diff --git a/LICENSE b/LICENSE new file mode 100644 --- /dev/null +++ b/LICENSE @@ -0,0 +1,12 @@ +This modile is under this "3 clause" BSD license: + +Copyright (c) 2007, 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: + + * 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. + * The names of the contributors may not 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. diff --git a/Setup.hs b/Setup.hs new file mode 100644 --- /dev/null +++ b/Setup.hs @@ -0,0 +1,11 @@ +#!/usr/bin/env runhaskell + +-- I usually compile this with "ghc --make -o setup Setup.hs" + +import Distribution.Simple(defaultMain) +main = do putStrLn msg + defaultMain + +msg = "\nregex-pcre needs to compile against the libpcre libary from http://www.pcre.org/\n\ + \You might also need to edit the end of the regex-pcre.cabal file to point at\n\ + \the directories where libpcre 'include' and 'lib' have been installed.\n" diff --git a/Text/Regex/PCRE.hs b/Text/Regex/PCRE.hs new file mode 100644 --- /dev/null +++ b/Text/Regex/PCRE.hs @@ -0,0 +1,62 @@ +{-# OPTIONS_GHC -fno-warn-unused-imports #-} +{-| +The "Text.Regex.PCRE" module provides a backend for regular +expressions. If you import this along with other backends, then +you should do so with qualified imports, perhaps renamed for +convenience. + +You will need to have libpcre, from www.pcre.org, to use +this module. The haskell must be compiled with -DHAVE_PCRE_H and +linked with pcre. This is the default in the cabal file. + +If you do not compile with HAVE_PCRE_H then the functions will still +exist, but using them will create a run time error. You can test for +the existance of PCRE by checking 'getVersion' which is 'Nothing' if +not compiled with PCRE or 'Just' 'String' if PCRE is present. + +Using the provided 'CompOption' and 'ExecOption' values and if +'configUTF8' is True, then you might be able to send UTF8 encoded +ByteStrings to PCRE and get sensible results. This is currently +untested. + +The regular expression can be provided as a 'ByteString', but it will +be copied and a NUL byte appended to make a 'CString' unless such a +byte is already present. Thus the regular expression cannot contain +an explicit NUL byte. The search string is passed as a 'CStringLen' +and may contain NUL bytes and does not need to end in a NUL +byte. 'ByteString's are searched in place (via unsafeUseAsCStringLen). + +A 'String' will be converted into a 'CString' or 'CStringLen' for +processing. Doing this repeatedly will be very inefficient. + +The "Text.Regex.PCRE.String", "Text.Regex.PCRE.ByteString", and +"Text.Regex.PCRE.Wrap" modules provides both the high level interface +exported by this module and medium- and low-level interfaces that +returns error using Either structures. +-} +{- Copyright : (c) Chris Kuklewicz 2007 -} +module Text.Regex.PCRE(getVersion_Text_Regex_PCRE + ,module Text.Regex.Base + -- ** Wrap, for '=~' and '=~~', types and constants + ,module Text.Regex.PCRE.Wrap) where + +import Text.Regex.PCRE.Wrap(Regex, CompOption(CompOption), ExecOption(ExecOption), (=~), (=~~), + unusedOffset, getNumSubs, configUTF8, getVersion, + compBlank, compAnchored, compAutoCallout, compCaseless, + compDollarEndOnly, compDotAll, compExtended, compExtra, + compFirstLine, compMultiline, compNoAutoCapture, compUngreedy, + compUTF8, compNoUTF8Check, + execBlank, execAnchored, execNotBOL, execNotEOL, execNotEmpty, + execNoUTF8Check, execPartial) +import Text.Regex.PCRE.String() +import Text.Regex.PCRE.Sequence() +import Text.Regex.PCRE.ByteString() +import Text.Regex.PCRE.ByteString.Lazy() +import Data.Version(Version(..)) +import Text.Regex.Base + +getVersion_Text_Regex_PCRE :: Version +getVersion_Text_Regex_PCRE = + Version { versionBranch = [0,94,1] -- Keep in sync with regex-pcre.cabal + , versionTags = ["unstable"] + } diff --git a/Text/Regex/PCRE/ByteString.hs b/Text/Regex/PCRE/ByteString.hs new file mode 100644 --- /dev/null +++ b/Text/Regex/PCRE/ByteString.hs @@ -0,0 +1,145 @@ +{-# OPTIONS_GHC -fno-warn-orphans #-} +{-| +This exports instances of the high level API and the medium level +API of 'compile','execute', and 'regexec'. +-} +{- Copyright : (c) Chris Kuklewicz 2007 -} +module Text.Regex.PCRE.ByteString( + -- ** Types + Regex, + MatchOffset, + MatchLength, + CompOption(CompOption), + ExecOption(ExecOption), + ReturnCode, + WrapError, + -- ** Miscellaneous + unusedOffset, + getVersion, + -- ** Medium level API functions + compile, + execute, + regexec, + -- ** CompOption flags + compBlank, + compAnchored, + compAutoCallout, + compCaseless, + compDollarEndOnly, + compDotAll, + compExtended, + compExtra, + compFirstLine, + compMultiline, + compNoAutoCapture, + compUngreedy, + compUTF8, + compNoUTF8Check, + -- ** ExecOption flags + execBlank, + execAnchored, + execNotBOL, + execNotEOL, + execNotEmpty, + execNoUTF8Check, + execPartial + ) where + +import Text.Regex.PCRE.Wrap -- all +import Data.Array(Array,listArray) +import Data.ByteString(ByteString) +import qualified Data.ByteString as B(empty,useAsCString,last,take,drop,null,pack) +#ifdef SPLIT_BASE +import qualified Data.ByteString.Unsafe as B(unsafeUseAsCString,unsafeUseAsCStringLen) +#else +import qualified Data.ByteString.Base as B(unsafeUseAsCString,unsafeUseAsCStringLen) +#endif +import System.IO.Unsafe(unsafePerformIO) +import Text.Regex.Base.RegexLike(RegexContext(..),RegexMaker(..),RegexLike(..),MatchOffset,MatchLength) +import Text.Regex.Base.Impl(polymatch,polymatchM) +import Foreign.C.String(CStringLen) +import Foreign(nullPtr) + +instance RegexContext Regex ByteString ByteString where + match = polymatch + matchM = polymatchM + +unwrap :: (Show e) => Either e v -> IO v +unwrap x = case x of Left err -> fail ("Text.Regex.PCRE.ByteString died: "++ show err) + Right v -> return v + +{-# INLINE asCStringLen #-} +asCStringLen :: ByteString -> (CStringLen -> IO a) -> IO a +asCStringLen s op = B.unsafeUseAsCStringLen s checked + where checked cs@(ptr,_) | ptr == nullPtr = B.unsafeUseAsCStringLen myEmpty (op . trim) + | otherwise = op cs + myEmpty = B.pack [0] + trim (ptr,_) = (ptr,0) + +instance RegexMaker Regex CompOption ExecOption ByteString where + makeRegexOpts c e pattern = unsafePerformIO $ + compile c e pattern >>= unwrap + makeRegexOptsM c e pattern = either (fail.show) return $ unsafePerformIO $ + compile c e pattern + +instance RegexLike Regex ByteString where + matchTest regex bs = unsafePerformIO $ + asCStringLen bs (wrapTest 0 regex) >>= unwrap + matchOnce regex bs = unsafePerformIO $ + execute regex bs >>= unwrap + matchAll regex bs = unsafePerformIO $ + asCStringLen bs (wrapMatchAll regex) >>= unwrap + matchCount regex bs = unsafePerformIO $ + asCStringLen bs (wrapCount regex) >>= unwrap + +-- --------------------------------------------------------------------- +-- | Compiles a regular expression +-- +compile :: CompOption -- ^ (summed together) + -> ExecOption -- ^ (summed together) + -> ByteString -- ^ The regular expression to compile + -> IO (Either (MatchOffset,String) Regex) -- ^ Returns: the compiled regular expression +compile c e pattern = do + -- PCRE does not allow one to specify a length for the regular expression, it must by 0 terminated + let asCString bs = if (not (B.null bs)) && (0==B.last bs) + then B.unsafeUseAsCString bs + else B.useAsCString bs + asCString pattern (wrapCompile c e) + +-- --------------------------------------------------------------------- +-- | Matches a regular expression against a buffer, returning the buffer +-- indicies of the match, and any submatches +-- +-- | Matches a regular expression against a string +execute :: Regex -- ^ Compiled regular expression + -> ByteString -- ^ String to match against + -> IO (Either WrapError (Maybe (Array Int (MatchOffset,MatchLength)))) + -- ^ Returns: 'Nothing' if the regex did not match the + -- string, or: + -- 'Just' an array of (offset,length) pairs where index 0 is whole match, and the rest are the captured subexpressions. +execute regex bs = do + maybeStartEnd <- asCStringLen bs (wrapMatch 0 regex) + case maybeStartEnd of + Right Nothing -> return (Right Nothing) + Right (Just parts) -> + return . Right . Just . listArray (0,pred (length parts)) + . map (\(s,e)->(fromIntegral s, fromIntegral (e-s))) $ parts + Left err -> return (Left err) + +regexec :: Regex -- ^ Compiled regular expression + -> ByteString -- ^ String to match against + -> IO (Either WrapError (Maybe (ByteString, ByteString, ByteString, [ByteString]))) +regexec regex bs = do + let getSub (start,stop) | start == unusedOffset = B.empty + | otherwise = B.take (stop-start) . B.drop start $ bs + matchedParts [] = (B.empty,B.empty,bs,[]) -- no information + matchedParts (matchedStartStop@(start,stop):subStartStop) = + (B.take start bs + ,getSub matchedStartStop + ,B.drop stop bs + ,map getSub subStartStop) + maybeStartEnd <- asCStringLen bs (wrapMatch 0 regex) + case maybeStartEnd of + Right Nothing -> return (Right Nothing) + Right (Just parts) -> return . Right . Just . matchedParts $ parts + Left err -> return (Left err) diff --git a/Text/Regex/PCRE/ByteString/Lazy.hs b/Text/Regex/PCRE/ByteString/Lazy.hs new file mode 100644 --- /dev/null +++ b/Text/Regex/PCRE/ByteString/Lazy.hs @@ -0,0 +1,141 @@ +{-# OPTIONS_GHC -fno-warn-orphans #-} +{-| +This exports instances of the high level API and the medium level +API of 'compile','execute', and 'regexec'. +-} +{- Copyright : (c) Chris Kuklewicz 2007 -} +module Text.Regex.PCRE.ByteString.Lazy( + -- ** Types + Regex, + MatchOffset, + MatchLength, + CompOption(CompOption), + ExecOption(ExecOption), + ReturnCode, + WrapError, + -- ** Miscellaneous + unusedOffset, + getVersion, + -- ** Medium level API functions + compile, + execute, + regexec, + -- ** CompOption flags + compBlank, + compAnchored, + compAutoCallout, + compCaseless, + compDollarEndOnly, + compDotAll, + compExtended, + compExtra, + compFirstLine, + compMultiline, + compNoAutoCapture, + compUngreedy, + compUTF8, + compNoUTF8Check, + -- ** ExecOption flags + execBlank, + execAnchored, + execNotBOL, + execNotEOL, + execNotEmpty, + execNoUTF8Check, + execPartial + ) where + +import Text.Regex.PCRE.Wrap -- all +import Data.Array(Array) +import qualified Data.ByteString.Lazy as L(ByteString,toChunks,fromChunks,last,null,snoc) +import qualified Data.ByteString as B(ByteString,concat,pack) +#ifdef SPLIT_BASE +import qualified Data.ByteString.Unsafe as B(unsafeUseAsCString,unsafeUseAsCStringLen) +#else +import qualified Data.ByteString.Base as B(unsafeUseAsCString,unsafeUseAsCStringLen) +#endif +import System.IO.Unsafe(unsafePerformIO) +import Text.Regex.Base.RegexLike(RegexContext(..),RegexMaker(..),RegexLike(..),MatchOffset,MatchLength) +import Text.Regex.Base.Impl(polymatch,polymatchM) +import qualified Text.Regex.PCRE.ByteString as BS(execute,regexec) +import Foreign.C.String(CString,CStringLen) +import Foreign(nullPtr) + +instance RegexContext Regex L.ByteString L.ByteString where + match = polymatch + matchM = polymatchM + +{-# INLINE fromLazy #-} +fromLazy :: L.ByteString -> B.ByteString +fromLazy = B.concat . L.toChunks + +{-# INLINE toLazy #-} +toLazy :: B.ByteString -> L.ByteString +toLazy = L.fromChunks . return + +unwrap :: (Show e) => Either e v -> IO v +unwrap x = case x of Left err -> fail ("Text.Regex.PCRE.ByteString.Lazy died: "++ show err) + Right v -> return v + +{-# INLINE asCString #-} +asCString :: L.ByteString -> (CString -> IO a) -> IO a +asCString s = if (not (L.null s)) && (0==L.last s) + then B.unsafeUseAsCString (fromLazy s) + else B.unsafeUseAsCString (fromLazy (L.snoc s 0)) + +{-# INLINE asCStringLen #-} +asCStringLen :: L.ByteString -> (CStringLen -> IO a) -> IO a +asCStringLen ls op = B.unsafeUseAsCStringLen (fromLazy ls) checked + where checked cs@(ptr,_) | ptr == nullPtr = B.unsafeUseAsCStringLen myEmpty (op . trim) + | otherwise = op cs + myEmpty = B.pack [0] + trim (ptr,_) = (ptr,0) + +instance RegexMaker Regex CompOption ExecOption L.ByteString where + makeRegexOpts c e pattern = unsafePerformIO $ + compile c e pattern >>= unwrap + makeRegexOptsM c e pattern = either (fail.show) return $ unsafePerformIO $ + compile c e pattern + +instance RegexLike Regex L.ByteString where + matchTest regex bs = unsafePerformIO $ + asCStringLen bs (wrapTest 0 regex) >>= unwrap + matchOnce regex bs = unsafePerformIO $ + execute regex bs >>= unwrap + matchAll regex bs = unsafePerformIO $ + asCStringLen bs (wrapMatchAll regex) >>= unwrap + matchCount regex bs = unsafePerformIO $ + asCStringLen bs (wrapCount regex) >>= unwrap + +-- --------------------------------------------------------------------- +-- | Compiles a regular expression +-- +compile :: CompOption -- ^ (summed together) + -> ExecOption -- ^ (summed together) + -> L.ByteString -- ^ The regular expression to compile + -> IO (Either (MatchOffset,String) Regex) -- ^ Returns: the compiled regular expression +compile c e pattern = do + asCString pattern (wrapCompile c e) + +-- --------------------------------------------------------------------- +-- | Matches a regular expression against a buffer, returning the buffer +-- indicies of the match, and any submatches +-- +-- | Matches a regular expression against a string +execute :: Regex -- ^ Compiled regular expression + -> L.ByteString -- ^ String to match against + -> IO (Either WrapError (Maybe (Array Int (MatchOffset,MatchLength)))) + -- ^ Returns: 'Nothing' if the regex did not match the + -- string, or: + -- 'Just' an array of (offset,length) pairs where index 0 is whole match, and the rest are the captured subexpressions. +execute regex bs = BS.execute regex (fromLazy bs) + +regexec :: Regex -- ^ Compiled regular expression + -> L.ByteString -- ^ String to match against + -> IO (Either WrapError (Maybe (L.ByteString, L.ByteString, L.ByteString, [L.ByteString]))) +regexec regex bs = do + x <- BS.regexec regex (fromLazy bs) + return $ case x of + Left e -> Left e + Right Nothing -> Right Nothing + Right (Just (a,b,c,ds)) -> Right (Just (toLazy a,toLazy b,toLazy c,map toLazy ds)) diff --git a/Text/Regex/PCRE/Sequence.hs b/Text/Regex/PCRE/Sequence.hs new file mode 100644 --- /dev/null +++ b/Text/Regex/PCRE/Sequence.hs @@ -0,0 +1,154 @@ +{-# OPTIONS_GHC -fno-warn-orphans #-} +{-| +This exports instances of the high level API and the medium level +API of 'compile','execute', and 'regexec'. +-} +{- Copyright : (c) Chris Kuklewicz 2007 -} +module Text.Regex.PCRE.Sequence( + -- ** Types + Regex, + MatchOffset, + MatchLength, + CompOption(CompOption), + ExecOption(ExecOption), + ReturnCode, + WrapError, + -- ** Miscellaneous + unusedOffset, + getVersion, + -- ** Medium level API functions + compile, + execute, + regexec, + -- ** Constants for CompOption + compBlank, + compAnchored, + compAutoCallout, + compCaseless, + compDollarEndOnly, + compDotAll, + compExtended, + compExtra, + compFirstLine, + compMultiline, + compNoAutoCapture, + compUngreedy, + compUTF8, + compNoUTF8Check, + -- ** Constants for ExecOption + execBlank, + execAnchored, + execNotBOL, + execNotEOL, + execNotEmpty, + execNoUTF8Check, + execPartial + ) where + +import Text.Regex.PCRE.Wrap -- all +--import Foreign.C.String(withCStringLen,withCString) +import Data.Array(Array,listArray) +import System.IO.Unsafe(unsafePerformIO) +import Text.Regex.Base.RegexLike(RegexMaker(..),RegexLike(..),RegexContext(..),MatchLength,MatchOffset,Extract(..)) +import Text.Regex.Base.Impl(polymatch,polymatchM) +import Data.Sequence as S hiding (length) +import qualified Data.Sequence as S (length) +import Foreign.C.String +import Foreign.Marshal.Array +import Foreign.Marshal.Alloc +import Foreign.Storable + +instance RegexContext Regex (Seq Char) (Seq Char) where + match = polymatch + matchM = polymatchM + +unwrap :: (Show e) => Either e v -> IO v +unwrap x = case x of Left err -> fail ("Text.Regex.PCRE.Sequence died: "++ show err) + Right v -> return v + +instance RegexMaker Regex CompOption ExecOption (Seq Char) where + makeRegexOpts c e pattern = unsafePerformIO $ + compile c e pattern >>= unwrap + makeRegexOptsM c e pattern = either (fail.show) return $ unsafePerformIO $ + compile c e pattern + +instance RegexLike Regex (Seq Char) where + matchTest regex str = unsafePerformIO $ + withSeq str (wrapTest 0 regex) >>= unwrap + matchOnce regex str = unsafePerformIO $ + execute regex str >>= unwrap + matchAll regex str = unsafePerformIO $ + withSeq str (wrapMatchAll regex) >>= unwrap + matchCount regex str = unsafePerformIO $ + withSeq str (wrapCount regex) >>= unwrap + +-- | Compiles a regular expression +compile :: CompOption -- ^ Flags (summed together) + -> ExecOption -- ^ Flags (summed together) + -> (Seq Char) -- ^ The regular expression to compile + -> IO (Either (MatchOffset,String) Regex) -- ^ Returns: an error string and offset or the compiled regular expression +compile c e pattern = withSeq0 pattern (wrapCompile c e) + +-- | Matches a regular expression against a string +execute :: Regex -- ^ Compiled regular expression + -> (Seq Char) -- ^ (Seq Char) to match against + -> IO (Either WrapError (Maybe (Array Int (MatchOffset,MatchLength)))) + -- ^ Returns: 'Nothing' if the regex did not match the + -- string, or: + -- 'Just' an array of (offset,length) pairs where index 0 is whole match, and the rest are the captured subexpressions. +execute regex str = do + maybeStartEnd <- withSeq str (wrapMatch 0 regex) + case maybeStartEnd of + Right Nothing -> return (Right Nothing) +-- Right (Just []) -> fail "got [] back!" -- should never happen + Right (Just parts) -> + return . Right . Just . listArray (0,pred (length parts)) + . map (\(s,e)->(fromIntegral s, fromIntegral (e-s))) $ parts + Left err -> return (Left err) + +-- | execute match and extract substrings rather than just offsets +regexec :: Regex -- ^ compiled regular expression + -> (Seq Char) -- ^ string to match + -> IO (Either WrapError (Maybe ((Seq Char), (Seq Char),(Seq Char), [(Seq Char)]))) + -- ^ Returns: Nothing if no match, else + -- (text before match, text after match, array of matches with 0 being the whole match) +regexec regex str = do + let getSub (start,stop) | start == unusedOffset = S.empty + | otherwise = extract (start,stop-start) str + matchedParts [] = (S.empty,S.empty,str,[]) -- no information + matchedParts (matchedStartStop@(start,stop):subStartStop) = + (before start str + ,getSub matchedStartStop + ,after stop str + ,map getSub subStartStop) + maybeStartEnd <- withSeq str (wrapMatch 0 regex) + case maybeStartEnd of + Right Nothing -> return (Right Nothing) +-- Right (Just []) -> fail "got [] back!" -- should never happen + Right (Just parts) -> return . Right . Just . matchedParts $ parts + Left err -> return (Left err) + +withSeq :: Seq Char -> (CStringLen -> IO a) -> IO a +withSeq s f = + let -- Ensure null at end of s + len = S.length s + pokes p a | seq p (seq a False) = undefined + | otherwise = + case viewl a of + EmptyL -> return () + c :< a' -> poke p (castCharToCChar c) >> pokes (advancePtr p 1) a' + in allocaBytes (S.length s) (\ptr -> pokes ptr s >> f (ptr,len)) + +withSeq0 :: Seq Char -> (CString -> IO a) -> IO a +withSeq0 s f = + let -- Ensure null at end of s + s' = case viewr s of -- bang !s' + EmptyR -> singleton '\0' + _ :> '\0' -> s + _ -> s |> '\0' + pokes p a | seq p (seq a False) = undefined + | otherwise = + case viewl a of -- bang pokes !p !a + EmptyL -> return () + c :< a' -> poke p (castCharToCChar c) >> pokes (advancePtr p 1) a' + in allocaBytes (S.length s') (\ptr -> pokes ptr s' >> f ptr) diff --git a/Text/Regex/PCRE/String.hs b/Text/Regex/PCRE/String.hs new file mode 100644 --- /dev/null +++ b/Text/Regex/PCRE/String.hs @@ -0,0 +1,123 @@ +{-# OPTIONS_GHC -fno-warn-orphans #-} +{-| +This exports instances of the high level API and the medium level +API of 'compile','execute', and 'regexec'. +-} +{- Copyright : (c) Chris Kuklewicz 2007 -} +module Text.Regex.PCRE.String( + -- ** Types + Regex, + MatchOffset, + MatchLength, + CompOption(CompOption), + ExecOption(ExecOption), + ReturnCode, + WrapError, + -- ** Miscellaneous + unusedOffset, + getVersion, + -- ** Medium level API functions + compile, + execute, + regexec, + -- ** Constants for CompOption + compBlank, + compAnchored, + compAutoCallout, + compCaseless, + compDollarEndOnly, + compDotAll, + compExtended, + compExtra, + compFirstLine, + compMultiline, + compNoAutoCapture, + compUngreedy, + compUTF8, + compNoUTF8Check, + -- ** Constants for ExecOption + execBlank, + execAnchored, + execNotBOL, + execNotEOL, + execNotEmpty, + execNoUTF8Check, + execPartial + ) where + +import Text.Regex.PCRE.Wrap -- all +import Foreign.C.String(withCStringLen,withCString) +import Data.Array(Array,listArray) +import System.IO.Unsafe(unsafePerformIO) +import Text.Regex.Base.RegexLike(RegexMaker(..),RegexLike(..),RegexContext(..),MatchLength,MatchOffset) +import Text.Regex.Base.Impl(polymatch,polymatchM) + +instance RegexContext Regex String String where + match = polymatch + matchM = polymatchM + +unwrap :: (Show e) => Either e v -> IO v +unwrap x = case x of Left err -> fail ("Text.Regex.PCRE.String died: "++ show err) + Right v -> return v + +instance RegexMaker Regex CompOption ExecOption String where + makeRegexOpts c e pattern = unsafePerformIO $ + compile c e pattern >>= unwrap + makeRegexOptsM c e pattern = either (fail.show) return $ unsafePerformIO $ + compile c e pattern + +instance RegexLike Regex String where + matchTest regex str = unsafePerformIO $ + withCStringLen str (wrapTest 0 regex) >>= unwrap + matchOnce regex str = unsafePerformIO $ + execute regex str >>= unwrap + matchAll regex str = unsafePerformIO $ + withCStringLen str (wrapMatchAll regex) >>= unwrap + matchCount regex str = unsafePerformIO $ + withCStringLen str (wrapCount regex) >>= unwrap + +-- | Compiles a regular expression +compile :: CompOption -- ^ Flags (summed together) + -> ExecOption -- ^ Flags (summed together) + -> String -- ^ The regular expression to compile + -> IO (Either (MatchOffset,String) Regex) -- ^ Returns: an error string and offset or the compiled regular expression +compile c e pattern = withCString pattern (wrapCompile c e) + +-- | Matches a regular expression against a string +execute :: Regex -- ^ Compiled regular expression + -> String -- ^ String to match against + -> IO (Either WrapError (Maybe (Array Int (MatchOffset,MatchLength)))) + -- ^ Returns: 'Nothing' if the regex did not match the + -- string, or: + -- 'Just' an array of (offset,length) pairs where index 0 is whole match, and the rest are the captured subexpressions. +execute regex str = do + maybeStartEnd <- withCStringLen str (wrapMatch 0 regex) + case maybeStartEnd of + Right Nothing -> return (Right Nothing) +-- Right (Just []) -> fail "got [] back!" -- should never happen + Right (Just parts) -> + return . Right . Just . listArray (0,pred (length parts)) + . map (\(s,e)->(fromIntegral s, fromIntegral (e-s))) $ parts + Left err -> return (Left err) + +-- | execute match and extract substrings rather than just offsets +regexec :: Regex -- ^ compiled regular expression + -> String -- ^ string to match + -> IO (Either WrapError (Maybe (String, String,String, [String]))) + -- ^ Returns: Nothing if no match, else + -- (text before match, text after match, array of matches with 0 being the whole match) +regexec regex str = do + let getSub (start,stop) | start == unusedOffset = "" + | otherwise = take (stop-start) . drop start $ str + matchedParts [] = ("","",str,[]) -- no information + matchedParts (matchedStartStop@(start,stop):subStartStop) = + (take start str + ,getSub matchedStartStop + ,drop stop str + ,map getSub subStartStop) + maybeStartEnd <- withCStringLen str (wrapMatch 0 regex) + case maybeStartEnd of + Right Nothing -> return (Right Nothing) +-- Right (Just []) -> fail "got [] back!" -- should never happen + Right (Just parts) -> return . Right . Just . matchedParts $ parts + Left err -> return (Left err) diff --git a/Text/Regex/PCRE/Wrap.hsc b/Text/Regex/PCRE/Wrap.hsc new file mode 100644 --- /dev/null +++ b/Text/Regex/PCRE/Wrap.hsc @@ -0,0 +1,490 @@ +-- The exported symbols are the same whether HAVE_PCRE_H is defined, +-- but when if it is not defined then 'getVersion == Nothing' and all +-- other exported values will call error or fail. + +-- | This will fail or error only if allocation fails or a nullPtr is passed in. + +-- TODO :: Consider wrapMatchAll using list of start/end offsets and not MatchArray +-- + +{- Copyright : (c) Chris Kuklewicz 2007 -} +module Text.Regex.PCRE.Wrap( + -- ** High-level interface + Regex, + CompOption(CompOption), + ExecOption(ExecOption), + (=~), + (=~~), + + -- ** Low-level interface + StartOffset, + EndOffset, + ReturnCode(ReturnCode), + WrapError, + wrapCompile, + wrapTest, + wrapMatch, + wrapMatchAll, + wrapCount, + + -- ** Miscellaneous + getVersion, + configUTF8, + getNumSubs, + unusedOffset, + + -- ** CompOption values + compBlank, + compAnchored, + compAutoCallout, + compCaseless, + compDollarEndOnly, + compDotAll, + compExtended, + compExtra, + compFirstLine, + compMultiline, + compNoAutoCapture, + compUngreedy, + compUTF8, + compNoUTF8Check, + + -- ** ExecOption values + execBlank, + execAnchored, + execNotBOL, + execNotEOL, + execNotEmpty, + execNoUTF8Check, + execPartial, + + -- ** ReturnCode values + retOk, + retNoMatch, + retNull, + retBadOption, + retBadMagic, + retUnknownNode, + retNoMemory, + retNoSubstring + ) where + +#if defined(HAVE_PCRE_H) +import Control.Monad(when) +import Data.Array(Array,accumArray) +import Data.Bits(Bits((.|.))) -- ((.&.),(.|.),complement)) +import Foreign(unsafePerformIO + ,Ptr,ForeignPtr,FinalizerPtr -- ,FunPtr + ,alloca,allocaBytes,nullPtr + ,peek,peekElemOff + ,newForeignPtr,withForeignPtr) +import Foreign.C(CInt,CChar) +import Foreign.C.String(CString,CStringLen,peekCString) +import Text.Regex.Base.RegexLike(RegexOptions(..),RegexMaker(..),RegexContext(..),MatchArray,MatchOffset) +#else +import Data.Array(Array) +import Data.Bits(Bits) +import Foreign(ForeignPtr) +import Foreign.C(CInt) +import Foreign.C.String(CString,CStringLen) +import Text.Regex.Base.RegexLike(RegexOptions(..),RegexMaker(..),RegexContext(..),MatchArray,MatchOffset) +#endif + + +-- | return version of pcre used or Nothing if pcre is not available. +getVersion :: Maybe String + +type PCRE = () +type StartOffset = MatchOffset +type EndOffset = MatchOffset +type WrapError = (ReturnCode,String) + +newtype CompOption = CompOption CInt deriving (Eq,Show,Num,Bits) +newtype ExecOption = ExecOption CInt deriving (Eq,Show,Num,Bits) +newtype ReturnCode = ReturnCode CInt deriving (Eq,Show) + +-- | A compiled regular expression +data Regex = Regex (ForeignPtr PCRE) CompOption ExecOption + +compBlank :: CompOption +execBlank :: ExecOption +unusedOffset :: MatchOffset +retOk :: ReturnCode + +wrapCompile :: CompOption -- ^ Flags (summed together) + -> ExecOption -- ^ Flags (summed together) + -> CString -- ^ The regular expression to compile + -> IO (Either (MatchOffset,String) Regex) -- ^ Returns: an error offset and string or the compiled regular expression +wrapTest :: StartOffset -- ^ Starting index in CStringLen + -> Regex -- ^ Compiled regular expression + -> CStringLen -- ^ String to match against and length in bytes + -> IO (Either WrapError Bool) +wrapMatch :: StartOffset -- ^ Starting index in CStringLen + -> Regex -- ^ Compiled regular expression + -> CStringLen -- ^ String to match against and length in bytes + -> IO (Either WrapError (Maybe [(StartOffset,EndOffset)])) + -- ^ Returns: 'Right Nothing' if the regex did not match the + -- string, or: + -- 'Right Just' an array of (offset,length) pairs where index 0 is whole match, and the rest are the captured subexpressions, or: + -- 'Left ReturnCode' if there is some strange error +wrapMatchAll :: Regex -> CStringLen -> IO (Either WrapError [ MatchArray ]) +wrapCount :: Regex -> CStringLen -> IO (Either WrapError Int) + +getNumSubs :: Regex -> Int +configUTF8 :: Bool + +(=~) :: (RegexMaker Regex CompOption ExecOption source,RegexContext Regex source1 target) + => source1 -> source -> target +(=~~) :: (RegexMaker Regex CompOption ExecOption source,RegexContext Regex source1 target,Monad m) + => source1 -> source -> m target + +#if defined(HAVE_PCRE_H) +#include +#include + +instance RegexOptions Regex CompOption ExecOption where + blankCompOpt = compBlank + blankExecOpt = execBlank + defaultCompOpt = compMultiline + defaultExecOpt = execBlank + setExecOpts e' (Regex r c _) = Regex r c e' + getExecOpts (Regex _ _ e) = e + +-- (=~) :: (RegexMaker Regex CompOption ExecOption source,RegexContext Regex source1 target) => source1 -> source -> target +(=~) x r = let q :: Regex + q = makeRegex r + in match q x + +-- (=~~) ::(RegexMaker Regex CompOption ExecOption source,RegexContext Regex source1 target,Monad m) => source1 -> source -> m target +(=~~) x r = do (q :: Regex) <- makeRegexM r + matchM q x + +type PCRE_Extra = () + +fi :: (Integral i,Num n ) => i -> n +fi x = fromIntegral x + +compBlank = CompOption 0 +execBlank = ExecOption 0 +unusedOffset = (-1) +retOk = ReturnCode 0 + +retNeededMoreSpace :: ReturnCode +retNeededMoreSpace = ReturnCode 0 + +newtype InfoWhat = InfoWhat CInt deriving (Eq,Show) +newtype ConfigWhat = ConfigWhat CInt deriving (Eq,Show) + +nullTest' :: Ptr a -> String -> IO (Either (MatchOffset,String) b) -> IO (Either (MatchOffset,String) b) +{-# INLINE nullTest' #-} +nullTest' ptr msg io = do + if nullPtr == ptr + then return (Left (0,"Ptr parameter was nullPtr in Text.Regex.PCRE.Wrap."++msg)) + else io + +nullTest :: Ptr a -> String -> IO (Either WrapError b) -> IO (Either WrapError b) +{-# INLINE nullTest #-} +nullTest ptr msg io = do + if nullPtr == ptr + then return (Left (retOk,"Ptr parameter was nullPtr in Text.Regex.PCRE.Wrap."