" }, + { MUA, 0 | F_PROPERTY, "\\X+..", "\xcc\x8d#\xcc\x8d#\xcc\x8d\xcc\x8d" }, + { MUA, 0 | F_PROPERTY, "\\X{2,4}", "abcdef" }, + { MUA, 0 | F_PROPERTY, "\\X{2,4}?", "abcdef" }, + { MUA, 0 | F_NOMATCH | F_PROPERTY, "\\X{2,4}..", "#\xcc\x8d##" }, + { MUA, 0 | F_PROPERTY, "\\X{2,4}..", "#\xcc\x8d#\xcc\x8d##" }, + { MUA, 0, "(c(ab)?+ab)+", "cabcababcab" }, + { MUA, 0, "(?>(a+)b)+aabab", "aaaabaaabaabab" }, + + /* Possessive quantifiers. */ + { MUA, 0, "(?:a|b)++m", "mababbaaxababbaam" }, + { MUA, 0, "(?:a|b)*+m", "mababbaaxababbaam" }, + { MUA, 0, "(?:a|b)*+m", "ababbaaxababbaam" }, + { MUA, 0, "(a|b)++m", "mababbaaxababbaam" }, + { MUA, 0, "(a|b)*+m", "mababbaaxababbaam" }, + { MUA, 0, "(a|b)*+m", "ababbaaxababbaam" }, + { MUA, 0, "(a|b(*ACCEPT))++m", "maaxab" }, + { MUA, 0, "(?:b*)++m", "bxbbxbbbxm" }, + { MUA, 0, "(?:b*)++m", "bxbbxbbbxbbm" }, + { MUA, 0, "(?:b*)*+m", "bxbbxbbbxm" }, + { MUA, 0, "(?:b*)*+m", "bxbbxbbbxbbm" }, + { MUA, 0, "(b*)++m", "bxbbxbbbxm" }, + { MUA, 0, "(b*)++m", "bxbbxbbbxbbm" }, + { MUA, 0, "(b*)*+m", "bxbbxbbbxm" }, + { MUA, 0, "(b*)*+m", "bxbbxbbbxbbm" }, + { MUA, 0, "(?:a|(b))++m", "mababbaaxababbaam" }, + { MUA, 0, "(?:(a)|b)*+m", "mababbaaxababbaam" }, + { MUA, 0, "(?:(a)|(b))*+m", "ababbaaxababbaam" }, + { MUA, 0, "(a|(b))++m", "mababbaaxababbaam" }, + { MUA, 0, "((a)|b)*+m", "mababbaaxababbaam" }, + { MUA, 0, "((a)|(b))*+m", "ababbaaxababbaam" }, + { MUA, 0, "(a|(b)(*ACCEPT))++m", "maaxab" }, + { MUA, 0, "(?:(b*))++m", "bxbbxbbbxm" }, + { MUA, 0, "(?:(b*))++m", "bxbbxbbbxbbm" }, + { MUA, 0, "(?:(b*))*+m", "bxbbxbbbxm" }, + { MUA, 0, "(?:(b*))*+m", "bxbbxbbbxbbm" }, + { MUA, 0, "((b*))++m", "bxbbxbbbxm" }, + { MUA, 0, "((b*))++m", "bxbbxbbbxbbm" }, + { MUA, 0, "((b*))*+m", "bxbbxbbbxm" }, + { MUA, 0, "((b*))*+m", "bxbbxbbbxbbm" }, + { MUA, 0 | F_NOMATCH, "(?>(b{2,4}))(?:(?:(aa|c))++m|(?:(aa|c))+n)", "bbaacaaccaaaacxbbbmbn" }, + { MUA, 0, "((?:b)++a)+(cd)*+m", "bbababbacdcdnbbababbacdcdm" }, + { MUA, 0, "((?:(b))++a)+((c)d)*+m", "bbababbacdcdnbbababbacdcdm" }, + { MUA, 0, "(?:(?:(?:ab)*+k)++(?:n(?:cd)++)*+)*+m", "ababkkXababkkabkncXababkkabkncdcdncdXababkkabkncdcdncdkkabkncdXababkkabkncdcdncdkkabkncdm" }, + { MUA, 0, "(?:((ab)*+(k))++(n(?:c(d))++)*+)*+m", "ababkkXababkkabkncXababkkabkncdcdncdXababkkabkncdcdncdkkabkncdXababkkabkncdcdncdkkabkncdm" }, + + /* Back references. */ + { MUA, 0, "(aa|bb)(\\1*)(ll|)(\\3*)bbbbbbc", "aaaaaabbbbbbbbc" }, + { CMUA, 0, "(aa|bb)(\\1+)(ll|)(\\3+)bbbbbbc", "bBbbBbCbBbbbBbbcbbBbbbBBbbC" }, + { CMA, 0, "(a{2,4})\\1", "AaAaaAaA" }, + { MUA, 0, "(aa|bb)(\\1?)aa(\\1?)(ll|)(\\4+)bbc", "aaaaaaaabbaabbbbaabbbbc" }, + { MUA, 0, "(aa|bb)(\\1{0,5})(ll|)(\\3{0,5})cc", "bbxxbbbbxxaaaaaaaaaaaaaaaacc" }, + { MUA, 0, "(aa|bb)(\\1{3,5})(ll|)(\\3{3,5})cc", "bbbbbbbbbbbbaaaaaaccbbbbbbbbbbbbbbcc" }, + { MUA, 0, "(aa|bb)(\\1{3,})(ll|)(\\3{3,})cc", "bbbbbbbbbbbbaaaaaaccbbbbbbbbbbbbbbcc" }, + { MUA, 0, "(\\w+)b(\\1+)c", "GabGaGaDbGaDGaDc" }, + { MUA, 0, "(?:(aa)|b)\\1?b", "bb" }, + { CMUA, 0, "(aa|bb)(\\1*?)aa(\\1+?)", "bBBbaaAAaaAAaa" }, + { MUA, 0, "(aa|bb)(\\1*?)(dd|)cc(\\3+?)", "aaaaaccdd" }, + { CMUA, 0, "(?:(aa|bb)(\\1?\?)cc){2}(\\1?\?)", "aAaABBbbAAaAcCaAcCaA" }, + { MUA, 0, "(?:(aa|bb)(\\1{3,5}?)){2}(dd|)(\\3{3,5}?)", "aaaaaabbbbbbbbbbaaaaaaaaaaaaaa" }, + { CMA, 0, "(?:(aa|bb)(\\1{3,}?)){2}(dd|)(\\3{3,}?)", "aaaaaabbbbbbbbbbaaaaaaaaaaaaaa" }, + { MUA, 0, "(?:(aa|bb)(\\1{0,3}?)){2}(dd|)(\\3{0,3}?)b(\\1{0,3}?)(\\1{0,3})", "aaaaaaaaaaaaaaabaaaaa" }, + { MUA, 0, "(a(?:\\1|)a){3}b", "aaaaaaaaaaab" }, + { MA, 0, "(a?)b(\\1\\1*\\1+\\1?\\1*?\\1+?\\1??\\1*+\\1++\\1?+\\1{4}\\1{3,5}\\1{4,}\\1{0,5}\\1{3,5}?\\1{4,}?\\1{0,5}?\\1{3,5}+\\1{4,}+\\1{0,5}+#){2}d", "bb#b##d" }, + { MUAP, 0 | F_PROPERTY, "(\\P{N})\\1{2,}", ".www." }, + { MUAP, 0 | F_PROPERTY, "(\\P{N})\\1{0,2}", "wwwww." }, + { MUAP, 0 | F_PROPERTY, "(\\P{N})\\1{1,2}ww", "wwww" }, + { MUAP, 0 | F_PROPERTY, "(\\P{N})\\1{1,2}ww", "wwwww" }, + { PCRE_UCP, 0 | F_PROPERTY, "(\\P{N})\\1{2,}", ".www." }, + { CMUAP, 0, "(\xf0\x90\x90\x80)\\1", "\xf0\x90\x90\xa8\xf0\x90\x90\xa8" }, + + /* Assertions. */ + { MUA, 0, "(?=xx|yy|zz)\\w{4}", "abczzdefg" }, + { MUA, 0, "(?=((\\w+)b){3}|ab)", "dbbbb ab" }, + { MUA, 0, "(?!ab|bc|cd)[a-z]{2}", "Xabcdef" }, + { MUA, 0, "(?<=aaa|aa|a)a", "aaa" }, + { MUA, 2, "(?<=aaa|aa|a)a", "aaa" }, + { MA, 0, "(?<=aaa|aa|a)a", "aaa" }, + { MA, 2, "(?<=aaa|aa|a)a", "aaa" }, + { MUA, 0, "(\\d{2})(?!\\w+c|(((\\w?)m){2}n)+|\\1)", "x5656" }, + { MUA, 0, "((?=((\\d{2,6}\\w){2,}))\\w{5,20}K){2,}", "567v09708K12l00M00 567v09708K12l00M00K45K" }, + { MUA, 0, "(?=(?:(?=\\S+a)\\w*(b)){3})\\w+\\d", "bba bbab nbbkba nbbkba0kl" }, + { MUA, 0, "(?>a(?>(b+))a(?=(..)))*?k", "acabbcabbaabacabaabbakk" }, + { MUA, 0, "((?(?=(a))a)+k)", "bbak" }, + { MUA, 0, "((?(?=a)a)+k)", "bbak" }, + { MUA, 0 | F_NOMATCH, "(?=(?>(a))m)amk", "a k" }, + { MUA, 0 | F_NOMATCH, "(?!(?>(a))m)amk", "a k" }, + { MUA, 0 | F_NOMATCH, "(?>(?=(a))am)amk", "a k" }, + { MUA, 0, "(?=(?>a|(?=(?>(b+))a|c)[a-c]+)*?m)[a-cm]+k", "aaam bbam baaambaam abbabba baaambaamk" }, + { MUA, 0, "(?> ?\?\\b(?(?=\\w{1,4}(a))m)\\w{0,8}bc){2,}?", "bca ssbc mabd ssbc mabc" }, + { MUA, 0, "(?:(?=ab)?[^n][^n])+m", "ababcdabcdcdabnababcdabcdcdabm" }, + { MUA, 0, "(?:(?=a(b))?[^n][^n])+m", "ababcdabcdcdabnababcdabcdcdabm" }, + { MUA, 0, "(?:(?=.(.))??\\1.)+m", "aabbbcbacccanaabbbcbacccam" }, + { MUA, 0, "(?:(?=.)??[a-c])+m", "abacdcbacacdcaccam" }, + { MUA, 0, "((?!a)?(?!([^a]))?)+$", "acbab" }, + { MUA, 0, "((?!a)?\?(?!([^a]))?\?)+$", "acbab" }, + + /* Not empty, ACCEPT, FAIL */ + { MUA | PCRE_NOTEMPTY, 0 | F_NOMATCH, "a*", "bcx" }, + { MUA | PCRE_NOTEMPTY, 0, "a*", "bcaad" }, + { MUA | PCRE_NOTEMPTY, 0, "a*?", "bcaad" }, + { MUA | PCRE_NOTEMPTY_ATSTART, 0, "a*", "bcaad" }, + { MUA, 0, "a(*ACCEPT)b", "ab" }, + { MUA | PCRE_NOTEMPTY, 0 | F_NOMATCH, "a*(*ACCEPT)b", "bcx" }, + { MUA | PCRE_NOTEMPTY, 0, "a*(*ACCEPT)b", "bcaad" }, + { MUA | PCRE_NOTEMPTY, 0, "a*?(*ACCEPT)b", "bcaad" }, + { MUA | PCRE_NOTEMPTY, 0 | F_NOMATCH, "(?:z|a*(*ACCEPT)b)", "bcx" }, + { MUA | PCRE_NOTEMPTY, 0, "(?:z|a*(*ACCEPT)b)", "bcaad" }, + { MUA | PCRE_NOTEMPTY, 0, "(?:z|a*?(*ACCEPT)b)", "bcaad" }, + { MUA | PCRE_NOTEMPTY_ATSTART, 0, "a*(*ACCEPT)b", "bcx" }, + { MUA | PCRE_NOTEMPTY_ATSTART, 0 | F_NOMATCH, "a*(*ACCEPT)b", "" }, + { MUA, 0, "((a(*ACCEPT)b))", "ab" }, + { MUA, 0, "(a(*FAIL)a|a)", "aaa" }, + { MUA, 0, "(?=ab(*ACCEPT)b)a", "ab" }, + { MUA, 0, "(?=(?:x|ab(*ACCEPT)b))", "ab" }, + { MUA, 0, "(?=(a(b(*ACCEPT)b)))a", "ab" }, + { MUA | PCRE_NOTEMPTY, 0, "(?=a*(*ACCEPT))c", "c" }, + + /* Conditional blocks. */ + { MUA, 0, "(?(?=(a))a|b)+k", "ababbalbbadabak" }, + { MUA, 0, "(?(?!(b))a|b)+k", "ababbalbbadabak" }, + { MUA, 0, "(?(?=a)a|b)+k", "ababbalbbadabak" }, + { MUA, 0, "(?(?!b)a|b)+k", "ababbalbbadabak" }, + { MUA, 0, "(?(?=(a))a*|b*)+k", "ababbalbbadabak" }, + { MUA, 0, "(?(?!(b))a*|b*)+k", "ababbalbbadabak" }, + { MUA, 0, "(?(?!(b))(?:aaaaaa|a)|(?:bbbbbb|b))+aaaak", "aaaaaaaaaaaaaa bbbbbbbbbbbbbbb aaaaaaak" }, + { MUA, 0, "(?(?!b)(?:aaaaaa|a)|(?:bbbbbb|b))+aaaak", "aaaaaaaaaaaaaa bbbbbbbbbbbbbbb aaaaaaak" }, + { MUA, 0 | F_DIFF, "(?(?!(b))(?:aaaaaa|a)|(?:bbbbbb|b))+bbbbk", "aaaaaaaaaaaaaa bbbbbbbbbbbbbbb bbbbbbbk" }, + { MUA, 0, "(?(?!b)(?:aaaaaa|a)|(?:bbbbbb|b))+bbbbk", "aaaaaaaaaaaaaa bbbbbbbbbbbbbbb bbbbbbbk" }, + { MUA, 0, "(?(?=a)a*|b*)+k", "ababbalbbadabak" }, + { MUA, 0, "(?(?!b)a*|b*)+k", "ababbalbbadabak" }, + { MUA, 0, "(?(?=a)ab)", "a" }, + { MUA, 0, "(?(?a)?(?Pb)?(?(Name)c|d)*l", "bc ddd abccabccl" }, + { MUA, 0, "(?Pa)?(?Pb)?(?(Name)c|d)+?dd", "bcabcacdb bdddd" }, + { MUA, 0, "(?Pa)?(?Pb)?(?(Name)c|d)+l", "ababccddabdbccd abcccl" }, + + /* Set start of match. */ + { MUA, 0, "(?:\\Ka)*aaaab", "aaaaaaaa aaaaaaabb" }, + { MUA, 0, "(?>\\Ka\\Ka)*aaaab", "aaaaaaaa aaaaaaaaaabb" }, + { MUA, 0, "a+\\K(?<=\\Gaa)a", "aaaaaa" }, + { MUA | PCRE_NOTEMPTY, 0 | F_NOMATCH, "a\\K(*ACCEPT)b", "aa" }, + { MUA | PCRE_NOTEMPTY_ATSTART, 0, "a\\K(*ACCEPT)b", "aa" }, + + /* First line. */ + { MUA | PCRE_FIRSTLINE, 0 | F_PROPERTY, "\\p{Any}a", "bb\naaa" }, + { MUA | PCRE_FIRSTLINE, 0 | F_NOMATCH | F_PROPERTY, "\\p{Any}a", "bb\r\naaa" }, + { MUA | PCRE_FIRSTLINE, 0, "(?<=a)", "a" }, + { MUA | PCRE_FIRSTLINE, 0 | F_NOMATCH, "[^a][^b]", "ab" }, + { MUA | PCRE_FIRSTLINE, 0 | F_NOMATCH, "a", "\na" }, + { MUA | PCRE_FIRSTLINE, 0 | F_NOMATCH, "[abc]", "\na" }, + { MUA | PCRE_FIRSTLINE, 0 | F_NOMATCH, "^a", "\na" }, + { MUA | PCRE_FIRSTLINE, 0 | F_NOMATCH, "^(?<=\n)", "\na" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_ANY | PCRE_FIRSTLINE, 0 | F_NOMATCH, "#", "\xc2\x85#" }, + { PCRE_MULTILINE | PCRE_NEWLINE_ANY | PCRE_FIRSTLINE, 0 | F_NOMATCH, "#", "\x85#" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_ANY | PCRE_FIRSTLINE, 0 | F_NOMATCH, "^#", "\xe2\x80\xa8#" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_CRLF | PCRE_FIRSTLINE, 0 | F_PROPERTY, "\\p{Any}", "\r\na" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_CRLF | PCRE_FIRSTLINE, 0, ".", "\r" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_CRLF | PCRE_FIRSTLINE, 0, "a", "\ra" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_CRLF | PCRE_FIRSTLINE, 0 | F_NOMATCH, "ba", "bbb\r\nba" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_CRLF | PCRE_FIRSTLINE, 0 | F_NOMATCH | F_PROPERTY, "\\p{Any}{4}|a", "\r\na" }, + { PCRE_MULTILINE | PCRE_UTF8 | PCRE_NEWLINE_CRLF | PCRE_FIRSTLINE, 1, ".", "\r\n" }, + + /* Recurse. */ + { MUA, 0, "(a)(?1)", "aa" }, + { MUA, 0, "((a))(?1)", "aa" }, + { MUA, 0, "(b|a)(?1)", "aa" }, + { MUA, 0, "(b|(a))(?1)", "aa" }, + { MUA, 0 | F_NOMATCH, "((a)(b)(?:a*))(?1)", "aba" }, + { MUA, 0, "((a)(b)(?:a*))(?1)", "abab" }, + { MUA, 0, "((a+)c(?2))b(?1)", "aacaabaca" }, + { MUA, 0, "((?2)b|(a)){2}(?1)", "aabab" }, + { MUA, 0, "(?1)(a)*+(?2)(b(?1))", "aababa" }, + { MUA, 0, "(?1)(((a(*ACCEPT)))b)", "axaa" }, + { MUA, 0, "(?1)(?(DEFINE) (((ac(*ACCEPT)))b) )", "akaac" }, + { MUA, 0, "(a+)b(?1)b\\1", "abaaabaaaaa" }, + { MUA, 0 | F_NOMATCH, "(?(DEFINE)(aa|a))(?1)ab", "aab" }, + { MUA, 0, "(?(DEFINE)(a\\Kb))(?1)+ababc", "abababxabababc" }, + { MUA, 0, "(a\\Kb)(?1)+ababc", "abababxababababc" }, + { MUA, 0 | F_NOMATCH, "(a\\Kb)(?1)+ababc", "abababxababababxc" }, + { MUA, 0, "b|<(?R)*>", "<" }, + { MUA, 0, "(a\\K){0}(?:(?1)b|ac)", "ac" }, + { MUA, 0, "(?(DEFINE)(a(?2)|b)(b(?1)|(a)))(?:(?1)|(?2))m", "ababababnababababaam" }, + { MUA, 0, "(a)((?(R)a|b))(?2)", "aabbabaa" }, + { MUA, 0, "(a)((?(R2)a|b))(?2)", "aabbabaa" }, + { MUA, 0, "(a)((?(R1)a|b))(?2)", "ababba" }, + { MUA, 0, "(?(R0)aa|bb(?R))", "abba aabb bbaa" }, + { MUA, 0, "((?(R)(?:aaaa|a)|(?:(aaaa)|(a)))+)(?1)$", "aaaaaaaaaa aaaa" }, + { MUA, 0, "(?Pa(?(R&Name)a|b))(?1)", "aab abb abaa" }, + + /* 16 bit specific tests. */ + { CMA, 0 | F_FORCECONV, "\xc3\xa1", "\xc3\x81\xc3\xa1" }, + { CMA, 0 | F_FORCECONV, "\xe1\xbd\xb8", "\xe1\xbf\xb8\xe1\xbd\xb8" }, + { CMA, 0 | F_FORCECONV, "[\xc3\xa1]", "\xc3\x81\xc3\xa1" }, + { CMA, 0 | F_FORCECONV, "[\xe1\xbd\xb8]", "\xe1\xbf\xb8\xe1\xbd\xb8" }, + { CMA, 0 | F_FORCECONV, "[a-\xed\xb0\x80]", "A" }, + { CMA, 0 | F_NO8 | F_FORCECONV, "[a-\\x{dc00}]", "B" }, + { CMA, 0 | F_NO8 | F_NOMATCH | F_FORCECONV, "[b-\\x{dc00}]", "a" }, + { CMA, 0 | F_NO8 | F_FORCECONV, "\xed\xa0\x80\\x{d800}\xed\xb0\x80\\x{dc00}", "\xed\xa0\x80\xed\xa0\x80\xed\xb0\x80\xed\xb0\x80" }, + { CMA, 0 | F_NO8 | F_FORCECONV, "[\xed\xa0\x80\\x{d800}]{1,2}?[\xed\xb0\x80\\x{dc00}]{1,2}?#", "\xed\xa0\x80\xed\xa0\x80\xed\xb0\x80\xed\xb0\x80#" }, + { CMA, 0 | F_FORCECONV, "[\xed\xa0\x80\xed\xb0\x80#]{0,3}(?<=\xed\xb0\x80.)", "\xed\xa0\x80#\xed\xa0\x80##\xed\xb0\x80\xed\xa0\x80" }, + { CMA, 0 | F_FORCECONV, "[\xed\xa0\x80-\xed\xb3\xbf]", "\xed\x9f\xbf\xed\xa0\x83" }, + { CMA, 0 | F_FORCECONV, "[\xed\xa0\x80-\xed\xb3\xbf]", "\xed\xb4\x80\xed\xb3\xb0" }, + { CMA, 0 | F_NO8 | F_FORCECONV, "[\\x{d800}-\\x{dcff}]", "\xed\x9f\xbf\xed\xa0\x83" }, + { CMA, 0 | F_NO8 | F_FORCECONV, "[\\x{d800}-\\x{dcff}]", "\xed\xb4\x80\xed\xb3\xb0" }, + { CMA, 0 | F_FORCECONV, "[\xed\xa0\x80-\xef\xbf\xbf]+[\x1-\xed\xb0\x80]+#", "\xed\xa0\x85\xc3\x81\xed\xa0\x85\xef\xbf\xb0\xc2\x85\xed\xa9\x89#" }, + { CMA, 0 | F_FORCECONV, "[\xed\xa0\x80][\xed\xb0\x80]{2,}", "\xed\xa0\x80\xed\xb0\x80\xed\xa0\x80\xed\xb0\x80\xed\xb0\x80\xed\xb0\x80" }, + { MA, 0 | F_FORCECONV, "[^\xed\xb0\x80]{3,}?", "##\xed\xb0\x80#\xed\xb0\x80#\xc3\x89#\xed\xb0\x80" }, + { MA, 0 | F_NO8 | F_FORCECONV, "[^\\x{dc00}]{3,}?", "##\xed\xb0\x80#\xed\xb0\x80#\xc3\x89#\xed\xb0\x80" }, + { CMA, 0 | F_FORCECONV, ".\\B.", "\xed\xa0\x80\xed\xb0\x80" }, + { CMA, 0 | F_FORCECONV, "\\D+(?:\\d+|.)\\S+(?:\\s+|.)\\W+(?:\\w+|.)\xed\xa0\x80\xed\xa0\x80", "\xed\xa0\x80\xed\xa0\x80\xed\xa0\x80\xed\xa0\x80\xed\xa0\x80\xed\xa0\x80\xed\xa0\x80\xed\xa0\x80" }, + { CMA, 0 | F_FORCECONV, "\\d*\\s*\\w*\xed\xa0\x80\xed\xa0\x80", "\xed\xa0\x80\xed\xa0\x80" }, + { CMA, 0 | F_FORCECONV | F_NOMATCH, "\\d*?\\D*?\\s*?\\S*?\\w*?\\W*?##", "\xed\xa0\x80\xed\xa0\x80\xed\xa0\x80\xed\xa0\x80#" }, + { CMA | PCRE_EXTENDED, 0 | F_FORCECONV, "\xed\xa0\x80 \xed\xb0\x80 !", "\xed\xa0\x80\xed\xb0\x80!" }, + { CMA, 0 | F_FORCECONV, "\xed\xa0\x80+#[^#]+\xed\xa0\x80", "\xed\xa0\x80#a\xed\xa0\x80" }, + { CMA, 0 | F_FORCECONV, "(\xed\xa0\x80+)#\\1", "\xed\xa0\x80\xed\xa0\x80#\xed\xa0\x80\xed\xa0\x80" }, + { PCRE_MULTILINE | PCRE_NEWLINE_ANY, 0 | F_NO8 | F_FORCECONV, "^-", "a--\xe2\x80\xa8--" }, + { PCRE_BSR_UNICODE, 0 | F_NO8 | F_FORCECONV, "\\R", "ab\xe2\x80\xa8" }, + { 0, 0 | F_NO8 | F_FORCECONV, "\\v", "ab\xe2\x80\xa9" }, + { 0, 0 | F_NO8 | F_FORCECONV, "\\h", "ab\xe1\xa0\x8e" }, + { 0, 0 | F_NO8 | F_FORCECONV, "\\v+?\\V+?#", "\xe2\x80\xa9\xe2\x80\xa9\xef\xbf\xbf\xef\xbf\xbf#" }, + { 0, 0 | F_NO8 | F_FORCECONV, "\\h+?\\H+?#", "\xe1\xa0\x8e\xe1\xa0\x8e\xef\xbf\xbf\xef\xbf\xbf#" }, + + /* Partial matching. */ + { MUA | PCRE_PARTIAL_SOFT, 0, "ab", "a" }, + { MUA | PCRE_PARTIAL_SOFT, 0, "ab|a", "a" }, + { MUA | PCRE_PARTIAL_HARD, 0, "ab|a", "a" }, + { MUA | PCRE_PARTIAL_SOFT, 0, "\\b#", "a" }, + { MUA | PCRE_PARTIAL_SOFT, 0, "(?<=a)b", "a" }, + { MUA | PCRE_PARTIAL_SOFT, 0, "abc|(?<=xxa)bc", "xxab" }, + { MUA | PCRE_PARTIAL_SOFT, 0, "a\\B", "a" }, + { MUA | PCRE_PARTIAL_HARD, 0, "a\\b", "a" }, + + /* (*MARK) verb. */ + { MUA, 0, "a(*MARK:aa)a", "ababaa" }, + { MUA, 0 | F_NOMATCH, "a(*:aa)a", "abab" }, + { MUA, 0, "a(*:aa)(b(*:bb)b|bc)", "abc" }, + { MUA, 0 | F_NOMATCH, "a(*:1)x|b(*:2)y", "abc" }, + { MUA, 0, "(?>a(*:aa))b|ac", "ac" }, + { MUA, 0, "(?(DEFINE)(a(*:aa)))(?1)", "a" }, + { MUA, 0 | F_NOMATCH, "(?(DEFINE)((a)(*:aa)))(?1)b", "aa" }, + { MUA, 0, "(?(DEFINE)(a(*:aa)))a(?1)b|aac", "aac" }, + { MUA, 0, "(a(*:aa)){0}(?:b(?1)b|c)+c", "babbab cc" }, + { MUA, 0, "(a(*:aa)){0}(?:b(?1)b)+", "babba" }, + { MUA, 0 | F_NOMATCH, "(a(*:aa)){0}(?:b(?1)b)+", "ba" }, + { MUA, 0, "(a\\K(*:aa)){0}(?:b(?1)b|c)+c", "babbab cc" }, + { MUA, 0, "(a\\K(*:aa)){0}(?:b(?1)b)+", "babba" }, + { MUA, 0 | F_NOMATCH, "(a\\K(*:aa)){0}(?:b(?1)b)+", "ba" }, + + /* (*COMMIT) verb. */ + { MUA, 0 | F_NOMATCH, "a(*COMMIT)b", "ac" }, + { MUA, 0, "aa(*COMMIT)b", "xaxaab" }, + { MUA, 0 | F_NOMATCH, "a(*COMMIT)(*:msg)b|ac", "ac" }, + { MUA, 0, "(?=a(*COMMIT)b|ac)ac|(*:m)(a)c", "ac" }, + { MUA, 0, "(?!a(*COMMIT)(*:msg)b)a(c)|cd", "acd" }, + + /* Deep recursion. */ + { MUA, 0, "((((?:(?:(?:\\w)+)?)*|(?>\\w)+?)+|(?>\\w)?\?)*)?\\s", "aaaaa+ " }, + { MUA, 0, "(?:((?:(?:(?:\\w*?)+)??|(?>\\w)?|\\w*+)*)+)+?\\s", "aa+ " }, + { MUA, 0, "((a?)+)+b", "aaaaaaaaaaaa b" }, + + /* Deep recursion: Stack limit reached. */ + { MA, 0 | F_NOMATCH, "a?a?a?a?a?a?a?a?a?a?a?a?a?a?a?a?a?a?a?a?a?a?a?aaaaaaaaaaaaaaaaaaaaaaa", "aaaaaaaaaaaaaaaaaaaaaaa" }, + { MA, 0 | F_NOMATCH, "(?:a+)+b", "aaaaaaaaaaaaaaaaaaaaaaaa b" }, + { MA, 0 | F_NOMATCH, "(?:a+?)+?b", "aaaaaaaaaaaaaaaaaaaaaaaa b" }, + { MA, 0 | F_NOMATCH, "(?:a*)*b", "aaaaaaaaaaaaaaaaaaaaaaaa b" }, + { MA, 0 | F_NOMATCH, "(?:a*?)*?b", "aaaaaaaaaaaaaaaaaaaaaaaa b" }, + + { 0, 0, NULL, NULL } +}; + +static const unsigned char *tables(int mode) +{ + /* The purpose of this function to allow valgrind + for reporting invalid reads and writes. */ + static unsigned char *tables_copy; + const char *errorptr; + int erroroffset; + unsigned char *default_tables; +#ifdef SUPPORT_PCRE8 + pcre *regex; + char null_str[1] = { 0 }; +#else + pcre16 *regex; + PCRE_UCHAR16 null_str[1] = { 0 }; +#endif + + if (mode) { + if (tables_copy) + free(tables_copy); + tables_copy = NULL; + return NULL; + } + + if (tables_copy) + return tables_copy; + + default_tables = NULL; +#ifdef SUPPORT_PCRE8 + regex = pcre_compile(null_str, 0, &errorptr, &erroroffset, NULL); + if (regex) { + pcre_fullinfo(regex, NULL, PCRE_INFO_DEFAULT_TABLES, &default_tables); + pcre_free(regex); + } +#else + regex = pcre16_compile(null_str, 0, &errorptr, &erroroffset, NULL); + if (regex) { + pcre16_fullinfo(regex, NULL, PCRE_INFO_DEFAULT_TABLES, &default_tables); + pcre16_free(regex); + } +#endif + /* Shouldn't ever happen. */ + if (!default_tables) + return NULL; + + /* Unfortunately this value cannot get from pcre_fullinfo. + Since this is a test program, this is acceptable at the moment. */ + tables_copy = (unsigned char *)malloc(1088); + if (!tables_copy) + return NULL; + + memcpy(tables_copy, default_tables, 1088); + return tables_copy; +} + +#ifdef SUPPORT_PCRE8 +static pcre_jit_stack* callback8(void *arg) +{ + return (pcre_jit_stack *)arg; +} +#endif + +#ifdef SUPPORT_PCRE16 +static pcre16_jit_stack* callback16(void *arg) +{ + return (pcre16_jit_stack *)arg; +} +#endif + +#ifdef SUPPORT_PCRE8 +static void setstack8(pcre_extra *extra) +{ + static pcre_jit_stack *stack; + + if (!extra) { + if (stack) + pcre_jit_stack_free(stack); + stack = NULL; + return; + } + + if (!stack) + stack = pcre_jit_stack_alloc(1, 1024 * 1024); + /* Extra can be NULL. */ + pcre_assign_jit_stack(extra, callback8, stack); +} +#endif /* SUPPORT_PCRE8 */ + +#ifdef SUPPORT_PCRE16 +static void setstack16(pcre16_extra *extra) +{ + static pcre16_jit_stack *stack; + + if (!extra) { + if (stack) + pcre16_jit_stack_free(stack); + stack = NULL; + return; + } + + if (!stack) + stack = pcre16_jit_stack_alloc(1, 1024 * 1024); + /* Extra can be NULL. */ + pcre16_assign_jit_stack(extra, callback16, stack); +} +#endif /* SUPPORT_PCRE8 */ + +#ifdef SUPPORT_PCRE16 + +static int convert_utf8_to_utf16(const char *input, PCRE_UCHAR16 *output, int *offsetmap, int max_length) +{ + unsigned char *iptr = (unsigned char*)input; + unsigned short *optr = (unsigned short *)output; + unsigned int c; + + if (max_length == 0) + return 0; + + while (*iptr && max_length > 1) { + c = 0; + if (offsetmap) + *offsetmap++ = (int)(iptr - (unsigned char*)input); + + if (!(*iptr & 0x80)) + c = *iptr++; + else if (!(*iptr & 0x20)) { + c = ((iptr[0] & 0x1f) << 6) | (iptr[1] & 0x3f); + iptr += 2; + } else if (!(*iptr & 0x10)) { + c = ((iptr[0] & 0x0f) << 12) | ((iptr[1] & 0x3f) << 6) | (iptr[2] & 0x3f); + iptr += 3; + } else if (!(*iptr & 0x08)) { + c = ((iptr[0] & 0x07) << 18) | ((iptr[1] & 0x3f) << 12) | ((iptr[2] & 0x3f) << 6) | (iptr[3] & 0x3f); + iptr += 4; + } + + if (c < 65536) { + *optr++ = c; + max_length--; + } else if (max_length <= 2) { + *optr = '\0'; + return (int)(optr - (unsigned short *)output); + } else { + c -= 0x10000; + *optr++ = 0xd800 | ((c >> 10) & 0x3ff); + *optr++ = 0xdc00 | (c & 0x3ff); + max_length -= 2; + if (offsetmap) + offsetmap++; + } + } + if (offsetmap) + *offsetmap = (int)(iptr - (unsigned char*)input); + *optr = '\0'; + return (int)(optr - (unsigned short *)output); +} + +static int copy_char8_to_char16(const char *input, PCRE_UCHAR16 *output, int max_length) +{ + unsigned char *iptr = (unsigned char*)input; + unsigned short *optr = (unsigned short *)output; + + if (max_length == 0) + return 0; + + while (*iptr && max_length > 1) { + *optr++ = *iptr++; + max_length--; + } + *optr = '\0'; + return (int)(optr - (unsigned short *)output); +} + +#define REGTEST_MAX_LENGTH 4096 +static PCRE_UCHAR16 regtest_buf[REGTEST_MAX_LENGTH]; +static int regtest_offsetmap[REGTEST_MAX_LENGTH]; + +#endif /* SUPPORT_PCRE16 */ + +static int check_ascii(const char *input) +{ + const unsigned char *ptr = (unsigned char *)input; + while (*ptr) { + if (*ptr > 127) + return 0; + ptr++; + } + return 1; +} + +static int regression_tests(void) +{ + struct regression_test_case *current = regression_test_cases; + const char *error; + char *cpu_info; + int i, err_offs; + int is_successful, is_ascii_pattern, is_ascii_input; + int total = 0; + int successful = 0; + int successful_row = 0; + int counter = 0; + int study_mode; +#ifdef SUPPORT_PCRE8 + pcre *re8; + pcre_extra *extra8; + pcre_extra dummy_extra8; + int ovector8_1[32]; + int ovector8_2[32]; + int return_value8_1, return_value8_2; + unsigned char *mark8_1, *mark8_2; + int utf8 = 0, ucp8 = 0; + int disabled_flags8 = 0; +#endif +#ifdef SUPPORT_PCRE16 + pcre16 *re16; + pcre16_extra *extra16; + pcre16_extra dummy_extra16; + int ovector16_1[32]; + int ovector16_2[32]; + int return_value16_1, return_value16_2; + PCRE_UCHAR16 *mark16_1, *mark16_2; + int utf16 = 0, ucp16 = 0; + int disabled_flags16 = 0; + int length16; +#endif + + /* This test compares the behaviour of interpreter and JIT. Although disabling + utf or ucp may make tests fail, if the pcre_exec result is the SAME, it is + still considered successful from pcre_jit_test point of view. */ + +#ifdef SUPPORT_PCRE8 + pcre_config(PCRE_CONFIG_JITTARGET, &cpu_info); +#else + pcre16_config(PCRE_CONFIG_JITTARGET, &cpu_info); +#endif + + printf("Running JIT regression tests\n"); + printf(" target CPU of SLJIT compiler: %s\n", cpu_info); + +#ifdef SUPPORT_PCRE8 + pcre_config(PCRE_CONFIG_UTF8, &utf8); + pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &ucp8); + if (!utf8) + disabled_flags8 |= PCRE_UTF8; + if (!ucp8) + disabled_flags8 |= PCRE_UCP; + printf(" in 8 bit mode with utf8 %s and ucp %s:\n", utf8 ? "enabled" : "disabled", ucp8 ? "enabled" : "disabled"); +#endif +#ifdef SUPPORT_PCRE16 + pcre16_config(PCRE_CONFIG_UTF16, &utf16); + pcre16_config(PCRE_CONFIG_UNICODE_PROPERTIES, &ucp16); + if (!utf16) + disabled_flags16 |= PCRE_UTF8; + if (!ucp16) + disabled_flags16 |= PCRE_UCP; + printf(" in 16 bit mode with utf16 %s and ucp %s:\n", utf16 ? "enabled" : "disabled", ucp16 ? "enabled" : "disabled"); +#endif + + while (current->pattern) { + /* printf("\nPattern: %s :\n", current->pattern); */ + total++; + if (current->start_offset & F_PROPERTY) { + is_ascii_pattern = 0; + is_ascii_input = 0; + } else { + is_ascii_pattern = check_ascii(current->pattern); + is_ascii_input = check_ascii(current->input); + } + + if (current->flags & PCRE_PARTIAL_SOFT) + study_mode = PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE; + else if (current->flags & PCRE_PARTIAL_HARD) + study_mode = PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE; + else + study_mode = PCRE_STUDY_JIT_COMPILE; + error = NULL; +#ifdef SUPPORT_PCRE8 + re8 = NULL; + if (!(current->start_offset & F_NO8)) + re8 = pcre_compile(current->pattern, + current->flags & ~(PCRE_NOTBOL | PCRE_NOTEOL | PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART | PCRE_PARTIAL_SOFT | PCRE_PARTIAL_HARD | disabled_flags8), + &error, &err_offs, tables(0)); + + extra8 = NULL; + if (re8) { + error = NULL; + extra8 = pcre_study(re8, study_mode, &error); + if (!extra8) { + printf("\n8 bit: Cannot study pattern: %s\n", current->pattern); + pcre_free(re8); + re8 = NULL; + } + if (!(extra8->flags & PCRE_EXTRA_EXECUTABLE_JIT)) { + printf("\n8 bit: JIT compiler does not support: %s\n", current->pattern); + pcre_free_study(extra8); + pcre_free(re8); + re8 = NULL; + } + extra8->flags |= PCRE_EXTRA_MARK; + } else if (((utf8 && ucp8) || is_ascii_pattern) && !(current->start_offset & F_NO8)) + printf("\n8 bit: Cannot compile pattern: %s\n", current->pattern); +#endif +#ifdef SUPPORT_PCRE16 + if ((current->flags & PCRE_UTF8) || (current->start_offset & F_FORCECONV)) + convert_utf8_to_utf16(current->pattern, regtest_buf, NULL, REGTEST_MAX_LENGTH); + else + copy_char8_to_char16(current->pattern, regtest_buf, REGTEST_MAX_LENGTH); + + re16 = NULL; + if (!(current->start_offset & F_NO16)) + re16 = pcre16_compile(regtest_buf, + current->flags & ~(PCRE_NOTBOL | PCRE_NOTEOL | PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART | PCRE_PARTIAL_SOFT | PCRE_PARTIAL_HARD | disabled_flags16), + &error, &err_offs, tables(0)); + + extra16 = NULL; + if (re16) { + error = NULL; + extra16 = pcre16_study(re16, study_mode, &error); + if (!extra16) { + printf("\n16 bit: Cannot study pattern: %s\n", current->pattern); + pcre16_free(re16); + re16 = NULL; + } + if (!(extra16->flags & PCRE_EXTRA_EXECUTABLE_JIT)) { + printf("\n16 bit: JIT compiler does not support: %s\n", current->pattern); + pcre16_free_study(extra16); + pcre16_free(re16); + re16 = NULL; + } + extra16->flags |= PCRE_EXTRA_MARK; + } else if (((utf16 && ucp16) || is_ascii_pattern) && !(current->start_offset & F_NO16)) + printf("\n16 bit: Cannot compile pattern: %s\n", current->pattern); +#endif + + counter++; + if ((counter & 0x3) != 0) { +#ifdef SUPPORT_PCRE8 + setstack8(NULL); +#endif +#ifdef SUPPORT_PCRE16 + setstack16(NULL); +#endif + } + +#ifdef SUPPORT_PCRE8 + return_value8_1 = -1000; + return_value8_2 = -1000; + for (i = 0; i < 32; ++i) + ovector8_1[i] = -2; + for (i = 0; i < 32; ++i) + ovector8_2[i] = -2; + if (re8) { + mark8_1 = NULL; + mark8_2 = NULL; + setstack8(extra8); + extra8->mark = &mark8_1; + return_value8_1 = pcre_exec(re8, extra8, current->input, strlen(current->input), current->start_offset & OFFSET_MASK, + current->flags & (PCRE_NOTBOL | PCRE_NOTEOL | PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART | PCRE_PARTIAL_SOFT | PCRE_PARTIAL_HARD), ovector8_1, 32); + memset(&dummy_extra8, 0, sizeof(pcre_extra)); + dummy_extra8.flags = PCRE_EXTRA_MARK; + dummy_extra8.mark = &mark8_2; + return_value8_2 = pcre_exec(re8, &dummy_extra8, current->input, strlen(current->input), current->start_offset & OFFSET_MASK, + current->flags & (PCRE_NOTBOL | PCRE_NOTEOL | PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART | PCRE_PARTIAL_SOFT | PCRE_PARTIAL_HARD), ovector8_2, 32); + } +#endif + +#ifdef SUPPORT_PCRE16 + return_value16_1 = -1000; + return_value16_2 = -1000; + for (i = 0; i < 32; ++i) + ovector16_1[i] = -2; + for (i = 0; i < 32; ++i) + ovector16_2[i] = -2; + if (re16) { + mark16_1 = NULL; + mark16_2 = NULL; + setstack16(extra16); + if ((current->flags & PCRE_UTF8) || (current->start_offset & F_FORCECONV)) + length16 = convert_utf8_to_utf16(current->input, regtest_buf, regtest_offsetmap, REGTEST_MAX_LENGTH); + else + length16 = copy_char8_to_char16(current->input, regtest_buf, REGTEST_MAX_LENGTH); + extra16->mark = &mark16_1; + return_value16_1 = pcre16_exec(re16, extra16, regtest_buf, length16, current->start_offset & OFFSET_MASK, + current->flags & (PCRE_NOTBOL | PCRE_NOTEOL | PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART | PCRE_PARTIAL_SOFT | PCRE_PARTIAL_HARD), ovector16_1, 32); + memset(&dummy_extra16, 0, sizeof(pcre16_extra)); + dummy_extra16.flags = PCRE_EXTRA_MARK; + dummy_extra16.mark = &mark16_2; + return_value16_2 = pcre16_exec(re16, &dummy_extra16, regtest_buf, length16, current->start_offset & OFFSET_MASK, + current->flags & (PCRE_NOTBOL | PCRE_NOTEOL | PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART | PCRE_PARTIAL_SOFT | PCRE_PARTIAL_HARD), ovector16_2, 32); + } +#endif + + /* printf("[%d-%d|%d-%d|%d-%d]%s", return_value8_1, return_value16_1, ovector8_1[0], ovector8_1[1], ovector16_1[0], ovector16_1[1], (current->flags & PCRE_CASELESS) ? "C" : ""); */ + + /* If F_DIFF is set, just run the test, but do not compare the results. + Segfaults can still be captured. */ + + is_successful = 1; + if (!(current->start_offset & F_DIFF)) { +#if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16 + if (utf8 == utf16 && !(current->start_offset & F_FORCECONV)) { + /* All results must be the same. */ + if (return_value8_1 != return_value8_2 || return_value8_1 != return_value16_1 || return_value8_1 != return_value16_2) { + printf("\n8 and 16 bit: Return value differs(%d:%d:%d:%d): [%d] '%s' @ '%s'\n", + return_value8_1, return_value8_2, return_value16_1, return_value16_2, + total, current->pattern, current->input); + is_successful = 0; + } else if (return_value8_1 >= 0 || return_value8_1 == PCRE_ERROR_PARTIAL) { + if (return_value8_1 == PCRE_ERROR_PARTIAL) { + return_value8_1 = 2; + return_value16_1 = 2; + } else { + return_value8_1 *= 2; + return_value16_1 *= 2; + } + + /* Transform back the results. */ + if (current->flags & PCRE_UTF8) { + for (i = 0; i < return_value8_1; ++i) { + if (ovector16_1[i] >= 0) + ovector16_1[i] = regtest_offsetmap[ovector16_1[i]]; + if (ovector16_2[i] >= 0) + ovector16_2[i] = regtest_offsetmap[ovector16_2[i]]; + } + } + + for (i = 0; i < return_value8_1; ++i) + if (ovector8_1[i] != ovector8_2[i] || ovector8_1[i] != ovector16_1[i] || ovector8_1[i] != ovector16_2[i]) { + printf("\n8 and 16 bit: Ovector[%d] value differs(%d:%d:%d:%d): [%d] '%s' @ '%s' \n", + i, ovector8_1[i], ovector8_2[i], ovector16_1[i], ovector16_2[i], + total, current->pattern, current->input); + is_successful = 0; + } + } + } else { +#endif /* SUPPORT_PCRE8 && SUPPORT_PCRE16 */ + /* Only the 8 bit and 16 bit results must be equal. */ +#ifdef SUPPORT_PCRE8 + if (return_value8_1 != return_value8_2) { + printf("\n8 bit: Return value differs(%d:%d): [%d] '%s' @ '%s'\n", + return_value8_1, return_value8_2, total, current->pattern, current->input); + is_successful = 0; + } else if (return_value8_1 >= 0 || return_value8_1 == PCRE_ERROR_PARTIAL) { + if (return_value8_1 == PCRE_ERROR_PARTIAL) + return_value8_1 = 2; + else + return_value8_1 *= 2; + + for (i = 0; i < return_value8_1; ++i) + if (ovector8_1[i] != ovector8_2[i]) { + printf("\n8 bit: Ovector[%d] value differs(%d:%d): [%d] '%s' @ '%s'\n", + i, ovector8_1[i], ovector8_2[i], total, current->pattern, current->input); + is_successful = 0; + } + } +#endif + +#ifdef SUPPORT_PCRE16 + if (return_value16_1 != return_value16_2) { + printf("\n16 bit: Return value differs(%d:%d): [%d] '%s' @ '%s'\n", + return_value16_1, return_value16_2, total, current->pattern, current->input); + is_successful = 0; + } else if (return_value16_1 >= 0 || return_value16_1 == PCRE_ERROR_PARTIAL) { + if (return_value16_1 == PCRE_ERROR_PARTIAL) + return_value16_1 = 2; + else + return_value16_1 *= 2; + + for (i = 0; i < return_value16_1; ++i) + if (ovector16_1[i] != ovector16_2[i]) { + printf("\n16 bit: Ovector[%d] value differs(%d:%d): [%d] '%s' @ '%s'\n", + i, ovector16_1[i], ovector16_2[i], total, current->pattern, current->input); + is_successful = 0; + } + } +#endif + +#if defined SUPPORT_PCRE8 && defined SUPPORT_PCRE16 + } +#endif /* SUPPORT_PCRE8 && SUPPORT_PCRE16 */ + } + + if (is_successful) { +#ifdef SUPPORT_PCRE8 + if (!(current->start_offset & F_NO8) && ((utf8 && ucp8) || is_ascii_input)) { + if (return_value8_1 < 0 && !(current->start_offset & F_NOMATCH)) { + printf("8 bit: Test should match: [%d] '%s' @ '%s'\n", + total, current->pattern, current->input); + is_successful = 0; + } + + if (return_value8_1 >= 0 && (current->start_offset & F_NOMATCH)) { + printf("8 bit: Test should not match: [%d] '%s' @ '%s'\n", + total, current->pattern, current->input); + is_successful = 0; + } + } +#endif +#ifdef SUPPORT_PCRE16 + if (!(current->start_offset & F_NO16) && ((utf16 && ucp16) || is_ascii_input)) { + if (return_value16_1 < 0 && !(current->start_offset & F_NOMATCH)) { + printf("16 bit: Test should match: [%d] '%s' @ '%s'\n", + total, current->pattern, current->input); + is_successful = 0; + } + + if (return_value16_1 >= 0 && (current->start_offset & F_NOMATCH)) { + printf("16 bit: Test should not match: [%d] '%s' @ '%s'\n", + total, current->pattern, current->input); + is_successful = 0; + } + } +#endif + } + + if (is_successful) { +#ifdef SUPPORT_PCRE8 + if (mark8_1 != mark8_2) { + printf("8 bit: Mark value mismatch: [%d] '%s' @ '%s'\n", + total, current->pattern, current->input); + is_successful = 0; + } +#endif +#ifdef SUPPORT_PCRE16 + if (mark16_1 != mark16_2) { + printf("16 bit: Mark value mismatch: [%d] '%s' @ '%s'\n", + total, current->pattern, current->input); + is_successful = 0; + } +#endif + } + +#ifdef SUPPORT_PCRE8 + if (re8) { + pcre_free_study(extra8); + pcre_free(re8); + } +#endif +#ifdef SUPPORT_PCRE16 + if (re16) { + pcre16_free_study(extra16); + pcre16_free(re16); + } +#endif + + if (is_successful) { + successful++; + successful_row++; + printf("."); + if (successful_row >= 60) { + successful_row = 0; + printf("\n"); + } + } else + successful_row = 0; + + fflush(stdout); + current++; + } + tables(1); +#ifdef SUPPORT_PCRE8 + setstack8(NULL); +#endif +#ifdef SUPPORT_PCRE16 + setstack16(NULL); +#endif + + if (total == successful) { + printf("\nAll JIT regression tests are successfully passed.\n"); + return 0; + } else { + printf("\nSuccessful test ratio: %d%% (%d failed)\n", successful * 100 / total, total - successful); + return 1; + } +} + +/* End of pcre_jit_test.c */ diff --git a/pcre/pcre_maketables.c b/pcre/pcre_maketables.c --- a/pcre/pcre_maketables.c +++ b/pcre/pcre_maketables.c @@ -6,7 +6,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -59,21 +59,26 @@ /* This function builds a set of character tables for use by PCRE and returns a pointer to them. They are build using the ctype functions, and consequently their contents will depend upon the current locale setting. When compiled as -part of the library, the store is obtained via pcre_malloc(), but when compiled -inside dftables, use malloc(). +part of the library, the store is obtained via PUBL(malloc)(), but when +compiled inside dftables, use malloc(). Arguments: none Returns: pointer to the contiguous block of data */ +#ifdef COMPILE_PCRE8 const unsigned char * pcre_maketables(void) +#else +const unsigned char * +pcre16_maketables(void) +#endif { unsigned char *yield, *p; int i; #ifndef DFTABLES -yield = (unsigned char*)(pcre_malloc)(tables_length); +yield = (unsigned char*)(PUBL(malloc))(tables_length); #else yield = (unsigned char*)malloc(tables_length); #endif diff --git a/pcre/pcre_newline.c b/pcre/pcre_newline.c --- a/pcre/pcre_newline.c +++ b/pcre/pcre_newline.c @@ -6,7 +6,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -67,17 +67,25 @@ type the newline type endptr pointer to the end of the string lenptr where to return the length - utf8 TRUE if in utf8 mode + utf TRUE if in utf mode Returns: TRUE or FALSE */ BOOL -_pcre_is_newline(const uschar *ptr, int type, const uschar *endptr, - int *lenptr, BOOL utf8) +PRIV(is_newline)(PCRE_PUCHAR ptr, int type, PCRE_PUCHAR endptr, int *lenptr, + BOOL utf) { int c; -if (utf8) { GETCHAR(c, ptr); } else c = *ptr; +(void)utf; +#ifdef SUPPORT_UTF +if (utf) + { + GETCHAR(c, ptr); + } +else +#endif /* SUPPORT_UTF */ + c = *ptr; if (type == NLTYPE_ANYCRLF) switch(c) { @@ -96,9 +104,15 @@ case 0x000c: *lenptr = 1; return TRUE; /* FF */ case 0x000d: *lenptr = (ptr < endptr - 1 && ptr[1] == 0x0a)? 2 : 1; return TRUE; /* CR */ - case 0x0085: *lenptr = utf8? 2 : 1; return TRUE; /* NEL */ +#ifdef COMPILE_PCRE8 + case 0x0085: *lenptr = utf? 2 : 1; return TRUE; /* NEL */ case 0x2028: /* LS */ case 0x2029: *lenptr = 3; return TRUE; /* PS */ +#else + case 0x0085: /* NEL */ + case 0x2028: /* LS */ + case 0x2029: *lenptr = 1; return TRUE; /* PS */ +#endif /* COMPILE_PCRE8 */ default: return FALSE; } } @@ -117,27 +131,27 @@ type the newline type startptr pointer to the start of the string lenptr where to return the length - utf8 TRUE if in utf8 mode + utf TRUE if in utf mode Returns: TRUE or FALSE */ BOOL -_pcre_was_newline(const uschar *ptr, int type, const uschar *startptr, - int *lenptr, BOOL utf8) +PRIV(was_newline)(PCRE_PUCHAR ptr, int type, PCRE_PUCHAR startptr, int *lenptr, + BOOL utf) { int c; +(void)utf; ptr--; -#ifdef SUPPORT_UTF8 -if (utf8) +#ifdef SUPPORT_UTF +if (utf) { BACKCHAR(ptr); GETCHAR(c, ptr); } -else c = *ptr; -#else /* no UTF-8 support */ -c = *ptr; -#endif /* SUPPORT_UTF8 */ +else +#endif /* SUPPORT_UTF */ + c = *ptr; if (type == NLTYPE_ANYCRLF) switch(c) { @@ -154,9 +168,15 @@ case 0x000b: /* VT */ case 0x000c: /* FF */ case 0x000d: *lenptr = 1; return TRUE; /* CR */ - case 0x0085: *lenptr = utf8? 2 : 1; return TRUE; /* NEL */ +#ifdef COMPILE_PCRE8 + case 0x0085: *lenptr = utf? 2 : 1; return TRUE; /* NEL */ case 0x2028: /* LS */ case 0x2029: *lenptr = 3; return TRUE; /* PS */ +#else + case 0x0085: /* NEL */ + case 0x2028: /* LS */ + case 0x2029: *lenptr = 1; return TRUE; /* PS */ +#endif /* COMPILE_PCRE8 */ default: return FALSE; } } diff --git a/pcre/pcre_ord2utf8.c b/pcre/pcre_ord2utf8.c --- a/pcre/pcre_ord2utf8.c +++ b/pcre/pcre_ord2utf8.c @@ -6,7 +6,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -52,34 +52,46 @@ * Convert character value to UTF-8 * *************************************************/ -/* This function takes an integer value in the range 0 - 0x7fffffff -and encodes it as a UTF-8 character in 0 to 6 bytes. +/* This function takes an integer value in the range 0 - 0x10ffff +and encodes it as a UTF-8 character in 1 to 6 pcre_uchars. Arguments: cvalue the character value - buffer pointer to buffer for result - at least 6 bytes long + buffer pointer to buffer for result - at least 6 pcre_uchars long Returns: number of characters placed in the buffer */ int -_pcre_ord2utf8(int cvalue, uschar *buffer) +PRIV(ord2utf)(pcre_uint32 cvalue, pcre_uchar *buffer) { -#ifdef SUPPORT_UTF8 +#ifdef SUPPORT_UTF + register int i, j; -for (i = 0; i < _pcre_utf8_table1_size; i++) - if (cvalue <= _pcre_utf8_table1[i]) break; + +/* Checking invalid cvalue character, encoded as invalid UTF-16 character. +Should never happen in practice. */ +if ((cvalue & 0xf800) == 0xd800 || cvalue >= 0x110000) + cvalue = 0xfffe; + +for (i = 0; i < PRIV(utf8_table1_size); i++) + if ((int)cvalue <= PRIV(utf8_table1)[i]) break; buffer += i; for (j = i; j > 0; j--) { *buffer-- = 0x80 | (cvalue & 0x3f); cvalue >>= 6; } -*buffer = _pcre_utf8_table2[i] | cvalue; +*buffer = PRIV(utf8_table2)[i] | cvalue; return i + 1; + #else -return 0; /* Keep compiler happy; this function won't ever be */ -#endif /* called when SUPPORT_UTF8 is not defined. */ + +(void)(cvalue); /* Keep compiler happy; this function won't ever be */ +(void)(buffer); /* called when SUPPORT_UTF is not defined. */ +return 0; + +#endif } /* End of pcre_ord2utf8.c */ diff --git a/pcre/pcre_printint.c b/pcre/pcre_printint.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_printint.c @@ -0,0 +1,710 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Written by Philip Hazel + Copyright (c) 1997-2012 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + + +/* This module contains a PCRE private debugging function for printing out the +internal form of a compiled regular expression, along with some supporting +local functions. This source file is used in two places: + +(1) It is #included by pcre_compile.c when it is compiled in debugging mode +(PCRE_DEBUG defined in pcre_internal.h). It is not included in production +compiles. In this case PCRE_INCLUDED is defined. + +(2) It is also compiled separately and linked with pcretest.c, which can be +asked to print out a compiled regex for debugging purposes. */ + +#ifndef PCRE_INCLUDED + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +/* For pcretest program. */ +#define PRIV(name) name + +/* We have to include pcre_internal.h because we need the internal info for +displaying the results of pcre_study() and we also need to know about the +internal macros, structures, and other internal data values; pcretest has +"inside information" compared to a program that strictly follows the PCRE API. + +Although pcre_internal.h does itself include pcre.h, we explicitly include it +here before pcre_internal.h so that the PCRE_EXP_xxx macros get set +appropriately for an application, not for building PCRE. */ + +#include "pcre.h" +#include "pcre_internal.h" + +/* These are the funtions that are contained within. It doesn't seem worth +having a separate .h file just for this. */ + +#endif /* PCRE_INCLUDED */ + +#ifdef PCRE_INCLUDED +static /* Keep the following function as private. */ +#endif +#ifdef COMPILE_PCRE8 +void pcre_printint(pcre *external_re, FILE *f, BOOL print_lengths); +#else +void pcre16_printint(pcre *external_re, FILE *f, BOOL print_lengths); +#endif + +/* Macro that decides whether a character should be output as a literal or in +hexadecimal. We don't use isprint() because that can vary from system to system +(even without the use of locales) and we want the output always to be the same, +for testing purposes. */ + +#ifdef EBCDIC +#define PRINTABLE(c) ((c) >= 64 && (c) < 255) +#else +#define PRINTABLE(c) ((c) >= 32 && (c) < 127) +#endif + +/* The table of operator names. */ + +static const char *priv_OP_names[] = { OP_NAME_LIST }; + +/* This table of operator lengths is not actually used by the working code, +but its size is needed for a check that ensures it is the correct size for the +number of opcodes (thus catching update omissions). */ + +static const pcre_uint8 priv_OP_lengths[] = { OP_LENGTHS }; + + + +/************************************************* +* Print single- or multi-byte character * +*************************************************/ + +static int +print_char(FILE *f, pcre_uchar *ptr, BOOL utf) +{ +int c = *ptr; + +#ifndef SUPPORT_UTF + +(void)utf; /* Avoid compiler warning */ +if (PRINTABLE(c)) fprintf(f, "%c", c); +else if (c <= 0xff) fprintf(f, "\\x%02x", c); +else fprintf(f, "\\x{%x}", c); +return 0; + +#else + +#ifdef COMPILE_PCRE8 + +if (!utf || (c & 0xc0) != 0xc0) + { + if (PRINTABLE(c)) fprintf(f, "%c", c); else fprintf(f, "\\x%02x", c); + return 0; + } +else + { + int i; + int a = PRIV(utf8_table4)[c & 0x3f]; /* Number of additional bytes */ + int s = 6*a; + c = (c & PRIV(utf8_table3)[a]) << s; + for (i = 1; i <= a; i++) + { + /* This is a check for malformed UTF-8; it should only occur if the sanity + check has been turned off. Rather than swallow random bytes, just stop if + we hit a bad one. Print it with \X instead of \x as an indication. */ + + if ((ptr[i] & 0xc0) != 0x80) + { + fprintf(f, "\\X{%x}", c); + return i - 1; + } + + /* The byte is OK */ + + s -= 6; + c |= (ptr[i] & 0x3f) << s; + } + fprintf(f, "\\x{%x}", c); + return a; + } + +#else + +#ifdef COMPILE_PCRE16 + +if (!utf || (c & 0xfc00) != 0xd800) + { + if (PRINTABLE(c)) fprintf(f, "%c", c); + else if (c <= 0xff) fprintf(f, "\\x%02x", c); + else fprintf(f, "\\x{%x}", c); + return 0; + } +else + { + /* This is a check for malformed UTF-16; it should only occur if the sanity + check has been turned off. Rather than swallow a low surrogate, just stop if + we hit a bad one. Print it with \X instead of \x as an indication. */ + + if ((ptr[1] & 0xfc00) != 0xdc00) + { + fprintf(f, "\\X{%x}", c); + return 0; + } + + c = (((c & 0x3ff) << 10) | (ptr[1] & 0x3ff)) + 0x10000; + fprintf(f, "\\x{%x}", c); + return 1; + } + +#endif /* COMPILE_PCRE16 */ + +#endif /* COMPILE_PCRE8 */ + +#endif /* SUPPORT_UTF */ +} + +/************************************************* +* Print uchar string (regardless of utf) * +*************************************************/ + +static void +print_puchar(FILE *f, PCRE_PUCHAR ptr) +{ +while (*ptr != '\0') + { + register int c = *ptr++; + if (PRINTABLE(c)) fprintf(f, "%c", c); else fprintf(f, "\\x{%x}", c); + } +} + +/************************************************* +* Find Unicode property name * +*************************************************/ + +static const char * +get_ucpname(int ptype, int pvalue) +{ +#ifdef SUPPORT_UCP +int i; +for (i = PRIV(utt_size) - 1; i >= 0; i--) + { + if (ptype == PRIV(utt)[i].type && pvalue == PRIV(utt)[i].value) break; + } +return (i >= 0)? PRIV(utt_names) + PRIV(utt)[i].name_offset : "??"; +#else +/* It gets harder and harder to shut off unwanted compiler warnings. */ +ptype = ptype * pvalue; +return (ptype == pvalue)? "??" : "??"; +#endif +} + + + +/************************************************* +* Print compiled regex * +*************************************************/ + +/* Make this function work for a regex with integers either byte order. +However, we assume that what we are passed is a compiled regex. The +print_lengths flag controls whether offsets and lengths of items are printed. +They can be turned off from pcretest so that automatic tests on bytecode can be +written that do not depend on the value of LINK_SIZE. */ + +#ifdef PCRE_INCLUDED +static /* Keep the following function as private. */ +#endif +#ifdef COMPILE_PCRE8 +void +pcre_printint(pcre *external_re, FILE *f, BOOL print_lengths) +#else +void +pcre16_printint(pcre *external_re, FILE *f, BOOL print_lengths) +#endif +{ +REAL_PCRE *re = (REAL_PCRE *)external_re; +pcre_uchar *codestart, *code; +BOOL utf; + +unsigned int options = re->options; +int offset = re->name_table_offset; +int count = re->name_count; +int size = re->name_entry_size; + +if (re->magic_number != MAGIC_NUMBER) + { + offset = ((offset << 8) & 0xff00) | ((offset >> 8) & 0xff); + count = ((count << 8) & 0xff00) | ((count >> 8) & 0xff); + size = ((size << 8) & 0xff00) | ((size >> 8) & 0xff); + options = ((options << 24) & 0xff000000) | + ((options << 8) & 0x00ff0000) | + ((options >> 8) & 0x0000ff00) | + ((options >> 24) & 0x000000ff); + } + +code = codestart = (pcre_uchar *)re + offset + count * size; +/* PCRE_UTF16 has the same value as PCRE_UTF8. */ +utf = (options & PCRE_UTF8) != 0; + +for(;;) + { + pcre_uchar *ccode; + const char *flag = " "; + int c; + int extra = 0; + + if (print_lengths) + fprintf(f, "%3d ", (int)(code - codestart)); + else + fprintf(f, " "); + + switch(*code) + { +/* ========================================================================== */ + /* These cases are never obeyed. This is a fudge that causes a compile- + time error if the vectors OP_names or OP_lengths, which are indexed + by opcode, are not the correct length. It seems to be the only way to do + such a check at compile time, as the sizeof() operator does not work in + the C preprocessor. */ + + case OP_TABLE_LENGTH: + case OP_TABLE_LENGTH + + ((sizeof(priv_OP_names)/sizeof(const char *) == OP_TABLE_LENGTH) && + (sizeof(priv_OP_lengths) == OP_TABLE_LENGTH)): + break; +/* ========================================================================== */ + + case OP_END: + fprintf(f, " %s\n", priv_OP_names[*code]); + fprintf(f, "------------------------------------------------------------------\n"); + return; + + case OP_CHAR: + fprintf(f, " "); + do + { + code++; + code += 1 + print_char(f, code, utf); + } + while (*code == OP_CHAR); + fprintf(f, "\n"); + continue; + + case OP_CHARI: + fprintf(f, " /i "); + do + { + code++; + code += 1 + print_char(f, code, utf); + } + while (*code == OP_CHARI); + fprintf(f, "\n"); + continue; + + case OP_CBRA: + case OP_CBRAPOS: + case OP_SCBRA: + case OP_SCBRAPOS: + if (print_lengths) fprintf(f, "%3d ", GET(code, 1)); + else fprintf(f, " "); + fprintf(f, "%s %d", priv_OP_names[*code], GET2(code, 1+LINK_SIZE)); + break; + + case OP_BRA: + case OP_BRAPOS: + case OP_SBRA: + case OP_SBRAPOS: + case OP_KETRMAX: + case OP_KETRMIN: + case OP_KETRPOS: + case OP_ALT: + case OP_KET: + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + case OP_ONCE: + case OP_ONCE_NC: + case OP_COND: + case OP_SCOND: + case OP_REVERSE: + if (print_lengths) fprintf(f, "%3d ", GET(code, 1)); + else fprintf(f, " "); + fprintf(f, "%s", priv_OP_names[*code]); + break; + + case OP_CLOSE: + fprintf(f, " %s %d", priv_OP_names[*code], GET2(code, 1)); + break; + + case OP_CREF: + case OP_NCREF: + fprintf(f, "%3d %s", GET2(code,1), priv_OP_names[*code]); + break; + + case OP_RREF: + c = GET2(code, 1); + if (c == RREF_ANY) + fprintf(f, " Cond recurse any"); + else + fprintf(f, " Cond recurse %d", c); + break; + + case OP_NRREF: + c = GET2(code, 1); + if (c == RREF_ANY) + fprintf(f, " Cond nrecurse any"); + else + fprintf(f, " Cond nrecurse %d", c); + break; + + case OP_DEF: + fprintf(f, " Cond def"); + break; + + case OP_STARI: + case OP_MINSTARI: + case OP_POSSTARI: + case OP_PLUSI: + case OP_MINPLUSI: + case OP_POSPLUSI: + case OP_QUERYI: + case OP_MINQUERYI: + case OP_POSQUERYI: + flag = "/i"; + /* Fall through */ + case OP_STAR: + case OP_MINSTAR: + case OP_POSSTAR: + case OP_PLUS: + case OP_MINPLUS: + case OP_POSPLUS: + case OP_QUERY: + case OP_MINQUERY: + case OP_POSQUERY: + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPOSSTAR: + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEPOSPLUS: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSQUERY: + fprintf(f, " %s ", flag); + if (*code >= OP_TYPESTAR) + { + fprintf(f, "%s", priv_OP_names[code[1]]); + if (code[1] == OP_PROP || code[1] == OP_NOTPROP) + { + fprintf(f, " %s ", get_ucpname(code[2], code[3])); + extra = 2; + } + } + else extra = print_char(f, code+1, utf); + fprintf(f, "%s", priv_OP_names[*code]); + break; + + case OP_EXACTI: + case OP_UPTOI: + case OP_MINUPTOI: + case OP_POSUPTOI: + flag = "/i"; + /* Fall through */ + case OP_EXACT: + case OP_UPTO: + case OP_MINUPTO: + case OP_POSUPTO: + fprintf(f, " %s ", flag); + extra = print_char(f, code + 1 + IMM2_SIZE, utf); + fprintf(f, "{"); + if (*code != OP_EXACT && *code != OP_EXACTI) fprintf(f, "0,"); + fprintf(f, "%d}", GET2(code,1)); + if (*code == OP_MINUPTO || *code == OP_MINUPTOI) fprintf(f, "?"); + else if (*code == OP_POSUPTO || *code == OP_POSUPTOI) fprintf(f, "+"); + break; + + case OP_TYPEEXACT: + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEPOSUPTO: + fprintf(f, " %s", priv_OP_names[code[1 + IMM2_SIZE]]); + if (code[1 + IMM2_SIZE] == OP_PROP || code[1 + IMM2_SIZE] == OP_NOTPROP) + { + fprintf(f, " %s ", get_ucpname(code[1 + IMM2_SIZE + 1], + code[1 + IMM2_SIZE + 2])); + extra = 2; + } + fprintf(f, "{"); + if (*code != OP_TYPEEXACT) fprintf(f, "0,"); + fprintf(f, "%d}", GET2(code,1)); + if (*code == OP_TYPEMINUPTO) fprintf(f, "?"); + else if (*code == OP_TYPEPOSUPTO) fprintf(f, "+"); + break; + + case OP_NOTI: + flag = "/i"; + /* Fall through */ + case OP_NOT: + fprintf(f, " %s [^", flag); + extra = print_char(f, code + 1, utf); + fprintf(f, "]"); + break; + + case OP_NOTSTARI: + case OP_NOTMINSTARI: + case OP_NOTPOSSTARI: + case OP_NOTPLUSI: + case OP_NOTMINPLUSI: + case OP_NOTPOSPLUSI: + case OP_NOTQUERYI: + case OP_NOTMINQUERYI: + case OP_NOTPOSQUERYI: + flag = "/i"; + /* Fall through */ + + case OP_NOTSTAR: + case OP_NOTMINSTAR: + case OP_NOTPOSSTAR: + case OP_NOTPLUS: + case OP_NOTMINPLUS: + case OP_NOTPOSPLUS: + case OP_NOTQUERY: + case OP_NOTMINQUERY: + case OP_NOTPOSQUERY: + fprintf(f, " %s [^", flag); + extra = print_char(f, code + 1, utf); + fprintf(f, "]%s", priv_OP_names[*code]); + break; + + case OP_NOTEXACTI: + case OP_NOTUPTOI: + case OP_NOTMINUPTOI: + case OP_NOTPOSUPTOI: + flag = "/i"; + /* Fall through */ + + case OP_NOTEXACT: + case OP_NOTUPTO: + case OP_NOTMINUPTO: + case OP_NOTPOSUPTO: + fprintf(f, " %s [^", flag); + extra = print_char(f, code + 1 + IMM2_SIZE, utf); + fprintf(f, "]{"); + if (*code != OP_NOTEXACT && *code != OP_NOTEXACTI) fprintf(f, "0,"); + fprintf(f, "%d}", GET2(code,1)); + if (*code == OP_NOTMINUPTO || *code == OP_NOTMINUPTOI) fprintf(f, "?"); + else + if (*code == OP_NOTPOSUPTO || *code == OP_NOTPOSUPTOI) fprintf(f, "+"); + break; + + case OP_RECURSE: + if (print_lengths) fprintf(f, "%3d ", GET(code, 1)); + else fprintf(f, " "); + fprintf(f, "%s", priv_OP_names[*code]); + break; + + case OP_REFI: + flag = "/i"; + /* Fall through */ + case OP_REF: + fprintf(f, " %s \\%d", flag, GET2(code,1)); + ccode = code + priv_OP_lengths[*code]; + goto CLASS_REF_REPEAT; + + case OP_CALLOUT: + fprintf(f, " %s %d %d %d", priv_OP_names[*code], code[1], GET(code,2), + GET(code, 2 + LINK_SIZE)); + break; + + case OP_PROP: + case OP_NOTPROP: + fprintf(f, " %s %s", priv_OP_names[*code], get_ucpname(code[1], code[2])); + break; + + /* OP_XCLASS can only occur in UTF or PCRE16 modes. However, there's no + harm in having this code always here, and it makes it less messy without + all those #ifdefs. */ + + case OP_CLASS: + case OP_NCLASS: + case OP_XCLASS: + { + int i, min, max; + BOOL printmap; + pcre_uint8 *map; + + fprintf(f, " ["); + + if (*code == OP_XCLASS) + { + extra = GET(code, 1); + ccode = code + LINK_SIZE + 1; + printmap = (*ccode & XCL_MAP) != 0; + if ((*ccode++ & XCL_NOT) != 0) fprintf(f, "^"); + } + else + { + printmap = TRUE; + ccode = code + 1; + } + + /* Print a bit map */ + + if (printmap) + { + map = (pcre_uint8 *)ccode; + for (i = 0; i < 256; i++) + { + if ((map[i/8] & (1 << (i&7))) != 0) + { + int j; + for (j = i+1; j < 256; j++) + if ((map[j/8] & (1 << (j&7))) == 0) break; + if (i == '-' || i == ']') fprintf(f, "\\"); + if (PRINTABLE(i)) fprintf(f, "%c", i); + else fprintf(f, "\\x%02x", i); + if (--j > i) + { + if (j != i + 1) fprintf(f, "-"); + if (j == '-' || j == ']') fprintf(f, "\\"); + if (PRINTABLE(j)) fprintf(f, "%c", j); + else fprintf(f, "\\x%02x", j); + } + i = j; + } + } + ccode += 32 / sizeof(pcre_uchar); + } + + /* For an XCLASS there is always some additional data */ + + if (*code == OP_XCLASS) + { + int ch; + while ((ch = *ccode++) != XCL_END) + { + if (ch == XCL_PROP) + { + int ptype = *ccode++; + int pvalue = *ccode++; + fprintf(f, "\\p{%s}", get_ucpname(ptype, pvalue)); + } + else if (ch == XCL_NOTPROP) + { + int ptype = *ccode++; + int pvalue = *ccode++; + fprintf(f, "\\P{%s}", get_ucpname(ptype, pvalue)); + } + else + { + ccode += 1 + print_char(f, ccode, utf); + if (ch == XCL_RANGE) + { + fprintf(f, "-"); + ccode += 1 + print_char(f, ccode, utf); + } + } + } + } + + /* Indicate a non-UTF class which was created by negation */ + + fprintf(f, "]%s", (*code == OP_NCLASS)? " (neg)" : ""); + + /* Handle repeats after a class or a back reference */ + + CLASS_REF_REPEAT: + switch(*ccode) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRPLUS: + case OP_CRMINPLUS: + case OP_CRQUERY: + case OP_CRMINQUERY: + fprintf(f, "%s", priv_OP_names[*ccode]); + extra += priv_OP_lengths[*ccode]; + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + min = GET2(ccode,1); + max = GET2(ccode,1 + IMM2_SIZE); + if (max == 0) fprintf(f, "{%d,}", min); + else fprintf(f, "{%d,%d}", min, max); + if (*ccode == OP_CRMINRANGE) fprintf(f, "?"); + extra += priv_OP_lengths[*ccode]; + break; + + /* Do nothing if it's not a repeat; this code stops picky compilers + warning about the lack of a default code path. */ + + default: + break; + } + } + break; + + case OP_MARK: + case OP_PRUNE_ARG: + case OP_SKIP_ARG: + case OP_THEN_ARG: + fprintf(f, " %s ", priv_OP_names[*code]); + print_puchar(f, code + 2); + extra += code[1]; + break; + + case OP_THEN: + fprintf(f, " %s", priv_OP_names[*code]); + break; + + case OP_CIRCM: + case OP_DOLLM: + flag = "/m"; + /* Fall through */ + + /* Anything else is just an item with no data, but possibly a flag. */ + + default: + fprintf(f, " %s %s", flag, priv_OP_names[*code]); + break; + } + + code += priv_OP_lengths[*code] + extra; + fprintf(f, "\n"); + } +} + +/* End of pcre_printint.src */ diff --git a/pcre/pcre_refcount.c b/pcre/pcre_refcount.c --- a/pcre/pcre_refcount.c +++ b/pcre/pcre_refcount.c @@ -6,7 +6,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -68,11 +68,18 @@ a negative error number */ -PCRE_EXP_DEFN int +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION pcre_refcount(pcre *argument_re, int adjust) +#else +PCRE_EXP_DEFN int PCRE_CALL_CONVENTION +pcre16_refcount(pcre16 *argument_re, int adjust) +#endif { -real_pcre *re = (real_pcre *)argument_re; +REAL_PCRE *re = (REAL_PCRE *)argument_re; if (re == NULL) return PCRE_ERROR_NULL; +if (re->magic_number != MAGIC_NUMBER) return PCRE_ERROR_BADMAGIC; +if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE; re->ref_count = (-adjust > re->ref_count)? 0 : (adjust + re->ref_count > 65535)? 65535 : re->ref_count + adjust; diff --git a/pcre/pcre_string_utils.c b/pcre/pcre_string_utils.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_string_utils.c @@ -0,0 +1,168 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* PCRE is a library of functions to support regular expressions whose syntax +and semantics are as close as possible to those of the Perl 5 language. + + Written by Philip Hazel + Copyright (c) 1997-2012 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + + +/* This module contains an internal function that is used to match an extended +class. It is used by both pcre_exec() and pcre_def_exec(). */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + +#ifndef COMPILE_PCRE8 + +/************************************************* +* Compare string utilities * +*************************************************/ + +/* The following two functions compares two strings. Basically an strcmp +for non 8 bit characters. + +Arguments: + str1 first string + str2 second string + +Returns: 0 if both string are equal (like strcmp), 1 otherwise +*/ + +int +PRIV(strcmp_uc_uc)(const pcre_uchar *str1, const pcre_uchar *str2) +{ +pcre_uchar c1; +pcre_uchar c2; + +while (*str1 != '\0' || *str2 != '\0') + { + c1 = *str1++; + c2 = *str2++; + if (c1 != c2) + return ((c1 > c2) << 1) - 1; + } +/* Both length and characters must be equal. */ +return 0; +} + +int +PRIV(strcmp_uc_c8)(const pcre_uchar *str1, const char *str2) +{ +const pcre_uint8 *ustr2 = (pcre_uint8 *)str2; +pcre_uchar c1; +pcre_uchar c2; + +while (*str1 != '\0' || *ustr2 != '\0') + { + c1 = *str1++; + c2 = (pcre_uchar)*ustr2++; + if (c1 != c2) + return ((c1 > c2) << 1) - 1; + } +/* Both length and characters must be equal. */ +return 0; +} + +/* The following two functions compares two, fixed length +strings. Basically an strncmp for non 8 bit characters. + +Arguments: + str1 first string + str2 second string + num size of the string + +Returns: 0 if both string are equal (like strcmp), 1 otherwise +*/ + +int +PRIV(strncmp_uc_uc)(const pcre_uchar *str1, const pcre_uchar *str2, unsigned int num) +{ +pcre_uchar c1; +pcre_uchar c2; + +while (num-- > 0) + { + c1 = *str1++; + c2 = *str2++; + if (c1 != c2) + return ((c1 > c2) << 1) - 1; + } +/* Both length and characters must be equal. */ +return 0; +} + +int +PRIV(strncmp_uc_c8)(const pcre_uchar *str1, const char *str2, unsigned int num) +{ +const pcre_uint8 *ustr2 = (pcre_uint8 *)str2; +pcre_uchar c1; +pcre_uchar c2; + +while (num-- > 0) + { + c1 = *str1++; + c2 = (pcre_uchar)*ustr2++; + if (c1 != c2) + return ((c1 > c2) << 1) - 1; + } +/* Both length and characters must be equal. */ +return 0; +} + +/* The following function returns with the length of +a zero terminated string. Basically an strlen for non 8 bit characters. + +Arguments: + str string + +Returns: length of the string +*/ + +unsigned int +PRIV(strlen_uc)(const pcre_uchar *str) +{ +unsigned int len = 0; +while (*str++ != 0) + len++; +return len; +} + +#endif /* COMPILE_PCRE8 */ + +/* End of pcre_string_utils.c */ diff --git a/pcre/pcre_stringpiece.h b/pcre/pcre_stringpiece.h deleted file mode 100644 --- a/pcre/pcre_stringpiece.h +++ /dev/null @@ -1,177 +0,0 @@ -// Copyright (c) 2005, Google Inc. -// All rights reserved. -// -// Redistribution and use in source and binary forms, with or without -// modification, are permitted provided that the following conditions are -// met: -// -// * Redistributions of source code must retain the above copyright -// notice, this list of conditions and the following disclaimer. -// * Redistributions in binary form must reproduce the above -// copyright notice, this list of conditions and the following disclaimer -// in the documentation and/or other materials provided with the -// distribution. -// * Neither the name of Google Inc. nor the names of its -// contributors may be used to endorse or promote products derived from -// this software without specific prior written permission. -// -// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS -// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR -// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT -// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, -// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT -// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, -// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY -// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT -// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE -// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. -// -// Author: Sanjay Ghemawat -// -// A string like object that points into another piece of memory. -// Useful for providing an interface that allows clients to easily -// pass in either a "const char*" or a "string". -// -// Arghh! I wish C++ literals were automatically of type "string". - -#ifndef _PCRE_STRINGPIECE_H -#define _PCRE_STRINGPIECE_H - -#include -#include -#include // for ostream forward-declaration - -#if 0 -#define HAVE_TYPE_TRAITS -#include -#elif 0 -#define HAVE_TYPE_TRAITS -#include -#endif - -#include - -using std::string; - -namespace pcrecpp { - -class PCRECPP_EXP_DEFN StringPiece { - private: - const char* ptr_; - int length_; - - public: - // We provide non-explicit singleton constructors so users can pass - // in a "const char*" or a "string" wherever a "StringPiece" is - // expected. - StringPiece() - : ptr_(NULL), length_(0) { } - StringPiece(const char* str) - : ptr_(str), length_(static_cast(strlen(ptr_))) { } - StringPiece(const unsigned char* str) - : ptr_(reinterpret_cast(str)), - length_(static_cast(strlen(ptr_))) { } - StringPiece(const string& str) - : ptr_(str.data()), length_(static_cast(str.size())) { } - StringPiece(const char* offset, int len) - : ptr_(offset), length_(len) { } - - // data() may return a pointer to a buffer with embedded NULs, and the - // returned buffer may or may not be null terminated. Therefore it is - // typically a mistake to pass data() to a routine that expects a NUL - // terminated string. Use "as_string().c_str()" if you really need to do - // this. Or better yet, change your routine so it does not rely on NUL - // termination. - const char* data() const { return ptr_; } - int size() const { return length_; } - bool empty() const { return length_ == 0; } - - void clear() { ptr_ = NULL; length_ = 0; } - void set(const char* buffer, int len) { ptr_ = buffer; length_ = len; } - void set(const char* str) { - ptr_ = str; - length_ = static_cast(strlen(str)); - } - void set(const void* buffer, int len) { - ptr_ = reinterpret_cast(buffer); - length_ = len; - } - - char operator[](int i) const { return ptr_[i]; } - - void remove_prefix(int n) { - ptr_ += n; - length_ -= n; - } - - void remove_suffix(int n) { - length_ -= n; - } - - bool operator==(const StringPiece& x) const { - return ((length_ == x.length_) && - (memcmp(ptr_, x.ptr_, length_) == 0)); - } - bool operator!=(const StringPiece& x) const { - return !(*this == x); - } - -#define STRINGPIECE_BINARY_PREDICATE(cmp,auxcmp) \ - bool operator cmp (const StringPiece& x) const { \ - int r = memcmp(ptr_, x.ptr_, length_ < x.length_ ? length_ : x.length_); \ - return ((r auxcmp 0) || ((r == 0) && (length_ cmp x.length_))); \ - } - STRINGPIECE_BINARY_PREDICATE(<, <); - STRINGPIECE_BINARY_PREDICATE(<=, <); - STRINGPIECE_BINARY_PREDICATE(>=, >); - STRINGPIECE_BINARY_PREDICATE(>, >); -#undef STRINGPIECE_BINARY_PREDICATE - - int compare(const StringPiece& x) const { - int r = memcmp(ptr_, x.ptr_, length_ < x.length_ ? length_ : x.length_); - if (r == 0) { - if (length_ < x.length_) r = -1; - else if (length_ > x.length_) r = +1; - } - return r; - } - - string as_string() const { - return string(data(), size()); - } - - void CopyToString(string* target) const { - target->assign(ptr_, length_); - } - - // Does "this" start with "x" - bool starts_with(const StringPiece& x) const { - return ((length_ >= x.length_) && (memcmp(ptr_, x.ptr_, x.length_) == 0)); - } -}; - -} // namespace pcrecpp - -// ------------------------------------------------------------------ -// Functions used to create STL containers that use StringPiece -// Remember that a StringPiece's lifetime had better be less than -// that of the underlying string or char*. If it is not, then you -// cannot safely store a StringPiece into an STL container -// ------------------------------------------------------------------ - -#ifdef HAVE_TYPE_TRAITS -// This makes vector really fast for some STL implementations -template<> struct __type_traits { - typedef __true_type has_trivial_default_constructor; - typedef __true_type has_trivial_copy_constructor; - typedef __true_type has_trivial_assignment_operator; - typedef __true_type has_trivial_destructor; - typedef __true_type is_POD_type; -}; -#endif - -// allow StringPiece to be logged -std::ostream& operator<<(std::ostream& o, const pcrecpp::StringPiece& piece); - -#endif /* _PCRE_STRINGPIECE_H */ diff --git a/pcre/pcre_study.c b/pcre/pcre_study.c --- a/pcre/pcre_study.c +++ b/pcre/pcre_study.c @@ -6,582 +6,1529 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* This module contains the external function pcre_study(), along with local -supporting functions. */ - - -#ifdef HAVE_CONFIG_H -#include "config.h" -#endif - -#include "pcre_internal.h" - - -/* Returns from set_start_bits() */ - -enum { SSB_FAIL, SSB_DONE, SSB_CONTINUE }; - - -/************************************************* -* Set a bit and maybe its alternate case * -*************************************************/ - -/* Given a character, set its bit in the table, and also the bit for the other -version of a letter if we are caseless. - -Arguments: - start_bits points to the bit map - c is the character - caseless the caseless flag - cd the block with char table pointers - -Returns: nothing -*/ - -static void -set_bit(uschar *start_bits, unsigned int c, BOOL caseless, compile_data *cd) -{ -start_bits[c/8] |= (1 << (c&7)); -if (caseless && (cd->ctypes[c] & ctype_letter) != 0) - start_bits[cd->fcc[c]/8] |= (1 << (cd->fcc[c]&7)); -} - - - -/************************************************* -* Create bitmap of starting bytes * -*************************************************/ - -/* This function scans a compiled unanchored expression recursively and -attempts to build a bitmap of the set of possible starting bytes. As time goes -by, we may be able to get more clever at doing this. The SSB_CONTINUE return is -useful for parenthesized groups in patterns such as (a*)b where the group -provides some optional starting bytes but scanning must continue at the outer -level to find at least one mandatory byte. At the outermost level, this -function fails unless the result is SSB_DONE. - -Arguments: - code points to an expression - start_bits points to a 32-byte table, initialized to 0 - caseless the current state of the caseless flag - utf8 TRUE if in UTF-8 mode - cd the block with char table pointers - -Returns: SSB_FAIL => Failed to find any starting bytes - SSB_DONE => Found mandatory starting bytes - SSB_CONTINUE => Found optional starting bytes -*/ - -static int -set_start_bits(const uschar *code, uschar *start_bits, BOOL caseless, - BOOL utf8, compile_data *cd) -{ -register int c; -int yield = SSB_DONE; - -#if 0 -/* ========================================================================= */ -/* The following comment and code was inserted in January 1999. In May 2006, -when it was observed to cause compiler warnings about unused values, I took it -out again. If anybody is still using OS/2, they will have to put it back -manually. */ - -/* This next statement and the later reference to dummy are here in order to -trick the optimizer of the IBM C compiler for OS/2 into generating correct -code. Apparently IBM isn't going to fix the problem, and we would rather not -disable optimization (in this module it actually makes a big difference, and -the pcre module can use all the optimization it can get). */ - -volatile int dummy; -/* ========================================================================= */ -#endif - -do - { - const uschar *tcode = code + (((int)*code == OP_CBRA)? 