++msg)) + else io + +wrapRC :: ReturnCode -> IO (Either WrapError b) +{-# INLINE wrapRC #-} +wrapRC r = return (Left (r,"Error in Text.Regex.PCRE.Wrap: "++show r)) + +-- | Compiles a regular expression +wrapCompile flags e pattern = do + nullTest' pattern "wrapCompile pattern" $ do + alloca $ \errOffset -> alloca $ \errPtr -> do + nullTest' errPtr "wrapCompile errPtr" $ do + pcre_ptr <- c_pcre_compile pattern flags errPtr errOffset nullPtr + if pcre_ptr == nullPtr + then do + -- No need to use c_ptr_free in the error case (e.g. pcredemo.c) + offset <- peek errOffset + string <- peekCString =<< peek errPtr + return (Left (fi offset,string)) + else do regex <- newForeignPtr c_ptr_free pcre_ptr + return . Right $ Regex regex flags e + +getNumSubs (Regex pcre_fptr _ _) = fi . unsafePerformIO $ withForeignPtr pcre_fptr getNumSubs' + +getNumSubs' :: Ptr PCRE -> IO CInt +{-# INLINE getNumSubs' #-} +getNumSubs' pcre_ptr = + alloca $ \st -> do -- (st :: Ptr CInt) + when (st == nullPtr) (fail "Text.Regex.PCRE.Wrap.getNumSubs' could not allocate a CInt!!!") + c_pcre_fullinfo pcre_ptr nullPtr pcreInfoCapturecount st + peek st + +wrapTest startOffset (Regex pcre_fptr _ flags) (cstr,len) = do + nullTest cstr "wrapTest cstr" $ do + withForeignPtr pcre_fptr $ \pcre_ptr -> do + r@(ReturnCode r') <- c_pcre_exec pcre_ptr nullPtr cstr (fi len) (fi startOffset) flags nullPtr 0 + if r == retNoMatch + then return (Right False) + else if r' < 0 + then wrapRC r + else return (Right True) + +-- | Matches a regular expression against a string +-- +-- Should never return (Right (Just [])) +wrapMatch startOffset (Regex pcre_fptr _ flags) (cstr,len) = do + nullTest cstr "wrapMatch cstr" $ do + withForeignPtr pcre_fptr $ \pcre_ptr -> do + nsub <- getNumSubs' pcre_ptr + let nsub_int :: Int + nsub_int = fi nsub + ovec_size :: CInt + ovec_size = ((nsub + 1) * 3) -- "man pcreapi" for explanation + ovec_bytes :: Int + ovec_bytes = (fi ovec_size) * (#const sizeof(int)) + allocaBytes ovec_bytes $ \ovec -> do + nullTest ovec "wrapMatch ovec" $ do + r@(ReturnCode r') <- c_pcre_exec pcre_ptr nullPtr cstr (fi len) (fi startOffset) flags ovec ovec_size + if r == retNoMatch + then return (Right Nothing) + else if r' < 0 + then wrapRC r + else do + let pairsSet :: Int + pairsSet = if r == retNeededMoreSpace -- if r == ReturnCode 0 + then nsub_int + 1 -- should not happen + else fi r' -- implies pairsSet > 0 + extraPairs :: [(Int,Int)] + extraPairs = replicate (nsub_int + 1 - pairsSet) + (unusedOffset,unusedOffset) + pairs <- return . toPairs =<< mapM (peekElemOff ovec) [0 .. ((pairsSet*2)-1)] + return . Right . Just $ (pairs ++ extraPairs) + +-- | wrapMatchAll is an improvement over wrapMatch since it only +-- allocates memory with allocaBytes once at the start. +-- +-- +wrapMatchAll (Regex pcre_fptr _ flags) (cstr,len) = do + nullTest cstr "wrapMatchAll cstr" $ do + withForeignPtr pcre_fptr $ \regex -> do + nsub <- getNumSubs' regex + let nsub_int :: Int + nsub_int = fi nsub + ovec_size :: CInt + ovec_size = ((nsub + 1) * 3) -- "man pcreapi" for explanation + ovec_bytes :: Int + ovec_bytes = (fi ovec_size) * (#const sizeof(int)) + clen = fi len + flags' = (execNotEmpty .|. execAnchored .|. flags) + allocaBytes ovec_bytes $ \ovec -> + nullTest ovec "wrapMatchAll ovec" $ + let loop acc flags_in_use pos = do + r@(ReturnCode r') <- c_pcre_exec regex nullPtr cstr clen (fi pos) flags_in_use ovec ovec_size + if r == retNoMatch + then return (Right (acc [])) + else if r' < 0 + then wrapRC r + else do + let pairsSet = if r == retNeededMoreSpace then nsub_int+1 else fi r' + pairs <- return . toPairs =<< mapM (peekElemOff ovec) [0 .. ((pairsSet*2)-1)] + let acc' = acc . (toMatchArray nsub_int pairs:) + case pairs of + [] -> return (Right (acc' [])) + ((s,e):_) | s==e -> if s == len + then return (Right (acc' [])) + else loop acc' flags' e + | otherwise -> loop acc' flags e + in loop id flags 0 +toMatchArray :: Int -> [(Int,Int)] -> Array Int (Int,Int) +toMatchArray n pairs = accumArray (\_ (s,e) -> (s,(e-s))) (-1,0) (0,n) (zip [0..] pairs) + +toPairs :: [CInt] -> [(Int,Int)] +toPairs [] = [] +toPairs (a:b:rest) = (fi a,fi b):toPairs rest +toPairs [_] = error "Should not have just one element in WrapPCRE.wrapMatchAll.toPairs" + +wrapCount (Regex pcre_fptr _ flags) (cstr,len) = do + nullTest cstr "wrapCount cstr" $ do + withForeignPtr pcre_fptr $ \pcre_ptr -> do + nsub <- getNumSubs' pcre_ptr + let ovec_size :: CInt + ovec_size = ((nsub + 1) * 3) -- "man pcreapi" for explanation + ovec_bytes :: Int + ovec_bytes = (fi ovec_size) * (#const sizeof(int)) + clen = fi len + allocaBytes ovec_bytes $ \ovec -> + nullTest ovec "wrapCount ovec" $ + let act pos = c_pcre_exec pcre_ptr nullPtr cstr clen (fi pos) flags ovec ovec_size + loop acc pos | acc `seq` pos `seq` False = undefined + | otherwise = do + r@(ReturnCode r') <- act pos + if r == retNoMatch + then return (Right acc) + else if r' < 0 + then wrapRC r + else do + pairs <- return . toPairs =<< mapM (peekElemOff ovec) [0,1] + case pairs of + [] -> return (Right (succ acc)) + ((s,e):_) | s==e -> return (Right (succ acc)) + | otherwise -> loop (succ acc) e + in loop 0 0 + +getVersion = unsafePerformIO $ do + version <- c_pcre_version + if version == nullPtr + then return (Just "pcre_version was null") + else return . Just =<< peekCString version + +configUTF8 = unsafePerformIO $ + alloca $ \ptrVal -> do -- (ptrVal :: Ptr CInt) + when (ptrVal == nullPtr) (fail "Text.Regex.PCRE.Wrap.configUTF8 could not alloca CInt!!!") + c_pcre_config pcreConfigUtf8 ptrVal + val <- peek ptrVal + case val of + (1 :: CInt) -> return True + 0 -> return False + _ -> return False -- should not happen + +foreign import ccall unsafe "pcre.h pcre_compile" + c_pcre_compile :: CString -> CompOption -> Ptr CString -> Ptr CInt -> CString -> IO (Ptr PCRE) +foreign import ccall unsafe "&free" + c_ptr_free :: FinalizerPtr a -- FunPtr (Ptr a -> IO ()) +foreign import ccall unsafe "pcre.h pcre_exec" + c_pcre_exec :: Ptr PCRE -> Ptr PCRE_Extra -> CString -> CInt -> CInt -> ExecOption -> Ptr CInt -> CInt -> IO ReturnCode +foreign import ccall unsafe "pcre.h pcre_fullinfo" + c_pcre_fullinfo :: Ptr PCRE -> Ptr PCRE_Extra -> InfoWhat -> Ptr a -> IO CInt +foreign import ccall unsafe "pcre.h pcre_version" + c_pcre_version :: IO (Ptr CChar) +foreign import ccall unsafe "pcre.h pcre_config" + c_pcre_config :: ConfigWhat -> Ptr a -> IO CInt + + +#enum CompOption,CompOption, \ + compAnchored = PCRE_ANCHORED, \ + compAutoCallout = PCRE_AUTO_CALLOUT, \ + compCaseless = PCRE_CASELESS, \ + compDollarEndOnly = PCRE_DOLLAR_ENDONLY, \ + compDotAll = PCRE_DOTALL, \ + compExtended = PCRE_EXTENDED, \ + compExtra = PCRE_EXTRA, \ + compFirstLine = PCRE_FIRSTLINE, \ + compMultiline = PCRE_MULTILINE, \ + compNoAutoCapture = PCRE_NO_AUTO_CAPTURE, \ + compUngreedy = PCRE_UNGREEDY, \ + compUTF8 = PCRE_UTF8, \ + compNoUTF8Check = PCRE_NO_UTF8_CHECK + +#enum ExecOption,ExecOption, \ + execAnchored = PCRE_ANCHORED, \ + execNotBOL = PCRE_NOTBOL, \ + execNotEOL = PCRE_NOTEOL, \ + execNotEmpty = PCRE_NOTEMPTY, \ + execNoUTF8Check = PCRE_NO_UTF8_CHECK, \ + execPartial = PCRE_PARTIAL + +#enum ReturnCode,ReturnCode, \ + retNoMatch = PCRE_ERROR_NOMATCH, \ + retNull = PCRE_ERROR_NULL, \ + retBadOption = PCRE_ERROR_BADOPTION, \ + retBadMagic = PCRE_ERROR_BADMAGIC, \ + retUnknownNode = PCRE_ERROR_UNKNOWN_NODE, \ + retNoMemory = PCRE_ERROR_NOMEMORY, \ + retNoSubstring = PCRE_ERROR_NOSUBSTRING + +-- Comment out most of these to avoid unused binding warnings + +-- PCRE_INFO_FIRSTCHAR is deprecated, use PCRE_INFO_FIRSTBYTE instead. +#enum InfoWhat,InfoWhat, \ + PCRE_INFO_CAPTURECOUNT +{- + PCRE_INFO_BACKREFMAX, \ + PCRE_INFO_DEFAULT_TABLES, \ + PCRE_INFO_FIRSTBYTE, \ + PCRE_INFO_FIRSTCHAR, \ + PCRE_INFO_FIRSTTABLE, \ + PCRE_INFO_LASTLITERAL, \ + PCRE_INFO_NAMECOUNT, \ + PCRE_INFO_NAMEENTRYSIZE, \ + PCRE_INFO_NAMETABLE, \ + PCRE_INFO_OPTIONS, \ + PCRE_INFO_SIZE, \ + PCRE_INFO_STUDYSIZE +-} +#enum ConfigWhat,ConfigWhat, \ + PCRE_CONFIG_UTF8 +{- + PCRE_CONFIG_UNICODE_PROPERTIES, \ + PCRE_CONFIG_NEWLINE, \ + PCRE_CONFIG_LINK_SIZE, \ + PCRE_CONFIG_POSIX_MALLOC_THRESHOLD, \ + PCRE_CONFIG_MATCH_LIMIT, \ + PCRE_CONFIG_MATCH_LIMIT_RECURSION, \ + PCRE_CONFIG_STACKRECURSE +-} + +#else /* do not HAVE_PCRE_H */ + +instance RegexOptions Regex CompOption ExecOption where + blankCompOpt = err + blankExecOpt = err + defaultCompOpt = err + defaultExecOpt = err + getExecOpts = err + setExecOpts = err + +msg :: String +msg = "WrapPCRE.hsc was not compiled against pcre library with HAVE_PCRE_H defined" +err :: a +err = error msg + +(=~) = err +(=~~) = err + +-- Hack to avoid the constructor from being unused +wrapCompile _ _ _ = err >> return (Right (Regex err err err)) +wrapTest = err +wrapMatch = err +wrapMatchAll = err +wrapCount = err + +compAnchored, compAutoCallout, compCaseless, compDollarEndOnly, compDotAll, compExtended, compExtra, compFirstLine, compMultiline, compNoAutoCapture, compUngreedy, compUTF8, compNoUTF8Check :: CompOption +compBlank = err +compAnchored = err +compAutoCallout = err +compCaseless = err +compDollarEndOnly = err +compDotAll = err +compExtended = err +compExtra = err +compFirstLine = err +compMultiline = err +compNoAutoCapture = err +compUngreedy = err +compUTF8 = err +compNoUTF8Check = err + +execAnchored, execNotBOL, execNotEOL, execNotEmpty, execNoUTF8Check, execPartial :: ExecOption +execBlank = err +execAnchored = err +execNotBOL = err +execNotEOL = err +execNotEmpty = err +execNoUTF8Check = err +execPartial = err + +retNoMatch, retNull, retBadOption, retBadMagic, retUnknownNode, retNoMemory, retNoSubstring :: ReturnCode +retNoMatch = err +retNull = err +retBadOption = err +retBadMagic = err +retUnknownNode = err +retNoMemory = err +retNoSubstring = err + +getNumSubs = err +unusedOffset = err +configUTF8 = err + +getVersion = Nothing + +#endif /* HAVE_PCRE_H */ diff --git a/pcre/config.h b/pcre/config.h new file mode 100644 --- /dev/null +++ b/pcre/config.h @@ -0,0 +1,307 @@ +/* config.h. Generated from config.h.in by configure. */ +/* config.h.in. Generated from configure.ac by autoheader. */ + + +/* On Unix-like systems config.h.in is converted by "configure" into config.h. +Some other environments also support the use of "configure". PCRE is written in +Standard C, but there are a few non-standard things it can cope with, allowing +it to run on SunOS4 and other "close to standard" systems. + +If you are going to build PCRE "by hand" on a system without "configure" you +should copy the distributed config.h.generic to config.h, and then set up the +macro definitions the way you need them. You must then add -DHAVE_CONFIG_H to +all of your compile commands, so that config.h is included at the start of +every source. + +Alternatively, you can avoid editing by using -D on the compiler command line +to set the macro values. In this case, you do not have to set -DHAVE_CONFIG_H. + +PCRE uses memmove() if HAVE_MEMMOVE is set to 1; otherwise it uses bcopy() if +HAVE_BCOPY is set to 1. If your system has neither bcopy() nor memmove(), set +them both to 0; an emulation function will be used. */ + +/* By default, the \R escape sequence matches any Unicode line ending + character or sequence of characters. If BSR_ANYCRLF is defined, this is + changed so that backslash-R matches only CR, LF, or CRLF. The build- time + default can be overridden by the user of PCRE at runtime. On systems that + support it, "configure" can be used to override the default. */ +/* #undef BSR_ANYCRLF */ + +/* 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. */ +/* #undef EBCDIC */ + +/* Define to 1 if you have the `bcopy' function. */ +#ifndef HAVE_BCOPY +#define HAVE_BCOPY 1 +#endif + +/* Define to 1 if you have the header file. */ +/* #undef HAVE_BITS_TYPE_TRAITS_H */ + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_BZLIB_H +#define HAVE_BZLIB_H 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_DIRENT_H +#define HAVE_DIRENT_H 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_DLFCN_H +#define HAVE_DLFCN_H 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_INTTYPES_H +#define HAVE_INTTYPES_H 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_LIMITS_H +#define HAVE_LIMITS_H 1 +#endif + +/* Define to 1 if the system has the type `long long'. */ +#ifndef HAVE_LONG_LONG +#define HAVE_LONG_LONG 1 +#endif + +/* Define to 1 if you have the `memmove' function. */ +#ifndef HAVE_MEMMOVE +#define HAVE_MEMMOVE 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_MEMORY_H +#define HAVE_MEMORY_H 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_READLINE_HISTORY_H +#define HAVE_READLINE_HISTORY_H 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_READLINE_READLINE_H +#define HAVE_READLINE_READLINE_H 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_STDINT_H +#define HAVE_STDINT_H 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_STDLIB_H +#define HAVE_STDLIB_H 1 +#endif + +/* Define to 1 if you have the `strerror' function. */ +#ifndef HAVE_STRERROR +#define HAVE_STRERROR 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_STRING +#define HAVE_STRING 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_STRINGS_H +#define HAVE_STRINGS_H 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_STRING_H +#define HAVE_STRING_H 1 +#endif + +/* Define to 1 if you have the `strtoll' function. */ +/* #undef HAVE_STRTOLL */ + +/* Define to 1 if you have the `strtoq' function. */ +#ifndef HAVE_STRTOQ +#define HAVE_STRTOQ 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_SYS_STAT_H +#define HAVE_SYS_STAT_H 1 +#endif + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_SYS_TYPES_H +#define HAVE_SYS_TYPES_H 1 +#endif + +/* Define to 1 if you have the header file. */ +/* #undef HAVE_TYPE_TRAITS_H */ + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_UNISTD_H +#define HAVE_UNISTD_H 1 +#endif + +/* Define to 1 if the system has the type `unsigned long long'. */ +#ifndef HAVE_UNSIGNED_LONG_LONG +#define HAVE_UNSIGNED_LONG_LONG 1 +#endif + +/* Define to 1 if you have the header file. */ +/* #undef HAVE_WINDOWS_H */ + +/* Define to 1 if you have the header file. */ +#ifndef HAVE_ZLIB_H +#define HAVE_ZLIB_H 1 +#endif + +/* Define to 1 if you have the `_strtoi64' function. */ +/* #undef HAVE__STRTOI64 */ + +/* The value of LINK_SIZE determines the number of bytes used to store links + as offsets within the compiled regex. The default is 2, which allows for + compiled patterns up to 64K long. This covers the vast majority of cases. + However, PCRE can also be compiled to use 3 or 4 bytes instead. This allows + for longer patterns in extreme cases. On systems that support it, + "configure" can be used to override this default. */ +#ifndef LINK_SIZE +#define LINK_SIZE 2 +#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. + The limit exists in order to catch runaway regular expressions that take + for ever to determine that they do not match. The default is set very large + so that it does not accidentally catch legitimate cases. On systems that + support it, "configure" can be used to override this default default. */ +#ifndef MATCH_LIMIT +#define MATCH_LIMIT 10000000 +#endif + +/* The above limit applies to all calls of match(), whether or not they + increase the recursion depth. In some environments it is desirable to limit + the depth of recursive calls of match() more strictly, in order to restrict + the maximum amount of stack (or heap, if NO_RECURSE is defined) that is + used. The value of MATCH_LIMIT_RECURSION applies only to recursive calls of + match(). To have any useful effect, it must be less than the value of + MATCH_LIMIT. The default is to use the same value as MATCH_LIMIT. There is + a runtime method for setting a different limit. On systems that support it, + "configure" can be used to override the default. */ +#ifndef MATCH_LIMIT_RECURSION +#define MATCH_LIMIT_RECURSION MATCH_LIMIT +#endif + +/* This limit is parameterized just in case anybody ever wants to change it. + Care must be taken if it is increased, because it guards against integer + overflow caused by enormously large patterns. */ +#ifndef MAX_NAME_COUNT +#define MAX_NAME_COUNT 10000 +#endif + +/* This limit is parameterized just in case anybody ever wants to change it. + Care must be taken if it is increased, because it guards against integer + overflow caused by enormously large patterns. */ +#ifndef MAX_NAME_SIZE +#define MAX_NAME_SIZE 32 +#endif + +/* The value of NEWLINE determines the newline character sequence. On systems + that support it, "configure" can be used to override the default, which is + 10. The possible values are 10 (LF), 13 (CR), 3338 (CRLF), -1 (ANY), or -2 + (ANYCRLF). */ +#ifndef NEWLINE +#define NEWLINE 10 +#endif + +/* PCRE uses recursive function calls to handle backtracking while matching. + This can sometimes be a problem on systems that have stacks of limited + size. Define NO_RECURSE to get a version that doesn't use recursion in the + match() function; instead it creates its own stack by steam using + pcre_recurse_malloc() to obtain memory from the heap. For more detail, see + the comments and other stuff just above the match() function. On systems + that support it, "configure" can be used to set this in the Makefile (use + --disable-stack-for-recursion). */ +/* #undef NO_RECURSE */ + +/* Name of package */ +#define PACKAGE "pcre" + +/* Define to the address where bug reports for this package should be sent. */ +#define PACKAGE_BUGREPORT "" + +/* Define to the full name of this package. */ +#define PACKAGE_NAME "PCRE" + +/* Define to the full name and version of this package. */ +#define PACKAGE_STRING "PCRE 7.7" + +/* Define to the one symbol short name of this package. */ +#define PACKAGE_TARNAME "pcre" + +/* Define to the version of this package. */ +#define PACKAGE_VERSION "7.7" + + +/* 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 + contain the relevant magic. If you do not define this macro, it + defaults to "extern" for a C compiler and "extern C" for a C++ + compiler on non-Win32 systems. This macro apears at the start of + every exported function that is part of the external API. It does + not appear on functions that are "external" in the C sense, but + which are internal to the library. */ +/* #undef PCRE_EXP_DEFN */ + +/* Define if linking statically (TODO: make nice with Libtool) */ +/* #undef PCRE_STATIC */ + +/* When calling PCRE via the POSIX interface, additional working storage is + required for holding the pointers to capturing substrings because PCRE + requires three integers per substring, whereas the POSIX interface provides + only two. If the number of expected substrings is small, the wrapper + function uses space on the stack, because this is faster than using + malloc() for each call. The threshold above which the stack is no longer + used is defined by POSIX_MALLOC_THRESHOLD. On systems that support it, + "configure" can be used to override this default. */ +#ifndef POSIX_MALLOC_THRESHOLD +#define POSIX_MALLOC_THRESHOLD 10 +#endif + +/* Define to 1 if you have the ANSI C header files. */ +#ifndef STDC_HEADERS +#define STDC_HEADERS 1 +#endif + +/* 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 libreadline. */ +/* #undef SUPPORT_LIBREADLINE */ + +/* Define to allow pcregrep to be linked with libz, so that it is able to + handle .gz files. */ +/* #undef SUPPORT_LIBZ */ + +/* Define to enable support for Unicode properties */ +#define SUPPORT_UCP + +/* Define to enable support for the UTF-8 Unicode encoding. */ +#define SUPPORT_UTF8 + +/* Version number of package */ +#ifndef VERSION +#define VERSION "7.7" +#endif + +/* Define to empty if `const' does not conform to ANSI C. */ +/* #undef const */ + +/* Define to `unsigned int' if does not define. */ +/* #undef size_t */ diff --git a/pcre/dftables.c b/pcre/dftables.c new file mode 100644 --- /dev/null +++ b/pcre/dftables.c @@ -0,0 +1,199 @@ +/************************************************* +* 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 new file mode 100644 --- /dev/null +++ b/pcre/pcre.h @@ -0,0 +1,304 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* 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 + +----------------------------------------------------------------------------- +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. +----------------------------------------------------------------------------- +*/ + +#ifndef _PCRE_H +#define _PCRE_H + +/* The current PCRE version information. */ + +#define PCRE_MAJOR 7 +#define PCRE_MINOR 7 +#define PCRE_PRERELEASE +#define PCRE_DATE 2008-05-07 + +/* 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 +export setting is defined in pcre_internal.h, which includes this file. So we +don't change existing definitions of PCRE_EXP_DECL and PCRECPP_EXP_DECL. */ + +#if defined(_WIN32) && !defined(PCRE_STATIC) +# ifndef PCRE_EXP_DECL +# define PCRE_EXP_DECL extern __declspec(dllimport) +# endif +# ifdef __cplusplus +# ifndef PCRECPP_EXP_DECL +# define PCRECPP_EXP_DECL extern __declspec(dllimport) +# endif +# ifndef PCRECPP_EXP_DEFN +# define PCRECPP_EXP_DEFN __declspec(dllimport) +# endif +# endif +#endif + +/* By default, we use the standard "extern" declarations. */ + +#ifndef PCRE_EXP_DECL +# ifdef __cplusplus +# define PCRE_EXP_DECL extern "C" +# else +# define PCRE_EXP_DECL extern +# endif +#endif + +#ifdef __cplusplus +# ifndef PCRECPP_EXP_DECL +# define PCRECPP_EXP_DECL extern +# endif +# ifndef PCRECPP_EXP_DEFN +# define PCRECPP_EXP_DEFN +# endif +#endif + +/* Have to include stdlib.h in order to ensure that size_t is defined; +it is needed here for malloc. */ + +#include + +/* Allow for C++ users */ + +#ifdef __cplusplus +extern "C" { +#endif + +/* Options */ + +#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 + +/* 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) + +/* Request types for pcre_fullinfo() */ + +#define PCRE_INFO_OPTIONS 0 +#define PCRE_INFO_SIZE 1 +#define PCRE_INFO_CAPTURECOUNT 2 +#define PCRE_INFO_BACKREFMAX 3 +#define PCRE_INFO_FIRSTBYTE 4 +#define PCRE_INFO_FIRSTCHAR 4 /* For backwards compatibility */ +#define PCRE_INFO_FIRSTTABLE 5 +#define PCRE_INFO_LASTLITERAL 6 +#define PCRE_INFO_NAMEENTRYSIZE 7 +#define PCRE_INFO_NAMECOUNT 8 +#define PCRE_INFO_NAMETABLE 9 +#define PCRE_INFO_STUDYSIZE 10 +#define PCRE_INFO_DEFAULT_TABLES 11 +#define PCRE_INFO_OKPARTIAL 12 +#define PCRE_INFO_JCHANGED 13 +#define PCRE_INFO_HASCRORLF 14 + +/* Request types for pcre_config(). Do not re-arrange, in order to remain +compatible. */ + +#define PCRE_CONFIG_UTF8 0 +#define PCRE_CONFIG_NEWLINE 1 +#define PCRE_CONFIG_LINK_SIZE 2 +#define PCRE_CONFIG_POSIX_MALLOC_THRESHOLD 3 +#define PCRE_CONFIG_MATCH_LIMIT 4 +#define PCRE_CONFIG_STACKRECURSE 5 +#define PCRE_CONFIG_UNICODE_PROPERTIES 6 +#define PCRE_CONFIG_MATCH_LIMIT_RECURSION 7 +#define PCRE_CONFIG_BSR 8 + +/* Bit flags for the pcre_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 +#define PCRE_EXTRA_MATCH_LIMIT 0x0002 +#define PCRE_EXTRA_CALLOUT_DATA 0x0004 +#define PCRE_EXTRA_TABLES 0x0008 +#define PCRE_EXTRA_MATCH_LIMIT_RECURSION 0x0010 + +/* Types */ + +struct real_pcre; /* declaration; the definition is private */ +typedef struct real_pcre pcre; + +/* 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 *. */ + +#ifndef PCRE_SPTR +#define PCRE_SPTR const char * +#endif + +/* The structure for passing additional data to pcre_exec(). This is defined in +such as way as to be extensible. Always add new fields at the end, in order to +remain compatible. */ + +typedef struct pcre_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_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 +without modification. */ + +typedef struct pcre_callout_block { + int version; /* Identifies version of block */ + /* ------------------------ Version 0 ------------------------------- */ + int callout_number; /* Number compiled into pattern */ + int *offset_vector; /* The offset vector */ + PCRE_SPTR 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 */ + /* ------------------------------------------------------------------ */ +} pcre_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 +that is triggered by the (?) regex item. For Virtual Pascal, these definitions +have to take another form. */ + +#ifndef 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 *); +#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 *); +#endif /* VPCOMPAT */ + +/* Exported PCRE functions */ + +PCRE_EXP_DECL pcre *pcre_compile(const char *, 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 int pcre_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 pcre_dfa_exec(const pcre *, const pcre_extra *, + const char *, 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 void pcre_free_substring(const char *); +PCRE_EXP_DECL void pcre_free_substring_list(const char **); +PCRE_EXP_DECL int pcre_fullinfo(const pcre *, const pcre_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 pcre_get_stringnumber(const pcre *, const char *); +PCRE_EXP_DECL int pcre_get_stringtable_entries(const pcre *, const char *, + char **, char **); +PCRE_EXP_DECL int pcre_get_substring(const char *, int *, int, int, + const char **); +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 const unsigned char *pcre_maketables(void); +PCRE_EXP_DECL int pcre_refcount(pcre *, int); +PCRE_EXP_DECL pcre_extra *pcre_study(const pcre *, int, const char **); +PCRE_EXP_DECL const char *pcre_version(void); + +#ifdef __cplusplus +} /* extern "C" */ +#endif + +#endif /* End of pcre.h */ diff --git a/pcre/pcre_chartables.c b/pcre/pcre_chartables.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_chartables.c @@ -0,0 +1,198 @@ +/************************************************* +* 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 new file mode 100644 --- /dev/null +++ b/pcre/pcre_compile.c @@ -0,0 +1,6379 @@ +/************************************************* +* 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_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; +} + +/* End of pcre_compile.c */ diff --git a/pcre/pcre_config.c b/pcre/pcre_config.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_config.c @@ -0,0 +1,128 @@ +/************************************************* +* 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_config(). */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + +/************************************************* +* Return info about what features are configured * +*************************************************/ + +/* This function has an extensible interface so that additional items can be +added compatibly. + +Arguments: + what what information is required + where where to put the information + +Returns: 0 if data returned, negative on error +*/ + +PCRE_EXP_DEFN int +pcre_config(int what, void *where) +{ +switch (what) + { + case PCRE_CONFIG_UTF8: +#ifdef SUPPORT_UTF8 + *((int *)where) = 1; +#else + *((int *)where) = 0; +#endif + break; + + case PCRE_CONFIG_UNICODE_PROPERTIES: +#ifdef SUPPORT_UCP + *((int *)where) = 1; +#else + *((int *)where) = 0; +#endif + break; + + case PCRE_CONFIG_NEWLINE: + *((int *)where) = NEWLINE; + break; + + case PCRE_CONFIG_BSR: +#ifdef BSR_ANYCRLF + *((int *)where) = 1; +#else + *((int *)where) = 0; +#endif + break; + + case PCRE_CONFIG_LINK_SIZE: + *((int *)where) = LINK_SIZE; + break; + + case PCRE_CONFIG_POSIX_MALLOC_THRESHOLD: + *((int *)where) = POSIX_MALLOC_THRESHOLD; + break; + + case PCRE_CONFIG_MATCH_LIMIT: + *((unsigned int *)where) = MATCH_LIMIT; + break; + + case PCRE_CONFIG_MATCH_LIMIT_RECURSION: + *((unsigned int *)where) = MATCH_LIMIT_RECURSION; + break; + + case PCRE_CONFIG_STACKRECURSE: +#ifdef NO_RECURSE + *((int *)where) = 0; +#else + *((int *)where) = 1; +#endif + break; + + default: return PCRE_ERROR_BADOPTION; + } + +return 0; +} + +/* End of pcre_config.c */ diff --git a/pcre/pcre_dfa_exec.c b/pcre/pcre_dfa_exec.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_dfa_exec.c @@ -0,0 +1,2899 @@ +/************************************************* +* 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_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' && + (re->flags & PCRE_HASCRORLF) == 0 && + (md->nltype == NLTYPE_ANY || + md->nltype == NLTYPE_ANYCRLF || + md->nllen == 2)) + current_subject++; + + } /* "Bumpalong" loop */ + +return PCRE_ERROR_NOMATCH; +} + +/* End of pcre_dfa_exec.c */ diff --git a/pcre/pcre_exec.c b/pcre/pcre_exec.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_exec.c @@ -0,0 +1,4953 @@ +/************************************************* +* 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 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; + } +} + +/* End of pcre_exec.c */ diff --git a/pcre/pcre_fullinfo.c b/pcre/pcre_fullinfo.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_fullinfo.c @@ -0,0 +1,165 @@ +/************************************************* +* 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_fullinfo(), which returns +information about a compiled pattern. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + +/************************************************* +* Return info about compiled pattern * +*************************************************/ + +/* This is a newer "info" function which has an extensible interface so +that additional items can be added compatibly. + +Arguments: + argument_re points to compiled code + extra_data points extra data, or NULL + what what information is required + where where to put the information + +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) +{ +real_pcre internal_re; +pcre_study_data internal_study; +const real_pcre *re = (const real_pcre *)argument_re; +const pcre_study_data *study = NULL; + +if (re == NULL || where == NULL) return PCRE_ERROR_NULL; + +if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_STUDY_DATA) != 0) + study = (const pcre_study_data *)extra_data->study_data; + +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; + } + +switch (what) + { + case PCRE_INFO_OPTIONS: + *((unsigned long int *)where) = re->options & PUBLIC_OPTIONS; + break; + + case PCRE_INFO_SIZE: + *((size_t *)where) = re->size; + break; + + case PCRE_INFO_STUDYSIZE: + *((size_t *)where) = (study == NULL)? 0 : study->size; + break; + + case PCRE_INFO_CAPTURECOUNT: + *((int *)where) = re->top_bracket; + break; + + case PCRE_INFO_BACKREFMAX: + *((int *)where) = re->top_backref; + break; + + case PCRE_INFO_FIRSTBYTE: + *((int *)where) = + ((re->flags & PCRE_FIRSTSET) != 0)? re->first_byte : + ((re->flags & PCRE_STARTLINE) != 0)? -1 : -2; + break; + + /* Make sure we pass back the pointer to the bit vector in the external + 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_study_data *)extra_data->study_data)->start_bits : NULL; + break; + + case PCRE_INFO_LASTLITERAL: + *((int *)where) = + ((re->flags & PCRE_REQCHSET) != 0)? re->req_byte : -1; + break; + + case PCRE_INFO_NAMEENTRYSIZE: + *((int *)where) = re->name_entry_size; + break; + + case PCRE_INFO_NAMECOUNT: + *((int *)where) = re->name_count; + break; + + case PCRE_INFO_NAMETABLE: + *((const uschar **)where) = (const uschar *)re + re->name_table_offset; + break; + + case PCRE_INFO_DEFAULT_TABLES: + *((const uschar **)where) = (const uschar *)(_pcre_default_tables); + break; + + case PCRE_INFO_OKPARTIAL: + *((int *)where) = (re->flags & PCRE_NOPARTIAL) == 0; + break; + + case PCRE_INFO_JCHANGED: + *((int *)where) = (re->flags & PCRE_JCHANGED) != 0; + break; + + case PCRE_INFO_HASCRORLF: + *((int *)where) = (re->flags & PCRE_HASCRORLF) != 0; + break; + + default: return PCRE_ERROR_BADOPTION; + } + +return 0; +} + +/* End of pcre_fullinfo.c */ diff --git a/pcre/pcre_get.c b/pcre/pcre_get.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_get.c @@ -0,0 +1,465 @@ +/************************************************* +* 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 some convenience functions for extracting substrings +from the subject string after a regex match has succeeded. The original idea +for these functions came from Scott Wimer. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + +/************************************************* +* Find number for named string * +*************************************************/ + +/* This function is used by the get_first_set() function below, as well +as being generally available. It assumes that names are unique. + +Arguments: + code the compiled regex + stringname the name whose number is required + +Returns: the number of the named parentheses, or a negative number + (PCRE_ERROR_NOSUBSTRING) if not found +*/ + +int +pcre_get_stringnumber(const pcre *code, const char *stringname) +{ +int rc; +int entrysize; +int top, bot; +uschar *nametable; + +if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMECOUNT, &top)) != 0) + return rc; +if (top <= 0) return PCRE_ERROR_NOSUBSTRING; + +if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMEENTRYSIZE, &entrysize)) != 0) + return rc; +if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMETABLE, &nametable)) != 0) + return rc; + +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]; + if (c > 0) bot = mid + 1; else top = mid; + } + +return PCRE_ERROR_NOSUBSTRING; +} + + + +/************************************************* +* Find (multiple) entries for named string * +*************************************************/ + +/* This is used by the get_first_set() function below, as well as being +generally available. It is used when duplicated names are permitted. + +Arguments: + code the compiled regex + stringname the name whose entries required + firstptr where to put the pointer to the first entry + lastptr where to put the pointer to the last entry + +Returns: the length of each entry, or a negative number + (PCRE_ERROR_NOSUBSTRING) if not found +*/ + +int +pcre_get_stringtable_entries(const pcre *code, const char *stringname, + char **firstptr, char **lastptr) +{ +int rc; +int entrysize; +int top, bot; +uschar *nametable, *lastentry; + +if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMECOUNT, &top)) != 0) + return rc; +if (top <= 0) return PCRE_ERROR_NOSUBSTRING; + +if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMEENTRYSIZE, &entrysize)) != 0) + return rc; +if ((rc = pcre_fullinfo(code, NULL, PCRE_INFO_NAMETABLE, &nametable)) != 0) + return rc; + +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)); + if (c == 0) + { + uschar *first = entry; + uschar *last = entry; + while (first > nametable) + { + if (strcmp(stringname, (char *)(first - entrysize + 2)) != 0) break; + first -= entrysize; + } + while (last < lastentry) + { + if (strcmp(stringname, (char *)(last + entrysize + 2)) != 0) break; + last += entrysize; + } + *firstptr = (char *)first; + *lastptr = (char *)last; + return entrysize; + } + if (c > 0) bot = mid + 1; else top = mid; + } + +return PCRE_ERROR_NOSUBSTRING; +} + + + +/************************************************* +* Find first set of multiple named strings * +*************************************************/ + +/* This function allows for duplicate names in the table of named substrings. +It returns the number of the first one that was set in a pattern match. + +Arguments: + code the compiled regex + stringname the name of the capturing substring + ovector the vector of matched substrings + +Returns: the number of the first that is set, + or the number of the last one if none are set, + or a negative number on error +*/ + +static int +get_first_set(const pcre *code, const char *stringname, int *ovector) +{ +const real_pcre *re = (const real_pcre *)code; +int entrysize; +char *first, *last; +uschar *entry; +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); +if (entrysize <= 0) return entrysize; +for (entry = (uschar *)first; entry <= (uschar *)last; entry += entrysize) + { + int n = (entry[0] << 8) + entry[1]; + if (ovector[n*2] >= 0) return n; + } +return (first[0] << 8) + first[1]; +} + + + + +/************************************************* +* Copy captured string to given buffer * +*************************************************/ + +/* This function copies a single captured substring into a given buffer. +Note that we use memcpy() rather than strncpy() in case there are binary zeros +in the string. + +Arguments: + subject the subject string that was matched + ovector pointer to the offsets table + stringcount the number of substrings that were captured + (i.e. the yield of the pcre_exec call, unless + that was zero, in which case it should be 1/3 + of the offset table size) + stringnumber the number of the required substring + buffer where to put the substring + size the size of the buffer + +Returns: if successful: + the length of the copied string, not including the zero + that is put on the end; can be zero + if not successful: + PCRE_ERROR_NOMEMORY (-6) buffer too small + PCRE_ERROR_NOSUBSTRING (-7) no such captured substring +*/ + +int +pcre_copy_substring(const char *subject, int *ovector, int stringcount, + int stringnumber, char *buffer, int size) +{ +int yield; +if (stringnumber < 0 || stringnumber >= stringcount) + return PCRE_ERROR_NOSUBSTRING; +stringnumber *= 2; +yield = ovector[stringnumber+1] - ovector[stringnumber]; +if (size < yield + 1) return PCRE_ERROR_NOMEMORY; +memcpy(buffer, subject + ovector[stringnumber], yield); +buffer[yield] = 0; +return yield; +} + + + +/************************************************* +* Copy named captured string to given buffer * +*************************************************/ + +/* This function copies a single captured substring into a given buffer, +identifying it by name. If the regex permits duplicate names, the first +substring that is set is chosen. + +Arguments: + code the compiled regex + subject the subject string that was matched + ovector pointer to the offsets table + stringcount the number of substrings that were captured + (i.e. the yield of the pcre_exec call, unless + that was zero, in which case it should be 1/3 + of the offset table size) + stringname the name of the required substring + buffer where to put the substring + size the size of the buffer + +Returns: if successful: + the length of the copied string, not including the zero + that is put on the end; can be zero + if not successful: + PCRE_ERROR_NOMEMORY (-6) buffer too small + 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) +{ +int n = get_first_set(code, stringname, ovector); +if (n <= 0) return n; +return pcre_copy_substring(subject, ovector, stringcount, n, buffer, size); +} + + + +/************************************************* +* Copy all captured strings to new store * +*************************************************/ + +/* This function gets one chunk of store and builds a list of pointers and all +of the captured substrings in it. A NULL pointer is put on the end of the list. + +Arguments: + subject the subject string that was matched + ovector pointer to the offsets table + stringcount the number of substrings that were captured + (i.e. the yield of the pcre_exec call, unless + that was zero, in which case it should be 1/3 + of the offset table size) + listptr set to point to the list of pointers + +Returns: if successful: 0 + if not successful: + PCRE_ERROR_NOMEMORY (-6) failed to get store +*/ + +int +pcre_get_substring_list(const char *subject, int *ovector, int stringcount, + const char ***listptr) +{ +int i; +int size = sizeof(char *); +int double_count = stringcount * 2; +char **stringlist; +char *p; + +for (i = 0; i < double_count; i += 2) + size += sizeof(char *) + ovector[i+1] - ovector[i] + 1; + +stringlist = (char **)(pcre_malloc)(size); +if (stringlist == NULL) return PCRE_ERROR_NOMEMORY; + +*listptr = (const char **)stringlist; +p = (char *)(stringlist + stringcount + 1); + +for (i = 0; i < double_count; i += 2) + { + int len = ovector[i+1] - ovector[i]; + memcpy(p, subject + ovector[i], len); + *stringlist++ = p; + p += len; + *p++ = 0; + } + +*stringlist = NULL; +return 0; +} + + + +/************************************************* +* Free store obtained by get_substring_list * +*************************************************/ + +/* 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. + +Argument: the result of a previous pcre_get_substring_list() +Returns: nothing +*/ + +void +pcre_free_substring_list(const char **pointer) +{ +(pcre_free)((void *)pointer); +} + + + +/************************************************* +* Copy captured string to new store * +*************************************************/ + +/* This function copies a single captured substring into a piece of new +store + +Arguments: + subject the subject string that was matched + ovector pointer to the offsets table + stringcount the number of substrings that were captured + (i.e. the yield of the pcre_exec call, unless + that was zero, in which case it should be 1/3 + of the offset table size) + stringnumber the number of the required substring + stringptr where to put a pointer to the substring + +Returns: if successful: + the length of the string, not including the zero that + is put on the end; can be zero + if not successful: + PCRE_ERROR_NOMEMORY (-6) failed to get store + PCRE_ERROR_NOSUBSTRING (-7) substring not present +*/ + +int +pcre_get_substring(const char *subject, int *ovector, int stringcount, + int stringnumber, const char **stringptr) +{ +int yield; +char *substring; +if (stringnumber < 0 || stringnumber >= stringcount) + return PCRE_ERROR_NOSUBSTRING; +stringnumber *= 2; +yield = ovector[stringnumber+1] - ovector[stringnumber]; +substring = (char *)(pcre_malloc)(yield + 1); +if (substring == NULL) return PCRE_ERROR_NOMEMORY; +memcpy(substring, subject + ovector[stringnumber], yield); +substring[yield] = 0; +*stringptr = substring; +return yield; +} + + + +/************************************************* +* Copy named captured string to new store * +*************************************************/ + +/* This function copies a single captured substring, identified by name, into +new store. If the regex permits duplicate names, the first substring that is +set is chosen. + +Arguments: + code the compiled regex + subject the subject string that was matched + ovector pointer to the offsets table + stringcount the number of substrings that were captured + (i.e. the yield of the pcre_exec call, unless + that was zero, in which case it should be 1/3 + of the offset table size) + stringname the name of the required substring + stringptr where to put the pointer + +Returns: if successful: + the length of the copied string, not including the zero + that is put on the end; can be zero + if not successful: + PCRE_ERROR_NOMEMORY (-6) couldn't get memory + 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) +{ +int n = get_first_set(code, stringname, ovector); +if (n <= 0) return n; +return pcre_get_substring(subject, ovector, stringcount, n, stringptr); +} + + + + +/************************************************* +* Free store obtained by get_substring * +*************************************************/ + +/* 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. + +Argument: the result of a previous pcre_get_substring() +Returns: nothing +*/ + +void +pcre_free_substring(const char *pointer) +{ +(pcre_free)((void *)pointer); +} + +/* End of pcre_get.c */ diff --git a/pcre/pcre_globals.c b/pcre/pcre_globals.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_globals.c @@ -0,0 +1,63 @@ +/************************************************* +* 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 global variables that are exported by the PCRE library. +PCRE is thread-clean and doesn't use any global variables in the normal sense. +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). */ + +#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; +#endif + +/* End of pcre_globals.c */ diff --git a/pcre/pcre_info.c b/pcre/pcre_info.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_info.c @@ -0,0 +1,93 @@ +/************************************************* +* 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 new file mode 100644 --- /dev/null +++ b/pcre/pcre_internal.h @@ -0,0 +1,1140 @@ +/************************************************* +* 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 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 *); + +#endif + +/* End of pcre_internal.h */ diff --git a/pcre/pcre_maketables.c b/pcre/pcre_maketables.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_maketables.c @@ -0,0 +1,143 @@ +/************************************************* +* 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_maketables(), which builds +character tables for PCRE in the current locale. The file is compiled on its +own as part of the PCRE library. However, it is also included in the +compilation of dftables.c, in which case the macro DFTABLES is defined. */ + + +#ifndef DFTABLES +# ifdef HAVE_CONFIG_H +# include "config.h" +# endif +# include "pcre_internal.h" +#endif + + +/************************************************* +* Create PCRE character tables * +*************************************************/ + +/* 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(). + +Arguments: none +Returns: pointer to the contiguous block of data +*/ + +const unsigned char * +pcre_maketables(void) +{ +unsigned char *yield, *p; +int i; + +#ifndef DFTABLES +yield = (unsigned char*)(pcre_malloc)(tables_length); +#else +yield = (unsigned char*)malloc(tables_length); +#endif + +if (yield == NULL) return NULL; +p = yield; + +/* First comes the lower casing table */ + +for (i = 0; i < 256; i++) *p++ = tolower(i); + +/* Next the case-flipping table */ + +for (i = 0; i < 256; i++) *p++ = islower(i)? toupper(i) : tolower(i); + +/* Then the character class tables. Don't try to be clever and save effort on +exclusive ones - in some locales things may be different. Note that the table +for "space" includes everything "isspace" gives, including VT in the default +locale. This makes it work for the POSIX class [:space:]. Note also that it is +possible for a character to be alnum or alpha without being lower or upper, +such as "male and female ordinals" (\xAA and \xBA) in the fr_FR locale (at +least under Debian Linux's locales as of 12/2005). So we must test for alnum +specially. */ + +memset(p, 0, cbit_length); +for (i = 0; i < 256; i++) + { + if (isdigit(i)) p[cbit_digit + i/8] |= 1 << (i&7); + if (isupper(i)) p[cbit_upper + i/8] |= 1 << (i&7); + if (islower(i)) p[cbit_lower + i/8] |= 1 << (i&7); + if (isalnum(i)) p[cbit_word + i/8] |= 1 << (i&7); + if (i == '_') p[cbit_word + i/8] |= 1 << (i&7); + if (isspace(i)) p[cbit_space + i/8] |= 1 << (i&7); + if (isxdigit(i))p[cbit_xdigit + i/8] |= 1 << (i&7); + if (isgraph(i)) p[cbit_graph + i/8] |= 1 << (i&7); + if (isprint(i)) p[cbit_print + i/8] |= 1 << (i&7); + if (ispunct(i)) p[cbit_punct + i/8] |= 1 << (i&7); + if (iscntrl(i)) p[cbit_cntrl + i/8] |= 1 << (i&7); + } +p += cbit_length; + +/* Finally, the character type table. In this, we exclude VT from the white +space chars, because Perl doesn't recognize it as such for \s and for comments +within regexes. */ + +for (i = 0; i < 256; i++) + { + int x = 0; + if (i != 0x0b && isspace(i)) x += ctype_space; + if (isalpha(i)) x += ctype_letter; + if (isdigit(i)) x += ctype_digit; + if (isxdigit(i)) x += ctype_xdigit; + if (isalnum(i) || i == '_') x += ctype_word; + + /* Note: strchr includes the terminating zero in the characters it considers. + In this instance, that is ok because we want binary zero to be flagged as a + meta-character, which in this sense is any character that terminates a run + of data characters. */ + + if (strchr("\\*+?{^.$|()[", i) != 0) x += ctype_meta; + *p++ = x; + } + +return yield; +} + +/* End of pcre_maketables.c */ diff --git a/pcre/pcre_newline.c b/pcre/pcre_newline.