3:1) + LINK_SIZE; - BOOL try_next = TRUE; - - while (try_next) /* Loop for items in this branch */ - { - int rc; - switch(*tcode) - { - /* Fail if we reach something we don't understand */ - - default: - return SSB_FAIL; - - /* If we hit a bracket or a positive lookahead assertion, recurse to set - bits from within the subpattern. If it can't find anything, we have to - give up. If it finds some mandatory character(s), we are done for this - branch. Otherwise, carry on scanning after the subpattern. */ - - case OP_BRA: - case OP_SBRA: - case OP_CBRA: - case OP_SCBRA: - case OP_ONCE: - case OP_ASSERT: - rc = set_start_bits(tcode, start_bits, caseless, utf8, cd); - if (rc == SSB_FAIL) return SSB_FAIL; - if (rc == SSB_DONE) try_next = FALSE; else - { - do tcode += GET(tcode, 1); while (*tcode == OP_ALT); - tcode += 1 + LINK_SIZE; - } - break; - - /* If we hit ALT or KET, it means we haven't found anything mandatory in - this branch, though we might have found something optional. For ALT, we - continue with the next alternative, but we have to arrange that the final - result from subpattern is SSB_CONTINUE rather than SSB_DONE. For KET, - return SSB_CONTINUE: if this is the top level, that indicates failure, - but after a nested subpattern, it causes scanning to continue. */ - - case OP_ALT: - yield = SSB_CONTINUE; - try_next = FALSE; - break; - - case OP_KET: - case OP_KETRMAX: - case OP_KETRMIN: - return SSB_CONTINUE; - - /* Skip over callout */ - - case OP_CALLOUT: - tcode += 2 + 2*LINK_SIZE; - break; - - /* Skip over lookbehind and negative lookahead assertions */ - - case OP_ASSERT_NOT: - case OP_ASSERTBACK: - case OP_ASSERTBACK_NOT: - do tcode += GET(tcode, 1); while (*tcode == OP_ALT); - tcode += 1 + LINK_SIZE; - break; - - /* Skip over an option setting, changing the caseless flag */ - - case OP_OPT: - caseless = (tcode[1] & PCRE_CASELESS) != 0; - tcode += 2; - break; - - /* BRAZERO does the bracket, but carries on. */ - - case OP_BRAZERO: - case OP_BRAMINZERO: - if (set_start_bits(++tcode, start_bits, caseless, utf8, cd) == SSB_FAIL) - return SSB_FAIL; -/* ========================================================================= - See the comment at the head of this function concerning the next line, - which was an old fudge for the benefit of OS/2. - dummy = 1; - ========================================================================= */ - do tcode += GET(tcode,1); while (*tcode == OP_ALT); - tcode += 1 + LINK_SIZE; - break; - - /* SKIPZERO skips the bracket. */ - - case OP_SKIPZERO: - do tcode += GET(tcode,1); while (*tcode == OP_ALT); - tcode += 1 + LINK_SIZE; - break; - - /* Single-char * or ? sets the bit and tries the next item */ - - case OP_STAR: - case OP_MINSTAR: - case OP_POSSTAR: - case OP_QUERY: - case OP_MINQUERY: - case OP_POSQUERY: - set_bit(start_bits, tcode[1], caseless, cd); - tcode += 2; -#ifdef SUPPORT_UTF8 - if (utf8 && tcode[-1] >= 0xc0) - tcode += _pcre_utf8_table4[tcode[-1] & 0x3f]; -#endif - break; - - /* Single-char upto sets the bit and tries the next */ - - case OP_UPTO: - case OP_MINUPTO: - case OP_POSUPTO: - set_bit(start_bits, tcode[3], caseless, cd); - tcode += 4; -#ifdef SUPPORT_UTF8 - if (utf8 && tcode[-1] >= 0xc0) - tcode += _pcre_utf8_table4[tcode[-1] & 0x3f]; -#endif - break; - - /* At least one single char sets the bit and stops */ - - case OP_EXACT: /* Fall through */ - tcode += 2; - - case OP_CHAR: - case OP_CHARNC: - case OP_PLUS: - case OP_MINPLUS: - case OP_POSPLUS: - set_bit(start_bits, tcode[1], caseless, cd); - try_next = FALSE; - break; - - /* Single character type sets the bits and stops */ - - case OP_NOT_DIGIT: - for (c = 0; c < 32; c++) - start_bits[c] |= ~cd->cbits[c+cbit_digit]; - try_next = FALSE; - break; - - case OP_DIGIT: - for (c = 0; c < 32; c++) - start_bits[c] |= cd->cbits[c+cbit_digit]; - try_next = FALSE; - break; - - /* The cbit_space table has vertical tab as whitespace; we have to - discard it. */ - - case OP_NOT_WHITESPACE: - for (c = 0; c < 32; c++) - { - int d = cd->cbits[c+cbit_space]; - if (c == 1) d &= ~0x08; - start_bits[c] |= ~d; - } - try_next = FALSE; - break; - - /* The cbit_space table has vertical tab as whitespace; we have to - discard it. */ - - case OP_WHITESPACE: - for (c = 0; c < 32; c++) - { - int d = cd->cbits[c+cbit_space]; - if (c == 1) d &= ~0x08; - start_bits[c] |= d; - } - try_next = FALSE; - break; - - case OP_NOT_WORDCHAR: - for (c = 0; c < 32; c++) - start_bits[c] |= ~cd->cbits[c+cbit_word]; - try_next = FALSE; - break; - - case OP_WORDCHAR: - for (c = 0; c < 32; c++) - start_bits[c] |= cd->cbits[c+cbit_word]; - try_next = FALSE; - break; - - /* One or more character type fudges the pointer and restarts, knowing - it will hit a single character type and stop there. */ - - case OP_TYPEPLUS: - case OP_TYPEMINPLUS: - tcode++; - break; - - case OP_TYPEEXACT: - tcode += 3; - break; - - /* Zero or more repeats of character types set the bits and then - try again. */ - - case OP_TYPEUPTO: - case OP_TYPEMINUPTO: - case OP_TYPEPOSUPTO: - tcode += 2; /* Fall through */ - - case OP_TYPESTAR: - case OP_TYPEMINSTAR: - case OP_TYPEPOSSTAR: - case OP_TYPEQUERY: - case OP_TYPEMINQUERY: - case OP_TYPEPOSQUERY: - switch(tcode[1]) - { - case OP_ANY: - case OP_ALLANY: - return SSB_FAIL; - - case OP_NOT_DIGIT: - for (c = 0; c < 32; c++) - start_bits[c] |= ~cd->cbits[c+cbit_digit]; - break; - - case OP_DIGIT: - for (c = 0; c < 32; c++) - start_bits[c] |= cd->cbits[c+cbit_digit]; - break; - - /* The cbit_space table has vertical tab as whitespace; we have to - discard it. */ - - case OP_NOT_WHITESPACE: - for (c = 0; c < 32; c++) - { - int d = cd->cbits[c+cbit_space]; - if (c == 1) d &= ~0x08; - start_bits[c] |= ~d; - } - break; - - /* The cbit_space table has vertical tab as whitespace; we have to - discard it. */ - - case OP_WHITESPACE: - for (c = 0; c < 32; c++) - { - int d = cd->cbits[c+cbit_space]; - if (c == 1) d &= ~0x08; - start_bits[c] |= d; - } - break; - - case OP_NOT_WORDCHAR: - for (c = 0; c < 32; c++) - start_bits[c] |= ~cd->cbits[c+cbit_word]; - break; - - case OP_WORDCHAR: - for (c = 0; c < 32; c++) - start_bits[c] |= cd->cbits[c+cbit_word]; - break; - } - - tcode += 2; - break; - - /* Character class where all the information is in a bit map: set the - bits and either carry on or not, according to the repeat count. If it was - a negative class, and we are operating with UTF-8 characters, any byte - with a value >= 0xc4 is a potentially valid starter because it starts a - character with a value > 255. */ - - case OP_NCLASS: -#ifdef SUPPORT_UTF8 - if (utf8) - { - start_bits[24] |= 0xf0; /* Bits for 0xc4 - 0xc8 */ - memset(start_bits+25, 0xff, 7); /* Bits for 0xc9 - 0xff */ - } -#endif - /* Fall through */ - - case OP_CLASS: - { - tcode++; - - /* In UTF-8 mode, the bits in a bit map correspond to character - values, not to byte values. However, the bit map we are constructing is - for byte values. So we have to do a conversion for characters whose - value is > 127. In fact, there are only two possible starting bytes for - characters in the range 128 - 255. */ - -#ifdef SUPPORT_UTF8 - if (utf8) - { - for (c = 0; c < 16; c++) start_bits[c] |= tcode[c]; - for (c = 128; c < 256; c++) - { - if ((tcode[c/8] && (1 << (c&7))) != 0) - { - int d = (c >> 6) | 0xc0; /* Set bit for this starter */ - start_bits[d/8] |= (1 << (d&7)); /* and then skip on to the */ - c = (c & 0xc0) + 0x40 - 1; /* next relevant character. */ - } - } - } - - /* In non-UTF-8 mode, the two bit maps are completely compatible. */ - - else -#endif - { - for (c = 0; c < 32; c++) start_bits[c] |= tcode[c]; - } - - /* Advance past the bit map, and act on what follows */ - - tcode += 32; - switch (*tcode) - { - case OP_CRSTAR: - case OP_CRMINSTAR: - case OP_CRQUERY: - case OP_CRMINQUERY: - tcode++; - break; - - case OP_CRRANGE: - case OP_CRMINRANGE: - if (((tcode[1] << 8) + tcode[2]) == 0) tcode += 5; - else try_next = FALSE; - break; - - default: - try_next = FALSE; - break; - } - } - break; /* End of bitmap class handling */ - - } /* End of switch */ - } /* End of try_next loop */ - - code += GET(code, 1); /* Advance to next branch */ - } -while (*code == OP_ALT); -return yield; -} - - - -/************************************************* -* Study a compiled expression * -*************************************************/ - -/* This function is handed a compiled expression that it must study to produce -information that will speed up the matching. It returns a pcre_extra block -which then gets handed back to pcre_exec(). - -Arguments: - re points to the compiled expression - options contains option bits - errorptr points to where to place error messages; - set NULL unless error - -Returns: pointer to a pcre_extra block, with study_data filled in and the - appropriate flag set; - NULL on error or if no optimization possible -*/ - -PCRE_EXP_DEFN pcre_extra * -pcre_study(const pcre *external_re, int options, const char **errorptr) -{ -uschar start_bits[32]; -pcre_extra *extra; -pcre_study_data *study; -const uschar *tables; -uschar *code; -compile_data compile_block; -const real_pcre *re = (const real_pcre *)external_re; - -*errorptr = NULL; - -if (re == NULL || re->magic_number != MAGIC_NUMBER) - { - *errorptr = "argument is not a compiled regular expression"; - return NULL; - } - -if ((options & ~PUBLIC_STUDY_OPTIONS) != 0) - { - *errorptr = "unknown or incorrect option bit(s) set"; - return NULL; - } - -code = (uschar *)re + re->name_table_offset + - (re->name_count * re->name_entry_size); - -/* For an anchored pattern, or an unanchored pattern that has a first char, or -a multiline pattern that matches only at "line starts", no further processing -at present. */ - -if ((re->options & PCRE_ANCHORED) != 0 || - (re->flags & (PCRE_FIRSTSET|PCRE_STARTLINE)) != 0) - return NULL; - -/* Set the character tables in the block that is passed around */ - -tables = re->tables; -if (tables == NULL) - (void)pcre_fullinfo(external_re, NULL, PCRE_INFO_DEFAULT_TABLES, - (void *)(&tables)); - -compile_block.lcc = tables + lcc_offset; -compile_block.fcc = tables + fcc_offset; -compile_block.cbits = tables + cbits_offset; -compile_block.ctypes = tables + ctypes_offset; - -/* See if we can find a fixed set of initial characters for the pattern. */ - -memset(start_bits, 0, 32 * sizeof(uschar)); -if (set_start_bits(code, start_bits, (re->options & PCRE_CASELESS) != 0, - (re->options & PCRE_UTF8) != 0, &compile_block) != SSB_DONE) return NULL; - -/* Get a pcre_extra block and a pcre_study_data block. The study data is put in -the latter, which is pointed to by the former, which may also get additional -data set later by the calling program. At the moment, the size of -pcre_study_data is fixed. We nevertheless save it in a field for returning via -the pcre_fullinfo() function so that if it becomes variable in the future, we -don't have to change that code. */ - -extra = (pcre_extra *)(pcre_malloc) - (sizeof(pcre_extra) + sizeof(pcre_study_data)); - -if (extra == NULL) - { - *errorptr = "failed to get memory"; - return NULL; - } - -study = (pcre_study_data *)((char *)extra + sizeof(pcre_extra)); -extra->flags = PCRE_EXTRA_STUDY_DATA; -extra->study_data = study; - -study->size = sizeof(pcre_study_data); -study->options = PCRE_STUDY_MAPPED; -memcpy(study->start_bits, start_bits, sizeof(start_bits)); - -return extra; + Copyright (c) 1997-2012 University of Cambridge + +----------------------------------------------------------------------------- +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, + this list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright + notice, this list of conditions and the following disclaimer in the + documentation and/or other materials provided with the distribution. + + * Neither the name of the University of Cambridge nor the names of its + contributors may be used to endorse or promote products derived from + this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE +LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF +SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN +CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) +ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE +POSSIBILITY OF SUCH DAMAGE. +----------------------------------------------------------------------------- +*/ + + +/* This module contains the external function pcre_study(), along with local +supporting functions. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + +#define SET_BIT(c) start_bits[c/8] |= (1 << (c&7)) + +/* Returns from set_start_bits() */ + +enum { SSB_FAIL, SSB_DONE, SSB_CONTINUE, SSB_UNKNOWN }; + + + +/************************************************* +* Find the minimum subject length for a group * +*************************************************/ + +/* Scan a parenthesized group and compute the minimum length of subject that +is needed to match it. This is a lower bound; it does not mean there is a +string of that length that matches. In UTF8 mode, the result is in characters +rather than bytes. + +Arguments: + code pointer to start of group (the bracket) + startcode pointer to start of the whole pattern + options the compiling options + int RECURSE depth + +Returns: the minimum length + -1 if \C in UTF-8 mode or (*ACCEPT) was encountered + -2 internal error (missing capturing bracket) + -3 internal error (opcode not listed) +*/ + +static int +find_minlength(const pcre_uchar *code, const pcre_uchar *startcode, int options, + int recurse_depth) +{ +int length = -1; +/* PCRE_UTF16 has the same value as PCRE_UTF8. */ +BOOL utf = (options & PCRE_UTF8) != 0; +BOOL had_recurse = FALSE; +register int branchlength = 0; +register pcre_uchar *cc = (pcre_uchar *)code + 1 + LINK_SIZE; + +if (*code == OP_CBRA || *code == OP_SCBRA || + *code == OP_CBRAPOS || *code == OP_SCBRAPOS) cc += IMM2_SIZE; + +/* Scan along the opcodes for this branch. If we get to the end of the +branch, check the length against that of the other branches. */ + +for (;;) + { + int d, min; + pcre_uchar *cs, *ce; + register int op = *cc; + + switch (op) + { + case OP_COND: + case OP_SCOND: + + /* If there is only one branch in a condition, the implied branch has zero + length, so we don't add anything. This covers the DEFINE "condition" + automatically. */ + + cs = cc + GET(cc, 1); + if (*cs != OP_ALT) + { + cc = cs + 1 + LINK_SIZE; + break; + } + + /* Otherwise we can fall through and treat it the same as any other + subpattern. */ + + case OP_CBRA: + case OP_SCBRA: + case OP_BRA: + case OP_SBRA: + case OP_CBRAPOS: + case OP_SCBRAPOS: + case OP_BRAPOS: + case OP_SBRAPOS: + case OP_ONCE: + case OP_ONCE_NC: + d = find_minlength(cc, startcode, options, recurse_depth); + if (d < 0) return d; + branchlength += d; + do cc += GET(cc, 1); while (*cc == OP_ALT); + cc += 1 + LINK_SIZE; + break; + + /* ACCEPT makes things far too complicated; we have to give up. */ + + case OP_ACCEPT: + case OP_ASSERT_ACCEPT: + return -1; + + /* Reached end of a branch; if it's a ket it is the end of a nested + call. If it's ALT it is an alternation in a nested call. If it is END it's + the end of the outer call. All can be handled by the same code. If an + ACCEPT was previously encountered, use the length that was in force at that + time, and pass back the shortest ACCEPT length. */ + + case OP_ALT: + case OP_KET: + case OP_KETRMAX: + case OP_KETRMIN: + case OP_KETRPOS: + case OP_END: + if (length < 0 || (!had_recurse && branchlength < length)) + length = branchlength; + if (op != OP_ALT) return length; + cc += 1 + LINK_SIZE; + branchlength = 0; + had_recurse = FALSE; + break; + + /* Skip over assertive subpatterns */ + + case OP_ASSERT: + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + do cc += GET(cc, 1); while (*cc == OP_ALT); + /* Fall through */ + + /* Skip over things that don't match chars */ + + case OP_REVERSE: + case OP_CREF: + case OP_NCREF: + case OP_RREF: + case OP_NRREF: + case OP_DEF: + case OP_CALLOUT: + case OP_SOD: + case OP_SOM: + case OP_EOD: + case OP_EODN: + case OP_CIRC: + case OP_CIRCM: + case OP_DOLL: + case OP_DOLLM: + case OP_NOT_WORD_BOUNDARY: + case OP_WORD_BOUNDARY: + cc += PRIV(OP_lengths)[*cc]; + break; + + /* Skip over a subpattern that has a {0} or {0,x} quantifier */ + + case OP_BRAZERO: + case OP_BRAMINZERO: + case OP_BRAPOSZERO: + case OP_SKIPZERO: + cc += PRIV(OP_lengths)[*cc]; + do cc += GET(cc, 1); while (*cc == OP_ALT); + cc += 1 + LINK_SIZE; + break; + + /* Handle literal characters and + repetitions */ + + case OP_CHAR: + case OP_CHARI: + case OP_NOT: + case OP_NOTI: + case OP_PLUS: + case OP_PLUSI: + case OP_MINPLUS: + case OP_MINPLUSI: + case OP_POSPLUS: + case OP_POSPLUSI: + case OP_NOTPLUS: + case OP_NOTPLUSI: + case OP_NOTMINPLUS: + case OP_NOTMINPLUSI: + case OP_NOTPOSPLUS: + case OP_NOTPOSPLUSI: + branchlength++; + cc += 2; +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]); +#endif + break; + + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEPOSPLUS: + branchlength++; + cc += (cc[1] == OP_PROP || cc[1] == OP_NOTPROP)? 4 : 2; + break; + + /* Handle exact repetitions. The count is already in characters, but we + need to skip over a multibyte character in UTF8 mode. */ + + case OP_EXACT: + case OP_EXACTI: + case OP_NOTEXACT: + case OP_NOTEXACTI: + branchlength += GET2(cc,1); + cc += 2 + IMM2_SIZE; +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]); +#endif + break; + + case OP_TYPEEXACT: + branchlength += GET2(cc,1); + cc += 2 + IMM2_SIZE + ((cc[1 + IMM2_SIZE] == OP_PROP + || cc[1 + IMM2_SIZE] == OP_NOTPROP)? 2 : 0); + break; + + /* Handle single-char non-literal matchers */ + + case OP_PROP: + case OP_NOTPROP: + cc += 2; + /* Fall through */ + + case OP_NOT_DIGIT: + case OP_DIGIT: + case OP_NOT_WHITESPACE: + case OP_WHITESPACE: + case OP_NOT_WORDCHAR: + case OP_WORDCHAR: + case OP_ANY: + case OP_ALLANY: + case OP_EXTUNI: + case OP_HSPACE: + case OP_NOT_HSPACE: + case OP_VSPACE: + case OP_NOT_VSPACE: + branchlength++; + cc++; + break; + + /* "Any newline" might match two characters, but it also might match just + one. */ + + case OP_ANYNL: + branchlength += 1; + cc++; + break; + + /* The single-byte matcher means we can't proceed in UTF-8 mode. (In + non-UTF-8 mode \C will actually be turned into OP_ALLANY, so won't ever + appear, but leave the code, just in case.) */ + + case OP_ANYBYTE: +#ifdef SUPPORT_UTF + if (utf) return -1; +#endif + branchlength++; + cc++; + break; + + /* For repeated character types, we have to test for \p and \P, which have + an extra two bytes of parameters. */ + + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSSTAR: + case OP_TYPEPOSQUERY: + if (cc[1] == OP_PROP || cc[1] == OP_NOTPROP) cc += 2; + cc += PRIV(OP_lengths)[op]; + break; + + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEPOSUPTO: + if (cc[1 + IMM2_SIZE] == OP_PROP + || cc[1 + IMM2_SIZE] == OP_NOTPROP) cc += 2; + cc += PRIV(OP_lengths)[op]; + break; + + /* Check a class for variable quantification */ + +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: + cc += GET(cc, 1) - PRIV(OP_lengths)[OP_CLASS]; + /* Fall through */ +#endif + + case OP_CLASS: + case OP_NCLASS: + cc += PRIV(OP_lengths)[OP_CLASS]; + + switch (*cc) + { + case OP_CRPLUS: + case OP_CRMINPLUS: + branchlength++; + /* Fall through */ + + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRQUERY: + case OP_CRMINQUERY: + cc++; + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + branchlength += GET2(cc,1); + cc += 1 + 2 * IMM2_SIZE; + break; + + default: + branchlength++; + break; + } + break; + + /* Backreferences and subroutine calls are treated in the same way: we find + the minimum length for the subpattern. A recursion, however, causes an + a flag to be set that causes the length of this branch to be ignored. The + logic is that a recursion can only make sense if there is another + alternation that stops the recursing. That will provide the minimum length + (when no recursion happens). A backreference within the group that it is + referencing behaves in the same way. + + If PCRE_JAVASCRIPT_COMPAT is set, a backreference to an unset bracket + matches an empty string (by default it causes a matching failure), so in + that case we must set the minimum length to zero. */ + + case OP_REF: + case OP_REFI: + if ((options & PCRE_JAVASCRIPT_COMPAT) == 0) + { + ce = cs = (pcre_uchar *)PRIV(find_bracket)(startcode, utf, GET2(cc, 1)); + if (cs == NULL) return -2; + do ce += GET(ce, 1); while (*ce == OP_ALT); + if (cc > cs && cc < ce) + { + d = 0; + had_recurse = TRUE; + } + else + { + d = find_minlength(cs, startcode, options, recurse_depth); + } + } + else d = 0; + cc += 1 + IMM2_SIZE; + + /* Handle repeated back references */ + + switch (*cc) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRQUERY: + case OP_CRMINQUERY: + min = 0; + cc++; + break; + + case OP_CRPLUS: + case OP_CRMINPLUS: + min = 1; + cc++; + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + min = GET2(cc, 1); + cc += 1 + 2 * IMM2_SIZE; + break; + + default: + min = 1; + break; + } + + branchlength += min * d; + break; + + /* We can easily detect direct recursion, but not mutual recursion. This is + caught by a recursion depth count. */ + + case OP_RECURSE: + cs = ce = (pcre_uchar *)startcode + GET(cc, 1); + do ce += GET(ce, 1); while (*ce == OP_ALT); + if ((cc > cs && cc < ce) || recurse_depth > 10) + had_recurse = TRUE; + else + { + branchlength += find_minlength(cs, startcode, options, recurse_depth + 1); + } + cc += 1 + LINK_SIZE; + break; + + /* Anything else does not or need not match a character. We can get the + item's length from the table, but for those that can match zero occurrences + of a character, we must take special action for UTF-8 characters. As it + happens, the "NOT" versions of these opcodes are used at present only for + ASCII characters, so they could be omitted from this list. However, in + future that may change, so we include them here so as not to leave a + gotcha for a future maintainer. */ + + case OP_UPTO: + case OP_UPTOI: + case OP_NOTUPTO: + case OP_NOTUPTOI: + case OP_MINUPTO: + case OP_MINUPTOI: + case OP_NOTMINUPTO: + case OP_NOTMINUPTOI: + case OP_POSUPTO: + case OP_POSUPTOI: + case OP_NOTPOSUPTO: + case OP_NOTPOSUPTOI: + + case OP_STAR: + case OP_STARI: + case OP_NOTSTAR: + case OP_NOTSTARI: + case OP_MINSTAR: + case OP_MINSTARI: + case OP_NOTMINSTAR: + case OP_NOTMINSTARI: + case OP_POSSTAR: + case OP_POSSTARI: + case OP_NOTPOSSTAR: + case OP_NOTPOSSTARI: + + case OP_QUERY: + case OP_QUERYI: + case OP_NOTQUERY: + case OP_NOTQUERYI: + case OP_MINQUERY: + case OP_MINQUERYI: + case OP_NOTMINQUERY: + case OP_NOTMINQUERYI: + case OP_POSQUERY: + case OP_POSQUERYI: + case OP_NOTPOSQUERY: + case OP_NOTPOSQUERYI: + + cc += PRIV(OP_lengths)[op]; +#ifdef SUPPORT_UTF + if (utf && HAS_EXTRALEN(cc[-1])) cc += GET_EXTRALEN(cc[-1]); +#endif + break; + + /* Skip these, but we need to add in the name length. */ + + case OP_MARK: + case OP_PRUNE_ARG: + case OP_SKIP_ARG: + case OP_THEN_ARG: + cc += PRIV(OP_lengths)[op] + cc[1]; + break; + + /* The remaining opcodes are just skipped over. */ + + case OP_CLOSE: + case OP_COMMIT: + case OP_FAIL: + case OP_PRUNE: + case OP_SET_SOM: + case OP_SKIP: + case OP_THEN: + cc += PRIV(OP_lengths)[op]; + break; + + /* This should not occur: we list all opcodes explicitly so that when + new ones get added they are properly considered. */ + + default: + return -3; + } + } +/* Control never gets here */ +} + + + +/************************************************* +* Set a bit and maybe its alternate case * +*************************************************/ + +/* Given a character, set its first byte's bit in the table, and also the +corresponding bit for the other version of a letter if we are caseless. In +UTF-8 mode, for characters greater than 127, we can only do the caseless thing +when Unicode property support is available. + +Arguments: + start_bits points to the bit map + p points to the character + caseless the caseless flag + cd the block with char table pointers + utf TRUE for UTF-8 / UTF-16 mode + +Returns: pointer after the character +*/ + +static const pcre_uchar * +set_table_bit(pcre_uint8 *start_bits, const pcre_uchar *p, BOOL caseless, + compile_data *cd, BOOL utf) +{ +unsigned int c = *p; + +#ifdef COMPILE_PCRE8 +SET_BIT(c); + +#ifdef SUPPORT_UTF +if (utf && c > 127) + { + GETCHARINC(c, p); +#ifdef SUPPORT_UCP + if (caseless) + { + pcre_uchar buff[6]; + c = UCD_OTHERCASE(c); + (void)PRIV(ord2utf)(c, buff); + SET_BIT(buff[0]); + } +#endif + return p; + } +#endif + +/* Not UTF-8 mode, or character is less than 127. */ + +if (caseless && (cd->ctypes[c] & ctype_letter) != 0) SET_BIT(cd->fcc[c]); +return p + 1; +#endif + +#ifdef COMPILE_PCRE16 +if (c > 0xff) + { + c = 0xff; + caseless = FALSE; + } +SET_BIT(c); + +#ifdef SUPPORT_UTF +if (utf && c > 127) + { + GETCHARINC(c, p); +#ifdef SUPPORT_UCP + if (caseless) + { + c = UCD_OTHERCASE(c); + if (c > 0xff) + c = 0xff; + SET_BIT(c); + } +#endif + return p; + } +#endif + +if (caseless && (cd->ctypes[c] & ctype_letter) != 0) SET_BIT(cd->fcc[c]); +return p + 1; +#endif +} + + + +/************************************************* +* Set bits for a positive character type * +*************************************************/ + +/* This function sets starting bits for a character type. In UTF-8 mode, we can +only do a direct setting for bytes less than 128, as otherwise there can be +confusion with bytes in the middle of UTF-8 characters. In a "traditional" +environment, the tables will only recognize ASCII characters anyway, but in at +least one Windows environment, some higher bytes bits were set in the tables. +So we deal with that case by considering the UTF-8 encoding. + +Arguments: + start_bits the starting bitmap + cbit type the type of character wanted + table_limit 32 for non-UTF-8; 16 for UTF-8 + cd the block with char table pointers + +Returns: nothing +*/ + +static void +set_type_bits(pcre_uint8 *start_bits, int cbit_type, int table_limit, + compile_data *cd) +{ +register int c; +for (c = 0; c < table_limit; c++) start_bits[c] |= cd->cbits[c+cbit_type]; +#if defined SUPPORT_UTF && defined COMPILE_PCRE8 +if (table_limit == 32) return; +for (c = 128; c < 256; c++) + { + if ((cd->cbits[c/8] & (1 << (c&7))) != 0) + { + pcre_uchar buff[6]; + (void)PRIV(ord2utf)(c, buff); + SET_BIT(buff[0]); + } + } +#endif +} + + +/************************************************* +* Set bits for a negative character type * +*************************************************/ + +/* This function sets starting bits for a negative character type such as \D. +In UTF-8 mode, we can only do a direct setting for bytes less than 128, as +otherwise there can be confusion with bytes in the middle of UTF-8 characters. +Unlike in the positive case, where we can set appropriate starting bits for +specific high-valued UTF-8 characters, in this case we have to set the bits for +all high-valued characters. The lowest is 0xc2, but we overkill by starting at +0xc0 (192) for simplicity. + +Arguments: + start_bits the starting bitmap + cbit type the type of character wanted + table_limit 32 for non-UTF-8; 16 for UTF-8 + cd the block with char table pointers + +Returns: nothing +*/ + +static void +set_nottype_bits(pcre_uint8 *start_bits, int cbit_type, int table_limit, + compile_data *cd) +{ +register int c; +for (c = 0; c < table_limit; c++) start_bits[c] |= ~cd->cbits[c+cbit_type]; +#if defined SUPPORT_UTF && defined COMPILE_PCRE8 +if (table_limit != 32) for (c = 24; c < 32; c++) start_bits[c] = 0xff; +#endif +} + + + +/************************************************* +* Create bitmap of starting bytes * +*************************************************/ + +/* This function scans a compiled unanchored expression recursively and +attempts to build a bitmap of the set of possible starting bytes. As time goes +by, we may be able to get more clever at doing this. The SSB_CONTINUE return is +useful for parenthesized groups in patterns such as (a*)b where the group +provides some optional starting bytes but scanning must continue at the outer +level to find at least one mandatory byte. At the outermost level, this +function fails unless the result is SSB_DONE. + +Arguments: + code points to an expression + start_bits points to a 32-byte table, initialized to 0 + utf TRUE if in UTF-8 / UTF-16 mode + cd the block with char table pointers + +Returns: SSB_FAIL => Failed to find any starting bytes + SSB_DONE => Found mandatory starting bytes + SSB_CONTINUE => Found optional starting bytes + SSB_UNKNOWN => Hit an unrecognized opcode +*/ + +static int +set_start_bits(const pcre_uchar *code, pcre_uint8 *start_bits, BOOL utf, + compile_data *cd) +{ +register int c; +int yield = SSB_DONE; +#if defined SUPPORT_UTF && defined COMPILE_PCRE8 +int table_limit = utf? 16:32; +#else +int table_limit = 32; +#endif + +#if 0 +/* ========================================================================= */ +/* The following comment and code was inserted in January 1999. In May 2006, +when it was observed to cause compiler warnings about unused values, I took it +out again. If anybody is still using OS/2, they will have to put it back +manually. */ + +/* This next statement and the later reference to dummy are here in order to +trick the optimizer of the IBM C compiler for OS/2 into generating correct +code. Apparently IBM isn't going to fix the problem, and we would rather not +disable optimization (in this module it actually makes a big difference, and +the pcre module can use all the optimization it can get). */ + +volatile int dummy; +/* ========================================================================= */ +#endif + +do + { + BOOL try_next = TRUE; + const pcre_uchar *tcode = code + 1 + LINK_SIZE; + + if (*code == OP_CBRA || *code == OP_SCBRA || + *code == OP_CBRAPOS || *code == OP_SCBRAPOS) tcode += IMM2_SIZE; + + while (try_next) /* Loop for items in this branch */ + { + int rc; + + switch(*tcode) + { + /* If we reach something we don't understand, it means a new opcode has + been created that hasn't been added to this code. Hopefully this problem + will be discovered during testing. */ + + default: + return SSB_UNKNOWN; + + /* Fail for a valid opcode that implies no starting bits. */ + + case OP_ACCEPT: + case OP_ASSERT_ACCEPT: + case OP_ALLANY: + case OP_ANY: + case OP_ANYBYTE: + case OP_CIRC: + case OP_CIRCM: + case OP_CLOSE: + case OP_COMMIT: + case OP_COND: + case OP_CREF: + case OP_DEF: + case OP_DOLL: + case OP_DOLLM: + case OP_END: + case OP_EOD: + case OP_EODN: + case OP_EXTUNI: + case OP_FAIL: + case OP_MARK: + case OP_NCREF: + case OP_NOT: + case OP_NOTEXACT: + case OP_NOTEXACTI: + case OP_NOTI: + case OP_NOTMINPLUS: + case OP_NOTMINPLUSI: + case OP_NOTMINQUERY: + case OP_NOTMINQUERYI: + case OP_NOTMINSTAR: + case OP_NOTMINSTARI: + case OP_NOTMINUPTO: + case OP_NOTMINUPTOI: + case OP_NOTPLUS: + case OP_NOTPLUSI: + case OP_NOTPOSPLUS: + case OP_NOTPOSPLUSI: + case OP_NOTPOSQUERY: + case OP_NOTPOSQUERYI: + case OP_NOTPOSSTAR: + case OP_NOTPOSSTARI: + case OP_NOTPOSUPTO: + case OP_NOTPOSUPTOI: + case OP_NOTPROP: + case OP_NOTQUERY: + case OP_NOTQUERYI: + case OP_NOTSTAR: + case OP_NOTSTARI: + case OP_NOTUPTO: + case OP_NOTUPTOI: + case OP_NOT_HSPACE: + case OP_NOT_VSPACE: + case OP_NRREF: + case OP_PROP: + case OP_PRUNE: + case OP_PRUNE_ARG: + case OP_RECURSE: + case OP_REF: + case OP_REFI: + case OP_REVERSE: + case OP_RREF: + case OP_SCOND: + case OP_SET_SOM: + case OP_SKIP: + case OP_SKIP_ARG: + case OP_SOD: + case OP_SOM: + case OP_THEN: + case OP_THEN_ARG: +#if defined SUPPORT_UTF || !defined COMPILE_PCRE8 + case OP_XCLASS: +#endif + return SSB_FAIL; + + /* We can ignore word boundary tests. */ + + case OP_WORD_BOUNDARY: + case OP_NOT_WORD_BOUNDARY: + tcode++; + break; + + /* If we hit a bracket or a positive lookahead assertion, recurse to set + bits from within the subpattern. If it can't find anything, we have to + give up. If it finds some mandatory character(s), we are done for this + branch. Otherwise, carry on scanning after the subpattern. */ + + case OP_BRA: + case OP_SBRA: + case OP_CBRA: + case OP_SCBRA: + case OP_BRAPOS: + case OP_SBRAPOS: + case OP_CBRAPOS: + case OP_SCBRAPOS: + case OP_ONCE: + case OP_ONCE_NC: + case OP_ASSERT: + rc = set_start_bits(tcode, start_bits, utf, cd); + if (rc == SSB_FAIL || rc == SSB_UNKNOWN) return rc; + if (rc == SSB_DONE) try_next = FALSE; else + { + do tcode += GET(tcode, 1); while (*tcode == OP_ALT); + tcode += 1 + LINK_SIZE; + } + break; + + /* If we hit ALT or KET, it means we haven't found anything mandatory in + this branch, though we might have found something optional. For ALT, we + continue with the next alternative, but we have to arrange that the final + result from subpattern is SSB_CONTINUE rather than SSB_DONE. For KET, + return SSB_CONTINUE: if this is the top level, that indicates failure, + but after a nested subpattern, it causes scanning to continue. */ + + case OP_ALT: + yield = SSB_CONTINUE; + try_next = FALSE; + break; + + case OP_KET: + case OP_KETRMAX: + case OP_KETRMIN: + case OP_KETRPOS: + return SSB_CONTINUE; + + /* Skip over callout */ + + case OP_CALLOUT: + tcode += 2 + 2*LINK_SIZE; + break; + + /* Skip over lookbehind and negative lookahead assertions */ + + case OP_ASSERT_NOT: + case OP_ASSERTBACK: + case OP_ASSERTBACK_NOT: + do tcode += GET(tcode, 1); while (*tcode == OP_ALT); + tcode += 1 + LINK_SIZE; + break; + + /* BRAZERO does the bracket, but carries on. */ + + case OP_BRAZERO: + case OP_BRAMINZERO: + case OP_BRAPOSZERO: + rc = set_start_bits(++tcode, start_bits, utf, cd); + if (rc == SSB_FAIL || rc == SSB_UNKNOWN) return rc; +/* ========================================================================= + See the comment at the head of this function concerning the next line, + which was an old fudge for the benefit of OS/2. + dummy = 1; + ========================================================================= */ + do tcode += GET(tcode,1); while (*tcode == OP_ALT); + tcode += 1 + LINK_SIZE; + break; + + /* SKIPZERO skips the bracket. */ + + case OP_SKIPZERO: + tcode++; + do tcode += GET(tcode,1); while (*tcode == OP_ALT); + tcode += 1 + LINK_SIZE; + break; + + /* Single-char * or ? sets the bit and tries the next item */ + + case OP_STAR: + case OP_MINSTAR: + case OP_POSSTAR: + case OP_QUERY: + case OP_MINQUERY: + case OP_POSQUERY: + tcode = set_table_bit(start_bits, tcode + 1, FALSE, cd, utf); + break; + + case OP_STARI: + case OP_MINSTARI: + case OP_POSSTARI: + case OP_QUERYI: + case OP_MINQUERYI: + case OP_POSQUERYI: + tcode = set_table_bit(start_bits, tcode + 1, TRUE, cd, utf); + break; + + /* Single-char upto sets the bit and tries the next */ + + case OP_UPTO: + case OP_MINUPTO: + case OP_POSUPTO: + tcode = set_table_bit(start_bits, tcode + 1 + IMM2_SIZE, FALSE, cd, utf); + break; + + case OP_UPTOI: + case OP_MINUPTOI: + case OP_POSUPTOI: + tcode = set_table_bit(start_bits, tcode + 1 + IMM2_SIZE, TRUE, cd, utf); + break; + + /* At least one single char sets the bit and stops */ + + case OP_EXACT: + tcode += IMM2_SIZE; + /* Fall through */ + case OP_CHAR: + case OP_PLUS: + case OP_MINPLUS: + case OP_POSPLUS: + (void)set_table_bit(start_bits, tcode + 1, FALSE, cd, utf); + try_next = FALSE; + break; + + case OP_EXACTI: + tcode += IMM2_SIZE; + /* Fall through */ + case OP_CHARI: + case OP_PLUSI: + case OP_MINPLUSI: + case OP_POSPLUSI: + (void)set_table_bit(start_bits, tcode + 1, TRUE, cd, utf); + try_next = FALSE; + break; + + /* Special spacing and line-terminating items. These recognize specific + lists of characters. The difference between VSPACE and ANYNL is that the + latter can match the two-character CRLF sequence, but that is not + relevant for finding the first character, so their code here is + identical. */ + + case OP_HSPACE: + SET_BIT(0x09); + SET_BIT(0x20); +#ifdef SUPPORT_UTF + if (utf) + { +#ifdef COMPILE_PCRE8 + SET_BIT(0xC2); /* For U+00A0 */ + SET_BIT(0xE1); /* For U+1680, U+180E */ + SET_BIT(0xE2); /* For U+2000 - U+200A, U+202F, U+205F */ + SET_BIT(0xE3); /* For U+3000 */ +#endif +#ifdef COMPILE_PCRE16 + SET_BIT(0xA0); + SET_BIT(0xFF); /* For characters > 255 */ +#endif + } + else +#endif /* SUPPORT_UTF */ + { + SET_BIT(0xA0); +#ifdef COMPILE_PCRE16 + SET_BIT(0xFF); /* For characters > 255 */ +#endif + } + try_next = FALSE; + break; + + case OP_ANYNL: + case OP_VSPACE: + SET_BIT(0x0A); + SET_BIT(0x0B); + SET_BIT(0x0C); + SET_BIT(0x0D); +#ifdef SUPPORT_UTF + if (utf) + { +#ifdef COMPILE_PCRE8 + SET_BIT(0xC2); /* For U+0085 */ + SET_BIT(0xE2); /* For U+2028, U+2029 */ +#endif +#ifdef COMPILE_PCRE16 + SET_BIT(0x85); + SET_BIT(0xFF); /* For characters > 255 */ +#endif + } + else +#endif /* SUPPORT_UTF */ + { + SET_BIT(0x85); +#ifdef COMPILE_PCRE16 + SET_BIT(0xFF); /* For characters > 255 */ +#endif + } + try_next = FALSE; + break; + + /* Single character types set the bits and stop. Note that if PCRE_UCP + is set, we do not see these op codes because \d etc are converted to + properties. Therefore, these apply in the case when only characters less + than 256 are recognized to match the types. */ + + case OP_NOT_DIGIT: + set_nottype_bits(start_bits, cbit_digit, table_limit, cd); + try_next = FALSE; + break; + + case OP_DIGIT: + set_type_bits(start_bits, cbit_digit, table_limit, cd); + try_next = FALSE; + break; + + /* The cbit_space table has vertical tab as whitespace; we have to + ensure it is set as not whitespace. */ + + case OP_NOT_WHITESPACE: + set_nottype_bits(start_bits, cbit_space, table_limit, cd); + start_bits[1] |= 0x08; + try_next = FALSE; + break; + + /* The cbit_space table has vertical tab as whitespace; we have to + not set it from the table. */ + + case OP_WHITESPACE: + c = start_bits[1]; /* Save in case it was already set */ + set_type_bits(start_bits, cbit_space, table_limit, cd); + start_bits[1] = (start_bits[1] & ~0x08) | c; + try_next = FALSE; + break; + + case OP_NOT_WORDCHAR: + set_nottype_bits(start_bits, cbit_word, table_limit, cd); + try_next = FALSE; + break; + + case OP_WORDCHAR: + set_type_bits(start_bits, cbit_word, table_limit, cd); + try_next = FALSE; + break; + + /* One or more character type fudges the pointer and restarts, knowing + it will hit a single character type and stop there. */ + + case OP_TYPEPLUS: + case OP_TYPEMINPLUS: + case OP_TYPEPOSPLUS: + tcode++; + break; + + case OP_TYPEEXACT: + tcode += 1 + IMM2_SIZE; + break; + + /* Zero or more repeats of character types set the bits and then + try again. */ + + case OP_TYPEUPTO: + case OP_TYPEMINUPTO: + case OP_TYPEPOSUPTO: + tcode += IMM2_SIZE; /* Fall through */ + + case OP_TYPESTAR: + case OP_TYPEMINSTAR: + case OP_TYPEPOSSTAR: + case OP_TYPEQUERY: + case OP_TYPEMINQUERY: + case OP_TYPEPOSQUERY: + switch(tcode[1]) + { + default: + case OP_ANY: + case OP_ALLANY: + return SSB_FAIL; + + case OP_HSPACE: + SET_BIT(0x09); + SET_BIT(0x20); +#ifdef SUPPORT_UTF + if (utf) + { +#ifdef COMPILE_PCRE8 + SET_BIT(0xC2); /* For U+00A0 */ + SET_BIT(0xE1); /* For U+1680, U+180E */ + SET_BIT(0xE2); /* For U+2000 - U+200A, U+202F, U+205F */ + SET_BIT(0xE3); /* For U+3000 */ +#endif +#ifdef COMPILE_PCRE16 + SET_BIT(0xA0); + SET_BIT(0xFF); /* For characters > 255 */ +#endif + } + else +#endif /* SUPPORT_UTF */ + SET_BIT(0xA0); + break; + + case OP_ANYNL: + case OP_VSPACE: + SET_BIT(0x0A); + SET_BIT(0x0B); + SET_BIT(0x0C); + SET_BIT(0x0D); +#ifdef SUPPORT_UTF + if (utf) + { +#ifdef COMPILE_PCRE8 + SET_BIT(0xC2); /* For U+0085 */ + SET_BIT(0xE2); /* For U+2028, U+2029 */ +#endif +#ifdef COMPILE_PCRE16 + SET_BIT(0x85); + SET_BIT(0xFF); /* For characters > 255 */ +#endif + } + else +#endif /* SUPPORT_UTF */ + SET_BIT(0x85); + break; + + case OP_NOT_DIGIT: + set_nottype_bits(start_bits, cbit_digit, table_limit, cd); + break; + + case OP_DIGIT: + set_type_bits(start_bits, cbit_digit, table_limit, cd); + break; + + /* The cbit_space table has vertical tab as whitespace; we have to + ensure it gets set as not whitespace. */ + + case OP_NOT_WHITESPACE: + set_nottype_bits(start_bits, cbit_space, table_limit, cd); + start_bits[1] |= 0x08; + break; + + /* The cbit_space table has vertical tab as whitespace; we have to + avoid setting it. */ + + case OP_WHITESPACE: + c = start_bits[1]; /* Save in case it was already set */ + set_type_bits(start_bits, cbit_space, table_limit, cd); + start_bits[1] = (start_bits[1] & ~0x08) | c; + break; + + case OP_NOT_WORDCHAR: + set_nottype_bits(start_bits, cbit_word, table_limit, cd); + break; + + case OP_WORDCHAR: + set_type_bits(start_bits, cbit_word, table_limit, cd); + break; + } + + tcode += 2; + break; + + /* Character class where all the information is in a bit map: set the + bits and either carry on or not, according to the repeat count. If it was + a negative class, and we are operating with UTF-8 characters, any byte + with a value >= 0xc4 is a potentially valid starter because it starts a + character with a value > 255. */ + + case OP_NCLASS: +#if defined SUPPORT_UTF && defined COMPILE_PCRE8 + if (utf) + { + start_bits[24] |= 0xf0; /* Bits for 0xc4 - 0xc8 */ + memset(start_bits+25, 0xff, 7); /* Bits for 0xc9 - 0xff */ + } +#endif +#ifdef COMPILE_PCRE16 + SET_BIT(0xFF); /* For characters > 255 */ +#endif + /* Fall through */ + + case OP_CLASS: + { + pcre_uint8 *map; + tcode++; + map = (pcre_uint8 *)tcode; + + /* In UTF-8 mode, the bits in a bit map correspond to character + values, not to byte values. However, the bit map we are constructing is + for byte values. So we have to do a conversion for characters whose + value is > 127. In fact, there are only two possible starting bytes for + characters in the range 128 - 255. */ + +#if defined SUPPORT_UTF && defined COMPILE_PCRE8 + if (utf) + { + for (c = 0; c < 16; c++) start_bits[c] |= map[c]; + for (c = 128; c < 256; c++) + { + if ((map[c/8] && (1 << (c&7))) != 0) + { + int d = (c >> 6) | 0xc0; /* Set bit for this starter */ + start_bits[d/8] |= (1 << (d&7)); /* and then skip on to the */ + c = (c & 0xc0) + 0x40 - 1; /* next relevant character. */ + } + } + } + else +#endif + { + /* In non-UTF-8 mode, the two bit maps are completely compatible. */ + for (c = 0; c < 32; c++) start_bits[c] |= map[c]; + } + + /* Advance past the bit map, and act on what follows. For a zero + minimum repeat, continue; otherwise stop processing. */ + + tcode += 32 / sizeof(pcre_uchar); + switch (*tcode) + { + case OP_CRSTAR: + case OP_CRMINSTAR: + case OP_CRQUERY: + case OP_CRMINQUERY: + tcode++; + break; + + case OP_CRRANGE: + case OP_CRMINRANGE: + if (GET2(tcode, 1) == 0) tcode += 1 + 2 * IMM2_SIZE; + else try_next = FALSE; + break; + + default: + try_next = FALSE; + break; + } + } + break; /* End of bitmap class handling */ + + } /* End of switch */ + } /* End of try_next loop */ + + code += GET(code, 1); /* Advance to next branch */ + } +while (*code == OP_ALT); +return yield; +} + + + + + +/************************************************* +* Study a compiled expression * +*************************************************/ + +/* This function is handed a compiled expression that it must study to produce +information that will speed up the matching. It returns a pcre[16]_extra block +which then gets handed back to pcre_exec(). + +Arguments: + re points to the compiled expression + options contains option bits + errorptr points to where to place error messages; + set NULL unless error + +Returns: pointer to a pcre[16]_extra block, with study_data filled in and + the appropriate flags set; + NULL on error or if no optimization possible +*/ + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN pcre_extra * PCRE_CALL_CONVENTION +pcre_study(const pcre *external_re, int options, const char **errorptr) +#else +PCRE_EXP_DEFN pcre16_extra * PCRE_CALL_CONVENTION +pcre16_study(const pcre16 *external_re, int options, const char **errorptr) +#endif +{ +int min; +BOOL bits_set = FALSE; +pcre_uint8 start_bits[32]; +PUBL(extra) *extra = NULL; +pcre_study_data *study; +const pcre_uint8 *tables; +pcre_uchar *code; +compile_data compile_block; +const REAL_PCRE *re = (const REAL_PCRE *)external_re; + +*errorptr = NULL; + +if (re == NULL || re->magic_number != MAGIC_NUMBER) + { + *errorptr = "argument is not a compiled regular expression"; + return NULL; + } + +if ((re->flags & PCRE_MODE) == 0) + { +#ifdef COMPILE_PCRE8 + *errorptr = "argument is compiled in 16 bit mode"; +#else + *errorptr = "argument is compiled in 8 bit mode"; +#endif + return NULL; + } + +if ((options & ~PUBLIC_STUDY_OPTIONS) != 0) + { + *errorptr = "unknown or incorrect option bit(s) set"; + return NULL; + } + +code = (pcre_uchar *)re + re->name_table_offset + + (re->name_count * re->name_entry_size); + +/* For an anchored pattern, or an unanchored pattern that has a first char, or +a multiline pattern that matches only at "line starts", there is no point in +seeking a list of starting bytes. */ + +if ((re->options & PCRE_ANCHORED) == 0 && + (re->flags & (PCRE_FIRSTSET|PCRE_STARTLINE)) == 0) + { + int rc; + + /* Set the character tables in the block that is passed around */ + + tables = re->tables; + +#ifdef COMPILE_PCRE8 + if (tables == NULL) + (void)pcre_fullinfo(external_re, NULL, PCRE_INFO_DEFAULT_TABLES, + (void *)(&tables)); +#else + if (tables == NULL) + (void)pcre16_fullinfo(external_re, NULL, PCRE_INFO_DEFAULT_TABLES, + (void *)(&tables)); +#endif + + compile_block.lcc = tables + lcc_offset; + compile_block.fcc = tables + fcc_offset; + compile_block.cbits = tables + cbits_offset; + compile_block.ctypes = tables + ctypes_offset; + + /* See if we can find a fixed set of initial characters for the pattern. */ + + memset(start_bits, 0, 32 * sizeof(pcre_uint8)); + rc = set_start_bits(code, start_bits, (re->options & PCRE_UTF8) != 0, + &compile_block); + bits_set = rc == SSB_DONE; + if (rc == SSB_UNKNOWN) + { + *errorptr = "internal error: opcode not recognized"; + return NULL; + } + } + +/* Find the minimum length of subject string. */ + +switch(min = find_minlength(code, code, re->options, 0)) + { + case -2: *errorptr = "internal error: missing capturing bracket"; return NULL; + case -3: *errorptr = "internal error: opcode not recognized"; return NULL; + default: break; + } + +/* If a set of starting bytes has been identified, or if the minimum length is +greater than zero, or if JIT optimization has been requested, get a +pcre[16]_extra block and a pcre_study_data block. The study data is put in the +latter, which is pointed to by the former, which may also get additional data +set later by the calling program. At the moment, the size of pcre_study_data +is fixed. We nevertheless save it in a field for returning via the +pcre_fullinfo() function so that if it becomes variable in the future, +we don't have to change that code. */ + +if (bits_set || min > 0 +#ifdef SUPPORT_JIT + || (options & (PCRE_STUDY_JIT_COMPILE | PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE + | PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE)) != 0 +#endif + ) + { + extra = (PUBL(extra) *)(PUBL(malloc)) + (sizeof(PUBL(extra)) + sizeof(pcre_study_data)); + if (extra == NULL) + { + *errorptr = "failed to get memory"; + return NULL; + } + + study = (pcre_study_data *)((char *)extra + sizeof(PUBL(extra))); + extra->flags = PCRE_EXTRA_STUDY_DATA; + extra->study_data = study; + + study->size = sizeof(pcre_study_data); + study->flags = 0; + + /* Set the start bits always, to avoid unset memory errors if the + study data is written to a file, but set the flag only if any of the bits + are set, to save time looking when none are. */ + + if (bits_set) + { + study->flags |= PCRE_STUDY_MAPPED; + memcpy(study->start_bits, start_bits, sizeof(start_bits)); + } + else memset(study->start_bits, 0, 32 * sizeof(pcre_uint8)); + +#ifdef PCRE_DEBUG + if (bits_set) + { + pcre_uint8 *ptr = start_bits; + int i; + + printf("Start bits:\n"); + for (i = 0; i < 32; i++) + printf("%3d: %02x%s", i * 8, *ptr++, ((i + 1) & 0x7) != 0? " " : "\n"); + } +#endif + + /* Always set the minlength value in the block, because the JIT compiler + makes use of it. However, don't set the bit unless the length is greater than + zero - the interpretive pcre_exec() and pcre_dfa_exec() needn't waste time + checking the zero case. */ + + if (min > 0) + { + study->flags |= PCRE_STUDY_MINLEN; + study->minlength = min; + } + else study->minlength = 0; + + /* If JIT support was compiled and requested, attempt the JIT compilation. + If no starting bytes were found, and the minimum length is zero, and JIT + compilation fails, abandon the extra block and return NULL. */ + +#ifdef SUPPORT_JIT + extra->executable_jit = NULL; + if ((options & PCRE_STUDY_JIT_COMPILE) != 0) + PRIV(jit_compile)(re, extra, JIT_COMPILE); + if ((options & PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE) != 0) + PRIV(jit_compile)(re, extra, JIT_PARTIAL_SOFT_COMPILE); + if ((options & PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE) != 0) + PRIV(jit_compile)(re, extra, JIT_PARTIAL_HARD_COMPILE); + + if (study->flags == 0 && (extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) == 0) + { +#ifdef COMPILE_PCRE8 + pcre_free_study(extra); +#endif +#ifdef COMPILE_PCRE16 + pcre16_free_study(extra); +#endif + extra = NULL; + } +#endif + } + +return extra; +} + + +/************************************************* +* Free the study data * +*************************************************/ + +/* This function frees the memory that was obtained by pcre_study(). + +Argument: a pointer to the pcre[16]_extra block +Returns: nothing +*/ + +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN void +pcre_free_study(pcre_extra *extra) +#else +PCRE_EXP_DEFN void +pcre16_free_study(pcre16_extra *extra) +#endif +{ +if (extra == NULL) + return; +#ifdef SUPPORT_JIT +if ((extra->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0 && + extra->executable_jit != NULL) + PRIV(jit_free)(extra->executable_jit); +#endif +PUBL(free)(extra); } /* End of pcre_study.c */ diff --git a/pcre/pcre_tables.c b/pcre/pcre_tables.c --- a/pcre/pcre_tables.c +++ b/pcre/pcre_tables.c @@ -6,7 +6,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -37,6 +37,7 @@ ----------------------------------------------------------------------------- */ +#ifndef PCRE_INCLUDED /* This module contains some fixed tables that are used by more than one of the PCRE code modules. The tables are also #included by the pcretest program, which @@ -50,11 +51,12 @@ #include "pcre_internal.h" +#endif /* PCRE_INCLUDED */ /* Table of sizes for the fixed-length opcodes. It's defined in a macro so that the definition is next to the definition of the opcodes in pcre_internal.h. */ -const uschar _pcre_OP_lengths[] = { OP_LENGTHS }; +const pcre_uint8 PRIV(OP_lengths)[] = { OP_LENGTHS }; @@ -65,28 +67,63 @@ /* These are the breakpoints for different numbers of bytes in a UTF-8 character. */ -#ifdef SUPPORT_UTF8 +#if (defined SUPPORT_UTF && defined COMPILE_PCRE8) \ + || (defined PCRE_INCLUDED && defined SUPPORT_PCRE16) -const int _pcre_utf8_table1[] = +/* These tables are also required by pcretest in 16 bit mode. */ + +const int PRIV(utf8_table1)[] = { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff}; -const int _pcre_utf8_table1_size = sizeof(_pcre_utf8_table1)/sizeof(int); +const int PRIV(utf8_table1_size) = sizeof(PRIV(utf8_table1)) / sizeof(int); /* These are the indicator bits and the mask for the data bits to set in the first byte of a character, indexed by the number of additional bytes. */ -const int _pcre_utf8_table2[] = { 0, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc}; -const int _pcre_utf8_table3[] = { 0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01}; +const int PRIV(utf8_table2)[] = { 0, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc}; +const int PRIV(utf8_table3)[] = { 0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01}; /* Table of the number of extra bytes, indexed by the first byte masked with 0x3f. The highest number for a valid UTF-8 first byte is in fact 0x3d. */ -const uschar _pcre_utf8_table4[] = { +const pcre_uint8 PRIV(utf8_table4)[] = { 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, 3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5 }; +#endif /* (SUPPORT_UTF && COMPILE_PCRE8) || (PCRE_INCLUDED && SUPPORT_PCRE16)*/ + +#ifdef SUPPORT_UTF + +/* Table to translate from particular type value to the general value. */ + +const int PRIV(ucp_gentype)[] = { + ucp_C, ucp_C, ucp_C, ucp_C, ucp_C, /* Cc, Cf, Cn, Co, Cs */ + ucp_L, ucp_L, ucp_L, ucp_L, ucp_L, /* Ll, Lu, Lm, Lo, Lt */ + ucp_M, ucp_M, ucp_M, /* Mc, Me, Mn */ + ucp_N, ucp_N, ucp_N, /* Nd, Nl, No */ + ucp_P, ucp_P, ucp_P, ucp_P, ucp_P, /* Pc, Pd, Pe, Pf, Pi */ + ucp_P, ucp_P, /* Ps, Po */ + ucp_S, ucp_S, ucp_S, ucp_S, /* Sc, Sk, Sm, So */ + ucp_Z, ucp_Z, ucp_Z /* Zl, Zp, Zs */ +}; + +#ifdef SUPPORT_JIT +/* This table reverses PRIV(ucp_gentype). We can save the cost +of a memory load. */ + +const int PRIV(ucp_typerange)[] = { + ucp_Cc, ucp_Cs, + ucp_Ll, ucp_Lu, + ucp_Mc, ucp_Mn, + ucp_Nd, ucp_No, + ucp_Pc, ucp_Ps, + ucp_Sc, ucp_So, + ucp_Zl, ucp_Zs, +}; +#endif /* SUPPORT_JIT */ + /* The pcre_utt[] table below translates Unicode property names into type and code values. It is searched by binary chop, so must be in collating sequence of name. Originally, the table contained pointers to the name strings in the first @@ -94,225 +131,459 @@ a shared library is dynamically loaded. A significant reduction is made by putting all the names into a single, large string and then using offsets in the table itself. Maintenance is more error-prone, but frequent changes to this -data is unlikely. */ +data are unlikely. -const char _pcre_utt_names[] = - "Any\0" - "Arabic\0" - "Armenian\0" - "Balinese\0" - "Bengali\0" - "Bopomofo\0" - "Braille\0" - "Buginese\0" - "Buhid\0" - "C\0" - "Canadian_Aboriginal\0" - "Cc\0" - "Cf\0" - "Cherokee\0" - "Cn\0" - "Co\0" - "Common\0" - "Coptic\0" - "Cs\0" - "Cuneiform\0" - "Cypriot\0" - "Cyrillic\0" - "Deseret\0" - "Devanagari\0" - "Ethiopic\0" - "Georgian\0" - "Glagolitic\0" - "Gothic\0" - "Greek\0" - "Gujarati\0" - "Gurmukhi\0" - "Han\0" - "Hangul\0" - "Hanunoo\0" - "Hebrew\0" - "Hiragana\0" - "Inherited\0" - "Kannada\0" - "Katakana\0" - "Kharoshthi\0" - "Khmer\0" - "L\0" - "L&\0" - "Lao\0" - "Latin\0" - "Limbu\0" - "Linear_B\0" - "Ll\0" - "Lm\0" - "Lo\0" - "Lt\0" - "Lu\0" - "M\0" - "Malayalam\0" - "Mc\0" - "Me\0" - "Mn\0" - "Mongolian\0" - "Myanmar\0" - "N\0" - "Nd\0" - "New_Tai_Lue\0" - "Nko\0" - "Nl\0" - "No\0" - "Ogham\0" - "Old_Italic\0" - "Old_Persian\0" - "Oriya\0" - "Osmanya\0" - "P\0" - "Pc\0" - "Pd\0" - "Pe\0" - "Pf\0" - "Phags_Pa\0" - "Phoenician\0" - "Pi\0" - "Po\0" - "Ps\0" - "Runic\0" - "S\0" - "Sc\0" - "Shavian\0" - "Sinhala\0" - "Sk\0" - "Sm\0" - "So\0" - "Syloti_Nagri\0" - "Syriac\0" - "Tagalog\0" - "Tagbanwa\0" - "Tai_Le\0" - "Tamil\0" - "Telugu\0" - "Thaana\0" - "Thai\0" - "Tibetan\0" - "Tifinagh\0" - "Ugaritic\0" - "Yi\0" - "Z\0" - "Zl\0" - "Zp\0" - "Zs\0"; +July 2008: There is now a script called maint/GenerateUtt.py that can be used +to generate this data automatically instead of maintaining it by hand. -const ucp_type_table _pcre_utt[] = { - { 0, PT_ANY, 0 }, - { 4, PT_SC, ucp_Arabic }, - { 11, PT_SC, ucp_Armenian }, - { 20, PT_SC, ucp_Balinese }, - { 29, PT_SC, ucp_Bengali }, - { 37, PT_SC, ucp_Bopomofo }, - { 46, PT_SC, ucp_Braille }, - { 54, PT_SC, ucp_Buginese }, - { 63, PT_SC, ucp_Buhid }, - { 69, PT_GC, ucp_C }, - { 71, PT_SC, ucp_Canadian_Aboriginal }, - { 91, PT_PC, ucp_Cc }, - { 94, PT_PC, ucp_Cf }, - { 97, PT_SC, ucp_Cherokee }, - { 106, PT_PC, ucp_Cn }, - { 109, PT_PC, ucp_Co }, - { 112, PT_SC, ucp_Common }, - { 119, PT_SC, ucp_Coptic }, - { 126, PT_PC, ucp_Cs }, - { 129, PT_SC, ucp_Cuneiform }, - { 139, PT_SC, ucp_Cypriot }, - { 147, PT_SC, ucp_Cyrillic }, - { 156, PT_SC, ucp_Deseret }, - { 164, PT_SC, ucp_Devanagari }, - { 175, PT_SC, ucp_Ethiopic }, - { 184, PT_SC, ucp_Georgian }, - { 193, PT_SC, ucp_Glagolitic }, - { 204, PT_SC, ucp_Gothic }, - { 211, PT_SC, ucp_Greek }, - { 217, PT_SC, ucp_Gujarati }, - { 226, PT_SC, ucp_Gurmukhi }, - { 235, PT_SC, ucp_Han }, - { 239, PT_SC, ucp_Hangul }, - { 246, PT_SC, ucp_Hanunoo }, - { 254, PT_SC, ucp_Hebrew }, - { 261, PT_SC, ucp_Hiragana }, - { 270, PT_SC, ucp_Inherited }, - { 280, PT_SC, ucp_Kannada }, - { 288, PT_SC, ucp_Katakana }, - { 297, PT_SC, ucp_Kharoshthi }, - { 308, PT_SC, ucp_Khmer }, - { 314, PT_GC, ucp_L }, - { 316, PT_LAMP, 0 }, - { 319, PT_SC, ucp_Lao }, - { 323, PT_SC, ucp_Latin }, - { 329, PT_SC, ucp_Limbu }, - { 335, PT_SC, ucp_Linear_B }, - { 344, PT_PC, ucp_Ll }, - { 347, PT_PC, ucp_Lm }, - { 350, PT_PC, ucp_Lo }, - { 353, PT_PC, ucp_Lt }, - { 356, PT_PC, ucp_Lu }, - { 359, PT_GC, ucp_M }, - { 361, PT_SC, ucp_Malayalam }, - { 371, PT_PC, ucp_Mc }, - { 374, PT_PC, ucp_Me }, - { 377, PT_PC, ucp_Mn }, - { 380, PT_SC, ucp_Mongolian }, - { 390, PT_SC, ucp_Myanmar }, - { 398, PT_GC, ucp_N }, - { 400, PT_PC, ucp_Nd }, - { 403, PT_SC, ucp_New_Tai_Lue }, - { 415, PT_SC, ucp_Nko }, - { 419, PT_PC, ucp_Nl }, - { 422, PT_PC, ucp_No }, - { 425, PT_SC, ucp_Ogham }, - { 431, PT_SC, ucp_Old_Italic }, - { 442, PT_SC, ucp_Old_Persian }, - { 454, PT_SC, ucp_Oriya }, - { 460, PT_SC, ucp_Osmanya }, - { 468, PT_GC, ucp_P }, - { 470, PT_PC, ucp_Pc }, - { 473, PT_PC, ucp_Pd }, - { 476, PT_PC, ucp_Pe }, - { 479, PT_PC, ucp_Pf }, - { 482, PT_SC, ucp_Phags_Pa }, - { 491, PT_SC, ucp_Phoenician }, - { 502, PT_PC, ucp_Pi }, - { 505, PT_PC, ucp_Po }, - { 508, PT_PC, ucp_Ps }, - { 511, PT_SC, ucp_Runic }, - { 517, PT_GC, ucp_S }, - { 519, PT_PC, ucp_Sc }, - { 522, PT_SC, ucp_Shavian }, - { 530, PT_SC, ucp_Sinhala }, - { 538, PT_PC, ucp_Sk }, - { 541, PT_PC, ucp_Sm }, - { 544, PT_PC, ucp_So }, - { 547, PT_SC, ucp_Syloti_Nagri }, - { 560, PT_SC, ucp_Syriac }, - { 567, PT_SC, ucp_Tagalog }, - { 575, PT_SC, ucp_Tagbanwa }, - { 584, PT_SC, ucp_Tai_Le }, - { 591, PT_SC, ucp_Tamil }, - { 597, PT_SC, ucp_Telugu }, - { 604, PT_SC, ucp_Thaana }, - { 611, PT_SC, ucp_Thai }, - { 616, PT_SC, ucp_Tibetan }, - { 624, PT_SC, ucp_Tifinagh }, - { 633, PT_SC, ucp_Ugaritic }, - { 642, PT_SC, ucp_Yi }, - { 645, PT_GC, ucp_Z }, - { 647, PT_PC, ucp_Zl }, - { 650, PT_PC, ucp_Zp }, - { 653, PT_PC, ucp_Zs } +The script was updated in March 2009 to generate a new EBCDIC-compliant +version. Like all other character and string literals that are compared against +the regular expression pattern, we must use STR_ macros instead of literal +strings to make sure that UTF-8 support works on EBCDIC platforms. */ + +#define STRING_Any0 STR_A STR_n STR_y "\0" +#define STRING_Arabic0 STR_A STR_r STR_a STR_b STR_i STR_c "\0" +#define STRING_Armenian0 STR_A STR_r STR_m STR_e STR_n STR_i STR_a STR_n "\0" +#define STRING_Avestan0 STR_A STR_v STR_e STR_s STR_t STR_a STR_n "\0" +#define STRING_Balinese0 STR_B STR_a STR_l STR_i STR_n STR_e STR_s STR_e "\0" +#define STRING_Bamum0 STR_B STR_a STR_m STR_u STR_m "\0" +#define STRING_Batak0 STR_B STR_a STR_t STR_a STR_k "\0" +#define STRING_Bengali0 STR_B STR_e STR_n STR_g STR_a STR_l STR_i "\0" +#define STRING_Bopomofo0 STR_B STR_o STR_p STR_o STR_m STR_o STR_f STR_o "\0" +#define STRING_Brahmi0 STR_B STR_r STR_a STR_h STR_m STR_i "\0" +#define STRING_Braille0 STR_B STR_r STR_a STR_i STR_l STR_l STR_e "\0" +#define STRING_Buginese0 STR_B STR_u STR_g STR_i STR_n STR_e STR_s STR_e "\0" +#define STRING_Buhid0 STR_B STR_u STR_h STR_i STR_d "\0" +#define STRING_C0 STR_C "\0" +#define STRING_Canadian_Aboriginal0 STR_C STR_a STR_n STR_a STR_d STR_i STR_a STR_n STR_UNDERSCORE STR_A STR_b STR_o STR_r STR_i STR_g STR_i STR_n STR_a STR_l "\0" +#define STRING_Carian0 STR_C STR_a STR_r STR_i STR_a STR_n "\0" +#define STRING_Cc0 STR_C STR_c "\0" +#define STRING_Cf0 STR_C STR_f "\0" +#define STRING_Chakma0 STR_C STR_h STR_a STR_k STR_m STR_a "\0" +#define STRING_Cham0 STR_C STR_h STR_a STR_m "\0" +#define STRING_Cherokee0 STR_C STR_h STR_e STR_r STR_o STR_k STR_e STR_e "\0" +#define STRING_Cn0 STR_C STR_n "\0" +#define STRING_Co0 STR_C STR_o "\0" +#define STRING_Common0 STR_C STR_o STR_m STR_m STR_o STR_n "\0" +#define STRING_Coptic0 STR_C STR_o STR_p STR_t STR_i STR_c "\0" +#define STRING_Cs0 STR_C STR_s "\0" +#define STRING_Cuneiform0 STR_C STR_u STR_n STR_e STR_i STR_f STR_o STR_r STR_m "\0" +#define STRING_Cypriot0 STR_C STR_y STR_p STR_r STR_i STR_o STR_t "\0" +#define STRING_Cyrillic0 STR_C STR_y STR_r STR_i STR_l STR_l STR_i STR_c "\0" +#define STRING_Deseret0 STR_D STR_e STR_s STR_e STR_r STR_e STR_t "\0" +#define STRING_Devanagari0 STR_D STR_e STR_v STR_a STR_n STR_a STR_g STR_a STR_r STR_i "\0" +#define STRING_Egyptian_Hieroglyphs0 STR_E STR_g STR_y STR_p STR_t STR_i STR_a STR_n STR_UNDERSCORE STR_H STR_i STR_e STR_r STR_o STR_g STR_l STR_y STR_p STR_h STR_s "\0" +#define STRING_Ethiopic0 STR_E STR_t STR_h STR_i STR_o STR_p STR_i STR_c "\0" +#define STRING_Georgian0 STR_G STR_e STR_o STR_r STR_g STR_i STR_a STR_n "\0" +#define STRING_Glagolitic0 STR_G STR_l STR_a STR_g STR_o STR_l STR_i STR_t STR_i STR_c "\0" +#define STRING_Gothic0 STR_G STR_o STR_t STR_h STR_i STR_c "\0" +#define STRING_Greek0 STR_G STR_r STR_e STR_e STR_k "\0" +#define STRING_Gujarati0 STR_G STR_u STR_j STR_a STR_r STR_a STR_t STR_i "\0" +#define STRING_Gurmukhi0 STR_G STR_u STR_r STR_m STR_u STR_k STR_h STR_i "\0" +#define STRING_Han0 STR_H STR_a STR_n "\0" +#define STRING_Hangul0 STR_H STR_a STR_n STR_g STR_u STR_l "\0" +#define STRING_Hanunoo0 STR_H STR_a STR_n STR_u STR_n STR_o STR_o "\0" +#define STRING_Hebrew0 STR_H STR_e STR_b STR_r STR_e STR_w "\0" +#define STRING_Hiragana0 STR_H STR_i STR_r STR_a STR_g STR_a STR_n STR_a "\0" +#define STRING_Imperial_Aramaic0 STR_I STR_m STR_p STR_e STR_r STR_i STR_a STR_l STR_UNDERSCORE STR_A STR_r STR_a STR_m STR_a STR_i STR_c "\0" +#define STRING_Inherited0 STR_I STR_n STR_h STR_e STR_r STR_i STR_t STR_e STR_d "\0" +#define STRING_Inscriptional_Pahlavi0 STR_I STR_n STR_s STR_c STR_r STR_i STR_p STR_t STR_i STR_o STR_n STR_a STR_l STR_UNDERSCORE STR_P STR_a STR_h STR_l STR_a STR_v STR_i "\0" +#define STRING_Inscriptional_Parthian0 STR_I STR_n STR_s STR_c STR_r STR_i STR_p STR_t STR_i STR_o STR_n STR_a STR_l STR_UNDERSCORE STR_P STR_a STR_r STR_t STR_h STR_i STR_a STR_n "\0" +#define STRING_Javanese0 STR_J STR_a STR_v STR_a STR_n STR_e STR_s STR_e "\0" +#define STRING_Kaithi0 STR_K STR_a STR_i STR_t STR_h STR_i "\0" +#define STRING_Kannada0 STR_K STR_a STR_n STR_n STR_a STR_d STR_a "\0" +#define STRING_Katakana0 STR_K STR_a STR_t STR_a STR_k STR_a STR_n STR_a "\0" +#define STRING_Kayah_Li0 STR_K STR_a STR_y STR_a STR_h STR_UNDERSCORE STR_L STR_i "\0" +#define STRING_Kharoshthi0 STR_K STR_h STR_a STR_r STR_o STR_s STR_h STR_t STR_h STR_i "\0" +#define STRING_Khmer0 STR_K STR_h STR_m STR_e STR_r "\0" +#define STRING_L0 STR_L "\0" +#define STRING_L_AMPERSAND0 STR_L STR_AMPERSAND "\0" +#define STRING_Lao0 STR_L STR_a STR_o "\0" +#define STRING_Latin0 STR_L STR_a STR_t STR_i STR_n "\0" +#define STRING_Lepcha0 STR_L STR_e STR_p STR_c STR_h STR_a "\0" +#define STRING_Limbu0 STR_L STR_i STR_m STR_b STR_u "\0" +#define STRING_Linear_B0 STR_L STR_i STR_n STR_e STR_a STR_r STR_UNDERSCORE STR_B "\0" +#define STRING_Lisu0 STR_L STR_i STR_s STR_u "\0" +#define STRING_Ll0 STR_L STR_l "\0" +#define STRING_Lm0 STR_L STR_m "\0" +#define STRING_Lo0 STR_L STR_o "\0" +#define STRING_Lt0 STR_L STR_t "\0" +#define STRING_Lu0 STR_L STR_u "\0" +#define STRING_Lycian0 STR_L STR_y STR_c STR_i STR_a STR_n "\0" +#define STRING_Lydian0 STR_L STR_y STR_d STR_i STR_a STR_n "\0" +#define STRING_M0 STR_M "\0" +#define STRING_Malayalam0 STR_M STR_a STR_l STR_a STR_y STR_a STR_l STR_a STR_m "\0" +#define STRING_Mandaic0 STR_M STR_a STR_n STR_d STR_a STR_i STR_c "\0" +#define STRING_Mc0 STR_M STR_c "\0" +#define STRING_Me0 STR_M STR_e "\0" +#define STRING_Meetei_Mayek0 STR_M STR_e STR_e STR_t STR_e STR_i STR_UNDERSCORE STR_M STR_a STR_y STR_e STR_k "\0" +#define STRING_Meroitic_Cursive0 STR_M STR_e STR_r STR_o STR_i STR_t STR_i STR_c STR_UNDERSCORE STR_C STR_u STR_r STR_s STR_i STR_v STR_e "\0" +#define STRING_Meroitic_Hieroglyphs0 STR_M STR_e STR_r STR_o STR_i STR_t STR_i STR_c STR_UNDERSCORE STR_H STR_i STR_e STR_r STR_o STR_g STR_l STR_y STR_p STR_h STR_s "\0" +#define STRING_Miao0 STR_M STR_i STR_a STR_o "\0" +#define STRING_Mn0 STR_M STR_n "\0" +#define STRING_Mongolian0 STR_M STR_o STR_n STR_g STR_o STR_l STR_i STR_a STR_n "\0" +#define STRING_Myanmar0 STR_M STR_y STR_a STR_n STR_m STR_a STR_r "\0" +#define STRING_N0 STR_N "\0" +#define STRING_Nd0 STR_N STR_d "\0" +#define STRING_New_Tai_Lue0 STR_N STR_e STR_w STR_UNDERSCORE STR_T STR_a STR_i STR_UNDERSCORE STR_L STR_u STR_e "\0" +#define STRING_Nko0 STR_N STR_k STR_o "\0" +#define STRING_Nl0 STR_N STR_l "\0" +#define STRING_No0 STR_N STR_o "\0" +#define STRING_Ogham0 STR_O STR_g STR_h STR_a STR_m "\0" +#define STRING_Ol_Chiki0 STR_O STR_l STR_UNDERSCORE STR_C STR_h STR_i STR_k STR_i "\0" +#define STRING_Old_Italic0 STR_O STR_l STR_d STR_UNDERSCORE STR_I STR_t STR_a STR_l STR_i STR_c "\0" +#define STRING_Old_Persian0 STR_O STR_l STR_d STR_UNDERSCORE STR_P STR_e STR_r STR_s STR_i STR_a STR_n "\0" +#define STRING_Old_South_Arabian0 STR_O STR_l STR_d STR_UNDERSCORE STR_S STR_o STR_u STR_t STR_h STR_UNDERSCORE STR_A STR_r STR_a STR_b STR_i STR_a STR_n "\0" +#define STRING_Old_Turkic0 STR_O STR_l STR_d STR_UNDERSCORE STR_T STR_u STR_r STR_k STR_i STR_c "\0" +#define STRING_Oriya0 STR_O STR_r STR_i STR_y STR_a "\0" +#define STRING_Osmanya0 STR_O STR_s STR_m STR_a STR_n STR_y STR_a "\0" +#define STRING_P0 STR_P "\0" +#define STRING_Pc0 STR_P STR_c "\0" +#define STRING_Pd0 STR_P STR_d "\0" +#define STRING_Pe0 STR_P STR_e "\0" +#define STRING_Pf0 STR_P STR_f "\0" +#define STRING_Phags_Pa0 STR_P STR_h STR_a STR_g STR_s STR_UNDERSCORE STR_P STR_a "\0" +#define STRING_Phoenician0 STR_P STR_h STR_o STR_e STR_n STR_i STR_c STR_i STR_a STR_n "\0" +#define STRING_Pi0 STR_P STR_i "\0" +#define STRING_Po0 STR_P STR_o "\0" +#define STRING_Ps0 STR_P STR_s "\0" +#define STRING_Rejang0 STR_R STR_e STR_j STR_a STR_n STR_g "\0" +#define STRING_Runic0 STR_R STR_u STR_n STR_i STR_c "\0" +#define STRING_S0 STR_S "\0" +#define STRING_Samaritan0 STR_S STR_a STR_m STR_a STR_r STR_i STR_t STR_a STR_n "\0" +#define STRING_Saurashtra0 STR_S STR_a STR_u STR_r STR_a STR_s STR_h STR_t STR_r STR_a "\0" +#define STRING_Sc0 STR_S STR_c "\0" +#define STRING_Sharada0 STR_S STR_h STR_a STR_r STR_a STR_d STR_a "\0" +#define STRING_Shavian0 STR_S STR_h STR_a STR_v STR_i STR_a STR_n "\0" +#define STRING_Sinhala0 STR_S STR_i STR_n STR_h STR_a STR_l STR_a "\0" +#define STRING_Sk0 STR_S STR_k "\0" +#define STRING_Sm0 STR_S STR_m "\0" +#define STRING_So0 STR_S STR_o "\0" +#define STRING_Sora_Sompeng0 STR_S STR_o STR_r STR_a STR_UNDERSCORE STR_S STR_o STR_m STR_p STR_e STR_n STR_g "\0" +#define STRING_Sundanese0 STR_S STR_u STR_n STR_d STR_a STR_n STR_e STR_s STR_e "\0" +#define STRING_Syloti_Nagri0 STR_S STR_y STR_l STR_o STR_t STR_i STR_UNDERSCORE STR_N STR_a STR_g STR_r STR_i "\0" +#define STRING_Syriac0 STR_S STR_y STR_r STR_i STR_a STR_c "\0" +#define STRING_Tagalog0 STR_T STR_a STR_g STR_a STR_l STR_o STR_g "\0" +#define STRING_Tagbanwa0 STR_T STR_a STR_g STR_b STR_a STR_n STR_w STR_a "\0" +#define STRING_Tai_Le0 STR_T STR_a STR_i STR_UNDERSCORE STR_L STR_e "\0" +#define STRING_Tai_Tham0 STR_T STR_a STR_i STR_UNDERSCORE STR_T STR_h STR_a STR_m "\0" +#define STRING_Tai_Viet0 STR_T STR_a STR_i STR_UNDERSCORE STR_V STR_i STR_e STR_t "\0" +#define STRING_Takri0 STR_T STR_a STR_k STR_r STR_i "\0" +#define STRING_Tamil0 STR_T STR_a STR_m STR_i STR_l "\0" +#define STRING_Telugu0 STR_T STR_e STR_l STR_u STR_g STR_u "\0" +#define STRING_Thaana0 STR_T STR_h STR_a STR_a STR_n STR_a "\0" +#define STRING_Thai0 STR_T STR_h STR_a STR_i "\0" +#define STRING_Tibetan0 STR_T STR_i STR_b STR_e STR_t STR_a STR_n "\0" +#define STRING_Tifinagh0 STR_T STR_i STR_f STR_i STR_n STR_a STR_g STR_h "\0" +#define STRING_Ugaritic0 STR_U STR_g STR_a STR_r STR_i STR_t STR_i STR_c "\0" +#define STRING_Vai0 STR_V STR_a STR_i "\0" +#define STRING_Xan0 STR_X STR_a STR_n "\0" +#define STRING_Xps0 STR_X STR_p STR_s "\0" +#define STRING_Xsp0 STR_X STR_s STR_p "\0" +#define STRING_Xwd0 STR_X STR_w STR_d "\0" +#define STRING_Yi0 STR_Y STR_i "\0" +#define STRING_Z0 STR_Z "\0" +#define STRING_Zl0 STR_Z STR_l "\0" +#define STRING_Zp0 STR_Z STR_p "\0" +#define STRING_Zs0 STR_Z STR_s "\0" + +const char PRIV(utt_names)[] = + STRING_Any0 + STRING_Arabic0 + STRING_Armenian0 + STRING_Avestan0 + STRING_Balinese0 + STRING_Bamum0 + STRING_Batak0 + STRING_Bengali0 + STRING_Bopomofo0 + STRING_Brahmi0 + STRING_Braille0 + STRING_Buginese0 + STRING_Buhid0 + STRING_C0 + STRING_Canadian_Aboriginal0 + STRING_Carian0 + STRING_Cc0 + STRING_Cf0 + STRING_Chakma0 + STRING_Cham0 + STRING_Cherokee0 + STRING_Cn0 + STRING_Co0 + STRING_Common0 + STRING_Coptic0 + STRING_Cs0 + STRING_Cuneiform0 + STRING_Cypriot0 + STRING_Cyrillic0 + STRING_Deseret0 + STRING_Devanagari0 + STRING_Egyptian_Hieroglyphs0 + STRING_Ethiopic0 + STRING_Georgian0 + STRING_Glagolitic0 + STRING_Gothic0 + STRING_Greek0 + STRING_Gujarati0 + STRING_Gurmukhi0 + STRING_Han0 + STRING_Hangul0 + STRING_Hanunoo0 + STRING_Hebrew0 + STRING_Hiragana0 + STRING_Imperial_Aramaic0 + STRING_Inherited0 + STRING_Inscriptional_Pahlavi0 + STRING_Inscriptional_Parthian0 + STRING_Javanese0 + STRING_Kaithi0 + STRING_Kannada0 + STRING_Katakana0 + STRING_Kayah_Li0 + STRING_Kharoshthi0 + STRING_Khmer0 + STRING_L0 + STRING_L_AMPERSAND0 + STRING_Lao0 + STRING_Latin0 + STRING_Lepcha0 + STRING_Limbu0 + STRING_Linear_B0 + STRING_Lisu0 + STRING_Ll0 + STRING_Lm0 + STRING_Lo0 + STRING_Lt0 + STRING_Lu0 + STRING_Lycian0 + STRING_Lydian0 + STRING_M0 + STRING_Malayalam0 + STRING_Mandaic0 + STRING_Mc0 + STRING_Me0 + STRING_Meetei_Mayek0 + STRING_Meroitic_Cursive0 + STRING_Meroitic_Hieroglyphs0 + STRING_Miao0 + STRING_Mn0 + STRING_Mongolian0 + STRING_Myanmar0 + STRING_N0 + STRING_Nd0 + STRING_New_Tai_Lue0 + STRING_Nko0 + STRING_Nl0 + STRING_No0 + STRING_Ogham0 + STRING_Ol_Chiki0 + STRING_Old_Italic0 + STRING_Old_Persian0 + STRING_Old_South_Arabian0 + STRING_Old_Turkic0 + STRING_Oriya0 + STRING_Osmanya0 + STRING_P0 + STRING_Pc0 + STRING_Pd0 + STRING_Pe0 + STRING_Pf0 + STRING_Phags_Pa0 + STRING_Phoenician0 + STRING_Pi0 + STRING_Po0 + STRING_Ps0 + STRING_Rejang0 + STRING_Runic0 + STRING_S0 + STRING_Samaritan0 + STRING_Saurashtra0 + STRING_Sc0 + STRING_Sharada0 + STRING_Shavian0 + STRING_Sinhala0 + STRING_Sk0 + STRING_Sm0 + STRING_So0 + STRING_Sora_Sompeng0 + STRING_Sundanese0 + STRING_Syloti_Nagri0 + STRING_Syriac0 + STRING_Tagalog0 + STRING_Tagbanwa0 + STRING_Tai_Le0 + STRING_Tai_Tham0 + STRING_Tai_Viet0 + STRING_Takri0 + STRING_Tamil0 + STRING_Telugu0 + STRING_Thaana0 + STRING_Thai0 + STRING_Tibetan0 + STRING_Tifinagh0 + STRING_Ugaritic0 + STRING_Vai0 + STRING_Xan0 + STRING_Xps0 + STRING_Xsp0 + STRING_Xwd0 + STRING_Yi0 + STRING_Z0 + STRING_Zl0 + STRING_Zp0 + STRING_Zs0; + +const ucp_type_table PRIV(utt)[] = { + { 0, PT_ANY, 0 }, + { 4, PT_SC, ucp_Arabic }, + { 11, PT_SC, ucp_Armenian }, + { 20, PT_SC, ucp_Avestan }, + { 28, PT_SC, ucp_Balinese }, + { 37, PT_SC, ucp_Bamum }, + { 43, PT_SC, ucp_Batak }, + { 49, PT_SC, ucp_Bengali }, + { 57, PT_SC, ucp_Bopomofo }, + { 66, PT_SC, ucp_Brahmi }, + { 73, PT_SC, ucp_Braille }, + { 81, PT_SC, ucp_Buginese }, + { 90, PT_SC, ucp_Buhid }, + { 96, PT_GC, ucp_C }, + { 98, PT_SC, ucp_Canadian_Aboriginal }, + { 118, PT_SC, ucp_Carian }, + { 125, PT_PC, ucp_Cc }, + { 128, PT_PC, ucp_Cf }, + { 131, PT_SC, ucp_Chakma }, + { 138, PT_SC, ucp_Cham }, + { 143, PT_SC, ucp_Cherokee }, + { 152, PT_PC, ucp_Cn }, + { 155, PT_PC, ucp_Co }, + { 158, PT_SC, ucp_Common }, + { 165, PT_SC, ucp_Coptic }, + { 172, PT_PC, ucp_Cs }, + { 175, PT_SC, ucp_Cuneiform }, + { 185, PT_SC, ucp_Cypriot }, + { 193, PT_SC, ucp_Cyrillic }, + { 202, PT_SC, ucp_Deseret }, + { 210, PT_SC, ucp_Devanagari }, + { 221, PT_SC, ucp_Egyptian_Hieroglyphs }, + { 242, PT_SC, ucp_Ethiopic }, + { 251, PT_SC, ucp_Georgian }, + { 260, PT_SC, ucp_Glagolitic }, + { 271, PT_SC, ucp_Gothic }, + { 278, PT_SC, ucp_Greek }, + { 284, PT_SC, ucp_Gujarati }, + { 293, PT_SC, ucp_Gurmukhi }, + { 302, PT_SC, ucp_Han }, + { 306, PT_SC, ucp_Hangul }, + { 313, PT_SC, ucp_Hanunoo }, + { 321, PT_SC, ucp_Hebrew }, + { 328, PT_SC, ucp_Hiragana }, + { 337, PT_SC, ucp_Imperial_Aramaic }, + { 354, PT_SC, ucp_Inherited }, + { 364, PT_SC, ucp_Inscriptional_Pahlavi }, + { 386, PT_SC, ucp_Inscriptional_Parthian }, + { 409, PT_SC, ucp_Javanese }, + { 418, PT_SC, ucp_Kaithi }, + { 425, PT_SC, ucp_Kannada }, + { 433, PT_SC, ucp_Katakana }, + { 442, PT_SC, ucp_Kayah_Li }, + { 451, PT_SC, ucp_Kharoshthi }, + { 462, PT_SC, ucp_Khmer }, + { 468, PT_GC, ucp_L }, + { 470, PT_LAMP, 0 }, + { 473, PT_SC, ucp_Lao }, + { 477, PT_SC, ucp_Latin }, + { 483, PT_SC, ucp_Lepcha }, + { 490, PT_SC, ucp_Limbu }, + { 496, PT_SC, ucp_Linear_B }, + { 505, PT_SC, ucp_Lisu }, + { 510, PT_PC, ucp_Ll }, + { 513, PT_PC, ucp_Lm }, + { 516, PT_PC, ucp_Lo }, + { 519, PT_PC, ucp_Lt }, + { 522, PT_PC, ucp_Lu }, + { 525, PT_SC, ucp_Lycian }, + { 532, PT_SC, ucp_Lydian }, + { 539, PT_GC, ucp_M }, + { 541, PT_SC, ucp_Malayalam }, + { 551, PT_SC, ucp_Mandaic }, + { 559, PT_PC, ucp_Mc }, + { 562, PT_PC, ucp_Me }, + { 565, PT_SC, ucp_Meetei_Mayek }, + { 578, PT_SC, ucp_Meroitic_Cursive }, + { 595, PT_SC, ucp_Meroitic_Hieroglyphs }, + { 616, PT_SC, ucp_Miao }, + { 621, PT_PC, ucp_Mn }, + { 624, PT_SC, ucp_Mongolian }, + { 634, PT_SC, ucp_Myanmar }, + { 642, PT_GC, ucp_N }, + { 644, PT_PC, ucp_Nd }, + { 647, PT_SC, ucp_New_Tai_Lue }, + { 659, PT_SC, ucp_Nko }, + { 663, PT_PC, ucp_Nl }, + { 666, PT_PC, ucp_No }, + { 669, PT_SC, ucp_Ogham }, + { 675, PT_SC, ucp_Ol_Chiki }, + { 684, PT_SC, ucp_Old_Italic }, + { 695, PT_SC, ucp_Old_Persian }, + { 707, PT_SC, ucp_Old_South_Arabian }, + { 725, PT_SC, ucp_Old_Turkic }, + { 736, PT_SC, ucp_Oriya }, + { 742, PT_SC, ucp_Osmanya }, + { 750, PT_GC, ucp_P }, + { 752, PT_PC, ucp_Pc }, + { 755, PT_PC, ucp_Pd }, + { 758, PT_PC, ucp_Pe }, + { 761, PT_PC, ucp_Pf }, + { 764, PT_SC, ucp_Phags_Pa }, + { 773, PT_SC, ucp_Phoenician }, + { 784, PT_PC, ucp_Pi }, + { 787, PT_PC, ucp_Po }, + { 790, PT_PC, ucp_Ps }, + { 793, PT_SC, ucp_Rejang }, + { 800, PT_SC, ucp_Runic }, + { 806, PT_GC, ucp_S }, + { 808, PT_SC, ucp_Samaritan }, + { 818, PT_SC, ucp_Saurashtra }, + { 829, PT_PC, ucp_Sc }, + { 832, PT_SC, ucp_Sharada }, + { 840, PT_SC, ucp_Shavian }, + { 848, PT_SC, ucp_Sinhala }, + { 856, PT_PC, ucp_Sk }, + { 859, PT_PC, ucp_Sm }, + { 862, PT_PC, ucp_So }, + { 865, PT_SC, ucp_Sora_Sompeng }, + { 878, PT_SC, ucp_Sundanese }, + { 888, PT_SC, ucp_Syloti_Nagri }, + { 901, PT_SC, ucp_Syriac }, + { 908, PT_SC, ucp_Tagalog }, + { 916, PT_SC, ucp_Tagbanwa }, + { 925, PT_SC, ucp_Tai_Le }, + { 932, PT_SC, ucp_Tai_Tham }, + { 941, PT_SC, ucp_Tai_Viet }, + { 950, PT_SC, ucp_Takri }, + { 956, PT_SC, ucp_Tamil }, + { 962, PT_SC, ucp_Telugu }, + { 969, PT_SC, ucp_Thaana }, + { 976, PT_SC, ucp_Thai }, + { 981, PT_SC, ucp_Tibetan }, + { 989, PT_SC, ucp_Tifinagh }, + { 998, PT_SC, ucp_Ugaritic }, + { 1007, PT_SC, ucp_Vai }, + { 1011, PT_ALNUM, 0 }, + { 1015, PT_PXSPACE, 0 }, + { 1019, PT_SPACE, 0 }, + { 1023, PT_WORD, 0 }, + { 1027, PT_SC, ucp_Yi }, + { 1030, PT_GC, ucp_Z }, + { 1032, PT_PC, ucp_Zl }, + { 1035, PT_PC, ucp_Zp }, + { 1038, PT_PC, ucp_Zs } }; -const int _pcre_utt_size = sizeof(_pcre_utt)/sizeof(ucp_type_table); +const int PRIV(utt_size) = sizeof(PRIV(utt)) / sizeof(ucp_type_table); -#endif /* SUPPORT_UTF8 */ +#endif /* SUPPORT_UTF */ /* End of pcre_tables.c */ diff --git a/pcre/pcre_try_flipped.c b/pcre/pcre_try_flipped.c deleted file mode 100644 --- a/pcre/pcre_try_flipped.c +++ /dev/null @@ -1,137 +0,0 @@ -/************************************************* -* Perl-Compatible Regular Expressions * -*************************************************/ - -/* PCRE is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. - - Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* This module contains an internal function that tests a compiled pattern to -see if it was compiled with the opposite endianness. If so, it uses an -auxiliary local function to flip the appropriate bytes. */ - - -#ifdef HAVE_CONFIG_H -#include "config.h" -#endif - -#include "pcre_internal.h" - - -/************************************************* -* Flip bytes in an integer * -*************************************************/ - -/* This function is called when the magic number in a regex doesn't match, in -order to flip its bytes to see if we are dealing with a pattern that was -compiled on a host of different endianness. If so, this function is used to -flip other byte values. - -Arguments: - value the number to flip - n the number of bytes to flip (assumed to be 2 or 4) - -Returns: the flipped value -*/ - -static unsigned long int -byteflip(unsigned long int value, int n) -{ -if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8); -return ((value & 0x000000ff) << 24) | - ((value & 0x0000ff00) << 8) | - ((value & 0x00ff0000) >> 8) | - ((value & 0xff000000) >> 24); -} - - - -/************************************************* -* Test for a byte-flipped compiled regex * -*************************************************/ - -/* This function is called from pcre_exec(), pcre_dfa_exec(), and also from -pcre_fullinfo(). Its job is to test whether the regex is byte-flipped - that -is, it was compiled on a system of opposite endianness. The function is called -only when the native MAGIC_NUMBER test fails. If the regex is indeed flipped, -we flip all the relevant values into a different data block, and return it. - -Arguments: - re points to the regex - study points to study data, or NULL - internal_re points to a new regex block - internal_study points to a new study block - -Returns: the new block if is is indeed a byte-flipped regex - NULL if it is not -*/ - -real_pcre * -_pcre_try_flipped(const real_pcre *re, real_pcre *internal_re, - const pcre_study_data *study, pcre_study_data *internal_study) -{ -if (byteflip(re->magic_number, sizeof(re->magic_number)) != MAGIC_NUMBER) - return NULL; - -*internal_re = *re; /* To copy other fields */ -internal_re->size = byteflip(re->size, sizeof(re->size)); -internal_re->options = byteflip(re->options, sizeof(re->options)); -internal_re->flags = (pcre_uint16)byteflip(re->flags, sizeof(re->flags)); -internal_re->top_bracket = - (pcre_uint16)byteflip(re->top_bracket, sizeof(re->top_bracket)); -internal_re->top_backref = - (pcre_uint16)byteflip(re->top_backref, sizeof(re->top_backref)); -internal_re->first_byte = - (pcre_uint16)byteflip(re->first_byte, sizeof(re->first_byte)); -internal_re->req_byte = - (pcre_uint16)byteflip(re->req_byte, sizeof(re->req_byte)); -internal_re->name_table_offset = - (pcre_uint16)byteflip(re->name_table_offset, sizeof(re->name_table_offset)); -internal_re->name_entry_size = - (pcre_uint16)byteflip(re->name_entry_size, sizeof(re->name_entry_size)); -internal_re->name_count = - (pcre_uint16)byteflip(re->name_count, sizeof(re->name_count)); - -if (study != NULL) - { - *internal_study = *study; /* To copy other fields */ - internal_study->size = byteflip(study->size, sizeof(study->size)); - internal_study->options = byteflip(study->options, sizeof(study->options)); - } - -return internal_re; -} - -/* End of pcre_tryflipped.c */ diff --git a/pcre/pcre_ucd.c b/pcre/pcre_ucd.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_ucd.c @@ -0,0 +1,3098 @@ +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + +/* Unicode character database. */ +/* This file was autogenerated by the MultiStage2.py script. */ +/* Total size: 62904 bytes, block size: 128. */ + +/* The tables herein are needed only when UCP support is built */ +/* into PCRE. This module should not be referenced otherwise, so */ +/* it should not matter whether it is compiled or not. However */ +/* a comment was received about space saving - maybe the guy linked */ +/* all the modules rather than using a library - so we include a */ +/* condition to cut out the tables when not needed. But don't leave */ +/* a totally empty module because some compilers barf at that. */ +/* Instead, just supply small dummy tables. */ + +#ifndef SUPPORT_UCP +const ucd_record PRIV(ucd_records)[] = {{0,0,0 }}; +const pcre_uint8 PRIV(ucd_stage1)[] = {0}; +const pcre_uint16 PRIV(ucd_stage2)[] = {0}; +#else + +/* When recompiling tables with a new Unicode version, +please check types in the structure definition from pcre_internal.h: +typedef struct { +pcre_uint8 property_0; +pcre_uint8 property_1; +pcre_int32 property_2; +} ucd_record; */ + + +const ucd_record PRIV(ucd_records)[] = { /* 4536 bytes, record size 8 */ + { 9, 0, 0, }, /* 0 */ + { 9, 29, 0, }, /* 1 */ + { 9, 21, 0, }, /* 2 */ + { 9, 23, 0, }, /* 3 */ + { 9, 22, 0, }, /* 4 */ + { 9, 18, 0, }, /* 5 */ + { 9, 25, 0, }, /* 6 */ + { 9, 17, 0, }, /* 7 */ + { 9, 13, 0, }, /* 8 */ + { 33, 9, 32, }, /* 9 */ + { 9, 24, 0, }, /* 10 */ + { 9, 16, 0, }, /* 11 */ + { 33, 5, -32, }, /* 12 */ + { 9, 26, 0, }, /* 13 */ + { 33, 7, 0, }, /* 14 */ + { 9, 20, 0, }, /* 15 */ + { 9, 1, 0, }, /* 16 */ + { 9, 15, 0, }, /* 17 */ + { 9, 5, 743, }, /* 18 */ + { 9, 19, 0, }, /* 19 */ + { 33, 5, 0, }, /* 20 */ + { 33, 5, 121, }, /* 21 */ + { 33, 9, 1, }, /* 22 */ + { 33, 5, -1, }, /* 23 */ + { 33, 9, -199, }, /* 24 */ + { 33, 5, -232, }, /* 25 */ + { 33, 9, -121, }, /* 26 */ + { 33, 5, -300, }, /* 27 */ + { 33, 5, 195, }, /* 28 */ + { 33, 9, 210, }, /* 29 */ + { 33, 9, 206, }, /* 30 */ + { 33, 9, 205, }, /* 31 */ + { 33, 9, 79, }, /* 32 */ + { 33, 9, 202, }, /* 33 */ + { 33, 9, 203, }, /* 34 */ + { 33, 9, 207, }, /* 35 */ + { 33, 5, 97, }, /* 36 */ + { 33, 9, 211, }, /* 37 */ + { 33, 9, 209, }, /* 38 */ + { 33, 5, 163, }, /* 39 */ + { 33, 9, 213, }, /* 40 */ + { 33, 5, 130, }, /* 41 */ + { 33, 9, 214, }, /* 42 */ + { 33, 9, 218, }, /* 43 */ + { 33, 9, 217, }, /* 44 */ + { 33, 9, 219, }, /* 45 */ + { 33, 5, 56, }, /* 46 */ + { 33, 9, 2, }, /* 47 */ + { 33, 8, -1, }, /* 48 */ + { 33, 5, -2, }, /* 49 */ + { 33, 5, -79, }, /* 50 */ + { 33, 9, -97, }, /* 51 */ + { 33, 9, -56, }, /* 52 */ + { 33, 9, -130, }, /* 53 */ + { 33, 9, 10795, }, /* 54 */ + { 33, 9, -163, }, /* 55 */ + { 33, 9, 10792, }, /* 56 */ + { 33, 5, 10815, }, /* 57 */ + { 33, 9, -195, }, /* 58 */ + { 33, 9, 69, }, /* 59 */ + { 33, 9, 71, }, /* 60 */ + { 33, 5, 10783, }, /* 61 */ + { 33, 5, 10780, }, /* 62 */ + { 33, 5, 10782, }, /* 63 */ + { 33, 5, -210, }, /* 64 */ + { 33, 5, -206, }, /* 65 */ + { 33, 5, -205, }, /* 66 */ + { 33, 5, -202, }, /* 67 */ + { 33, 5, -203, }, /* 68 */ + { 33, 5, -207, }, /* 69 */ + { 33, 5, 42280, }, /* 70 */ + { 33, 5, 42308, }, /* 71 */ + { 33, 5, -209, }, /* 72 */ + { 33, 5, -211, }, /* 73 */ + { 33, 5, 10743, }, /* 74 */ + { 33, 5, 10749, }, /* 75 */ + { 33, 5, -213, }, /* 76 */ + { 33, 5, -214, }, /* 77 */ + { 33, 5, 10727, }, /* 78 */ + { 33, 5, -218, }, /* 79 */ + { 33, 5, -69, }, /* 80 */ + { 33, 5, -217, }, /* 81 */ + { 33, 5, -71, }, /* 82 */ + { 33, 5, -219, }, /* 83 */ + { 33, 6, 0, }, /* 84 */ + { 9, 6, 0, }, /* 85 */ + { 3, 24, 0, }, /* 86 */ + { 27, 12, 0, }, /* 87 */ + { 27, 12, 84, }, /* 88 */ + { 19, 9, 1, }, /* 89 */ + { 19, 5, -1, }, /* 90 */ + { 19, 24, 0, }, /* 91 */ + { 9, 2, 0, }, /* 92 */ + { 19, 6, 0, }, /* 93 */ + { 19, 5, 130, }, /* 94 */ + { 19, 9, 38, }, /* 95 */ + { 19, 9, 37, }, /* 96 */ + { 19, 9, 64, }, /* 97 */ + { 19, 9, 63, }, /* 98 */ + { 19, 5, 0, }, /* 99 */ + { 19, 9, 32, }, /* 100 */ + { 19, 5, -38, }, /* 101 */ + { 19, 5, -37, }, /* 102 */ + { 19, 5, -32, }, /* 103 */ + { 19, 5, -31, }, /* 104 */ + { 19, 5, -64, }, /* 105 */ + { 19, 5, -63, }, /* 106 */ + { 19, 9, 8, }, /* 107 */ + { 19, 5, -62, }, /* 108 */ + { 19, 5, -57, }, /* 109 */ + { 19, 9, 0, }, /* 110 */ + { 19, 5, -47, }, /* 111 */ + { 19, 5, -54, }, /* 112 */ + { 19, 5, -8, }, /* 113 */ + { 10, 9, 1, }, /* 114 */ + { 10, 5, -1, }, /* 115 */ + { 19, 5, -86, }, /* 116 */ + { 19, 5, -80, }, /* 117 */ + { 19, 5, 7, }, /* 118 */ + { 19, 9, -60, }, /* 119 */ + { 19, 5, -96, }, /* 120 */ + { 19, 25, 0, }, /* 121 */ + { 19, 9, -7, }, /* 122 */ + { 19, 9, -130, }, /* 123 */ + { 12, 9, 80, }, /* 124 */ + { 12, 9, 32, }, /* 125 */ + { 12, 5, -32, }, /* 126 */ + { 12, 5, -80, }, /* 127 */ + { 12, 9, 1, }, /* 128 */ + { 12, 5, -1, }, /* 129 */ + { 12, 26, 0, }, /* 130 */ + { 12, 12, 0, }, /* 131 */ + { 12, 11, 0, }, /* 132 */ + { 12, 9, 15, }, /* 133 */ + { 12, 5, -15, }, /* 134 */ + { 1, 9, 48, }, /* 135 */ + { 1, 6, 0, }, /* 136 */ + { 1, 21, 0, }, /* 137 */ + { 1, 5, -48, }, /* 138 */ + { 1, 5, 0, }, /* 139 */ + { 1, 17, 0, }, /* 140 */ + { 1, 23, 0, }, /* 141 */ + { 25, 12, 0, }, /* 142 */ + { 25, 17, 0, }, /* 143 */ + { 25, 21, 0, }, /* 144 */ + { 25, 7, 0, }, /* 145 */ + { 0, 1, 0, }, /* 146 */ + { 0, 25, 0, }, /* 147 */ + { 0, 21, 0, }, /* 148 */ + { 0, 23, 0, }, /* 149 */ + { 0, 26, 0, }, /* 150 */ + { 0, 12, 0, }, /* 151 */ + { 0, 7, 0, }, /* 152 */ + { 0, 6, 0, }, /* 153 */ + { 0, 13, 0, }, /* 154 */ + { 49, 21, 0, }, /* 155 */ + { 49, 1, 0, }, /* 156 */ + { 49, 7, 0, }, /* 157 */ + { 49, 12, 0, }, /* 158 */ + { 55, 7, 0, }, /* 159 */ + { 55, 12, 0, }, /* 160 */ + { 63, 13, 0, }, /* 161 */ + { 63, 7, 0, }, /* 162 */ + { 63, 12, 0, }, /* 163 */ + { 63, 6, 0, }, /* 164 */ + { 63, 26, 0, }, /* 165 */ + { 63, 21, 0, }, /* 166 */ + { 89, 7, 0, }, /* 167 */ + { 89, 12, 0, }, /* 168 */ + { 89, 6, 0, }, /* 169 */ + { 89, 21, 0, }, /* 170 */ + { 94, 7, 0, }, /* 171 */ + { 94, 12, 0, }, /* 172 */ + { 94, 21, 0, }, /* 173 */ + { 14, 12, 0, }, /* 174 */ + { 14, 10, 0, }, /* 175 */ + { 14, 7, 0, }, /* 176 */ + { 14, 13, 0, }, /* 177 */ + { 14, 21, 0, }, /* 178 */ + { 14, 6, 0, }, /* 179 */ + { 2, 12, 0, }, /* 180 */ + { 2, 10, 0, }, /* 181 */ + { 2, 7, 0, }, /* 182 */ + { 2, 13, 0, }, /* 183 */ + { 2, 23, 0, }, /* 184 */ + { 2, 15, 0, }, /* 185 */ + { 2, 26, 0, }, /* 186 */ + { 21, 12, 0, }, /* 187 */ + { 21, 10, 0, }, /* 188 */ + { 21, 7, 0, }, /* 189 */ + { 21, 13, 0, }, /* 190 */ + { 20, 12, 0, }, /* 191 */ + { 20, 10, 0, }, /* 192 */ + { 20, 7, 0, }, /* 193 */ + { 20, 13, 0, }, /* 194 */ + { 20, 21, 0, }, /* 195 */ + { 20, 23, 0, }, /* 196 */ + { 43, 12, 0, }, /* 197 */ + { 43, 10, 0, }, /* 198 */ + { 43, 7, 0, }, /* 199 */ + { 43, 13, 0, }, /* 200 */ + { 43, 26, 0, }, /* 201 */ + { 43, 15, 0, }, /* 202 */ + { 53, 12, 0, }, /* 203 */ + { 53, 7, 0, }, /* 204 */ + { 53, 10, 0, }, /* 205 */ + { 53, 13, 0, }, /* 206 */ + { 53, 15, 0, }, /* 207 */ + { 53, 26, 0, }, /* 208 */ + { 53, 23, 0, }, /* 209 */ + { 54, 10, 0, }, /* 210 */ + { 54, 7, 0, }, /* 211 */ + { 54, 12, 0, }, /* 212 */ + { 54, 13, 0, }, /* 213 */ + { 54, 15, 0, }, /* 214 */ + { 54, 26, 0, }, /* 215 */ + { 28, 10, 0, }, /* 216 */ + { 28, 7, 0, }, /* 217 */ + { 28, 12, 0, }, /* 218 */ + { 28, 13, 0, }, /* 219 */ + { 36, 10, 0, }, /* 220 */ + { 36, 7, 0, }, /* 221 */ + { 36, 12, 0, }, /* 222 */ + { 36, 13, 0, }, /* 223 */ + { 36, 15, 0, }, /* 224 */ + { 36, 26, 0, }, /* 225 */ + { 47, 10, 0, }, /* 226 */ + { 47, 7, 0, }, /* 227 */ + { 47, 12, 0, }, /* 228 */ + { 47, 21, 0, }, /* 229 */ + { 56, 7, 0, }, /* 230 */ + { 56, 12, 0, }, /* 231 */ + { 56, 6, 0, }, /* 232 */ + { 56, 21, 0, }, /* 233 */ + { 56, 13, 0, }, /* 234 */ + { 32, 7, 0, }, /* 235 */ + { 32, 12, 0, }, /* 236 */ + { 32, 6, 0, }, /* 237 */ + { 32, 13, 0, }, /* 238 */ + { 57, 7, 0, }, /* 239 */ + { 57, 26, 0, }, /* 240 */ + { 57, 21, 0, }, /* 241 */ + { 57, 12, 0, }, /* 242 */ + { 57, 13, 0, }, /* 243 */ + { 57, 15, 0, }, /* 244 */ + { 57, 22, 0, }, /* 245 */ + { 57, 18, 0, }, /* 246 */ + { 57, 10, 0, }, /* 247 */ + { 38, 7, 0, }, /* 248 */ + { 38, 10, 0, }, /* 249 */ + { 38, 12, 0, }, /* 250 */ + { 38, 13, 0, }, /* 251 */ + { 38, 21, 0, }, /* 252 */ + { 38, 26, 0, }, /* 253 */ + { 16, 9, 7264, }, /* 254 */ + { 16, 7, 0, }, /* 255 */ + { 16, 6, 0, }, /* 256 */ + { 23, 7, 0, }, /* 257 */ + { 15, 7, 0, }, /* 258 */ + { 15, 12, 0, }, /* 259 */ + { 15, 21, 0, }, /* 260 */ + { 15, 15, 0, }, /* 261 */ + { 15, 26, 0, }, /* 262 */ + { 8, 7, 0, }, /* 263 */ + { 7, 17, 0, }, /* 264 */ + { 7, 7, 0, }, /* 265 */ + { 7, 21, 0, }, /* 266 */ + { 40, 29, 0, }, /* 267 */ + { 40, 7, 0, }, /* 268 */ + { 40, 22, 0, }, /* 269 */ + { 40, 18, 0, }, /* 270 */ + { 45, 7, 0, }, /* 271 */ + { 45, 14, 0, }, /* 272 */ + { 50, 7, 0, }, /* 273 */ + { 50, 12, 0, }, /* 274 */ + { 24, 7, 0, }, /* 275 */ + { 24, 12, 0, }, /* 276 */ + { 6, 7, 0, }, /* 277 */ + { 6, 12, 0, }, /* 278 */ + { 51, 7, 0, }, /* 279 */ + { 51, 12, 0, }, /* 280 */ + { 31, 7, 0, }, /* 281 */ + { 31, 12, 0, }, /* 282 */ + { 31, 10, 0, }, /* 283 */ + { 31, 21, 0, }, /* 284 */ + { 31, 6, 0, }, /* 285 */ + { 31, 23, 0, }, /* 286 */ + { 31, 13, 0, }, /* 287 */ + { 31, 15, 0, }, /* 288 */ + { 37, 21, 0, }, /* 289 */ + { 37, 17, 0, }, /* 290 */ + { 37, 12, 0, }, /* 291 */ + { 37, 29, 0, }, /* 292 */ + { 37, 13, 0, }, /* 293 */ + { 37, 7, 0, }, /* 294 */ + { 37, 6, 0, }, /* 295 */ + { 34, 7, 0, }, /* 296 */ + { 34, 12, 0, }, /* 297 */ + { 34, 10, 0, }, /* 298 */ + { 34, 26, 0, }, /* 299 */ + { 34, 21, 0, }, /* 300 */ + { 34, 13, 0, }, /* 301 */ + { 52, 7, 0, }, /* 302 */ + { 39, 7, 0, }, /* 303 */ + { 39, 10, 0, }, /* 304 */ + { 39, 13, 0, }, /* 305 */ + { 39, 15, 0, }, /* 306 */ + { 39, 26, 0, }, /* 307 */ + { 31, 26, 0, }, /* 308 */ + { 5, 7, 0, }, /* 309 */ + { 5, 12, 0, }, /* 310 */ + { 5, 10, 0, }, /* 311 */ + { 5, 21, 0, }, /* 312 */ + { 90, 7, 0, }, /* 313 */ + { 90, 10, 0, }, /* 314 */ + { 90, 12, 0, }, /* 315 */ + { 90, 13, 0, }, /* 316 */ + { 90, 21, 0, }, /* 317 */ + { 90, 6, 0, }, /* 318 */ + { 61, 12, 0, }, /* 319 */ + { 61, 10, 0, }, /* 320 */ + { 61, 7, 0, }, /* 321 */ + { 61, 13, 0, }, /* 322 */ + { 61, 21, 0, }, /* 323 */ + { 61, 26, 0, }, /* 324 */ + { 75, 12, 0, }, /* 325 */ + { 75, 10, 0, }, /* 326 */ + { 75, 7, 0, }, /* 327 */ + { 75, 13, 0, }, /* 328 */ + { 92, 7, 0, }, /* 329 */ + { 92, 12, 0, }, /* 330 */ + { 92, 10, 0, }, /* 331 */ + { 92, 21, 0, }, /* 332 */ + { 69, 7, 0, }, /* 333 */ + { 69, 10, 0, }, /* 334 */ + { 69, 12, 0, }, /* 335 */ + { 69, 21, 0, }, /* 336 */ + { 69, 13, 0, }, /* 337 */ + { 72, 13, 0, }, /* 338 */ + { 72, 7, 0, }, /* 339 */ + { 72, 6, 0, }, /* 340 */ + { 72, 21, 0, }, /* 341 */ + { 75, 21, 0, }, /* 342 */ + { 9, 10, 0, }, /* 343 */ + { 9, 7, 0, }, /* 344 */ + { 12, 5, 0, }, /* 345 */ + { 12, 6, 0, }, /* 346 */ + { 33, 5, 35332, }, /* 347 */ + { 33, 5, 3814, }, /* 348 */ + { 33, 5, -59, }, /* 349 */ + { 33, 9, -7615, }, /* 350 */ + { 19, 5, 8, }, /* 351 */ + { 19, 9, -8, }, /* 352 */ + { 19, 5, 74, }, /* 353 */ + { 19, 5, 86, }, /* 354 */ + { 19, 5, 100, }, /* 355 */ + { 19, 5, 128, }, /* 356 */ + { 19, 5, 112, }, /* 357 */ + { 19, 5, 126, }, /* 358 */ + { 19, 8, -8, }, /* 359 */ + { 19, 5, 9, }, /* 360 */ + { 19, 9, -74, }, /* 361 */ + { 19, 8, -9, }, /* 362 */ + { 19, 5, -7205, }, /* 363 */ + { 19, 9, -86, }, /* 364 */ + { 19, 9, -100, }, /* 365 */ + { 19, 9, -112, }, /* 366 */ + { 19, 9, -128, }, /* 367 */ + { 19, 9, -126, }, /* 368 */ + { 27, 1, 0, }, /* 369 */ + { 9, 27, 0, }, /* 370 */ + { 9, 28, 0, }, /* 371 */ + { 27, 11, 0, }, /* 372 */ + { 9, 9, 0, }, /* 373 */ + { 9, 5, 0, }, /* 374 */ + { 19, 9, -7517, }, /* 375 */ + { 33, 9, -8383, }, /* 376 */ + { 33, 9, -8262, }, /* 377 */ + { 33, 9, 28, }, /* 378 */ + { 33, 5, -28, }, /* 379 */ + { 33, 14, 16, }, /* 380 */ + { 33, 14, -16, }, /* 381 */ + { 33, 14, 0, }, /* 382 */ + { 9, 26, 26, }, /* 383 */ + { 9, 26, -26, }, /* 384 */ + { 4, 26, 0, }, /* 385 */ + { 17, 9, 48, }, /* 386 */ + { 17, 5, -48, }, /* 387 */ + { 33, 9, -10743, }, /* 388 */ + { 33, 9, -3814, }, /* 389 */ + { 33, 9, -10727, }, /* 390 */ + { 33, 5, -10795, }, /* 391 */ + { 33, 5, -10792, }, /* 392 */ + { 33, 9, -10780, }, /* 393 */ + { 33, 9, -10749, }, /* 394 */ + { 33, 9, -10783, }, /* 395 */ + { 33, 9, -10782, }, /* 396 */ + { 33, 9, -10815, }, /* 397 */ + { 10, 5, 0, }, /* 398 */ + { 10, 26, 0, }, /* 399 */ + { 10, 12, 0, }, /* 400 */ + { 10, 21, 0, }, /* 401 */ + { 10, 15, 0, }, /* 402 */ + { 16, 5, -7264, }, /* 403 */ + { 58, 7, 0, }, /* 404 */ + { 58, 6, 0, }, /* 405 */ + { 58, 21, 0, }, /* 406 */ + { 58, 12, 0, }, /* 407 */ + { 22, 26, 0, }, /* 408 */ + { 22, 6, 0, }, /* 409 */ + { 22, 14, 0, }, /* 410 */ + { 23, 10, 0, }, /* 411 */ + { 26, 7, 0, }, /* 412 */ + { 26, 6, 0, }, /* 413 */ + { 29, 7, 0, }, /* 414 */ + { 29, 6, 0, }, /* 415 */ + { 3, 7, 0, }, /* 416 */ + { 23, 26, 0, }, /* 417 */ + { 29, 26, 0, }, /* 418 */ + { 22, 7, 0, }, /* 419 */ + { 60, 7, 0, }, /* 420 */ + { 60, 6, 0, }, /* 421 */ + { 60, 26, 0, }, /* 422 */ + { 85, 7, 0, }, /* 423 */ + { 85, 6, 0, }, /* 424 */ + { 85, 21, 0, }, /* 425 */ + { 76, 7, 0, }, /* 426 */ + { 76, 6, 0, }, /* 427 */ + { 76, 21, 0, }, /* 428 */ + { 76, 13, 0, }, /* 429 */ + { 12, 7, 0, }, /* 430 */ + { 12, 21, 0, }, /* 431 */ + { 78, 7, 0, }, /* 432 */ + { 78, 14, 0, }, /* 433 */ + { 78, 12, 0, }, /* 434 */ + { 78, 21, 0, }, /* 435 */ + { 33, 9, -35332, }, /* 436 */ + { 33, 9, -42280, }, /* 437 */ + { 33, 9, -42308, }, /* 438 */ + { 48, 7, 0, }, /* 439 */ + { 48, 12, 0, }, /* 440 */ + { 48, 10, 0, }, /* 441 */ + { 48, 26, 0, }, /* 442 */ + { 64, 7, 0, }, /* 443 */ + { 64, 21, 0, }, /* 444 */ + { 74, 10, 0, }, /* 445 */ + { 74, 7, 0, }, /* 446 */ + { 74, 12, 0, }, /* 447 */ + { 74, 21, 0, }, /* 448 */ + { 74, 13, 0, }, /* 449 */ + { 68, 13, 0, }, /* 450 */ + { 68, 7, 0, }, /* 451 */ + { 68, 12, 0, }, /* 452 */ + { 68, 21, 0, }, /* 453 */ + { 73, 7, 0, }, /* 454 */ + { 73, 12, 0, }, /* 455 */ + { 73, 10, 0, }, /* 456 */ + { 73, 21, 0, }, /* 457 */ + { 83, 12, 0, }, /* 458 */ + { 83, 10, 0, }, /* 459 */ + { 83, 7, 0, }, /* 460 */ + { 83, 21, 0, }, /* 461 */ + { 83, 6, 0, }, /* 462 */ + { 83, 13, 0, }, /* 463 */ + { 67, 7, 0, }, /* 464 */ + { 67, 12, 0, }, /* 465 */ + { 67, 10, 0, }, /* 466 */ + { 67, 13, 0, }, /* 467 */ + { 67, 21, 0, }, /* 468 */ + { 38, 6, 0, }, /* 469 */ + { 91, 7, 0, }, /* 470 */ + { 91, 12, 0, }, /* 471 */ + { 91, 6, 0, }, /* 472 */ + { 91, 21, 0, }, /* 473 */ + { 86, 7, 0, }, /* 474 */ + { 86, 10, 0, }, /* 475 */ + { 86, 12, 0, }, /* 476 */ + { 86, 21, 0, }, /* 477 */ + { 86, 6, 0, }, /* 478 */ + { 86, 13, 0, }, /* 479 */ + { 9, 4, 0, }, /* 480 */ + { 9, 3, 0, }, /* 481 */ + { 25, 25, 0, }, /* 482 */ + { 0, 24, 0, }, /* 483 */ + { 35, 7, 0, }, /* 484 */ + { 19, 14, 0, }, /* 485 */ + { 19, 15, 0, }, /* 486 */ + { 19, 26, 0, }, /* 487 */ + { 70, 7, 0, }, /* 488 */ + { 66, 7, 0, }, /* 489 */ + { 41, 7, 0, }, /* 490 */ + { 41, 15, 0, }, /* 491 */ + { 18, 7, 0, }, /* 492 */ + { 18, 14, 0, }, /* 493 */ + { 59, 7, 0, }, /* 494 */ + { 59, 21, 0, }, /* 495 */ + { 42, 7, 0, }, /* 496 */ + { 42, 21, 0, }, /* 497 */ + { 42, 14, 0, }, /* 498 */ + { 13, 9, 40, }, /* 499 */ + { 13, 5, -40, }, /* 500 */ + { 46, 7, 0, }, /* 501 */ + { 44, 7, 0, }, /* 502 */ + { 44, 13, 0, }, /* 503 */ + { 11, 7, 0, }, /* 504 */ + { 80, 7, 0, }, /* 505 */ + { 80, 21, 0, }, /* 506 */ + { 80, 15, 0, }, /* 507 */ + { 65, 7, 0, }, /* 508 */ + { 65, 15, 0, }, /* 509 */ + { 65, 21, 0, }, /* 510 */ + { 71, 7, 0, }, /* 511 */ + { 71, 21, 0, }, /* 512 */ + { 97, 7, 0, }, /* 513 */ + { 96, 7, 0, }, /* 514 */ + { 30, 7, 0, }, /* 515 */ + { 30, 12, 0, }, /* 516 */ + { 30, 15, 0, }, /* 517 */ + { 30, 21, 0, }, /* 518 */ + { 87, 7, 0, }, /* 519 */ + { 87, 15, 0, }, /* 520 */ + { 87, 21, 0, }, /* 521 */ + { 77, 7, 0, }, /* 522 */ + { 77, 21, 0, }, /* 523 */ + { 82, 7, 0, }, /* 524 */ + { 82, 15, 0, }, /* 525 */ + { 81, 7, 0, }, /* 526 */ + { 81, 15, 0, }, /* 527 */ + { 88, 7, 0, }, /* 528 */ + { 0, 15, 0, }, /* 529 */ + { 93, 10, 0, }, /* 530 */ + { 93, 12, 0, }, /* 531 */ + { 93, 7, 0, }, /* 532 */ + { 93, 21, 0, }, /* 533 */ + { 93, 15, 0, }, /* 534 */ + { 93, 13, 0, }, /* 535 */ + { 84, 12, 0, }, /* 536 */ + { 84, 10, 0, }, /* 537 */ + { 84, 7, 0, }, /* 538 */ + { 84, 21, 0, }, /* 539 */ + { 84, 1, 0, }, /* 540 */ + { 100, 7, 0, }, /* 541 */ + { 100, 13, 0, }, /* 542 */ + { 95, 12, 0, }, /* 543 */ + { 95, 7, 0, }, /* 544 */ + { 95, 10, 0, }, /* 545 */ + { 95, 13, 0, }, /* 546 */ + { 95, 21, 0, }, /* 547 */ + { 99, 12, 0, }, /* 548 */ + { 99, 10, 0, }, /* 549 */ + { 99, 7, 0, }, /* 550 */ + { 99, 21, 0, }, /* 551 */ + { 99, 13, 0, }, /* 552 */ + { 101, 7, 0, }, /* 553 */ + { 101, 12, 0, }, /* 554 */ + { 101, 10, 0, }, /* 555 */ + { 101, 13, 0, }, /* 556 */ + { 62, 7, 0, }, /* 557 */ + { 62, 14, 0, }, /* 558 */ + { 62, 21, 0, }, /* 559 */ + { 79, 7, 0, }, /* 560 */ + { 98, 7, 0, }, /* 561 */ + { 98, 10, 0, }, /* 562 */ + { 98, 12, 0, }, /* 563 */ + { 98, 6, 0, }, /* 564 */ + { 19, 12, 0, }, /* 565 */ + { 26, 26, 0, }, /* 566 */ +}; + +const pcre_uint8 PRIV(ucd_stage1)[] = { /* 8704 bytes */ + 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, /* U+0000 */ + 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, /* U+0800 */ + 32, 33, 34, 34, 35, 36, 37, 38, 39, 40, 40, 40, 41, 42, 43, 44, /* U+1000 */ + 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, /* U+1800 */ + 61, 62, 63, 64, 65, 65, 66, 67, 68, 69, 70, 71, 72, 70, 73, 74, /* U+2000 */ + 75, 75, 65, 76, 65, 65, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, /* U+2800 */ + 87, 88, 89, 90, 91, 92, 93, 70, 94, 94, 94, 94, 94, 94, 94, 94, /* U+3000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+3800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+4000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 95, 94, 94, 94, 94, /* U+4800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+5000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+5800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+6000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+6800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+7000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+7800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+8000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+8800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+9000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 96, /* U+9800 */ + 97, 98, 98, 98, 98, 98, 98, 98, 98, 99,100,100,101,102,103,104, /* U+A000 */ +105,106,107,108,109,110,111,112, 34, 34, 34, 34, 34, 34, 34, 34, /* U+A800 */ + 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, /* U+B000 */ + 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, /* U+B800 */ + 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, /* U+C000 */ + 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, /* U+C800 */ + 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34,113, /* U+D000 */ +114,114,114,114,114,114,114,114,114,114,114,114,114,114,114,114, /* U+D800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+E000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+E800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F000 */ +115,115, 94, 94,116,117,118,119,120,120,121,122,123,124,125,126, /* U+F800 */ +127,128,129,130, 78,131,132,133,134,135, 78, 78, 78, 78, 78, 78, /* U+10000 */ +136, 78,137,138,139, 78,140, 78,141, 78, 78, 78,142, 78, 78, 78, /* U+10800 */ +143,144,145,146, 78, 78, 78, 78, 78, 78, 78, 78, 78,147, 78, 78, /* U+11000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+11800 */ +148,148,148,148,148,148,149, 78,150, 78, 78, 78, 78, 78, 78, 78, /* U+12000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+12800 */ +151,151,151,151,151,151,151,151,152, 78, 78, 78, 78, 78, 78, 78, /* U+13000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+13800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+14000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+14800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+15000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+15800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+16000 */ +153,153,153,153,154, 78, 78, 78, 78, 78, 78, 78, 78, 78,155,156, /* U+16800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+17000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+17800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+18000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+18800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+19000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+19800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+1A000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+1A800 */ +157, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+1B000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+1B800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+1C000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+1C800 */ + 70,158,159,160,161, 78,162, 78,163,164,165,166,167,168,169,170, /* U+1D000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+1D800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+1E000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78,171,172, 78, 78, /* U+1E800 */ +173,174,175,176,177, 78,178,179,180,181,182,183,184,185,186, 78, /* U+1F000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+1F800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+20000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+20800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+21000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+21800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+22000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+22800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+23000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+23800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+24000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+24800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+25000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+25800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+26000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+26800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+27000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+27800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+28000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+28800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+29000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+29800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94,187, 94, 94, /* U+2A000 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, /* U+2A800 */ + 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94, 94,188, 94, /* U+2B000 */ +189, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+2B800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+2C000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+2C800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+2D000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+2D800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+2E000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+2E800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+2F000 */ + 94, 94, 94, 94,189, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+2F800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+30000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+30800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+31000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+31800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+32000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+32800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+33000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+33800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+34000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+34800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+35000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+35800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+36000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+36800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+37000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+37800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+38000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+38800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+39000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+39800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+3A000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+3A800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+3B000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+3B800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+3C000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+3C800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+3D000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+3D800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+3E000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+3E800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+3F000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+3F800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+40000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+40800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+41000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+41800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+42000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+42800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+43000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+43800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+44000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+44800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+45000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+45800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+46000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+46800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+47000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+47800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+48000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+48800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+49000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+49800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+4A000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+4A800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+4B000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+4B800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+4C000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+4C800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+4D000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+4D800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+4E000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+4E800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+4F000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+4F800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+50000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+50800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+51000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+51800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+52000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+52800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+53000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+53800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+54000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+54800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+55000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+55800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+56000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+56800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+57000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+57800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+58000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+58800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+59000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+59800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+5A000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+5A800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+5B000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+5B800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+5C000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+5C800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+5D000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+5D800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+5E000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+5E800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+5F000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+5F800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+60000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+60800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+61000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+61800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+62000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+62800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+63000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+63800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+64000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+64800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+65000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+65800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+66000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+66800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+67000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+67800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+68000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+68800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+69000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+69800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+6A000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+6A800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+6B000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+6B800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+6C000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+6C800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+6D000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+6D800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+6E000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+6E800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+6F000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+6F800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+70000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+70800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+71000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+71800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+72000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+72800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+73000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+73800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+74000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+74800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+75000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+75800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+76000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+76800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+77000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+77800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+78000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+78800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+79000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+79800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+7A000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+7A800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+7B000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+7B800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+7C000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+7C800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+7D000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+7D800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+7E000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+7E800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+7F000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+7F800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+80000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+80800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+81000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+81800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+82000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+82800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+83000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+83800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+84000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+84800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+85000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+85800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+86000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+86800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+87000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+87800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+88000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+88800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+89000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+89800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+8A000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+8A800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+8B000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+8B800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+8C000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+8C800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+8D000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+8D800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+8E000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+8E800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+8F000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+8F800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+90000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+90800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+91000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+91800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+92000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+92800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+93000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+93800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+94000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+94800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+95000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+95800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+96000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+96800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+97000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+97800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+98000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+98800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+99000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+99800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+9A000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+9A800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+9B000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+9B800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+9C000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+9C800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+9D000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+9D800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+9E000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+9E800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+9F000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+9F800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A0000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A0800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A1000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A1800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A2000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A2800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A3000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A3800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A4000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A4800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A5000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A5800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A6000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A6800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A7000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A7800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A8000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A8800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A9000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+A9800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+AA000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+AA800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+AB000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+AB800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+AC000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+AC800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+AD000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+AD800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+AE000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+AE800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+AF000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+AF800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B0000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B0800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B1000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B1800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B2000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B2800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B3000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B3800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B4000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B4800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B5000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B5800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B6000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B6800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B7000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B7800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B8000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B8800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B9000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+B9800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+BA000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+BA800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+BB000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+BB800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+BC000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+BC800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+BD000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+BD800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+BE000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+BE800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+BF000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+BF800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C0000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C0800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C1000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C1800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C2000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C2800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C3000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C3800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C4000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C4800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C5000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C5800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C6000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C6800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C7000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C7800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C8000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C8800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C9000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+C9800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+CA000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+CA800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+CB000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+CB800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+CC000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+CC800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+CD000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+CD800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+CE000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+CE800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+CF000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+CF800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D0000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D0800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D1000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D1800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D2000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D2800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D3000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D3800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D4000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D4800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D5000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D5800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D6000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D6800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D7000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D7800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D8000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D8800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D9000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+D9800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+DA000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+DA800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+DB000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+DB800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+DC000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+DC800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+DD000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+DD800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+DE000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+DE800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+DF000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+DF800 */ +190, 78,191,192, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E0000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E0800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E1000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E1800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E2000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E2800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E3000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E3800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E4000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E4800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E5000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E5800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E6000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E6800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E7000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E7800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E8000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E8800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E9000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+E9800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+EA000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+EA800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+EB000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+EB800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+EC000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+EC800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+ED000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+ED800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+EE000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+EE800 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+EF000 */ + 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, /* U+EF800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F0000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F0800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F1000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F1800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F2000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F2800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F3000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F3800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F4000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F4800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F5000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F5800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F6000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F6800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F7000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F7800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F8000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F8800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F9000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+F9800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+FA000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+FA800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+FB000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+FB800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+FC000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+FC800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+FD000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+FD800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+FE000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+FE800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+FF000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,193, /* U+FF800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+100000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+100800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+101000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+101800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+102000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+102800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+103000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+103800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+104000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+104800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+105000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+105800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+106000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+106800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+107000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+107800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+108000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+108800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+109000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+109800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+10A000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+10A800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+10B000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+10B800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+10C000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+10C800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+10D000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+10D800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+10E000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+10E800 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,115, /* U+10F000 */ +115,115,115,115,115,115,115,115,115,115,115,115,115,115,115,193, /* U+10F800 */ +}; + +const pcre_uint16 PRIV(ucd_stage2)[] = { /* 49664 bytes, block = 128 */ +/* block 0 */ + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 1, 2, 2, 2, 3, 2, 2, 2, 4, 5, 2, 6, 2, 7, 2, 2, + 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2, 2, 6, 6, 6, 2, + 2, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, + 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 4, 2, 5, 10, 11, + 10, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, + 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 4, 6, 5, 6, 0, + +/* block 1 */ + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 1, 2, 3, 3, 3, 3, 13, 2, 10, 13, 14, 15, 6, 16, 13, 10, + 13, 6, 17, 17, 10, 18, 2, 2, 10, 17, 14, 19, 17, 17, 17, 2, + 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, + 9, 9, 9, 9, 9, 9, 9, 6, 9, 9, 9, 9, 9, 9, 9, 20, + 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, + 12, 12, 12, 12, 12, 12, 12, 6, 12, 12, 12, 12, 12, 12, 12, 21, + +/* block 2 */ + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 24, 25, 22, 23, 22, 23, 22, 23, 20, 22, 23, 22, 23, 22, 23, 22, + 23, 22, 23, 22, 23, 22, 23, 22, 23, 20, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 26, 22, 23, 22, 23, 22, 23, 27, + +/* block 3 */ + 28, 29, 22, 23, 22, 23, 30, 22, 23, 31, 31, 22, 23, 20, 32, 33, + 34, 22, 23, 31, 35, 36, 37, 38, 22, 23, 39, 20, 37, 40, 41, 42, + 22, 23, 22, 23, 22, 23, 43, 22, 23, 43, 20, 20, 22, 23, 43, 22, + 23, 44, 44, 22, 23, 22, 23, 45, 22, 23, 20, 14, 22, 23, 20, 46, + 14, 14, 14, 14, 47, 48, 49, 47, 48, 49, 47, 48, 49, 22, 23, 22, + 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 50, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 20, 47, 48, 49, 22, 23, 51, 52, 22, 23, 22, 23, 22, 23, 22, 23, + +/* block 4 */ + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 53, 20, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 20, 20, 20, 20, 20, 20, 54, 22, 23, 55, 56, 57, + 57, 22, 23, 58, 59, 60, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 61, 62, 63, 64, 65, 20, 66, 66, 20, 67, 20, 68, 20, 20, 20, 20, + 66, 20, 20, 69, 20, 70, 71, 20, 72, 73, 20, 74, 20, 20, 20, 73, + 20, 75, 76, 20, 20, 77, 20, 20, 20, 20, 20, 20, 20, 78, 20, 20, + +/* block 5 */ + 79, 20, 20, 79, 20, 20, 20, 20, 79, 80, 81, 81, 82, 20, 20, 20, + 20, 20, 83, 20, 14, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, + 84, 84, 84, 84, 84, 84, 84, 84, 84, 85, 85, 85, 85, 85, 85, 85, + 85, 85, 10, 10, 10, 10, 85, 85, 85, 85, 85, 85, 85, 85, 85, 85, + 85, 85, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, + 84, 84, 84, 84, 84, 10, 10, 10, 10, 10, 86, 86, 85, 10, 85, 10, + 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, + +/* block 6 */ + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 88, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 89, 90, 89, 90, 85, 91, 89, 90, 92, 92, 93, 94, 94, 94, 2, 92, + +/* block 7 */ + 92, 92, 92, 92, 91, 10, 95, 2, 96, 96, 96, 92, 97, 92, 98, 98, + 99,100,100,100,100,100,100,100,100,100,100,100,100,100,100,100, +100,100, 92,100,100,100,100,100,100,100,100,100,101,102,102,102, + 99,103,103,103,103,103,103,103,103,103,103,103,103,103,103,103, +103,103,104,103,103,103,103,103,103,103,103,103,105,106,106,107, +108,109,110,110,110,111,112,113, 89, 90, 89, 90, 89, 90, 89, 90, + 89, 90,114,115,114,115,114,115,114,115,114,115,114,115,114,115, +116,117,118, 99,119,120,121, 89, 90,122, 89, 90, 99,123,123,123, + +/* block 8 */ +124,124,124,124,124,124,124,124,124,124,124,124,124,124,124,124, +125,125,125,125,125,125,125,125,125,125,125,125,125,125,125,125, +125,125,125,125,125,125,125,125,125,125,125,125,125,125,125,125, +126,126,126,126,126,126,126,126,126,126,126,126,126,126,126,126, +126,126,126,126,126,126,126,126,126,126,126,126,126,126,126,126, +127,127,127,127,127,127,127,127,127,127,127,127,127,127,127,127, +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, + +/* block 9 */ +128,129,130,131,131, 87, 87,131,132,132,128,129,128,129,128,129, +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, +133,128,129,128,129,128,129,128,129,128,129,128,129,128,129,134, +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, + +/* block 10 */ +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, +128,129,128,129,128,129,128,129, 92, 92, 92, 92, 92, 92, 92, 92, + 92,135,135,135,135,135,135,135,135,135,135,135,135,135,135,135, +135,135,135,135,135,135,135,135,135,135,135,135,135,135,135,135, +135,135,135,135,135,135,135, 92, 92,136,137,137,137,137,137,137, + 92,138,138,138,138,138,138,138,138,138,138,138,138,138,138,138, +138,138,138,138,138,138,138,138,138,138,138,138,138,138,138,138, + +/* block 11 */ +138,138,138,138,138,138,138,139, 92, 2,140, 92, 92, 92, 92,141, + 92,142,142,142,142,142,142,142,142,142,142,142,142,142,142,142, +142,142,142,142,142,142,142,142,142,142,142,142,142,142,142,142, +142,142,142,142,142,142,142,142,142,142,142,142,142,142,143,142, +144,142,142,144,142,142,144,142, 92, 92, 92, 92, 92, 92, 92, 92, +145,145,145,145,145,145,145,145,145,145,145,145,145,145,145,145, +145,145,145,145,145,145,145,145,145,145,145, 92, 92, 92, 92, 92, +145,145,145,144,144, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 12 */ +146,146,146,146,146, 92,147,147,147,148,148,149, 2,148,150,150, +151,151,151,151,151,151,151,151,151,151,151, 2, 92, 92,148, 2, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, + 85,152,152,152,152,152,152,152,152,152,152, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87,151,151,151,151,151,151,151,151,151, 87, + 8, 8, 8, 8, 8, 8, 8, 8, 8, 8,148,148,148,148,152,152, + 87,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, + +/* block 13 */ +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,148,152,151,151,151,151,151,151,151, 16,150,151, +151,151,151,151,151,153,153,151,151,150,151,151,151,151,152,152, +154,154,154,154,154,154,154,154,154,154,152,152,152,150,150,152, + +/* block 14 */ +155,155,155,155,155,155,155,155,155,155,155,155,155,155, 92,156, +157,158,157,157,157,157,157,157,157,157,157,157,157,157,157,157, +157,157,157,157,157,157,157,157,157,157,157,157,157,157,157,157, +158,158,158,158,158,158,158,158,158,158,158,158,158,158,158,158, +158,158,158,158,158,158,158,158,158,158,158, 92, 92,157,157,157, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, + +/* block 15 */ +159,159,159,159,159,159,159,159,159,159,159,159,159,159,159,159, +159,159,159,159,159,159,159,159,159,159,159,159,159,159,159,159, +159,159,159,159,159,159,160,160,160,160,160,160,160,160,160,160, +160,159, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +161,161,161,161,161,161,161,161,161,161,162,162,162,162,162,162, +162,162,162,162,162,162,162,162,162,162,162,162,162,162,162,162, +162,162,162,162,162,162,162,162,162,162,162,163,163,163,163,163, +163,163,163,163,164,164,165,166,166,166,164, 92, 92, 92, 92, 92, + +/* block 16 */ +167,167,167,167,167,167,167,167,167,167,167,167,167,167,167,167, +167,167,167,167,167,167,168,168,168,168,169,168,168,168,168,168, +168,168,168,168,169,168,168,168,169,168,168,168,168,168, 92, 92, +170,170,170,170,170,170,170,170,170,170,170,170,170,170,170, 92, +171,171,171,171,171,171,171,171,171,171,171,171,171,171,171,171, +171,171,171,171,171,171,171,171,171,172,172,172, 92, 92,173, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 17 */ + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +152, 92,152,152,152,152,152,152,152,152,152,152,152, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92,151,151,151,151,151,151,151,151,151,151,151,151, +151,151,151,151,151,151,151,151,151,151,151,151,151,151,151, 92, + +/* block 18 */ +174,174,174,175,176,176,176,176,176,176,176,176,176,176,176,176, +176,176,176,176,176,176,176,176,176,176,176,176,176,176,176,176, +176,176,176,176,176,176,176,176,176,176,176,176,176,176,176,176, +176,176,176,176,176,176,176,176,176,176,174,175,174,176,175,175, +175,174,174,174,174,174,174,174,174,175,175,175,175,174,175,175, +176, 87, 87,174,174,174,174,174,176,176,176,176,176,176,176,176, +176,176,174,174, 2, 2,177,177,177,177,177,177,177,177,177,177, +178,179,176,176,176,176,176,176, 92,176,176,176,176,176,176,176, + +/* block 19 */ + 92,180,181,181, 92,182,182,182,182,182,182,182,182, 92, 92,182, +182, 92, 92,182,182,182,182,182,182,182,182,182,182,182,182,182, +182,182,182,182,182,182,182,182,182, 92,182,182,182,182,182,182, +182, 92,182, 92, 92, 92,182,182,182,182, 92, 92,180,182,181,181, +181,180,180,180,180, 92, 92,181,181, 92, 92,181,181,180,182, 92, + 92, 92, 92, 92, 92, 92, 92,181, 92, 92, 92, 92,182,182, 92,182, +182,182,180,180, 92, 92,183,183,183,183,183,183,183,183,183,183, +182,182,184,184,185,185,185,185,185,185,186,184, 92, 92, 92, 92, + +/* block 20 */ + 92,187,187,188, 92,189,189,189,189,189,189, 92, 92, 92, 92,189, +189, 92, 92,189,189,189,189,189,189,189,189,189,189,189,189,189, +189,189,189,189,189,189,189,189,189, 92,189,189,189,189,189,189, +189, 92,189,189, 92,189,189, 92,189,189, 92, 92,187, 92,188,188, +188,187,187, 92, 92, 92, 92,187,187, 92, 92,187,187,187, 92, 92, + 92,187, 92, 92, 92, 92, 92, 92, 92,189,189,189,189, 92,189, 92, + 92, 92, 92, 92, 92, 92,190,190,190,190,190,190,190,190,190,190, +187,187,189,189,189,187, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 21 */ + 92,191,191,192, 92,193,193,193,193,193,193,193,193,193, 92,193, +193,193, 92,193,193,193,193,193,193,193,193,193,193,193,193,193, +193,193,193,193,193,193,193,193,193, 92,193,193,193,193,193,193, +193, 92,193,193, 92,193,193,193,193,193, 92, 92,191,193,192,192, +192,191,191,191,191,191, 92,191,191,192, 92,192,192,191, 92, 92, +193, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +193,193,191,191, 92, 92,194,194,194,194,194,194,194,194,194,194, +195,196, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 22 */ + 92,197,198,198, 92,199,199,199,199,199,199,199,199, 92, 92,199, +199, 92, 92,199,199,199,199,199,199,199,199,199,199,199,199,199, +199,199,199,199,199,199,199,199,199, 92,199,199,199,199,199,199, +199, 92,199,199, 92,199,199,199,199,199, 92, 92,197,199,198,197, +198,197,197,197,197, 92, 92,198,198, 92, 92,198,198,197, 92, 92, + 92, 92, 92, 92, 92, 92,197,198, 92, 92, 92, 92,199,199, 92,199, +199,199,197,197, 92, 92,200,200,200,200,200,200,200,200,200,200, +201,199,202,202,202,202,202,202, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 23 */ + 92, 92,203,204, 92,204,204,204,204,204,204, 92, 92, 92,204,204, +204, 92,204,204,204,204, 92, 92, 92,204,204, 92,204, 92,204,204, + 92, 92, 92,204,204, 92, 92, 92,204,204,204, 92, 92, 92,204,204, +204,204,204,204,204,204,204,204,204,204, 92, 92, 92, 92,205,205, +203,205,205, 92, 92, 92,205,205,205, 92,205,205,205,203, 92, 92, +204, 92, 92, 92, 92, 92, 92,205, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92,206,206,206,206,206,206,206,206,206,206, +207,207,207,208,208,208,208,208,208,209,208, 92, 92, 92, 92, 92, + +/* block 24 */ + 92,210,210,210, 92,211,211,211,211,211,211,211,211, 92,211,211, +211, 92,211,211,211,211,211,211,211,211,211,211,211,211,211,211, +211,211,211,211,211,211,211,211,211, 92,211,211,211,211,211,211, +211,211,211,211, 92,211,211,211,211,211, 92, 92, 92,211,212,212, +212,210,210,210,210, 92,212,212,212, 92,212,212,212,212, 92, 92, + 92, 92, 92, 92, 92,212,212, 92,211,211, 92, 92, 92, 92, 92, 92, +211,211,212,212, 92, 92,213,213,213,213,213,213,213,213,213,213, + 92, 92, 92, 92, 92, 92, 92, 92,214,214,214,214,214,214,214,215, + +/* block 25 */ + 92, 92,216,216, 92,217,217,217,217,217,217,217,217, 92,217,217, +217, 92,217,217,217,217,217,217,217,217,217,217,217,217,217,217, +217,217,217,217,217,217,217,217,217, 92,217,217,217,217,217,217, +217,217,217,217, 92,217,217,217,217,217, 92, 92,218,217,216,218, +216,216,216,216,216, 92,218,216,216, 92,216,216,218,218, 92, 92, + 92, 92, 92, 92, 92,216,216, 92, 92, 92, 92, 92, 92, 92,217, 92, +217,217,218,218, 92, 92,219,219,219,219,219,219,219,219,219,219, + 92,217,217, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 26 */ + 92, 92,220,220, 92,221,221,221,221,221,221,221,221, 92,221,221, +221, 92,221,221,221,221,221,221,221,221,221,221,221,221,221,221, +221,221,221,221,221,221,221,221,221,221,221,221,221,221,221,221, +221,221,221,221,221,221,221,221,221,221,221, 92, 92,221,220,220, +220,222,222,222,222, 92,220,220,220, 92,220,220,220,222,221, 92, + 92, 92, 92, 92, 92, 92, 92,220, 92, 92, 92, 92, 92, 92, 92, 92, +221,221,222,222, 92, 92,223,223,223,223,223,223,223,223,223,223, +224,224,224,224,224,224, 92, 92, 92,225,221,221,221,221,221,221, + +/* block 27 */ + 92, 92,226,226, 92,227,227,227,227,227,227,227,227,227,227,227, +227,227,227,227,227,227,227, 92, 92, 92,227,227,227,227,227,227, +227,227,227,227,227,227,227,227,227,227,227,227,227,227,227,227, +227,227, 92,227,227,227,227,227,227,227,227,227, 92,227, 92, 92, +227,227,227,227,227,227,227, 92, 92, 92,228, 92, 92, 92, 92,226, +226,226,228,228,228, 92,228, 92,226,226,226,226,226,226,226,226, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92,226,226,229, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 28 */ + 92,230,230,230,230,230,230,230,230,230,230,230,230,230,230,230, +230,230,230,230,230,230,230,230,230,230,230,230,230,230,230,230, +230,230,230,230,230,230,230,230,230,230,230,230,230,230,230,230, +230,231,230,230,231,231,231,231,231,231,231, 92, 92, 92, 92, 3, +230,230,230,230,230,230,232,231,231,231,231,231,231,231,231,233, +234,234,234,234,234,234,234,234,234,234,233,233, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 29 */ + 92,235,235, 92,235, 92, 92,235,235, 92,235, 92, 92,235, 92, 92, + 92, 92, 92, 92,235,235,235,235, 92,235,235,235,235,235,235,235, + 92,235,235,235, 92,235, 92,235, 92, 92,235,235, 92,235,235,235, +235,236,235,235,236,236,236,236,236,236, 92,236,236,235, 92, 92, +235,235,235,235,235, 92,237, 92,236,236,236,236,236,236, 92, 92, +238,238,238,238,238,238,238,238,238,238, 92, 92,235,235,235,235, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 30 */ +239,240,240,240,241,241,241,241,241,241,241,241,241,241,241,241, +241,241,241,240,241,240,240,240,242,242,240,240,240,240,240,240, +243,243,243,243,243,243,243,243,243,243,244,244,244,244,244,244, +244,244,244,244,240,242,240,242,240,242,245,246,245,246,247,247, +239,239,239,239,239,239,239,239, 92,239,239,239,239,239,239,239, +239,239,239,239,239,239,239,239,239,239,239,239,239,239,239,239, +239,239,239,239,239,239,239,239,239,239,239,239,239, 92, 92, 92, + 92,242,242,242,242,242,242,242,242,242,242,242,242,242,242,247, + +/* block 31 */ +242,242,242,242,242,241,242,242,239,239,239,239,239,242,242,242, +242,242,242,242,242,242,242,242, 92,242,242,242,242,242,242,242, +242,242,242,242,242,242,242,242,242,242,242,242,242,242,242,242, +242,242,242,242,242,242,242,242,242,242,242,242,242, 92,240,240, +240,240,240,240,240,240,242,240,240,240,240,240,240, 92,240,240, +241,241,241,241,241, 13, 13, 13, 13,241,241, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 32 */ +248,248,248,248,248,248,248,248,248,248,248,248,248,248,248,248, +248,248,248,248,248,248,248,248,248,248,248,248,248,248,248,248, +248,248,248,248,248,248,248,248,248,248,248,249,249,250,250,250, +250,249,250,250,250,250,250,250,249,250,250,249,249,250,250,248, +251,251,251,251,251,251,251,251,251,251,252,252,252,252,252,252, +248,248,248,248,248,248,249,249,250,250,248,248,248,248,250,250, +250,248,249,249,249,248,248,249,249,249,249,249,249,249,248,248, +248,250,250,250,250,248,248,248,248,248,248,248,248,248,248,248, + +/* block 33 */ +248,248,250,249,249,250,250,249,249,249,249,249,249,250,248,249, +251,251,251,251,251,251,251,251,251,251,249,249,249,250,253,253, +254,254,254,254,254,254,254,254,254,254,254,254,254,254,254,254, +254,254,254,254,254,254,254,254,254,254,254,254,254,254,254,254, +254,254,254,254,254,254, 92,254, 92, 92, 92, 92, 92,254, 92, 92, +255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, +255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, +255,255,255,255,255,255,255,255,255,255,255, 2,256,255,255,255, + +/* block 34 */ +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, + +/* block 35 */ +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258,258,258, 92,258,258,258,258, 92, 92, +258,258,258,258,258,258,258, 92,258, 92,258,258,258,258, 92, 92, +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, + +/* block 36 */ +258,258,258,258,258,258,258,258,258, 92,258,258,258,258, 92, 92, +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258, 92,258,258,258,258, 92, 92,258,258,258,258,258,258,258, 92, +258, 92,258,258,258,258, 92, 92,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258, 92,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, + +/* block 37 */ +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258, 92,258,258,258,258, 92, 92,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258,258,258,258,258, 92, 92,259,259,259, +260,260,260,260,260,260,260,260,260,261,261,261,261,261,261,261, +261,261,261,261,261,261,261,261,261,261,261,261,261, 92, 92, 92, + +/* block 38 */ +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +262,262,262,262,262,262,262,262,262,262, 92, 92, 92, 92, 92, 92, +263,263,263,263,263,263,263,263,263,263,263,263,263,263,263,263, +263,263,263,263,263,263,263,263,263,263,263,263,263,263,263,263, +263,263,263,263,263,263,263,263,263,263,263,263,263,263,263,263, +263,263,263,263,263,263,263,263,263,263,263,263,263,263,263,263, +263,263,263,263,263,263,263,263,263,263,263,263,263,263,263,263, +263,263,263,263,263, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 39 */ +264,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, + +/* block 40 */ +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, + +/* block 41 */ +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,266,266,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, + +/* block 42 */ +267,268,268,268,268,268,268,268,268,268,268,268,268,268,268,268, +268,268,268,268,268,268,268,268,268,268,268,269,270, 92, 92, 92, +271,271,271,271,271,271,271,271,271,271,271,271,271,271,271,271, +271,271,271,271,271,271,271,271,271,271,271,271,271,271,271,271, +271,271,271,271,271,271,271,271,271,271,271,271,271,271,271,271, +271,271,271,271,271,271,271,271,271,271,271,271,271,271,271,271, +271,271,271,271,271,271,271,271,271,271,271, 2, 2, 2,272,272, +272, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 43 */ +273,273,273,273,273,273,273,273,273,273,273,273,273, 92,273,273, +273,273,274,274,274, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +275,275,275,275,275,275,275,275,275,275,275,275,275,275,275,275, +275,275,276,276,276, 2, 2, 92, 92, 92, 92, 92, 92, 92, 92, 92, +277,277,277,277,277,277,277,277,277,277,277,277,277,277,277,277, +277,277,278,278, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +279,279,279,279,279,279,279,279,279,279,279,279,279, 92,279,279, +279, 92,280,280, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 44 */ +281,281,281,281,281,281,281,281,281,281,281,281,281,281,281,281, +281,281,281,281,281,281,281,281,281,281,281,281,281,281,281,281, +281,281,281,281,281,281,281,281,281,281,281,281,281,281,281,281, +281,281,281,281,282,282,283,282,282,282,282,282,282,282,283,283, +283,283,283,283,283,283,282,283,283,282,282,282,282,282,282,282, +282,282,282,282,284,284,284,285,284,284,284,286,281,282, 92, 92, +287,287,287,287,287,287,287,287,287,287, 92, 92, 92, 92, 92, 92, +288,288,288,288,288,288,288,288,288,288, 92, 92, 92, 92, 92, 92, + +/* block 45 */ +289,289, 2, 2,289, 2,290,289,289,289,289,291,291,291,292, 92, +293,293,293,293,293,293,293,293,293,293, 92, 92, 92, 92, 92, 92, +294,294,294,294,294,294,294,294,294,294,294,294,294,294,294,294, +294,294,294,294,294,294,294,294,294,294,294,294,294,294,294,294, +294,294,294,295,294,294,294,294,294,294,294,294,294,294,294,294, +294,294,294,294,294,294,294,294,294,294,294,294,294,294,294,294, +294,294,294,294,294,294,294,294,294,294,294,294,294,294,294,294, +294,294,294,294,294,294,294,294, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 46 */ +294,294,294,294,294,294,294,294,294,294,294,294,294,294,294,294, +294,294,294,294,294,294,294,294,294,294,294,294,294,294,294,294, +294,294,294,294,294,294,294,294,294,291,294, 92, 92, 92, 92, 92, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265,265,265,265,265,265,265,265,265,265,265, +265,265,265,265,265,265, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 47 */ +296,296,296,296,296,296,296,296,296,296,296,296,296,296,296,296, +296,296,296,296,296,296,296,296,296,296,296,296,296, 92, 92, 92, +297,297,297,298,298,298,298,297,297,298,298,298, 92, 92, 92, 92, +298,298,297,298,298,298,298,298,298,297,297,297, 92, 92, 92, 92, +299, 92, 92, 92,300,300,301,301,301,301,301,301,301,301,301,301, +302,302,302,302,302,302,302,302,302,302,302,302,302,302,302,302, +302,302,302,302,302,302,302,302,302,302,302,302,302,302, 92, 92, +302,302,302,302,302, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 48 */ +303,303,303,303,303,303,303,303,303,303,303,303,303,303,303,303, +303,303,303,303,303,303,303,303,303,303,303,303,303,303,303,303, +303,303,303,303,303,303,303,303,303,303,303,303, 92, 92, 92, 92, +304,304,304,304,304,304,304,304,304,304,304,304,304,304,304,304, +304,303,303,303,303,303,303,303,304,304, 92, 92, 92, 92, 92, 92, +305,305,305,305,305,305,305,305,305,305,306, 92, 92, 92,307,307, +308,308,308,308,308,308,308,308,308,308,308,308,308,308,308,308, +308,308,308,308,308,308,308,308,308,308,308,308,308,308,308,308, + +/* block 49 */ +309,309,309,309,309,309,309,309,309,309,309,309,309,309,309,309, +309,309,309,309,309,309,309,310,310,311,311,311, 92, 92,312,312, +313,313,313,313,313,313,313,313,313,313,313,313,313,313,313,313, +313,313,313,313,313,313,313,313,313,313,313,313,313,313,313,313, +313,313,313,313,313,313,313,313,313,313,313,313,313,313,313,313, +313,313,313,313,313,314,315,314,315,315,315,315,315,315,315, 92, +315,314,315,314,314,315,315,315,315,315,315,315,315,314,314,314, +314,314,314,315,315,315,315,315,315,315,315,315,315, 92, 92,315, + +/* block 50 */ +316,316,316,316,316,316,316,316,316,316, 92, 92, 92, 92, 92, 92, +316,316,316,316,316,316,316,316,316,316, 92, 92, 92, 92, 92, 92, +317,317,317,317,317,317,317,318,317,317,317,317,317,317, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 51 */ +319,319,319,319,320,321,321,321,321,321,321,321,321,321,321,321, +321,321,321,321,321,321,321,321,321,321,321,321,321,321,321,321, +321,321,321,321,321,321,321,321,321,321,321,321,321,321,321,321, +321,321,321,321,319,320,319,319,319,319,319,320,319,320,320,320, +320,320,319,320,320,321,321,321,321,321,321,321, 92, 92, 92, 92, +322,322,322,322,322,322,322,322,322,322,323,323,323,323,323,323, +323,324,324,324,324,324,324,324,324,324,324,319,319,319,319,319, +319,319,319,319,324,324,324,324,324,324,324,324,324, 92, 92, 92, + +/* block 52 */ +325,325,326,327,327,327,327,327,327,327,327,327,327,327,327,327, +327,327,327,327,327,327,327,327,327,327,327,327,327,327,327,327, +327,326,325,325,325,325,326,326,325,325,326,325,326,326,327,327, +328,328,328,328,328,328,328,328,328,328,327,327,327,327,327,327, +329,329,329,329,329,329,329,329,329,329,329,329,329,329,329,329, +329,329,329,329,329,329,329,329,329,329,329,329,329,329,329,329, +329,329,329,329,329,329,330,331,330,330,331,331,331,330,331,330, +330,330,331,331, 92, 92, 92, 92, 92, 92, 92, 92,332,332,332,332, + +/* block 53 */ +333,333,333,333,333,333,333,333,333,333,333,333,333,333,333,333, +333,333,333,333,333,333,333,333,333,333,333,333,333,333,333,333, +333,333,333,333,334,334,334,334,334,334,334,334,335,335,335,335, +335,335,335,335,334,334,335,335, 92, 92, 92,336,336,336,336,336, +337,337,337,337,337,337,337,337,337,337, 92, 92, 92,333,333,333, +338,338,338,338,338,338,338,338,338,338,339,339,339,339,339,339, +339,339,339,339,339,339,339,339,339,339,339,339,339,339,339,339, +339,339,339,339,339,339,339,339,340,340,340,340,340,340,341,341, + +/* block 54 */ + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +342,342,342,342,342,342,342,342, 92, 92, 92, 92, 92, 92, 92, 92, + 87, 87, 87, 2, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87,343, 87, 87, 87, 87, 87, 87, 87,344,344,344,344, 87,344,344, +344,344,343,343, 87,344,344, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 55 */ + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, + 20, 20, 20, 20, 20, 20, 99, 99, 99, 99, 99,345, 84, 84, 84, 84, + 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, + 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, + 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 93, 93, 93, + 93, 93, 84, 84, 84, 84, 93, 93, 93, 93, 93, 20, 20, 20, 20, 20, + 20, 20, 20, 20, 20, 20, 20, 20,346,347, 20, 20, 20,348, 20, 20, + +/* block 56 */ + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 84, 84, 84, 84, 84, + 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, + 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 93, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 87, 87, 87, 87, + +/* block 57 */ + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + +/* block 58 */ + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 20, 20, 20, 20, 20,349, 20, 20,350, 20, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + +/* block 59 */ +351,351,351,351,351,351,351,351,352,352,352,352,352,352,352,352, +351,351,351,351,351,351, 92, 92,352,352,352,352,352,352, 92, 92, +351,351,351,351,351,351,351,351,352,352,352,352,352,352,352,352, +351,351,351,351,351,351,351,351,352,352,352,352,352,352,352,352, +351,351,351,351,351,351, 92, 92,352,352,352,352,352,352, 92, 92, + 99,351, 99,351, 99,351, 99,351, 92,352, 92,352, 92,352, 92,352, +351,351,351,351,351,351,351,351,352,352,352,352,352,352,352,352, +353,353,354,354,354,354,355,355,356,356,357,357,358,358, 92, 92, + +/* block 60 */ +351,351,351,351,351,351,351,351,359,359,359,359,359,359,359,359, +351,351,351,351,351,351,351,351,359,359,359,359,359,359,359,359, +351,351,351,351,351,351,351,351,359,359,359,359,359,359,359,359, +351,351, 99,360, 99, 92, 99, 99,352,352,361,361,362, 91,363, 91, + 91, 91, 99,360, 99, 92, 99, 99,364,364,364,364,362, 91, 91, 91, +351,351, 99, 99, 92, 92, 99, 99,352,352,365,365, 92, 91, 91, 91, +351,351, 99, 99, 99,118, 99, 99,352,352,366,366,122, 91, 91, 91, + 92, 92, 99,360, 99, 92, 99, 99,367,367,368,368,362, 91, 91, 92, + +/* block 61 */ + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 16,369,369, 16, 16, + 7, 7, 7, 7, 7, 7, 2, 2, 15, 19, 4, 15, 15, 19, 4, 15, + 2, 2, 2, 2, 2, 2, 2, 2,370,371, 16, 16, 16, 16, 16, 1, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 15, 19, 2, 2, 2, 2, 11, + 11, 2, 2, 2, 6, 4, 5, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 6, 2, 11, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, + 16, 16, 16, 16, 16, 92, 92, 92, 92, 92, 16, 16, 16, 16, 16, 16, + 17, 84, 92, 92, 17, 17, 17, 17, 17, 17, 6, 6, 6, 4, 5, 84, + +/* block 62 */ + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 6, 6, 6, 4, 5, 92, + 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 84, 92, 92, 92, + 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, + 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87,372,372,372, +372, 87,372,372,372, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 63 */ + 13, 13,373, 13, 13, 13, 13,373, 13, 13,374,373,373,373,374,374, +373,373,373,374, 13,373, 13, 13, 6,373,373,373,373,373, 13, 13, + 13, 13, 13, 13,373, 13,375, 13,373, 13,376,377,373,373, 13,374, +373,373,378,373,374,344,344,344,344,374, 13, 13,374,374,373,373, + 6, 6, 6, 6, 6,373,374,374,374,374, 13, 6, 13, 13,379, 13, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, +380,380,380,380,380,380,380,380,380,380,380,380,380,380,380,380, +381,381,381,381,381,381,381,381,381,381,381,381,381,381,381,381, + +/* block 64 */ +382,382,382, 22, 23,382,382,382,382, 17, 92, 92, 92, 92, 92, 92, + 6, 6, 6, 6, 6, 13, 13, 13, 13, 13, 6, 6, 13, 13, 13, 13, + 6, 13, 13, 6, 13, 13, 6, 13, 13, 13, 13, 13, 13, 13, 6, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 6, 6, + 13, 13, 6, 13, 6, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + +/* block 65 */ + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + +/* block 66 */ + 13, 13, 13, 13, 13, 13, 13, 13, 6, 6, 6, 6, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 6, 6, 13, 13, 13, 13, 13, 13, 13, 4, 5, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 6, 13, 13, 13, + +/* block 67 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 6, 6, 6, 6, + 6, 6, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 68 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + +/* block 69 */ + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13,383,383,383,383,383,383,383,383,383,383, +383,383,383,383,383,383,383,383,383,383,383,383,383,383,383,383, +384,384,384,384,384,384,384,384,384,384,384,384,384,384,384,384, +384,384,384,384,384,384,384,384,384,384, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + +/* block 70 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + +/* block 71 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 6, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 6, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 6, 6, 6, 6, 6, 6, 6, 6, + +/* block 72 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 6, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + +/* block 73 */ + 92, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 4, 5, 4, 5, 4, 5, 4, 5, + 4, 5, 4, 5, 4, 5, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + +/* block 74 */ + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 6, 6, 6, 6, 6, 4, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + +/* block 75 */ +385,385,385,385,385,385,385,385,385,385,385,385,385,385,385,385, +385,385,385,385,385,385,385,385,385,385,385,385,385,385,385,385, +385,385,385,385,385,385,385,385,385,385,385,385,385,385,385,385, +385,385,385,385,385,385,385,385,385,385,385,385,385,385,385,385, +385,385,385,385,385,385,385,385,385,385,385,385,385,385,385,385, +385,385,385,385,385,385,385,385,385,385,385,385,385,385,385,385, +385,385,385,385,385,385,385,385,385,385,385,385,385,385,385,385, +385,385,385,385,385,385,385,385,385,385,385,385,385,385,385,385, + +/* block 76 */ + 6, 6, 6, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, + 5, 4, 5, 4, 5, 4, 5, 4, 5, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 4, 5, 4, 5, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 4, 5, 6, 6, + +/* block 77 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, + 6, 6, 6, 6, 6, 13, 13, 6, 6, 6, 6, 6, 6, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 78 */ + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 79 */ +386,386,386,386,386,386,386,386,386,386,386,386,386,386,386,386, +386,386,386,386,386,386,386,386,386,386,386,386,386,386,386,386, +386,386,386,386,386,386,386,386,386,386,386,386,386,386,386, 92, +387,387,387,387,387,387,387,387,387,387,387,387,387,387,387,387, +387,387,387,387,387,387,387,387,387,387,387,387,387,387,387,387, +387,387,387,387,387,387,387,387,387,387,387,387,387,387,387, 92, + 22, 23,388,389,390,391,392, 22, 23, 22, 23, 22, 23,393,394,395, +396, 20, 22, 23, 20, 22, 23, 20, 20, 20, 20, 20, 84, 84,397,397, + +/* block 80 */ +114,115,114,115,114,115,114,115,114,115,114,115,114,115,114,115, +114,115,114,115,114,115,114,115,114,115,114,115,114,115,114,115, +114,115,114,115,114,115,114,115,114,115,114,115,114,115,114,115, +114,115,114,115,114,115,114,115,114,115,114,115,114,115,114,115, +114,115,114,115,114,115,114,115,114,115,114,115,114,115,114,115, +114,115,114,115,114,115,114,115,114,115,114,115,114,115,114,115, +114,115,114,115,398,399,399,399,399,399,399,114,115,114,115,400, +400,400,114,115, 92, 92, 92, 92, 92,401,401,401,401,402,401,401, + +/* block 81 */ +403,403,403,403,403,403,403,403,403,403,403,403,403,403,403,403, +403,403,403,403,403,403,403,403,403,403,403,403,403,403,403,403, +403,403,403,403,403,403, 92,403, 92, 92, 92, 92, 92,403, 92, 92, +404,404,404,404,404,404,404,404,404,404,404,404,404,404,404,404, +404,404,404,404,404,404,404,404,404,404,404,404,404,404,404,404, +404,404,404,404,404,404,404,404,404,404,404,404,404,404,404,404, +404,404,404,404,404,404,404,404, 92, 92, 92, 92, 92, 92, 92,405, +406, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92,407, + +/* block 82 */ +258,258,258,258,258,258,258,258,258,258,258,258,258,258,258,258, +258,258,258,258,258,258,258, 92, 92, 92, 92, 92, 92, 92, 92, 92, +258,258,258,258,258,258,258, 92,258,258,258,258,258,258,258, 92, +258,258,258,258,258,258,258, 92,258,258,258,258,258,258,258, 92, +258,258,258,258,258,258,258, 92,258,258,258,258,258,258,258, 92, +258,258,258,258,258,258,258, 92,258,258,258,258,258,258,258, 92, +131,131,131,131,131,131,131,131,131,131,131,131,131,131,131,131, +131,131,131,131,131,131,131,131,131,131,131,131,131,131,131,131, + +/* block 83 */ + 2, 2, 15, 19, 15, 19, 2, 2, 2, 15, 19, 2, 15, 19, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 7, 2, 2, 7, 2, 15, 19, 2, 2, + 15, 19, 4, 5, 4, 5, 4, 5, 4, 5, 2, 2, 2, 2, 2, 85, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 7, 7, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 84 */ +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408, 92,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 85 */ +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, + +/* block 86 */ +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408,408,408,408,408,408,408,408,408,408,408, +408,408,408,408,408,408, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, + +/* block 87 */ + 1, 2, 2, 2, 13,409,344,410, 4, 5, 4, 5, 4, 5, 4, 5, + 4, 5, 13, 13, 4, 5, 4, 5, 4, 5, 4, 5, 7, 4, 5, 5, + 13,410,410,410,410,410,410,410,410,410, 87, 87, 87, 87,411,411, + 7, 85, 85, 85, 85, 85, 13, 13,410,410,410,409,344, 2, 13, 13, + 92,412,412,412,412,412,412,412,412,412,412,412,412,412,412,412, +412,412,412,412,412,412,412,412,412,412,412,412,412,412,412,412, +412,412,412,412,412,412,412,412,412,412,412,412,412,412,412,412, +412,412,412,412,412,412,412,412,412,412,412,412,412,412,412,412, + +/* block 88 */ +412,412,412,412,412,412,412,412,412,412,412,412,412,412,412,412, +412,412,412,412,412,412,412, 92, 92, 87, 87, 10, 10,413,413,412, + 7,414,414,414,414,414,414,414,414,414,414,414,414,414,414,414, +414,414,414,414,414,414,414,414,414,414,414,414,414,414,414,414, +414,414,414,414,414,414,414,414,414,414,414,414,414,414,414,414, +414,414,414,414,414,414,414,414,414,414,414,414,414,414,414,414, +414,414,414,414,414,414,414,414,414,414,414,414,414,414,414,414, +414,414,414,414,414,414,414,414,414,414,414, 2, 85,415,415,414, + +/* block 89 */ + 92, 92, 92, 92, 92,416,416,416,416,416,416,416,416,416,416,416, +416,416,416,416,416,416,416,416,416,416,416,416,416,416,416,416, +416,416,416,416,416,416,416,416,416,416,416,416,416,416, 92, 92, + 92,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, + +/* block 90 */ +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, 92, + 13, 13, 17, 17, 17, 17, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, +416,416,416,416,416,416,416,416,416,416,416,416,416,416,416,416, +416,416,416,416,416,416,416,416,416,416,416, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +414,414,414,414,414,414,414,414,414,414,414,414,414,414,414,414, + +/* block 91 */ +417,417,417,417,417,417,417,417,417,417,417,417,417,417,417,417, +417,417,417,417,417,417,417,417,417,417,417,417,417,417,417, 92, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 17, 17, 17, 17, 17, 17, 17, 17, + 13, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, +417,417,417,417,417,417,417,417,417,417,417,417,417,417,417,417, +417,417,417,417,417,417,417,417,417,417,417,417,417,417,417, 13, + +/* block 92 */ + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, +418,418,418,418,418,418,418,418,418,418,418,418,418,418,418,418, +418,418,418,418,418,418,418,418,418,418,418,418,418,418,418,418, +418,418,418,418,418,418,418,418,418,418,418,418,418,418,418, 92, + +/* block 93 */ +418,418,418,418,418,418,418,418,418,418,418,418,418,418,418,418, +418,418,418,418,418,418,418,418,418,418,418,418,418,418,418,418, +418,418,418,418,418,418,418,418,418,418,418,418,418,418,418,418, +418,418,418,418,418,418,418,418,418,418,418,418,418,418,418,418, +418,418,418,418,418,418,418,418,418,418,418,418,418,418,418,418, +418,418,418,418,418,418,418,418, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + +/* block 94 */ +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, + +/* block 95 */ +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + +/* block 96 */ +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 97 */ +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,421,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, + +/* block 98 */ +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, +420,420,420,420,420,420,420,420,420,420,420,420,420,420,420,420, + +/* block 99 */ +420,420,420,420,420,420,420,420,420,420,420,420,420, 92, 92, 92, +422,422,422,422,422,422,422,422,422,422,422,422,422,422,422,422, +422,422,422,422,422,422,422,422,422,422,422,422,422,422,422,422, +422,422,422,422,422,422,422,422,422,422,422,422,422,422,422,422, +422,422,422,422,422,422,422, 92, 92, 92, 92, 92, 92, 92, 92, 92, +423,423,423,423,423,423,423,423,423,423,423,423,423,423,423,423, +423,423,423,423,423,423,423,423,423,423,423,423,423,423,423,423, +423,423,423,423,423,423,423,423,424,424,424,424,424,424,425,425, + +/* block 100 */ +426,426,426,426,426,426,426,426,426,426,426,426,426,426,426,426, +426,426,426,426,426,426,426,426,426,426,426,426,426,426,426,426, +426,426,426,426,426,426,426,426,426,426,426,426,426,426,426,426, +426,426,426,426,426,426,426,426,426,426,426,426,426,426,426,426, +426,426,426,426,426,426,426,426,426,426,426,426,426,426,426,426, +426,426,426,426,426,426,426,426,426,426,426,426,426,426,426,426, +426,426,426,426,426,426,426,426,426,426,426,426,426,426,426,426, +426,426,426,426,426,426,426,426,426,426,426,426,426,426,426,426, + +/* block 101 */ +426,426,426,426,426,426,426,426,426,426,426,426,427,428,428,428, +426,426,426,426,426,426,426,426,426,426,426,426,426,426,426,426, +429,429,429,429,429,429,429,429,429,429,426,426, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, +128,129,128,129,128,129,128,129,128,129,128,129,128,129,430,131, +132,132,132,431,131,131,131,131,131,131,131,131,131,131,431,346, + +/* block 102 */ +128,129,128,129,128,129,128,129,128,129,128,129,128,129,128,129, +128,129,128,129,128,129,128,129, 92, 92, 92, 92, 92, 92, 92,131, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,433,433,433,433,433,433,433,433,433,433, +434,434,435,435,435,435,435,435, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 103 */ + 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, + 10, 10, 10, 10, 10, 10, 10, 85, 85, 85, 85, 85, 85, 85, 85, 85, + 10, 10, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 20, 20, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, 22, 23, + 84, 20, 20, 20, 20, 20, 20, 20, 20, 22, 23, 22, 23,436, 22, 23, + +/* block 104 */ + 22, 23, 22, 23, 22, 23, 22, 23, 85, 10, 10, 22, 23,437, 20, 92, + 22, 23, 22, 23, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 22, 23, 22, 23, 22, 23, 22, 23, 22, 23,438, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 84, 84, 20, 14, 14, 14, 14, 14, + +/* block 105 */ +439,439,440,439,439,439,440,439,439,439,439,440,439,439,439,439, +439,439,439,439,439,439,439,439,439,439,439,439,439,439,439,439, +439,439,439,441,441,440,440,441,442,442,442,442, 92, 92, 92, 92, + 17, 17, 17, 17, 17, 17, 13, 13, 3, 13, 92, 92, 92, 92, 92, 92, +443,443,443,443,443,443,443,443,443,443,443,443,443,443,443,443, +443,443,443,443,443,443,443,443,443,443,443,443,443,443,443,443, +443,443,443,443,443,443,443,443,443,443,443,443,443,443,443,443, +443,443,443,443,444,444,444,444, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 106 */ +445,445,446,446,446,446,446,446,446,446,446,446,446,446,446,446, +446,446,446,446,446,446,446,446,446,446,446,446,446,446,446,446, +446,446,446,446,446,446,446,446,446,446,446,446,446,446,446,446, +446,446,446,446,445,445,445,445,445,445,445,445,445,445,445,445, +445,445,445,445,447, 92, 92, 92, 92, 92, 92, 92, 92, 92,448,448, +449,449,449,449,449,449,449,449,449,449, 92, 92, 92, 92, 92, 92, +174,174,174,174,174,174,174,174,174,174,174,174,174,174,174,174, +174,174,176,176,176,176,176,176,178,178,178,176, 92, 92, 92, 92, + +/* block 107 */ +450,450,450,450,450,450,450,450,450,450,451,451,451,451,451,451, +451,451,451,451,451,451,451,451,451,451,451,451,451,451,451,451, +451,451,451,451,451,451,452,452,452,452,452,452,452,452,453,453, +454,454,454,454,454,454,454,454,454,454,454,454,454,454,454,454, +454,454,454,454,454,454,454,455,455,455,455,455,455,455,455,455, +455,455,456,456, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92,457, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257, 92, 92, 92, + +/* block 108 */ +458,458,458,459,460,460,460,460,460,460,460,460,460,460,460,460, +460,460,460,460,460,460,460,460,460,460,460,460,460,460,460,460, +460,460,460,460,460,460,460,460,460,460,460,460,460,460,460,460, +460,460,460,458,459,459,458,458,458,458,459,459,458,459,459,459, +459,461,461,461,461,461,461,461,461,461,461,461,461,461, 92,462, +463,463,463,463,463,463,463,463,463,463, 92, 92, 92, 92,461,461, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 109 */ +464,464,464,464,464,464,464,464,464,464,464,464,464,464,464,464, +464,464,464,464,464,464,464,464,464,464,464,464,464,464,464,464, +464,464,464,464,464,464,464,464,464,465,465,465,465,465,465,466, +466,465,465,466,466,465,465, 92, 92, 92, 92, 92, 92, 92, 92, 92, +464,464,464,465,464,464,464,464,464,464,464,464,465,466, 92, 92, +467,467,467,467,467,467,467,467,467,467, 92, 92,468,468,468,468, +248,248,248,248,248,248,248,248,248,248,248,248,248,248,248,248, +469,248,248,248,248,248,248,253,253,253,248,249, 92, 92, 92, 92, + +/* block 110 */ +470,470,470,470,470,470,470,470,470,470,470,470,470,470,470,470, +470,470,470,470,470,470,470,470,470,470,470,470,470,470,470,470, +470,470,470,470,470,470,470,470,470,470,470,470,470,470,470,470, +471,470,471,471,471,470,470,471,471,470,470,470,470,470,471,471, +470,471,470, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92,470,470,472,473,473, +474,474,474,474,474,474,474,474,474,474,474,475,476,476,475,475, +477,477,474,478,478,475,476, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 111 */ + 92,258,258,258,258,258,258, 92, 92,258,258,258,258,258,258, 92, + 92,258,258,258,258,258,258, 92, 92, 92, 92, 92, 92, 92, 92, 92, +258,258,258,258,258,258,258, 92,258,258,258,258,258,258,258, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 112 */ + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +474,474,474,474,474,474,474,474,474,474,474,474,474,474,474,474, +474,474,474,474,474,474,474,474,474,474,474,474,474,474,474,474, +474,474,474,475,475,476,475,475,476,475,475,477,475,476, 92, 92, +479,479,479,479,479,479,479,479,479,479, 92, 92, 92, 92, 92, 92, + +/* block 113 */ +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257, 92, 92, 92, 92,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257, 92, 92, 92, 92, + +/* block 114 */ +480,480,480,480,480,480,480,480,480,480,480,480,480,480,480,480, +480,480,480,480,480,480,480,480,480,480,480,480,480,480,480,480, +480,480,480,480,480,480,480,480,480,480,480,480,480,480,480,480, +480,480,480,480,480,480,480,480,480,480,480,480,480,480,480,480, +480,480,480,480,480,480,480,480,480,480,480,480,480,480,480,480, +480,480,480,480,480,480,480,480,480,480,480,480,480,480,480,480, +480,480,480,480,480,480,480,480,480,480,480,480,480,480,480,480, +480,480,480,480,480,480,480,480,480,480,480,480,480,480,480,480, + +/* block 115 */ +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, + +/* block 116 */ +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419, 92, 92, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, + +/* block 117 */ +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 118 */ + 20, 20, 20, 20, 20, 20, 20, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92,139,139,139,139,139, 92, 92, 92, 92, 92,145,142,145, +145,145,145,145,145,145,145,145,145,482,145,145,145,145,145,145, +145,145,145,145,145,145,145, 92,145,145,145,145,145, 92,145, 92, +145,145, 92,145,145, 92,145,145,145,145,145,145,145,145,145,145, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, + +/* block 119 */ +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,483,483,483,483,483,483,483,483,483,483,483,483,483,483, +483,483, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, + +/* block 120 */ +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, + +/* block 121 */ +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152, 4, 5, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, + +/* block 122 */ +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, + 92, 92,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +152,152,152,152,152,152,152,152,152,152,152,152,149, 13, 92, 92, + +/* block 123 */ + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 2, 2, 2, 2, 2, 2, 2, 4, 5, 2, 92, 92, 92, 92, 92, 92, + 87, 87, 87, 87, 87, 87, 87, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 2, 7, 7, 11, 11, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, + 5, 4, 5, 4, 5, 2, 2, 4, 5, 2, 2, 2, 2, 11, 11, 11, + 2, 2, 2, 92, 2, 2, 2, 2, 7, 4, 5, 4, 5, 4, 5, 2, + 2, 2, 6, 7, 6, 6, 6, 92, 2, 3, 2, 2, 92, 92, 92, 92, +152,152,152,152,152, 92,152,152,152,152,152,152,152,152,152,152, + +/* block 124 */ +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152, 92, 92, 16, + +/* block 125 */ + 92, 2, 2, 2, 3, 2, 2, 2, 4, 5, 2, 6, 2, 7, 2, 2, + 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2, 2, 6, 6, 6, 2, + 2, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, + 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 4, 2, 5, 10, 11, + 10, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, + 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 4, 6, 5, 6, 4, + 5, 2, 4, 5, 2, 2,414,414,414,414,414,414,414,414,414,414, + 85,414,414,414,414,414,414,414,414,414,414,414,414,414,414,414, + +/* block 126 */ +414,414,414,414,414,414,414,414,414,414,414,414,414,414,414,414, +414,414,414,414,414,414,414,414,414,414,414,414,414,414, 85, 85, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, +257,257,257,257,257,257,257,257,257,257,257,257,257,257,257, 92, + 92, 92,257,257,257,257,257,257, 92, 92,257,257,257,257,257,257, + 92, 92,257,257,257,257,257,257, 92, 92,257,257,257, 92, 92, 92, + 3, 3, 6, 10, 13, 3, 3, 92, 13, 6, 6, 6, 6, 13, 13, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 16, 16, 16, 13, 13, 92, 92, + +/* block 127 */ +484,484,484,484,484,484,484,484,484,484,484,484, 92,484,484,484, +484,484,484,484,484,484,484,484,484,484,484,484,484,484,484,484, +484,484,484,484,484,484,484, 92,484,484,484,484,484,484,484,484, +484,484,484,484,484,484,484,484,484,484,484, 92,484,484, 92,484, +484,484,484,484,484,484,484,484,484,484,484,484,484,484, 92, 92, +484,484,484,484,484,484,484,484,484,484,484,484,484,484, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 128 */ +484,484,484,484,484,484,484,484,484,484,484,484,484,484,484,484, +484,484,484,484,484,484,484,484,484,484,484,484,484,484,484,484, +484,484,484,484,484,484,484,484,484,484,484,484,484,484,484,484, +484,484,484,484,484,484,484,484,484,484,484,484,484,484,484,484, +484,484,484,484,484,484,484,484,484,484,484,484,484,484,484,484, +484,484,484,484,484,484,484,484,484,484,484,484,484,484,484,484, +484,484,484,484,484,484,484,484,484,484,484,484,484,484,484,484, +484,484,484,484,484,484,484,484,484,484,484, 92, 92, 92, 92, 92, + +/* block 129 */ + 2, 2, 2, 92, 92, 92, 92, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 17, 17, 92, 92, 92, 13, 13, 13, 13, 13, 13, 13, 13, 13, +485,485,485,485,485,485,485,485,485,485,485,485,485,485,485,485, +485,485,485,485,485,485,485,485,485,485,485,485,485,485,485,485, +485,485,485,485,485,485,485,485,485,485,485,485,485,485,485,485, +485,485,485,485,485,486,486,486,486,487,487,487,487,487,487,487, + +/* block 130 */ +487,487,487,487,487,487,487,487,487,487,486, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 87, 92, 92, + +/* block 131 */ +488,488,488,488,488,488,488,488,488,488,488,488,488,488,488,488, +488,488,488,488,488,488,488,488,488,488,488,488,488, 92, 92, 92, +489,489,489,489,489,489,489,489,489,489,489,489,489,489,489,489, +489,489,489,489,489,489,489,489,489,489,489,489,489,489,489,489, +489,489,489,489,489,489,489,489,489,489,489,489,489,489,489,489, +489, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 132 */ +490,490,490,490,490,490,490,490,490,490,490,490,490,490,490,490, +490,490,490,490,490,490,490,490,490,490,490,490,490,490,490, 92, +491,491,491,491, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +492,492,492,492,492,492,492,492,492,492,492,492,492,492,492,492, +492,493,492,492,492,492,492,492,492,492,493, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 133 */ +494,494,494,494,494,494,494,494,494,494,494,494,494,494,494,494, +494,494,494,494,494,494,494,494,494,494,494,494,494,494, 92,495, +496,496,496,496,496,496,496,496,496,496,496,496,496,496,496,496, +496,496,496,496,496,496,496,496,496,496,496,496,496,496,496,496, +496,496,496,496, 92, 92, 92, 92,496,496,496,496,496,496,496,496, +497,498,498,498,498,498, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 134 */ +499,499,499,499,499,499,499,499,499,499,499,499,499,499,499,499, +499,499,499,499,499,499,499,499,499,499,499,499,499,499,499,499, +499,499,499,499,499,499,499,499,500,500,500,500,500,500,500,500, +500,500,500,500,500,500,500,500,500,500,500,500,500,500,500,500, +500,500,500,500,500,500,500,500,500,500,500,500,500,500,500,500, +501,501,501,501,501,501,501,501,501,501,501,501,501,501,501,501, +501,501,501,501,501,501,501,501,501,501,501,501,501,501,501,501, +501,501,501,501,501,501,501,501,501,501,501,501,501,501,501,501, + +/* block 135 */ +502,502,502,502,502,502,502,502,502,502,502,502,502,502,502,502, +502,502,502,502,502,502,502,502,502,502,502,502,502,502, 92, 92, +503,503,503,503,503,503,503,503,503,503, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 136 */ +504,504,504,504,504,504, 92, 92,504, 92,504,504,504,504,504,504, +504,504,504,504,504,504,504,504,504,504,504,504,504,504,504,504, +504,504,504,504,504,504,504,504,504,504,504,504,504,504,504,504, +504,504,504,504,504,504, 92,504,504, 92, 92, 92,504, 92, 92,504, +505,505,505,505,505,505,505,505,505,505,505,505,505,505,505,505, +505,505,505,505,505,505, 92,506,507,507,507,507,507,507,507,507, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 137 */ +508,508,508,508,508,508,508,508,508,508,508,508,508,508,508,508, +508,508,508,508,508,508,509,509,509,509,509,509, 92, 92, 92,510, +511,511,511,511,511,511,511,511,511,511,511,511,511,511,511,511, +511,511,511,511,511,511,511,511,511,511, 92, 92, 92, 92, 92,512, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 138 */ +513,513,513,513,513,513,513,513,513,513,513,513,513,513,513,513, +513,513,513,513,513,513,513,513,513,513,513,513,513,513,513,513, +514,514,514,514,514,514,514,514,514,514,514,514,514,514,514,514, +514,514,514,514,514,514,514,514, 92, 92, 92, 92, 92, 92,514,514, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 139 */ +515,516,516,516, 92,516,516, 92, 92, 92, 92, 92,516,516,516,516, +515,515,515,515, 92,515,515,515, 92,515,515,515,515,515,515,515, +515,515,515,515,515,515,515,515,515,515,515,515,515,515,515,515, +515,515,515,515, 92, 92, 92, 92,516,516,516, 92, 92, 92, 92,516, +517,517,517,517,517,517,517,517, 92, 92, 92, 92, 92, 92, 92, 92, +518,518,518,518,518,518,518,518,518, 92, 92, 92, 92, 92, 92, 92, +519,519,519,519,519,519,519,519,519,519,519,519,519,519,519,519, +519,519,519,519,519,519,519,519,519,519,519,519,519,520,520,521, + +/* block 140 */ +522,522,522,522,522,522,522,522,522,522,522,522,522,522,522,522, +522,522,522,522,522,522,522,522,522,522,522,522,522,522,522,522, +522,522,522,522,522,522,522,522,522,522,522,522,522,522,522,522, +522,522,522,522,522,522, 92, 92, 92,523,523,523,523,523,523,523, +524,524,524,524,524,524,524,524,524,524,524,524,524,524,524,524, +524,524,524,524,524,524, 92, 92,525,525,525,525,525,525,525,525, +526,526,526,526,526,526,526,526,526,526,526,526,526,526,526,526, +526,526,526, 92, 92, 92, 92, 92,527,527,527,527,527,527,527,527, + +/* block 141 */ +528,528,528,528,528,528,528,528,528,528,528,528,528,528,528,528, +528,528,528,528,528,528,528,528,528,528,528,528,528,528,528,528, +528,528,528,528,528,528,528,528,528,528,528,528,528,528,528,528, +528,528,528,528,528,528,528,528,528,528,528,528,528,528,528,528, +528,528,528,528,528,528,528,528,528, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 142 */ + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +529,529,529,529,529,529,529,529,529,529,529,529,529,529,529,529, +529,529,529,529,529,529,529,529,529,529,529,529,529,529,529, 92, + +/* block 143 */ +530,531,530,532,532,532,532,532,532,532,532,532,532,532,532,532, +532,532,532,532,532,532,532,532,532,532,532,532,532,532,532,532, +532,532,532,532,532,532,532,532,532,532,532,532,532,532,532,532, +532,532,532,532,532,532,532,532,531,531,531,531,531,531,531,531, +531,531,531,531,531,531,531,533,533,533,533,533,533,533, 92, 92, + 92, 92,534,534,534,534,534,534,534,534,534,534,534,534,534,534, +534,534,534,534,534,534,535,535,535,535,535,535,535,535,535,535, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 144 */ +536,536,537,538,538,538,538,538,538,538,538,538,538,538,538,538, +538,538,538,538,538,538,538,538,538,538,538,538,538,538,538,538, +538,538,538,538,538,538,538,538,538,538,538,538,538,538,538,538, +537,537,537,536,536,536,536,537,537,536,536,539,539,540,539,539, +539,539, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +541,541,541,541,541,541,541,541,541,541,541,541,541,541,541,541, +541,541,541,541,541,541,541,541,541, 92, 92, 92, 92, 92, 92, 92, +542,542,542,542,542,542,542,542,542,542, 92, 92, 92, 92, 92, 92, + +/* block 145 */ +543,543,543,544,544,544,544,544,544,544,544,544,544,544,544,544, +544,544,544,544,544,544,544,544,544,544,544,544,544,544,544,544, +544,544,544,544,544,544,544,543,543,543,543,543,545,543,543,543, +543,543,543,543,543, 92,546,546,546,546,546,546,546,546,546,546, +547,547,547,547, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 146 */ +548,548,549,550,550,550,550,550,550,550,550,550,550,550,550,550, +550,550,550,550,550,550,550,550,550,550,550,550,550,550,550,550, +550,550,550,550,550,550,550,550,550,550,550,550,550,550,550,550, +550,550,550,549,549,549,548,548,548,548,548,548,548,548,548,549, +549,550,550,550,550,551,551,551,551, 92, 92, 92, 92, 92, 92, 92, +552,552,552,552,552,552,552,552,552,552, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 147 */ +553,553,553,553,553,553,553,553,553,553,553,553,553,553,553,553, +553,553,553,553,553,553,553,553,553,553,553,553,553,553,553,553, +553,553,553,553,553,553,553,553,553,553,553,554,555,554,555,555, +554,554,554,554,554,554,555,554, 92, 92, 92, 92, 92, 92, 92, 92, +556,556,556,556,556,556,556,556,556,556, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 148 */ +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, + +/* block 149 */ +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, +557,557,557,557,557,557,557,557,557,557,557,557,557,557,557, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 150 */ +558,558,558,558,558,558,558,558,558,558,558,558,558,558,558,558, +558,558,558,558,558,558,558,558,558,558,558,558,558,558,558,558, +558,558,558,558,558,558,558,558,558,558,558,558,558,558,558,558, +558,558,558,558,558,558,558,558,558,558,558,558,558,558,558,558, +558,558,558,558,558,558,558,558,558,558,558,558,558,558,558,558, +558,558,558,558,558,558,558,558,558,558,558,558,558,558,558,558, +558,558,558, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +559,559,559,559, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 151 */ +560,560,560,560,560,560,560,560,560,560,560,560,560,560,560,560, +560,560,560,560,560,560,560,560,560,560,560,560,560,560,560,560, +560,560,560,560,560,560,560,560,560,560,560,560,560,560,560,560, +560,560,560,560,560,560,560,560,560,560,560,560,560,560,560,560, +560,560,560,560,560,560,560,560,560,560,560,560,560,560,560,560, +560,560,560,560,560,560,560,560,560,560,560,560,560,560,560,560, +560,560,560,560,560,560,560,560,560,560,560,560,560,560,560,560, +560,560,560,560,560,560,560,560,560,560,560,560,560,560,560,560, + +/* block 152 */ +560,560,560,560,560,560,560,560,560,560,560,560,560,560,560,560, +560,560,560,560,560,560,560,560,560,560,560,560,560,560,560,560, +560,560,560,560,560,560,560,560,560,560,560,560,560,560,560, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 153 */ +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, + +/* block 154 */ +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432,432,432,432,432,432,432,432, +432,432,432,432,432,432,432,432,432, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 155 */ +561,561,561,561,561,561,561,561,561,561,561,561,561,561,561,561, +561,561,561,561,561,561,561,561,561,561,561,561,561,561,561,561, +561,561,561,561,561,561,561,561,561,561,561,561,561,561,561,561, +561,561,561,561,561,561,561,561,561,561,561,561,561,561,561,561, +561,561,561,561,561, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +561,562,562,562,562,562,562,562,562,562,562,562,562,562,562,562, +562,562,562,562,562,562,562,562,562,562,562,562,562,562,562,562, +562,562,562,562,562,562,562,562,562,562,562,562,562,562,562, 92, + +/* block 156 */ + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92,563, +563,563,563,564,564,564,564,564,564,564,564,564,564,564,564,564, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 157 */ +414,412, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 158 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 159 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 92, 92, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13,343,343, 87, 87, 87, 13, 13, 13,343,343,343, +343,343,343, 16, 16, 16, 16, 16, 16, 16, 16, 87, 87, 87, 87, 87, + +/* block 160 */ + 87, 87, 87, 13, 13, 87, 87, 87, 87, 87, 87, 87, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 87, 87, 87, 87, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 161 */ +487,487,487,487,487,487,487,487,487,487,487,487,487,487,487,487, +487,487,487,487,487,487,487,487,487,487,487,487,487,487,487,487, +487,487,487,487,487,487,487,487,487,487,487,487,487,487,487,487, +487,487,487,487,487,487,487,487,487,487,487,487,487,487,487,487, +487,487,565,565,565,487, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 162 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, + 17, 17, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 163 */ +373,373,373,373,373,373,373,373,373,373,373,373,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,373,373,373,373,373,373,373,373,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,374,374, +374,374,374,374,374, 92,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,373,373,373,373,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,373,373, + +/* block 164 */ +373,373,374,374,374,374,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,373, 92,373,373, + 92, 92,373, 92, 92,373,373, 92, 92,373,373,373,373, 92,373,373, +373,373,373,373,373,373,374,374,374,374, 92,374, 92,374,374,374, +374,374,374,374, 92,374,374,374,374,374,374,374,374,374,374,374, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,374,374,374,374, + +/* block 165 */ +374,374,374,374,373,373, 92,373,373,373,373, 92, 92,373,373,373, +373,373,373,373,373, 92,373,373,373,373,373,373,373, 92,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,373,373, 92,373,373,373,373, 92, +373,373,373,373,373, 92,373, 92, 92, 92,373,373,373,373,373,373, +373, 92,374,374,374,374,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,373,373, + +/* block 166 */ +373,373,373,373,373,373,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,374,374,374,374, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,373,373,373,373,373,373,373,373,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,374,374,374,374, + +/* block 167 */ +374,374,374,374,374,374,374,374,373,373,373,373,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,373,373, +373,373,374,374,374,374,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,373,373, +373,373,373,373,373,373,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,374,374,374,374, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,373,373, + +/* block 168 */ +373,373,373,373,373,373,373,373,373,373,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,374,374, 92, 92,373,373,373,373,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,373,373, +373, 6,374,374,374,374,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,374,374,374, 6,374,374,374,374, +374,374,373,373,373,373,373,373,373,373,373,373,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,373, 6,374,374,374,374, + +/* block 169 */ +374,374,374,374,374,374,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,374, 6,374,374,374,374,374,374,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,373,373, +373,373,373,373,373, 6,374,374,374,374,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,374,374,374, 6, +374,374,374,374,374,374,373,373,373,373,373,373,373,373,373,373, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,373, 6, +374,374,374,374,374,374,374,374,374,374,374,374,374,374,374,374, + +/* block 170 */ +374,374,374,374,374,374,374,374,374, 6,374,374,374,374,374,374, +373,373,373,373,373,373,373,373,373,373,373,373,373,373,373,373, +373,373,373,373,373,373,373,373,373, 6,374,374,374,374,374,374, +374,374,374,374,374,374,374,374,374,374,374,374,374,374,374,374, +374,374,374, 6,374,374,374,374,374,374,373,374, 92, 92, 8, 8, + 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, + 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, + 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, + +/* block 171 */ +152,152,152,152, 92,152,152,152,152,152,152,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152,152,152,152,152, + 92,152,152, 92,152, 92, 92,152, 92,152,152,152,152,152,152,152, +152,152,152, 92,152,152,152,152, 92,152, 92,152, 92, 92, 92, 92, + 92, 92,152, 92, 92, 92, 92,152, 92,152, 92,152, 92,152,152,152, + 92,152,152, 92,152, 92, 92,152, 92,152, 92,152, 92,152, 92,152, + 92,152,152, 92,152, 92, 92,152,152,152,152, 92,152,152,152,152, +152,152,152, 92,152,152,152,152, 92,152,152,152,152, 92,152, 92, + +/* block 172 */ +152,152,152,152,152,152,152,152,152,152, 92,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152, 92, 92, 92, 92, + 92,152,152,152, 92,152,152,152,152,152, 92,152,152,152,152,152, +152,152,152,152,152,152,152,152,152,152,152,152, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +147,147, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 173 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + +/* block 174 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, + 92, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, + 92, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 92, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 175 */ + 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + +/* block 176 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + +/* block 177 */ +566, 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, 92, + 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 178 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 92, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, + +/* block 179 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 92, 13, 13, 13, 13, 13, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 180 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, + 13, 92, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + +/* block 181 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 92, 13, 13, 13, 13, 92, 92, 92, + +/* block 182 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, + 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 183 */ + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 13, 13, 13, 13, 13, + +/* block 184 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 92, 92, 92, 92, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 185 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 186 */ + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, + 13, 13, 13, 13, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 187 */ +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 188 */ +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, + +/* block 189 */ +419,419,419,419,419,419,419,419,419,419,419,419,419,419,419,419, +419,419,419,419,419,419,419,419,419,419,419,419,419,419, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 190 */ + 92, 16, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, + +/* block 191 */ + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + +/* block 192 */ + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, 87, + 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, 92, + +/* block 193 */ +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481,481,481, +481,481,481,481,481,481,481,481,481,481,481,481,481,481, 92, 92, + +}; + +#if UCD_BLOCK_SIZE != 128 +#error Please correct UCD_BLOCK_SIZE in pcre_internal.h +#endif +#endif /* SUPPORT_UCP */ diff --git a/pcre/pcre_ucp_searchfuncs.c b/pcre/pcre_ucp_searchfuncs.c deleted file mode 100644 --- a/pcre/pcre_ucp_searchfuncs.c +++ /dev/null @@ -1,179 +0,0 @@ -/************************************************* -* Perl-Compatible Regular Expressions * -*************************************************/ - -/* PCRE is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. - - Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* This module contains code for searching the table of Unicode character -properties. */ - -#ifdef HAVE_CONFIG_H -#include "config.h" -#endif - -#include "pcre_internal.h" - -#include "ucp.h" /* Category definitions */ -#include "ucpinternal.h" /* Internal table details */ -#include "ucptable.h" /* The table itself */ - - -/* Table to translate from particular type value to the general value. */ - -static const int ucp_gentype[] = { - ucp_C, ucp_C, ucp_C, ucp_C, ucp_C, /* Cc, Cf, Cn, Co, Cs */ - ucp_L, ucp_L, ucp_L, ucp_L, ucp_L, /* Ll, Lu, Lm, Lo, Lt */ - ucp_M, ucp_M, ucp_M, /* Mc, Me, Mn */ - ucp_N, ucp_N, ucp_N, /* Nd, Nl, No */ - ucp_P, ucp_P, ucp_P, ucp_P, ucp_P, /* Pc, Pd, Pe, Pf, Pi */ - ucp_P, ucp_P, /* Ps, Po */ - ucp_S, ucp_S, ucp_S, ucp_S, /* Sc, Sk, Sm, So */ - ucp_Z, ucp_Z, ucp_Z /* Zl, Zp, Zs */ -}; - - - -/************************************************* -* Search table and return type * -*************************************************/ - -/* Three values are returned: the category is ucp_C, ucp_L, etc. The detailed -character type is ucp_Lu, ucp_Nd, etc. The script is ucp_Latin, etc. - -Arguments: - c the character value - type_ptr the detailed character type is returned here - script_ptr the script is returned here - -Returns: the character type category -*/ - -int -_pcre_ucp_findprop(const unsigned int c, int *type_ptr, int *script_ptr) -{ -int bot = 0; -int top = sizeof(ucp_table)/sizeof(cnode); -int mid; - -/* The table is searched using a binary chop. You might think that using -intermediate variables to hold some of the common expressions would speed -things up, but tests with gcc 3.4.4 on Linux showed that, on the contrary, it -makes things a lot slower. */ - -for (;;) - { - if (top <= bot) - { - *type_ptr = ucp_Cn; - *script_ptr = ucp_Common; - return ucp_C; - } - mid = (bot + top) >> 1; - if (c == (ucp_table[mid].f0 & f0_charmask)) break; - if (c < (ucp_table[mid].f0 & f0_charmask)) top = mid; - else - { - if ((ucp_table[mid].f0 & f0_rangeflag) != 0 && - c <= (ucp_table[mid].f0 & f0_charmask) + - (ucp_table[mid].f1 & f1_rangemask)) break; - bot = mid + 1; - } - } - -/* Found an entry in the table. Set the script and detailed type values, and -return the general type. */ - -*script_ptr = (ucp_table[mid].f0 & f0_scriptmask) >> f0_scriptshift; -*type_ptr = (ucp_table[mid].f1 & f1_typemask) >> f1_typeshift; - -return ucp_gentype[*type_ptr]; -} - - - -/************************************************* -* Search table and return other case * -*************************************************/ - -/* If the given character is a letter, and there is another case for the -letter, return the other case. Otherwise, return -1. - -Arguments: - c the character value - -Returns: the other case or NOTACHAR if none -*/ - -unsigned int -_pcre_ucp_othercase(const unsigned int c) -{ -int bot = 0; -int top = sizeof(ucp_table)/sizeof(cnode); -int mid, offset; - -/* The table is searched using a binary chop. You might think that using -intermediate variables to hold some of the common expressions would speed -things up, but tests with gcc 3.4.4 on Linux showed that, on the contrary, it -makes things a lot slower. */ - -for (;;) - { - if (top <= bot) return -1; - mid = (bot + top) >> 1; - if (c == (ucp_table[mid].f0 & f0_charmask)) break; - if (c < (ucp_table[mid].f0 & f0_charmask)) top = mid; - else - { - if ((ucp_table[mid].f0 & f0_rangeflag) != 0 && - c <= (ucp_table[mid].f0 & f0_charmask) + - (ucp_table[mid].f1 & f1_rangemask)) break; - bot = mid + 1; - } - } - -/* Found an entry in the table. Return NOTACHAR for a range entry. Otherwise -return the other case if there is one, else NOTACHAR. */ - -if ((ucp_table[mid].f0 & f0_rangeflag) != 0) return NOTACHAR; - -offset = ucp_table[mid].f1 & f1_casemask; -if ((offset & f1_caseneg) != 0) offset |= f1_caseneg; -return (offset == 0)? NOTACHAR : c + offset; -} - - -/* End of pcre_ucp_searchfuncs.c */ diff --git a/pcre/pcre_valid_utf8.c b/pcre/pcre_valid_utf8.c --- a/pcre/pcre_valid_utf8.c +++ b/pcre/pcre_valid_utf8.c @@ -6,7 +6,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -54,109 +54,246 @@ *************************************************/ /* This function is called (optionally) at the start of compile or match, to -validate that a supposed UTF-8 string is actually valid. The early check means +check that a supposed UTF-8 string is actually valid. The early check means that subsequent code can assume it is dealing with a valid string. The check -can be turned off for maximum performance, but the consequences of supplying -an invalid string are then undefined. +can be turned off for maximum performance, but the consequences of supplying an +invalid string are then undefined. Originally, this function checked according to RFC 2279, allowing for values in the range 0 to 0x7fffffff, up to 6 bytes long, but ensuring that they were in the canonical format. Once somebody had pointed out RFC 3629 to me (it obsoletes 2279), additional restrictions were applied. The values are now limited to be between 0 and 0x0010ffff, no more than 4 bytes long, and the -subrange 0xd000 to 0xdfff is excluded. +subrange 0xd000 to 0xdfff is excluded. However, the format of 5-byte and 6-byte +characters is still checked. +From release 8.13 more information about the details of the error are passed +back in the returned value: + +PCRE_UTF8_ERR0 No error +PCRE_UTF8_ERR1 Missing 1 byte at the end of the string +PCRE_UTF8_ERR2 Missing 2 bytes at the end of the string +PCRE_UTF8_ERR3 Missing 3 bytes at the end of the string +PCRE_UTF8_ERR4 Missing 4 bytes at the end of the string +PCRE_UTF8_ERR5 Missing 5 bytes at the end of the string +PCRE_UTF8_ERR6 2nd-byte's two top bits are not 0x80 +PCRE_UTF8_ERR7 3rd-byte's two top bits are not 0x80 +PCRE_UTF8_ERR8 4th-byte's two top bits are not 0x80 +PCRE_UTF8_ERR9 5th-byte's two top bits are not 0x80 +PCRE_UTF8_ERR10 6th-byte's two top bits are not 0x80 +PCRE_UTF8_ERR11 5-byte character is not permitted by RFC 3629 +PCRE_UTF8_ERR12 6-byte character is not permitted by RFC 3629 +PCRE_UTF8_ERR13 4-byte character with value > 0x10ffff is not permitted +PCRE_UTF8_ERR14 3-byte character with value 0xd000-0xdfff is not permitted +PCRE_UTF8_ERR15 Overlong 2-byte sequence +PCRE_UTF8_ERR16 Overlong 3-byte sequence +PCRE_UTF8_ERR17 Overlong 4-byte sequence +PCRE_UTF8_ERR18 Overlong 5-byte sequence (won't ever occur) +PCRE_UTF8_ERR19 Overlong 6-byte sequence (won't ever occur) +PCRE_UTF8_ERR20 Isolated 0x80 byte (not within UTF-8 character) +PCRE_UTF8_ERR21 Byte with the illegal value 0xfe or 0xff + Arguments: string points to the string length length of string, or -1 if the string is zero-terminated + errp pointer to an error position offset variable -Returns: < 0 if the string is a valid UTF-8 string - >= 0 otherwise; the value is the offset of the bad byte +Returns: = 0 if the string is a valid UTF-8 string + > 0 otherwise, setting the offset of the bad character */ int -_pcre_valid_utf8(const uschar *string, int length) +PRIV(valid_utf)(PCRE_PUCHAR string, int length, int *erroroffset) { -#ifdef SUPPORT_UTF8 -register const uschar *p; +#ifdef SUPPORT_UTF +register PCRE_PUCHAR p; if (length < 0) { for (p = string; *p != 0; p++); - length = p - string; + length = (int)(p - string); } for (p = string; length-- > 0; p++) { - register int ab; - register int c = *p; - if (c < 128) continue; - if (c < 0xc0) return p - string; - ab = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ - if (length < ab || ab > 3) return p - string; - length -= ab; + register int ab, c, d; + c = *p; + if (c < 128) continue; /* ASCII character */ + + if (c < 0xc0) /* Isolated 10xx xxxx byte */ + { + *erroroffset = (int)(p - string); + return PCRE_UTF8_ERR20; + } + + if (c >= 0xfe) /* Invalid 0xfe or 0xff bytes */ + { + *erroroffset = (int)(p - string); + return PCRE_UTF8_ERR21; + } + + ab = PRIV(utf8_table4)[c & 0x3f]; /* Number of additional bytes */ + if (length < ab) + { + *erroroffset = (int)(p - string); /* Missing bytes */ + return ab - length; /* Codes ERR1 to ERR5 */ + } + length -= ab; /* Length remaining */ + /* Check top bits in the second byte */ - if ((*(++p) & 0xc0) != 0x80) return p - string; - /* Check for overlong sequences for each different length, and for the - excluded range 0xd000 to 0xdfff. */ + if (((d = *(++p)) & 0xc0) != 0x80) + { + *erroroffset = (int)(p - string) - 1; + return PCRE_UTF8_ERR6; + } + /* For each length, check that the remaining bytes start with the 0x80 bit + set and not the 0x40 bit. Then check for an overlong sequence, and for the + excluded range 0xd800 to 0xdfff. */ + switch (ab) { - /* Check for xx00 000x (overlong sequence) */ + /* 2-byte character. No further bytes to check for 0x80. Check first byte + for for xx00 000x (overlong sequence). */ - case 1: - if ((c & 0x3e) == 0) return p - string; - continue; /* We know there aren't any more bytes to check */ + case 1: if ((c & 0x3e) == 0) + { + *erroroffset = (int)(p - string) - 1; + return PCRE_UTF8_ERR15; + } + break; - /* Check for 1110 0000, xx0x xxxx (overlong sequence) or - 1110 1101, 1010 xxxx (0xd000 - 0xdfff) */ + /* 3-byte character. Check third byte for 0x80. Then check first 2 bytes + for 1110 0000, xx0x xxxx (overlong sequence) or + 1110 1101, 1010 xxxx (0xd800 - 0xdfff) */ case 2: - if ((c == 0xe0 && (*p & 0x20) == 0) || - (c == 0xed && *p >= 0xa0)) - return p - string; + if ((*(++p) & 0xc0) != 0x80) /* Third byte */ + { + *erroroffset = (int)(p - string) - 2; + return PCRE_UTF8_ERR7; + } + if (c == 0xe0 && (d & 0x20) == 0) + { + *erroroffset = (int)(p - string) - 2; + return PCRE_UTF8_ERR16; + } + if (c == 0xed && d >= 0xa0) + { + *erroroffset = (int)(p - string) - 2; + return PCRE_UTF8_ERR14; + } break; - /* Check for 1111 0000, xx00 xxxx (overlong sequence) or - greater than 0x0010ffff (f4 8f bf bf) */ + /* 4-byte character. Check 3rd and 4th bytes for 0x80. Then check first 2 + bytes for for 1111 0000, xx00 xxxx (overlong sequence), then check for a + character greater than 0x0010ffff (f4 8f bf bf) */ case 3: - if ((c == 0xf0 && (*p & 0x30) == 0) || - (c > 0xf4 ) || - (c == 0xf4 && *p > 0x8f)) - return p - string; + if ((*(++p) & 0xc0) != 0x80) /* Third byte */ + { + *erroroffset = (int)(p - string) - 2; + return PCRE_UTF8_ERR7; + } + if ((*(++p) & 0xc0) != 0x80) /* Fourth byte */ + { + *erroroffset = (int)(p - string) - 3; + return PCRE_UTF8_ERR8; + } + if (c == 0xf0 && (d & 0x30) == 0) + { + *erroroffset = (int)(p - string) - 3; + return PCRE_UTF8_ERR17; + } + if (c > 0xf4 || (c == 0xf4 && d > 0x8f)) + { + *erroroffset = (int)(p - string) - 3; + return PCRE_UTF8_ERR13; + } break; -#if 0 - /* These cases can no longer occur, as we restrict to a maximum of four - bytes nowadays. Leave the code here in case we ever want to add an option - for longer sequences. */ + /* 5-byte and 6-byte characters are not allowed by RFC 3629, and will be + rejected by the length test below. However, we do the appropriate tests + here so that overlong sequences get diagnosed, and also in case there is + ever an option for handling these larger code points. */ - /* Check for 1111 1000, xx00 0xxx */ + /* 5-byte character. Check 3rd, 4th, and 5th bytes for 0x80. Then check for + 1111 1000, xx00 0xxx */ + case 4: - if (c == 0xf8 && (*p & 0x38) == 0) return p - string; + if ((*(++p) & 0xc0) != 0x80) /* Third byte */ + { + *erroroffset = (int)(p - string) - 2; + return PCRE_UTF8_ERR7; + } + if ((*(++p) & 0xc0) != 0x80) /* Fourth byte */ + { + *erroroffset = (int)(p - string) - 3; + return PCRE_UTF8_ERR8; + } + if ((*(++p) & 0xc0) != 0x80) /* Fifth byte */ + { + *erroroffset = (int)(p - string) - 4; + return PCRE_UTF8_ERR9; + } + if (c == 0xf8 && (d & 0x38) == 0) + { + *erroroffset = (int)(p - string) - 4; + return PCRE_UTF8_ERR18; + } break; - /* Check for leading 0xfe or 0xff, and then for 1111 1100, xx00 00xx */ + /* 6-byte character. Check 3rd-6th bytes for 0x80. Then check for + 1111 1100, xx00 00xx. */ + case 5: - if (c == 0xfe || c == 0xff || - (c == 0xfc && (*p & 0x3c) == 0)) return p - string; + if ((*(++p) & 0xc0) != 0x80) /* Third byte */ + { + *erroroffset = (int)(p - string) - 2; + return PCRE_UTF8_ERR7; + } + if ((*(++p) & 0xc0) != 0x80) /* Fourth byte */ + { + *erroroffset = (int)(p - string) - 3; + return PCRE_UTF8_ERR8; + } + if ((*(++p) & 0xc0) != 0x80) /* Fifth byte */ + { + *erroroffset = (int)(p - string) - 4; + return PCRE_UTF8_ERR9; + } + if ((*(++p) & 0xc0) != 0x80) /* Sixth byte */ + { + *erroroffset = (int)(p - string) - 5; + return PCRE_UTF8_ERR10; + } + if (c == 0xfc && (d & 0x3c) == 0) + { + *erroroffset = (int)(p - string) - 5; + return PCRE_UTF8_ERR19; + } break; -#endif - } - /* Check for valid bytes after the 2nd, if any; all must start 10 */ - while (--ab > 0) + /* Character is valid under RFC 2279, but 4-byte and 5-byte characters are + excluded by RFC 3629. The pointer p is currently at the last byte of the + character. */ + + if (ab > 3) { - if ((*(++p) & 0xc0) != 0x80) return p - string; + *erroroffset = (int)(p - string) - ab; + return (ab == 4)? PCRE_UTF8_ERR11 : PCRE_UTF8_ERR12; } } + +#else /* SUPPORT_UTF */ +(void)(string); /* Keep picky compilers happy */ +(void)(length); #endif -return -1; +return PCRE_UTF8_ERR0; /* This indicates success */ } /* End of pcre_valid_utf8.c */ diff --git a/pcre/pcre_version.c b/pcre/pcre_version.c --- a/pcre/pcre_version.c +++ b/pcre/pcre_version.c @@ -6,7 +6,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -79,8 +79,13 @@ pre-processor time. This hack uses a standard trick for avoiding calling the STRING macro with an empty argument when doing the test. */ -PCRE_EXP_DEFN const char * +#ifdef COMPILE_PCRE8 +PCRE_EXP_DEFN const char * PCRE_CALL_CONVENTION pcre_version(void) +#else +PCRE_EXP_DEFN const char * PCRE_CALL_CONVENTION +pcre16_version(void) +#endif { return (XSTRING(Z PCRE_PRERELEASE)[1] == 0)? XSTRING(PCRE_MAJOR.PCRE_MINOR PCRE_DATE) : diff --git a/pcre/pcre_xclass.c b/pcre/pcre_xclass.c --- a/pcre/pcre_xclass.c +++ b/pcre/pcre_xclass.c @@ -6,7 +6,7 @@ and semantics are as close as possible to those of the Perl 5 language. Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -39,8 +39,7 @@ /* This module contains an internal function that is used to match an extended -class (one that contains characters whose values are > 255). It is used by both -pcre_exec() and pcre_def_exec(). */ +class. It is used by both pcre_exec() and pcre_def_exec(). */ #ifdef HAVE_CONFIG_H @@ -55,7 +54,7 @@ *************************************************/ /* This function is called to match a character against an extended class that -might contain values > 255. +might contain values > 255 and/or Unicode properties. Arguments: c the character @@ -65,47 +64,70 @@ */ BOOL -_pcre_xclass(int c, const uschar *data) +PRIV(xclass)(int c, const pcre_uchar *data, BOOL utf) { int t; BOOL negated = (*data & XCL_NOT) != 0; +(void)utf; +#ifdef COMPILE_PCRE8 +/* In 8 bit mode, this must always be TRUE. Help the compiler to know that. */ +utf = TRUE; +#endif + /* Character values < 256 are matched against a bitmap, if one is present. If not, we still carry on, because there may be ranges that start below 256 in the additional data. */ if (c < 256) { - if ((*data & XCL_MAP) != 0 && (data[1 + c/8] & (1 << (c&7))) != 0) - return !negated; /* char found */ + if ((*data & XCL_MAP) != 0 && + (((pcre_uint8 *)(data + 1))[c/8] & (1 << (c&7))) != 0) + return !negated; /* char found */ } /* First skip the bit map if present. Then match against the list of Unicode properties or large chars or ranges that end with a large char. We won't ever encounter XCL_PROP or XCL_NOTPROP when UCP support is not compiled. */ -if ((*data++ & XCL_MAP) != 0) data += 32; +if ((*data++ & XCL_MAP) != 0) data += 32 / sizeof(pcre_uchar); while ((t = *data++) != XCL_END) { int x, y; if (t == XCL_SINGLE) { - GETCHARINC(x, data); +#ifdef SUPPORT_UTF + if (utf) + { + GETCHARINC(x, data); /* macro generates multiple statements */ + } + else +#endif + x = *data++; if (c == x) return !negated; } else if (t == XCL_RANGE) { - GETCHARINC(x, data); - GETCHARINC(y, data); +#ifdef SUPPORT_UTF + if (utf) + { + GETCHARINC(x, data); /* macro generates multiple statements */ + GETCHARINC(y, data); /* macro generates multiple statements */ + } + else +#endif + { + x = *data++; + y = *data++; + } if (c >= x && c <= y) return !negated; } #ifdef SUPPORT_UCP else /* XCL_PROP & XCL_NOTPROP */ { - int chartype, script; - int category = _pcre_ucp_findprop(c, &chartype, &script); + const ucd_record *prop = GET_UCD(c); switch(*data) { @@ -114,20 +136,48 @@ break; case PT_LAMP: - if ((chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt) == - (t == XCL_PROP)) return !negated; + if ((prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || + prop->chartype == ucp_Lt) == (t == XCL_PROP)) return !negated; break; case PT_GC: - if ((data[1] == category) == (t == XCL_PROP)) return !negated; + if ((data[1] == PRIV(ucp_gentype)[prop->chartype]) == (t == XCL_PROP)) + return !negated; break; case PT_PC: - if ((data[1] == chartype) == (t == XCL_PROP)) return !negated; + if ((data[1] == prop->chartype) == (t == XCL_PROP)) return !negated; break; case PT_SC: - if ((data[1] == script) == (t == XCL_PROP)) return !negated; + if ((data[1] == prop->script) == (t == XCL_PROP)) return !negated; + break; + + case PT_ALNUM: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (t == XCL_PROP)) + return !negated; + break; + + case PT_SPACE: /* Perl space */ + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR) + == (t == XCL_PROP)) + return !negated; + break; + + case PT_PXSPACE: /* POSIX space */ + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z || + c == CHAR_HT || c == CHAR_NL || c == CHAR_VT || + c == CHAR_FF || c == CHAR_CR) == (t == XCL_PROP)) + return !negated; + break; + + case PT_WORD: + if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L || + PRIV(ucp_gentype)[prop->chartype] == ucp_N || c == CHAR_UNDERSCORE) + == (t == XCL_PROP)) + return !negated; break; /* This should never occur, but compilers may mutter if there is no diff --git a/pcre/pcrecpp.h b/pcre/pcrecpp.h --- a/pcre/pcrecpp.h +++ b/pcre/pcrecpp.h @@ -646,7 +646,13 @@ // regexp wasn't valid on construction. int NumberOfCapturingGroups() const; - // The default value for an argument, to indicate no arg was passed in + // The default value for an argument, to indicate the end of the argument + // list. This must be used only in optional argument defaults. It should NOT + // be passed explicitly. Some people have tried to use it like this: + // + // FullMatch(x, y, &z, no_arg, &w); + // + // This is a mistake, and will not work. static Arg no_arg; private: @@ -668,6 +674,7 @@ int TryMatch(const StringPiece& text, int startpos, Anchor anchor, + bool empty_ok, int *vec, int vecsize) const; diff --git a/pcre/pcrecpp_internal.h b/pcre/pcrecpp_internal.h --- a/pcre/pcrecpp_internal.h +++ b/pcre/pcrecpp_internal.h @@ -3,6 +3,9 @@ *************************************************/ /* +Copyright (c) 2005, Google Inc. +All rights reserved. + ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: diff --git a/pcre/pcrecpparg.h b/pcre/pcrecpparg.h deleted file mode 100644 --- a/pcre/pcrecpparg.h +++ /dev/null @@ -1,174 +0,0 @@ -// Copyright (c) 2005, Google Inc. -// All rights reserved. -// -// Redistribution and use in source and binary forms, with or without -// modification, are permitted provided that the following conditions are -// met: -// -// * Redistributions of source code must retain the above copyright -// notice, this list of conditions and the following disclaimer. -// * Redistributions in binary form must reproduce the above -// copyright notice, this list of conditions and the following disclaimer -// in the documentation and/or other materials provided with the -// distribution. -// * Neither the name of Google Inc. nor the names of its -// contributors may be used to endorse or promote products derived from -// this software without specific prior written permission. -// -// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS -// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR -// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT -// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, -// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT -// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, -// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY -// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT -// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE -// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. -// -// Author: Sanjay Ghemawat - -#ifndef _PCRECPPARG_H -#define _PCRECPPARG_H - -#include // for NULL -#include - -#include - -namespace pcrecpp { - -class StringPiece; - -// Hex/Octal/Binary? - -// Special class for parsing into objects that define a ParseFrom() method -template -class _RE_MatchObject { - public: - static inline bool Parse(const char* str, int n, void* dest) { - if (dest == NULL) return true; - T* object = reinterpret_cast(dest); - return object->ParseFrom(str, n); - } -}; - -class PCRECPP_EXP_DEFN Arg { - public: - // Empty constructor so we can declare arrays of Arg - Arg(); - - // Constructor specially designed for NULL arguments - Arg(void*); - - typedef bool (*Parser)(const char* str, int n, void* dest); - -// Type-specific parsers -#define PCRE_MAKE_PARSER(type,name) \ - Arg(type* p) : arg_(p), parser_(name) { } \ - Arg(type* p, Parser parser) : arg_(p), parser_(parser) { } - - - PCRE_MAKE_PARSER(char, parse_char); - PCRE_MAKE_PARSER(unsigned char, parse_uchar); - PCRE_MAKE_PARSER(short, parse_short); - PCRE_MAKE_PARSER(unsigned short, parse_ushort); - PCRE_MAKE_PARSER(int, parse_int); - PCRE_MAKE_PARSER(unsigned int, parse_uint); - PCRE_MAKE_PARSER(long, parse_long); - PCRE_MAKE_PARSER(unsigned long, parse_ulong); -#if 1 - PCRE_MAKE_PARSER(long long, parse_longlong); -#endif -#if 1 - PCRE_MAKE_PARSER(unsigned long long, parse_ulonglong); -#endif - PCRE_MAKE_PARSER(float, parse_float); - PCRE_MAKE_PARSER(double, parse_double); - PCRE_MAKE_PARSER(std::string, parse_string); - PCRE_MAKE_PARSER(StringPiece, parse_stringpiece); - -#undef PCRE_MAKE_PARSER - - // Generic constructor - template Arg(T*, Parser parser); - // Generic constructor template - template Arg(T* p) - : arg_(p), parser_(_RE_MatchObject::Parse) { - } - - // Parse the data - bool Parse(const char* str, int n) const; - - private: - void* arg_; - Parser parser_; - - static bool parse_null (const char* str, int n, void* dest); - static bool parse_char (const char* str, int n, void* dest); - static bool parse_uchar (const char* str, int n, void* dest); - static bool parse_float (const char* str, int n, void* dest); - static bool parse_double (const char* str, int n, void* dest); - static bool parse_string (const char* str, int n, void* dest); - static bool parse_stringpiece (const char* str, int n, void* dest); - -#define PCRE_DECLARE_INTEGER_PARSER(name) \ - private: \ - static bool parse_ ## name(const char* str, int n, void* dest); \ - static bool parse_ ## name ## _radix( \ - const char* str, int n, void* dest, int radix); \ - public: \ - static bool parse_ ## name ## _hex(const char* str, int n, void* dest); \ - static bool parse_ ## name ## _octal(const char* str, int n, void* dest); \ - static bool parse_ ## name ## _cradix(const char* str, int n, void* dest) - - PCRE_DECLARE_INTEGER_PARSER(short); - PCRE_DECLARE_INTEGER_PARSER(ushort); - PCRE_DECLARE_INTEGER_PARSER(int); - PCRE_DECLARE_INTEGER_PARSER(uint); - PCRE_DECLARE_INTEGER_PARSER(long); - PCRE_DECLARE_INTEGER_PARSER(ulong); - PCRE_DECLARE_INTEGER_PARSER(longlong); - PCRE_DECLARE_INTEGER_PARSER(ulonglong); - -#undef PCRE_DECLARE_INTEGER_PARSER -}; - -inline Arg::Arg() : arg_(NULL), parser_(parse_null) { } -inline Arg::Arg(void* p) : arg_(p), parser_(parse_null) { } - -inline bool Arg::Parse(const char* str, int n) const { - return (*parser_)(str, n, arg_); -} - -// This part of the parser, appropriate only for ints, deals with bases -#define MAKE_INTEGER_PARSER(type, name) \ - inline Arg Hex(type* ptr) { \ - return Arg(ptr, Arg::parse_ ## name ## _hex); } \ - inline Arg Octal(type* ptr) { \ - return Arg(ptr, Arg::parse_ ## name ## _octal); } \ - inline Arg CRadix(type* ptr) { \ - return Arg(ptr, Arg::parse_ ## name ## _cradix); } - -MAKE_INTEGER_PARSER(short, short) /* */ -MAKE_INTEGER_PARSER(unsigned short, ushort) /* */ -MAKE_INTEGER_PARSER(int, int) /* Don't use semicolons */ -MAKE_INTEGER_PARSER(unsigned int, uint) /* after these statement */ -MAKE_INTEGER_PARSER(long, long) /* because they can cause */ -MAKE_INTEGER_PARSER(unsigned long, ulong) /* compiler warnings if */ -#if 1 /* the checking level is */ -MAKE_INTEGER_PARSER(long long, longlong) /* turned up high enough. */ -#endif /* */ -#if 1 /* */ -MAKE_INTEGER_PARSER(unsigned long long, ulonglong) /* */ -#endif - -#undef PCRE_IS_SET -#undef PCRE_SET_OR_CLEAR -#undef MAKE_INTEGER_PARSER - -} // namespace pcrecpp - - -#endif /* _PCRECPPARG_H */ diff --git a/pcre/pcredemo.c b/pcre/pcredemo.c deleted file mode 100644 --- a/pcre/pcredemo.c +++ /dev/null @@ -1,334 +0,0 @@ -/************************************************* -* PCRE DEMONSTRATION PROGRAM * -*************************************************/ - -/* This is a demonstration program to illustrate the most straightforward ways -of calling the PCRE regular expression library from a C program. See the -pcresample documentation for a short discussion ("man pcresample" if you have -the PCRE man pages installed). - -In Unix-like environments, compile this program thuswise: - - gcc -Wall pcredemo.c -I/usr/local/include -L/usr/local/lib \ - -R/usr/local/lib -lpcre - -Replace "/usr/local/include" and "/usr/local/lib" with wherever the include and -library files for PCRE are installed on your system. You don't need -I and -L -if PCRE is installed in the standard system libraries. Only some operating -systems (e.g. Solaris) use the -R option. - -Building under Windows: - -If you want to statically link this program against a non-dll .a file, you must -define PCRE_STATIC before including pcre.h, otherwise the pcre_malloc() and -pcre_free() exported functions will be declared __declspec(dllimport), with -unwanted results. So in this environment, uncomment the following line. */ - -/* #define PCRE_STATIC */ - -#include -#include -#include - -#define OVECCOUNT 30 /* should be a multiple of 3 */ - - -int main(int argc, char **argv) -{ -pcre *re; -const char *error; -char *pattern; -char *subject; -unsigned char *name_table; -int erroffset; -int find_all; -int namecount; -int name_entry_size; -int ovector[OVECCOUNT]; -int subject_length; -int rc, i; - - -/************************************************************************** -* First, sort out the command line. There is only one possible option at * -* the moment, "-g" to request repeated matching to find all occurrences, * -* like Perl's /g option. We set the variable find_all to a non-zero value * -* if the -g option is present. Apart from that, there must be exactly two * -* arguments. * -**************************************************************************/ - -find_all = 0; -for (i = 1; i < argc; i++) - { - if (strcmp(argv[i], "-g") == 0) find_all = 1; - else break; - } - -/* After the options, we require exactly two arguments, which are the pattern, -and the subject string. */ - -if (argc - i != 2) - { - printf("Two arguments required: a regex and a subject string\n"); - return 1; - } - -pattern = argv[i]; -subject = argv[i+1]; -subject_length = (int)strlen(subject); - - -/************************************************************************* -* Now we are going to compile the regular expression pattern, and handle * -* and errors that are detected. * -*************************************************************************/ - -re = pcre_compile( - pattern, /* the pattern */ - 0, /* default options */ - &error, /* for error message */ - &erroffset, /* for error offset */ - NULL); /* use default character tables */ - -/* Compilation failed: print the error message and exit */ - -if (re == NULL) - { - printf("PCRE compilation failed at offset %d: %s\n", erroffset, error); - return 1; - } - - -/************************************************************************* -* If the compilation succeeded, we call PCRE again, in order to do a * -* pattern match against the subject string. This does just ONE match. If * -* further matching is needed, it will be done below. * -*************************************************************************/ - -rc = pcre_exec( - re, /* the compiled pattern */ - NULL, /* no extra data - we didn't study the pattern */ - subject, /* the subject string */ - subject_length, /* the length of the subject */ - 0, /* start at offset 0 in the subject */ - 0, /* default options */ - ovector, /* output vector for substring information */ - OVECCOUNT); /* number of elements in the output vector */ - -/* Matching failed: handle error cases */ - -if (rc < 0) - { - switch(rc) - { - case PCRE_ERROR_NOMATCH: printf("No match\n"); break; - /* - Handle other special cases if you like - */ - default: printf("Matching error %d\n", rc); break; - } - pcre_free(re); /* Release memory used for the compiled pattern */ - return 1; - } - -/* Match succeded */ - -printf("\nMatch succeeded at offset %d\n", ovector[0]); - - -/************************************************************************* -* We have found the first match within the subject string. If the output * -* vector wasn't big enough, say so. Then output any substrings that were * -* captured. * -*************************************************************************/ - -/* The output vector wasn't big enough */ - -if (rc == 0) - { - rc = OVECCOUNT/3; - printf("ovector only has room for %d captured substrings\n", rc - 1); - } - -/* Show substrings stored in the output vector by number. Obviously, in a real -application you might want to do things other than print them. */ - -for (i = 0; i < rc; i++) - { - char *substring_start = subject + ovector[2*i]; - int substring_length = ovector[2*i+1] - ovector[2*i]; - printf("%2d: %.*s\n", i, substring_length, substring_start); - } - - -/************************************************************************** -* That concludes the basic part of this demonstration program. We have * -* compiled a pattern, and performed a single match. The code that follows * -* shows first how to access named substrings, and then how to code for * -* repeated matches on the same subject. * -**************************************************************************/ - -/* See if there are any named substrings, and if so, show them by name. First -we have to extract the count of named parentheses from the pattern. */ - -(void)pcre_fullinfo( - re, /* the compiled pattern */ - NULL, /* no extra data - we didn't study the pattern */ - PCRE_INFO_NAMECOUNT, /* number of named substrings */ - &namecount); /* where to put the answer */ - -if (namecount <= 0) printf("No named substrings\n"); else - { - unsigned char *tabptr; - printf("Named substrings\n"); - - /* Before we can access the substrings, we must extract the table for - translating names to numbers, and the size of each entry in the table. */ - - (void)pcre_fullinfo( - re, /* the compiled pattern */ - NULL, /* no extra data - we didn't study the pattern */ - PCRE_INFO_NAMETABLE, /* address of the table */ - &name_table); /* where to put the answer */ - - (void)pcre_fullinfo( - re, /* the compiled pattern */ - NULL, /* no extra data - we didn't study the pattern */ - PCRE_INFO_NAMEENTRYSIZE, /* size of each entry in the table */ - &name_entry_size); /* where to put the answer */ - - /* Now we can scan the table and, for each entry, print the number, the name, - and the substring itself. */ - - tabptr = name_table; - for (i = 0; i < namecount; i++) - { - int n = (tabptr[0] << 8) | tabptr[1]; - printf("(%d) %*s: %.*s\n", n, name_entry_size - 3, tabptr + 2, - ovector[2*n+1] - ovector[2*n], subject + ovector[2*n]); - tabptr += name_entry_size; - } - } - - -/************************************************************************* -* If the "-g" option was given on the command line, we want to continue * -* to search for additional matches in the subject string, in a similar * -* way to the /g option in Perl. This turns out to be trickier than you * -* might think because of the possibility of matching an empty string. * -* What happens is as follows: * -* * -* If the previous match was NOT for an empty string, we can just start * -* the next match at the end of the previous one. * -* * -* If the previous match WAS for an empty string, we can't do that, as it * -* would lead to an infinite loop. Instead, a special call of pcre_exec() * -* is made with the PCRE_NOTEMPTY and PCRE_ANCHORED flags set. The first * -* of these tells PCRE that an empty string is not a valid match; other * -* possibilities must be tried. The second flag restricts PCRE to one * -* match attempt at the initial string position. If this match succeeds, * -* an alternative to the empty string match has been found, and we can * -* proceed round the loop. * -*************************************************************************/ - -if (!find_all) - { - pcre_free(re); /* Release the memory used for the compiled pattern */ - return 0; /* Finish unless -g was given */ - } - -/* Loop for second and subsequent matches */ - -for (;;) - { - int options = 0; /* Normally no options */ - int start_offset = ovector[1]; /* Start at end of previous match */ - - /* If the previous match was for an empty string, we are finished if we are - at the end of the subject. Otherwise, arrange to run another match at the - same point to see if a non-empty match can be found. */ - - if (ovector[0] == ovector[1]) - { - if (ovector[0] == subject_length) break; - options = PCRE_NOTEMPTY | PCRE_ANCHORED; - } - - /* Run the next matching operation */ - - rc = pcre_exec( - re, /* the compiled pattern */ - NULL, /* no extra data - we didn't study the pattern */ - subject, /* the subject string */ - subject_length, /* the length of the subject */ - start_offset, /* starting offset in the subject */ - options, /* options */ - ovector, /* output vector for substring information */ - OVECCOUNT); /* number of elements in the output vector */ - - /* This time, a result of NOMATCH isn't an error. If the value in "options" - is zero, it just means we have found all possible matches, so the loop ends. - Otherwise, it means we have failed to find a non-empty-string match at a - point where there was a previous empty-string match. In this case, we do what - Perl does: advance the matching position by one, and continue. We do this by - setting the "end of previous match" offset, because that is picked up at the - top of the loop as the point at which to start again. */ - - if (rc == PCRE_ERROR_NOMATCH) - { - if (options == 0) break; - ovector[1] = start_offset + 1; - continue; /* Go round the loop again */ - } - - /* Other matching errors are not recoverable. */ - - if (rc < 0) - { - printf("Matching error %d\n", rc); - pcre_free(re); /* Release memory used for the compiled pattern */ - return 1; - } - - /* Match succeded */ - - printf("\nMatch succeeded again at offset %d\n", ovector[0]); - - /* The match succeeded, but the output vector wasn't big enough. */ - - if (rc == 0) - { - rc = OVECCOUNT/3; - printf("ovector only has room for %d captured substrings\n", rc - 1); - } - - /* As before, show substrings stored in the output vector by number, and then - also any named substrings. */ - - for (i = 0; i < rc; i++) - { - char *substring_start = subject + ovector[2*i]; - int substring_length = ovector[2*i+1] - ovector[2*i]; - printf("%2d: %.*s\n", i, substring_length, substring_start); - } - - if (namecount <= 0) printf("No named substrings\n"); else - { - unsigned char *tabptr = name_table; - printf("Named substrings\n"); - for (i = 0; i < namecount; i++) - { - int n = (tabptr[0] << 8) | tabptr[1]; - printf("(%d) %*s: %.*s\n", n, name_entry_size - 3, tabptr + 2, - ovector[2*n+1] - ovector[2*n], subject + ovector[2*n]); - tabptr += name_entry_size; - } - } - } /* End of loop to find second and subsequent matches */ - -printf("\n"); -pcre_free(re); /* Release memory used for the compiled pattern */ -return 0; -} - -/* End of pcredemo.c */ diff --git a/pcre/pcregrep.c b/pcre/pcregrep.c deleted file mode 100644 --- a/pcre/pcregrep.c +++ /dev/null @@ -1,2379 +0,0 @@ -/************************************************* -* pcregrep program * -*************************************************/ - -/* This is a grep program that uses the PCRE regular expression library to do -its pattern matching. On a Unix or Win32 system it can recurse into -directories. - - Copyright (c) 1997-2008 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - -#ifdef HAVE_CONFIG_H -#include "config.h" -#endif - -#include -#include -#include -#include -#include -#include - -#include -#include - -#ifdef HAVE_UNISTD_H -#include -#endif - -#ifdef SUPPORT_LIBZ -#include -#endif - -#ifdef SUPPORT_LIBBZ2 -#include -#endif - -#include "pcre.h" - -#define FALSE 0 -#define TRUE 1 - -typedef int BOOL; - -#define MAX_PATTERN_COUNT 100 - -#if BUFSIZ > 8192 -#define MBUFTHIRD BUFSIZ -#else -#define MBUFTHIRD 8192 -#endif - -/* Values for the "filenames" variable, which specifies options for file name -output. The order is important; it is assumed that a file name is wanted for -all values greater than FN_DEFAULT. */ - -enum { FN_NONE, FN_DEFAULT, FN_ONLY, FN_NOMATCH_ONLY, FN_FORCE }; - -/* File reading styles */ - -enum { FR_PLAIN, FR_LIBZ, FR_LIBBZ2 }; - -/* Actions for the -d and -D options */ - -enum { dee_READ, dee_SKIP, dee_RECURSE }; -enum { DEE_READ, DEE_SKIP }; - -/* Actions for special processing options (flag bits) */ - -#define PO_WORD_MATCH 0x0001 -#define PO_LINE_MATCH 0x0002 -#define PO_FIXED_STRINGS 0x0004 - -/* Line ending types */ - -enum { EL_LF, EL_CR, EL_CRLF, EL_ANY, EL_ANYCRLF }; - - - -/************************************************* -* Global variables * -*************************************************/ - -/* Jeffrey Friedl has some debugging requirements that are not part of the -regular code. */ - -#ifdef JFRIEDL_DEBUG -static int S_arg = -1; -static unsigned int jfriedl_XR = 0; /* repeat regex attempt this many times */ -static unsigned int jfriedl_XT = 0; /* replicate text this many times */ -static const char *jfriedl_prefix = ""; -static const char *jfriedl_postfix = ""; -#endif - -static int endlinetype; - -static char *colour_string = (char *)"1;31"; -static char *colour_option = NULL; -static char *dee_option = NULL; -static char *DEE_option = NULL; -static char *newline = NULL; -static char *pattern_filename = NULL; -static char *stdin_name = (char *)"(standard input)"; -static char *locale = NULL; - -static const unsigned char *pcretables = NULL; - -static int pattern_count = 0; -static pcre **pattern_list = NULL; -static pcre_extra **hints_list = NULL; - -static char *include_pattern = NULL; -static char *exclude_pattern = NULL; -static char *include_dir_pattern = NULL; -static char *exclude_dir_pattern = NULL; - -static pcre *include_compiled = NULL; -static pcre *exclude_compiled = NULL; -static pcre *include_dir_compiled = NULL; -static pcre *exclude_dir_compiled = NULL; - -static int after_context = 0; -static int before_context = 0; -static int both_context = 0; -static int dee_action = dee_READ; -static int DEE_action = DEE_READ; -static int error_count = 0; -static int filenames = FN_DEFAULT; -static int process_options = 0; - -static BOOL count_only = FALSE; -static BOOL do_colour = FALSE; -static BOOL file_offsets = FALSE; -static BOOL hyphenpending = FALSE; -static BOOL invert = FALSE; -static BOOL line_offsets = FALSE; -static BOOL multiline = FALSE; -static BOOL number = FALSE; -static BOOL only_matching = FALSE; -static BOOL quiet = FALSE; -static BOOL silent = FALSE; -static BOOL utf8 = FALSE; - -/* Structure for options and list of them */ - -enum { OP_NODATA, OP_STRING, OP_OP_STRING, OP_NUMBER, OP_OP_NUMBER, - OP_PATLIST }; - -typedef struct option_item { - int type; - int one_char; - void *dataptr; - const char *long_name; - const char *help_text; -} option_item; - -/* Options without a single-letter equivalent get a negative value. This can be -used to identify them. */ - -#define N_COLOUR (-1) -#define N_EXCLUDE (-2) -#define N_EXCLUDE_DIR (-3) -#define N_HELP (-4) -#define N_INCLUDE (-5) -#define N_INCLUDE_DIR (-6) -#define N_LABEL (-7) -#define N_LOCALE (-8) -#define N_NULL (-9) -#define N_LOFFSETS (-10) -#define N_FOFFSETS (-11) - -static option_item optionlist[] = { - { OP_NODATA, N_NULL, NULL, "", " terminate options" }, - { OP_NODATA, N_HELP, NULL, "help", "display this help and exit" }, - { OP_NUMBER, 'A', &after_context, "after-context=number", "set number of following context lines" }, - { OP_NUMBER, 'B', &before_context, "before-context=number", "set number of prior context lines" }, - { OP_OP_STRING, N_COLOUR, &colour_option, "color=option", "matched text color option" }, - { OP_NUMBER, 'C', &both_context, "context=number", "set number of context lines, before & after" }, - { OP_NODATA, 'c', NULL, "count", "print only a count of matching lines per FILE" }, - { OP_OP_STRING, N_COLOUR, &colour_option, "colour=option", "matched text colour option" }, - { OP_STRING, 'D', &DEE_option, "devices=action","how to handle devices, FIFOs, and sockets" }, - { OP_STRING, 'd', &dee_option, "directories=action", "how to handle directories" }, - { OP_PATLIST, 'e', NULL, "regex(p)", "specify pattern (may be used more than once)" }, - { OP_NODATA, 'F', NULL, "fixed_strings", "patterns are sets of newline-separated strings" }, - { OP_STRING, 'f', &pattern_filename, "file=path", "read patterns from file" }, - { OP_NODATA, N_FOFFSETS, NULL, "file-offsets", "output file offsets, not text" }, - { OP_NODATA, 'H', NULL, "with-filename", "force the prefixing filename on output" }, - { OP_NODATA, 'h', NULL, "no-filename", "suppress the prefixing filename on output" }, - { OP_NODATA, 'i', NULL, "ignore-case", "ignore case distinctions" }, - { OP_NODATA, 'l', NULL, "files-with-matches", "print only FILE names containing matches" }, - { OP_NODATA, 'L', NULL, "files-without-match","print only FILE names not containing matches" }, - { OP_STRING, N_LABEL, &stdin_name, "label=name", "set name for standard input" }, - { OP_NODATA, N_LOFFSETS, NULL, "line-offsets", "output line numbers and offsets, not text" }, - { OP_STRING, N_LOCALE, &locale, "locale=locale", "use the named locale" }, - { OP_NODATA, 'M', NULL, "multiline", "run in multiline mode" }, - { OP_STRING, 'N', &newline, "newline=type", "set newline type (CR, LF, CRLF, ANYCRLF or ANY)" }, - { OP_NODATA, 'n', NULL, "line-number", "print line number with output lines" }, - { OP_NODATA, 'o', NULL, "only-matching", "show only the part of the line that matched" }, - { OP_NODATA, 'q', NULL, "quiet", "suppress output, just set return code" }, - { OP_NODATA, 'r', NULL, "recursive", "recursively scan sub-directories" }, - { OP_STRING, N_EXCLUDE,&exclude_pattern, "exclude=pattern","exclude matching files when recursing" }, - { OP_STRING, N_INCLUDE,&include_pattern, "include=pattern","include matching files when recursing" }, - { OP_STRING, N_EXCLUDE_DIR,&exclude_dir_pattern, "exclude_dir=pattern","exclude matching directories when recursing" }, - { OP_STRING, N_INCLUDE_DIR,&include_dir_pattern, "include_dir=pattern","include matching directories when recursing" }, -#ifdef JFRIEDL_DEBUG - { OP_OP_NUMBER, 'S', &S_arg, "jeffS", "replace matched (sub)string with X" }, -#endif - { OP_NODATA, 's', NULL, "no-messages", "suppress error messages" }, - { OP_NODATA, 'u', NULL, "utf-8", "use UTF-8 mode" }, - { OP_NODATA, 'V', NULL, "version", "print version information and exit" }, - { OP_NODATA, 'v', NULL, "invert-match", "select non-matching lines" }, - { OP_NODATA, 'w', NULL, "word-regex(p)", "force patterns to match only as words" }, - { OP_NODATA, 'x', NULL, "line-regex(p)", "force patterns to match only whole lines" }, - { OP_NODATA, 0, NULL, NULL, NULL } -}; - -/* Tables for prefixing and suffixing patterns, according to the -w, -x, and -F -options. These set the 1, 2, and 4 bits in process_options, respectively. Note -that the combination of -w and -x has the same effect as -x on its own, so we -can treat them as the same. */ - -static const char *prefix[] = { - "", "\\b", "^(?:", "^(?:", "\\Q", "\\b\\Q", "^(?:\\Q", "^(?:\\Q" }; - -static const char *suffix[] = { - "", "\\b", ")$", ")$", "\\E", "\\E\\b", "\\E)$", "\\E)$" }; - -/* UTF-8 tables - used only when the newline setting is "any". */ - -const int utf8_table3[] = { 0xff, 0x1f, 0x0f, 0x07, 0x03, 0x01}; - -const char utf8_table4[] = { - 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, - 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, - 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, - 3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5 }; - - - -/************************************************* -* OS-specific functions * -*************************************************/ - -/* These functions are defined so that they can be made system specific, -although at present the only ones are for Unix, Win32, and for "no support". */ - - -/************* Directory scanning in Unix ***********/ - -#if defined HAVE_SYS_STAT_H && defined HAVE_DIRENT_H && defined HAVE_SYS_TYPES_H -#include -#include -#include - -typedef DIR directory_type; - -static int -isdirectory(char *filename) -{ -struct stat statbuf; -if (stat(filename, &statbuf) < 0) - return 0; /* In the expectation that opening as a file will fail */ -return ((statbuf.st_mode & S_IFMT) == S_IFDIR)? '/' : 0; -} - -static directory_type * -opendirectory(char *filename) -{ -return opendir(filename); -} - -static char * -readdirectory(directory_type *dir) -{ -for (;;) - { - struct dirent *dent = readdir(dir); - if (dent == NULL) return NULL; - if (strcmp(dent->d_name, ".") != 0 && strcmp(dent->d_name, "..") != 0) - return dent->d_name; - } -/* Control never reaches here */ -} - -static void -closedirectory(directory_type *dir) -{ -closedir(dir); -} - - -/************* Test for regular file in Unix **********/ - -static int -isregfile(char *filename) -{ -struct stat statbuf; -if (stat(filename, &statbuf) < 0) - return 1; /* In the expectation that opening as a file will fail */ -return (statbuf.st_mode & S_IFMT) == S_IFREG; -} - - -/************* Test stdout for being a terminal in Unix **********/ - -static BOOL -is_stdout_tty(void) -{ -return isatty(fileno(stdout)); -} - - -/************* Directory scanning in Win32 ***********/ - -/* I (Philip Hazel) have no means of testing this code. It was contributed by -Lionel Fourquaux. David Burgess added a patch to define INVALID_FILE_ATTRIBUTES -when it did not exist. David Byron added a patch that moved the #include of - to before the INVALID_FILE_ATTRIBUTES definition rather than after. -*/ - -#elif HAVE_WINDOWS_H - -#ifndef STRICT -# define STRICT -#endif -#ifndef WIN32_LEAN_AND_MEAN -# define WIN32_LEAN_AND_MEAN -#endif - -#include - -#ifndef INVALID_FILE_ATTRIBUTES -#define INVALID_FILE_ATTRIBUTES 0xFFFFFFFF -#endif - -typedef struct directory_type -{ -HANDLE handle; -BOOL first; -WIN32_FIND_DATA data; -} directory_type; - -int -isdirectory(char *filename) -{ -DWORD attr = GetFileAttributes(filename); -if (attr == INVALID_FILE_ATTRIBUTES) - return 0; -return ((attr & FILE_ATTRIBUTE_DIRECTORY) != 0) ? '/' : 0; -} - -directory_type * -opendirectory(char *filename) -{ -size_t len; -char *pattern; -directory_type *dir; -DWORD err; -len = strlen(filename); -pattern = (char *) malloc(len + 3); -dir = (directory_type *) malloc(sizeof(*dir)); -if ((pattern == NULL) || (dir == NULL)) - { - fprintf(stderr, "pcregrep: malloc failed\n"); - exit(2); - } -memcpy(pattern, filename, len); -memcpy(&(pattern[len]), "\\*", 3); -dir->handle = FindFirstFile(pattern, &(dir->data)); -if (dir->handle != INVALID_HANDLE_VALUE) - { - free(pattern); - dir->first = TRUE; - return dir; - } -err = GetLastError(); -free(pattern); -free(dir); -errno = (err == ERROR_ACCESS_DENIED) ? EACCES : ENOENT; -return NULL; -} - -char * -readdirectory(directory_type *dir) -{ -for (;;) - { - if (!dir->first) - { - if (!FindNextFile(dir->handle, &(dir->data))) - return NULL; - } - else - { - dir->first = FALSE; - } - if (strcmp(dir->data.cFileName, ".") != 0 && strcmp(dir->data.cFileName, "..") != 0) - return dir->data.cFileName; - } -#ifndef _MSC_VER -return NULL; /* Keep compiler happy; never executed */ -#endif -} - -void -closedirectory(directory_type *dir) -{ -FindClose(dir->handle); -free(dir); -} - - -/************* Test for regular file in Win32 **********/ - -/* I don't know how to do this, or if it can be done; assume all paths are -regular if they are not directories. */ - -int isregfile(char *filename) -{ -return !isdirectory(filename); -} - - -/************* Test stdout for being a terminal in Win32 **********/ - -/* I don't know how to do this; assume never */ - -static BOOL -is_stdout_tty(void) -{ -return FALSE; -} - - -/************* Directory scanning when we can't do it ***********/ - -/* The type is void, and apart from isdirectory(), the functions do nothing. */ - -#else - -typedef void directory_type; - -int isdirectory(char *filename) { return 0; } -directory_type * opendirectory(char *filename) { return (directory_type*)0;} -char *readdirectory(directory_type *dir) { return (char*)0;} -void closedirectory(directory_type *dir) {} - - -/************* Test for regular when we can't do it **********/ - -/* Assume all files are regular. */ - -int isregfile(char *filename) { return 1; } - - -/************* Test stdout for being a terminal when we can't do it **********/ - -static BOOL -is_stdout_tty(void) -{ -return FALSE; -} - - -#endif - - - -#ifndef HAVE_STRERROR -/************************************************* -* Provide strerror() for non-ANSI libraries * -*************************************************/ - -/* Some old-fashioned systems still around (e.g. SunOS4) don't have strerror() -in their libraries, but can provide the same facility by this simple -alternative function. */ - -extern int sys_nerr; -extern char *sys_errlist[]; - -char * -strerror(int n) -{ -if (n < 0 || n >= sys_nerr) return "unknown error number"; -return sys_errlist[n]; -} -#endif /* HAVE_STRERROR */ - - - -/************************************************* -* Find end of line * -*************************************************/ - -/* The length of the endline sequence that is found is set via lenptr. This may -be zero at the very end of the file if there is no line-ending sequence there. - -Arguments: - p current position in line - endptr end of available data - lenptr where to put the length of the eol sequence - -Returns: pointer to the last byte of the line -*/ - -static char * -end_of_line(char *p, char *endptr, int *lenptr) -{ -switch(endlinetype) - { - default: /* Just in case */ - case EL_LF: - while (p < endptr && *p != '\n') p++; - if (p < endptr) - { - *lenptr = 1; - return p + 1; - } - *lenptr = 0; - return endptr; - - case EL_CR: - while (p < endptr && *p != '\r') p++; - if (p < endptr) - { - *lenptr = 1; - return p + 1; - } - *lenptr = 0; - return endptr; - - case EL_CRLF: - for (;;) - { - while (p < endptr && *p != '\r') p++; - if (++p >= endptr) - { - *lenptr = 0; - return endptr; - } - if (*p == '\n') - { - *lenptr = 2; - return p + 1; - } - } - break; - - case EL_ANYCRLF: - while (p < endptr) - { - int extra = 0; - register int c = *((unsigned char *)p); - - if (utf8 && c >= 0xc0) - { - int gcii, gcss; - extra = utf8_table4[c & 0x3f]; /* Number of additional bytes */ - gcss = 6*extra; - c = (c & utf8_table3[extra]) << gcss; - for (gcii = 1; gcii <= extra; gcii++) - { - gcss -= 6; - c |= (p[gcii] & 0x3f) << gcss; - } - } - - p += 1 + extra; - - switch (c) - { - case 0x0a: /* LF */ - *lenptr = 1; - return p; - - case 0x0d: /* CR */ - if (p < endptr && *p == 0x0a) - { - *lenptr = 2; - p++; - } - else *lenptr = 1; - return p; - - default: - break; - } - } /* End of loop for ANYCRLF case */ - - *lenptr = 0; /* Must have hit the end */ - return endptr; - - case EL_ANY: - while (p < endptr) - { - int extra = 0; - register int c = *((unsigned char *)p); - - if (utf8 && c >= 0xc0) - { - int gcii, gcss; - extra = utf8_table4[c & 0x3f]; /* Number of additional bytes */ - gcss = 6*extra; - c = (c & utf8_table3[extra]) << gcss; - for (gcii = 1; gcii <= extra; gcii++) - { - gcss -= 6; - c |= (p[gcii] & 0x3f) << gcss; - } - } - - p += 1 + extra; - - switch (c) - { - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - *lenptr = 1; - return p; - - case 0x0d: /* CR */ - if (p < endptr && *p == 0x0a) - { - *lenptr = 2; - p++; - } - else *lenptr = 1; - return p; - - case 0x85: /* NEL */ - *lenptr = utf8? 2 : 1; - return p; - - case 0x2028: /* LS */ - case 0x2029: /* PS */ - *lenptr = 3; - return p; - - default: - break; - } - } /* End of loop for ANY case */ - - *lenptr = 0; /* Must have hit the end */ - return endptr; - } /* End of overall switch */ -} - - - -/************************************************* -* Find start of previous line * -*************************************************/ - -/* This is called when looking back for before lines to print. - -Arguments: - p start of the subsequent line - startptr start of available data - -Returns: pointer to the start of the previous line -*/ - -static char * -previous_line(char *p, char *startptr) -{ -switch(endlinetype) - { - default: /* Just in case */ - case EL_LF: - p--; - while (p > startptr && p[-1] != '\n') p--; - return p; - - case EL_CR: - p--; - while (p > startptr && p[-1] != '\n') p--; - return p; - - case EL_CRLF: - for (;;) - { - p -= 2; - while (p > startptr && p[-1] != '\n') p--; - if (p <= startptr + 1 || p[-2] == '\r') return p; - } - return p; /* But control should never get here */ - - case EL_ANY: - case EL_ANYCRLF: - if (*(--p) == '\n' && p > startptr && p[-1] == '\r') p--; - if (utf8) while ((*p & 0xc0) == 0x80) p--; - - while (p > startptr) - { - register int c; - char *pp = p - 1; - - if (utf8) - { - int extra = 0; - while ((*pp & 0xc0) == 0x80) pp--; - c = *((unsigned char *)pp); - if (c >= 0xc0) - { - int gcii, gcss; - extra = utf8_table4[c & 0x3f]; /* Number of additional bytes */ - gcss = 6*extra; - c = (c & utf8_table3[extra]) << gcss; - for (gcii = 1; gcii <= extra; gcii++) - { - gcss -= 6; - c |= (pp[gcii] & 0x3f) << gcss; - } - } - } - else c = *((unsigned char *)pp); - - if (endlinetype == EL_ANYCRLF) switch (c) - { - case 0x0a: /* LF */ - case 0x0d: /* CR */ - return p; - - default: - break; - } - - else switch (c) - { - case 0x0a: /* LF */ - case 0x0b: /* VT */ - case 0x0c: /* FF */ - case 0x0d: /* CR */ - case 0x85: /* NEL */ - case 0x2028: /* LS */ - case 0x2029: /* PS */ - return p; - - default: - break; - } - - p = pp; /* Back one character */ - } /* End of loop for ANY case */ - - return startptr; /* Hit start of data */ - } /* End of overall switch */ -} - - - - - -/************************************************* -* Print the previous "after" lines * -*************************************************/ - -/* This is called if we are about to lose said lines because of buffer filling, -and at the end of the file. The data in the line is written using fwrite() so -that a binary zero does not terminate it. - -Arguments: - lastmatchnumber the number of the last matching line, plus one - lastmatchrestart where we restarted after the last match - endptr end of available data - printname filename for printing - -Returns: nothing -*/ - -static void do_after_lines(int lastmatchnumber, char *lastmatchrestart, - char *endptr, char *printname) -{ -if (after_context > 0 && lastmatchnumber > 0) - { - int count = 0; - while (lastmatchrestart < endptr && count++ < after_context) - { - int ellength; - char *pp = lastmatchrestart; - if (printname != NULL) fprintf(stdout, "%s-", printname); - if (number) fprintf(stdout, "%d-", lastmatchnumber++); - pp = end_of_line(pp, endptr, &ellength); - fwrite(lastmatchrestart, 1, pp - lastmatchrestart, stdout); - lastmatchrestart = pp; - } - hyphenpending = TRUE; - } -} - - - -/************************************************* -* Grep an individual file * -*************************************************/ - -/* This is called from grep_or_recurse() below. It uses a buffer that is three -times the value of MBUFTHIRD. The matching point is never allowed to stray into -the top third of the buffer, thus keeping more of the file available for -context printing or for multiline scanning. For large files, the pointer will -be in the middle third most of the time, so the bottom third is available for -"before" context printing. - -Arguments: - handle the fopened FILE stream for a normal file - the gzFile pointer when reading is via libz - the BZFILE pointer when reading is via libbz2 - frtype FR_PLAIN, FR_LIBZ, or FR_LIBBZ2 - printname the file name if it is to be printed for each match - or NULL if the file name is not to be printed - it cannot be NULL if filenames[_nomatch]_only is set - -Returns: 0 if there was at least one match - 1 otherwise (no matches) - 2 if there is a read error on a .bz2 file -*/ - -static int -pcregrep(void *handle, int frtype, char *printname) -{ -int rc = 1; -int linenumber = 1; -int lastmatchnumber = 0; -int count = 0; -int filepos = 0; -int offsets[99]; -char *lastmatchrestart = NULL; -char buffer[3*MBUFTHIRD]; -char *ptr = buffer; -char *endptr; -size_t bufflength; -BOOL endhyphenpending = FALSE; -FILE *in = NULL; /* Ensure initialized */ - -#ifdef SUPPORT_LIBZ -gzFile ingz = NULL; -#endif - -#ifdef SUPPORT_LIBBZ2 -BZFILE *inbz2 = NULL; -#endif - - -/* Do the first read into the start of the buffer and set up the pointer to end -of what we have. In the case of libz, a non-zipped .gz file will be read as a -plain file. However, if a .bz2 file isn't actually bzipped, the first read will -fail. */ - -#ifdef SUPPORT_LIBZ -if (frtype == FR_LIBZ) - { - ingz = (gzFile)handle; - bufflength = gzread (ingz, buffer, 3*MBUFTHIRD); - } -else -#endif - -#ifdef SUPPORT_LIBBZ2 -if (frtype == FR_LIBBZ2) - { - inbz2 = (BZFILE *)handle; - bufflength = BZ2_bzread(inbz2, buffer, 3*MBUFTHIRD); - if ((int)bufflength < 0) return 2; /* Gotcha: bufflength is size_t; */ - } /* without the cast it is unsigned. */ -else -#endif - - { - in = (FILE *)handle; - bufflength = fread(buffer, 1, 3*MBUFTHIRD, in); - } - -endptr = buffer + bufflength; - -/* Loop while the current pointer is not at the end of the file. For large -files, endptr will be at the end of the buffer when we are in the middle of the -file, but ptr will never get there, because as soon as it gets over 2/3 of the -way, the buffer is shifted left and re-filled. */ - -while (ptr < endptr) - { - int i, endlinelength; - int mrc = 0; - BOOL match = FALSE; - char *matchptr = ptr; - char *t = ptr; - size_t length, linelength; - - /* At this point, ptr is at the start of a line. We need to find the length - of the subject string to pass to pcre_exec(). In multiline mode, it is the - length remainder of the data in the buffer. Otherwise, it is the length of - the next line. After matching, we always advance by the length of the next - line. In multiline mode the PCRE_FIRSTLINE option is used for compiling, so - that any match is constrained to be in the first line. */ - - t = end_of_line(t, endptr, &endlinelength); - linelength = t - ptr - endlinelength; - length = multiline? (size_t)(endptr - ptr) : linelength; - - /* Extra processing for Jeffrey Friedl's debugging. */ - -#ifdef JFRIEDL_DEBUG - if (jfriedl_XT || jfriedl_XR) - { - #include - #include - struct timeval start_time, end_time; - struct timezone dummy; - - if (jfriedl_XT) - { - unsigned long newlen = length * jfriedl_XT + strlen(jfriedl_prefix) + strlen(jfriedl_postfix); - const char *orig = ptr; - ptr = malloc(newlen + 1); - if (!ptr) { - printf("out of memory"); - exit(2); - } - endptr = ptr; - strcpy(endptr, jfriedl_prefix); endptr += strlen(jfriedl_prefix); - for (i = 0; i < jfriedl_XT; i++) { - strncpy(endptr, orig, length); - endptr += length; - } - strcpy(endptr, jfriedl_postfix); endptr += strlen(jfriedl_postfix); - length = newlen; - } - - if (gettimeofday(&start_time, &dummy) != 0) - perror("bad gettimeofday"); - - - for (i = 0; i < jfriedl_XR; i++) - match = (pcre_exec(pattern_list[0], hints_list[0], ptr, length, 0, 0, offsets, 99) >= 0); - - if (gettimeofday(&end_time, &dummy) != 0) - perror("bad gettimeofday"); - - double delta = ((end_time.tv_sec + (end_time.tv_usec / 1000000.0)) - - - (start_time.tv_sec + (start_time.tv_usec / 1000000.0))); - - printf("%s TIMER[%.4f]\n", match ? "MATCH" : "FAIL", delta); - return 0; - } -#endif - - /* We come back here after a match when the -o option (only_matching) is set, - in order to find any further matches in the same line. */ - - ONLY_MATCHING_RESTART: - - /* Run through all the patterns until one matches. Note that we don't include - the final newline in the subject string. */ - - for (i = 0; i < pattern_count; i++) - { - mrc = pcre_exec(pattern_list[i], hints_list[i], matchptr, length, 0, 0, - offsets, 99); - if (mrc >= 0) { match = TRUE; break; } - if (mrc != PCRE_ERROR_NOMATCH) - { - fprintf(stderr, "pcregrep: pcre_exec() error %d while matching ", mrc); - if (pattern_count > 1) fprintf(stderr, "pattern number %d to ", i+1); - fprintf(stderr, "this line:\n"); - fwrite(matchptr, 1, linelength, stderr); /* In case binary zero included */ - fprintf(stderr, "\n"); - if (error_count == 0 && - (mrc == PCRE_ERROR_MATCHLIMIT || mrc == PCRE_ERROR_RECURSIONLIMIT)) - { - fprintf(stderr, "pcregrep: error %d means that a resource limit " - "was exceeded\n", mrc); - fprintf(stderr, "pcregrep: check your regex for nested unlimited loops\n"); - } - if (error_count++ > 20) - { - fprintf(stderr, "pcregrep: too many errors - abandoned\n"); - exit(2); - } - match = invert; /* No more matching; don't show the line again */ - break; - } - } - - /* If it's a match or a not-match (as required), do what's wanted. */ - - if (match != invert) - { - BOOL hyphenprinted = FALSE; - - /* We've failed if we want a file that doesn't have any matches. */ - - if (filenames == FN_NOMATCH_ONLY) return 1; - - /* Just count if just counting is wanted. */ - - if (count_only) count++; - - /* If all we want is a file name, there is no need to scan any more lines - in the file. */ - - else if (filenames == FN_ONLY) - { - fprintf(stdout, "%s\n", printname); - return 0; - } - - /* Likewise, if all we want is a yes/no answer. */ - - else if (quiet) return 0; - - /* The --only-matching option prints just the substring that matched, and - the --file-offsets and --line-offsets options output offsets for the - matching substring (they both force --only-matching). None of these options - prints any context. Afterwards, adjust the start and length, and then jump - back to look for further matches in the same line. If we are in invert - mode, however, nothing is printed - this could be still useful because the - return code is set. */ - - else if (only_matching) - { - if (!invert) - { - if (printname != NULL) fprintf(stdout, "%s:", printname); - if (number) fprintf(stdout, "%d:", linenumber); - if (line_offsets) - fprintf(stdout, "%d,%d", matchptr + offsets[0] - ptr, - offsets[1] - offsets[0]); - else if (file_offsets) - fprintf(stdout, "%d,%d", filepos + matchptr + offsets[0] - ptr, - offsets[1] - offsets[0]); - else - fwrite(matchptr + offsets[0], 1, offsets[1] - offsets[0], stdout); - fprintf(stdout, "\n"); - matchptr += offsets[1]; - length -= offsets[1]; - match = FALSE; - goto ONLY_MATCHING_RESTART; - } - } - - /* This is the default case when none of the above options is set. We print - the matching lines(s), possibly preceded and/or followed by other lines of - context. */ - - else - { - /* See if there is a requirement to print some "after" lines from a - previous match. We never print any overlaps. */ - - if (after_context > 0 && lastmatchnumber > 0) - { - int ellength; - int linecount = 0; - char *p = lastmatchrestart; - - while (p < ptr && linecount < after_context) - { - p = end_of_line(p, ptr, &ellength); - linecount++; - } - - /* It is important to advance lastmatchrestart during this printing so - that it interacts correctly with any "before" printing below. Print - each line's data using fwrite() in case there are binary zeroes. */ - - while (lastmatchrestart < p) - { - char *pp = lastmatchrestart; - if (printname != NULL) fprintf(stdout, "%s-", printname); - if (number) fprintf(stdout, "%d-", lastmatchnumber++); - pp = end_of_line(pp, endptr, &ellength); - fwrite(lastmatchrestart, 1, pp - lastmatchrestart, stdout); - lastmatchrestart = pp; - } - if (lastmatchrestart != ptr) hyphenpending = TRUE; - } - - /* If there were non-contiguous lines printed above, insert hyphens. */ - - if (hyphenpending) - { - fprintf(stdout, "--\n"); - hyphenpending = FALSE; - hyphenprinted = TRUE; - } - - /* See if there is a requirement to print some "before" lines for this - match. Again, don't print overlaps. */ - - if (before_context > 0) - { - int linecount = 0; - char *p = ptr; - - while (p > buffer && (lastmatchnumber == 0 || p > lastmatchrestart) && - linecount < before_context) - { - linecount++; - p = previous_line(p, buffer); - } - - if (lastmatchnumber > 0 && p > lastmatchrestart && !hyphenprinted) - fprintf(stdout, "--\n"); - - while (p < ptr) - { - int ellength; - char *pp = p; - if (printname != NULL) fprintf(stdout, "%s-", printname); - if (number) fprintf(stdout, "%d-", linenumber - linecount--); - pp = end_of_line(pp, endptr, &ellength); - fwrite(p, 1, pp - p, stdout); - p = pp; - } - } - - /* Now print the matching line(s); ensure we set hyphenpending at the end - of the file if any context lines are being output. */ - - if (after_context > 0 || before_context > 0) - endhyphenpending = TRUE; - - if (printname != NULL) fprintf(stdout, "%s:", printname); - if (number) fprintf(stdout, "%d:", linenumber); - - /* In multiline mode, we want to print to the end of the line in which - the end of the matched string is found, so we adjust linelength and the - line number appropriately, but only when there actually was a match - (invert not set). Because the PCRE_FIRSTLINE option is set, the start of - the match will always be before the first newline sequence. */ - - if (multiline) - { - int ellength; - char *endmatch = ptr; - if (!invert) - { - endmatch += offsets[1]; - t = ptr; - while (t < endmatch) - { - t = end_of_line(t, endptr, &ellength); - if (t <= endmatch) linenumber++; else break; - } - } - endmatch = end_of_line(endmatch, endptr, &ellength); - linelength = endmatch - ptr - ellength; - } - - /*** NOTE: Use only fwrite() to output the data line, so that binary - zeroes are treated as just another data character. */ - - /* This extra option, for Jeffrey Friedl's debugging requirements, - replaces the matched string, or a specific captured string if it exists, - with X. When this happens, colouring is ignored. */ - -#ifdef JFRIEDL_DEBUG - if (S_arg >= 0 && S_arg < mrc) - { - int first = S_arg * 2; - int last = first + 1; - fwrite(ptr, 1, offsets[first], stdout); - fprintf(stdout, "X"); - fwrite(ptr + offsets[last], 1, linelength - offsets[last], stdout); - } - else -#endif - - /* We have to split the line(s) up if colouring. */ - - if (do_colour) - { - fwrite(ptr, 1, offsets[0], stdout); - fprintf(stdout, "%c[%sm", 0x1b, colour_string); - fwrite(ptr + offsets[0], 1, offsets[1] - offsets[0], stdout); - fprintf(stdout, "%c[00m", 0x1b); - fwrite(ptr + offsets[1], 1, (linelength + endlinelength) - offsets[1], - stdout); - } - else fwrite(ptr, 1, linelength + endlinelength, stdout); - } - - /* End of doing what has to be done for a match */ - - rc = 0; /* Had some success */ - - /* Remember where the last match happened for after_context. We remember - where we are about to restart, and that line's number. */ - - lastmatchrestart = ptr + linelength + endlinelength; - lastmatchnumber = linenumber + 1; - } - - /* For a match in multiline inverted mode (which of course did not cause - anything to be printed), we have to move on to the end of the match before - proceeding. */ - - if (multiline && invert && match) - { - int ellength; - char *endmatch = ptr + offsets[1]; - t = ptr; - while (t < endmatch) - { - t = end_of_line(t, endptr, &ellength); - if (t <= endmatch) linenumber++; else break; - } - endmatch = end_of_line(endmatch, endptr, &ellength); - linelength = endmatch - ptr - ellength; - } - - /* Advance to after the newline and increment the line number. The file - offset to the current line is maintained in filepos. */ - - ptr += linelength + endlinelength; - filepos += linelength + endlinelength; - linenumber++; - - /* If we haven't yet reached the end of the file (the buffer is full), and - the current point is in the top 1/3 of the buffer, slide the buffer down by - 1/3 and refill it. Before we do this, if some unprinted "after" lines are - about to be lost, print them. */ - - if (bufflength >= sizeof(buffer) && ptr > buffer + 2*MBUFTHIRD) - { - if (after_context > 0 && - lastmatchnumber > 0 && - lastmatchrestart < buffer + MBUFTHIRD) - { - do_after_lines(lastmatchnumber, lastmatchrestart, endptr, printname); - lastmatchnumber = 0; - } - - /* Now do the shuffle */ - - memmove(buffer, buffer + MBUFTHIRD, 2*MBUFTHIRD); - ptr -= MBUFTHIRD; - -#ifdef SUPPORT_LIBZ - if (frtype == FR_LIBZ) - bufflength = 2*MBUFTHIRD + - gzread (ingz, buffer + 2*MBUFTHIRD, MBUFTHIRD); - else -#endif - -#ifdef SUPPORT_LIBBZ2 - if (frtype == FR_LIBBZ2) - bufflength = 2*MBUFTHIRD + - BZ2_bzread(inbz2, buffer + 2*MBUFTHIRD, MBUFTHIRD); - else -#endif - - bufflength = 2*MBUFTHIRD + fread(buffer + 2*MBUFTHIRD, 1, MBUFTHIRD, in); - - endptr = buffer + bufflength; - - /* Adjust any last match point */ - - if (lastmatchnumber > 0) lastmatchrestart -= MBUFTHIRD; - } - } /* Loop through the whole file */ - -/* End of file; print final "after" lines if wanted; do_after_lines sets -hyphenpending if it prints something. */ - -if (!only_matching && !count_only) - { - do_after_lines(lastmatchnumber, lastmatchrestart, endptr, printname); - hyphenpending |= endhyphenpending; - } - -/* Print the file name if we are looking for those without matches and there -were none. If we found a match, we won't have got this far. */ - -if (filenames == FN_NOMATCH_ONLY) - { - fprintf(stdout, "%s\n", printname); - return 0; - } - -/* Print the match count if wanted */ - -if (count_only) - { - if (printname != NULL) fprintf(stdout, "%s:", printname); - fprintf(stdout, "%d\n", count); - } - -return rc; -} - - - -/************************************************* -* Grep a file or recurse into a directory * -*************************************************/ - -/* Given a path name, if it's a directory, scan all the files if we are -recursing; if it's a file, grep it. - -Arguments: - pathname the path to investigate - dir_recurse TRUE if recursing is wanted (-r or -drecurse) - only_one_at_top TRUE if the path is the only one at toplevel - -Returns: 0 if there was at least one match - 1 if there were no matches - 2 there was some kind of error - -However, file opening failures are suppressed if "silent" is set. -*/ - -static int -grep_or_recurse(char *pathname, BOOL dir_recurse, BOOL only_one_at_top) -{ -int rc = 1; -int sep; -int frtype; -int pathlen; -void *handle; -FILE *in = NULL; /* Ensure initialized */ - -#ifdef SUPPORT_LIBZ -gzFile ingz = NULL; -#endif - -#ifdef SUPPORT_LIBBZ2 -BZFILE *inbz2 = NULL; -#endif - -/* If the file name is "-" we scan stdin */ - -if (strcmp(pathname, "-") == 0) - { - return pcregrep(stdin, FR_PLAIN, - (filenames > FN_DEFAULT || (filenames == FN_DEFAULT && !only_one_at_top))? - stdin_name : NULL); - } - -/* If the file is a directory, skip if skipping or if we are recursing, scan -each file and directory within it, subject to any include or exclude patterns -that were set. The scanning code is localized so it can be made -system-specific. */ - -if ((sep = isdirectory(pathname)) != 0) - { - if (dee_action == dee_SKIP) return 1; - if (dee_action == dee_RECURSE) - { - char buffer[1024]; - char *nextfile; - directory_type *dir = opendirectory(pathname); - - if (dir == NULL) - { - if (!silent) - fprintf(stderr, "pcregrep: Failed to open directory %s: %s\n", pathname, - strerror(errno)); - return 2; - } - - while ((nextfile = readdirectory(dir)) != NULL) - { - int frc, nflen; - sprintf(buffer, "%.512s%c%.128s", pathname, sep, nextfile); - nflen = strlen(nextfile); - - if (isdirectory(buffer)) - { - if (exclude_dir_compiled != NULL && - pcre_exec(exclude_dir_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) >= 0) - continue; - - if (include_dir_compiled != NULL && - pcre_exec(include_dir_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) < 0) - continue; - } - else - { - if (exclude_compiled != NULL && - pcre_exec(exclude_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) >= 0) - continue; - - if (include_compiled != NULL && - pcre_exec(include_compiled, NULL, nextfile, nflen, 0, 0, NULL, 0) < 0) - continue; - } - - frc = grep_or_recurse(buffer, dir_recurse, FALSE); - if (frc > 1) rc = frc; - else if (frc == 0 && rc == 1) rc = 0; - } - - closedirectory(dir); - return rc; - } - } - -/* If the file is not a directory and not a regular file, skip it if that's -been requested. */ - -else if (!isregfile(pathname) && DEE_action == DEE_SKIP) return 1; - -/* Control reaches here if we have a regular file, or if we have a directory -and recursion or skipping was not requested, or if we have anything else and -skipping was not requested. The scan proceeds. If this is the first and only -argument at top level, we don't show the file name, unless we are only showing -the file name, or the filename was forced (-H). */ - -pathlen = strlen(pathname); - -/* Open using zlib if it is supported and the file name ends with .gz. */ - -#ifdef SUPPORT_LIBZ -if (pathlen > 3 && strcmp(pathname + pathlen - 3, ".gz") == 0) - { - ingz = gzopen(pathname, "rb"); - if (ingz == NULL) - { - if (!silent) - fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname, - strerror(errno)); - return 2; - } - handle = (void *)ingz; - frtype = FR_LIBZ; - } -else -#endif - -/* Otherwise open with bz2lib if it is supported and the name ends with .bz2. */ - -#ifdef SUPPORT_LIBBZ2 -if (pathlen > 4 && strcmp(pathname + pathlen - 4, ".bz2") == 0) - { - inbz2 = BZ2_bzopen(pathname, "rb"); - handle = (void *)inbz2; - frtype = FR_LIBBZ2; - } -else -#endif - -/* Otherwise use plain fopen(). The label is so that we can come back here if -an attempt to read a .bz2 file indicates that it really is a plain file. */ - -#ifdef SUPPORT_LIBBZ2 -PLAIN_FILE: -#endif - { - in = fopen(pathname, "r"); - handle = (void *)in; - frtype = FR_PLAIN; - } - -/* All the opening methods return errno when they fail. */ - -if (handle == NULL) - { - if (!silent) - fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pathname, - strerror(errno)); - return 2; - } - -/* Now grep the file */ - -rc = pcregrep(handle, frtype, (filenames > FN_DEFAULT || - (filenames == FN_DEFAULT && !only_one_at_top))? pathname : NULL); - -/* Close in an appropriate manner. */ - -#ifdef SUPPORT_LIBZ -if (frtype == FR_LIBZ) - gzclose(ingz); -else -#endif - -/* If it is a .bz2 file and the result is 2, it means that the first attempt to -read failed. If the error indicates that the file isn't in fact bzipped, try -again as a normal file. */ - -#ifdef SUPPORT_LIBBZ2 -if (frtype == FR_LIBBZ2) - { - if (rc == 2) - { - int errnum; - const char *err = BZ2_bzerror(inbz2, &errnum); - if (errnum == BZ_DATA_ERROR_MAGIC) - { - BZ2_bzclose(inbz2); - goto PLAIN_FILE; - } - else if (!silent) - fprintf(stderr, "pcregrep: Failed to read %s using bzlib: %s\n", - pathname, err); - } - BZ2_bzclose(inbz2); - } -else -#endif - -/* Normal file close */ - -fclose(in); - -/* Pass back the yield from pcregrep(). */ - -return rc; -} - - - - -/************************************************* -* Usage function * -*************************************************/ - -static int -usage(int rc) -{ -option_item *op; -fprintf(stderr, "Usage: pcregrep [-"); -for (op = optionlist; op->one_char != 0; op++) - { - if (op->one_char > 0) fprintf(stderr, "%c", op->one_char); - } -fprintf(stderr, "] [long options] [pattern] [files]\n"); -fprintf(stderr, "Type `pcregrep --help' for more information and the long " - "options.