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_newline.c @@ -0,0 +1,164 @@ +/************************************************* +* 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 internal functions for testing newlines when more than +one kind of newline is to be recognized. When a newline is found, its length is +returned. In principle, we could implement several newline "types", each +referring to a different set of newline characters. At present, PCRE supports +only NLTYPE_FIXED, which gets handled without these functions, NLTYPE_ANYCRLF, +and NLTYPE_ANY. The full list of Unicode newline characters is taken from +http://unicode.org/unicode/reports/tr18/. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + + +/************************************************* +* Check for newline at given position * +*************************************************/ + +/* It is guaranteed that the initial value of ptr is less than the end of the +string that is being processed. + +Arguments: + ptr pointer to possible newline + type the newline type + endptr pointer to the end of the string + lenptr where to return the length + utf8 TRUE if in utf8 mode + +Returns: TRUE or FALSE +*/ + +BOOL +_pcre_is_newline(const uschar *ptr, int type, const uschar *endptr, + int *lenptr, BOOL utf8) +{ +int c; +if (utf8) { GETCHAR(c, ptr); } else c = *ptr; + +if (type == NLTYPE_ANYCRLF) switch(c) + { + case 0x000a: *lenptr = 1; return TRUE; /* LF */ + case 0x000d: *lenptr = (ptr < endptr - 1 && ptr[1] == 0x0a)? 2 : 1; + return TRUE; /* CR */ + default: return FALSE; + } + +/* NLTYPE_ANY */ + +else switch(c) + { + case 0x000a: /* LF */ + case 0x000b: /* VT */ + 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 */ + case 0x2028: /* LS */ + case 0x2029: *lenptr = 3; return TRUE; /* PS */ + default: return FALSE; + } +} + + + +/************************************************* +* Check for newline at previous position * +*************************************************/ + +/* It is guaranteed that the initial value of ptr is greater than the start of +the string that is being processed. + +Arguments: + ptr pointer to possible newline + type the newline type + startptr pointer to the start of the string + lenptr where to return the length + utf8 TRUE if in utf8 mode + +Returns: TRUE or FALSE +*/ + +BOOL +_pcre_was_newline(const uschar *ptr, int type, const uschar *startptr, + int *lenptr, BOOL utf8) +{ +int c; +ptr--; +#ifdef SUPPORT_UTF8 +if (utf8) + { + BACKCHAR(ptr); + GETCHAR(c, ptr); + } +else c = *ptr; +#else /* no UTF-8 support */ +c = *ptr; +#endif /* SUPPORT_UTF8 */ + +if (type == NLTYPE_ANYCRLF) switch(c) + { + case 0x000a: *lenptr = (ptr > startptr && ptr[-1] == 0x0d)? 2 : 1; + return TRUE; /* LF */ + case 0x000d: *lenptr = 1; return TRUE; /* CR */ + default: return FALSE; + } + +else switch(c) + { + case 0x000a: *lenptr = (ptr > startptr && ptr[-1] == 0x0d)? 2 : 1; + return TRUE; /* LF */ + case 0x000b: /* VT */ + case 0x000c: /* FF */ + case 0x000d: *lenptr = 1; return TRUE; /* CR */ + case 0x0085: *lenptr = utf8? 2 : 1; return TRUE; /* NEL */ + case 0x2028: /* LS */ + case 0x2029: *lenptr = 3; return TRUE; /* PS */ + default: return FALSE; + } +} + +/* End of pcre_newline.c */ diff --git a/pcre/pcre_ord2utf8.c b/pcre/pcre_ord2utf8.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_ord2utf8.c @@ -0,0 +1,85 @@ +/************************************************* +* 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 file contains a private PCRE function that converts an ordinal +character value into a UTF8 string. */ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + +/************************************************* +* 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 + buffer pointer to buffer for result - at least 6 bytes long + +Returns: number of characters placed in the buffer +*/ + +int +_pcre_ord2utf8(int cvalue, uschar *buffer) +{ +#ifdef SUPPORT_UTF8 +register int i, j; +for (i = 0; i < _pcre_utf8_table1_size; i++) + if (cvalue <= _pcre_utf8_table1[i]) break; +buffer += i; +for (j = i; j > 0; j--) + { + *buffer-- = 0x80 | (cvalue & 0x3f); + cvalue >>= 6; + } +*buffer = _pcre_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. */ +} + +/* End of pcre_ord2utf8.c */ diff --git a/pcre/pcre_refcount.c b/pcre/pcre_refcount.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_refcount.c @@ -0,0 +1,82 @@ +/************************************************* +* 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_refcount(), which is an +auxiliary function that can be used to maintain a reference count in a compiled +pattern data block. This might be helpful in applications where the block is +shared by different users. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + +/************************************************* +* Maintain reference count * +*************************************************/ + +/* The reference count is a 16-bit field, initialized to zero. It is not +possible to transfer a non-zero count from one host to a different host that +has a different byte order - though I can't see why anyone in their right mind +would ever want to do that! + +Arguments: + argument_re points to compiled code + adjust value to add to the count + +Returns: the (possibly updated) count value (a non-negative number), or + a negative error number +*/ + +PCRE_EXP_DEFN int +pcre_refcount(pcre *argument_re, int adjust) +{ +real_pcre *re = (real_pcre *)argument_re; +if (re == NULL) return PCRE_ERROR_NULL; +re->ref_count = (-adjust > re->ref_count)? 0 : + (adjust + re->ref_count > 65535)? 65535 : + re->ref_count + adjust; +return re->ref_count; +} + +/* End of pcre_refcount.c */ diff --git a/pcre/pcre_scanner.h b/pcre/pcre_scanner.h new file mode 100644 --- /dev/null +++ b/pcre/pcre_scanner.h @@ -0,0 +1,172 @@ +// 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 +// +// Regular-expression based scanner for parsing an input stream. +// +// Example 1: parse a sequence of "var = number" entries from input: +// +// Scanner scanner(input); +// string var; +// int number; +// scanner.SetSkipExpression("\\s+"); // Skip any white space we encounter +// while (scanner.Consume("(\\w+) = (\\d+)", &var, &number)) { +// ...; +// } + +#ifndef _PCRE_SCANNER_H +#define _PCRE_SCANNER_H + +#include +#include +#include + +#include +#include + +namespace pcrecpp { + +class PCRECPP_EXP_DEFN Scanner { + public: + Scanner(); + explicit Scanner(const std::string& input); + ~Scanner(); + + // Return current line number. The returned line-number is + // one-based. I.e. it returns 1 + the number of consumed newlines. + // + // Note: this method may be slow. It may take time proportional to + // the size of the input. + int LineNumber() const; + + // Return the byte-offset that the scanner is looking in the + // input data; + int Offset() const; + + // Return true iff the start of the remaining input matches "re" + bool LookingAt(const RE& re) const; + + // Return true iff all of the following are true + // a. the start of the remaining input matches "re", + // b. if any arguments are supplied, matched sub-patterns can be + // parsed and stored into the arguments. + // If it returns true, it skips over the matched input and any + // following input that matches the "skip" regular expression. + bool Consume(const RE& re, + const Arg& arg0 = RE::no_arg, + const Arg& arg1 = RE::no_arg, + const Arg& arg2 = RE::no_arg + // TODO: Allow more arguments? + ); + + // Set the "skip" regular expression. If after consuming some data, + // a prefix of the input matches this RE, it is automatically + // skipped. For example, a programming language scanner would use + // a skip RE that matches white space and comments. + // + // scanner.SetSkipExpression("\\s+|//.*|/[*](.|\n)*?[*]/"); + // + // Skipping repeats as long as it succeeds. We used to let people do + // this by writing "(...)*" in the regular expression, but that added + // up to lots of recursive calls within the pcre library, so now we + // control repetition explicitly via the function call API. + // + // You can pass NULL for "re" if you do not want any data to be skipped. + void Skip(const char* re); // DEPRECATED; does *not* repeat + void SetSkipExpression(const char* re); + + // Temporarily pause "skip"ing. This + // Skip("Foo"); code ; DisableSkip(); code; EnableSkip() + // is similar to + // Skip("Foo"); code ; Skip(NULL); code ; Skip("Foo"); + // but avoids creating/deleting new RE objects. + void DisableSkip(); + + // Reenable previously paused skipping. Any prefix of the input + // that matches the skip pattern is immediately dropped. + void EnableSkip(); + + /***** Special wrappers around SetSkip() for some common idioms *****/ + + // Arranges to skip whitespace, C comments, C++ comments. + // The overall RE is a disjunction of the following REs: + // \\s whitespace + // //.*\n C++ comment + // /[*](.|\n)*?[*]/ C comment (x*? means minimal repetitions of x) + // We get repetition via the semantics of SetSkipExpression, not by using * + void SkipCXXComments() { + SetSkipExpression("\\s|//.*\n|/[*](?:\n|.)*?[*]/"); + } + + void set_save_comments(bool comments) { + save_comments_ = comments; + } + + bool save_comments() { + return save_comments_; + } + + // Append to vector ranges the comments found in the + // byte range [start,end] (inclusive) of the input data. + // Only comments that were extracted entirely within that + // range are returned: no range splitting of atomically-extracted + // comments is performed. + void GetComments(int start, int end, std::vector *ranges); + + // Append to vector ranges the comments added + // since the last time this was called. This + // functionality is provided for efficiency when + // interleaving scanning with parsing. + void GetNextComments(std::vector *ranges); + + private: + std::string data_; // All the input data + StringPiece input_; // Unprocessed input + RE* skip_; // If non-NULL, RE for skipping input + bool should_skip_; // If true, use skip_ + bool skip_repeat_; // If true, repeat skip_ as long as it works + bool save_comments_; // If true, aggregate the skip expression + + // the skipped comments + // TODO: later consider requiring that the StringPieces be added + // in order by their start position + std::vector *comments_; + + // the offset into comments_ that has been returned by GetNextComments + int comments_offset_; + + // helper function to consume *skip_ and honour + // save_comments_ + void ConsumeSkip(); +}; + +} // namespace pcrecpp + +#endif /* _PCRE_SCANNER_H */ diff --git a/pcre/pcre_stringpiece.h b/pcre/pcre_stringpiece.h new file mode 100644 --- /dev/null +++ b/pcre/pcre_stringpiece.h @@ -0,0 +1,177 @@ +// 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 new file mode 100644 --- /dev/null +++ b/pcre/pcre_study.c @@ -0,0 +1,587 @@ +/************************************************* +* 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_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; +} + +/* End of pcre_study.c */ diff --git a/pcre/pcre_tables.c b/pcre/pcre_tables.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_tables.c @@ -0,0 +1,318 @@ +/************************************************* +* 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 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 +uses macros to change their names from _pcre_xxx to xxxx, thereby avoiding name +clashes with the library. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + +/* 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 }; + + + +/************************************************* +* Tables for UTF-8 support * +*************************************************/ + +/* These are the breakpoints for different numbers of bytes in a UTF-8 +character. */ + +#ifdef SUPPORT_UTF8 + +const int _pcre_utf8_table1[] = + { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff}; + +const int _pcre_utf8_table1_size = sizeof(_pcre_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}; + +/* 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[] = { + 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 }; + +/* 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 +field of each entry. However, that leads to a large number of relocations when +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. */ + +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"; + +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 } +}; + +const int _pcre_utt_size = sizeof(_pcre_utt)/sizeof(ucp_type_table); + +#endif /* SUPPORT_UTF8 */ + +/* End of pcre_tables.c */ diff --git a/pcre/pcre_try_flipped.c b/pcre/pcre_try_flipped.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_try_flipped.c @@ -0,0 +1,137 @@ +/************************************************* +* 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_ucp_searchfuncs.c b/pcre/pcre_ucp_searchfuncs.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_ucp_searchfuncs.c @@ -0,0 +1,179 @@ +/************************************************* +* 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 new file mode 100644 --- /dev/null +++ b/pcre/pcre_valid_utf8.c @@ -0,0 +1,162 @@ +/************************************************* +* 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 for validating UTF-8 character +strings. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + +/************************************************* +* Validate a UTF-8 string * +*************************************************/ + +/* 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 +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. + +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. + +Arguments: + string points to the string + length length of string, or -1 if the string is zero-terminated + +Returns: < 0 if the string is a valid UTF-8 string + >= 0 otherwise; the value is the offset of the bad byte +*/ + +int +_pcre_valid_utf8(const uschar *string, int length) +{ +#ifdef SUPPORT_UTF8 +register const uschar *p; + +if (length < 0) + { + for (p = string; *p != 0; p++); + length = 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; + + /* 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. */ + + switch (ab) + { + /* Check 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 */ + + /* Check for 1110 0000, xx0x xxxx (overlong sequence) or + 1110 1101, 1010 xxxx (0xd000 - 0xdfff) */ + + case 2: + if ((c == 0xe0 && (*p & 0x20) == 0) || + (c == 0xed && *p >= 0xa0)) + return p - string; + break; + + /* Check for 1111 0000, xx00 xxxx (overlong sequence) or + greater than 0x0010ffff (f4 8f bf bf) */ + + case 3: + if ((c == 0xf0 && (*p & 0x30) == 0) || + (c > 0xf4 ) || + (c == 0xf4 && *p > 0x8f)) + return p - string; + 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. */ + + /* Check for 1111 1000, xx00 0xxx */ + case 4: + if (c == 0xf8 && (*p & 0x38) == 0) return p - string; + break; + + /* Check for leading 0xfe or 0xff, and then for 1111 1100, xx00 00xx */ + case 5: + if (c == 0xfe || c == 0xff || + (c == 0xfc && (*p & 0x3c) == 0)) return p - string; + break; +#endif + + } + + /* Check for valid bytes after the 2nd, if any; all must start 10 */ + while (--ab > 0) + { + if ((*(++p) & 0xc0) != 0x80) return p - string; + } + } +#endif + +return -1; +} + +/* End of pcre_valid_utf8.c */ diff --git a/pcre/pcre_version.c b/pcre/pcre_version.