\n"); -return rc; -} - - - - -/************************************************* -* Help function * -*************************************************/ - -static void -help(void) -{ -option_item *op; - -printf("Usage: pcregrep [OPTION]... [PATTERN] [FILE1 FILE2 ...]\n"); -printf("Search for PATTERN in each FILE or standard input.\n"); -printf("PATTERN must be present if neither -e nor -f is used.\n"); -printf("\"-\" can be used as a file name to mean STDIN.\n"); - -#ifdef SUPPORT_LIBZ -printf("Files whose names end in .gz are read using zlib.\n"); -#endif - -#ifdef SUPPORT_LIBBZ2 -printf("Files whose names end in .bz2 are read using bzlib2.\n"); -#endif - -#if defined SUPPORT_LIBZ || defined SUPPORT_LIBBZ2 -printf("Other files and the standard input are read as plain files.\n\n"); -#else -printf("All files are read as plain files, without any interpretation.\n\n"); -#endif - -printf("Example: pcregrep -i 'hello.*world' menu.h main.c\n\n"); -printf("Options:\n"); - -for (op = optionlist; op->one_char != 0; op++) - { - int n; - char s[4]; - if (op->one_char > 0) sprintf(s, "-%c,", op->one_char); else strcpy(s, " "); - n = 30 - printf(" %s --%s", s, op->long_name); - if (n < 1) n = 1; - printf("%.*s%s\n", n, " ", op->help_text); - } - -printf("\nWhen reading patterns from a file instead of using a command line option,\n"); -printf("trailing white space is removed and blank lines are ignored.\n"); -printf("There is a maximum of %d patterns.\n", MAX_PATTERN_COUNT); - -printf("\nWith no FILEs, read standard input. If fewer than two FILEs given, assume -h.\n"); -printf("Exit status is 0 if any matches, 1 if no matches, and 2 if trouble.\n"); -} - - - - -/************************************************* -* Handle a single-letter, no data option * -*************************************************/ - -static int -handle_option(int letter, int options) -{ -switch(letter) - { - case N_FOFFSETS: file_offsets = TRUE; break; - case N_HELP: help(); exit(0); - case N_LOFFSETS: line_offsets = number = TRUE; break; - case 'c': count_only = TRUE; break; - case 'F': process_options |= PO_FIXED_STRINGS; break; - case 'H': filenames = FN_FORCE; break; - case 'h': filenames = FN_NONE; break; - case 'i': options |= PCRE_CASELESS; break; - case 'l': filenames = FN_ONLY; break; - case 'L': filenames = FN_NOMATCH_ONLY; break; - case 'M': multiline = TRUE; options |= PCRE_MULTILINE|PCRE_FIRSTLINE; break; - case 'n': number = TRUE; break; - case 'o': only_matching = TRUE; break; - case 'q': quiet = TRUE; break; - case 'r': dee_action = dee_RECURSE; break; - case 's': silent = TRUE; break; - case 'u': options |= PCRE_UTF8; utf8 = TRUE; break; - case 'v': invert = TRUE; break; - case 'w': process_options |= PO_WORD_MATCH; break; - case 'x': process_options |= PO_LINE_MATCH; break; - - case 'V': - fprintf(stderr, "pcregrep version %s\n", pcre_version()); - exit(0); - break; - - default: - fprintf(stderr, "pcregrep: Unknown option -%c\n", letter); - exit(usage(2)); - } - -return options; -} - - - - -/************************************************* -* Construct printed ordinal * -*************************************************/ - -/* This turns a number into "1st", "3rd", etc. */ - -static char * -ordin(int n) -{ -static char buffer[8]; -char *p = buffer; -sprintf(p, "%d", n); -while (*p != 0) p++; -switch (n%10) - { - case 1: strcpy(p, "st"); break; - case 2: strcpy(p, "nd"); break; - case 3: strcpy(p, "rd"); break; - default: strcpy(p, "th"); break; - } -return buffer; -} - - - -/************************************************* -* Compile a single pattern * -*************************************************/ - -/* When the -F option has been used, this is called for each substring. -Otherwise it's called for each supplied pattern. - -Arguments: - pattern the pattern string - options the PCRE options - filename the file name, or NULL for a command-line pattern - count 0 if this is the only command line pattern, or - number of the command line pattern, or - linenumber for a pattern from a file - -Returns: TRUE on success, FALSE after an error -*/ - -static BOOL -compile_single_pattern(char *pattern, int options, char *filename, int count) -{ -char buffer[MBUFTHIRD + 16]; -const char *error; -int errptr; - -if (pattern_count >= MAX_PATTERN_COUNT) - { - fprintf(stderr, "pcregrep: Too many %spatterns (max %d)\n", - (filename == NULL)? "command-line " : "", MAX_PATTERN_COUNT); - return FALSE; - } - -sprintf(buffer, "%s%.*s%s", prefix[process_options], MBUFTHIRD, pattern, - suffix[process_options]); -pattern_list[pattern_count] = - pcre_compile(buffer, options, &error, &errptr, pcretables); -if (pattern_list[pattern_count] != NULL) - { - pattern_count++; - return TRUE; - } - -/* Handle compile errors */ - -errptr -= (int)strlen(prefix[process_options]); -if (errptr > (int)strlen(pattern)) errptr = (int)strlen(pattern); - -if (filename == NULL) - { - if (count == 0) - fprintf(stderr, "pcregrep: Error in command-line regex " - "at offset %d: %s\n", errptr, error); - else - fprintf(stderr, "pcregrep: Error in %s command-line regex " - "at offset %d: %s\n", ordin(count), errptr, error); - } -else - { - fprintf(stderr, "pcregrep: Error in regex in line %d of %s " - "at offset %d: %s\n", count, filename, errptr, error); - } - -return FALSE; -} - - - -/************************************************* -* Compile one supplied pattern * -*************************************************/ - -/* When the -F option has been used, each string may be a list of strings, -separated by line breaks. They will be matched literally. - -Arguments: - pattern the pattern string - options the PCRE options - filename the file name, or NULL for a command-line pattern - count 0 if this is the only command line pattern, or - number of the command line pattern, or - linenumber for a pattern from a file - -Returns: TRUE on success, FALSE after an error -*/ - -static BOOL -compile_pattern(char *pattern, int options, char *filename, int count) -{ -if ((process_options & PO_FIXED_STRINGS) != 0) - { - char *eop = pattern + strlen(pattern); - char buffer[MBUFTHIRD]; - for(;;) - { - int ellength; - char *p = end_of_line(pattern, eop, &ellength); - if (ellength == 0) - return compile_single_pattern(pattern, options, filename, count); - sprintf(buffer, "%.*s", (int)(p - pattern - ellength), pattern); - pattern = p; - if (!compile_single_pattern(buffer, options, filename, count)) - return FALSE; - } - } -else return compile_single_pattern(pattern, options, filename, count); -} - - - -/************************************************* -* Main program * -*************************************************/ - -/* Returns 0 if something matched, 1 if nothing matched, 2 after an error. */ - -int -main(int argc, char **argv) -{ -int i, j; -int rc = 1; -int pcre_options = 0; -int cmd_pattern_count = 0; -int hint_count = 0; -int errptr; -BOOL only_one_at_top; -char *patterns[MAX_PATTERN_COUNT]; -const char *locale_from = "--locale"; -const char *error; - -/* Set the default line ending value from the default in the PCRE library; -"lf", "cr", "crlf", and "any" are supported. Anything else is treated as "lf". -*/ - -(void)pcre_config(PCRE_CONFIG_NEWLINE, &i); -switch(i) - { - default: newline = (char *)"lf"; break; - case '\r': newline = (char *)"cr"; break; - case ('\r' << 8) | '\n': newline = (char *)"crlf"; break; - case -1: newline = (char *)"any"; break; - case -2: newline = (char *)"anycrlf"; break; - } - -/* Process the options */ - -for (i = 1; i < argc; i++) - { - option_item *op = NULL; - char *option_data = (char *)""; /* default to keep compiler happy */ - BOOL longop; - BOOL longopwasequals = FALSE; - - if (argv[i][0] != '-') break; - - /* If we hit an argument that is just "-", it may be a reference to STDIN, - but only if we have previously had -e or -f to define the patterns. */ - - if (argv[i][1] == 0) - { - if (pattern_filename != NULL || pattern_count > 0) break; - else exit(usage(2)); - } - - /* Handle a long name option, or -- to terminate the options */ - - if (argv[i][1] == '-') - { - char *arg = argv[i] + 2; - char *argequals = strchr(arg, '='); - - if (*arg == 0) /* -- terminates options */ - { - i++; - break; /* out of the options-handling loop */ - } - - longop = TRUE; - - /* Some long options have data that follows after =, for example file=name. - Some options have variations in the long name spelling: specifically, we - allow "regexp" because GNU grep allows it, though I personally go along - with Jeffrey Friedl and Larry Wall in preferring "regex" without the "p". - These options are entered in the table as "regex(p)". No option is in both - these categories, fortunately. */ - - for (op = optionlist; op->one_char != 0; op++) - { - char *opbra = strchr(op->long_name, '('); - char *equals = strchr(op->long_name, '='); - if (opbra == NULL) /* Not a (p) case */ - { - if (equals == NULL) /* Not thing=data case */ - { - if (strcmp(arg, op->long_name) == 0) break; - } - else /* Special case xxx=data */ - { - int oplen = equals - op->long_name; - int arglen = (argequals == NULL)? (int)strlen(arg) : argequals - arg; - if (oplen == arglen && strncmp(arg, op->long_name, oplen) == 0) - { - option_data = arg + arglen; - if (*option_data == '=') - { - option_data++; - longopwasequals = TRUE; - } - break; - } - } - } - else /* Special case xxxx(p) */ - { - char buff1[24]; - char buff2[24]; - int baselen = opbra - op->long_name; - sprintf(buff1, "%.*s", baselen, op->long_name); - sprintf(buff2, "%s%.*s", buff1, - (int)strlen(op->long_name) - baselen - 2, opbra + 1); - if (strcmp(arg, buff1) == 0 || strcmp(arg, buff2) == 0) - break; - } - } - - if (op->one_char == 0) - { - fprintf(stderr, "pcregrep: Unknown option %s\n", argv[i]); - exit(usage(2)); - } - } - - - /* Jeffrey Friedl's debugging harness uses these additional options which - are not in the right form for putting in the option table because they use - only one hyphen, yet are more than one character long. By putting them - separately here, they will not get displayed as part of the help() output, - but I don't think Jeffrey will care about that. */ - -#ifdef JFRIEDL_DEBUG - else if (strcmp(argv[i], "-pre") == 0) { - jfriedl_prefix = argv[++i]; - continue; - } else if (strcmp(argv[i], "-post") == 0) { - jfriedl_postfix = argv[++i]; - continue; - } else if (strcmp(argv[i], "-XT") == 0) { - sscanf(argv[++i], "%d", &jfriedl_XT); - continue; - } else if (strcmp(argv[i], "-XR") == 0) { - sscanf(argv[++i], "%d", &jfriedl_XR); - continue; - } -#endif - - - /* One-char options; many that have no data may be in a single argument; we - continue till we hit the last one or one that needs data. */ - - else - { - char *s = argv[i] + 1; - longop = FALSE; - while (*s != 0) - { - for (op = optionlist; op->one_char != 0; op++) - { if (*s == op->one_char) break; } - if (op->one_char == 0) - { - fprintf(stderr, "pcregrep: Unknown option letter '%c' in \"%s\"\n", - *s, argv[i]); - exit(usage(2)); - } - if (op->type != OP_NODATA || s[1] == 0) - { - option_data = s+1; - break; - } - pcre_options = handle_option(*s++, pcre_options); - } - } - - /* At this point we should have op pointing to a matched option. If the type - is NO_DATA, it means that there is no data, and the option might set - something in the PCRE options. */ - - if (op->type == OP_NODATA) - { - pcre_options = handle_option(op->one_char, pcre_options); - continue; - } - - /* If the option type is OP_OP_STRING or OP_OP_NUMBER, it's an option that - either has a value or defaults to something. It cannot have data in a - separate item. At the moment, the only such options are "colo(u)r" and - Jeffrey Friedl's special -S debugging option. */ - - if (*option_data == 0 && - (op->type == OP_OP_STRING || op->type == OP_OP_NUMBER)) - { - switch (op->one_char) - { - case N_COLOUR: - colour_option = (char *)"auto"; - break; -#ifdef JFRIEDL_DEBUG - case 'S': - S_arg = 0; - break; -#endif - } - continue; - } - - /* Otherwise, find the data string for the option. */ - - if (*option_data == 0) - { - if (i >= argc - 1 || longopwasequals) - { - fprintf(stderr, "pcregrep: Data missing after %s\n", argv[i]); - exit(usage(2)); - } - option_data = argv[++i]; - } - - /* If the option type is OP_PATLIST, it's the -e option, which can be called - multiple times to create a list of patterns. */ - - if (op->type == OP_PATLIST) - { - if (cmd_pattern_count >= MAX_PATTERN_COUNT) - { - fprintf(stderr, "pcregrep: Too many command-line patterns (max %d)\n", - MAX_PATTERN_COUNT); - return 2; - } - patterns[cmd_pattern_count++] = option_data; - } - - /* Otherwise, deal with single string or numeric data values. */ - - else if (op->type != OP_NUMBER && op->type != OP_OP_NUMBER) - { - *((char **)op->dataptr) = option_data; - } - else - { - char *endptr; - int n = strtoul(option_data, &endptr, 10); - if (*endptr != 0) - { - if (longop) - { - char *equals = strchr(op->long_name, '='); - int nlen = (equals == NULL)? (int)strlen(op->long_name) : - equals - op->long_name; - fprintf(stderr, "pcregrep: Malformed number \"%s\" after --%.*s\n", - option_data, nlen, op->long_name); - } - else - fprintf(stderr, "pcregrep: Malformed number \"%s\" after -%c\n", - option_data, op->one_char); - exit(usage(2)); - } - *((int *)op->dataptr) = n; - } - } - -/* Options have been decoded. If -C was used, its value is used as a default -for -A and -B. */ - -if (both_context > 0) - { - if (after_context == 0) after_context = both_context; - if (before_context == 0) before_context = both_context; - } - -/* Only one of --only-matching, --file-offsets, or --line-offsets is permitted. -However, the latter two set the only_matching flag. */ - -if ((only_matching && (file_offsets || line_offsets)) || - (file_offsets && line_offsets)) - { - fprintf(stderr, "pcregrep: Cannot mix --only-matching, --file-offsets " - "and/or --line-offsets\n"); - exit(usage(2)); - } - -if (file_offsets || line_offsets) only_matching = TRUE; - -/* If a locale has not been provided as an option, see if the LC_CTYPE or -LC_ALL environment variable is set, and if so, use it. */ - -if (locale == NULL) - { - locale = getenv("LC_ALL"); - locale_from = "LCC_ALL"; - } - -if (locale == NULL) - { - locale = getenv("LC_CTYPE"); - locale_from = "LC_CTYPE"; - } - -/* If a locale has been provided, set it, and generate the tables the PCRE -needs. Otherwise, pcretables==NULL, which causes the use of default tables. */ - -if (locale != NULL) - { - if (setlocale(LC_CTYPE, locale) == NULL) - { - fprintf(stderr, "pcregrep: Failed to set locale %s (obtained from %s)\n", - locale, locale_from); - return 2; - } - pcretables = pcre_maketables(); - } - -/* Sort out colouring */ - -if (colour_option != NULL && strcmp(colour_option, "never") != 0) - { - if (strcmp(colour_option, "always") == 0) do_colour = TRUE; - else if (strcmp(colour_option, "auto") == 0) do_colour = is_stdout_tty(); - else - { - fprintf(stderr, "pcregrep: Unknown colour setting \"%s\"\n", - colour_option); - return 2; - } - if (do_colour) - { - char *cs = getenv("PCREGREP_COLOUR"); - if (cs == NULL) cs = getenv("PCREGREP_COLOR"); - if (cs != NULL) colour_string = cs; - } - } - -/* Interpret the newline type; the default settings are Unix-like. */ - -if (strcmp(newline, "cr") == 0 || strcmp(newline, "CR") == 0) - { - pcre_options |= PCRE_NEWLINE_CR; - endlinetype = EL_CR; - } -else if (strcmp(newline, "lf") == 0 || strcmp(newline, "LF") == 0) - { - pcre_options |= PCRE_NEWLINE_LF; - endlinetype = EL_LF; - } -else if (strcmp(newline, "crlf") == 0 || strcmp(newline, "CRLF") == 0) - { - pcre_options |= PCRE_NEWLINE_CRLF; - endlinetype = EL_CRLF; - } -else if (strcmp(newline, "any") == 0 || strcmp(newline, "ANY") == 0) - { - pcre_options |= PCRE_NEWLINE_ANY; - endlinetype = EL_ANY; - } -else if (strcmp(newline, "anycrlf") == 0 || strcmp(newline, "ANYCRLF") == 0) - { - pcre_options |= PCRE_NEWLINE_ANYCRLF; - endlinetype = EL_ANYCRLF; - } -else - { - fprintf(stderr, "pcregrep: Invalid newline specifier \"%s\"\n", newline); - return 2; - } - -/* Interpret the text values for -d and -D */ - -if (dee_option != NULL) - { - if (strcmp(dee_option, "read") == 0) dee_action = dee_READ; - else if (strcmp(dee_option, "recurse") == 0) dee_action = dee_RECURSE; - else if (strcmp(dee_option, "skip") == 0) dee_action = dee_SKIP; - else - { - fprintf(stderr, "pcregrep: Invalid value \"%s\" for -d\n", dee_option); - return 2; - } - } - -if (DEE_option != NULL) - { - if (strcmp(DEE_option, "read") == 0) DEE_action = DEE_READ; - else if (strcmp(DEE_option, "skip") == 0) DEE_action = DEE_SKIP; - else - { - fprintf(stderr, "pcregrep: Invalid value \"%s\" for -D\n", DEE_option); - return 2; - } - } - -/* Check the values for Jeffrey Friedl's debugging options. */ - -#ifdef JFRIEDL_DEBUG -if (S_arg > 9) - { - fprintf(stderr, "pcregrep: bad value for -S option\n"); - return 2; - } -if (jfriedl_XT != 0 || jfriedl_XR != 0) - { - if (jfriedl_XT == 0) jfriedl_XT = 1; - if (jfriedl_XR == 0) jfriedl_XR = 1; - } -#endif - -/* Get memory to store the pattern and hints lists. */ - -pattern_list = (pcre **)malloc(MAX_PATTERN_COUNT * sizeof(pcre *)); -hints_list = (pcre_extra **)malloc(MAX_PATTERN_COUNT * sizeof(pcre_extra *)); - -if (pattern_list == NULL || hints_list == NULL) - { - fprintf(stderr, "pcregrep: malloc failed\n"); - goto EXIT2; - } - -/* If no patterns were provided by -e, and there is no file provided by -f, -the first argument is the one and only pattern, and it must exist. */ - -if (cmd_pattern_count == 0 && pattern_filename == NULL) - { - if (i >= argc) return usage(2); - patterns[cmd_pattern_count++] = argv[i++]; - } - -/* Compile the patterns that were provided on the command line, either by -multiple uses of -e or as a single unkeyed pattern. */ - -for (j = 0; j < cmd_pattern_count; j++) - { - if (!compile_pattern(patterns[j], pcre_options, NULL, - (j == 0 && cmd_pattern_count == 1)? 0 : j + 1)) - goto EXIT2; - } - -/* Compile the regular expressions that are provided in a file. */ - -if (pattern_filename != NULL) - { - int linenumber = 0; - FILE *f; - char *filename; - char buffer[MBUFTHIRD]; - - if (strcmp(pattern_filename, "-") == 0) - { - f = stdin; - filename = stdin_name; - } - else - { - f = fopen(pattern_filename, "r"); - if (f == NULL) - { - fprintf(stderr, "pcregrep: Failed to open %s: %s\n", pattern_filename, - strerror(errno)); - goto EXIT2; - } - filename = pattern_filename; - } - - while (fgets(buffer, MBUFTHIRD, f) != NULL) - { - char *s = buffer + (int)strlen(buffer); - while (s > buffer && isspace((unsigned char)(s[-1]))) s--; - *s = 0; - linenumber++; - if (buffer[0] == 0) continue; /* Skip blank lines */ - if (!compile_pattern(buffer, pcre_options, filename, linenumber)) - goto EXIT2; - } - - if (f != stdin) fclose(f); - } - -/* Study the regular expressions, as we will be running them many times */ - -for (j = 0; j < pattern_count; j++) - { - hints_list[j] = pcre_study(pattern_list[j], 0, &error); - if (error != NULL) - { - char s[16]; - if (pattern_count == 1) s[0] = 0; else sprintf(s, " number %d", j); - fprintf(stderr, "pcregrep: Error while studying regex%s: %s\n", s, error); - goto EXIT2; - } - hint_count++; - } - -/* If there are include or exclude patterns, compile them. */ - -if (exclude_pattern != NULL) - { - exclude_compiled = pcre_compile(exclude_pattern, 0, &error, &errptr, - pcretables); - if (exclude_compiled == NULL) - { - fprintf(stderr, "pcregrep: Error in 'exclude' regex at offset %d: %s\n", - errptr, error); - goto EXIT2; - } - } - -if (include_pattern != NULL) - { - include_compiled = pcre_compile(include_pattern, 0, &error, &errptr, - pcretables); - if (include_compiled == NULL) - { - fprintf(stderr, "pcregrep: Error in 'include' regex at offset %d: %s\n", - errptr, error); - goto EXIT2; - } - } - -if (exclude_dir_pattern != NULL) - { - exclude_dir_compiled = pcre_compile(exclude_dir_pattern, 0, &error, &errptr, - pcretables); - if (exclude_dir_compiled == NULL) - { - fprintf(stderr, "pcregrep: Error in 'exclude_dir' regex at offset %d: %s\n", - errptr, error); - goto EXIT2; - } - } - -if (include_dir_pattern != NULL) - { - include_dir_compiled = pcre_compile(include_dir_pattern, 0, &error, &errptr, - pcretables); - if (include_dir_compiled == NULL) - { - fprintf(stderr, "pcregrep: Error in 'include_dir' regex at offset %d: %s\n", - errptr, error); - goto EXIT2; - } - } - -/* If there are no further arguments, do the business on stdin and exit. */ - -if (i >= argc) - { - rc = pcregrep(stdin, FR_PLAIN, (filenames > FN_DEFAULT)? stdin_name : NULL); - goto EXIT; - } - -/* Otherwise, work through the remaining arguments as files or directories. -Pass in the fact that there is only one argument at top level - this suppresses -the file name if the argument is not a directory and filenames are not -otherwise forced. */ - -only_one_at_top = i == argc - 1; /* Catch initial value of i */ - -for (; i < argc; i++) - { - int frc = grep_or_recurse(argv[i], dee_action == dee_RECURSE, - only_one_at_top); - if (frc > 1) rc = frc; - else if (frc == 0 && rc == 1) rc = 0; - } - -EXIT: -if (pattern_list != NULL) - { - for (i = 0; i < pattern_count; i++) free(pattern_list[i]); - free(pattern_list); - } -if (hints_list != NULL) - { - for (i = 0; i < hint_count; i++) free(hints_list[i]); - free(hints_list); - } -return rc; - -EXIT2: -rc = 2; -goto EXIT; -} - -/* End of pcregrep */ diff --git a/pcre/pcreposix.c b/pcre/pcreposix.c deleted file mode 100644 --- a/pcre/pcreposix.c +++ /dev/null @@ -1,356 +0,0 @@ -/************************************************* -* Perl-Compatible Regular Expressions * -*************************************************/ - -/* PCRE is a library of functions to support regular expressions whose syntax -and semantics are as close as possible to those of the Perl 5 language. - - Written by Philip Hazel - Copyright (c) 1997-2008 University of Cambridge - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -/* This module is a wrapper that provides a POSIX API to the underlying PCRE -functions. */ - - -#ifdef HAVE_CONFIG_H -#include "config.h" -#endif - - -/* Ensure that the PCREPOSIX_EXP_xxx macros are set appropriately for -compiling these functions. This must come before including pcreposix.h, where -they are set for an application (using these functions) if they have not -previously been set. */ - -#if defined(_WIN32) && !defined(PCRE_STATIC) -# define PCREPOSIX_EXP_DECL extern __declspec(dllexport) -# define PCREPOSIX_EXP_DEFN __declspec(dllexport) -#endif - -#include "pcre.h" -#include "pcre_internal.h" -#include "pcreposix.h" - - -/* Table to translate PCRE compile time error codes into POSIX error codes. */ - -static const int eint[] = { - 0, /* no error */ - REG_EESCAPE, /* \ at end of pattern */ - REG_EESCAPE, /* \c at end of pattern */ - REG_EESCAPE, /* unrecognized character follows \ */ - REG_BADBR, /* numbers out of order in {} quantifier */ - REG_BADBR, /* number too big in {} quantifier */ - REG_EBRACK, /* missing terminating ] for character class */ - REG_ECTYPE, /* invalid escape sequence in character class */ - REG_ERANGE, /* range out of order in character class */ - REG_BADRPT, /* nothing to repeat */ - REG_BADRPT, /* operand of unlimited repeat could match the empty string */ - REG_ASSERT, /* internal error: unexpected repeat */ - REG_BADPAT, /* unrecognized character after (? */ - REG_BADPAT, /* POSIX named classes are supported only within a class */ - REG_EPAREN, /* missing ) */ - REG_ESUBREG, /* reference to non-existent subpattern */ - REG_INVARG, /* erroffset passed as NULL */ - REG_INVARG, /* unknown option bit(s) set */ - REG_EPAREN, /* missing ) after comment */ - REG_ESIZE, /* parentheses nested too deeply */ - REG_ESIZE, /* regular expression too large */ - REG_ESPACE, /* failed to get memory */ - REG_EPAREN, /* unmatched brackets */ - REG_ASSERT, /* internal error: code overflow */ - REG_BADPAT, /* unrecognized character after (?< */ - REG_BADPAT, /* lookbehind assertion is not fixed length */ - REG_BADPAT, /* malformed number or name after (?( */ - REG_BADPAT, /* conditional group contains more than two branches */ - REG_BADPAT, /* assertion expected after (?( */ - REG_BADPAT, /* (?R or (?[+-]digits must be followed by ) */ - REG_ECTYPE, /* unknown POSIX class name */ - REG_BADPAT, /* POSIX collating elements are not supported */ - REG_INVARG, /* this version of PCRE is not compiled with PCRE_UTF8 support */ - REG_BADPAT, /* spare error */ - REG_BADPAT, /* character value in \x{...} sequence is too large */ - REG_BADPAT, /* invalid condition (?(0) */ - REG_BADPAT, /* \C not allowed in lookbehind assertion */ - REG_EESCAPE, /* PCRE does not support \L, \l, \N, \U, or \u */ - REG_BADPAT, /* number after (?C is > 255 */ - REG_BADPAT, /* closing ) for (?C expected */ - REG_BADPAT, /* recursive call could loop indefinitely */ - REG_BADPAT, /* unrecognized character after (?P */ - REG_BADPAT, /* syntax error in subpattern name (missing terminator) */ - REG_BADPAT, /* two named subpatterns have the same name */ - REG_BADPAT, /* invalid UTF-8 string */ - REG_BADPAT, /* support for \P, \p, and \X has not been compiled */ - REG_BADPAT, /* malformed \P or \p sequence */ - REG_BADPAT, /* unknown property name after \P or \p */ - REG_BADPAT, /* subpattern name is too long (maximum 32 characters) */ - REG_BADPAT, /* too many named subpatterns (maximum 10,000) */ - REG_BADPAT, /* repeated subpattern is too long */ - REG_BADPAT, /* octal value is greater than \377 (not in UTF-8 mode) */ - REG_BADPAT, /* internal error: overran compiling workspace */ - REG_BADPAT, /* internal error: previously-checked referenced subpattern not found */ - REG_BADPAT, /* DEFINE group contains more than one branch */ - REG_BADPAT, /* repeating a DEFINE group is not allowed */ - REG_INVARG, /* inconsistent NEWLINE options */ - REG_BADPAT, /* \g is not followed followed by an (optionally braced) non-zero number */ - REG_BADPAT, /* (?+ or (?- must be followed by a non-zero number */ - REG_BADPAT, /* number is too big */ - REG_BADPAT, /* subpattern name expected */ - REG_BADPAT, /* digit expected after (?+ */ - REG_BADPAT /* ] is an invalid data character in JavaScript compatibility mode */ -}; - -/* Table of texts corresponding to POSIX error codes */ - -static const char *const pstring[] = { - "", /* Dummy for value 0 */ - "internal error", /* REG_ASSERT */ - "invalid repeat counts in {}", /* BADBR */ - "pattern error", /* BADPAT */ - "? * + invalid", /* BADRPT */ - "unbalanced {}", /* EBRACE */ - "unbalanced []", /* EBRACK */ - "collation error - not relevant", /* ECOLLATE */ - "bad class", /* ECTYPE */ - "bad escape sequence", /* EESCAPE */ - "empty expression", /* EMPTY */ - "unbalanced ()", /* EPAREN */ - "bad range inside []", /* ERANGE */ - "expression too big", /* ESIZE */ - "failed to get memory", /* ESPACE */ - "bad back reference", /* ESUBREG */ - "bad argument", /* INVARG */ - "match failed" /* NOMATCH */ -}; - - - - -/************************************************* -* Translate error code to string * -*************************************************/ - -PCREPOSIX_EXP_DEFN size_t -regerror(int errcode, const regex_t *preg, char *errbuf, size_t errbuf_size) -{ -const char *message, *addmessage; -size_t length, addlength; - -message = (errcode >= (int)(sizeof(pstring)/sizeof(char *)))? - "unknown error code" : pstring[errcode]; -length = strlen(message) + 1; - -addmessage = " at offset "; -addlength = (preg != NULL && (int)preg->re_erroffset != -1)? - strlen(addmessage) + 6 : 0; - -if (errbuf_size > 0) - { - if (addlength > 0 && errbuf_size >= length + addlength) - sprintf(errbuf, "%s%s%-6d", message, addmessage, (int)preg->re_erroffset); - else - { - strncpy(errbuf, message, errbuf_size - 1); - errbuf[errbuf_size-1] = 0; - } - } - -return length + addlength; -} - - - - -/************************************************* -* Free store held by a regex * -*************************************************/ - -PCREPOSIX_EXP_DEFN void -regfree(regex_t *preg) -{ -(pcre_free)(preg->re_pcre); -} - - - - -/************************************************* -* Compile a regular expression * -*************************************************/ - -/* -Arguments: - preg points to a structure for recording the compiled expression - pattern the pattern to compile - cflags compilation flags - -Returns: 0 on success - various non-zero codes on failure -*/ - -PCREPOSIX_EXP_DEFN int -regcomp(regex_t *preg, const char *pattern, int cflags) -{ -const char *errorptr; -int erroffset; -int errorcode; -int options = 0; - -if ((cflags & REG_ICASE) != 0) options |= PCRE_CASELESS; -if ((cflags & REG_NEWLINE) != 0) options |= PCRE_MULTILINE; -if ((cflags & REG_DOTALL) != 0) options |= PCRE_DOTALL; -if ((cflags & REG_NOSUB) != 0) options |= PCRE_NO_AUTO_CAPTURE; -if ((cflags & REG_UTF8) != 0) options |= PCRE_UTF8; - -preg->re_pcre = pcre_compile2(pattern, options, &errorcode, &errorptr, - &erroffset, NULL); -preg->re_erroffset = erroffset; - -if (preg->re_pcre == NULL) return eint[errorcode]; - -preg->re_nsub = pcre_info((const pcre *)preg->re_pcre, NULL, NULL); -return 0; -} - - - - -/************************************************* -* Match a regular expression * -*************************************************/ - -/* Unfortunately, PCRE requires 3 ints of working space for each captured -substring, so we have to get and release working store instead of just using -the POSIX structures as was done in earlier releases when PCRE needed only 2 -ints. However, if the number of possible capturing brackets is small, use a -block of store on the stack, to reduce the use of malloc/free. The threshold is -in a macro that can be changed at configure time. - -If REG_NOSUB was specified at compile time, the PCRE_NO_AUTO_CAPTURE flag will -be set. When this is the case, the nmatch and pmatch arguments are ignored, and -the only result is yes/no/error. */ - -PCREPOSIX_EXP_DEFN int -regexec(const regex_t *preg, const char *string, size_t nmatch, - regmatch_t pmatch[], int eflags) -{ -int rc, so, eo; -int options = 0; -int *ovector = NULL; -int small_ovector[POSIX_MALLOC_THRESHOLD * 3]; -BOOL allocated_ovector = FALSE; -BOOL nosub = - (((const pcre *)preg->re_pcre)->options & PCRE_NO_AUTO_CAPTURE) != 0; - -if ((eflags & REG_NOTBOL) != 0) options |= PCRE_NOTBOL; -if ((eflags & REG_NOTEOL) != 0) options |= PCRE_NOTEOL; - -((regex_t *)preg)->re_erroffset = (size_t)(-1); /* Only has meaning after compile */ - -/* When no string data is being returned, ensure that nmatch is zero. -Otherwise, ensure the vector for holding the return data is large enough. */ - -if (nosub) nmatch = 0; - -else if (nmatch > 0) - { - if (nmatch <= POSIX_MALLOC_THRESHOLD) - { - ovector = &(small_ovector[0]); - } - else - { - if (nmatch > INT_MAX/(sizeof(int) * 3)) return REG_ESPACE; - ovector = (int *)malloc(sizeof(int) * nmatch * 3); - if (ovector == NULL) return REG_ESPACE; - allocated_ovector = TRUE; - } - } - -/* REG_STARTEND is a BSD extension, to allow for non-NUL-terminated strings. -The man page from OS X says "REG_STARTEND affects only the location of the -string, not how it is matched". That is why the "so" value is used to bump the -start location rather than being passed as a PCRE "starting offset". */ - -if ((eflags & REG_STARTEND) != 0) - { - so = pmatch[0].rm_so; - eo = pmatch[0].rm_eo; - } -else - { - so = 0; - eo = strlen(string); - } - -rc = pcre_exec((const pcre *)preg->re_pcre, NULL, string + so, (eo - so), - 0, options, ovector, nmatch * 3); - -if (rc == 0) rc = nmatch; /* All captured slots were filled in */ - -if (rc >= 0) - { - size_t i; - if (!nosub) - { - for (i = 0; i < (size_t)rc; i++) - { - pmatch[i].rm_so = ovector[i*2]; - pmatch[i].rm_eo = ovector[i*2+1]; - } - if (allocated_ovector) free(ovector); - for (; i < nmatch; i++) pmatch[i].rm_so = pmatch[i].rm_eo = -1; - } - return 0; - } - -else - { - if (allocated_ovector) free(ovector); - switch(rc) - { - case PCRE_ERROR_NOMATCH: return REG_NOMATCH; - case PCRE_ERROR_NULL: return REG_INVARG; - case PCRE_ERROR_BADOPTION: return REG_INVARG; - case PCRE_ERROR_BADMAGIC: return REG_INVARG; - case PCRE_ERROR_UNKNOWN_NODE: return REG_ASSERT; - case PCRE_ERROR_NOMEMORY: return REG_ESPACE; - case PCRE_ERROR_MATCHLIMIT: return REG_ESPACE; - case PCRE_ERROR_BADUTF8: return REG_INVARG; - case PCRE_ERROR_BADUTF8_OFFSET: return REG_INVARG; - default: return REG_ASSERT; - } - } -} - -/* End of pcreposix.c */ diff --git a/pcre/pcreposix.h b/pcre/pcreposix.h --- a/pcre/pcreposix.h +++ b/pcre/pcreposix.h @@ -9,7 +9,7 @@ Compatible Regular Expression library. It defines the things POSIX says should be there. I hope. - Copyright (c) 1997-2008 University of Cambridge + Copyright (c) 1997-2012 University of Cambridge ----------------------------------------------------------------------------- Redistribution and use in source and binary forms, with or without @@ -50,16 +50,19 @@ extern "C" { #endif -/* Options, mostly defined by POSIX, but with a couple of extras. */ +/* Options, mostly defined by POSIX, but with some extras. */ -#define REG_ICASE 0x0001 -#define REG_NEWLINE 0x0002 -#define REG_NOTBOL 0x0004 -#define REG_NOTEOL 0x0008 -#define REG_DOTALL 0x0010 /* NOT defined by POSIX. */ -#define REG_NOSUB 0x0020 -#define REG_UTF8 0x0040 /* NOT defined by POSIX. */ +#define REG_ICASE 0x0001 /* Maps to PCRE_CASELESS */ +#define REG_NEWLINE 0x0002 /* Maps to PCRE_MULTILINE */ +#define REG_NOTBOL 0x0004 /* Maps to PCRE_NOTBOL */ +#define REG_NOTEOL 0x0008 /* Maps to PCRE_NOTEOL */ +#define REG_DOTALL 0x0010 /* NOT defined by POSIX; maps to PCRE_DOTALL */ +#define REG_NOSUB 0x0020 /* Maps to PCRE_NO_AUTO_CAPTURE */ +#define REG_UTF8 0x0040 /* NOT defined by POSIX; maps to PCRE_UTF8 */ #define REG_STARTEND 0x0080 /* BSD feature: pass subject string by so,eo */ +#define REG_NOTEMPTY 0x0100 /* NOT defined by POSIX; maps to PCRE_NOTEMPTY */ +#define REG_UNGREEDY 0x0200 /* NOT defined by POSIX; maps to PCRE_UNGREEDY */ +#define REG_UCP 0x0400 /* NOT defined by POSIX; maps to PCRE_UCP */ /* This is not used by PCRE, but by defining it we make it easier to slot PCRE into existing programs that make POSIX calls. */ diff --git a/pcre/pcretest.c b/pcre/pcretest.c deleted file mode 100644 --- a/pcre/pcretest.c +++ /dev/null @@ -1,2448 +0,0 @@ -/************************************************* -* PCRE testing program * -*************************************************/ - -/* This program was hacked up as a tester for PCRE. I really should have -written it more tidily in the first place. Will I ever learn? It has grown and -been extended and consequently is now rather, er, *very* untidy in places. - ------------------------------------------------------------------------------ -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - - * Redistributions of source code must retain the above copyright notice, - this list of conditions and the following disclaimer. - - * Redistributions in binary form must reproduce the above copyright - notice, this list of conditions and the following disclaimer in the - documentation and/or other materials provided with the distribution. - - * Neither the name of the University of Cambridge nor the names of its - contributors may be used to endorse or promote products derived from - this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE -ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE -LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR -CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF -SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS -INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN -CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) -ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -POSSIBILITY OF SUCH DAMAGE. ------------------------------------------------------------------------------ -*/ - - -#ifdef HAVE_CONFIG_H -#include "config.h" -#endif - -#include -#include -#include -#include -#include -#include -#include - -#ifdef SUPPORT_LIBREADLINE -#ifdef HAVE_UNISTD_H -#include -#endif -#include -#include -#endif - - -/* A number of things vary for Windows builds. Originally, pcretest opened its -input and output without "b"; then I was told that "b" was needed in some -environments, so it was added for release 5.0 to both the input and output. (It -makes no difference on Unix-like systems.) Later I was told that it is wrong -for the input on Windows. I've now abstracted the modes into two macros that -are set here, to make it easier to fiddle with them, and removed "b" from the -input mode under Windows. */ - -#if defined(_WIN32) || defined(WIN32) -#include /* For _setmode() */ -#include /* For _O_BINARY */ -#define INPUT_MODE "r" -#define OUTPUT_MODE "wb" - -#define isatty _isatty /* This is what Windows calls them, I'm told */ -#define fileno _fileno - -#else -#include /* These two includes are needed */ -#include /* for setrlimit(). */ -#define INPUT_MODE "rb" -#define OUTPUT_MODE "wb" -#endif - - -/* We have to include pcre_internal.h because we need the internal info for -displaying the results of pcre_study() and we also need to know about the -internal macros, structures, and other internal data values; pcretest has -"inside information" compared to a program that strictly follows the PCRE API. - -Although pcre_internal.h does itself include pcre.h, we explicitly include it -here before pcre_internal.h so that the PCRE_EXP_xxx macros get set -appropriately for an application, not for building PCRE. */ - -#include "pcre.h" -#include "pcre_internal.h" - -/* We need access to the data tables that PCRE uses. So as not to have to keep -two copies, we include the source file here, changing the names of the external -symbols to prevent clashes. */ - -#define _pcre_utf8_table1 utf8_table1 -#define _pcre_utf8_table1_size utf8_table1_size -#define _pcre_utf8_table2 utf8_table2 -#define _pcre_utf8_table3 utf8_table3 -#define _pcre_utf8_table4 utf8_table4 -#define _pcre_utt utt -#define _pcre_utt_size utt_size -#define _pcre_utt_names utt_names -#define _pcre_OP_lengths OP_lengths - -#include "pcre_tables.c" - -/* We also need the pcre_printint() function for printing out compiled -patterns. This function is in a separate file so that it can be included in -pcre_compile.c when that module is compiled with debugging enabled. - -The definition of the macro PRINTABLE, which determines whether to print an -output character as-is or as a hex value when showing compiled patterns, is -contained in this file. We uses it here also, in cases when the locale has not -been explicitly changed, so as to get consistent output from systems that -differ in their output from isprint() even in the "C" locale. */ - -#include "pcre_printint.src" - -#define PRINTHEX(c) (locale_set? isprint(c) : PRINTABLE(c)) - - -/* It is possible to compile this test program without including support for -testing the POSIX interface, though this is not available via the standard -Makefile. */ - -#if !defined NOPOSIX -#include "pcreposix.h" -#endif - -/* It is also possible, for the benefit of the version currently imported into -Exim, to build pcretest without support for UTF8 (define NOUTF8), without the -interface to the DFA matcher (NODFA), and without the doublecheck of the old -"info" function (define NOINFOCHECK). In fact, we automatically cut out the -UTF8 support if PCRE is built without it. */ - -#ifndef SUPPORT_UTF8 -#ifndef NOUTF8 -#define NOUTF8 -#endif -#endif - - -/* Other parameters */ - -#ifndef CLOCKS_PER_SEC -#ifdef CLK_TCK -#define CLOCKS_PER_SEC CLK_TCK -#else -#define CLOCKS_PER_SEC 100 -#endif -#endif - -/* This is the default loop count for timing. */ - -#define LOOPREPEAT 500000 - -/* Static variables */ - -static FILE *outfile; -static int log_store = 0; -static int callout_count; -static int callout_extra; -static int callout_fail_count; -static int callout_fail_id; -static int debug_lengths; -static int first_callout; -static int locale_set = 0; -static int show_malloc; -static int use_utf8; -static size_t gotten_store; - -/* The buffers grow automatically if very long input lines are encountered. */ - -static int buffer_size = 50000; -static uschar *buffer = NULL; -static uschar *dbuffer = NULL; -static uschar *pbuffer = NULL; - - - -/************************************************* -* Read or extend an input line * -*************************************************/ - -/* Input lines are read into buffer, but both patterns and data lines can be -continued over multiple input lines. In addition, if the buffer fills up, we -want to automatically expand it so as to be able to handle extremely large -lines that are needed for certain stress tests. When the input buffer is -expanded, the other two buffers must also be expanded likewise, and the -contents of pbuffer, which are a copy of the input for callouts, must be -preserved (for when expansion happens for a data line). This is not the most -optimal way of handling this, but hey, this is just a test program! - -Arguments: - f the file to read - start where in buffer to start (this *must* be within buffer) - prompt for stdin or readline() - -Returns: pointer to the start of new data - could be a copy of start, or could be moved - NULL if no data read and EOF reached -*/ - -static uschar * -extend_inputline(FILE *f, uschar *start, const char *prompt) -{ -uschar *here = start; - -for (;;) - { - int rlen = buffer_size - (here - buffer); - - if (rlen > 1000) - { - int dlen; - - /* If libreadline support is required, use readline() to read a line if the - input is a terminal. Note that readline() removes the trailing newline, so - we must put it back again, to be compatible with fgets(). */ - -#ifdef SUPPORT_LIBREADLINE - if (isatty(fileno(f))) - { - size_t len; - char *s = readline(prompt); - if (s == NULL) return (here == start)? NULL : start; - len = strlen(s); - if (len > 0) add_history(s); - if (len > rlen - 1) len = rlen - 1; - memcpy(here, s, len); - here[len] = '\n'; - here[len+1] = 0; - free(s); - } - else -#endif - - /* Read the next line by normal means, prompting if the file is stdin. */ - - { - if (f == stdin) printf(prompt); - if (fgets((char *)here, rlen, f) == NULL) - return (here == start)? NULL : start; - } - - dlen = (int)strlen((char *)here); - if (dlen > 0 && here[dlen - 1] == '\n') return start; - here += dlen; - } - - else - { - int new_buffer_size = 2*buffer_size; - uschar *new_buffer = (unsigned char *)malloc(new_buffer_size); - uschar *new_dbuffer = (unsigned char *)malloc(new_buffer_size); - uschar *new_pbuffer = (unsigned char *)malloc(new_buffer_size); - - if (new_buffer == NULL || new_dbuffer == NULL || new_pbuffer == NULL) - { - fprintf(stderr, "pcretest: malloc(%d) failed\n", new_buffer_size); - exit(1); - } - - memcpy(new_buffer, buffer, buffer_size); - memcpy(new_pbuffer, pbuffer, buffer_size); - - buffer_size = new_buffer_size; - - start = new_buffer + (start - buffer); - here = new_buffer + (here - buffer); - - free(buffer); - free(dbuffer); - free(pbuffer); - - buffer = new_buffer; - dbuffer = new_dbuffer; - pbuffer = new_pbuffer; - } - } - -return NULL; /* Control never gets here */ -} - - - - - - - -/************************************************* -* Read number from string * -*************************************************/ - -/* We don't use strtoul() because SunOS4 doesn't have it. Rather than mess -around with conditional compilation, just do the job by hand. It is only used -for unpicking arguments, so just keep it simple. - -Arguments: - str string to be converted - endptr where to put the end pointer - -Returns: the unsigned long -*/ - -static int -get_value(unsigned char *str, unsigned char **endptr) -{ -int result = 0; -while(*str != 0 && isspace(*str)) str++; -while (isdigit(*str)) result = result * 10 + (int)(*str++ - '0'); -*endptr = str; -return(result); -} - - - - -/************************************************* -* Convert UTF-8 string to value * -*************************************************/ - -/* This function takes one or more bytes that represents a UTF-8 character, -and returns the value of the character. - -Argument: - utf8bytes a pointer to the byte vector - vptr a pointer to an int to receive the value - -Returns: > 0 => the number of bytes consumed - -6 to 0 => malformed UTF-8 character at offset = (-return) -*/ - -#if !defined NOUTF8 - -static int -utf82ord(unsigned char *utf8bytes, int *vptr) -{ -int c = *utf8bytes++; -int d = c; -int i, j, s; - -for (i = -1; i < 6; i++) /* i is number of additional bytes */ - { - if ((d & 0x80) == 0) break; - d <<= 1; - } - -if (i == -1) { *vptr = c; return 1; } /* ascii character */ -if (i == 0 || i == 6) return 0; /* invalid UTF-8 */ - -/* i now has a value in the range 1-5 */ - -s = 6*i; -d = (c & utf8_table3[i]) << s; - -for (j = 0; j < i; j++) - { - c = *utf8bytes++; - if ((c & 0xc0) != 0x80) return -(j+1); - s -= 6; - d |= (c & 0x3f) << s; - } - -/* Check that encoding was the correct unique one */ - -for (j = 0; j < utf8_table1_size; j++) - if (d <= utf8_table1[j]) break; -if (j != i) return -(i+1); - -/* Valid value */ - -*vptr = d; -return i+1; -} - -#endif - - - -/************************************************* -* Convert character value to UTF-8 * -*************************************************/ - -/* This function takes an integer value in the range 0 - 0x7fffffff -and encodes it as a UTF-8 character in 0 to 6 bytes. - -Arguments: - cvalue the character value - utf8bytes pointer to buffer for result - at least 6 bytes long - -Returns: number of characters placed in the buffer -*/ - -#if !defined NOUTF8 - -static int -ord2utf8(int cvalue, uschar *utf8bytes) -{ -register int i, j; -for (i = 0; i < utf8_table1_size; i++) - if (cvalue <= utf8_table1[i]) break; -utf8bytes += i; -for (j = i; j > 0; j--) - { - *utf8bytes-- = 0x80 | (cvalue & 0x3f); - cvalue >>= 6; - } -*utf8bytes = utf8_table2[i] | cvalue; -return i + 1; -} - -#endif - - - -/************************************************* -* Print character string * -*************************************************/ - -/* Character string printing function. Must handle UTF-8 strings in utf8 -mode. Yields number of characters printed. If handed a NULL file, just counts -chars without printing. */ - -static int pchars(unsigned char *p, int length, FILE *f) -{ -int c = 0; -int yield = 0; - -while (length-- > 0) - { -#if !defined NOUTF8 - if (use_utf8) - { - int rc = utf82ord(p, &c); - - if (rc > 0 && rc <= length + 1) /* Mustn't run over the end */ - { - length -= rc - 1; - p += rc; - if (PRINTHEX(c)) - { - if (f != NULL) fprintf(f, "%c", c); - yield++; - } - else - { - int n = 4; - if (f != NULL) fprintf(f, "\\x{%02x}", c); - yield += (n <= 0x000000ff)? 2 : - (n <= 0x00000fff)? 3 : - (n <= 0x0000ffff)? 4 : - (n <= 0x000fffff)? 5 : 6; - } - continue; - } - } -#endif - - /* Not UTF-8, or malformed UTF-8 */ - - c = *p++; - if (PRINTHEX(c)) - { - if (f != NULL) fprintf(f, "%c", c); - yield++; - } - else - { - if (f != NULL) fprintf(f, "\\x%02x", c); - yield += 4; - } - } - -return yield; -} - - - -/************************************************* -* Callout function * -*************************************************/ - -/* Called from PCRE as a result of the (?C) item. We print out where we are in -the match. Yield zero unless more callouts than the fail count, or the callout -data is not zero. */ - -static int callout(pcre_callout_block *cb) -{ -FILE *f = (first_callout | callout_extra)? outfile : NULL; -int i, pre_start, post_start, subject_length; - -if (callout_extra) - { - fprintf(f, "Callout %d: last capture = %d\n", - cb->callout_number, cb->capture_last); - - for (i = 0; i < cb->capture_top * 2; i += 2) - { - if (cb->offset_vector[i] < 0) - fprintf(f, "%2d: \n", i/2); - else - { - fprintf(f, "%2d: ", i/2); - (void)pchars((unsigned char *)cb->subject + cb->offset_vector[i], - cb->offset_vector[i+1] - cb->offset_vector[i], f); - fprintf(f, "\n"); - } - } - } - -/* Re-print the subject in canonical form, the first time or if giving full -datails. On subsequent calls in the same match, we use pchars just to find the -printed lengths of the substrings. */ - -if (f != NULL) fprintf(f, "--->"); - -pre_start = pchars((unsigned char *)cb->subject, cb->start_match, f); -post_start = pchars((unsigned char *)(cb->subject + cb->start_match), - cb->current_position - cb->start_match, f); - -subject_length = pchars((unsigned char *)cb->subject, cb->subject_length, NULL); - -(void)pchars((unsigned char *)(cb->subject + cb->current_position), - cb->subject_length - cb->current_position, f); - -if (f != NULL) fprintf(f, "\n"); - -/* Always print appropriate indicators, with callout number if not already -shown. For automatic callouts, show the pattern offset. */ - -if (cb->callout_number == 255) - { - fprintf(outfile, "%+3d ", cb->pattern_position); - if (cb->pattern_position > 99) fprintf(outfile, "\n "); - } -else - { - if (callout_extra) fprintf(outfile, " "); - else fprintf(outfile, "%3d ", cb->callout_number); - } - -for (i = 0; i < pre_start; i++) fprintf(outfile, " "); -fprintf(outfile, "^"); - -if (post_start > 0) - { - for (i = 0; i < post_start - 1; i++) fprintf(outfile, " "); - fprintf(outfile, "^"); - } - -for (i = 0; i < subject_length - pre_start - post_start + 4; i++) - fprintf(outfile, " "); - -fprintf(outfile, "%.*s", (cb->next_item_length == 0)? 1 : cb->next_item_length, - pbuffer + cb->pattern_position); - -fprintf(outfile, "\n"); -first_callout = 0; - -if (cb->callout_data != NULL) - { - int callout_data = *((int *)(cb->callout_data)); - if (callout_data != 0) - { - fprintf(outfile, "Callout data = %d\n", callout_data); - return callout_data; - } - } - -return (cb->callout_number != callout_fail_id)? 0 : - (++callout_count >= callout_fail_count)? 1 : 0; -} - - -/************************************************* -* Local malloc functions * -*************************************************/ - -/* Alternative malloc function, to test functionality and show the size of the -compiled re. */ - -static void *new_malloc(size_t size) -{ -void *block = malloc(size); -gotten_store = size; -if (show_malloc) - fprintf(outfile, "malloc %3d %p\n", (int)size, block); -return block; -} - -static void new_free(void *block) -{ -if (show_malloc) - fprintf(outfile, "free %p\n", block); -free(block); -} - - -/* For recursion malloc/free, to test stacking calls */ - -static void *stack_malloc(size_t size) -{ -void *block = malloc(size); -if (show_malloc) - fprintf(outfile, "stack_malloc %3d %p\n", (int)size, block); -return block; -} - -static void stack_free(void *block) -{ -if (show_malloc) - fprintf(outfile, "stack_free %p\n", block); -free(block); -} - - -/************************************************* -* Call pcre_fullinfo() * -*************************************************/ - -/* Get one piece of information from the pcre_fullinfo() function */ - -static void new_info(pcre *re, pcre_extra *study, int option, void *ptr) -{ -int rc; -if ((rc = pcre_fullinfo(re, study, option, ptr)) < 0) - fprintf(outfile, "Error %d from pcre_fullinfo(%d)\n", rc, option); -} - - - -/************************************************* -* Byte flipping function * -*************************************************/ - -static unsigned long int -byteflip(unsigned long int value, int n) -{ -if (n == 2) return ((value & 0x00ff) << 8) | ((value & 0xff00) >> 8); -return ((value & 0x000000ff) << 24) | - ((value & 0x0000ff00) << 8) | - ((value & 0x00ff0000) >> 8) | - ((value & 0xff000000) >> 24); -} - - - - -/************************************************* -* Check match or recursion limit * -*************************************************/ - -static int -check_match_limit(pcre *re, pcre_extra *extra, uschar *bptr, int len, - int start_offset, int options, int *use_offsets, int use_size_offsets, - int flag, unsigned long int *limit, int errnumber, const char *msg) -{ -int count; -int min = 0; -int mid = 64; -int max = -1; - -extra->flags |= flag; - -for (;;) - { - *limit = mid; - - count = pcre_exec(re, extra, (char *)bptr, len, start_offset, options, - use_offsets, use_size_offsets); - - if (count == errnumber) - { - /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */ - min = mid; - mid = (mid == max - 1)? max : (max > 0)? (min + max)/2 : mid*2; - } - - else if (count >= 0 || count == PCRE_ERROR_NOMATCH || - count == PCRE_ERROR_PARTIAL) - { - if (mid == min + 1) - { - fprintf(outfile, "Minimum %s limit = %d\n", msg, mid); - break; - } - /* fprintf(outfile, "Testing %s limit = %d\n", msg, mid); */ - max = mid; - mid = (min + mid)/2; - } - else break; /* Some other error */ - } - -extra->flags &= ~flag; -return count; -} - - - -/************************************************* -* Case-independent strncmp() function * -*************************************************/ - -/* -Arguments: - s first string - t second string - n number of characters to compare - -Returns: < 0, = 0, or > 0, according to the comparison -*/ - -static int -strncmpic(uschar *s, uschar *t, int n) -{ -while (n--) - { - int c = tolower(*s++) - tolower(*t++); - if (c) return c; - } -return 0; -} - - - -/************************************************* -* Check newline indicator * -*************************************************/ - -/* This is used both at compile and run-time to check for escapes, where -xxx is LF, CR, CRLF, ANYCRLF, or ANY. Print a message and return 0 if there is -no match. - -Arguments: - p points after the leading '<' - f file for error message - -Returns: appropriate PCRE_NEWLINE_xxx flags, or 0 -*/ - -static int -check_newline(uschar *p, FILE *f) -{ -if (strncmpic(p, (uschar *)"cr>", 3) == 0) return PCRE_NEWLINE_CR; -if (strncmpic(p, (uschar *)"lf>", 3) == 0) return PCRE_NEWLINE_LF; -if (strncmpic(p, (uschar *)"crlf>", 5) == 0) return PCRE_NEWLINE_CRLF; -if (strncmpic(p, (uschar *)"anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF; -if (strncmpic(p, (uschar *)"any>", 4) == 0) return PCRE_NEWLINE_ANY; -if (strncmpic(p, (uschar *)"bsr_anycrlf>", 12) == 0) return PCRE_BSR_ANYCRLF; -if (strncmpic(p, (uschar *)"bsr_unicode>", 12) == 0) return PCRE_BSR_UNICODE; -fprintf(f, "Unknown newline type at: <%s\n", p); -return 0; -} - - - -/************************************************* -* Usage function * -*************************************************/ - -static void -usage(void) -{ -printf("Usage: pcretest [options] [ []]\n\n"); -printf("Input and output default to stdin and stdout.\n"); -#ifdef SUPPORT_LIBREADLINE -printf("If input is a terminal, readline() is used to read from it.\n"); -#else -printf("This version of pcretest is not linked with readline().\n"); -#endif -printf("\nOptions:\n"); -printf(" -b show compiled code (bytecode)\n"); -printf(" -C show PCRE compile-time options and exit\n"); -printf(" -d debug: show compiled code and information (-b and -i)\n"); -#if !defined NODFA -printf(" -dfa force DFA matching for all subjects\n"); -#endif -printf(" -help show usage information\n"); -printf(" -i show information about compiled patterns\n" - " -m output memory used information\n" - " -o set size of offsets vector to \n"); -#if !defined NOPOSIX -printf(" -p use POSIX interface\n"); -#endif -printf(" -q quiet: do not output PCRE version number at start\n"); -printf(" -S set stack size to megabytes\n"); -printf(" -s output store (memory) used information\n" - " -t time compilation and execution\n"); -printf(" -t time compilation and execution, repeating times\n"); -printf(" -tm time execution (matching) only\n"); -printf(" -tm time execution (matching) only, repeating times\n"); -} - - - -/************************************************* -* Main Program * -*************************************************/ - -/* Read lines from named file or stdin and write to named file or stdout; lines -consist of a regular expression, in delimiters and optionally followed by -options, followed by a set of test data, terminated by an empty line. */ - -int main(int argc, char **argv) -{ -FILE *infile = stdin; -int options = 0; -int study_options = 0; -int op = 1; -int timeit = 0; -int timeitm = 0; -int showinfo = 0; -int showstore = 0; -int quiet = 0; -int size_offsets = 45; -int size_offsets_max; -int *offsets = NULL; -#if !defined NOPOSIX -int posix = 0; -#endif -int debug = 0; -int done = 0; -int all_use_dfa = 0; -int yield = 0; -int stack_size; - -/* These vectors store, end-to-end, a list of captured substring names. Assume -that 1024 is plenty long enough for the few names we'll be testing. */ - -uschar copynames[1024]; -uschar getnames[1024]; - -uschar *copynamesptr; -uschar *getnamesptr; - -/* Get buffers from malloc() so that Electric Fence will check their misuse -when I am debugging. They grow automatically when very long lines are read. */ - -buffer = (unsigned char *)malloc(buffer_size); -dbuffer = (unsigned char *)malloc(buffer_size); -pbuffer = (unsigned char *)malloc(buffer_size); - -/* The outfile variable is static so that new_malloc can use it. */ - -outfile = stdout; - -/* The following _setmode() stuff is some Windows magic that tells its runtime -library to translate CRLF into a single LF character. At least, that's what -I've been told: never having used Windows I take this all on trust. Originally -it set 0x8000, but then I was advised that _O_BINARY was better. */ - -#if defined(_WIN32) || defined(WIN32) -_setmode( _fileno( stdout ), _O_BINARY ); -#endif - -/* Scan options */ - -while (argc > 1 && argv[op][0] == '-') - { - unsigned char *endptr; - - if (strcmp(argv[op], "-s") == 0 || strcmp(argv[op], "-m") == 0) - showstore = 1; - else if (strcmp(argv[op], "-q") == 0) quiet = 1; - else if (strcmp(argv[op], "-b") == 0) debug = 1; - else if (strcmp(argv[op], "-i") == 0) showinfo = 1; - else if (strcmp(argv[op], "-d") == 0) showinfo = debug = 1; -#if !defined NODFA - else if (strcmp(argv[op], "-dfa") == 0) all_use_dfa = 1; -#endif - else if (strcmp(argv[op], "-o") == 0 && argc > 2 && - ((size_offsets = get_value((unsigned char *)argv[op+1], &endptr)), - *endptr == 0)) - { - op++; - argc--; - } - else if (strcmp(argv[op], "-t") == 0 || strcmp(argv[op], "-tm") == 0) - { - int both = argv[op][2] == 0; - int temp; - if (argc > 2 && (temp = get_value((unsigned char *)argv[op+1], &endptr), - *endptr == 0)) - { - timeitm = temp; - op++; - argc--; - } - else timeitm = LOOPREPEAT; - if (both) timeit = timeitm; - } - else if (strcmp(argv[op], "-S") == 0 && argc > 2 && - ((stack_size = get_value((unsigned char *)argv[op+1], &endptr)), - *endptr == 0)) - { -#if defined(_WIN32) || defined(WIN32) - printf("PCRE: -S not supported on this OS\n"); - exit(1); -#else - int rc; - struct rlimit rlim; - getrlimit(RLIMIT_STACK, &rlim); - rlim.rlim_cur = stack_size * 1024 * 1024; - rc = setrlimit(RLIMIT_STACK, &rlim); - if (rc != 0) - { - printf("PCRE: setrlimit() failed with error %d\n", rc); - exit(1); - } - op++; - argc--; -#endif - } -#if !defined NOPOSIX - else if (strcmp(argv[op], "-p") == 0) posix = 1; -#endif - else if (strcmp(argv[op], "-C") == 0) - { - int rc; - printf("PCRE version %s\n", pcre_version()); - printf("Compiled with\n"); - (void)pcre_config(PCRE_CONFIG_UTF8, &rc); - printf(" %sUTF-8 support\n", rc? "" : "No "); - (void)pcre_config(PCRE_CONFIG_UNICODE_PROPERTIES, &rc); - printf(" %sUnicode properties support\n", rc? "" : "No "); - (void)pcre_config(PCRE_CONFIG_NEWLINE, &rc); - printf(" Newline sequence is %s\n", (rc == '\r')? "CR" : - (rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" : - (rc == -2)? "ANYCRLF" : - (rc == -1)? "ANY" : "???"); - (void)pcre_config(PCRE_CONFIG_BSR, &rc); - printf(" \\R matches %s\n", rc? "CR, LF, or CRLF only" : - "all Unicode newlines"); - (void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc); - printf(" Internal link size = %d\n", rc); - (void)pcre_config(PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, &rc); - printf(" POSIX malloc threshold = %d\n", rc); - (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT, &rc); - printf(" Default match limit = %d\n", rc); - (void)pcre_config(PCRE_CONFIG_MATCH_LIMIT_RECURSION, &rc); - printf(" Default recursion depth limit = %d\n", rc); - (void)pcre_config(PCRE_CONFIG_STACKRECURSE, &rc); - printf(" Match recursion uses %s\n", rc? "stack" : "heap"); - goto EXIT; - } - else if (strcmp(argv[op], "-help") == 0 || - strcmp(argv[op], "--help") == 0) - { - usage(); - goto EXIT; - } - else - { - printf("** Unknown or malformed option %s\n", argv[op]); - usage(); - yield = 1; - goto EXIT; - } - op++; - argc--; - } - -/* Get the store for the offsets vector, and remember what it was */ - -size_offsets_max = size_offsets; -offsets = (int *)malloc(size_offsets_max * sizeof(int)); -if (offsets == NULL) - { - printf("** Failed to get %d bytes of memory for offsets vector\n", - (int)(size_offsets_max * sizeof(int))); - yield = 1; - goto EXIT; - } - -/* Sort out the input and output files */ - -if (argc > 1) - { - infile = fopen(argv[op], INPUT_MODE); - if (infile == NULL) - { - printf("** Failed to open %s\n", argv[op]); - yield = 1; - goto EXIT; - } - } - -if (argc > 2) - { - outfile = fopen(argv[op+1], OUTPUT_MODE); - if (outfile == NULL) - { - printf("** Failed to open %s\n", argv[op+1]); - yield = 1; - goto EXIT; - } - } - -/* Set alternative malloc function */ - -pcre_malloc = new_malloc; -pcre_free = new_free; -pcre_stack_malloc = stack_malloc; -pcre_stack_free = stack_free; - -/* Heading line unless quiet, then prompt for first regex if stdin */ - -if (!quiet) fprintf(outfile, "PCRE version %s\n\n", pcre_version()); - -/* Main loop */ - -while (!done) - { - pcre *re = NULL; - pcre_extra *extra = NULL; - -#if !defined NOPOSIX /* There are still compilers that require no indent */ - regex_t preg; - int do_posix = 0; -#endif - - const char *error; - unsigned char *p, *pp, *ppp; - unsigned char *to_file = NULL; - const unsigned char *tables = NULL; - unsigned long int true_size, true_study_size = 0; - size_t size, regex_gotten_store; - int do_study = 0; - int do_debug = debug; - int do_G = 0; - int do_g = 0; - int do_showinfo = showinfo; - int do_showrest = 0; - int do_flip = 0; - int erroroffset, len, delimiter, poffset; - - use_utf8 = 0; - debug_lengths = 1; - - if (extend_inputline(infile, buffer, " re> ") == NULL) break; - if (infile != stdin) fprintf(outfile, "%s", (char *)buffer); - fflush(outfile); - - p = buffer; - while (isspace(*p)) p++; - if (*p == 0) continue; - - /* See if the pattern is to be loaded pre-compiled from a file. */ - - if (*p == '<' && strchr((char *)(p+1), '<') == NULL) - { - unsigned long int magic, get_options; - uschar sbuf[8]; - FILE *f; - - p++; - pp = p + (int)strlen((char *)p); - while (isspace(pp[-1])) pp--; - *pp = 0; - - f = fopen((char *)p, "rb"); - if (f == NULL) - { - fprintf(outfile, "Failed to open %s: %s\n", p, strerror(errno)); - continue; - } - - if (fread(sbuf, 1, 8, f) != 8) goto FAIL_READ; - - true_size = - (sbuf[0] << 24) | (sbuf[1] << 16) | (sbuf[2] << 8) | sbuf[3]; - true_study_size = - (sbuf[4] << 24) | (sbuf[5] << 16) | (sbuf[6] << 8) | sbuf[7]; - - re = (real_pcre *)new_malloc(true_size); - regex_gotten_store = gotten_store; - - if (fread(re, 1, true_size, f) != true_size) goto FAIL_READ; - - magic = ((real_pcre *)re)->magic_number; - if (magic != MAGIC_NUMBER) - { - if (byteflip(magic, sizeof(magic)) == MAGIC_NUMBER) - { - do_flip = 1; - } - else - { - fprintf(outfile, "Data in %s is not a compiled PCRE regex\n", p); - fclose(f); - continue; - } - } - - fprintf(outfile, "Compiled regex%s loaded from %s\n", - do_flip? " (byte-inverted)" : "", p); - - /* Need to know if UTF-8 for printing data strings */ - - new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options); - use_utf8 = (get_options & PCRE_UTF8) != 0; - - /* Now see if there is any following study data */ - - if (true_study_size != 0) - { - pcre_study_data *psd; - - extra = (pcre_extra *)new_malloc(sizeof(pcre_extra) + true_study_size); - extra->flags = PCRE_EXTRA_STUDY_DATA; - - psd = (pcre_study_data *)(((char *)extra) + sizeof(pcre_extra)); - extra->study_data = psd; - - if (fread(psd, 1, true_study_size, f) != true_study_size) - { - FAIL_READ: - fprintf(outfile, "Failed to read data from %s\n", p); - if (extra != NULL) new_free(extra); - if (re != NULL) new_free(re); - fclose(f); - continue; - } - fprintf(outfile, "Study data loaded from %s\n", p); - do_study = 1; /* To get the data output if requested */ - } - else fprintf(outfile, "No study data\n"); - - fclose(f); - goto SHOW_INFO; - } - - /* In-line pattern (the usual case). Get the delimiter and seek the end of - the pattern; if is isn't complete, read more. */ - - delimiter = *p++; - - if (isalnum(delimiter) || delimiter == '\\') - { - fprintf(outfile, "** Delimiter must not be alphanumeric or \\\n"); - goto SKIP_DATA; - } - - pp = p; - poffset = p - buffer; - - for(;;) - { - while (*pp != 0) - { - if (*pp == '\\' && pp[1] != 0) pp++; - else if (*pp == delimiter) break; - pp++; - } - if (*pp != 0) break; - if ((pp = extend_inputline(infile, pp, " > ")) == NULL) - { - fprintf(outfile, "** Unexpected EOF\n"); - done = 1; - goto CONTINUE; - } - if (infile != stdin) fprintf(outfile, "%s", (char *)pp); - } - - /* The buffer may have moved while being extended; reset the start of data - pointer to the correct relative point in the buffer. */ - - p = buffer + poffset; - - /* If the first character after the delimiter is backslash, make - the pattern end with backslash. This is purely to provide a way - of testing for the error message when a pattern ends with backslash. */ - - if (pp[1] == '\\') *pp++ = '\\'; - - /* Terminate the pattern at the delimiter, and save a copy of the pattern - for callouts. */ - - *pp++ = 0; - strcpy((char *)pbuffer, (char *)p); - - /* Look for options after final delimiter */ - - options = 0; - study_options = 0; - log_store = showstore; /* default from command line */ - - while (*pp != 0) - { - switch (*pp++) - { - case 'f': options |= PCRE_FIRSTLINE; break; - case 'g': do_g = 1; break; - case 'i': options |= PCRE_CASELESS; break; - case 'm': options |= PCRE_MULTILINE; break; - case 's': options |= PCRE_DOTALL; break; - case 'x': options |= PCRE_EXTENDED; break; - - case '+': do_showrest = 1; break; - case 'A': options |= PCRE_ANCHORED; break; - case 'B': do_debug = 1; break; - case 'C': options |= PCRE_AUTO_CALLOUT; break; - case 'D': do_debug = do_showinfo = 1; break; - case 'E': options |= PCRE_DOLLAR_ENDONLY; break; - case 'F': do_flip = 1; break; - case 'G': do_G = 1; break; - case 'I': do_showinfo = 1; break; - case 'J': options |= PCRE_DUPNAMES; break; - case 'M': log_store = 1; break; - case 'N': options |= PCRE_NO_AUTO_CAPTURE; break; - -#if !defined NOPOSIX - case 'P': do_posix = 1; break; -#endif - - case 'S': do_study = 1; break; - case 'U': options |= PCRE_UNGREEDY; break; - case 'X': options |= PCRE_EXTRA; break; - case 'Z': debug_lengths = 0; break; - case '8': options |= PCRE_UTF8; use_utf8 = 1; break; - case '?': options |= PCRE_NO_UTF8_CHECK; break; - - case 'L': - ppp = pp; - /* The '\r' test here is so that it works on Windows. */ - /* The '0' test is just in case this is an unterminated line. */ - while (*ppp != 0 && *ppp != '\n' && *ppp != '\r' && *ppp != ' ') ppp++; - *ppp = 0; - if (setlocale(LC_CTYPE, (const char *)pp) == NULL) - { - fprintf(outfile, "** Failed to set locale \"%s\"\n", pp); - goto SKIP_DATA; - } - locale_set = 1; - tables = pcre_maketables(); - pp = ppp; - break; - - case '>': - to_file = pp; - while (*pp != 0) pp++; - while (isspace(pp[-1])) pp--; - *pp = 0; - break; - - case '<': - { - if (strncmp((char *)pp, "JS>", 3) == 0) - { - options |= PCRE_JAVASCRIPT_COMPAT; - pp += 3; - } - else - { - int x = check_newline(pp, outfile); - if (x == 0) goto SKIP_DATA; - options |= x; - while (*pp++ != '>'); - } - } - break; - - case '\r': /* So that it works in Windows */ - case '\n': - case ' ': - break; - - default: - fprintf(outfile, "** Unknown option '%c'\n", pp[-1]); - goto SKIP_DATA; - } - } - - /* Handle compiling via the POSIX interface, which doesn't support the - timing, showing, or debugging options, nor the ability to pass over - local character tables. */ - -#if !defined NOPOSIX - if (posix || do_posix) - { - int rc; - int cflags = 0; - - if ((options & PCRE_CASELESS) != 0) cflags |= REG_ICASE; - if ((options & PCRE_MULTILINE) != 0) cflags |= REG_NEWLINE; - if ((options & PCRE_DOTALL) != 0) cflags |= REG_DOTALL; - if ((options & PCRE_NO_AUTO_CAPTURE) != 0) cflags |= REG_NOSUB; - if ((options & PCRE_UTF8) != 0) cflags |= REG_UTF8; - - rc = regcomp(&preg, (char *)p, cflags); - - /* Compilation failed; go back for another re, skipping to blank line - if non-interactive. */ - - if (rc != 0) - { - (void)regerror(rc, &preg, (char *)buffer, buffer_size); - fprintf(outfile, "Failed: POSIX code %d: %s\n", rc, buffer); - goto SKIP_DATA; - } - } - - /* Handle compiling via the native interface */ - - else -#endif /* !defined NOPOSIX */ - - { - if (timeit > 0) - { - register int i; - clock_t time_taken; - clock_t start_time = clock(); - for (i = 0; i < timeit; i++) - { - re = pcre_compile((char *)p, options, &error, &erroroffset, tables); - if (re != NULL) free(re); - } - time_taken = clock() - start_time; - fprintf(outfile, "Compile time %.4f milliseconds\n", - (((double)time_taken * 1000.0) / (double)timeit) / - (double)CLOCKS_PER_SEC); - } - - re = pcre_compile((char *)p, options, &error, &erroroffset, tables); - - /* Compilation failed; go back for another re, skipping to blank line - if non-interactive. */ - - if (re == NULL) - { - fprintf(outfile, "Failed: %s at offset %d\n", error, erroroffset); - SKIP_DATA: - if (infile != stdin) - { - for (;;) - { - if (extend_inputline(infile, buffer, NULL) == NULL) - { - done = 1; - goto CONTINUE; - } - len = (int)strlen((char *)buffer); - while (len > 0 && isspace(buffer[len-1])) len--; - if (len == 0) break; - } - fprintf(outfile, "\n"); - } - goto CONTINUE; - } - - /* Compilation succeeded; print data if required. There are now two - info-returning functions. The old one has a limited interface and - returns only limited data. Check that it agrees with the newer one. */ - - if (log_store) - fprintf(outfile, "Memory allocation (code space): %d\n", - (int)(gotten_store - - sizeof(real_pcre) - - ((real_pcre *)re)->name_count * ((real_pcre *)re)->name_entry_size)); - - /* Extract the size for possible writing before possibly flipping it, - and remember the store that was got. */ - - true_size = ((real_pcre *)re)->size; - regex_gotten_store = gotten_store; - - /* If /S was present, study the regexp to generate additional info to - help with the matching. */ - - if (do_study) - { - if (timeit > 0) - { - register int i; - clock_t time_taken; - clock_t start_time = clock(); - for (i = 0; i < timeit; i++) - extra = pcre_study(re, study_options, &error); - time_taken = clock() - start_time; - if (extra != NULL) free(extra); - fprintf(outfile, " Study time %.4f milliseconds\n", - (((double)time_taken * 1000.0) / (double)timeit) / - (double)CLOCKS_PER_SEC); - } - extra = pcre_study(re, study_options, &error); - if (error != NULL) - fprintf(outfile, "Failed to study: %s\n", error); - else if (extra != NULL) - true_study_size = ((pcre_study_data *)(extra->study_data))->size; - } - - /* If the 'F' option was present, we flip the bytes of all the integer - fields in the regex data block and the study block. This is to make it - possible to test PCRE's handling of byte-flipped patterns, e.g. those - compiled on a different architecture. */ - - if (do_flip) - { - real_pcre *rre = (real_pcre *)re; - rre->magic_number = - byteflip(rre->magic_number, sizeof(rre->magic_number)); - rre->size = byteflip(rre->size, sizeof(rre->size)); - rre->options = byteflip(rre->options, sizeof(rre->options)); - rre->flags = (pcre_uint16)byteflip(rre->flags, sizeof(rre->flags)); - rre->top_bracket = - (pcre_uint16)byteflip(rre->top_bracket, sizeof(rre->top_bracket)); - rre->top_backref = - (pcre_uint16)byteflip(rre->top_backref, sizeof(rre->top_backref)); - rre->first_byte = - (pcre_uint16)byteflip(rre->first_byte, sizeof(rre->first_byte)); - rre->req_byte = - (pcre_uint16)byteflip(rre->req_byte, sizeof(rre->req_byte)); - rre->name_table_offset = (pcre_uint16)byteflip(rre->name_table_offset, - sizeof(rre->name_table_offset)); - rre->name_entry_size = (pcre_uint16)byteflip(rre->name_entry_size, - sizeof(rre->name_entry_size)); - rre->name_count = (pcre_uint16)byteflip(rre->name_count, - sizeof(rre->name_count)); - - if (extra != NULL) - { - pcre_study_data *rsd = (pcre_study_data *)(extra->study_data); - rsd->size = byteflip(rsd->size, sizeof(rsd->size)); - rsd->options = byteflip(rsd->options, sizeof(rsd->options)); - } - } - - /* Extract information from the compiled data if required */ - - SHOW_INFO: - - if (do_debug) - { - fprintf(outfile, "------------------------------------------------------------------\n"); - pcre_printint(re, outfile, debug_lengths); - } - - if (do_showinfo) - { - unsigned long int get_options, all_options; -#if !defined NOINFOCHECK - int old_first_char, old_options, old_count; -#endif - int count, backrefmax, first_char, need_char, okpartial, jchanged, - hascrorlf; - int nameentrysize, namecount; - const uschar *nametable; - - new_info(re, NULL, PCRE_INFO_OPTIONS, &get_options); - new_info(re, NULL, PCRE_INFO_SIZE, &size); - new_info(re, NULL, PCRE_INFO_CAPTURECOUNT, &count); - new_info(re, NULL, PCRE_INFO_BACKREFMAX, &backrefmax); - new_info(re, NULL, PCRE_INFO_FIRSTBYTE, &first_char); - new_info(re, NULL, PCRE_INFO_LASTLITERAL, &need_char); - new_info(re, NULL, PCRE_INFO_NAMEENTRYSIZE, &nameentrysize); - new_info(re, NULL, PCRE_INFO_NAMECOUNT, &namecount); - new_info(re, NULL, PCRE_INFO_NAMETABLE, (void *)&nametable); - new_info(re, NULL, PCRE_INFO_OKPARTIAL, &okpartial); - new_info(re, NULL, PCRE_INFO_JCHANGED, &jchanged); - new_info(re, NULL, PCRE_INFO_HASCRORLF, &hascrorlf); - -#if !defined NOINFOCHECK - old_count = pcre_info(re, &old_options, &old_first_char); - if (count < 0) fprintf(outfile, - "Error %d from pcre_info()\n", count); - else - { - if (old_count != count) fprintf(outfile, - "Count disagreement: pcre_fullinfo=%d pcre_info=%d\n", count, - old_count); - - if (old_first_char != first_char) fprintf(outfile, - "First char disagreement: pcre_fullinfo=%d pcre_info=%d\n", - first_char, old_first_char); - - if (old_options != (int)get_options) fprintf(outfile, - "Options disagreement: pcre_fullinfo=%ld pcre_info=%d\n", - get_options, old_options); - } -#endif - - if (size != regex_gotten_store) fprintf(outfile, - "Size disagreement: pcre_fullinfo=%d call to malloc for %d\n", - (int)size, (int)regex_gotten_store); - - fprintf(outfile, "Capturing subpattern count = %d\n", count); - if (backrefmax > 0) - fprintf(outfile, "Max back reference = %d\n", backrefmax); - - if (namecount > 0) - { - fprintf(outfile, "Named capturing subpatterns:\n"); - while (namecount-- > 0) - { - fprintf(outfile, " %s %*s%3d\n", nametable + 2, - nameentrysize - 3 - (int)strlen((char *)nametable + 2), "", - GET2(nametable, 0)); - nametable += nameentrysize; - } - } - - if (!okpartial) fprintf(outfile, "Partial matching not supported\n"); - if (hascrorlf) fprintf(outfile, "Contains explicit CR or LF match\n"); - - all_options = ((real_pcre *)re)->options; - if (do_flip) all_options = byteflip(all_options, sizeof(all_options)); - - if (get_options == 0) fprintf(outfile, "No options\n"); - else fprintf(outfile, "Options:%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n", - ((get_options & PCRE_ANCHORED) != 0)? " anchored" : "", - ((get_options & PCRE_CASELESS) != 0)? " caseless" : "", - ((get_options & PCRE_EXTENDED) != 0)? " extended" : "", - ((get_options & PCRE_MULTILINE) != 0)? " multiline" : "", - ((get_options & PCRE_FIRSTLINE) != 0)? " firstline" : "", - ((get_options & PCRE_DOTALL) != 0)? " dotall" : "", - ((get_options & PCRE_BSR_ANYCRLF) != 0)? " bsr_anycrlf" : "", - ((get_options & PCRE_BSR_UNICODE) != 0)? " bsr_unicode" : "", - ((get_options & PCRE_DOLLAR_ENDONLY) != 0)? " dollar_endonly" : "", - ((get_options & PCRE_EXTRA) != 0)? " extra" : "", - ((get_options & PCRE_UNGREEDY) != 0)? " ungreedy" : "", - ((get_options & PCRE_NO_AUTO_CAPTURE) != 0)? " no_auto_capture" : "", - ((get_options & PCRE_UTF8) != 0)? " utf8" : "", - ((get_options & PCRE_NO_UTF8_CHECK) != 0)? " no_utf8_check" : "", - ((get_options & PCRE_DUPNAMES) != 0)? " dupnames" : ""); - - if (jchanged) fprintf(outfile, "Duplicate name status changes\n"); - - switch (get_options & PCRE_NEWLINE_BITS) - { - case PCRE_NEWLINE_CR: - fprintf(outfile, "Forced newline sequence: CR\n"); - break; - - case PCRE_NEWLINE_LF: - fprintf(outfile, "Forced newline sequence: LF\n"); - break; - - case PCRE_NEWLINE_CRLF: - fprintf(outfile, "Forced newline sequence: CRLF\n"); - break; - - case PCRE_NEWLINE_ANYCRLF: - fprintf(outfile, "Forced newline sequence: ANYCRLF\n"); - break; - - case PCRE_NEWLINE_ANY: - fprintf(outfile, "Forced newline sequence: ANY\n"); - break; - - default: - break; - } - - if (first_char == -1) - { - fprintf(outfile, "First char at start or follows newline\n"); - } - else if (first_char < 0) - { - fprintf(outfile, "No first char\n"); - } - else - { - int ch = first_char & 255; - const char *caseless = ((first_char & REQ_CASELESS) == 0)? - "" : " (caseless)"; - if (PRINTHEX(ch)) - fprintf(outfile, "First char = \'%c\'%s\n", ch, caseless); - else - fprintf(outfile, "First char = %d%s\n", ch, caseless); - } - - if (need_char < 0) - { - fprintf(outfile, "No need char\n"); - } - else - { - int ch = need_char & 255; - const char *caseless = ((need_char & REQ_CASELESS) == 0)? - "" : " (caseless)"; - if (PRINTHEX(ch)) - fprintf(outfile, "Need char = \'%c\'%s\n", ch, caseless); - else - fprintf(outfile, "Need char = %d%s\n", ch, caseless); - } - - /* Don't output study size; at present it is in any case a fixed - value, but it varies, depending on the computer architecture, and - so messes up the test suite. (And with the /F option, it might be - flipped.) */ - - if (do_study) - { - if (extra == NULL) - fprintf(outfile, "Study returned NULL\n"); - else - { - uschar *start_bits = NULL; - new_info(re, extra, PCRE_INFO_FIRSTTABLE, &start_bits); - - if (start_bits == NULL) - fprintf(outfile, "No starting byte set\n"); - else - { - int i; - int c = 24; - fprintf(outfile, "Starting byte set: "); - for (i = 0; i < 256; i++) - { - if ((start_bits[i/8] & (1<<(i&7))) != 0) - { - if (c > 75) - { - fprintf(outfile, "\n "); - c = 2; - } - if (PRINTHEX(i) && i != ' ') - { - fprintf(outfile, "%c ", i); - c += 2; - } - else - { - fprintf(outfile, "\\x%02x ", i); - c += 5; - } - } - } - fprintf(outfile, "\n"); - } - } - } - } - - /* If the '>' option was present, we write out the regex to a file, and - that is all. The first 8 bytes of the file are the regex length and then - the study length, in big-endian order. */ - - if (to_file != NULL) - { - FILE *f = fopen((char *)to_file, "wb"); - if (f == NULL) - { - fprintf(outfile, "Unable to open %s: %s\n", to_file, strerror(errno)); - } - else - { - uschar sbuf[8]; - sbuf[0] = (uschar)((true_size >> 24) & 255); - sbuf[1] = (uschar)((true_size >> 16) & 255); - sbuf[2] = (uschar)((true_size >> 8) & 255); - sbuf[3] = (uschar)((true_size) & 255); - - sbuf[4] = (uschar)((true_study_size >> 24) & 255); - sbuf[5] = (uschar)((true_study_size >> 16) & 255); - sbuf[6] = (uschar)((true_study_size >> 8) & 255); - sbuf[7] = (uschar)((true_study_size) & 255); - - if (fwrite(sbuf, 1, 8, f) < 8 || - fwrite(re, 1, true_size, f) < true_size) - { - fprintf(outfile, "Write error on %s: %s\n", to_file, strerror(errno)); - } - else - { - fprintf(outfile, "Compiled regex written to %s\n", to_file); - if (extra != NULL) - { - if (fwrite(extra->study_data, 1, true_study_size, f) < - true_study_size) - { - fprintf(outfile, "Write error on %s: %s\n", to_file, - strerror(errno)); - } - else fprintf(outfile, "Study data written to %s\n", to_file); - - } - } - fclose(f); - } - - new_free(re); - if (extra != NULL) new_free(extra); - if (tables != NULL) new_free((void *)tables); - continue; /* With next regex */ - } - } /* End of non-POSIX compile */ - - /* Read data lines and test them */ - - for (;;) - { - uschar *q; - uschar *bptr; - int *use_offsets = offsets; - int use_size_offsets = size_offsets; - int callout_data = 0; - int callout_data_set = 0; - int count, c; - int copystrings = 0; - int find_match_limit = 0; - int getstrings = 0; - int getlist = 0; - int gmatched = 0; - int start_offset = 0; - int g_notempty = 0; - int use_dfa = 0; - - options = 0; - - *copynames = 0; - *getnames = 0; - - copynamesptr = copynames; - getnamesptr = getnames; - - pcre_callout = callout; - first_callout = 1; - callout_extra = 0; - callout_count = 0; - callout_fail_count = 999999; - callout_fail_id = -1; - show_malloc = 0; - - if (extra != NULL) extra->flags &= - ~(PCRE_EXTRA_MATCH_LIMIT|PCRE_EXTRA_MATCH_LIMIT_RECURSION); - - len = 0; - for (;;) - { - if (extend_inputline(infile, buffer + len, "data> ") == NULL) - { - if (len > 0) break; - done = 1; - goto CONTINUE; - } - if (infile != stdin) fprintf(outfile, "%s", (char *)buffer); - len = (int)strlen((char *)buffer); - if (buffer[len-1] == '\n') break; - } - - while (len > 0 && isspace(buffer[len-1])) len--; - buffer[len] = 0; - if (len == 0) break; - - p = buffer; - while (isspace(*p)) p++; - - bptr = q = dbuffer; - while ((c = *p++) != 0) - { - int i = 0; - int n = 0; - - if (c == '\\') switch ((c = *p++)) - { - case 'a': c = 7; break; - case 'b': c = '\b'; break; - case 'e': c = 27; break; - case 'f': c = '\f'; break; - case 'n': c = '\n'; break; - case 'r': c = '\r'; break; - case 't': c = '\t'; break; - case 'v': c = '\v'; break; - - case '0': case '1': case '2': case '3': - case '4': case '5': case '6': case '7': - c -= '0'; - while (i++ < 2 && isdigit(*p) && *p != '8' && *p != '9') - c = c * 8 + *p++ - '0'; - -#if !defined NOUTF8 - if (use_utf8 && c > 255) - { - unsigned char buff8[8]; - int ii, utn; - utn = ord2utf8(c, buff8); - for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii]; - c = buff8[ii]; /* Last byte */ - } -#endif - break; - - case 'x': - - /* Handle \x{..} specially - new Perl thing for utf8 */ - -#if !defined NOUTF8 - if (*p == '{') - { - unsigned char *pt = p; - c = 0; - while (isxdigit(*(++pt))) - c = c * 16 + tolower(*pt) - ((isdigit(*pt))? '0' : 'W'); - if (*pt == '}') - { - unsigned char buff8[8]; - int ii, utn; - utn = ord2utf8(c, buff8); - for (ii = 0; ii < utn - 1; ii++) *q++ = buff8[ii]; - c = buff8[ii]; /* Last byte */ - p = pt + 1; - break; - } - /* Not correct form; fall through */ - } -#endif - - /* Ordinary \x */ - - c = 0; - while (i++ < 2 && isxdigit(*p)) - { - c = c * 16 + tolower(*p) - ((isdigit(*p))? '0' : 'W'); - p++; - } - break; - - case 0: /* \ followed by EOF allows for an empty line */ - p--; - continue; - - case '>': - while(isdigit(*p)) start_offset = start_offset * 10 + *p++ - '0'; - continue; - - case 'A': /* Option setting */ - options |= PCRE_ANCHORED; - continue; - - case 'B': - options |= PCRE_NOTBOL; - continue; - - case 'C': - if (isdigit(*p)) /* Set copy string */ - { - while(isdigit(*p)) n = n * 10 + *p++ - '0'; - copystrings |= 1 << n; - } - else if (isalnum(*p)) - { - uschar *npp = copynamesptr; - while (isalnum(*p)) *npp++ = *p++; - *npp++ = 0; - *npp = 0; - n = pcre_get_stringnumber(re, (char *)copynamesptr); - if (n < 0) - fprintf(outfile, "no parentheses with name \"%s\"\n", copynamesptr); - copynamesptr = npp; - } - else if (*p == '+') - { - callout_extra = 1; - p++; - } - else if (*p == '-') - { - pcre_callout = NULL; - p++; - } - else if (*p == '!') - { - callout_fail_id = 0; - p++; - while(isdigit(*p)) - callout_fail_id = callout_fail_id * 10 + *p++ - '0'; - callout_fail_count = 0; - if (*p == '!') - { - p++; - while(isdigit(*p)) - callout_fail_count = callout_fail_count * 10 + *p++ - '0'; - } - } - else if (*p == '*') - { - int sign = 1; - callout_data = 0; - if (*(++p) == '-') { sign = -1; p++; } - while(isdigit(*p)) - callout_data = callout_data * 10 + *p++ - '0'; - callout_data *= sign; - callout_data_set = 1; - } - continue; - -#if !defined NODFA - case 'D': -#if !defined NOPOSIX - if (posix || do_posix) - printf("** Can't use dfa matching in POSIX mode: \\D ignored\n"); - else -#endif - use_dfa = 1; - continue; - - case 'F': - options |= PCRE_DFA_SHORTEST; - continue; -#endif - - case 'G': - if (isdigit(*p)) - { - while(isdigit(*p)) n = n * 10 + *p++ - '0'; - getstrings |= 1 << n; - } - else if (isalnum(*p)) - { - uschar *npp = getnamesptr; - while (isalnum(*p)) *npp++ = *p++; - *npp++ = 0; - *npp = 0; - n = pcre_get_stringnumber(re, (char *)getnamesptr); - if (n < 0) - fprintf(outfile, "no parentheses with name \"%s\"\n", getnamesptr); - getnamesptr = npp; - } - continue; - - case 'L': - getlist = 1; - continue; - - case 'M': - find_match_limit = 1; - continue; - - case 'N': - options |= PCRE_NOTEMPTY; - continue; - - case 'O': - while(isdigit(*p)) n = n * 10 + *p++ - '0'; - if (n > size_offsets_max) - { - size_offsets_max = n; - free(offsets); - use_offsets = offsets = (int *)malloc(size_offsets_max * sizeof(int)); - if (offsets == NULL) - { - printf("** Failed to get %d bytes of memory for offsets vector\n", - (int)(size_offsets_max * sizeof(int))); - yield = 1; - goto EXIT; - } - } - use_size_offsets = n; - if (n == 0) use_offsets = NULL; /* Ensures it can't write to it */ - continue; - - case 'P': - options |= PCRE_PARTIAL; - continue; - - case 'Q': - while(isdigit(*p)) n = n * 10 + *p++ - '0'; - if (extra == NULL) - { - extra = (pcre_extra *)malloc(sizeof(pcre_extra)); - extra->flags = 0; - } - extra->flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION; - extra->match_limit_recursion = n; - continue; - - case 'q': - while(isdigit(*p)) n = n * 10 + *p++ - '0'; - if (extra == NULL) - { - extra = (pcre_extra *)malloc(sizeof(pcre_extra)); - extra->flags = 0; - } - extra->flags |= PCRE_EXTRA_MATCH_LIMIT; - extra->match_limit = n; - continue; - -#if !defined NODFA - case 'R': - options |= PCRE_DFA_RESTART; - continue; -#endif - - case 'S': - show_malloc = 1; - continue; - - case 'Z': - options |= PCRE_NOTEOL; - continue; - - case '?': - options |= PCRE_NO_UTF8_CHECK; - continue; - - case '<': - { - int x = check_newline(p, outfile); - if (x == 0) goto NEXT_DATA; - options |= x; - while (*p++ != '>'); - } - continue; - } - *q++ = c; - } - *q = 0; - len = q - dbuffer; - - if ((all_use_dfa || use_dfa) && find_match_limit) - { - printf("**Match limit not relevant for DFA matching: ignored\n"); - find_match_limit = 0; - } - - /* Handle matching via the POSIX interface, which does not - support timing or playing with the match limit or callout data. */ - -#if !defined NOPOSIX - if (posix || do_posix) - { - int rc; - int eflags = 0; - regmatch_t *pmatch = NULL; - if (use_size_offsets > 0) - pmatch = (regmatch_t *)malloc(sizeof(regmatch_t) * use_size_offsets); - if ((options & PCRE_NOTBOL) != 0) eflags |= REG_NOTBOL; - if ((options & PCRE_NOTEOL) != 0) eflags |= REG_NOTEOL; - - rc = regexec(&preg, (const char *)bptr, use_size_offsets, pmatch, eflags); - - if (rc != 0) - { - (void)regerror(rc, &preg, (char *)buffer, buffer_size); - fprintf(outfile, "No match: POSIX code %d: %s\n", rc, buffer); - } - else if ((((const pcre *)preg.re_pcre)->options & PCRE_NO_AUTO_CAPTURE) - != 0) - { - fprintf(outfile, "Matched with REG_NOSUB\n"); - } - else - { - size_t i; - for (i = 0; i < (size_t)use_size_offsets; i++) - { - if (pmatch[i].rm_so >= 0) - { - fprintf(outfile, "%2d: ", (int)i); - (void)pchars(dbuffer + pmatch[i].rm_so, - pmatch[i].rm_eo - pmatch[i].rm_so, outfile); - fprintf(outfile, "\n"); - if (i == 0 && do_showrest) - { - fprintf(outfile, " 0+ "); - (void)pchars(dbuffer + pmatch[i].rm_eo, len - pmatch[i].rm_eo, - outfile); - fprintf(outfile, "\n"); - } - } - } - } - free(pmatch); - } - - /* Handle matching via the native interface - repeats for /g and /G */ - - else -#endif /* !defined NOPOSIX */ - - for (;; gmatched++) /* Loop for /g or /G */ - { - if (timeitm > 0) - { - register int i; - clock_t time_taken; - clock_t start_time = clock(); - -#if !defined NODFA - if (all_use_dfa || use_dfa) - { - int workspace[1000]; - for (i = 0; i < timeitm; i++) - count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset, - options | g_notempty, use_offsets, use_size_offsets, workspace, - sizeof(workspace)/sizeof(int)); - } - else -#endif - - for (i = 0; i < timeitm; i++) - count = pcre_exec(re, extra, (char *)bptr, len, - start_offset, options | g_notempty, use_offsets, use_size_offsets); - - time_taken = clock() - start_time; - fprintf(outfile, "Execute time %.4f milliseconds\n", - (((double)time_taken * 1000.0) / (double)timeitm) / - (double)CLOCKS_PER_SEC); - } - - /* If find_match_limit is set, we want to do repeated matches with - varying limits in order to find the minimum value for the match limit and - for the recursion limit. */ - - if (find_match_limit) - { - if (extra == NULL) - { - extra = (pcre_extra *)malloc(sizeof(pcre_extra)); - extra->flags = 0; - } - - (void)check_match_limit(re, extra, bptr, len, start_offset, - options|g_notempty, use_offsets, use_size_offsets, - PCRE_EXTRA_MATCH_LIMIT, &(extra->match_limit), - PCRE_ERROR_MATCHLIMIT, "match()"); - - count = check_match_limit(re, extra, bptr, len, start_offset, - options|g_notempty, use_offsets, use_size_offsets, - PCRE_EXTRA_MATCH_LIMIT_RECURSION, &(extra->match_limit_recursion), - PCRE_ERROR_RECURSIONLIMIT, "match() recursion"); - } - - /* If callout_data is set, use the interface with additional data */ - - else if (callout_data_set) - { - if (extra == NULL) - { - extra = (pcre_extra *)malloc(sizeof(pcre_extra)); - extra->flags = 0; - } - extra->flags |= PCRE_EXTRA_CALLOUT_DATA; - extra->callout_data = &callout_data; - count = pcre_exec(re, extra, (char *)bptr, len, start_offset, - options | g_notempty, use_offsets, use_size_offsets); - extra->flags &= ~PCRE_EXTRA_CALLOUT_DATA; - } - - /* The normal case is just to do the match once, with the default - value of match_limit. */ - -#if !defined NODFA - else if (all_use_dfa || use_dfa) - { - int workspace[1000]; - count = pcre_dfa_exec(re, NULL, (char *)bptr, len, start_offset, - options | g_notempty, use_offsets, use_size_offsets, workspace, - sizeof(workspace)/sizeof(int)); - if (count == 0) - { - fprintf(outfile, "Matched, but too many subsidiary matches\n"); - count = use_size_offsets/2; - } - } -#endif - - else - { - count = pcre_exec(re, extra, (char *)bptr, len, - start_offset, options | g_notempty, use_offsets, use_size_offsets); - if (count == 0) - { - fprintf(outfile, "Matched, but too many substrings\n"); - count = use_size_offsets/3; - } - } - - /* Matched */ - - if (count >= 0) - { - int i, maxcount; - -#if !defined NODFA - if (all_use_dfa || use_dfa) maxcount = use_size_offsets/2; else -#endif - maxcount = use_size_offsets/3; - - /* This is a check against a lunatic return value. */ - - if (count > maxcount) - { - fprintf(outfile, - "** PCRE error: returned count %d is too big for offset size %d\n", - count, use_size_offsets); - count = use_size_offsets/3; - if (do_g || do_G) - { - fprintf(outfile, "** /%c loop abandoned\n", do_g? 'g' : 'G'); - do_g = do_G = FALSE; /* Break g/G loop */ - } - } - - for (i = 0; i < count * 2; i += 2) - { - if (use_offsets[i] < 0) - fprintf(outfile, "%2d: \n", i/2); - else - { - fprintf(outfile, "%2d: ", i/2); - (void)pchars(bptr + use_offsets[i], - use_offsets[i+1] - use_offsets[i], outfile); - fprintf(outfile, "\n"); - if (i == 0) - { - if (do_showrest) - { - fprintf(outfile, " 0+ "); - (void)pchars(bptr + use_offsets[i+1], len - use_offsets[i+1], - outfile); - fprintf(outfile, "\n"); - } - } - } - } - - for (i = 0; i < 32; i++) - { - if ((copystrings & (1 << i)) != 0) - { - char copybuffer[256]; - int rc = pcre_copy_substring((char *)bptr, use_offsets, count, - i, copybuffer, sizeof(copybuffer)); - if (rc < 0) - fprintf(outfile, "copy substring %d failed %d\n", i, rc); - else - fprintf(outfile, "%2dC %s (%d)\n", i, copybuffer, rc); - } - } - - for (copynamesptr = copynames; - *copynamesptr != 0; - copynamesptr += (int)strlen((char*)copynamesptr) + 1) - { - char copybuffer[256]; - int rc = pcre_copy_named_substring(re, (char *)bptr, use_offsets, - count, (char *)copynamesptr, copybuffer, sizeof(copybuffer)); - if (rc < 0) - fprintf(outfile, "copy substring %s failed %d\n", copynamesptr, rc); - else - fprintf(outfile, " C %s (%d) %s\n", copybuffer, rc, copynamesptr); - } - - for (i = 0; i < 32; i++) - { - if ((getstrings & (1 << i)) != 0) - { - const char *substring; - int rc = pcre_get_substring((char *)bptr, use_offsets, count, - i, &substring); - if (rc < 0) - fprintf(outfile, "get substring %d failed %d\n", i, rc); - else - { - fprintf(outfile, "%2dG %s (%d)\n", i, substring, rc); - pcre_free_substring(substring); - } - } - } - - for (getnamesptr = getnames; - *getnamesptr != 0; - getnamesptr += (int)strlen((char*)getnamesptr) + 1) - { - const char *substring; - int rc = pcre_get_named_substring(re, (char *)bptr, use_offsets, - count, (char *)getnamesptr, &substring); - if (rc < 0) - fprintf(outfile, "copy substring %s failed %d\n", getnamesptr, rc); - else - { - fprintf(outfile, " G %s (%d) %s\n", substring, rc, getnamesptr); - pcre_free_substring(substring); - } - } - - if (getlist) - { - const char **stringlist; - int rc = pcre_get_substring_list((char *)bptr, use_offsets, count, - &stringlist); - if (rc < 0) - fprintf(outfile, "get substring list failed %d\n", rc); - else - { - for (i = 0; i < count; i++) - fprintf(outfile, "%2dL %s\n", i, stringlist[i]); - if (stringlist[i] != NULL) - fprintf(outfile, "string list not terminated by NULL\n"); - /* free((void *)stringlist); */ - pcre_free_substring_list(stringlist); - } - } - } - - /* There was a partial match */ - - else if (count == PCRE_ERROR_PARTIAL) - { - fprintf(outfile, "Partial match"); -#if !defined NODFA - if ((all_use_dfa || use_dfa) && use_size_offsets > 2) - fprintf(outfile, ": %.*s", use_offsets[1] - use_offsets[0], - bptr + use_offsets[0]); -#endif - fprintf(outfile, "\n"); - break; /* Out of the /g loop */ - } - - /* Failed to match. If this is a /g or /G loop and we previously set - g_notempty after a null match, this is not necessarily the end. We want - to advance the start offset, and continue. We won't be at the end of the - string - that was checked before setting g_notempty. - - Complication arises in the case when the newline option is "any" or - "anycrlf". If the previous match was at the end of a line terminated by - CRLF, an advance of one character just passes the \r, whereas we should - prefer the longer newline sequence, as does the code in pcre_exec(). - Fudge the offset value to achieve this. - - Otherwise, in the case of UTF-8 matching, the advance must be one - character, not one byte. */ - - else - { - if (g_notempty != 0) - { - int onechar = 1; - unsigned int obits = ((real_pcre *)re)->options; - use_offsets[0] = start_offset; - if ((obits & PCRE_NEWLINE_BITS) == 0) - { - int d; - (void)pcre_config(PCRE_CONFIG_NEWLINE, &d); - obits = (d == '\r')? PCRE_NEWLINE_CR : - (d == '\n')? PCRE_NEWLINE_LF : - (d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF : - (d == -2)? PCRE_NEWLINE_ANYCRLF : - (d == -1)? PCRE_NEWLINE_ANY : 0; - } - if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY || - (obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF) - && - start_offset < len - 1 && - bptr[start_offset] == '\r' && - bptr[start_offset+1] == '\n') - onechar++; - else if (use_utf8) - { - while (start_offset + onechar < len) - { - int tb = bptr[start_offset+onechar]; - if (tb <= 127) break; - tb &= 0xc0; - if (tb != 0 && tb != 0xc0) onechar++; - } - } - use_offsets[1] = start_offset + onechar; - } - else - { - if (count == PCRE_ERROR_NOMATCH) - { - if (gmatched == 0) fprintf(outfile, "No match\n"); - } - else fprintf(outfile, "Error %d\n", count); - break; /* Out of the /g loop */ - } - } - - /* If not /g or /G we are done */ - - if (!do_g && !do_G) break; - - /* If we have matched an empty string, first check to see if we are at - the end of the subject. If so, the /g loop is over. Otherwise, mimic - what Perl's /g options does. This turns out to be rather cunning. First - we set PCRE_NOTEMPTY and PCRE_ANCHORED and try the match again at the - same point. If this fails (picked up above) we advance to the next - character. */ - - g_notempty = 0; - - if (use_offsets[0] == use_offsets[1]) - { - if (use_offsets[0] == len) break; - g_notempty = PCRE_NOTEMPTY | PCRE_ANCHORED; - } - - /* For /g, update the start offset, leaving the rest alone */ - - if (do_g) start_offset = use_offsets[1]; - - /* For /G, update the pointer and length */ - - else - { - bptr += use_offsets[1]; - len -= use_offsets[1]; - } - } /* End of loop for /g and /G */ - - NEXT_DATA: continue; - } /* End of loop for data lines */ - - CONTINUE: - -#if !defined NOPOSIX - if (posix || do_posix) regfree(&preg); -#endif - - if (re != NULL) new_free(re); - if (extra != NULL) new_free(extra); - if (tables != NULL) - { - new_free((void *)tables); - setlocale(LC_CTYPE, "C"); - locale_set = 0; - } - } - -if (infile == stdin) fprintf(outfile, "\n"); - -EXIT: - -if (infile != NULL && infile != stdin) fclose(infile); -if (outfile != NULL && outfile != stdout) fclose(outfile); - -free(buffer); -free(dbuffer); -free(pbuffer); -free(offsets); - -return yield; -} - -/* End of pcretest.c */ diff --git a/pcre/ucp.h b/pcre/ucp.h --- a/pcre/ucp.h +++ b/pcre/ucp.h @@ -6,9 +6,8 @@ #define _UCP_H /* This file contains definitions of the property values that are returned by -the function _pcre_ucp_findprop(). New values that are added for new releases -of Unicode should always be at the end of each enum, for backwards -compatibility. */ +the UCD access macros. New values that are added for new releases of Unicode +should always be at the end of each enum, for backwards compatibility. */ /* These are the general character categories. */ @@ -121,11 +120,52 @@ ucp_Tifinagh, ucp_Ugaritic, ucp_Yi, - ucp_Balinese, /* New for Unicode 5.0.0 */ - ucp_Cuneiform, /* New for Unicode 5.0.0 */ - ucp_Nko, /* New for Unicode 5.0.0 */ - ucp_Phags_Pa, /* New for Unicode 5.0.0 */ - ucp_Phoenician /* New for Unicode 5.0.0 */ + /* New for Unicode 5.0: */ + ucp_Balinese, + ucp_Cuneiform, + ucp_Nko, + ucp_Phags_Pa, + ucp_Phoenician, + /* New for Unicode 5.1: */ + ucp_Carian, + ucp_Cham, + ucp_Kayah_Li, + ucp_Lepcha, + ucp_Lycian, + ucp_Lydian, + ucp_Ol_Chiki, + ucp_Rejang, + ucp_Saurashtra, + ucp_Sundanese, + ucp_Vai, + /* New for Unicode 5.2: */ + ucp_Avestan, + ucp_Bamum, + ucp_Egyptian_Hieroglyphs, + ucp_Imperial_Aramaic, + ucp_Inscriptional_Pahlavi, + ucp_Inscriptional_Parthian, + ucp_Javanese, + ucp_Kaithi, + ucp_Lisu, + ucp_Meetei_Mayek, + ucp_Old_South_Arabian, + ucp_Old_Turkic, + ucp_Samaritan, + ucp_Tai_Tham, + ucp_Tai_Viet, + /* New for Unicode 6.0.0: */ + ucp_Batak, + ucp_Brahmi, + ucp_Mandaic, + /* New for Unicode 6.1.0: */ + ucp_Chakma, + ucp_Meroitic_Cursive, + ucp_Meroitic_Hieroglyphs, + ucp_Miao, + ucp_Sharada, + ucp_Sora_Sompeng, + ucp_Takri }; #endif diff --git a/pcre/ucpinternal.h b/pcre/ucpinternal.h deleted file mode 100644 --- a/pcre/ucpinternal.h +++ /dev/null @@ -1,92 +0,0 @@ -/************************************************* -* Unicode Property Table handler * -*************************************************/ - -#ifndef _UCPINTERNAL_H -#define _UCPINTERNAL_H - -/* Internal header file defining the layout of the bits in each pair of 32-bit -words that form a data item in the table. */ - -typedef struct cnode { - pcre_uint32 f0; - pcre_uint32 f1; -} cnode; - -/* Things for the f0 field */ - -#define f0_scriptmask 0xff000000 /* Mask for script field */ -#define f0_scriptshift 24 /* Shift for script value */ -#define f0_rangeflag 0x00800000 /* Flag for a range item */ -#define f0_charmask 0x001fffff /* Mask for code point value */ - -/* Things for the f1 field */ - -#define f1_typemask 0xfc000000 /* Mask for char type field */ -#define f1_typeshift 26 /* Shift for the type field */ -#define f1_rangemask 0x0000ffff /* Mask for a range offset */ -#define f1_casemask 0x0000ffff /* Mask for a case offset */ -#define f1_caseneg 0xffff8000 /* Bits for negation */ - -/* The data consists of a vector of structures of type cnode. The two unsigned -32-bit integers are used as follows: - -(f0) (1) The most significant byte holds the script number. The numbers are - defined by the enum in ucp.h. - - (2) The 0x00800000 bit is set if this entry defines a range of characters. - It is not set if this entry defines a single character - - (3) The 0x00600000 bits are spare. - - (4) The 0x001fffff bits contain the code point. No Unicode code point will - ever be greater than 0x0010ffff, so this should be OK for ever. - -(f1) (1) The 0xfc000000 bits contain the character type number. The numbers are - defined by an enum in ucp.h. - - (2) The 0x03ff0000 bits are spare. - - (3) The 0x0000ffff bits contain EITHER the unsigned offset to the top of - range if this entry defines a range, OR the *signed* offset to the - character's "other case" partner if this entry defines a single - character. There is no partner if the value is zero. - -------------------------------------------------------------------------------- -| script (8) |.|.|.| codepoint (21) || type (6) |.|.| spare (8) | offset (16) | -------------------------------------------------------------------------------- - | | | | | - | | |-> spare | |-> spare - | | | - | |-> spare |-> spare - | - |-> range flag - -The upper/lower casing information is set only for characters that come in -pairs. The non-one-to-one mappings in the Unicode data are ignored. - -When searching the data, proceed as follows: - -(1) Set up for a binary chop search. - -(2) If the top is not greater than the bottom, the character is not in the - table. Its type must therefore be "Cn" ("Undefined"). - -(3) Find the middle vector element. - -(4) Extract the code point and compare. If equal, we are done. - -(5) If the test character is smaller, set the top to the current point, and - goto (2). - -(6) If the current entry defines a range, compute the last character by adding - the offset, and see if the test character is within the range. If it is, - we are done. - -(7) Otherwise, set the bottom to one element past the current point and goto - (2). -*/ - -#endif /* _UCPINTERNAL_H */ - -/* End of ucpinternal.h */ diff --git a/pcre/ucptable.h b/pcre/ucptable.h deleted file mode 100644 --- a/pcre/ucptable.h +++ /dev/null @@ -1,3088 +0,0 @@ -/* This source module is automatically generated from the Unicode -property table. See ucpinternal.h for a description of the layout. -This version was made from the Unicode 5.0.0 tables. */ - -static const cnode ucp_table[] = { - { 0x09800000, 0x0000001f }, - { 0x09000020, 0x74000000 }, - { 0x09800021, 0x54000002 }, - { 0x09000024, 0x5c000000 }, - { 0x09800025, 0x54000002 }, - { 0x09000028, 0x58000000 }, - { 0x09000029, 0x48000000 }, - { 0x0900002a, 0x54000000 }, - { 0x0900002b, 0x64000000 }, - { 0x0900002c, 0x54000000 }, - { 0x0900002d, 0x44000000 }, - { 0x0980002e, 0x54000001 }, - { 0x09800030, 0x34000009 }, - { 0x0980003a, 0x54000001 }, - { 0x0980003c, 0x64000002 }, - { 0x0980003f, 0x54000001 }, - { 0x21000041, 0x24000020 }, - { 0x21000042, 0x24000020 }, - { 0x21000043, 0x24000020 }, - { 0x21000044, 0x24000020 }, - { 0x21000045, 0x24000020 }, - { 0x21000046, 0x24000020 }, - { 0x21000047, 0x24000020 }, - { 0x21000048, 0x24000020 }, - { 0x21000049, 0x24000020 }, - { 0x2100004a, 0x24000020 }, - { 0x2100004b, 0x24000020 }, - { 0x2100004c, 0x24000020 }, - { 0x2100004d, 0x24000020 }, - { 0x2100004e, 0x24000020 }, - { 0x2100004f, 0x24000020 }, - { 0x21000050, 0x24000020 }, - { 0x21000051, 0x24000020 }, - { 0x21000052, 0x24000020 }, - { 0x21000053, 0x24000020 }, - { 0x21000054, 0x24000020 }, - { 0x21000055, 0x24000020 }, - { 0x21000056, 0x24000020 }, - { 0x21000057, 0x24000020 }, - { 0x21000058, 0x24000020 }, - { 0x21000059, 0x24000020 }, - { 0x2100005a, 0x24000020 }, - { 0x0900005b, 0x58000000 }, - { 0x0900005c, 0x54000000 }, - { 0x0900005d, 0x48000000 }, - { 0x0900005e, 0x60000000 }, - { 0x0900005f, 0x40000000 }, - { 0x09000060, 0x60000000 }, - { 0x21000061, 0x1400ffe0 }, - { 0x21000062, 0x1400ffe0 }, - { 0x21000063, 0x1400ffe0 }, - { 0x21000064, 0x1400ffe0 }, - { 0x21000065, 0x1400ffe0 }, - { 0x21000066, 0x1400ffe0 }, - { 0x21000067, 0x1400ffe0 }, - { 0x21000068, 0x1400ffe0 }, - { 0x21000069, 0x1400ffe0 }, - { 0x2100006a, 0x1400ffe0 }, - { 0x2100006b, 0x1400ffe0 }, - { 0x2100006c, 0x1400ffe0 }, - { 0x2100006d, 0x1400ffe0 }, - { 0x2100006e, 0x1400ffe0 }, - { 0x2100006f, 0x1400ffe0 }, - { 0x21000070, 0x1400ffe0 }, - { 0x21000071, 0x1400ffe0 }, - { 0x21000072, 0x1400ffe0 }, - { 0x21000073, 0x1400ffe0 }, - { 0x21000074, 0x1400ffe0 }, - { 0x21000075, 0x1400ffe0 }, - { 0x21000076, 0x1400ffe0 }, - { 0x21000077, 0x1400ffe0 }, - { 0x21000078, 0x1400ffe0 }, - { 0x21000079, 0x1400ffe0 }, - { 0x2100007a, 0x1400ffe0 }, - { 0x0900007b, 0x58000000 }, - { 0x0900007c, 0x64000000 }, - { 0x0900007d, 0x48000000 }, - { 0x0900007e, 0x64000000 }, - { 0x0980007f, 0x00000020 }, - { 0x090000a0, 0x74000000 }, - { 0x090000a1, 0x54000000 }, - { 0x098000a2, 0x5c000003 }, - { 0x098000a6, 0x68000001 }, - { 0x090000a8, 0x60000000 }, - { 0x090000a9, 0x68000000 }, - { 0x210000aa, 0x14000000 }, - { 0x090000ab, 0x50000000 }, - { 0x090000ac, 0x64000000 }, - { 0x090000ad, 0x04000000 }, - { 0x090000ae, 0x68000000 }, - { 0x090000af, 0x60000000 }, - { 0x090000b0, 0x68000000 }, - { 0x090000b1, 0x64000000 }, - { 0x098000b2, 0x3c000001 }, - { 0x090000b4, 0x60000000 }, - { 0x090000b5, 0x140002e7 }, - { 0x090000b6, 0x68000000 }, - { 0x090000b7, 0x54000000 }, - { 0x090000b8, 0x60000000 }, - { 0x090000b9, 0x3c000000 }, - { 0x210000ba, 0x14000000 }, - { 0x090000bb, 0x4c000000 }, - { 0x098000bc, 0x3c000002 }, - { 0x090000bf, 0x54000000 }, - { 0x210000c0, 0x24000020 }, - { 0x210000c1, 0x24000020 }, - { 0x210000c2, 0x24000020 }, - { 0x210000c3, 0x24000020 }, - { 0x210000c4, 0x24000020 }, - { 0x210000c5, 0x24000020 }, - { 0x210000c6, 0x24000020 }, - { 0x210000c7, 0x24000020 }, - { 0x210000c8, 0x24000020 }, - { 0x210000c9, 0x24000020 }, - { 0x210000ca, 0x24000020 }, - { 0x210000cb, 0x24000020 }, - { 0x210000cc, 0x24000020 }, - { 0x210000cd, 0x24000020 }, - { 0x210000ce, 0x24000020 }, - { 0x210000cf, 0x24000020 }, - { 0x210000d0, 0x24000020 }, - { 0x210000d1, 0x24000020 }, - { 0x210000d2, 0x24000020 }, - { 0x210000d3, 0x24000020 }, - { 0x210000d4, 0x24000020 }, - { 0x210000d5, 0x24000020 }, - { 0x210000d6, 0x24000020 }, - { 0x090000d7, 0x64000000 }, - { 0x210000d8, 0x24000020 }, - { 0x210000d9, 0x24000020 }, - { 0x210000da, 0x24000020 }, - { 0x210000db, 0x24000020 }, - { 0x210000dc, 0x24000020 }, - { 0x210000dd, 0x24000020 }, - { 0x210000de, 0x24000020 }, - { 0x210000df, 0x14000000 }, - { 0x210000e0, 0x1400ffe0 }, - { 0x210000e1, 0x1400ffe0 }, - { 0x210000e2, 0x1400ffe0 }, - { 0x210000e3, 0x1400ffe0 }, - { 0x210000e4, 0x1400ffe0 }, - { 0x210000e5, 0x1400ffe0 }, - { 0x210000e6, 0x1400ffe0 }, - { 0x210000e7, 0x1400ffe0 }, - { 0x210000e8, 0x1400ffe0 }, - { 0x210000e9, 0x1400ffe0 }, - { 0x210000ea, 0x1400ffe0 }, - { 0x210000eb, 0x1400ffe0 }, - { 0x210000ec, 0x1400ffe0 }, - { 0x210000ed, 0x1400ffe0 }, - { 0x210000ee, 0x1400ffe0 }, - { 0x210000ef, 0x1400ffe0 }, - { 0x210000f0, 0x1400ffe0 }, - { 0x210000f1, 0x1400ffe0 }, - { 0x210000f2, 0x1400ffe0 }, - { 0x210000f3, 0x1400ffe0 }, - { 0x210000f4, 0x1400ffe0 }, - { 0x210000f5, 0x1400ffe0 }, - { 0x210000f6, 0x1400ffe0 }, - { 0x090000f7, 0x64000000 }, - { 0x210000f8, 0x1400ffe0 }, - { 0x210000f9, 0x1400ffe0 }, - { 0x210000fa, 0x1400ffe0 }, - { 0x210000fb, 0x1400ffe0 }, - { 0x210000fc, 0x1400ffe0 }, - { 0x210000fd, 0x1400ffe0 }, - { 0x210000fe, 0x1400ffe0 }, - { 0x210000ff, 0x14000079 }, - { 0x21000100, 0x24000001 }, - { 0x21000101, 0x1400ffff }, - { 0x21000102, 0x24000001 }, - { 0x21000103, 0x1400ffff }, - { 0x21000104, 0x24000001 }, - { 0x21000105, 0x1400ffff }, - { 0x21000106, 0x24000001 }, - { 0x21000107, 0x1400ffff }, - { 0x21000108, 0x24000001 }, - { 0x21000109, 0x1400ffff }, - { 0x2100010a, 0x24000001 }, - { 0x2100010b, 0x1400ffff }, - { 0x2100010c, 0x24000001 }, - { 0x2100010d, 0x1400ffff }, - { 0x2100010e, 0x24000001 }, - { 0x2100010f, 0x1400ffff }, - { 0x21000110, 0x24000001 }, - { 0x21000111, 0x1400ffff }, - { 0x21000112, 0x24000001 }, - { 0x21000113, 0x1400ffff }, - { 0x21000114, 0x24000001 }, - { 0x21000115, 0x1400ffff }, - { 0x21000116, 0x24000001 }, - { 0x21000117, 0x1400ffff }, - { 0x21000118, 0x24000001 }, - { 0x21000119, 0x1400ffff }, - { 0x2100011a, 0x24000001 }, - { 0x2100011b, 0x1400ffff }, - { 0x2100011c, 0x24000001 }, - { 0x2100011d, 0x1400ffff }, - { 0x2100011e, 0x24000001 }, - { 0x2100011f, 0x1400ffff }, - { 0x21000120, 0x24000001 }, - { 0x21000121, 0x1400ffff }, - { 0x21000122, 0x24000001 }, - { 0x21000123, 0x1400ffff }, - { 0x21000124, 0x24000001 }, - { 0x21000125, 0x1400ffff }, - { 0x21000126, 0x24000001 }, - { 0x21000127, 0x1400ffff }, - { 0x21000128, 0x24000001 }, - { 0x21000129, 0x1400ffff }, - { 0x2100012a, 0x24000001 }, - { 0x2100012b, 0x1400ffff }, - { 0x2100012c, 0x24000001 }, - { 0x2100012d, 0x1400ffff }, - { 0x2100012e, 0x24000001 }, - { 0x2100012f, 0x1400ffff }, - { 0x21000130, 0x2400ff39 }, - { 0x21000131, 0x1400ff18 }, - { 0x21000132, 0x24000001 }, - { 0x21000133, 0x1400ffff }, - { 0x21000134, 0x24000001 }, - { 0x21000135, 0x1400ffff }, - { 0x21000136, 0x24000001 }, - { 0x21000137, 0x1400ffff }, - { 0x21000138, 0x14000000 }, - { 0x21000139, 0x24000001 }, - { 0x2100013a, 0x1400ffff }, - { 0x2100013b, 0x24000001 }, - { 0x2100013c, 0x1400ffff }, - { 0x2100013d, 0x24000001 }, - { 0x2100013e, 0x1400ffff }, - { 0x2100013f, 0x24000001 }, - { 0x21000140, 0x1400ffff }, - { 0x21000141, 0x24000001 }, - { 0x21000142, 0x1400ffff }, - { 0x21000143, 0x24000001 }, - { 0x21000144, 0x1400ffff }, - { 0x21000145, 0x24000001 }, - { 0x21000146, 0x1400ffff }, - { 0x21000147, 0x24000001 }, - { 0x21000148, 0x1400ffff }, - { 0x21000149, 0x14000000 }, - { 0x2100014a, 0x24000001 }, - { 0x2100014b, 0x1400ffff }, - { 0x2100014c, 0x24000001 }, - { 0x2100014d, 0x1400ffff }, - { 0x2100014e, 0x24000001 }, - { 0x2100014f, 0x1400ffff }, - { 0x21000150, 0x24000001 }, - { 0x21000151, 0x1400ffff }, - { 0x21000152, 0x24000001 }, - { 0x21000153, 0x1400ffff }, - { 0x21000154, 0x24000001 }, - { 0x21000155, 0x1400ffff }, - { 0x21000156, 0x24000001 }, - { 0x21000157, 0x1400ffff }, - { 0x21000158, 0x24000001 }, - { 0x21000159, 0x1400ffff }, - { 0x2100015a, 0x24000001 }, - { 0x2100015b, 0x1400ffff }, - { 0x2100015c, 0x24000001 }, - { 0x2100015d, 0x1400ffff }, - { 0x2100015e, 0x24000001 }, - { 0x2100015f, 0x1400ffff }, - { 0x21000160, 0x24000001 }, - { 0x21000161, 0x1400ffff }, - { 0x21000162, 0x24000001 }, - { 0x21000163, 0x1400ffff }, - { 0x21000164, 0x24000001 }, - { 0x21000165, 0x1400ffff }, - { 0x21000166, 0x24000001 }, - { 0x21000167, 0x1400ffff }, - { 0x21000168, 0x24000001 }, - { 0x21000169, 0x1400ffff }, - { 0x2100016a, 0x24000001 }, - { 0x2100016b, 0x1400ffff }, - { 0x2100016c, 0x24000001 }, - { 0x2100016d, 0x1400ffff }, - { 0x2100016e, 0x24000001 }, - { 0x2100016f, 0x1400ffff }, - { 0x21000170, 0x24000001 }, - { 0x21000171, 0x1400ffff }, - { 0x21000172, 0x24000001 }, - { 0x21000173, 0x1400ffff }, - { 0x21000174, 0x24000001 }, - { 0x21000175, 0x1400ffff }, - { 0x21000176, 0x24000001 }, - { 0x21000177, 0x1400ffff }, - { 0x21000178, 0x2400ff87 }, - { 0x21000179, 0x24000001 }, - { 0x2100017a, 0x1400ffff }, - { 0x2100017b, 0x24000001 }, - { 0x2100017c, 0x1400ffff }, - { 0x2100017d, 0x24000001 }, - { 0x2100017e, 0x1400ffff }, - { 0x2100017f, 0x1400fed4 }, - { 0x21000180, 0x140000c3 }, - { 0x21000181, 0x240000d2 }, - { 0x21000182, 0x24000001 }, - { 0x21000183, 0x1400ffff }, - { 0x21000184, 0x24000001 }, - { 0x21000185, 0x1400ffff }, - { 0x21000186, 0x240000ce }, - { 0x21000187, 0x24000001 }, - { 0x21000188, 0x1400ffff }, - { 0x21000189, 0x240000cd }, - { 0x2100018a, 0x240000cd }, - { 0x2100018b, 0x24000001 }, - { 0x2100018c, 0x1400ffff }, - { 0x2100018d, 0x14000000 }, - { 0x2100018e, 0x2400004f }, - { 0x2100018f, 0x240000ca }, - { 0x21000190, 0x240000cb }, - { 0x21000191, 0x24000001 }, - { 0x21000192, 0x1400ffff }, - { 0x21000193, 0x240000cd }, - { 0x21000194, 0x240000cf }, - { 0x21000195, 0x14000061 }, - { 0x21000196, 0x240000d3 }, - { 0x21000197, 0x240000d1 }, - { 0x21000198, 0x24000001 }, - { 0x21000199, 0x1400ffff }, - { 0x2100019a, 0x140000a3 }, - { 0x2100019b, 0x14000000 }, - { 0x2100019c, 0x240000d3 }, - { 0x2100019d, 0x240000d5 }, - { 0x2100019e, 0x14000082 }, - { 0x2100019f, 0x240000d6 }, - { 0x210001a0, 0x24000001 }, - { 0x210001a1, 0x1400ffff }, - { 0x210001a2, 0x24000001 }, - { 0x210001a3, 0x1400ffff }, - { 0x210001a4, 0x24000001 }, - { 0x210001a5, 0x1400ffff }, - { 0x210001a6, 0x240000da }, - { 0x210001a7, 0x24000001 }, - { 0x210001a8, 0x1400ffff }, - { 0x210001a9, 0x240000da }, - { 0x218001aa, 0x14000001 }, - { 0x210001ac, 0x24000001 }, - { 0x210001ad, 0x1400ffff }, - { 0x210001ae, 0x240000da }, - { 0x210001af, 0x24000001 }, - { 0x210001b0, 0x1400ffff }, - { 0x210001b1, 0x240000d9 }, - { 0x210001b2, 0x240000d9 }, - { 0x210001b3, 0x24000001 }, - { 0x210001b4, 0x1400ffff }, - { 0x210001b5, 0x24000001 }, - { 0x210001b6, 0x1400ffff }, - { 0x210001b7, 0x240000db }, - { 0x210001b8, 0x24000001 }, - { 0x210001b9, 0x1400ffff }, - { 0x210001ba, 0x14000000 }, - { 0x210001bb, 0x1c000000 }, - { 0x210001bc, 0x24000001 }, - { 0x210001bd, 0x1400ffff }, - { 0x210001be, 0x14000000 }, - { 0x210001bf, 0x14000038 }, - { 0x218001c0, 0x1c000003 }, - { 0x210001c4, 0x24000002 }, - { 0x210001c5, 0x2000ffff }, - { 0x210001c6, 0x1400fffe }, - { 0x210001c7, 0x24000002 }, - { 0x210001c8, 0x2000ffff }, - { 0x210001c9, 0x1400fffe }, - { 0x210001ca, 0x24000002 }, - { 0x210001cb, 0x2000ffff }, - { 0x210001cc, 0x1400fffe }, - { 0x210001cd, 0x24000001 }, - { 0x210001ce, 0x1400ffff }, - { 0x210001cf, 0x24000001 }, - { 0x210001d0, 0x1400ffff }, - { 0x210001d1, 0x24000001 }, - { 0x210001d2, 0x1400ffff }, - { 0x210001d3, 0x24000001 }, - { 0x210001d4, 0x1400ffff }, - { 0x210001d5, 0x24000001 }, - { 0x210001d6, 0x1400ffff }, - { 0x210001d7, 0x24000001 }, - { 0x210001d8, 0x1400ffff }, - { 0x210001d9, 0x24000001 }, - { 0x210001da, 0x1400ffff }, - { 0x210001db, 0x24000001 }, - { 0x210001dc, 0x1400ffff }, - { 0x210001dd, 0x1400ffb1 }, - { 0x210001de, 0x24000001 }, - { 0x210001df, 0x1400ffff }, - { 0x210001e0, 0x24000001 }, - { 0x210001e1, 0x1400ffff }, - { 0x210001e2, 0x24000001 }, - { 0x210001e3, 0x1400ffff }, - { 0x210001e4, 0x24000001 }, - { 0x210001e5, 0x1400ffff }, - { 0x210001e6, 0x24000001 }, - { 0x210001e7, 0x1400ffff }, - { 0x210001e8, 0x24000001 }, - { 0x210001e9, 0x1400ffff }, - { 0x210001ea, 0x24000001 }, - { 0x210001eb, 0x1400ffff }, - { 0x210001ec, 0x24000001 }, - { 0x210001ed, 0x1400ffff }, - { 0x210001ee, 0x24000001 }, - { 0x210001ef, 0x1400ffff }, - { 0x210001f0, 0x14000000 }, - { 0x210001f1, 0x24000002 }, - { 0x210001f2, 0x2000ffff }, - { 0x210001f3, 0x1400fffe }, - { 0x210001f4, 0x24000001 }, - { 0x210001f5, 0x1400ffff }, - { 0x210001f6, 0x2400ff9f }, - { 0x210001f7, 0x2400ffc8 }, - { 0x210001f8, 0x24000001 }, - { 0x210001f9, 0x1400ffff }, - { 0x210001fa, 0x24000001 }, - { 0x210001fb, 0x1400ffff }, - { 0x210001fc, 0x24000001 }, - { 0x210001fd, 0x1400ffff }, - { 0x210001fe, 0x24000001 }, - { 0x210001ff, 0x1400ffff }, - { 0x21000200, 0x24000001 }, - { 0x21000201, 0x1400ffff }, - { 0x21000202, 0x24000001 }, - { 0x21000203, 0x1400ffff }, - { 0x21000204, 0x24000001 }, - { 0x21000205, 0x1400ffff }, - { 0x21000206, 0x24000001 }, - { 0x21000207, 0x1400ffff }, - { 0x21000208, 0x24000001 }, - { 0x21000209, 0x1400ffff }, - { 0x2100020a, 0x24000001 }, - { 0x2100020b, 0x1400ffff }, - { 0x2100020c, 0x24000001 }, - { 0x2100020d, 0x1400ffff }, - { 0x2100020e, 0x24000001 }, - { 0x2100020f, 0x1400ffff }, - { 0x21000210, 0x24000001 }, - { 0x21000211, 0x1400ffff }, - { 0x21000212, 0x24000001 }, - { 0x21000213, 0x1400ffff }, - { 0x21000214, 0x24000001 }, - { 0x21000215, 0x1400ffff }, - { 0x21000216, 0x24000001 }, - { 0x21000217, 0x1400ffff }, - { 0x21000218, 0x24000001 }, - { 0x21000219, 0x1400ffff }, - { 0x2100021a, 0x24000001 }, - { 0x2100021b, 0x1400ffff }, - { 0x2100021c, 0x24000001 }, - { 0x2100021d, 0x1400ffff }, - { 0x2100021e, 0x24000001 }, - { 0x2100021f, 0x1400ffff }, - { 0x21000220, 0x2400ff7e }, - { 0x21000221, 0x14000000 }, - { 0x21000222, 0x24000001 }, - { 0x21000223, 0x1400ffff }, - { 0x21000224, 0x24000001 }, - { 0x21000225, 0x1400ffff }, - { 0x21000226, 0x24000001 }, - { 0x21000227, 0x1400ffff }, - { 0x21000228, 0x24000001 }, - { 0x21000229, 0x1400ffff }, - { 0x2100022a, 0x24000001 }, - { 0x2100022b, 0x1400ffff }, - { 0x2100022c, 0x24000001 }, - { 0x2100022d, 0x1400ffff }, - { 0x2100022e, 0x24000001 }, - { 0x2100022f, 0x1400ffff }, - { 0x21000230, 0x24000001 }, - { 0x21000231, 0x1400ffff }, - { 0x21000232, 0x24000001 }, - { 0x21000233, 0x1400ffff }, - { 0x21800234, 0x14000005 }, - { 0x2100023a, 0x24002a2b }, - { 0x2100023b, 0x24000001 }, - { 0x2100023c, 0x1400ffff }, - { 0x2100023d, 0x2400ff5d }, - { 0x2100023e, 0x24002a28 }, - { 0x2180023f, 0x14000001 }, - { 0x21000241, 0x24000001 }, - { 0x21000242, 0x1400ffff }, - { 0x21000243, 0x2400ff3d }, - { 0x21000244, 0x24000045 }, - { 0x21000245, 0x24000047 }, - { 0x21000246, 0x24000001 }, - { 0x21000247, 0x1400ffff }, - { 0x21000248, 0x24000001 }, - { 0x21000249, 0x1400ffff }, - { 0x2100024a, 0x24000001 }, - { 0x2100024b, 0x1400ffff }, - { 0x2100024c, 0x24000001 }, - { 0x2100024d, 0x1400ffff }, - { 0x2100024e, 0x24000001 }, - { 0x2100024f, 0x1400ffff }, - { 0x21800250, 0x14000002 }, - { 0x21000253, 0x1400ff2e }, - { 0x21000254, 0x1400ff32 }, - { 0x21000255, 0x14000000 }, - { 0x21000256, 0x1400ff33 }, - { 0x21000257, 0x1400ff33 }, - { 0x21000258, 0x14000000 }, - { 0x21000259, 0x1400ff36 }, - { 0x2100025a, 0x14000000 }, - { 0x2100025b, 0x1400ff35 }, - { 0x2180025c, 0x14000003 }, - { 0x21000260, 0x1400ff33 }, - { 0x21800261, 0x14000001 }, - { 0x21000263, 0x1400ff31 }, - { 0x21800264, 0x14000003 }, - { 0x21000268, 0x1400ff2f }, - { 0x21000269, 0x1400ff2d }, - { 0x2100026a, 0x14000000 }, - { 0x2100026b, 0x140029f7 }, - { 0x2180026c, 0x14000002 }, - { 0x2100026f, 0x1400ff2d }, - { 0x21800270, 0x14000001 }, - { 0x21000272, 0x1400ff2b }, - { 0x21800273, 0x14000001 }, - { 0x21000275, 0x1400ff2a }, - { 0x21800276, 0x14000006 }, - { 0x2100027d, 0x140029e7 }, - { 0x2180027e, 0x14000001 }, - { 0x21000280, 0x1400ff26 }, - { 0x21800281, 0x14000001 }, - { 0x21000283, 0x1400ff26 }, - { 0x21800284, 0x14000003 }, - { 0x21000288, 0x1400ff26 }, - { 0x21000289, 0x1400ffbb }, - { 0x2100028a, 0x1400ff27 }, - { 0x2100028b, 0x1400ff27 }, - { 0x2100028c, 0x1400ffb9 }, - { 0x2180028d, 0x14000004 }, - { 0x21000292, 0x1400ff25 }, - { 0x21000293, 0x14000000 }, - { 0x21000294, 0x1c000000 }, - { 0x21800295, 0x1400001a }, - { 0x218002b0, 0x18000008 }, - { 0x098002b9, 0x18000008 }, - { 0x098002c2, 0x60000003 }, - { 0x098002c6, 0x1800000b }, - { 0x098002d2, 0x6000000d }, - { 0x218002e0, 0x18000004 }, - { 0x098002e5, 0x60000008 }, - { 0x090002ee, 0x18000000 }, - { 0x098002ef, 0x60000010 }, - { 0x1b800300, 0x30000044 }, - { 0x1b000345, 0x30000054 }, - { 0x1b800346, 0x30000029 }, - { 0x13800374, 0x60000001 }, - { 0x1300037a, 0x18000000 }, - { 0x1300037b, 0x14000082 }, - { 0x1300037c, 0x14000082 }, - { 0x1300037d, 0x14000082 }, - { 0x0900037e, 0x54000000 }, - { 0x13800384, 0x60000001 }, - { 0x13000386, 0x24000026 }, - { 0x09000387, 0x54000000 }, - { 0x13000388, 0x24000025 }, - { 0x13000389, 0x24000025 }, - { 0x1300038a, 0x24000025 }, - { 0x1300038c, 0x24000040 }, - { 0x1300038e, 0x2400003f }, - { 0x1300038f, 0x2400003f }, - { 0x13000390, 0x14000000 }, - { 0x13000391, 0x24000020 }, - { 0x13000392, 0x24000020 }, - { 0x13000393, 0x24000020 }, - { 0x13000394, 0x24000020 }, - { 0x13000395, 0x24000020 }, - { 0x13000396, 0x24000020 }, - { 0x13000397, 0x24000020 }, - { 0x13000398, 0x24000020 }, - { 0x13000399, 0x24000020 }, - { 0x1300039a, 0x24000020 }, - { 0x1300039b, 0x24000020 }, - { 0x1300039c, 0x24000020 }, - { 0x1300039d, 0x24000020 }, - { 0x1300039e, 0x24000020 }, - { 0x1300039f, 0x24000020 }, - { 0x130003a0, 0x24000020 }, - { 0x130003a1, 0x24000020 }, - { 0x130003a3, 0x24000020 }, - { 0x130003a4, 0x24000020 }, - { 0x130003a5, 0x24000020 }, - { 0x130003a6, 0x24000020 }, - { 0x130003a7, 0x24000020 }, - { 0x130003a8, 0x24000020 }, - { 0x130003a9, 0x24000020 }, - { 0x130003aa, 0x24000020 }, - { 0x130003ab, 0x24000020 }, - { 0x130003ac, 0x1400ffda }, - { 0x130003ad, 0x1400ffdb }, - { 0x130003ae, 0x1400ffdb }, - { 0x130003af, 0x1400ffdb }, - { 0x130003b0, 0x14000000 }, - { 0x130003b1, 0x1400ffe0 }, - { 0x130003b2, 0x1400ffe0 }, - { 0x130003b3, 0x1400ffe0 }, - { 0x130003b4, 0x1400ffe0 }, - { 0x130003b5, 0x1400ffe0 }, - { 0x130003b6, 0x1400ffe0 }, - { 0x130003b7, 0x1400ffe0 }, - { 0x130003b8, 0x1400ffe0 }, - { 0x130003b9, 0x1400ffe0 }, - { 0x130003ba, 0x1400ffe0 }, - { 0x130003bb, 0x1400ffe0 }, - { 0x130003bc, 0x1400ffe0 }, - { 0x130003bd, 0x1400ffe0 }, - { 0x130003be, 0x1400ffe0 }, - { 0x130003bf, 0x1400ffe0 }, - { 0x130003c0, 0x1400ffe0 }, - { 0x130003c1, 0x1400ffe0 }, - { 0x130003c2, 0x1400ffe1 }, - { 0x130003c3, 0x1400ffe0 }, - { 0x130003c4, 0x1400ffe0 }, - { 0x130003c5, 0x1400ffe0 }, - { 0x130003c6, 0x1400ffe0 }, - { 0x130003c7, 0x1400ffe0 }, - { 0x130003c8, 0x1400ffe0 }, - { 0x130003c9, 0x1400ffe0 }, - { 0x130003ca, 0x1400ffe0 }, - { 0x130003cb, 0x1400ffe0 }, - { 0x130003cc, 0x1400ffc0 }, - { 0x130003cd, 0x1400ffc1 }, - { 0x130003ce, 0x1400ffc1 }, - { 0x130003d0, 0x1400ffc2 }, - { 0x130003d1, 0x1400ffc7 }, - { 0x138003d2, 0x24000002 }, - { 0x130003d5, 0x1400ffd1 }, - { 0x130003d6, 0x1400ffca }, - { 0x130003d7, 0x14000000 }, - { 0x130003d8, 0x24000001 }, - { 0x130003d9, 0x1400ffff }, - { 0x130003da, 0x24000001 }, - { 0x130003db, 0x1400ffff }, - { 0x130003dc, 0x24000001 }, - { 0x130003dd, 0x1400ffff }, - { 0x130003de, 0x24000001 }, - { 0x130003df, 0x1400ffff }, - { 0x130003e0, 0x24000001 }, - { 0x130003e1, 0x1400ffff }, - { 0x0a0003e2, 0x24000001 }, - { 0x0a0003e3, 0x1400ffff }, - { 0x0a0003e4, 0x24000001 }, - { 0x0a0003e5, 0x1400ffff }, - { 0x0a0003e6, 0x24000001 }, - { 0x0a0003e7, 0x1400ffff }, - { 0x0a0003e8, 0x24000001 }, - { 0x0a0003e9, 0x1400ffff }, - { 0x0a0003ea, 0x24000001 }, - { 0x0a0003eb, 0x1400ffff }, - { 0x0a0003ec, 0x24000001 }, - { 0x0a0003ed, 0x1400ffff }, - { 0x0a0003ee, 0x24000001 }, - { 0x0a0003ef, 0x1400ffff }, - { 0x130003f0, 0x1400ffaa }, - { 0x130003f1, 0x1400ffb0 }, - { 0x130003f2, 0x14000007 }, - { 0x130003f3, 0x14000000 }, - { 0x130003f4, 0x2400ffc4 }, - { 0x130003f5, 0x1400ffa0 }, - { 0x130003f6, 0x64000000 }, - { 0x130003f7, 0x24000001 }, - { 0x130003f8, 0x1400ffff }, - { 0x130003f9, 0x2400fff9 }, - { 0x130003fa, 0x24000001 }, - { 0x130003fb, 0x1400ffff }, - { 0x130003fc, 0x14000000 }, - { 0x130003fd, 0x2400ff7e }, - { 0x130003fe, 0x2400ff7e }, - { 0x130003ff, 0x2400ff7e }, - { 0x0c000400, 0x24000050 }, - { 0x0c000401, 0x24000050 }, - { 0x0c000402, 0x24000050 }, - { 0x0c000403, 0x24000050 }, - { 0x0c000404, 0x24000050 }, - { 0x0c000405, 0x24000050 }, - { 0x0c000406, 0x24000050 }, - { 0x0c000407, 0x24000050 }, - { 0x0c000408, 0x24000050 }, - { 0x0c000409, 0x24000050 }, - { 0x0c00040a, 0x24000050 }, - { 0x0c00040b, 0x24000050 }, - { 0x0c00040c, 0x24000050 }, - { 0x0c00040d, 0x24000050 }, - { 0x0c00040e, 0x24000050 }, - { 0x0c00040f, 0x24000050 }, - { 0x0c000410, 0x24000020 }, - { 0x0c000411, 0x24000020 }, - { 0x0c000412, 0x24000020 }, - { 0x0c000413, 0x24000020 }, - { 0x0c000414, 0x24000020 }, - { 0x0c000415, 0x24000020 }, - { 0x0c000416, 0x24000020 }, - { 0x0c000417, 0x24000020 }, - { 0x0c000418, 0x24000020 }, - { 0x0c000419, 0x24000020 }, - { 0x0c00041a, 0x24000020 }, - { 0x0c00041b, 0x24000020 }, - { 0x0c00041c, 0x24000020 }, - { 0x0c00041d, 0x24000020 }, - { 0x0c00041e, 0x24000020 }, - { 0x0c00041f, 0x24000020 }, - { 0x0c000420, 0x24000020 }, - { 0x0c000421, 0x24000020 }, - { 0x0c000422, 0x24000020 }, - { 0x0c000423, 0x24000020 }, - { 0x0c000424, 0x24000020 }, - { 0x0c000425, 0x24000020 }, - { 0x0c000426, 0x24000020 }, - { 0x0c000427, 0x24000020 }, - { 0x0c000428, 0x24000020 }, - { 0x0c000429, 0x24000020 }, - { 0x0c00042a, 0x24000020 }, - { 0x0c00042b, 0x24000020 }, - { 0x0c00042c, 0x24000020 }, - { 0x0c00042d, 0x24000020 }, - { 0x0c00042e, 0x24000020 }, - { 0x0c00042f, 0x24000020 }, - { 0x0c000430, 0x1400ffe0 }, - { 0x0c000431, 0x1400ffe0 }, - { 0x0c000432, 0x1400ffe0 }, - { 0x0c000433, 0x1400ffe0 }, - { 0x0c000434, 0x1400ffe0 }, - { 0x0c000435, 0x1400ffe0 }, - { 0x0c000436, 0x1400ffe0 }, - { 0x0c000437, 0x1400ffe0 }, - { 0x0c000438, 0x1400ffe0 }, - { 0x0c000439, 0x1400ffe0 }, - { 0x0c00043a, 0x1400ffe0 }, - { 0x0c00043b, 0x1400ffe0 }, - { 0x0c00043c, 0x1400ffe0 }, - { 0x0c00043d, 0x1400ffe0 }, - { 0x0c00043e, 0x1400ffe0 }, - { 0x0c00043f, 0x1400ffe0 }, - { 0x0c000440, 0x1400ffe0 }, - { 0x0c000441, 0x1400ffe0 }, - { 0x0c000442, 0x1400ffe0 }, - { 0x0c000443, 0x1400ffe0 }, - { 0x0c000444, 0x1400ffe0 }, - { 0x0c000445, 0x1400ffe0 }, - { 0x0c000446, 0x1400ffe0 }, - { 0x0c000447, 0x1400ffe0 }, - { 0x0c000448, 0x1400ffe0 }, - { 0x0c000449, 0x1400ffe0 }, - { 0x0c00044a, 0x1400ffe0 }, - { 0x0c00044b, 0x1400ffe0 }, - { 0x0c00044c, 0x1400ffe0 }, - { 0x0c00044d, 0x1400ffe0 }, - { 0x0c00044e, 0x1400ffe0 }, - { 0x0c00044f, 0x1400ffe0 }, - { 0x0c000450, 0x1400ffb0 }, - { 0x0c000451, 0x1400ffb0 }, - { 0x0c000452, 0x1400ffb0 }, - { 0x0c000453, 0x1400ffb0 }, - { 0x0c000454, 0x1400ffb0 }, - { 0x0c000455, 0x1400ffb0 }, - { 0x0c000456, 0x1400ffb0 }, - { 0x0c000457, 0x1400ffb0 }, - { 0x0c000458, 0x1400ffb0 }, - { 0x0c000459, 0x1400ffb0 }, - { 0x0c00045a, 0x1400ffb0 }, - { 0x0c00045b, 0x1400ffb0 }, - { 0x0c00045c, 0x1400ffb0 }, - { 0x0c00045d, 0x1400ffb0 }, - { 0x0c00045e, 0x1400ffb0 }, - { 0x0c00045f, 0x1400ffb0 }, - { 0x0c000460, 0x24000001 }, - { 0x0c000461, 0x1400ffff }, - { 0x0c000462, 0x24000001 }, - { 0x0c000463, 0x1400ffff }, - { 0x0c000464, 0x24000001 }, - { 0x0c000465, 0x1400ffff }, - { 0x0c000466, 0x24000001 }, - { 0x0c000467, 0x1400ffff }, - { 0x0c000468, 0x24000001 }, - { 0x0c000469, 0x1400ffff }, - { 0x0c00046a, 0x24000001 }, - { 0x0c00046b, 0x1400ffff }, - { 0x0c00046c, 0x24000001 }, - { 0x0c00046d, 0x1400ffff }, - { 0x0c00046e, 0x24000001 }, - { 0x0c00046f, 0x1400ffff }, - { 0x0c000470, 0x24000001 }, - { 0x0c000471, 0x1400ffff }, - { 0x0c000472, 0x24000001 }, - { 0x0c000473, 0x1400ffff }, - { 0x0c000474, 0x24000001 }, - { 0x0c000475, 0x1400ffff }, - { 0x0c000476, 0x24000001 }, - { 0x0c000477, 0x1400ffff }, - { 0x0c000478, 0x24000001 }, - { 0x0c000479, 0x1400ffff }, - { 0x0c00047a, 0x24000001 }, - { 0x0c00047b, 0x1400ffff }, - { 0x0c00047c, 0x24000001 }, - { 0x0c00047d, 0x1400ffff }, - { 0x0c00047e, 0x24000001 }, - { 0x0c00047f, 0x1400ffff }, - { 0x0c000480, 0x24000001 }, - { 0x0c000481, 0x1400ffff }, - { 0x0c000482, 0x68000000 }, - { 0x0c800483, 0x30000003 }, - { 0x0c800488, 0x2c000001 }, - { 0x0c00048a, 0x24000001 }, - { 0x0c00048b, 0x1400ffff }, - { 0x0c00048c, 0x24000001 }, - { 0x0c00048d, 0x1400ffff }, - { 0x0c00048e, 0x24000001 }, - { 0x0c00048f, 0x1400ffff }, - { 0x0c000490, 0x24000001 }, - { 0x0c000491, 0x1400ffff }, - { 0x0c000492, 0x24000001 }, - { 0x0c000493, 0x1400ffff }, - { 0x0c000494, 0x24000001 }, - { 0x0c000495, 0x1400ffff }, - { 0x0c000496, 0x24000001 }, - { 0x0c000497, 0x1400ffff }, - { 0x0c000498, 0x24000001 }, - { 0x0c000499, 0x1400ffff }, - { 0x0c00049a, 0x24000001 }, - { 0x0c00049b, 0x1400ffff }, - { 0x0c00049c, 0x24000001 }, - { 0x0c00049d, 0x1400ffff }, - { 0x0c00049e, 0x24000001 }, - { 0x0c00049f, 0x1400ffff }, - { 0x0c0004a0, 0x24000001 }, - { 0x0c0004a1, 0x1400ffff }, - { 0x0c0004a2, 0x24000001 }, - { 0x0c0004a3, 0x1400ffff }, - { 0x0c0004a4, 0x24000001 }, - { 0x0c0004a5, 0x1400ffff }, - { 0x0c0004a6, 0x24000001 }, - { 0x0c0004a7, 0x1400ffff }, - { 0x0c0004a8, 0x24000001 }, - { 0x0c0004a9, 0x1400ffff }, - { 0x0c0004aa, 0x24000001 }, - { 0x0c0004ab, 0x1400ffff }, - { 0x0c0004ac, 0x24000001 }, - { 0x0c0004ad, 0x1400ffff }, - { 0x0c0004ae, 0x24000001 }, - { 0x0c0004af, 0x1400ffff }, - { 0x0c0004b0, 0x24000001 }, - { 0x0c0004b1, 0x1400ffff }, - { 0x0c0004b2, 0x24000001 }, - { 0x0c0004b3, 0x1400ffff }, - { 0x0c0004b4, 0x24000001 }, - { 0x0c0004b5, 0x1400ffff }, - { 0x0c0004b6, 0x24000001 }, - { 0x0c0004b7, 0x1400ffff }, - { 0x0c0004b8, 0x24000001 }, - { 0x0c0004b9, 0x1400ffff }, - { 0x0c0004ba, 0x24000001 }, - { 0x0c0004bb, 0x1400ffff }, - { 0x0c0004bc, 0x24000001 }, - { 0x0c0004bd, 0x1400ffff }, - { 0x0c0004be, 0x24000001 }, - { 0x0c0004bf, 0x1400ffff }, - { 0x0c0004c0, 0x2400000f }, - { 0x0c0004c1, 0x24000001 }, - { 0x0c0004c2, 0x1400ffff }, - { 0x0c0004c3, 0x24000001 }, - { 0x0c0004c4, 0x1400ffff }, - { 0x0c0004c5, 0x24000001 }, - { 0x0c0004c6, 0x1400ffff }, - { 0x0c0004c7, 0x24000001 }, - { 0x0c0004c8, 0x1400ffff }, - { 0x0c0004c9, 0x24000001 }, - { 0x0c0004ca, 0x1400ffff }, - { 0x0c0004cb, 0x24000001 }, - { 0x0c0004cc, 0x1400ffff }, - { 0x0c0004cd, 0x24000001 }, - { 0x0c0004ce, 0x1400ffff }, - { 0x0c0004cf, 0x1400fff1 }, - { 0x0c0004d0, 0x24000001 }, - { 0x0c0004d1, 0x1400ffff }, - { 0x0c0004d2, 0x24000001 }, - { 0x0c0004d3, 0x1400ffff }, - { 0x0c0004d4, 0x24000001 }, - { 0x0c0004d5, 0x1400ffff }, - { 0x0c0004d6, 0x24000001 }, - { 0x0c0004d7, 0x1400ffff }, - { 0x0c0004d8, 0x24000001 }, - { 0x0c0004d9, 0x1400ffff }, - { 0x0c0004da, 0x24000001 }, - { 0x0c0004db, 0x1400ffff }, - { 0x0c0004dc, 0x24000001 }, - { 0x0c0004dd, 0x1400ffff }, - { 0x0c0004de, 0x24000001 }, - { 0x0c0004df, 0x1400ffff }, - { 0x0c0004e0, 0x24000001 }, - { 0x0c0004e1, 0x1400ffff }, - { 0x0c0004e2, 0x24000001 }, - { 0x0c0004e3, 0x1400ffff }, - { 0x0c0004e4, 0x24000001 }, - { 0x0c0004e5, 0x1400ffff }, - { 0x0c0004e6, 0x24000001 }, - { 0x0c0004e7, 0x1400ffff }, - { 0x0c0004e8, 0x24000001 }, - { 0x0c0004e9, 0x1400ffff }, - { 0x0c0004ea, 0x24000001 }, - { 0x0c0004eb, 0x1400ffff }, - { 0x0c0004ec, 0x24000001 }, - { 0x0c0004ed, 0x1400ffff }, - { 0x0c0004ee, 0x24000001 }, - { 0x0c0004ef, 0x1400ffff }, - { 0x0c0004f0, 0x24000001 }, - { 0x0c0004f1, 0x1400ffff }, - { 0x0c0004f2, 0x24000001 }, - { 0x0c0004f3, 0x1400ffff }, - { 0x0c0004f4, 0x24000001 }, - { 0x0c0004f5, 0x1400ffff }, - { 0x0c0004f6, 0x24000001 }, - { 0x0c0004f7, 0x1400ffff }, - { 0x0c0004f8, 0x24000001 }, - { 0x0c0004f9, 0x1400ffff }, - { 0x0c0004fa, 0x24000001 }, - { 0x0c0004fb, 0x1400ffff }, - { 0x0c0004fc, 0x24000001 }, - { 0x0c0004fd, 0x1400ffff }, - { 0x0c0004fe, 0x24000001 }, - { 0x0c0004ff, 0x1400ffff }, - { 0x0c000500, 0x24000001 }, - { 0x0c000501, 0x1400ffff }, - { 0x0c000502, 0x24000001 }, - { 0x0c000503, 0x1400ffff }, - { 0x0c000504, 0x24000001 }, - { 0x0c000505, 0x1400ffff }, - { 0x0c000506, 0x24000001 }, - { 0x0c000507, 0x1400ffff }, - { 0x0c000508, 0x24000001 }, - { 0x0c000509, 0x1400ffff }, - { 0x0c00050a, 0x24000001 }, - { 0x0c00050b, 0x1400ffff }, - { 0x0c00050c, 0x24000001 }, - { 0x0c00050d, 0x1400ffff }, - { 0x0c00050e, 0x24000001 }, - { 0x0c00050f, 0x1400ffff }, - { 0x0c000510, 0x24000001 }, - { 0x0c000511, 0x1400ffff }, - { 0x0c000512, 0x24000001 }, - { 0x0c000513, 0x1400ffff }, - { 0x01000531, 0x24000030 }, - { 0x01000532, 0x24000030 }, - { 0x01000533, 0x24000030 }, - { 0x01000534, 0x24000030 }, - { 0x01000535, 0x24000030 }, - { 0x01000536, 0x24000030 }, - { 0x01000537, 0x24000030 }, - { 0x01000538, 0x24000030 }, - { 0x01000539, 0x24000030 }, - { 0x0100053a, 0x24000030 }, - { 0x0100053b, 0x24000030 }, - { 0x0100053c, 0x24000030 }, - { 0x0100053d, 0x24000030 }, - { 0x0100053e, 0x24000030 }, - { 0x0100053f, 0x24000030 }, - { 0x01000540, 0x24000030 }, - { 0x01000541, 0x24000030 }, - { 0x01000542, 0x24000030 }, - { 0x01000543, 0x24000030 }, - { 0x01000544, 0x24000030 }, - { 0x01000545, 0x24000030 }, - { 0x01000546, 0x24000030 }, - { 0x01000547, 0x24000030 }, - { 0x01000548, 0x24000030 }, - { 0x01000549, 0x24000030 }, - { 0x0100054a, 0x24000030 }, - { 0x0100054b, 0x24000030 }, - { 0x0100054c, 0x24000030 }, - { 0x0100054d, 0x24000030 }, - { 0x0100054e, 0x24000030 }, - { 0x0100054f, 0x24000030 }, - { 0x01000550, 0x24000030 }, - { 0x01000551, 0x24000030 }, - { 0x01000552, 0x24000030 }, - { 0x01000553, 0x24000030 }, - { 0x01000554, 0x24000030 }, - { 0x01000555, 0x24000030 }, - { 0x01000556, 0x24000030 }, - { 0x01000559, 0x18000000 }, - { 0x0180055a, 0x54000005 }, - { 0x01000561, 0x1400ffd0 }, - { 0x01000562, 0x1400ffd0 }, - { 0x01000563, 0x1400ffd0 }, - { 0x01000564, 0x1400ffd0 }, - { 0x01000565, 0x1400ffd0 }, - { 0x01000566, 0x1400ffd0 }, - { 0x01000567, 0x1400ffd0 }, - { 0x01000568, 0x1400ffd0 }, - { 0x01000569, 0x1400ffd0 }, - { 0x0100056a, 0x1400ffd0 }, - { 0x0100056b, 0x1400ffd0 }, - { 0x0100056c, 0x1400ffd0 }, - { 0x0100056d, 0x1400ffd0 }, - { 0x0100056e, 0x1400ffd0 }, - { 0x0100056f, 0x1400ffd0 }, - { 0x01000570, 0x1400ffd0 }, - { 0x01000571, 0x1400ffd0 }, - { 0x01000572, 0x1400ffd0 }, - { 0x01000573, 0x1400ffd0 }, - { 0x01000574, 0x1400ffd0 }, - { 0x01000575, 0x1400ffd0 }, - { 0x01000576, 0x1400ffd0 }, - { 0x01000577, 0x1400ffd0 }, - { 0x01000578, 0x1400ffd0 }, - { 0x01000579, 0x1400ffd0 }, - { 0x0100057a, 0x1400ffd0 }, - { 0x0100057b, 0x1400ffd0 }, - { 0x0100057c, 0x1400ffd0 }, - { 0x0100057d, 0x1400ffd0 }, - { 0x0100057e, 0x1400ffd0 }, - { 0x0100057f, 0x1400ffd0 }, - { 0x01000580, 0x1400ffd0 }, - { 0x01000581, 0x1400ffd0 }, - { 0x01000582, 0x1400ffd0 }, - { 0x01000583, 0x1400ffd0 }, - { 0x01000584, 0x1400ffd0 }, - { 0x01000585, 0x1400ffd0 }, - { 0x01000586, 0x1400ffd0 }, - { 0x01000587, 0x14000000 }, - { 0x09000589, 0x54000000 }, - { 0x0100058a, 0x44000000 }, - { 0x19800591, 0x3000002c }, - { 0x190005be, 0x54000000 }, - { 0x190005bf, 0x30000000 }, - { 0x190005c0, 0x54000000 }, - { 0x198005c1, 0x30000001 }, - { 0x190005c3, 0x54000000 }, - { 0x198005c4, 0x30000001 }, - { 0x190005c6, 0x54000000 }, - { 0x190005c7, 0x30000000 }, - { 0x198005d0, 0x1c00001a }, - { 0x198005f0, 0x1c000002 }, - { 0x198005f3, 0x54000001 }, - { 0x09800600, 0x04000003 }, - { 0x0000060b, 0x5c000000 }, - { 0x0900060c, 0x54000000 }, - { 0x0000060d, 0x54000000 }, - { 0x0080060e, 0x68000001 }, - { 0x00800610, 0x30000005 }, - { 0x0900061b, 0x54000000 }, - { 0x0000061e, 0x54000000 }, - { 0x0900061f, 0x54000000 }, - { 0x00800621, 0x1c000019 }, - { 0x09000640, 0x18000000 }, - { 0x00800641, 0x1c000009 }, - { 0x1b80064b, 0x3000000a }, - { 0x00800656, 0x30000008 }, - { 0x09800660, 0x34000009 }, - { 0x0080066a, 0x54000003 }, - { 0x0080066e, 0x1c000001 }, - { 0x1b000670, 0x30000000 }, - { 0x00800671, 0x1c000062 }, - { 0x000006d4, 0x54000000 }, - { 0x000006d5, 0x1c000000 }, - { 0x008006d6, 0x30000006 }, - { 0x090006dd, 0x04000000 }, - { 0x000006de, 0x2c000000 }, - { 0x008006df, 0x30000005 }, - { 0x008006e5, 0x18000001 }, - { 0x008006e7, 0x30000001 }, - { 0x000006e9, 0x68000000 }, - { 0x008006ea, 0x30000003 }, - { 0x008006ee, 0x1c000001 }, - { 0x008006f0, 0x34000009 }, - { 0x008006fa, 0x1c000002 }, - { 0x008006fd, 0x68000001 }, - { 0x000006ff, 0x1c000000 }, - { 0x31800700, 0x5400000d }, - { 0x3100070f, 0x04000000 }, - { 0x31000710, 0x1c000000 }, - { 0x31000711, 0x30000000 }, - { 0x31800712, 0x1c00001d }, - { 0x31800730, 0x3000001a }, - { 0x3180074d, 0x1c000002 }, - { 0x00800750, 0x1c00001d }, - { 0x37800780, 0x1c000025 }, - { 0x378007a6, 0x3000000a }, - { 0x370007b1, 0x1c000000 }, - { 0x3f8007c0, 0x34000009 }, - { 0x3f8007ca, 0x1c000020 }, - { 0x3f8007eb, 0x30000008 }, - { 0x3f8007f4, 0x18000001 }, - { 0x3f0007f6, 0x68000000 }, - { 0x3f8007f7, 0x54000002 }, - { 0x3f0007fa, 0x18000000 }, - { 0x0e800901, 0x30000001 }, - { 0x0e000903, 0x28000000 }, - { 0x0e800904, 0x1c000035 }, - { 0x0e00093c, 0x30000000 }, - { 0x0e00093d, 0x1c000000 }, - { 0x0e80093e, 0x28000002 }, - { 0x0e800941, 0x30000007 }, - { 0x0e800949, 0x28000003 }, - { 0x0e00094d, 0x30000000 }, - { 0x0e000950, 0x1c000000 }, - { 0x0e800951, 0x30000003 }, - { 0x0e800958, 0x1c000009 }, - { 0x0e800962, 0x30000001 }, - { 0x09800964, 0x54000001 }, - { 0x0e800966, 0x34000009 }, - { 0x09000970, 0x54000000 }, - { 0x0e80097b, 0x1c000004 }, - { 0x02000981, 0x30000000 }, - { 0x02800982, 0x28000001 }, - { 0x02800985, 0x1c000007 }, - { 0x0280098f, 0x1c000001 }, - { 0x02800993, 0x1c000015 }, - { 0x028009aa, 0x1c000006 }, - { 0x020009b2, 0x1c000000 }, - { 0x028009b6, 0x1c000003 }, - { 0x020009bc, 0x30000000 }, - { 0x020009bd, 0x1c000000 }, - { 0x028009be, 0x28000002 }, - { 0x028009c1, 0x30000003 }, - { 0x028009c7, 0x28000001 }, - { 0x028009cb, 0x28000001 }, - { 0x020009cd, 0x30000000 }, - { 0x020009ce, 0x1c000000 }, - { 0x020009d7, 0x28000000 }, - { 0x028009dc, 0x1c000001 }, - { 0x028009df, 0x1c000002 }, - { 0x028009e2, 0x30000001 }, - { 0x028009e6, 0x34000009 }, - { 0x028009f0, 0x1c000001 }, - { 0x028009f2, 0x5c000001 }, - { 0x028009f4, 0x3c000005 }, - { 0x020009fa, 0x68000000 }, - { 0x15800a01, 0x30000001 }, - { 0x15000a03, 0x28000000 }, - { 0x15800a05, 0x1c000005 }, - { 0x15800a0f, 0x1c000001 }, - { 0x15800a13, 0x1c000015 }, - { 0x15800a2a, 0x1c000006 }, - { 0x15800a32, 0x1c000001 }, - { 0x15800a35, 0x1c000001 }, - { 0x15800a38, 0x1c000001 }, - { 0x15000a3c, 0x30000000 }, - { 0x15800a3e, 0x28000002 }, - { 0x15800a41, 0x30000001 }, - { 0x15800a47, 0x30000001 }, - { 0x15800a4b, 0x30000002 }, - { 0x15800a59, 0x1c000003 }, - { 0x15000a5e, 0x1c000000 }, - { 0x15800a66, 0x34000009 }, - { 0x15800a70, 0x30000001 }, - { 0x15800a72, 0x1c000002 }, - { 0x14800a81, 0x30000001 }, - { 0x14000a83, 0x28000000 }, - { 0x14800a85, 0x1c000008 }, - { 0x14800a8f, 0x1c000002 }, - { 0x14800a93, 0x1c000015 }, - { 0x14800aaa, 0x1c000006 }, - { 0x14800ab2, 0x1c000001 }, - { 0x14800ab5, 0x1c000004 }, - { 0x14000abc, 0x30000000 }, - { 0x14000abd, 0x1c000000 }, - { 0x14800abe, 0x28000002 }, - { 0x14800ac1, 0x30000004 }, - { 0x14800ac7, 0x30000001 }, - { 0x14000ac9, 0x28000000 }, - { 0x14800acb, 0x28000001 }, - { 0x14000acd, 0x30000000 }, - { 0x14000ad0, 0x1c000000 }, - { 0x14800ae0, 0x1c000001 }, - { 0x14800ae2, 0x30000001 }, - { 0x14800ae6, 0x34000009 }, - { 0x14000af1, 0x5c000000 }, - { 0x2b000b01, 0x30000000 }, - { 0x2b800b02, 0x28000001 }, - { 0x2b800b05, 0x1c000007 }, - { 0x2b800b0f, 0x1c000001 }, - { 0x2b800b13, 0x1c000015 }, - { 0x2b800b2a, 0x1c000006 }, - { 0x2b800b32, 0x1c000001 }, - { 0x2b800b35, 0x1c000004 }, - { 0x2b000b3c, 0x30000000 }, - { 0x2b000b3d, 0x1c000000 }, - { 0x2b000b3e, 0x28000000 }, - { 0x2b000b3f, 0x30000000 }, - { 0x2b000b40, 0x28000000 }, - { 0x2b800b41, 0x30000002 }, - { 0x2b800b47, 0x28000001 }, - { 0x2b800b4b, 0x28000001 }, - { 0x2b000b4d, 0x30000000 }, - { 0x2b000b56, 0x30000000 }, - { 0x2b000b57, 0x28000000 }, - { 0x2b800b5c, 0x1c000001 }, - { 0x2b800b5f, 0x1c000002 }, - { 0x2b800b66, 0x34000009 }, - { 0x2b000b70, 0x68000000 }, - { 0x2b000b71, 0x1c000000 }, - { 0x35000b82, 0x30000000 }, - { 0x35000b83, 0x1c000000 }, - { 0x35800b85, 0x1c000005 }, - { 0x35800b8e, 0x1c000002 }, - { 0x35800b92, 0x1c000003 }, - { 0x35800b99, 0x1c000001 }, - { 0x35000b9c, 0x1c000000 }, - { 0x35800b9e, 0x1c000001 }, - { 0x35800ba3, 0x1c000001 }, - { 0x35800ba8, 0x1c000002 }, - { 0x35800bae, 0x1c00000b }, - { 0x35800bbe, 0x28000001 }, - { 0x35000bc0, 0x30000000 }, - { 0x35800bc1, 0x28000001 }, - { 0x35800bc6, 0x28000002 }, - { 0x35800bca, 0x28000002 }, - { 0x35000bcd, 0x30000000 }, - { 0x35000bd7, 0x28000000 }, - { 0x35800be6, 0x34000009 }, - { 0x35800bf0, 0x3c000002 }, - { 0x35800bf3, 0x68000005 }, - { 0x35000bf9, 0x5c000000 }, - { 0x35000bfa, 0x68000000 }, - { 0x36800c01, 0x28000002 }, - { 0x36800c05, 0x1c000007 }, - { 0x36800c0e, 0x1c000002 }, - { 0x36800c12, 0x1c000016 }, - { 0x36800c2a, 0x1c000009 }, - { 0x36800c35, 0x1c000004 }, - { 0x36800c3e, 0x30000002 }, - { 0x36800c41, 0x28000003 }, - { 0x36800c46, 0x30000002 }, - { 0x36800c4a, 0x30000003 }, - { 0x36800c55, 0x30000001 }, - { 0x36800c60, 0x1c000001 }, - { 0x36800c66, 0x34000009 }, - { 0x1c800c82, 0x28000001 }, - { 0x1c800c85, 0x1c000007 }, - { 0x1c800c8e, 0x1c000002 }, - { 0x1c800c92, 0x1c000016 }, - { 0x1c800caa, 0x1c000009 }, - { 0x1c800cb5, 0x1c000004 }, - { 0x1c000cbc, 0x30000000 }, - { 0x1c000cbd, 0x1c000000 }, - { 0x1c000cbe, 0x28000000 }, - { 0x1c000cbf, 0x30000000 }, - { 0x1c800cc0, 0x28000004 }, - { 0x1c000cc6, 0x30000000 }, - { 0x1c800cc7, 0x28000001 }, - { 0x1c800cca, 0x28000001 }, - { 0x1c800ccc, 0x30000001 }, - { 0x1c800cd5, 0x28000001 }, - { 0x1c000cde, 0x1c000000 }, - { 0x1c800ce0, 0x1c000001 }, - { 0x1c800ce2, 0x30000001 }, - { 0x1c800ce6, 0x34000009 }, - { 0x1c800cf1, 0x68000001 }, - { 0x24800d02, 0x28000001 }, - { 0x24800d05, 0x1c000007 }, - { 0x24800d0e, 0x1c000002 }, - { 0x24800d12, 0x1c000016 }, - { 0x24800d2a, 0x1c00000f }, - { 0x24800d3e, 0x28000002 }, - { 0x24800d41, 0x30000002 }, - { 0x24800d46, 0x28000002 }, - { 0x24800d4a, 0x28000002 }, - { 0x24000d4d, 0x30000000 }, - { 0x24000d57, 0x28000000 }, - { 0x24800d60, 0x1c000001 }, - { 0x24800d66, 0x34000009 }, - { 0x2f800d82, 0x28000001 }, - { 0x2f800d85, 0x1c000011 }, - { 0x2f800d9a, 0x1c000017 }, - { 0x2f800db3, 0x1c000008 }, - { 0x2f000dbd, 0x1c000000 }, - { 0x2f800dc0, 0x1c000006 }, - { 0x2f000dca, 0x30000000 }, - { 0x2f800dcf, 0x28000002 }, - { 0x2f800dd2, 0x30000002 }, - { 0x2f000dd6, 0x30000000 }, - { 0x2f800dd8, 0x28000007 }, - { 0x2f800df2, 0x28000001 }, - { 0x2f000df4, 0x54000000 }, - { 0x38800e01, 0x1c00002f }, - { 0x38000e31, 0x30000000 }, - { 0x38800e32, 0x1c000001 }, - { 0x38800e34, 0x30000006 }, - { 0x09000e3f, 0x5c000000 }, - { 0x38800e40, 0x1c000005 }, - { 0x38000e46, 0x18000000 }, - { 0x38800e47, 0x30000007 }, - { 0x38000e4f, 0x54000000 }, - { 0x38800e50, 0x34000009 }, - { 0x38800e5a, 0x54000001 }, - { 0x20800e81, 0x1c000001 }, - { 0x20000e84, 0x1c000000 }, - { 0x20800e87, 0x1c000001 }, - { 0x20000e8a, 0x1c000000 }, - { 0x20000e8d, 0x1c000000 }, - { 0x20800e94, 0x1c000003 }, - { 0x20800e99, 0x1c000006 }, - { 0x20800ea1, 0x1c000002 }, - { 0x20000ea5, 0x1c000000 }, - { 0x20000ea7, 0x1c000000 }, - { 0x20800eaa, 0x1c000001 }, - { 0x20800ead, 0x1c000003 }, - { 0x20000eb1, 0x30000000 }, - { 0x20800eb2, 0x1c000001 }, - { 0x20800eb4, 0x30000005 }, - { 0x20800ebb, 0x30000001 }, - { 0x20000ebd, 0x1c000000 }, - { 0x20800ec0, 0x1c000004 }, - { 0x20000ec6, 0x18000000 }, - { 0x20800ec8, 0x30000005 }, - { 0x20800ed0, 0x34000009 }, - { 0x20800edc, 0x1c000001 }, - { 0x39000f00, 0x1c000000 }, - { 0x39800f01, 0x68000002 }, - { 0x39800f04, 0x5400000e }, - { 0x39800f13, 0x68000004 }, - { 0x39800f18, 0x30000001 }, - { 0x39800f1a, 0x68000005 }, - { 0x39800f20, 0x34000009 }, - { 0x39800f2a, 0x3c000009 }, - { 0x39000f34, 0x68000000 }, - { 0x39000f35, 0x30000000 }, - { 0x39000f36, 0x68000000 }, - { 0x39000f37, 0x30000000 }, - { 0x39000f38, 0x68000000 }, - { 0x39000f39, 0x30000000 }, - { 0x39000f3a, 0x58000000 }, - { 0x39000f3b, 0x48000000 }, - { 0x39000f3c, 0x58000000 }, - { 0x39000f3d, 0x48000000 }, - { 0x39800f3e, 0x28000001 }, - { 0x39800f40, 0x1c000007 }, - { 0x39800f49, 0x1c000021 }, - { 0x39800f71, 0x3000000d }, - { 0x39000f7f, 0x28000000 }, - { 0x39800f80, 0x30000004 }, - { 0x39000f85, 0x54000000 }, - { 0x39800f86, 0x30000001 }, - { 0x39800f88, 0x1c000003 }, - { 0x39800f90, 0x30000007 }, - { 0x39800f99, 0x30000023 }, - { 0x39800fbe, 0x68000007 }, - { 0x39000fc6, 0x30000000 }, - { 0x39800fc7, 0x68000005 }, - { 0x39000fcf, 0x68000000 }, - { 0x39800fd0, 0x54000001 }, - { 0x26801000, 0x1c000021 }, - { 0x26801023, 0x1c000004 }, - { 0x26801029, 0x1c000001 }, - { 0x2600102c, 0x28000000 }, - { 0x2680102d, 0x30000003 }, - { 0x26001031, 0x28000000 }, - { 0x26001032, 0x30000000 }, - { 0x26801036, 0x30000001 }, - { 0x26001038, 0x28000000 }, - { 0x26001039, 0x30000000 }, - { 0x26801040, 0x34000009 }, - { 0x2680104a, 0x54000005 }, - { 0x26801050, 0x1c000005 }, - { 0x26801056, 0x28000001 }, - { 0x26801058, 0x30000001 }, - { 0x100010a0, 0x24001c60 }, - { 0x100010a1, 0x24001c60 }, - { 0x100010a2, 0x24001c60 }, - { 0x100010a3, 0x24001c60 }, - { 0x100010a4, 0x24001c60 }, - { 0x100010a5, 0x24001c60 }, - { 0x100010a6, 0x24001c60 }, - { 0x100010a7, 0x24001c60 }, - { 0x100010a8, 0x24001c60 }, - { 0x100010a9, 0x24001c60 }, - { 0x100010aa, 0x24001c60 }, - { 0x100010ab, 0x24001c60 }, - { 0x100010ac, 0x24001c60 }, - { 0x100010ad, 0x24001c60 }, - { 0x100010ae, 0x24001c60 }, - { 0x100010af, 0x24001c60 }, - { 0x100010b0, 0x24001c60 }, - { 0x100010b1, 0x24001c60 }, - { 0x100010b2, 0x24001c60 }, - { 0x100010b3, 0x24001c60 }, - { 0x100010b4, 0x24001c60 }, - { 0x100010b5, 0x24001c60 }, - { 0x100010b6, 0x24001c60 }, - { 0x100010b7, 0x24001c60 }, - { 0x100010b8, 0x24001c60 }, - { 0x100010b9, 0x24001c60 }, - { 0x100010ba, 0x24001c60 }, - { 0x100010bb, 0x24001c60 }, - { 0x100010bc, 0x24001c60 }, - { 0x100010bd, 0x24001c60 }, - { 0x100010be, 0x24001c60 }, - { 0x100010bf, 0x24001c60 }, - { 0x100010c0, 0x24001c60 }, - { 0x100010c1, 0x24001c60 }, - { 0x100010c2, 0x24001c60 }, - { 0x100010c3, 0x24001c60 }, - { 0x100010c4, 0x24001c60 }, - { 0x100010c5, 0x24001c60 }, - { 0x108010d0, 0x1c00002a }, - { 0x090010fb, 0x54000000 }, - { 0x100010fc, 0x18000000 }, - { 0x17801100, 0x1c000059 }, - { 0x1780115f, 0x1c000043 }, - { 0x178011a8, 0x1c000051 }, - { 0x0f801200, 0x1c000048 }, - { 0x0f80124a, 0x1c000003 }, - { 0x0f801250, 0x1c000006 }, - { 0x0f001258, 0x1c000000 }, - { 0x0f80125a, 0x1c000003 }, - { 0x0f801260, 0x1c000028 }, - { 0x0f80128a, 0x1c000003 }, - { 0x0f801290, 0x1c000020 }, - { 0x0f8012b2, 0x1c000003 }, - { 0x0f8012b8, 0x1c000006 }, - { 0x0f0012c0, 0x1c000000 }, - { 0x0f8012c2, 0x1c000003 }, - { 0x0f8012c8, 0x1c00000e }, - { 0x0f8012d8, 0x1c000038 }, - { 0x0f801312, 0x1c000003 }, - { 0x0f801318, 0x1c000042 }, - { 0x0f00135f, 0x30000000 }, - { 0x0f001360, 0x68000000 }, - { 0x0f801361, 0x54000007 }, - { 0x0f801369, 0x3c000013 }, - { 0x0f801380, 0x1c00000f }, - { 0x0f801390, 0x68000009 }, - { 0x088013a0, 0x1c000054 }, - { 0x07801401, 0x1c00026b }, - { 0x0780166d, 0x54000001 }, - { 0x0780166f, 0x1c000007 }, - { 0x28001680, 0x74000000 }, - { 0x28801681, 0x1c000019 }, - { 0x2800169b, 0x58000000 }, - { 0x2800169c, 0x48000000 }, - { 0x2d8016a0, 0x1c00004a }, - { 0x098016eb, 0x54000002 }, - { 0x2d8016ee, 0x38000002 }, - { 0x32801700, 0x1c00000c }, - { 0x3280170e, 0x1c000003 }, - { 0x32801712, 0x30000002 }, - { 0x18801720, 0x1c000011 }, - { 0x18801732, 0x30000002 }, - { 0x09801735, 0x54000001 }, - { 0x06801740, 0x1c000011 }, - { 0x06801752, 0x30000001 }, - { 0x33801760, 0x1c00000c }, - { 0x3380176e, 0x1c000002 }, - { 0x33801772, 0x30000001 }, - { 0x1f801780, 0x1c000033 }, - { 0x1f8017b4, 0x04000001 }, - { 0x1f0017b6, 0x28000000 }, - { 0x1f8017b7, 0x30000006 }, - { 0x1f8017be, 0x28000007 }, - { 0x1f0017c6, 0x30000000 }, - { 0x1f8017c7, 0x28000001 }, - { 0x1f8017c9, 0x3000000a }, - { 0x1f8017d4, 0x54000002 }, - { 0x1f0017d7, 0x18000000 }, - { 0x1f8017d8, 0x54000002 }, - { 0x1f0017db, 0x5c000000 }, - { 0x1f0017dc, 0x1c000000 }, - { 0x1f0017dd, 0x30000000 }, - { 0x1f8017e0, 0x34000009 }, - { 0x1f8017f0, 0x3c000009 }, - { 0x25801800, 0x54000001 }, - { 0x09801802, 0x54000001 }, - { 0x25001804, 0x54000000 }, - { 0x09001805, 0x54000000 }, - { 0x25001806, 0x44000000 }, - { 0x25801807, 0x54000003 }, - { 0x2580180b, 0x30000002 }, - { 0x2500180e, 0x74000000 }, - { 0x25801810, 0x34000009 }, - { 0x25801820, 0x1c000022 }, - { 0x25001843, 0x18000000 }, - { 0x25801844, 0x1c000033 }, - { 0x25801880, 0x1c000028 }, - { 0x250018a9, 0x30000000 }, - { 0x22801900, 0x1c00001c }, - { 0x22801920, 0x30000002 }, - { 0x22801923, 0x28000003 }, - { 0x22801927, 0x30000001 }, - { 0x22801929, 0x28000002 }, - { 0x22801930, 0x28000001 }, - { 0x22001932, 0x30000000 }, - { 0x22801933, 0x28000005 }, - { 0x22801939, 0x30000002 }, - { 0x22001940, 0x68000000 }, - { 0x22801944, 0x54000001 }, - { 0x22801946, 0x34000009 }, - { 0x34801950, 0x1c00001d }, - { 0x34801970, 0x1c000004 }, - { 0x27801980, 0x1c000029 }, - { 0x278019b0, 0x28000010 }, - { 0x278019c1, 0x1c000006 }, - { 0x278019c8, 0x28000001 }, - { 0x278019d0, 0x34000009 }, - { 0x278019de, 0x54000001 }, - { 0x1f8019e0, 0x6800001f }, - { 0x05801a00, 0x1c000016 }, - { 0x05801a17, 0x30000001 }, - { 0x05801a19, 0x28000002 }, - { 0x05801a1e, 0x54000001 }, - { 0x3d801b00, 0x30000003 }, - { 0x3d001b04, 0x28000000 }, - { 0x3d801b05, 0x1c00002e }, - { 0x3d001b34, 0x30000000 }, - { 0x3d001b35, 0x28000000 }, - { 0x3d801b36, 0x30000004 }, - { 0x3d001b3b, 0x28000000 }, - { 0x3d001b3c, 0x30000000 }, - { 0x3d801b3d, 0x28000004 }, - { 0x3d001b42, 0x30000000 }, - { 0x3d801b43, 0x28000001 }, - { 0x3d801b45, 0x1c000006 }, - { 0x3d801b50, 0x34000009 }, - { 0x3d801b5a, 0x54000006 }, - { 0x3d801b61, 0x68000009 }, - { 0x3d801b6b, 0x30000008 }, - { 0x3d801b74, 0x68000008 }, - { 0x21801d00, 0x14000025 }, - { 0x13801d26, 0x14000004 }, - { 0x0c001d2b, 0x14000000 }, - { 0x21801d2c, 0x18000030 }, - { 0x13801d5d, 0x18000004 }, - { 0x21801d62, 0x14000003 }, - { 0x13801d66, 0x14000004 }, - { 0x21801d6b, 0x1400000c }, - { 0x0c001d78, 0x18000000 }, - { 0x21801d79, 0x14000003 }, - { 0x21001d7d, 0x14000ee6 }, - { 0x21801d7e, 0x1400001c }, - { 0x21801d9b, 0x18000023 }, - { 0x13001dbf, 0x18000000 }, - { 0x1b801dc0, 0x3000000a }, - { 0x1b801dfe, 0x30000001 }, - { 0x21001e00, 0x24000001 }, - { 0x21001e01, 0x1400ffff }, - { 0x21001e02, 0x24000001 }, - { 0x21001e03, 0x1400ffff }, - { 0x21001e04, 0x24000001 }, - { 0x21001e05, 0x1400ffff }, - { 0x21001e06, 0x24000001 }, - { 0x21001e07, 0x1400ffff }, - { 0x21001e08, 0x24000001 }, - { 0x21001e09, 0x1400ffff }, - { 0x21001e0a, 0x24000001 }, - { 0x21001e0b, 0x1400ffff }, - { 0x21001e0c, 0x24000001 }, - { 0x21001e0d, 0x1400ffff }, - { 0x21001e0e, 0x24000001 }, - { 0x21001e0f, 0x1400ffff }, - { 0x21001e10, 0x24000001 }, - { 0x21001e11, 0x1400ffff }, - { 0x21001e12, 0x24000001 }, - { 0x21001e13, 0x1400ffff }, - { 0x21001e14, 0x24000001 }, - { 0x21001e15, 0x1400ffff }, - { 0x21001e16, 0x24000001 }, - { 0x21001e17, 0x1400ffff }, - { 0x21001e18, 0x24000001 }, - { 0x21001e19, 0x1400ffff }, - { 0x21001e1a, 0x24000001 }, - { 0x21001e1b, 0x1400ffff }, - { 0x21001e1c, 0x24000001 }, - { 0x21001e1d, 0x1400ffff }, - { 0x21001e1e, 0x24000001 }, - { 0x21001e1f, 0x1400ffff }, - { 0x21001e20, 0x24000001 }, - { 0x21001e21, 0x1400ffff }, - { 0x21001e22, 0x24000001 }, - { 0x21001e23, 0x1400ffff }, - { 0x21001e24, 0x24000001 }, - { 0x21001e25, 0x1400ffff }, - { 0x21001e26, 0x24000001 }, - { 0x21001e27, 0x1400ffff }, - { 0x21001e28, 0x24000001 }, - { 0x21001e29, 0x1400ffff }, - { 0x21001e2a, 0x24000001 }, - { 0x21001e2b, 0x1400ffff }, - { 0x21001e2c, 0x24000001 }, - { 0x21001e2d, 0x1400ffff }, - { 0x21001e2e, 0x24000001 }, - { 0x21001e2f, 0x1400ffff }, - { 0x21001e30, 0x24000001 }, - { 0x21001e31, 0x1400ffff }, - { 0x21001e32, 0x24000001 }, - { 0x21001e33, 0x1400ffff }, - { 0x21001e34, 0x24000001 }, - { 0x21001e35, 0x1400ffff }, - { 0x21001e36, 0x24000001 }, - { 0x21001e37, 0x1400ffff }, - { 0x21001e38, 0x24000001 }, - { 0x21001e39, 0x1400ffff }, - { 0x21001e3a, 0x24000001 }, - { 0x21001e3b, 0x1400ffff }, - { 0x21001e3c, 0x24000001 }, - { 0x21001e3d, 0x1400ffff }, - { 0x21001e3e, 0x24000001 }, - { 0x21001e3f, 0x1400ffff }, - { 0x21001e40, 0x24000001 }, - { 0x21001e41, 0x1400ffff }, - { 0x21001e42, 0x24000001 }, - { 0x21001e43, 0x1400ffff }, - { 0x21001e44, 0x24000001 }, - { 0x21001e45, 0x1400ffff }, - { 0x21001e46, 0x24000001 }, - { 0x21001e47, 0x1400ffff }, - { 0x21001e48, 0x24000001 }, - { 0x21001e49, 0x1400ffff }, - { 0x21001e4a, 0x24000001 }, - { 0x21001e4b, 0x1400ffff }, - { 0x21001e4c, 0x24000001 }, - { 0x21001e4d, 0x1400ffff }, - { 0x21001e4e, 0x24000001 }, - { 0x21001e4f, 0x1400ffff }, - { 0x21001e50, 0x24000001 }, - { 0x21001e51, 0x1400ffff }, - { 0x21001e52, 0x24000001 }, - { 0x21001e53, 0x1400ffff }, - { 0x21001e54, 0x24000001 }, - { 0x21001e55, 0x1400ffff }, - { 0x21001e56, 0x24000001 }, - { 0x21001e57, 0x1400ffff }, - { 0x21001e58, 0x24000001 }, - { 0x21001e59, 0x1400ffff }, - { 0x21001e5a, 0x24000001 }, - { 0x21001e5b, 0x1400ffff }, - { 0x21001e5c, 0x24000001 }, - { 0x21001e5d, 0x1400ffff }, - { 0x21001e5e, 0x24000001 }, - { 0x21001e5f, 0x1400ffff }, - { 0x21001e60, 0x24000001 }, - { 0x21001e61, 0x1400ffff }, - { 0x21001e62, 0x24000001 }, - { 0x21001e63, 0x1400ffff }, - { 0x21001e64, 0x24000001 }, - { 0x21001e65, 0x1400ffff }, - { 0x21001e66, 0x24000001 }, - { 0x21001e67, 0x1400ffff }, - { 0x21001e68, 0x24000001 }, - { 0x21001e69, 0x1400ffff }, - { 0x21001e6a, 0x24000001 }, - { 0x21001e6b, 0x1400ffff }, - { 0x21001e6c, 0x24000001 }, - { 0x21001e6d, 0x1400ffff }, - { 0x21001e6e, 0x24000001 }, - { 0x21001e6f, 0x1400ffff }, - { 0x21001e70, 0x24000001 }, - { 0x21001e71, 0x1400ffff }, - { 0x21001e72, 0x24000001 }, - { 0x21001e73, 0x1400ffff }, - { 0x21001e74, 0x24000001 }, - { 0x21001e75, 0x1400ffff }, - { 0x21001e76, 0x24000001 }, - { 0x21001e77, 0x1400ffff }, - { 0x21001e78, 0x24000001 }, - { 0x21001e79, 0x1400ffff }, - { 0x21001e7a, 0x24000001 }, - { 0x21001e7b, 0x1400ffff }, - { 0x21001e7c, 0x24000001 }, - { 0x21001e7d, 0x1400ffff }, - { 0x21001e7e, 0x24000001 }, - { 0x21001e7f, 0x1400ffff }, - { 0x21001e80, 0x24000001 }, - { 0x21001e81, 0x1400ffff }, - { 0x21001e82, 0x24000001 }, - { 0x21001e83, 0x1400ffff }, - { 0x21001e84, 0x24000001 }, - { 0x21001e85, 0x1400ffff }, - { 0x21001e86, 0x24000001 }, - { 0x21001e87, 0x1400ffff }, - { 0x21001e88, 0x24000001 }, - { 0x21001e89, 0x1400ffff }, - { 0x21001e8a, 0x24000001 }, - { 0x21001e8b, 0x1400ffff }, - { 0x21001e8c, 0x24000001 }, - { 0x21001e8d, 0x1400ffff }, - { 0x21001e8e, 0x24000001 }, - { 0x21001e8f, 0x1400ffff }, - { 0x21001e90, 0x24000001 }, - { 0x21001e91, 0x1400ffff }, - { 0x21001e92, 0x24000001 }, - { 0x21001e93, 0x1400ffff }, - { 0x21001e94, 0x24000001 }, - { 0x21001e95, 0x1400ffff }, - { 0x21801e96, 0x14000004 }, - { 0x21001e9b, 0x1400ffc5 }, - { 0x21001ea0, 0x24000001 }, - { 0x21001ea1, 0x1400ffff }, - { 0x21001ea2, 0x24000001 }, - { 0x21001ea3, 0x1400ffff }, - { 0x21001ea4, 0x24000001 }, - { 0x21001ea5, 0x1400ffff }, - { 0x21001ea6, 0x24000001 }, - { 0x21001ea7, 0x1400ffff }, - { 0x21001ea8, 0x24000001 }, - { 0x21001ea9, 0x1400ffff }, - { 0x21001eaa, 0x24000001 }, - { 0x21001eab, 0x1400ffff }, - { 0x21001eac, 0x24000001 }, - { 0x21001ead, 0x1400ffff }, - { 0x21001eae, 0x24000001 }, - { 0x21001eaf, 0x1400ffff }, - { 0x21001eb0, 0x24000001 }, - { 0x21001eb1, 0x1400ffff }, - { 0x21001eb2, 0x24000001 }, - { 0x21001eb3, 0x1400ffff }, - { 0x21001eb4, 0x24000001 }, - { 0x21001eb5, 0x1400ffff }, - { 0x21001eb6, 0x24000001 }, - { 0x21001eb7, 0x1400ffff }, - { 0x21001eb8, 0x24000001 }, - { 0x21001eb9, 0x1400ffff }, - { 0x21001eba, 0x24000001 }, - { 0x21001ebb, 0x1400ffff }, - { 0x21001ebc, 0x24000001 }, - { 0x21001ebd, 0x1400ffff }, - { 0x21001ebe, 0x24000001 }, - { 0x21001ebf, 0x1400ffff }, - { 0x21001ec0, 0x24000001 }, - { 0x21001ec1, 0x1400ffff }, - { 0x21001ec2, 0x24000001 }, - { 0x21001ec3, 0x1400ffff }, - { 0x21001ec4, 0x24000001 }, - { 0x21001ec5, 0x1400ffff }, - { 0x21001ec6, 0x24000001 }, - { 0x21001ec7, 0x1400ffff }, - { 0x21001ec8, 0x24000001 }, - { 0x21001ec9, 0x1400ffff }, - { 0x21001eca, 0x24000001 }, - { 0x21001ecb, 0x1400ffff }, - { 0x21001ecc, 0x24000001 }, - { 0x21001ecd, 0x1400ffff }, - { 0x21001ece, 0x24000001 }, - { 0x21001ecf, 0x1400ffff }, - { 0x21001ed0, 0x24000001 }, - { 0x21001ed1, 0x1400ffff }, - { 0x21001ed2, 0x24000001 }, - { 0x21001ed3, 0x1400ffff }, - { 0x21001ed4, 0x24000001 }, - { 0x21001ed5, 0x1400ffff }, - { 0x21001ed6, 0x24000001 }, - { 0x21001ed7, 0x1400ffff }, - { 0x21001ed8, 0x24000001 }, - { 0x21001ed9, 0x1400ffff }, - { 0x21001eda, 0x24000001 }, - { 0x21001edb, 0x1400ffff }, - { 0x21001edc, 0x24000001 }, - { 0x21001edd, 0x1400ffff }, - { 0x21001ede, 0x24000001 }, - { 0x21001edf, 0x1400ffff }, - { 0x21001ee0, 0x24000001 }, - { 0x21001ee1, 0x1400ffff }, - { 0x21001ee2, 0x24000001 }, - { 0x21001ee3, 0x1400ffff }, - { 0x21001ee4, 0x24000001 }, - { 0x21001ee5, 0x1400ffff }, - { 0x21001ee6, 0x24000001 }, - { 0x21001ee7, 0x1400ffff }, - { 0x21001ee8, 0x24000001 }, - { 0x21001ee9, 0x1400ffff }, - { 0x21001eea, 0x24000001 }, - { 0x21001eeb, 0x1400ffff }, - { 0x21001eec, 0x24000001 }, - { 0x21001eed, 0x1400ffff }, - { 0x21001eee, 0x24000001 }, - { 0x21001eef, 0x1400ffff }, - { 0x21001ef0, 0x24000001 }, - { 0x21001ef1, 0x1400ffff }, - { 0x21001ef2, 0x24000001 }, - { 0x21001ef3, 0x1400ffff }, - { 0x21001ef4, 0x24000001 }, - { 0x21001ef5, 0x1400ffff }, - { 0x21001ef6, 0x24000001 }, - { 0x21001ef7, 0x1400ffff }, - { 0x21001ef8, 0x24000001 }, - { 0x21001ef9, 0x1400ffff }, - { 0x13001f00, 0x14000008 }, - { 0x13001f01, 0x14000008 }, - { 0x13001f02, 0x14000008 }, - { 0x13001f03, 0x14000008 }, - { 0x13001f04, 0x14000008 }, - { 0x13001f05, 0x14000008 }, - { 0x13001f06, 0x14000008 }, - { 0x13001f07, 0x14000008 }, - { 0x13001f08, 0x2400fff8 }, - { 0x13001f09, 0x2400fff8 }, - { 0x13001f0a, 0x2400fff8 }, - { 0x13001f0b, 0x2400fff8 }, - { 0x13001f0c, 0x2400fff8 }, - { 0x13001f0d, 0x2400fff8 }, - { 0x13001f0e, 0x2400fff8 }, - { 0x13001f0f, 0x2400fff8 }, - { 0x13001f10, 0x14000008 }, - { 0x13001f11, 0x14000008 }, - { 0x13001f12, 0x14000008 }, - { 0x13001f13, 0x14000008 }, - { 0x13001f14, 0x14000008 }, - { 0x13001f15, 0x14000008 }, - { 0x13001f18, 0x2400fff8 }, - { 0x13001f19, 0x2400fff8 }, - { 0x13001f1a, 0x2400fff8 }, - { 0x13001f1b, 0x2400fff8 }, - { 0x13001f1c, 0x2400fff8 }, - { 0x13001f1d, 0x2400fff8 }, - { 0x13001f20, 0x14000008 }, - { 0x13001f21, 0x14000008 }, - { 0x13001f22, 0x14000008 }, - { 0x13001f23, 0x14000008 }, - { 0x13001f24, 0x14000008 }, - { 0x13001f25, 0x14000008 }, - { 0x13001f26, 0x14000008 }, - { 0x13001f27, 0x14000008 }, - { 0x13001f28, 0x2400fff8 }, - { 0x13001f29, 0x2400fff8 }, - { 0x13001f2a, 0x2400fff8 }, - { 0x13001f2b, 0x2400fff8 }, - { 0x13001f2c, 0x2400fff8 }, - { 0x13001f2d, 0x2400fff8 }, - { 0x13001f2e, 0x2400fff8 }, - { 0x13001f2f, 0x2400fff8 }, - { 0x13001f30, 0x14000008 }, - { 0x13001f31, 0x14000008 }, - { 0x13001f32, 0x14000008 }, - { 0x13001f33, 0x14000008 }, - { 0x13001f34, 0x14000008 }, - { 0x13001f35, 0x14000008 }, - { 0x13001f36, 0x14000008 }, - { 0x13001f37, 0x14000008 }, - { 0x13001f38, 0x2400fff8 }, - { 0x13001f39, 0x2400fff8 }, - { 0x13001f3a, 0x2400fff8 }, - { 0x13001f3b, 0x2400fff8 }, - { 0x13001f3c, 0x2400fff8 }, - { 0x13001f3d, 0x2400fff8 }, - { 0x13001f3e, 0x2400fff8 }, - { 0x13001f3f, 0x2400fff8 }, - { 0x13001f40, 0x14000008 }, - { 0x13001f41, 0x14000008 }, - { 0x13001f42, 0x14000008 }, - { 0x13001f43, 0x14000008 }, - { 0x13001f44, 0x14000008 }, - { 0x13001f45, 0x14000008 }, - { 0x13001f48, 0x2400fff8 }, - { 0x13001f49, 0x2400fff8 }, - { 0x13001f4a, 0x2400fff8 }, - { 0x13001f4b, 0x2400fff8 }, - { 0x13001f4c, 0x2400fff8 }, - { 0x13001f4d, 0x2400fff8 }, - { 0x13001f50, 0x14000000 }, - { 0x13001f51, 0x14000008 }, - { 0x13001f52, 0x14000000 }, - { 0x13001f53, 0x14000008 }, - { 0x13001f54, 0x14000000 }, - { 0x13001f55, 0x14000008 }, - { 0x13001f56, 0x14000000 }, - { 0x13001f57, 0x14000008 }, - { 0x13001f59, 0x2400fff8 }, - { 0x13001f5b, 0x2400fff8 }, - { 0x13001f5d, 0x2400fff8 }, - { 0x13001f5f, 0x2400fff8 }, - { 0x13001f60, 0x14000008 }, - { 0x13001f61, 0x14000008 }, - { 0x13001f62, 0x14000008 }, - { 0x13001f63, 0x14000008 }, - { 0x13001f64, 0x14000008 }, - { 0x13001f65, 0x14000008 }, - { 0x13001f66, 0x14000008 }, - { 0x13001f67, 0x14000008 }, - { 0x13001f68, 0x2400fff8 }, - { 0x13001f69, 0x2400fff8 }, - { 0x13001f6a, 0x2400fff8 }, - { 0x13001f6b, 0x2400fff8 }, - { 0x13001f6c, 0x2400fff8 }, - { 0x13001f6d, 0x2400fff8 }, - { 0x13001f6e, 0x2400fff8 }, - { 0x13001f6f, 0x2400fff8 }, - { 0x13001f70, 0x1400004a }, - { 0x13001f71, 0x1400004a }, - { 0x13001f72, 0x14000056 }, - { 0x13001f73, 0x14000056 }, - { 0x13001f74, 0x14000056 }, - { 0x13001f75, 0x14000056 }, - { 0x13001f76, 0x14000064 }, - { 0x13001f77, 0x14000064 }, - { 0x13001f78, 0x14000080 }, - { 0x13001f79, 0x14000080 }, - { 0x13001f7a, 0x14000070 }, - { 0x13001f7b, 0x14000070 }, - { 0x13001f7c, 0x1400007e }, - { 0x13001f7d, 0x1400007e }, - { 0x13001f80, 0x14000008 }, - { 0x13001f81, 0x14000008 }, - { 0x13001f82, 0x14000008 }, - { 0x13001f83, 0x14000008 }, - { 0x13001f84, 0x14000008 }, - { 0x13001f85, 0x14000008 }, - { 0x13001f86, 0x14000008 }, - { 0x13001f87, 0x14000008 }, - { 0x13001f88, 0x2000fff8 }, - { 0x13001f89, 0x2000fff8 }, - { 0x13001f8a, 0x2000fff8 }, - { 0x13001f8b, 0x2000fff8 }, - { 0x13001f8c, 0x2000fff8 }, - { 0x13001f8d, 0x2000fff8 }, - { 0x13001f8e, 0x2000fff8 }, - { 0x13001f8f, 0x2000fff8 }, - { 0x13001f90, 0x14000008 }, - { 0x13001f91, 0x14000008 }, - { 0x13001f92, 0x14000008 }, - { 0x13001f93, 0x14000008 }, - { 0x13001f94, 0x14000008 }, - { 0x13001f95, 0x14000008 }, - { 0x13001f96, 0x14000008 }, - { 0x13001f97, 0x14000008 }, - { 0x13001f98, 0x2000fff8 }, - { 0x13001f99, 0x2000fff8 }, - { 0x13001f9a, 0x2000fff8 }, - { 0x13001f9b, 0x2000fff8 }, - { 0x13001f9c, 0x2000fff8 }, - { 0x13001f9d, 0x2000fff8 }, - { 0x13001f9e, 0x2000fff8 }, - { 0x13001f9f, 0x2000fff8 }, - { 0x13001fa0, 0x14000008 }, - { 0x13001fa1, 0x14000008 }, - { 0x13001fa2, 0x14000008 }, - { 0x13001fa3, 0x14000008 }, - { 0x13001fa4, 0x14000008 }, - { 0x13001fa5, 0x14000008 }, - { 0x13001fa6, 0x14000008 }, - { 0x13001fa7, 0x14000008 }, - { 0x13001fa8, 0x2000fff8 }, - { 0x13001fa9, 0x2000fff8 }, - { 0x13001faa, 0x2000fff8 }, - { 0x13001fab, 0x2000fff8 }, - { 0x13001fac, 0x2000fff8 }, - { 0x13001fad, 0x2000fff8 }, - { 0x13001fae, 0x2000fff8 }, - { 0x13001faf, 0x2000fff8 }, - { 0x13001fb0, 0x14000008 }, - { 0x13001fb1, 0x14000008 }, - { 0x13001fb2, 0x14000000 }, - { 0x13001fb3, 0x14000009 }, - { 0x13001fb4, 0x14000000 }, - { 0x13801fb6, 0x14000001 }, - { 0x13001fb8, 0x2400fff8 }, - { 0x13001fb9, 0x2400fff8 }, - { 0x13001fba, 0x2400ffb6 }, - { 0x13001fbb, 0x2400ffb6 }, - { 0x13001fbc, 0x2000fff7 }, - { 0x13001fbd, 0x60000000 }, - { 0x13001fbe, 0x1400e3db }, - { 0x13801fbf, 0x60000002 }, - { 0x13001fc2, 0x14000000 }, - { 0x13001fc3, 0x14000009 }, - { 0x13001fc4, 0x14000000 }, - { 0x13801fc6, 0x14000001 }, - { 0x13001fc8, 0x2400ffaa }, - { 0x13001fc9, 0x2400ffaa }, - { 0x13001fca, 0x2400ffaa }, - { 0x13001fcb, 0x2400ffaa }, - { 0x13001fcc, 0x2000fff7 }, - { 0x13801fcd, 0x60000002 }, - { 0x13001fd0, 0x14000008 }, - { 0x13001fd1, 0x14000008 }, - { 0x13801fd2, 0x14000001 }, - { 0x13801fd6, 0x14000001 }, - { 0x13001fd8, 0x2400fff8 }, - { 0x13001fd9, 0x2400fff8 }, - { 0x13001fda, 0x2400ff9c }, - { 0x13001fdb, 0x2400ff9c }, - { 0x13801fdd, 0x60000002 }, - { 0x13001fe0, 0x14000008 }, - { 0x13001fe1, 0x14000008 }, - { 0x13801fe2, 0x14000002 }, - { 0x13001fe5, 0x14000007 }, - { 0x13801fe6, 0x14000001 }, - { 0x13001fe8, 0x2400fff8 }, - { 0x13001fe9, 0x2400fff8 }, - { 0x13001fea, 0x2400ff90 }, - { 0x13001feb, 0x2400ff90 }, - { 0x13001fec, 0x2400fff9 }, - { 0x13801fed, 0x60000002 }, - { 0x13001ff2, 0x14000000 }, - { 0x13001ff3, 0x14000009 }, - { 0x13001ff4, 0x14000000 }, - { 0x13801ff6, 0x14000001 }, - { 0x13001ff8, 0x2400ff80 }, - { 0x13001ff9, 0x2400ff80 }, - { 0x13001ffa, 0x2400ff82 }, - { 0x13001ffb, 0x2400ff82 }, - { 0x13001ffc, 0x2000fff7 }, - { 0x13801ffd, 0x60000001 }, - { 0x09802000, 0x7400000a }, - { 0x0900200b, 0x04000000 }, - { 0x1b80200c, 0x04000001 }, - { 0x0980200e, 0x04000001 }, - { 0x09802010, 0x44000005 }, - { 0x09802016, 0x54000001 }, - { 0x09002018, 0x50000000 }, - { 0x09002019, 0x4c000000 }, - { 0x0900201a, 0x58000000 }, - { 0x0980201b, 0x50000001 }, - { 0x0900201d, 0x4c000000 }, - { 0x0900201e, 0x58000000 }, - { 0x0900201f, 0x50000000 }, - { 0x09802020, 0x54000007 }, - { 0x09002028, 0x6c000000 }, - { 0x09002029, 0x70000000 }, - { 0x0980202a, 0x04000004 }, - { 0x0900202f, 0x74000000 }, - { 0x09802030, 0x54000008 }, - { 0x09002039, 0x50000000 }, - { 0x0900203a, 0x4c000000 }, - { 0x0980203b, 0x54000003 }, - { 0x0980203f, 0x40000001 }, - { 0x09802041, 0x54000002 }, - { 0x09002044, 0x64000000 }, - { 0x09002045, 0x58000000 }, - { 0x09002046, 0x48000000 }, - { 0x09802047, 0x5400000a }, - { 0x09002052, 0x64000000 }, - { 0x09002053, 0x54000000 }, - { 0x09002054, 0x40000000 }, - { 0x09802055, 0x54000009 }, - { 0x0900205f, 0x74000000 }, - { 0x09802060, 0x04000003 }, - { 0x0980206a, 0x04000005 }, - { 0x09002070, 0x3c000000 }, - { 0x21002071, 0x14000000 }, - { 0x09802074, 0x3c000005 }, - { 0x0980207a, 0x64000002 }, - { 0x0900207d, 0x58000000 }, - { 0x0900207e, 0x48000000 }, - { 0x2100207f, 0x14000000 }, - { 0x09802080, 0x3c000009 }, - { 0x0980208a, 0x64000002 }, - { 0x0900208d, 0x58000000 }, - { 0x0900208e, 0x48000000 }, - { 0x21802090, 0x18000004 }, - { 0x098020a0, 0x5c000015 }, - { 0x1b8020d0, 0x3000000c }, - { 0x1b8020dd, 0x2c000003 }, - { 0x1b0020e1, 0x30000000 }, - { 0x1b8020e2, 0x2c000002 }, - { 0x1b8020e5, 0x3000000a }, - { 0x09802100, 0x68000001 }, - { 0x09002102, 0x24000000 }, - { 0x09802103, 0x68000003 }, - { 0x09002107, 0x24000000 }, - { 0x09802108, 0x68000001 }, - { 0x0900210a, 0x14000000 }, - { 0x0980210b, 0x24000002 }, - { 0x0980210e, 0x14000001 }, - { 0x09802110, 0x24000002 }, - { 0x09002113, 0x14000000 }, - { 0x09002114, 0x68000000 }, - { 0x09002115, 0x24000000 }, - { 0x09802116, 0x68000002 }, - { 0x09802119, 0x24000004 }, - { 0x0980211e, 0x68000005 }, - { 0x09002124, 0x24000000 }, - { 0x09002125, 0x68000000 }, - { 0x13002126, 0x2400e2a3 }, - { 0x09002127, 0x68000000 }, - { 0x09002128, 0x24000000 }, - { 0x09002129, 0x68000000 }, - { 0x2100212a, 0x2400df41 }, - { 0x2100212b, 0x2400dfba }, - { 0x0980212c, 0x24000001 }, - { 0x0900212e, 0x68000000 }, - { 0x0900212f, 0x14000000 }, - { 0x09802130, 0x24000001 }, - { 0x21002132, 0x2400001c }, - { 0x09002133, 0x24000000 }, - { 0x09002134, 0x14000000 }, - { 0x09802135, 0x1c000003 }, - { 0x09002139, 0x14000000 }, - { 0x0980213a, 0x68000001 }, - { 0x0980213c, 0x14000001 }, - { 0x0980213e, 0x24000001 }, - { 0x09802140, 0x64000004 }, - { 0x09002145, 0x24000000 }, - { 0x09802146, 0x14000003 }, - { 0x0900214a, 0x68000000 }, - { 0x0900214b, 0x64000000 }, - { 0x0980214c, 0x68000001 }, - { 0x2100214e, 0x1400ffe4 }, - { 0x09802153, 0x3c00000c }, - { 0x09002160, 0x38000010 }, - { 0x09002161, 0x38000010 }, - { 0x09002162, 0x38000010 }, - { 0x09002163, 0x38000010 }, - { 0x09002164, 0x38000010 }, - { 0x09002165, 0x38000010 }, - { 0x09002166, 0x38000010 }, - { 0x09002167, 0x38000010 }, - { 0x09002168, 0x38000010 }, - { 0x09002169, 0x38000010 }, - { 0x0900216a, 0x38000010 }, - { 0x0900216b, 0x38000010 }, - { 0x0900216c, 0x38000010 }, - { 0x0900216d, 0x38000010 }, - { 0x0900216e, 0x38000010 }, - { 0x0900216f, 0x38000010 }, - { 0x09002170, 0x3800fff0 }, - { 0x09002171, 0x3800fff0 }, - { 0x09002172, 0x3800fff0 }, - { 0x09002173, 0x3800fff0 }, - { 0x09002174, 0x3800fff0 }, - { 0x09002175, 0x3800fff0 }, - { 0x09002176, 0x3800fff0 }, - { 0x09002177, 0x3800fff0 }, - { 0x09002178, 0x3800fff0 }, - { 0x09002179, 0x3800fff0 }, - { 0x0900217a, 0x3800fff0 }, - { 0x0900217b, 0x3800fff0 }, - { 0x0900217c, 0x3800fff0 }, - { 0x0900217d, 0x3800fff0 }, - { 0x0900217e, 0x3800fff0 }, - { 0x0900217f, 0x3800fff0 }, - { 0x09802180, 0x38000002 }, - { 0x09002183, 0x24000001 }, - { 0x21002184, 0x1400ffff }, - { 0x09802190, 0x64000004 }, - { 0x09802195, 0x68000004 }, - { 0x0980219a, 0x64000001 }, - { 0x0980219c, 0x68000003 }, - { 0x090021a0, 0x64000000 }, - { 0x098021a1, 0x68000001 }, - { 0x090021a3, 0x64000000 }, - { 0x098021a4, 0x68000001 }, - { 0x090021a6, 0x64000000 }, - { 0x098021a7, 0x68000006 }, - { 0x090021ae, 0x64000000 }, - { 0x098021af, 0x6800001e }, - { 0x098021ce, 0x64000001 }, - { 0x098021d0, 0x68000001 }, - { 0x090021d2, 0x64000000 }, - { 0x090021d3, 0x68000000 }, - { 0x090021d4, 0x64000000 }, - { 0x098021d5, 0x6800001e }, - { 0x098021f4, 0x6400010b }, - { 0x09802300, 0x68000007 }, - { 0x09802308, 0x64000003 }, - { 0x0980230c, 0x68000013 }, - { 0x09802320, 0x64000001 }, - { 0x09802322, 0x68000006 }, - { 0x09002329, 0x58000000 }, - { 0x0900232a, 0x48000000 }, - { 0x0980232b, 0x68000050 }, - { 0x0900237c, 0x64000000 }, - { 0x0980237d, 0x6800001d }, - { 0x0980239b, 0x64000018 }, - { 0x098023b4, 0x68000027 }, - { 0x098023dc, 0x64000005 }, - { 0x098023e2, 0x68000005 }, - { 0x09802400, 0x68000026 }, - { 0x09802440, 0x6800000a }, - { 0x09802460, 0x3c00003b }, - { 0x0980249c, 0x68000019 }, - { 0x090024b6, 0x6800001a }, - { 0x090024b7, 0x6800001a }, - { 0x090024b8, 0x6800001a }, - { 0x090024b9, 0x6800001a }, - { 0x090024ba, 0x6800001a }, - { 0x090024bb, 0x6800001a }, - { 0x090024bc, 0x6800001a }, - { 0x090024bd, 0x6800001a }, - { 0x090024be, 0x6800001a }, - { 0x090024bf, 0x6800001a }, - { 0x090024c0, 0x6800001a }, - { 0x090024c1, 0x6800001a }, - { 0x090024c2, 0x6800001a }, - { 0x090024c3, 0x6800001a }, - { 0x090024c4, 0x6800001a }, - { 0x090024c5, 0x6800001a }, - { 0x090024c6, 0x6800001a }, - { 0x090024c7, 0x6800001a }, - { 0x090024c8, 0x6800001a }, - { 0x090024c9, 0x6800001a }, - { 0x090024ca, 0x6800001a }, - { 0x090024cb, 0x6800001a }, - { 0x090024cc, 0x6800001a }, - { 0x090024cd, 0x6800001a }, - { 0x090024ce, 0x6800001a }, - { 0x090024cf, 0x6800001a }, - { 0x090024d0, 0x6800ffe6 }, - { 0x090024d1, 0x6800ffe6 }, - { 0x090024d2, 0x6800ffe6 }, - { 0x090024d3, 0x6800ffe6 }, - { 0x090024d4, 0x6800ffe6 }, - { 0x090024d5, 0x6800ffe6 }, - { 0x090024d6, 0x6800ffe6 }, - { 0x090024d7, 0x6800ffe6 }, - { 0x090024d8, 0x6800ffe6 }, - { 0x090024d9, 0x6800ffe6 }, - { 0x090024da, 0x6800ffe6 }, - { 0x090024db, 0x6800ffe6 }, - { 0x090024dc, 0x6800ffe6 }, - { 0x090024dd, 0x6800ffe6 }, - { 0x090024de, 0x6800ffe6 }, - { 0x090024df, 0x6800ffe6 }, - { 0x090024e0, 0x6800ffe6 }, - { 0x090024e1, 0x6800ffe6 }, - { 0x090024e2, 0x6800ffe6 }, - { 0x090024e3, 0x6800ffe6 }, - { 0x090024e4, 0x6800ffe6 }, - { 0x090024e5, 0x6800ffe6 }, - { 0x090024e6, 0x6800ffe6 }, - { 0x090024e7, 0x6800ffe6 }, - { 0x090024e8, 0x6800ffe6 }, - { 0x090024e9, 0x6800ffe6 }, - { 0x098024ea, 0x3c000015 }, - { 0x09802500, 0x680000b6 }, - { 0x090025b7, 0x64000000 }, - { 0x098025b8, 0x68000008 }, - { 0x090025c1, 0x64000000 }, - { 0x098025c2, 0x68000035 }, - { 0x098025f8, 0x64000007 }, - { 0x09802600, 0x6800006e }, - { 0x0900266f, 0x64000000 }, - { 0x09802670, 0x6800002c }, - { 0x098026a0, 0x68000012 }, - { 0x09802701, 0x68000003 }, - { 0x09802706, 0x68000003 }, - { 0x0980270c, 0x6800001b }, - { 0x09802729, 0x68000022 }, - { 0x0900274d, 0x68000000 }, - { 0x0980274f, 0x68000003 }, - { 0x09002756, 0x68000000 }, - { 0x09802758, 0x68000006 }, - { 0x09802761, 0x68000006 }, - { 0x09002768, 0x58000000 }, - { 0x09002769, 0x48000000 }, - { 0x0900276a, 0x58000000 }, - { 0x0900276b, 0x48000000 }, - { 0x0900276c, 0x58000000 }, - { 0x0900276d, 0x48000000 }, - { 0x0900276e, 0x58000000 }, - { 0x0900276f, 0x48000000 }, - { 0x09002770, 0x58000000 }, - { 0x09002771, 0x48000000 }, - { 0x09002772, 0x58000000 }, - { 0x09002773, 0x48000000 }, - { 0x09002774, 0x58000000 }, - { 0x09002775, 0x48000000 }, - { 0x09802776, 0x3c00001d }, - { 0x09002794, 0x68000000 }, - { 0x09802798, 0x68000017 }, - { 0x098027b1, 0x6800000d }, - { 0x098027c0, 0x64000004 }, - { 0x090027c5, 0x58000000 }, - { 0x090027c6, 0x48000000 }, - { 0x098027c7, 0x64000003 }, - { 0x098027d0, 0x64000015 }, - { 0x090027e6, 0x58000000 }, - { 0x090027e7, 0x48000000 }, - { 0x090027e8, 0x58000000 }, - { 0x090027e9, 0x48000000 }, - { 0x090027ea, 0x58000000 }, - { 0x090027eb, 0x48000000 }, - { 0x098027f0, 0x6400000f }, - { 0x04802800, 0x680000ff }, - { 0x09802900, 0x64000082 }, - { 0x09002983, 0x58000000 }, - { 0x09002984, 0x48000000 }, - { 0x09002985, 0x58000000 }, - { 0x09002986, 0x48000000 }, - { 0x09002987, 0x58000000 }, - { 0x09002988, 0x48000000 }, - { 0x09002989, 0x58000000 }, - { 0x0900298a, 0x48000000 }, - { 0x0900298b, 0x58000000 }, - { 0x0900298c, 0x48000000 }, - { 0x0900298d, 0x58000000 }, - { 0x0900298e, 0x48000000 }, - { 0x0900298f, 0x58000000 }, - { 0x09002990, 0x48000000 }, - { 0x09002991, 0x58000000 }, - { 0x09002992, 0x48000000 }, - { 0x09002993, 0x58000000 }, - { 0x09002994, 0x48000000 }, - { 0x09002995, 0x58000000 }, - { 0x09002996, 0x48000000 }, - { 0x09002997, 0x58000000 }, - { 0x09002998, 0x48000000 }, - { 0x09802999, 0x6400003e }, - { 0x090029d8, 0x58000000 }, - { 0x090029d9, 0x48000000 }, - { 0x090029da, 0x58000000 }, - { 0x090029db, 0x48000000 }, - { 0x098029dc, 0x6400001f }, - { 0x090029fc, 0x58000000 }, - { 0x090029fd, 0x48000000 }, - { 0x098029fe, 0x64000101 }, - { 0x09802b00, 0x6800001a }, - { 0x09802b20, 0x68000003 }, - { 0x11002c00, 0x24000030 }, - { 0x11002c01, 0x24000030 }, - { 0x11002c02, 0x24000030 }, - { 0x11002c03, 0x24000030 }, - { 0x11002c04, 0x24000030 }, - { 0x11002c05, 0x24000030 }, - { 0x11002c06, 0x24000030 }, - { 0x11002c07, 0x24000030 }, - { 0x11002c08, 0x24000030 }, - { 0x11002c09, 0x24000030 }, - { 0x11002c0a, 0x24000030 }, - { 0x11002c0b, 0x24000030 }, - { 0x11002c0c, 0x24000030 }, - { 0x11002c0d, 0x24000030 }, - { 0x11002c0e, 0x24000030 }, - { 0x11002c0f, 0x24000030 }, - { 0x11002c10, 0x24000030 }, - { 0x11002c11, 0x24000030 }, - { 0x11002c12, 0x24000030 }, - { 0x11002c13, 0x24000030 }, - { 0x11002c14, 0x24000030 }, - { 0x11002c15, 0x24000030 }, - { 0x11002c16, 0x24000030 }, - { 0x11002c17, 0x24000030 }, - { 0x11002c18, 0x24000030 }, - { 0x11002c19, 0x24000030 }, - { 0x11002c1a, 0x24000030 }, - { 0x11002c1b, 0x24000030 }, - { 0x11002c1c, 0x24000030 }, - { 0x11002c1d, 0x24000030 }, - { 0x11002c1e, 0x24000030 }, - { 0x11002c1f, 0x24000030 }, - { 0x11002c20, 0x24000030 }, - { 0x11002c21, 0x24000030 }, - { 0x11002c22, 0x24000030 }, - { 0x11002c23, 0x24000030 }, - { 0x11002c24, 0x24000030 }, - { 0x11002c25, 0x24000030 }, - { 0x11002c26, 0x24000030 }, - { 0x11002c27, 0x24000030 }, - { 0x11002c28, 0x24000030 }, - { 0x11002c29, 0x24000030 }, - { 0x11002c2a, 0x24000030 }, - { 0x11002c2b, 0x24000030 }, - { 0x11002c2c, 0x24000030 }, - { 0x11002c2d, 0x24000030 }, - { 0x11002c2e, 0x24000030 }, - { 0x11002c30, 0x1400ffd0 }, - { 0x11002c31, 0x1400ffd0 }, - { 0x11002c32, 0x1400ffd0 }, - { 0x11002c33, 0x1400ffd0 }, - { 0x11002c34, 0x1400ffd0 }, - { 0x11002c35, 0x1400ffd0 }, - { 0x11002c36, 0x1400ffd0 }, - { 0x11002c37, 0x1400ffd0 }, - { 0x11002c38, 0x1400ffd0 }, - { 0x11002c39, 0x1400ffd0 }, - { 0x11002c3a, 0x1400ffd0 }, - { 0x11002c3b, 0x1400ffd0 }, - { 0x11002c3c, 0x1400ffd0 }, - { 0x11002c3d, 0x1400ffd0 }, - { 0x11002c3e, 0x1400ffd0 }, - { 0x11002c3f, 0x1400ffd0 }, - { 0x11002c40, 0x1400ffd0 }, - { 0x11002c41, 0x1400ffd0 }, - { 0x11002c42, 0x1400ffd0 }, - { 0x11002c43, 0x1400ffd0 }, - { 0x11002c44, 0x1400ffd0 }, - { 0x11002c45, 0x1400ffd0 }, - { 0x11002c46, 0x1400ffd0 }, - { 0x11002c47, 0x1400ffd0 }, - { 0x11002c48, 0x1400ffd0 }, - { 0x11002c49, 0x1400ffd0 }, - { 0x11002c4a, 0x1400ffd0 }, - { 0x11002c4b, 0x1400ffd0 }, - { 0x11002c4c, 0x1400ffd0 }, - { 0x11002c4d, 0x1400ffd0 }, - { 0x11002c4e, 0x1400ffd0 }, - { 0x11002c4f, 0x1400ffd0 }, - { 0x11002c50, 0x1400ffd0 }, - { 0x11002c51, 0x1400ffd0 }, - { 0x11002c52, 0x1400ffd0 }, - { 0x11002c53, 0x1400ffd0 }, - { 0x11002c54, 0x1400ffd0 }, - { 0x11002c55, 0x1400ffd0 }, - { 0x11002c56, 0x1400ffd0 }, - { 0x11002c57, 0x1400ffd0 }, - { 0x11002c58, 0x1400ffd0 }, - { 0x11002c59, 0x1400ffd0 }, - { 0x11002c5a, 0x1400ffd0 }, - { 0x11002c5b, 0x1400ffd0 }, - { 0x11002c5c, 0x1400ffd0 }, - { 0x11002c5d, 0x1400ffd0 }, - { 0x11002c5e, 0x1400ffd0 }, - { 0x21002c60, 0x24000001 }, - { 0x21002c61, 0x1400ffff }, - { 0x21002c62, 0x2400d609 }, - { 0x21002c63, 0x2400f11a }, - { 0x21002c64, 0x2400d619 }, - { 0x21002c65, 0x1400d5d5 }, - { 0x21002c66, 0x1400d5d8 }, - { 0x21002c67, 0x24000001 }, - { 0x21002c68, 0x1400ffff }, - { 0x21002c69, 0x24000001 }, - { 0x21002c6a, 0x1400ffff }, - { 0x21002c6b, 0x24000001 }, - { 0x21002c6c, 0x1400ffff }, - { 0x21002c74, 0x14000000 }, - { 0x21002c75, 0x24000001 }, - { 0x21002c76, 0x1400ffff }, - { 0x21002c77, 0x14000000 }, - { 0x0a002c80, 0x24000001 }, - { 0x0a002c81, 0x1400ffff }, - { 0x0a002c82, 0x24000001 }, - { 0x0a002c83, 0x1400ffff }, - { 0x0a002c84, 0x24000001 }, - { 0x0a002c85, 0x1400ffff }, - { 0x0a002c86, 0x24000001 }, - { 0x0a002c87, 0x1400ffff }, - { 0x0a002c88, 0x24000001 }, - { 0x0a002c89, 0x1400ffff }, - { 0x0a002c8a, 0x24000001 }, - { 0x0a002c8b, 0x1400ffff }, - { 0x0a002c8c, 0x24000001 }, - { 0x0a002c8d, 0x1400ffff }, - { 0x0a002c8e, 0x24000001 }, - { 0x0a002c8f, 0x1400ffff }, - { 0x0a002c90, 0x24000001 }, - { 0x0a002c91, 0x1400ffff }, - { 0x0a002c92, 0x24000001 }, - { 0x0a002c93, 0x1400ffff }, - { 0x0a002c94, 0x24000001 }, - { 0x0a002c95, 0x1400ffff }, - { 0x0a002c96, 0x24000001 }, - { 0x0a002c97, 0x1400ffff }, - { 0x0a002c98, 0x24000001 }, - { 0x0a002c99, 0x1400ffff }, - { 0x0a002c9a, 0x24000001 }, - { 0x0a002c9b, 0x1400ffff }, - { 0x0a002c9c, 0x24000001 }, - { 0x0a002c9d, 0x1400ffff }, - { 0x0a002c9e, 0x24000001 }, - { 0x0a002c9f, 0x1400ffff }, - { 0x0a002ca0, 0x24000001 }, - { 0x0a002ca1, 0x1400ffff }, - { 0x0a002ca2, 0x24000001 }, - { 0x0a002ca3, 0x1400ffff }, - { 0x0a002ca4, 0x24000001 }, - { 0x0a002ca5, 0x1400ffff }, - { 0x0a002ca6, 0x24000001 }, - { 0x0a002ca7, 0x1400ffff }, - { 0x0a002ca8, 0x24000001 }, - { 0x0a002ca9, 0x1400ffff }, - { 0x0a002caa, 0x24000001 }, - { 0x0a002cab, 0x1400ffff }, - { 0x0a002cac, 0x24000001 }, - { 0x0a002cad, 0x1400ffff }, - { 0x0a002cae, 0x24000001 }, - { 0x0a002caf, 0x1400ffff }, - { 0x0a002cb0, 0x24000001 }, - { 0x0a002cb1, 0x1400ffff }, - { 0x0a002cb2, 0x24000001 }, - { 0x0a002cb3, 0x1400ffff }, - { 0x0a002cb4, 0x24000001 }, - { 0x0a002cb5, 0x1400ffff }, - { 0x0a002cb6, 0x24000001 }, - { 0x0a002cb7, 0x1400ffff }, - { 0x0a002cb8, 0x24000001 }, - { 0x0a002cb9, 0x1400ffff }, - { 0x0a002cba, 0x24000001 }, - { 0x0a002cbb, 0x1400ffff }, - { 0x0a002cbc, 0x24000001 }, - { 0x0a002cbd, 0x1400ffff }, - { 0x0a002cbe, 0x24000001 }, - { 0x0a002cbf, 0x1400ffff }, - { 0x0a002cc0, 0x24000001 }, - { 0x0a002cc1, 0x1400ffff }, - { 0x0a002cc2, 0x24000001 }, - { 0x0a002cc3, 0x1400ffff }, - { 0x0a002cc4, 0x24000001 }, - { 0x0a002cc5, 0x1400ffff }, - { 0x0a002cc6, 0x24000001 }, - { 0x0a002cc7, 0x1400ffff }, - { 0x0a002cc8, 0x24000001 }, - { 0x0a002cc9, 0x1400ffff }, - { 0x0a002cca, 0x24000001 }, - { 0x0a002ccb, 0x1400ffff }, - { 0x0a002ccc, 0x24000001 }, - { 0x0a002ccd, 0x1400ffff }, - { 0x0a002cce, 0x24000001 }, - { 0x0a002ccf, 0x1400ffff }, - { 0x0a002cd0, 0x24000001 }, - { 0x0a002cd1, 0x1400ffff }, - { 0x0a002cd2, 0x24000001 }, - { 0x0a002cd3, 0x1400ffff }, - { 0x0a002cd4, 0x24000001 }, - { 0x0a002cd5, 0x1400ffff }, - { 0x0a002cd6, 0x24000001 }, - { 0x0a002cd7, 0x1400ffff }, - { 0x0a002cd8, 0x24000001 }, - { 0x0a002cd9, 0x1400ffff }, - { 0x0a002cda, 0x24000001 }, - { 0x0a002cdb, 0x1400ffff }, - { 0x0a002cdc, 0x24000001 }, - { 0x0a002cdd, 0x1400ffff }, - { 0x0a002cde, 0x24000001 }, - { 0x0a002cdf, 0x1400ffff }, - { 0x0a002ce0, 0x24000001 }, - { 0x0a002ce1, 0x1400ffff }, - { 0x0a002ce2, 0x24000001 }, - { 0x0a002ce3, 0x1400ffff }, - { 0x0a002ce4, 0x14000000 }, - { 0x0a802ce5, 0x68000005 }, - { 0x0a802cf9, 0x54000003 }, - { 0x0a002cfd, 0x3c000000 }, - { 0x0a802cfe, 0x54000001 }, - { 0x10002d00, 0x1400e3a0 }, - { 0x10002d01, 0x1400e3a0 }, - { 0x10002d02, 0x1400e3a0 }, - { 0x10002d03, 0x1400e3a0 }, - { 0x10002d04, 0x1400e3a0 }, - { 0x10002d05, 0x1400e3a0 }, - { 0x10002d06, 0x1400e3a0 }, - { 0x10002d07, 0x1400e3a0 }, - { 0x10002d08, 0x1400e3a0 }, - { 0x10002d09, 0x1400e3a0 }, - { 0x10002d0a, 0x1400e3a0 }, - { 0x10002d0b, 0x1400e3a0 }, - { 0x10002d0c, 0x1400e3a0 }, - { 0x10002d0d, 0x1400e3a0 }, - { 0x10002d0e, 0x1400e3a0 }, - { 0x10002d0f, 0x1400e3a0 }, - { 0x10002d10, 0x1400e3a0 }, - { 0x10002d11, 0x1400e3a0 }, - { 0x10002d12, 0x1400e3a0 }, - { 0x10002d13, 0x1400e3a0 }, - { 0x10002d14, 0x1400e3a0 }, - { 0x10002d15, 0x1400e3a0 }, - { 0x10002d16, 0x1400e3a0 }, - { 0x10002d17, 0x1400e3a0 }, - { 0x10002d18, 0x1400e3a0 }, - { 0x10002d19, 0x1400e3a0 }, - { 0x10002d1a, 0x1400e3a0 }, - { 0x10002d1b, 0x1400e3a0 }, - { 0x10002d1c, 0x1400e3a0 }, - { 0x10002d1d, 0x1400e3a0 }, - { 0x10002d1e, 0x1400e3a0 }, - { 0x10002d1f, 0x1400e3a0 }, - { 0x10002d20, 0x1400e3a0 }, - { 0x10002d21, 0x1400e3a0 }, - { 0x10002d22, 0x1400e3a0 }, - { 0x10002d23, 0x1400e3a0 }, - { 0x10002d24, 0x1400e3a0 }, - { 0x10002d25, 0x1400e3a0 }, - { 0x3a802d30, 0x1c000035 }, - { 0x3a002d6f, 0x18000000 }, - { 0x0f802d80, 0x1c000016 }, - { 0x0f802da0, 0x1c000006 }, - { 0x0f802da8, 0x1c000006 }, - { 0x0f802db0, 0x1c000006 }, - { 0x0f802db8, 0x1c000006 }, - { 0x0f802dc0, 0x1c000006 }, - { 0x0f802dc8, 0x1c000006 }, - { 0x0f802dd0, 0x1c000006 }, - { 0x0f802dd8, 0x1c000006 }, - { 0x09802e00, 0x54000001 }, - { 0x09002e02, 0x50000000 }, - { 0x09002e03, 0x4c000000 }, - { 0x09002e04, 0x50000000 }, - { 0x09002e05, 0x4c000000 }, - { 0x09802e06, 0x54000002 }, - { 0x09002e09, 0x50000000 }, - { 0x09002e0a, 0x4c000000 }, - { 0x09002e0b, 0x54000000 }, - { 0x09002e0c, 0x50000000 }, - { 0x09002e0d, 0x4c000000 }, - { 0x09802e0e, 0x54000008 }, - { 0x09002e17, 0x44000000 }, - { 0x09002e1c, 0x50000000 }, - { 0x09002e1d, 0x4c000000 }, - { 0x16802e80, 0x68000019 }, - { 0x16802e9b, 0x68000058 }, - { 0x16802f00, 0x680000d5 }, - { 0x09802ff0, 0x6800000b }, - { 0x09003000, 0x74000000 }, - { 0x09803001, 0x54000002 }, - { 0x09003004, 0x68000000 }, - { 0x16003005, 0x18000000 }, - { 0x09003006, 0x1c000000 }, - { 0x16003007, 0x38000000 }, - { 0x09003008, 0x58000000 }, - { 0x09003009, 0x48000000 }, - { 0x0900300a, 0x58000000 }, - { 0x0900300b, 0x48000000 }, - { 0x0900300c, 0x58000000 }, - { 0x0900300d, 0x48000000 }, - { 0x0900300e, 0x58000000 }, - { 0x0900300f, 0x48000000 }, - { 0x09003010, 0x58000000 }, - { 0x09003011, 0x48000000 }, - { 0x09803012, 0x68000001 }, - { 0x09003014, 0x58000000 }, - { 0x09003015, 0x48000000 }, - { 0x09003016, 0x58000000 }, - { 0x09003017, 0x48000000 }, - { 0x09003018, 0x58000000 }, - { 0x09003019, 0x48000000 }, - { 0x0900301a, 0x58000000 }, - { 0x0900301b, 0x48000000 }, - { 0x0900301c, 0x44000000 }, - { 0x0900301d, 0x58000000 }, - { 0x0980301e, 0x48000001 }, - { 0x09003020, 0x68000000 }, - { 0x16803021, 0x38000008 }, - { 0x1b80302a, 0x30000005 }, - { 0x09003030, 0x44000000 }, - { 0x09803031, 0x18000004 }, - { 0x09803036, 0x68000001 }, - { 0x16803038, 0x38000002 }, - { 0x1600303b, 0x18000000 }, - { 0x0900303c, 0x1c000000 }, - { 0x0900303d, 0x54000000 }, - { 0x0980303e, 0x68000001 }, - { 0x1a803041, 0x1c000055 }, - { 0x1b803099, 0x30000001 }, - { 0x0980309b, 0x60000001 }, - { 0x1a80309d, 0x18000001 }, - { 0x1a00309f, 0x1c000000 }, - { 0x090030a0, 0x44000000 }, - { 0x1d8030a1, 0x1c000059 }, - { 0x090030fb, 0x54000000 }, - { 0x090030fc, 0x18000000 }, - { 0x1d8030fd, 0x18000001 }, - { 0x1d0030ff, 0x1c000000 }, - { 0x03803105, 0x1c000027 }, - { 0x17803131, 0x1c00005d }, - { 0x09803190, 0x68000001 }, - { 0x09803192, 0x3c000003 }, - { 0x09803196, 0x68000009 }, - { 0x038031a0, 0x1c000017 }, - { 0x098031c0, 0x6800000f }, - { 0x1d8031f0, 0x1c00000f }, - { 0x17803200, 0x6800001e }, - { 0x09803220, 0x3c000009 }, - { 0x0980322a, 0x68000019 }, - { 0x09003250, 0x68000000 }, - { 0x09803251, 0x3c00000e }, - { 0x17803260, 0x6800001d }, - { 0x0980327e, 0x68000001 }, - { 0x09803280, 0x3c000009 }, - { 0x0980328a, 0x68000026 }, - { 0x098032b1, 0x3c00000e }, - { 0x098032c0, 0x6800003e }, - { 0x09803300, 0x680000ff }, - { 0x16803400, 0x1c0019b5 }, - { 0x09804dc0, 0x6800003f }, - { 0x16804e00, 0x1c0051bb }, - { 0x3c80a000, 0x1c000014 }, - { 0x3c00a015, 0x18000000 }, - { 0x3c80a016, 0x1c000476 }, - { 0x3c80a490, 0x68000036 }, - { 0x0980a700, 0x60000016 }, - { 0x0980a717, 0x18000003 }, - { 0x0980a720, 0x60000001 }, - { 0x3080a800, 0x1c000001 }, - { 0x3000a802, 0x28000000 }, - { 0x3080a803, 0x1c000002 }, - { 0x3000a806, 0x30000000 }, - { 0x3080a807, 0x1c000003 }, - { 0x3000a80b, 0x30000000 }, - { 0x3080a80c, 0x1c000016 }, - { 0x3080a823, 0x28000001 }, - { 0x3080a825, 0x30000001 }, - { 0x3000a827, 0x28000000 }, - { 0x3080a828, 0x68000003 }, - { 0x4080a840, 0x1c000033 }, - { 0x4080a874, 0x54000003 }, - { 0x1780ac00, 0x1c002ba3 }, - { 0x0980d800, 0x1000037f }, - { 0x0980db80, 0x1000007f }, - { 0x0980dc00, 0x100003ff }, - { 0x0980e000, 0x0c0018ff }, - { 0x1680f900, 0x1c00012d }, - { 0x1680fa30, 0x1c00003a }, - { 0x1680fa70, 0x1c000069 }, - { 0x2180fb00, 0x14000006 }, - { 0x0180fb13, 0x14000004 }, - { 0x1900fb1d, 0x1c000000 }, - { 0x1900fb1e, 0x30000000 }, - { 0x1980fb1f, 0x1c000009 }, - { 0x1900fb29, 0x64000000 }, - { 0x1980fb2a, 0x1c00000c }, - { 0x1980fb38, 0x1c000004 }, - { 0x1900fb3e, 0x1c000000 }, - { 0x1980fb40, 0x1c000001 }, - { 0x1980fb43, 0x1c000001 }, - { 0x1980fb46, 0x1c000009 }, - { 0x0080fb50, 0x1c000061 }, - { 0x0080fbd3, 0x1c00016a }, - { 0x0900fd3e, 0x58000000 }, - { 0x0900fd3f, 0x48000000 }, - { 0x0080fd50, 0x1c00003f }, - { 0x0080fd92, 0x1c000035 }, - { 0x0080fdf0, 0x1c00000b }, - { 0x0000fdfc, 0x5c000000 }, - { 0x0900fdfd, 0x68000000 }, - { 0x1b80fe00, 0x3000000f }, - { 0x0980fe10, 0x54000006 }, - { 0x0900fe17, 0x58000000 }, - { 0x0900fe18, 0x48000000 }, - { 0x0900fe19, 0x54000000 }, - { 0x1b80fe20, 0x30000003 }, - { 0x0900fe30, 0x54000000 }, - { 0x0980fe31, 0x44000001 }, - { 0x0980fe33, 0x40000001 }, - { 0x0900fe35, 0x58000000 }, - { 0x0900fe36, 0x48000000 }, - { 0x0900fe37, 0x58000000 }, - { 0x0900fe38, 0x48000000 }, - { 0x0900fe39, 0x58000000 }, - { 0x0900fe3a, 0x48000000 }, - { 0x0900fe3b, 0x58000000 }, - { 0x0900fe3c, 0x48000000 }, - { 0x0900fe3d, 0x58000000 }, - { 0x0900fe3e, 0x48000000 }, - { 0x0900fe3f, 0x58000000 }, - { 0x0900fe40, 0x48000000 }, - { 0x0900fe41, 0x58000000 }, - { 0x0900fe42, 0x48000000 }, - { 0x0900fe43, 0x58000000 }, - { 0x0900fe44, 0x48000000 }, - { 0x0980fe45, 0x54000001 }, - { 0x0900fe47, 0x58000000 }, - { 0x0900fe48, 0x48000000 }, - { 0x0980fe49, 0x54000003 }, - { 0x0980fe4d, 0x40000002 }, - { 0x0980fe50, 0x54000002 }, - { 0x0980fe54, 0x54000003 }, - { 0x0900fe58, 0x44000000 }, - { 0x0900fe59, 0x58000000 }, - { 0x0900fe5a, 0x48000000 }, - { 0x0900fe5b, 0x58000000 }, - { 0x0900fe5c, 0x48000000 }, - { 0x0900fe5d, 0x58000000 }, - { 0x0900fe5e, 0x48000000 }, - { 0x0980fe5f, 0x54000002 }, - { 0x0900fe62, 0x64000000 }, - { 0x0900fe63, 0x44000000 }, - { 0x0980fe64, 0x64000002 }, - { 0x0900fe68, 0x54000000 }, - { 0x0900fe69, 0x5c000000 }, - { 0x0980fe6a, 0x54000001 }, - { 0x0080fe70, 0x1c000004 }, - { 0x0080fe76, 0x1c000086 }, - { 0x0900feff, 0x04000000 }, - { 0x0980ff01, 0x54000002 }, - { 0x0900ff04, 0x5c000000 }, - { 0x0980ff05, 0x54000002 }, - { 0x0900ff08, 0x58000000 }, - { 0x0900ff09, 0x48000000 }, - { 0x0900ff0a, 0x54000000 }, - { 0x0900ff0b, 0x64000000 }, - { 0x0900ff0c, 0x54000000 }, - { 0x0900ff0d, 0x44000000 }, - { 0x0980ff0e, 0x54000001 }, - { 0x0980ff10, 0x34000009 }, - { 0x0980ff1a, 0x54000001 }, - { 0x0980ff1c, 0x64000002 }, - { 0x0980ff1f, 0x54000001 }, - { 0x2100ff21, 0x24000020 }, - { 0x2100ff22, 0x24000020 }, - { 0x2100ff23, 0x24000020 }, - { 0x2100ff24, 0x24000020 }, - { 0x2100ff25, 0x24000020 }, - { 0x2100ff26, 0x24000020 }, - { 0x2100ff27, 0x24000020 }, - { 0x2100ff28, 0x24000020 }, - { 0x2100ff29, 0x24000020 }, - { 0x2100ff2a, 0x24000020 }, - { 0x2100ff2b, 0x24000020 }, - { 0x2100ff2c, 0x24000020 }, - { 0x2100ff2d, 0x24000020 }, - { 0x2100ff2e, 0x24000020 }, - { 0x2100ff2f, 0x24000020 }, - { 0x2100ff30, 0x24000020 }, - { 0x2100ff31, 0x24000020 }, - { 0x2100ff32, 0x24000020 }, - { 0x2100ff33, 0x24000020 }, - { 0x2100ff34, 0x24000020 }, - { 0x2100ff35, 0x24000020 }, - { 0x2100ff36, 0x24000020 }, - { 0x2100ff37, 0x24000020 }, - { 0x2100ff38, 0x24000020 }, - { 0x2100ff39, 0x24000020 }, - { 0x2100ff3a, 0x24000020 }, - { 0x0900ff3b, 0x58000000 }, - { 0x0900ff3c, 0x54000000 }, - { 0x0900ff3d, 0x48000000 }, - { 0x0900ff3e, 0x60000000 }, - { 0x0900ff3f, 0x40000000 }, - { 0x0900ff40, 0x60000000 }, - { 0x2100ff41, 0x1400ffe0 }, - { 0x2100ff42, 0x1400ffe0 }, - { 0x2100ff43, 0x1400ffe0 }, - { 0x2100ff44, 0x1400ffe0 }, - { 0x2100ff45, 0x1400ffe0 }, - { 0x2100ff46, 0x1400ffe0 }, - { 0x2100ff47, 0x1400ffe0 }, - { 0x2100ff48, 0x1400ffe0 }, - { 0x2100ff49, 0x1400ffe0 }, - { 0x2100ff4a, 0x1400ffe0 }, - { 0x2100ff4b, 0x1400ffe0 }, - { 0x2100ff4c, 0x1400ffe0 }, - { 0x2100ff4d, 0x1400ffe0 }, - { 0x2100ff4e, 0x1400ffe0 }, - { 0x2100ff4f, 0x1400ffe0 }, - { 0x2100ff50, 0x1400ffe0 }, - { 0x2100ff51, 0x1400ffe0 }, - { 0x2100ff52, 0x1400ffe0 }, - { 0x2100ff53, 0x1400ffe0 }, - { 0x2100ff54, 0x1400ffe0 }, - { 0x2100ff55, 0x1400ffe0 }, - { 0x2100ff56, 0x1400ffe0 }, - { 0x2100ff57, 0x1400ffe0 }, - { 0x2100ff58, 0x1400ffe0 }, - { 0x2100ff59, 0x1400ffe0 }, - { 0x2100ff5a, 0x1400ffe0 }, - { 0x0900ff5b, 0x58000000 }, - { 0x0900ff5c, 0x64000000 }, - { 0x0900ff5d, 0x48000000 }, - { 0x0900ff5e, 0x64000000 }, - { 0x0900ff5f, 0x58000000 }, - { 0x0900ff60, 0x48000000 }, - { 0x0900ff61, 0x54000000 }, - { 0x0900ff62, 0x58000000 }, - { 0x0900ff63, 0x48000000 }, - { 0x0980ff64, 0x54000001 }, - { 0x1d80ff66, 0x1c000009 }, - { 0x0900ff70, 0x18000000 }, - { 0x1d80ff71, 0x1c00002c }, - { 0x0980ff9e, 0x18000001 }, - { 0x1780ffa0, 0x1c00001e }, - { 0x1780ffc2, 0x1c000005 }, - { 0x1780ffca, 0x1c000005 }, - { 0x1780ffd2, 0x1c000005 }, - { 0x1780ffda, 0x1c000002 }, - { 0x0980ffe0, 0x5c000001 }, - { 0x0900ffe2, 0x64000000 }, - { 0x0900ffe3, 0x60000000 }, - { 0x0900ffe4, 0x68000000 }, - { 0x0980ffe5, 0x5c000001 }, - { 0x0900ffe8, 0x68000000 }, - { 0x0980ffe9, 0x64000003 }, - { 0x0980ffed, 0x68000001 }, - { 0x0980fff9, 0x04000002 }, - { 0x0980fffc, 0x68000001 }, - { 0x23810000, 0x1c00000b }, - { 0x2381000d, 0x1c000019 }, - { 0x23810028, 0x1c000012 }, - { 0x2381003c, 0x1c000001 }, - { 0x2381003f, 0x1c00000e }, - { 0x23810050, 0x1c00000d }, - { 0x23810080, 0x1c00007a }, - { 0x09810100, 0x54000001 }, - { 0x09010102, 0x68000000 }, - { 0x09810107, 0x3c00002c }, - { 0x09810137, 0x68000008 }, - { 0x13810140, 0x38000034 }, - { 0x13810175, 0x3c000003 }, - { 0x13810179, 0x68000010 }, - { 0x1301018a, 0x3c000000 }, - { 0x29810300, 0x1c00001e }, - { 0x29810320, 0x3c000003 }, - { 0x12810330, 0x1c000010 }, - { 0x12010341, 0x38000000 }, - { 0x12810342, 0x1c000007 }, - { 0x1201034a, 0x38000000 }, - { 0x3b810380, 0x1c00001d }, - { 0x3b01039f, 0x54000000 }, - { 0x2a8103a0, 0x1c000023 }, - { 0x2a8103c8, 0x1c000007 }, - { 0x2a0103d0, 0x54000000 }, - { 0x2a8103d1, 0x38000004 }, - { 0x0d010400, 0x24000028 }, - { 0x0d010401, 0x24000028 }, - { 0x0d010402, 0x24000028 }, - { 0x0d010403, 0x24000028 }, - { 0x0d010404, 0x24000028 }, - { 0x0d010405, 0x24000028 }, - { 0x0d010406, 0x24000028 }, - { 0x0d010407, 0x24000028 }, - { 0x0d010408, 0x24000028 }, - { 0x0d010409, 0x24000028 }, - { 0x0d01040a, 0x24000028 }, - { 0x0d01040b, 0x24000028 }, - { 0x0d01040c, 0x24000028 }, - { 0x0d01040d, 0x24000028 }, - { 0x0d01040e, 0x24000028 }, - { 0x0d01040f, 0x24000028 }, - { 0x0d010410, 0x24000028 }, - { 0x0d010411, 0x24000028 }, - { 0x0d010412, 0x24000028 }, - { 0x0d010413, 0x24000028 }, - { 0x0d010414, 0x24000028 }, - { 0x0d010415, 0x24000028 }, - { 0x0d010416, 0x24000028 }, - { 0x0d010417, 0x24000028 }, - { 0x0d010418, 0x24000028 }, - { 0x0d010419, 0x24000028 }, - { 0x0d01041a, 0x24000028 }, - { 0x0d01041b, 0x24000028 }, - { 0x0d01041c, 0x24000028 }, - { 0x0d01041d, 0x24000028 }, - { 0x0d01041e, 0x24000028 }, - { 0x0d01041f, 0x24000028 }, - { 0x0d010420, 0x24000028 }, - { 0x0d010421, 0x24000028 }, - { 0x0d010422, 0x24000028 }, - { 0x0d010423, 0x24000028 }, - { 0x0d010424, 0x24000028 }, - { 0x0d010425, 0x24000028 }, - { 0x0d010426, 0x24000028 }, - { 0x0d010427, 0x24000028 }, - { 0x0d010428, 0x1400ffd8 }, - { 0x0d010429, 0x1400ffd8 }, - { 0x0d01042a, 0x1400ffd8 }, - { 0x0d01042b, 0x1400ffd8 }, - { 0x0d01042c, 0x1400ffd8 }, - { 0x0d01042d, 0x1400ffd8 }, - { 0x0d01042e, 0x1400ffd8 }, - { 0x0d01042f, 0x1400ffd8 }, - { 0x0d010430, 0x1400ffd8 }, - { 0x0d010431, 0x1400ffd8 }, - { 0x0d010432, 0x1400ffd8 }, - { 0x0d010433, 0x1400ffd8 }, - { 0x0d010434, 0x1400ffd8 }, - { 0x0d010435, 0x1400ffd8 }, - { 0x0d010436, 0x1400ffd8 }, - { 0x0d010437, 0x1400ffd8 }, - { 0x0d010438, 0x1400ffd8 }, - { 0x0d010439, 0x1400ffd8 }, - { 0x0d01043a, 0x1400ffd8 }, - { 0x0d01043b, 0x1400ffd8 }, - { 0x0d01043c, 0x1400ffd8 }, - { 0x0d01043d, 0x1400ffd8 }, - { 0x0d01043e, 0x1400ffd8 }, - { 0x0d01043f, 0x1400ffd8 }, - { 0x0d010440, 0x1400ffd8 }, - { 0x0d010441, 0x1400ffd8 }, - { 0x0d010442, 0x1400ffd8 }, - { 0x0d010443, 0x1400ffd8 }, - { 0x0d010444, 0x1400ffd8 }, - { 0x0d010445, 0x1400ffd8 }, - { 0x0d010446, 0x1400ffd8 }, - { 0x0d010447, 0x1400ffd8 }, - { 0x0d010448, 0x1400ffd8 }, - { 0x0d010449, 0x1400ffd8 }, - { 0x0d01044a, 0x1400ffd8 }, - { 0x0d01044b, 0x1400ffd8 }, - { 0x0d01044c, 0x1400ffd8 }, - { 0x0d01044d, 0x1400ffd8 }, - { 0x0d01044e, 0x1400ffd8 }, - { 0x0d01044f, 0x1400ffd8 }, - { 0x2e810450, 0x1c00002f }, - { 0x2c810480, 0x1c00001d }, - { 0x2c8104a0, 0x34000009 }, - { 0x0b810800, 0x1c000005 }, - { 0x0b010808, 0x1c000000 }, - { 0x0b81080a, 0x1c00002b }, - { 0x0b810837, 0x1c000001 }, - { 0x0b01083c, 0x1c000000 }, - { 0x0b01083f, 0x1c000000 }, - { 0x41810900, 0x1c000015 }, - { 0x41810916, 0x3c000003 }, - { 0x4101091f, 0x54000000 }, - { 0x1e010a00, 0x1c000000 }, - { 0x1e810a01, 0x30000002 }, - { 0x1e810a05, 0x30000001 }, - { 0x1e810a0c, 0x30000003 }, - { 0x1e810a10, 0x1c000003 }, - { 0x1e810a15, 0x1c000002 }, - { 0x1e810a19, 0x1c00001a }, - { 0x1e810a38, 0x30000002 }, - { 0x1e010a3f, 0x30000000 }, - { 0x1e810a40, 0x3c000007 }, - { 0x1e810a50, 0x54000008 }, - { 0x3e812000, 0x1c00036e }, - { 0x3e812400, 0x38000062 }, - { 0x3e812470, 0x54000003 }, - { 0x0981d000, 0x680000f5 }, - { 0x0981d100, 0x68000026 }, - { 0x0981d12a, 0x6800003a }, - { 0x0981d165, 0x28000001 }, - { 0x1b81d167, 0x30000002 }, - { 0x0981d16a, 0x68000002 }, - { 0x0981d16d, 0x28000005 }, - { 0x0981d173, 0x04000007 }, - { 0x1b81d17b, 0x30000007 }, - { 0x0981d183, 0x68000001 }, - { 0x1b81d185, 0x30000006 }, - { 0x0981d18c, 0x6800001d }, - { 0x1b81d1aa, 0x30000003 }, - { 0x0981d1ae, 0x6800002f }, - { 0x1381d200, 0x68000041 }, - { 0x1381d242, 0x30000002 }, - { 0x1301d245, 0x68000000 }, - { 0x0981d300, 0x68000056 }, - { 0x0981d360, 0x3c000011 }, - { 0x0981d400, 0x24000019 }, - { 0x0981d41a, 0x14000019 }, - { 0x0981d434, 0x24000019 }, - { 0x0981d44e, 0x14000006 }, - { 0x0981d456, 0x14000011 }, - { 0x0981d468, 0x24000019 }, - { 0x0981d482, 0x14000019 }, - { 0x0901d49c, 0x24000000 }, - { 0x0981d49e, 0x24000001 }, - { 0x0901d4a2, 0x24000000 }, - { 0x0981d4a5, 0x24000001 }, - { 0x0981d4a9, 0x24000003 }, - { 0x0981d4ae, 0x24000007 }, - { 0x0981d4b6, 0x14000003 }, - { 0x0901d4bb, 0x14000000 }, - { 0x0981d4bd, 0x14000006 }, - { 0x0981d4c5, 0x1400000a }, - { 0x0981d4d0, 0x24000019 }, - { 0x0981d4ea, 0x14000019 }, - { 0x0981d504, 0x24000001 }, - { 0x0981d507, 0x24000003 }, - { 0x0981d50d, 0x24000007 }, - { 0x0981d516, 0x24000006 }, - { 0x0981d51e, 0x14000019 }, - { 0x0981d538, 0x24000001 }, - { 0x0981d53b, 0x24000003 }, - { 0x0981d540, 0x24000004 }, - { 0x0901d546, 0x24000000 }, - { 0x0981d54a, 0x24000006 }, - { 0x0981d552, 0x14000019 }, - { 0x0981d56c, 0x24000019 }, - { 0x0981d586, 0x14000019 }, - { 0x0981d5a0, 0x24000019 }, - { 0x0981d5ba, 0x14000019 }, - { 0x0981d5d4, 0x24000019 }, - { 0x0981d5ee, 0x14000019 }, - { 0x0981d608, 0x24000019 }, - { 0x0981d622, 0x14000019 }, - { 0x0981d63c, 0x24000019 }, - { 0x0981d656, 0x14000019 }, - { 0x0981d670, 0x24000019 }, - { 0x0981d68a, 0x1400001b }, - { 0x0981d6a8, 0x24000018 }, - { 0x0901d6c1, 0x64000000 }, - { 0x0981d6c2, 0x14000018 }, - { 0x0901d6db, 0x64000000 }, - { 0x0981d6dc, 0x14000005 }, - { 0x0981d6e2, 0x24000018 }, - { 0x0901d6fb, 0x64000000 }, - { 0x0981d6fc, 0x14000018 }, - { 0x0901d715, 0x64000000 }, - { 0x0981d716, 0x14000005 }, - { 0x0981d71c, 0x24000018 }, - { 0x0901d735, 0x64000000 }, - { 0x0981d736, 0x14000018 }, - { 0x0901d74f, 0x64000000 }, - { 0x0981d750, 0x14000005 }, - { 0x0981d756, 0x24000018 }, - { 0x0901d76f, 0x64000000 }, - { 0x0981d770, 0x14000018 }, - { 0x0901d789, 0x64000000 }, - { 0x0981d78a, 0x14000005 }, - { 0x0981d790, 0x24000018 }, - { 0x0901d7a9, 0x64000000 }, - { 0x0981d7aa, 0x14000018 }, - { 0x0901d7c3, 0x64000000 }, - { 0x0981d7c4, 0x14000005 }, - { 0x0901d7ca, 0x24000000 }, - { 0x0901d7cb, 0x14000000 }, - { 0x0981d7ce, 0x34000031 }, - { 0x16820000, 0x1c00a6d6 }, - { 0x1682f800, 0x1c00021d }, - { 0x090e0001, 0x04000000 }, - { 0x098e0020, 0x0400005f }, - { 0x1b8e0100, 0x300000ef }, - { 0x098f0000, 0x0c00fffd }, - { 0x09900000, 0x0c00fffd }, -}; diff --git a/regex-pcre-builtin.cabal b/regex-pcre-builtin.cabal --- a/regex-pcre-builtin.cabal +++ b/regex-pcre-builtin.cabal @@ -1,40 +1,46 @@ Name: regex-pcre-builtin -- Keep the Version below in sync with ./Text/Regex/PCRE.hs value getVersion_Text_Regex_PCRE :: Version -Version: 0.94.2.1.7.7 +Version: 0.94.4.0.8.31 Cabal-Version: >=1.2.3 License: BSD3 License-File: LICENSE -Copyright: Copyright (c) 2006, Christopher Kuklewicz -Author: Christopher Kuklewicz -Maintainer: TextRegexLazy@personal.mightyreason.com +Copyright: See LICENSE file +Author: Audrey Tang +Maintainer: audreyt@audreyt.org Stability: Seems to work, passes a few tests -Homepage: http://sourceforge.net/projects/lazy-regex -Package-URL: http://darcs.haskell.org/packages/regex-unstable/regex-pcre/ +Homepage: http://hackage.haskell.org/package/regex-pcre +Package-URL: http://code.haskell.org/regex-pcre/ Synopsis: Replaces/Enhances Text.Regex -Description: The PCRE backend to accompany regex-base, see www.pcre.org -Category: Text, Pugs +Description: The PCRE backend to accompany regex-base, with bundled code from www.pcre.org +Category: Text Tested-With: GHC Build-Type: Simple Extra-source-files: - pcre/config.h pcre/pcre_globals.c pcre/pcre_study.c - pcre/pcrecpparg.h pcre/dftables.c pcre/pcre_info.c - pcre/pcre_tables.c pcre/pcredemo.c pcre/pcre.h - pcre/pcre_internal.h pcre/pcre_try_flipped.c pcre/pcregrep.c - pcre/pcre_compile.c pcre/pcre_maketables.c pcre/pcre_ucp_searchfuncs.c - pcre/pcreposix.c pcre/pcre_config.c pcre/pcre_newline.c - pcre/pcre_valid_utf8.c pcre/pcreposix.h - pcre/pcre_dfa_exec.c pcre/pcre_ord2utf8.c pcre/pcre_version.c pcre/pcretest.c - pcre/pcre_exec.c pcre/pcre_refcount.c pcre/pcre_xclass.c pcre/ucp.h - pcre/pcre_fullinfo.c pcre/pcre_scanner.h pcre/pcrecpp.h pcre/ucpinternal.h - pcre/pcre_get.c pcre/pcre_stringpiece.h pcre/pcrecpp_internal.h pcre/ucptable.h + pcre/config.h pcre/pcre.h pcre/pcre_byte_order.c pcre/pcre_compile.c pcre/pcre_config.c pcre/pcre_dfa_exec.c pcre/pcre_exec.c pcre/pcre_fullinfo.c pcre/pcre_get.c pcre/pcre_globals.c pcre/pcre_internal.h pcre/pcre_jit_compile.c pcre/pcre_jit_test.c pcre/pcre_maketables.c pcre/pcre_newline.c pcre/pcre_ord2utf8.c pcre/pcre_printint.c pcre/pcre_refcount.c pcre/pcre_scanner.h pcre/pcre_string_utils.c pcre/pcre_study.c pcre/pcre_tables.c pcre/pcre_ucd.c pcre/pcre_valid_utf8.c pcre/pcre_version.c pcre/pcre_xclass.c pcre/pcrecpp.h pcre/pcrecpp_internal.h pcre/pcreposix.h pcre/ucp.h +flag newBase + description: Choose base >= 4 + default: True +flag splitBase + description: Choose the new smaller, split-up base package. + default: True library - Build-Depends: base >= 3.0, regex-base >= 0.93, array, containers, bytestring - Extensions: MultiParamTypeClasses, FunctionalDependencies, CPP, ForeignFunctionInterface, PatternSignatures, GeneralizedNewtypeDeriving, FlexibleContexts, TypeSynonymInstances, FlexibleInstances - - -- Data-Files: - -- Extra-Source-Files: - -- Extra-Tmp-Files: + if flag(newBase) + Build-Depends: base >= 4 && < 5, regex-base >= 0.93, array, containers, bytestring + -- Need the next symbol for using CPP to get Data.ByteString.Base|Unsafe in + -- ./Text/Regex/Posix/ByteString.hs and ./Text/Regex/Posix/ByteString/Lazy.hs + CPP-Options: "-DSPLIT_BASE=1" + Extensions: MultiParamTypeClasses, FunctionalDependencies, CPP, ForeignFunctionInterface, ScopedTypeVariables, GeneralizedNewtypeDeriving, FlexibleContexts, TypeSynonymInstances, FlexibleInstances + else + if flag(splitBase) + Build-Depends: base >= 3.0, regex-base >= 0.93, array, containers, bytestring + -- Need the next symbol for using CPP to get Data.ByteString.Base|Unsafe in + -- ./Text/Regex/Posix/ByteString.hs and ./Text/Regex/Posix/ByteString/Lazy.hs + CPP-Options: "-DSPLIT_BASE=1" + Extensions: MultiParamTypeClasses, FunctionalDependencies, CPP, ForeignFunctionInterface, ScopedTypeVariables, GeneralizedNewtypeDeriving, FlexibleContexts, TypeSynonymInstances, FlexibleInstances + else + Build-Depends: base < 3.0, regex-base >= 0.93 + Extensions: MultiParamTypeClasses, FunctionalDependencies, CPP Exposed-Modules: Text.Regex.PCRE Text.Regex.PCRE.Wrap Text.Regex.PCRE.String @@ -47,7 +53,7 @@ -- The CPP is for using -DSPLIT_BASE=1 to get Data.ByteString.Base|Unsafe -- And the CPP is for using -DHAVE_PCRE_H to get the local posix library -- GHC-Options: -Wall -Werror -O2 - -- GHC-Options: -Wall -O2 + GHC-Options: -Wall -O2 -- GHC-Options: -Wall -ddump-minimal-imports -- GHC-Prof-Options: -- Hugs-Options: @@ -57,19 +63,13 @@ -- LD-Options: -- Frameworks: -- The only reason to NOT define -DHAVE_PCRE_H is if you are on - -- a platform without a regex library but want to compile this package + -- a plotform without a regex library but want to compile this package -- anyway. The resulting regex-posix will exist, but throw errors. - CC-Options: -DHAVE_PCRE_H + CC-Options: -DHAVE_PCRE_H -DHAVE_CONFIG_H + Extra-Libraries: pcre --Include-Dirs: /opt/local/include --Extra-Lib-Dirs: /opt/local/lib Include-dirs: pcre - Includes: pcre.h + Includes: pcre.h config.h C-sources: - pcre/pcre_compile.c pcre/pcre_config.c - pcre/pcre_dfa_exec.c pcre/pcre_exec.c pcre/pcre_fullinfo.c pcre/pcre_get.c - pcre/pcre_globals.c pcre/pcre_info.c pcre/pcre_maketables.c - pcre/pcre_newline.c pcre/pcre_ord2utf8.c pcre/pcre_refcount.c - pcre/pcre_study.c pcre/pcre_tables.c pcre/pcre_try_flipped.c - pcre/pcre_ucp_searchfuncs.c pcre/pcre_valid_utf8.c pcre/pcre_version.c - pcre/pcre_xclass.c pcre/pcre_chartables.c - + pcre/pcre_byte_order.c pcre/pcre_compile.c pcre/pcre_config.c pcre/pcre_dfa_exec.c pcre/pcre_exec.c pcre/pcre_fullinfo.c pcre/pcre_get.c pcre/pcre_globals.c pcre/pcre_jit_compile.c pcre/pcre_jit_test.c pcre/pcre_maketables.c pcre/pcre_newline.c pcre/pcre_ord2utf8.c pcre/pcre_printint.c pcre/pcre_refcount.c pcre/pcre_string_utils.c pcre/pcre_study.c pcre/pcre_tables.c pcre/pcre_ucd.c pcre/pcre_valid_utf8.c pcre/pcre_version.c pcre/pcre_xclass.c