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_version.c @@ -0,0 +1,90 @@ +/************************************************* +* 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_version(), which returns a +string that identifies the PCRE version that is in use. */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + +/************************************************* +* Return version string * +*************************************************/ + +/* These macros are the standard way of turning unquoted text into C strings. +They allow macros like PCRE_MAJOR to be defined without quotes, which is +convenient for user programs that want to test its value. */ + +#define STRING(a) # a +#define XSTRING(s) STRING(s) + +/* A problem turned up with PCRE_PRERELEASE, which is defined empty for +production releases. Originally, it was used naively in this code: + + return XSTRING(PCRE_MAJOR) + "." XSTRING(PCRE_MINOR) + XSTRING(PCRE_PRERELEASE) + " " XSTRING(PCRE_DATE); + +However, when PCRE_PRERELEASE is empty, this leads to an attempted expansion of +STRING(). The C standard states: "If (before argument substitution) any +argument consists of no preprocessing tokens, the behavior is undefined." It +turns out the gcc treats this case as a single empty string - which is what we +really want - but Visual C grumbles about the lack of an argument for the +macro. Unfortunately, both are within their rights. To cope with both ways of +handling this, I had resort to some messy hackery that does a test at run time. +I could find no way of detecting that a macro is defined as an empty string at +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 * +pcre_version(void) +{ +return (XSTRING(Z PCRE_PRERELEASE)[1] == 0)? + XSTRING(PCRE_MAJOR.PCRE_MINOR PCRE_DATE) : + XSTRING(PCRE_MAJOR.PCRE_MINOR) XSTRING(PCRE_PRERELEASE PCRE_DATE); +} + +/* End of pcre_version.c */ diff --git a/pcre/pcre_xclass.c b/pcre/pcre_xclass.c new file mode 100644 --- /dev/null +++ b/pcre/pcre_xclass.c @@ -0,0 +1,148 @@ +/************************************************* +* 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 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(). */ + + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "pcre_internal.h" + + +/************************************************* +* Match character against an XCLASS * +*************************************************/ + +/* This function is called to match a character against an extended class that +might contain values > 255. + +Arguments: + c the character + data points to the flag byte of the XCLASS data + +Returns: TRUE if character matches, else FALSE +*/ + +BOOL +_pcre_xclass(int c, const uschar *data) +{ +int t; +BOOL negated = (*data & XCL_NOT) != 0; + +/* 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 */ + } + +/* 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; + +while ((t = *data++) != XCL_END) + { + int x, y; + if (t == XCL_SINGLE) + { + GETCHARINC(x, data); + if (c == x) return !negated; + } + else if (t == XCL_RANGE) + { + GETCHARINC(x, data); + GETCHARINC(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); + + switch(*data) + { + case PT_ANY: + if (t == XCL_PROP) return !negated; + break; + + case PT_LAMP: + if ((chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt) == + (t == XCL_PROP)) return !negated; + break; + + case PT_GC: + if ((data[1] == category) == (t == XCL_PROP)) return !negated; + break; + + case PT_PC: + if ((data[1] == chartype) == (t == XCL_PROP)) return !negated; + break; + + case PT_SC: + if ((data[1] == script) == (t == XCL_PROP)) return !negated; + break; + + /* This should never occur, but compilers may mutter if there is no + default. */ + + default: + return FALSE; + } + + data += 2; + } +#endif /* SUPPORT_UCP */ + } + +return negated; /* char did not match */ +} + +/* End of pcre_xclass.c */ diff --git a/pcre/pcrecpp.h b/pcre/pcrecpp.h new file mode 100644 --- /dev/null +++ b/pcre/pcrecpp.h @@ -0,0 +1,703 @@ +// 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 +// Support for PCRE_XXX modifiers added by Giuseppe Maxia, July 2005 + +#ifndef _PCRECPP_H +#define _PCRECPP_H + +// C++ interface to the pcre regular-expression library. RE supports +// Perl-style regular expressions (with extensions like \d, \w, \s, +// ...). +// +// ----------------------------------------------------------------------- +// REGEXP SYNTAX: +// +// This module is part of the pcre library and hence supports its syntax +// for regular expressions. +// +// The syntax is pretty similar to Perl's. For those not familiar +// with Perl's regular expressions, here are some examples of the most +// commonly used extensions: +// +// "hello (\\w+) world" -- \w matches a "word" character +// "version (\\d+)" -- \d matches a digit +// "hello\\s+world" -- \s matches any whitespace character +// "\\b(\\w+)\\b" -- \b matches empty string at a word boundary +// "(?i)hello" -- (?i) turns on case-insensitive matching +// "/\\*(.*?)\\*/" -- .*? matches . minimum no. of times possible +// +// ----------------------------------------------------------------------- +// MATCHING INTERFACE: +// +// The "FullMatch" operation checks that supplied text matches a +// supplied pattern exactly. +// +// Example: successful match +// pcrecpp::RE re("h.*o"); +// re.FullMatch("hello"); +// +// Example: unsuccessful match (requires full match): +// pcrecpp::RE re("e"); +// !re.FullMatch("hello"); +// +// Example: creating a temporary RE object: +// pcrecpp::RE("h.*o").FullMatch("hello"); +// +// You can pass in a "const char*" or a "string" for "text". The +// examples below tend to use a const char*. +// +// You can, as in the different examples above, store the RE object +// explicitly in a variable or use a temporary RE object. The +// examples below use one mode or the other arbitrarily. Either +// could correctly be used for any of these examples. +// +// ----------------------------------------------------------------------- +// MATCHING WITH SUB-STRING EXTRACTION: +// +// You can supply extra pointer arguments to extract matched subpieces. +// +// Example: extracts "ruby" into "s" and 1234 into "i" +// int i; +// string s; +// pcrecpp::RE re("(\\w+):(\\d+)"); +// re.FullMatch("ruby:1234", &s, &i); +// +// Example: does not try to extract any extra sub-patterns +// re.FullMatch("ruby:1234", &s); +// +// Example: does not try to extract into NULL +// re.FullMatch("ruby:1234", NULL, &i); +// +// Example: integer overflow causes failure +// !re.FullMatch("ruby:1234567891234", NULL, &i); +// +// Example: fails because there aren't enough sub-patterns: +// !pcrecpp::RE("\\w+:\\d+").FullMatch("ruby:1234", &s); +// +// Example: fails because string cannot be stored in integer +// !pcrecpp::RE("(.*)").FullMatch("ruby", &i); +// +// The provided pointer arguments can be pointers to any scalar numeric +// type, or one of +// string (matched piece is copied to string) +// StringPiece (StringPiece is mutated to point to matched piece) +// T (where "bool T::ParseFrom(const char*, int)" exists) +// NULL (the corresponding matched sub-pattern is not copied) +// +// CAVEAT: An optional sub-pattern that does not exist in the matched +// string is assigned the empty string. Therefore, the following will +// return false (because the empty string is not a valid number): +// int number; +// pcrecpp::RE::FullMatch("abc", "[a-z]+(\\d+)?", &number); +// +// ----------------------------------------------------------------------- +// DO_MATCH +// +// The matching interface supports at most 16 arguments per call. +// If you need more, consider using the more general interface +// pcrecpp::RE::DoMatch(). See pcrecpp.h for the signature for DoMatch. +// +// ----------------------------------------------------------------------- +// PARTIAL MATCHES +// +// You can use the "PartialMatch" operation when you want the pattern +// to match any substring of the text. +// +// Example: simple search for a string: +// pcrecpp::RE("ell").PartialMatch("hello"); +// +// Example: find first number in a string: +// int number; +// pcrecpp::RE re("(\\d+)"); +// re.PartialMatch("x*100 + 20", &number); +// assert(number == 100); +// +// ----------------------------------------------------------------------- +// UTF-8 AND THE MATCHING INTERFACE: +// +// By default, pattern and text are plain text, one byte per character. +// The UTF8 flag, passed to the constructor, causes both pattern +// and string to be treated as UTF-8 text, still a byte stream but +// potentially multiple bytes per character. In practice, the text +// is likelier to be UTF-8 than the pattern, but the match returned +// may depend on the UTF8 flag, so always use it when matching +// UTF8 text. E.g., "." will match one byte normally but with UTF8 +// set may match up to three bytes of a multi-byte character. +// +// Example: +// pcrecpp::RE_Options options; +// options.set_utf8(); +// pcrecpp::RE re(utf8_pattern, options); +// re.FullMatch(utf8_string); +// +// Example: using the convenience function UTF8(): +// pcrecpp::RE re(utf8_pattern, pcrecpp::UTF8()); +// re.FullMatch(utf8_string); +// +// NOTE: The UTF8 option is ignored if pcre was not configured with the +// --enable-utf8 flag. +// +// ----------------------------------------------------------------------- +// PASSING MODIFIERS TO THE REGULAR EXPRESSION ENGINE +// +// PCRE defines some modifiers to change the behavior of the regular +// expression engine. +// The C++ wrapper defines an auxiliary class, RE_Options, as a vehicle +// to pass such modifiers to a RE class. +// +// Currently, the following modifiers are supported +// +// modifier description Perl corresponding +// +// PCRE_CASELESS case insensitive match /i +// PCRE_MULTILINE multiple lines match /m +// PCRE_DOTALL dot matches newlines /s +// PCRE_DOLLAR_ENDONLY $ matches only at end N/A +// PCRE_EXTRA strict escape parsing N/A +// PCRE_EXTENDED ignore whitespaces /x +// PCRE_UTF8 handles UTF8 chars built-in +// PCRE_UNGREEDY reverses * and *? N/A +// PCRE_NO_AUTO_CAPTURE disables matching parens N/A (*) +// +// (For a full account on how each modifier works, please check the +// PCRE API reference manual). +// +// (*) Both Perl and PCRE allow non matching parentheses by means of the +// "?:" modifier within the pattern itself. e.g. (?:ab|cd) does not +// capture, while (ab|cd) does. +// +// For each modifier, there are two member functions whose name is made +// out of the modifier in lowercase, without the "PCRE_" prefix. For +// instance, PCRE_CASELESS is handled by +// bool caseless(), +// which returns true if the modifier is set, and +// RE_Options & set_caseless(bool), +// which sets or unsets the modifier. +// +// Moreover, PCRE_EXTRA_MATCH_LIMIT can be accessed through the +// set_match_limit() and match_limit() member functions. +// Setting match_limit to a non-zero value will limit the executation of +// pcre to keep it from doing bad things like blowing the stack or taking +// an eternity to return a result. A value of 5000 is good enough to stop +// stack blowup in a 2MB thread stack. Setting match_limit to zero will +// disable match limiting. Alternately, you can set match_limit_recursion() +// which uses PCRE_EXTRA_MATCH_LIMIT_RECURSION to limit how much pcre +// recurses. match_limit() caps the number of matches pcre does; +// match_limit_recrusion() caps the depth of recursion. +// +// Normally, to pass one or more modifiers to a RE class, you declare +// a RE_Options object, set the appropriate options, and pass this +// object to a RE constructor. Example: +// +// RE_options opt; +// opt.set_caseless(true); +// +// if (RE("HELLO", opt).PartialMatch("hello world")) ... +// +// RE_options has two constructors. The default constructor takes no +// arguments and creates a set of flags that are off by default. +// +// The optional parameter 'option_flags' is to facilitate transfer +// of legacy code from C programs. This lets you do +// RE(pattern, RE_Options(PCRE_CASELESS|PCRE_MULTILINE)).PartialMatch(str); +// +// But new code is better off doing +// RE(pattern, +// RE_Options().set_caseless(true).set_multiline(true)).PartialMatch(str); +// (See below) +// +// If you are going to pass one of the most used modifiers, there are some +// convenience functions that return a RE_Options class with the +// appropriate modifier already set: +// CASELESS(), UTF8(), MULTILINE(), DOTALL(), EXTENDED() +// +// If you need to set several options at once, and you don't want to go +// through the pains of declaring a RE_Options object and setting several +// options, there is a parallel method that give you such ability on the +// fly. You can concatenate several set_xxxxx member functions, since each +// of them returns a reference to its class object. e.g.: to pass +// PCRE_CASELESS, PCRE_EXTENDED, and PCRE_MULTILINE to a RE with one +// statement, you may write +// +// RE(" ^ xyz \\s+ .* blah$", RE_Options() +// .set_caseless(true) +// .set_extended(true) +// .set_multiline(true)).PartialMatch(sometext); +// +// ----------------------------------------------------------------------- +// SCANNING TEXT INCREMENTALLY +// +// The "Consume" operation may be useful if you want to repeatedly +// match regular expressions at the front of a string and skip over +// them as they match. This requires use of the "StringPiece" type, +// which represents a sub-range of a real string. Like RE, StringPiece +// is defined in the pcrecpp namespace. +// +// Example: read lines of the form "var = value" from a string. +// string contents = ...; // Fill string somehow +// pcrecpp::StringPiece input(contents); // Wrap in a StringPiece +// +// string var; +// int value; +// pcrecpp::RE re("(\\w+) = (\\d+)\n"); +// while (re.Consume(&input, &var, &value)) { +// ...; +// } +// +// Each successful call to "Consume" will set "var/value", and also +// advance "input" so it points past the matched text. +// +// The "FindAndConsume" operation is similar to "Consume" but does not +// anchor your match at the beginning of the string. For example, you +// could extract all words from a string by repeatedly calling +// pcrecpp::RE("(\\w+)").FindAndConsume(&input, &word) +// +// ----------------------------------------------------------------------- +// PARSING HEX/OCTAL/C-RADIX NUMBERS +// +// By default, if you pass a pointer to a numeric value, the +// corresponding text is interpreted as a base-10 number. You can +// instead wrap the pointer with a call to one of the operators Hex(), +// Octal(), or CRadix() to interpret the text in another base. The +// CRadix operator interprets C-style "0" (base-8) and "0x" (base-16) +// prefixes, but defaults to base-10. +// +// Example: +// int a, b, c, d; +// pcrecpp::RE re("(.*) (.*) (.*) (.*)"); +// re.FullMatch("100 40 0100 0x40", +// pcrecpp::Octal(&a), pcrecpp::Hex(&b), +// pcrecpp::CRadix(&c), pcrecpp::CRadix(&d)); +// will leave 64 in a, b, c, and d. +// +// ----------------------------------------------------------------------- +// REPLACING PARTS OF STRINGS +// +// You can replace the first match of "pattern" in "str" with +// "rewrite". Within "rewrite", backslash-escaped digits (\1 to \9) +// can be used to insert text matching corresponding parenthesized +// group from the pattern. \0 in "rewrite" refers to the entire +// matching text. E.g., +// +// string s = "yabba dabba doo"; +// pcrecpp::RE("b+").Replace("d", &s); +// +// will leave "s" containing "yada dabba doo". The result is true if +// the pattern matches and a replacement occurs, or false otherwise. +// +// GlobalReplace() is like Replace(), except that it replaces all +// occurrences of the pattern in the string with the rewrite. +// Replacements are not subject to re-matching. E.g., +// +// string s = "yabba dabba doo"; +// pcrecpp::RE("b+").GlobalReplace("d", &s); +// +// will leave "s" containing "yada dada doo". It returns the number +// of replacements made. +// +// Extract() is like Replace(), except that if the pattern matches, +// "rewrite" is copied into "out" (an additional argument) with +// substitutions. The non-matching portions of "text" are ignored. +// Returns true iff a match occurred and the extraction happened +// successfully. If no match occurs, the string is left unaffected. + + +#include +#include +#include // defines the Arg class +// This isn't technically needed here, but we include it +// anyway so folks who include pcrecpp.h don't have to. +#include + +namespace pcrecpp { + +#define PCRE_SET_OR_CLEAR(b, o) \ + if (b) all_options_ |= (o); else all_options_ &= ~(o); \ + return *this + +#define PCRE_IS_SET(o) \ + (all_options_ & o) == o + +/***** Compiling regular expressions: the RE class *****/ + +// RE_Options allow you to set options to be passed along to pcre, +// along with other options we put on top of pcre. +// Only 9 modifiers, plus match_limit and match_limit_recursion, +// are supported now. +class PCRECPP_EXP_DEFN RE_Options { + public: + // constructor + RE_Options() : match_limit_(0), match_limit_recursion_(0), all_options_(0) {} + + // alternative constructor. + // To facilitate transfer of legacy code from C programs + // + // This lets you do + // RE(pattern, RE_Options(PCRE_CASELESS|PCRE_MULTILINE)).PartialMatch(str); + // But new code is better off doing + // RE(pattern, + // RE_Options().set_caseless(true).set_multiline(true)).PartialMatch(str); + RE_Options(int option_flags) : match_limit_(0), match_limit_recursion_(0), + all_options_(option_flags) {} + // we're fine with the default destructor, copy constructor, etc. + + // accessors and mutators + int match_limit() const { return match_limit_; }; + RE_Options &set_match_limit(int limit) { + match_limit_ = limit; + return *this; + } + + int match_limit_recursion() const { return match_limit_recursion_; }; + RE_Options &set_match_limit_recursion(int limit) { + match_limit_recursion_ = limit; + return *this; + } + + bool caseless() const { + return PCRE_IS_SET(PCRE_CASELESS); + } + RE_Options &set_caseless(bool x) { + PCRE_SET_OR_CLEAR(x, PCRE_CASELESS); + } + + bool multiline() const { + return PCRE_IS_SET(PCRE_MULTILINE); + } + RE_Options &set_multiline(bool x) { + PCRE_SET_OR_CLEAR(x, PCRE_MULTILINE); + } + + bool dotall() const { + return PCRE_IS_SET(PCRE_DOTALL); + } + RE_Options &set_dotall(bool x) { + PCRE_SET_OR_CLEAR(x, PCRE_DOTALL); + } + + bool extended() const { + return PCRE_IS_SET(PCRE_EXTENDED); + } + RE_Options &set_extended(bool x) { + PCRE_SET_OR_CLEAR(x, PCRE_EXTENDED); + } + + bool dollar_endonly() const { + return PCRE_IS_SET(PCRE_DOLLAR_ENDONLY); + } + RE_Options &set_dollar_endonly(bool x) { + PCRE_SET_OR_CLEAR(x, PCRE_DOLLAR_ENDONLY); + } + + bool extra() const { + return PCRE_IS_SET(PCRE_EXTRA); + } + RE_Options &set_extra(bool x) { + PCRE_SET_OR_CLEAR(x, PCRE_EXTRA); + } + + bool ungreedy() const { + return PCRE_IS_SET(PCRE_UNGREEDY); + } + RE_Options &set_ungreedy(bool x) { + PCRE_SET_OR_CLEAR(x, PCRE_UNGREEDY); + } + + bool utf8() const { + return PCRE_IS_SET(PCRE_UTF8); + } + RE_Options &set_utf8(bool x) { + PCRE_SET_OR_CLEAR(x, PCRE_UTF8); + } + + bool no_auto_capture() const { + return PCRE_IS_SET(PCRE_NO_AUTO_CAPTURE); + } + RE_Options &set_no_auto_capture(bool x) { + PCRE_SET_OR_CLEAR(x, PCRE_NO_AUTO_CAPTURE); + } + + RE_Options &set_all_options(int opt) { + all_options_ = opt; + return *this; + } + int all_options() const { + return all_options_ ; + } + + // TODO: add other pcre flags + + private: + int match_limit_; + int match_limit_recursion_; + int all_options_; +}; + +// These functions return some common RE_Options +static inline RE_Options UTF8() { + return RE_Options().set_utf8(true); +} + +static inline RE_Options CASELESS() { + return RE_Options().set_caseless(true); +} +static inline RE_Options MULTILINE() { + return RE_Options().set_multiline(true); +} + +static inline RE_Options DOTALL() { + return RE_Options().set_dotall(true); +} + +static inline RE_Options EXTENDED() { + return RE_Options().set_extended(true); +} + +// Interface for regular expression matching. Also corresponds to a +// pre-compiled regular expression. An "RE" object is safe for +// concurrent use by multiple threads. +class PCRECPP_EXP_DEFN RE { + public: + // We provide implicit conversions from strings so that users can + // pass in a string or a "const char*" wherever an "RE" is expected. + RE(const string& pat) { Init(pat, NULL); } + RE(const string& pat, const RE_Options& option) { Init(pat, &option); } + RE(const char* pat) { Init(pat, NULL); } + RE(const char* pat, const RE_Options& option) { Init(pat, &option); } + RE(const unsigned char* pat) { + Init(reinterpret_cast(pat), NULL); + } + RE(const unsigned char* pat, const RE_Options& option) { + Init(reinterpret_cast(pat), &option); + } + + // Copy constructor & assignment - note that these are expensive + // because they recompile the expression. + RE(const RE& re) { Init(re.pattern_, &re.options_); } + const RE& operator=(const RE& re) { + if (this != &re) { + Cleanup(); + + // This is the code that originally came from Google + // Init(re.pattern_.c_str(), &re.options_); + + // This is the replacement from Ari Pollak + Init(re.pattern_, &re.options_); + } + return *this; + } + + + ~RE(); + + // The string specification for this RE. E.g. + // RE re("ab*c?d+"); + // re.pattern(); // "ab*c?d+" + const string& pattern() const { return pattern_; } + + // If RE could not be created properly, returns an error string. + // Else returns the empty string. + const string& error() const { return *error_; } + + /***** The useful part: the matching interface *****/ + + // This is provided so one can do pattern.ReplaceAll() just as + // easily as ReplaceAll(pattern-text, ....) + + bool FullMatch(const StringPiece& text, + const Arg& ptr1 = no_arg, + const Arg& ptr2 = no_arg, + const Arg& ptr3 = no_arg, + const Arg& ptr4 = no_arg, + const Arg& ptr5 = no_arg, + const Arg& ptr6 = no_arg, + const Arg& ptr7 = no_arg, + const Arg& ptr8 = no_arg, + const Arg& ptr9 = no_arg, + const Arg& ptr10 = no_arg, + const Arg& ptr11 = no_arg, + const Arg& ptr12 = no_arg, + const Arg& ptr13 = no_arg, + const Arg& ptr14 = no_arg, + const Arg& ptr15 = no_arg, + const Arg& ptr16 = no_arg) const; + + bool PartialMatch(const StringPiece& text, + const Arg& ptr1 = no_arg, + const Arg& ptr2 = no_arg, + const Arg& ptr3 = no_arg, + const Arg& ptr4 = no_arg, + const Arg& ptr5 = no_arg, + const Arg& ptr6 = no_arg, + const Arg& ptr7 = no_arg, + const Arg& ptr8 = no_arg, + const Arg& ptr9 = no_arg, + const Arg& ptr10 = no_arg, + const Arg& ptr11 = no_arg, + const Arg& ptr12 = no_arg, + const Arg& ptr13 = no_arg, + const Arg& ptr14 = no_arg, + const Arg& ptr15 = no_arg, + const Arg& ptr16 = no_arg) const; + + bool Consume(StringPiece* input, + const Arg& ptr1 = no_arg, + const Arg& ptr2 = no_arg, + const Arg& ptr3 = no_arg, + const Arg& ptr4 = no_arg, + const Arg& ptr5 = no_arg, + const Arg& ptr6 = no_arg, + const Arg& ptr7 = no_arg, + const Arg& ptr8 = no_arg, + const Arg& ptr9 = no_arg, + const Arg& ptr10 = no_arg, + const Arg& ptr11 = no_arg, + const Arg& ptr12 = no_arg, + const Arg& ptr13 = no_arg, + const Arg& ptr14 = no_arg, + const Arg& ptr15 = no_arg, + const Arg& ptr16 = no_arg) const; + + bool FindAndConsume(StringPiece* input, + const Arg& ptr1 = no_arg, + const Arg& ptr2 = no_arg, + const Arg& ptr3 = no_arg, + const Arg& ptr4 = no_arg, + const Arg& ptr5 = no_arg, + const Arg& ptr6 = no_arg, + const Arg& ptr7 = no_arg, + const Arg& ptr8 = no_arg, + const Arg& ptr9 = no_arg, + const Arg& ptr10 = no_arg, + const Arg& ptr11 = no_arg, + const Arg& ptr12 = no_arg, + const Arg& ptr13 = no_arg, + const Arg& ptr14 = no_arg, + const Arg& ptr15 = no_arg, + const Arg& ptr16 = no_arg) const; + + bool Replace(const StringPiece& rewrite, + string *str) const; + + int GlobalReplace(const StringPiece& rewrite, + string *str) const; + + bool Extract(const StringPiece &rewrite, + const StringPiece &text, + string *out) const; + + // Escapes all potentially meaningful regexp characters in + // 'unquoted'. The returned string, used as a regular expression, + // will exactly match the original string. For example, + // 1.5-2.0? + // may become: + // 1\.5\-2\.0\? + // Note QuoteMeta behaves the same as perl's QuoteMeta function, + // *except* that it escapes the NUL character (\0) as backslash + 0, + // rather than backslash + NUL. + static string QuoteMeta(const StringPiece& unquoted); + + + /***** Generic matching interface *****/ + + // Type of match (TODO: Should be restructured as part of RE_Options) + enum Anchor { + UNANCHORED, // No anchoring + ANCHOR_START, // Anchor at start only + ANCHOR_BOTH // Anchor at start and end + }; + + // General matching routine. Stores the length of the match in + // "*consumed" if successful. + bool DoMatch(const StringPiece& text, + Anchor anchor, + int* consumed, + const Arg* const* args, int n) const; + + // Return the number of capturing subpatterns, or -1 if the + // regexp wasn't valid on construction. + int NumberOfCapturingGroups() const; + + // The default value for an argument, to indicate no arg was passed in + static Arg no_arg; + + private: + + void Init(const string& pattern, const RE_Options* options); + void Cleanup(); + + // Match against "text", filling in "vec" (up to "vecsize" * 2/3) with + // pairs of integers for the beginning and end positions of matched + // text. The first pair corresponds to the entire matched text; + // subsequent pairs correspond, in order, to parentheses-captured + // matches. Returns the number of pairs (one more than the number of + // the last subpattern with a match) if matching was successful + // and zero if the match failed. + // I.e. for RE("(foo)|(bar)|(baz)") it will return 2, 3, and 4 when matching + // against "foo", "bar", and "baz" respectively. + // When matching RE("(foo)|hello") against "hello", it will return 1. + // But the values for all subpattern are filled in into "vec". + int TryMatch(const StringPiece& text, + int startpos, + Anchor anchor, + int *vec, + int vecsize) const; + + // Append the "rewrite" string, with backslash subsitutions from "text" + // and "vec", to string "out". + bool Rewrite(string *out, + const StringPiece& rewrite, + const StringPiece& text, + int *vec, + int veclen) const; + + // internal implementation for DoMatch + bool DoMatchImpl(const StringPiece& text, + Anchor anchor, + int* consumed, + const Arg* const args[], + int n, + int* vec, + int vecsize) const; + + // Compile the regexp for the specified anchoring mode + pcre* Compile(Anchor anchor); + + string pattern_; + RE_Options options_; + pcre* re_full_; // For full matches + pcre* re_partial_; // For partial matches + const string* error_; // Error indicator (or points to empty string) +}; + +} // namespace pcrecpp + +#endif /* _PCRECPP_H */ diff --git a/pcre/pcrecpp_internal.h b/pcre/pcrecpp_internal.h new file mode 100644 --- /dev/null +++ b/pcre/pcrecpp_internal.h @@ -0,0 +1,68 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +/* +----------------------------------------------------------------------------- +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. +----------------------------------------------------------------------------- +*/ + + +#ifndef PCRECPP_INTERNAL_H +#define PCRECPP_INTERNAL_H + +/* 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. We use: + + PCRECPP_EXP_DECL for declarations + PCRECPP_EXP_DEFN for definitions of exported functions + +*/ + +#ifndef PCRECPP_EXP_DECL +# ifdef _WIN32 +# ifndef PCRE_STATIC +# define PCRECPP_EXP_DECL extern __declspec(dllexport) +# define PCRECPP_EXP_DEFN __declspec(dllexport) +# else +# define PCRECPP_EXP_DECL extern +# define PCRECPP_EXP_DEFN +# endif +# else +# define PCRECPP_EXP_DECL extern +# define PCRECPP_EXP_DEFN +# endif +#endif + +#endif /* PCRECPP_INTERNAL_H */ + +/* End of pcrecpp_internal.h */ diff --git a/pcre/pcrecpparg.h b/pcre/pcrecpparg.h new file mode 100644 --- /dev/null +++ b/pcre/pcrecpparg.h @@ -0,0 +1,174 @@ +// 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 new file mode 100644 --- /dev/null +++ b/pcre/pcredemo.c @@ -0,0 +1,334 @@ +/************************************************* +* 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 new file mode 100644 --- /dev/null +++ b/pcre/pcregrep.c @@ -0,0 +1,2379 @@ +/************************************************* +* 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 new file mode 100644 --- /dev/null +++ b/pcre/pcreposix.c @@ -0,0 +1,356 @@ +/************************************************* +* 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 new file mode 100644 --- /dev/null +++ b/pcre/pcreposix.h @@ -0,0 +1,143 @@ +/************************************************* +* Perl-Compatible Regular Expressions * +*************************************************/ + +#ifndef _PCREPOSIX_H +#define _PCREPOSIX_H + +/* This is the header for the POSIX wrapper interface to the PCRE Perl- +Compatible Regular Expression library. It defines the things POSIX says should +be there. I hope. + + 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. +----------------------------------------------------------------------------- +*/ + +/* Have to include stdlib.h in order to ensure that size_t is defined. */ + +#include + +/* Allow for C++ users */ + +#ifdef __cplusplus +extern "C" { +#endif + +/* Options, mostly defined by POSIX, but with a couple of 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_STARTEND 0x0080 /* BSD feature: pass subject string by so,eo */ + +/* This is not used by PCRE, but by defining it we make it easier +to slot PCRE into existing programs that make POSIX calls. */ + +#define REG_EXTENDED 0 + +/* Error values. Not all these are relevant or used by the wrapper. */ + +enum { + REG_ASSERT = 1, /* internal error ? */ + REG_BADBR, /* invalid repeat counts in {} */ + REG_BADPAT, /* pattern error */ + REG_BADRPT, /* ? * + invalid */ + REG_EBRACE, /* unbalanced {} */ + REG_EBRACK, /* unbalanced [] */ + REG_ECOLLATE, /* collation error - not relevant */ + REG_ECTYPE, /* bad class */ + REG_EESCAPE, /* bad escape sequence */ + REG_EMPTY, /* empty expression */ + REG_EPAREN, /* unbalanced () */ + REG_ERANGE, /* bad range inside [] */ + REG_ESIZE, /* expression too big */ + REG_ESPACE, /* failed to get memory */ + REG_ESUBREG, /* bad back reference */ + REG_INVARG, /* bad argument */ + REG_NOMATCH /* match failed */ +}; + + +/* The structure representing a compiled regular expression. */ + +typedef struct { + void *re_pcre; + size_t re_nsub; + size_t re_erroffset; +} regex_t; + +/* The structure in which a captured offset is returned. */ + +typedef int regoff_t; + +typedef struct { + regoff_t rm_so; + regoff_t rm_eo; +} regmatch_t; + +/* 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 +export settings are needed, and are set in pcreposix.c before including this +file. */ + +#if defined(_WIN32) && !defined(PCRE_STATIC) && !defined(PCREPOSIX_EXP_DECL) +# define PCREPOSIX_EXP_DECL extern __declspec(dllimport) +# define PCREPOSIX_EXP_DEFN __declspec(dllimport) +#endif + +/* By default, we use the standard "extern" declarations. */ + +#ifndef PCREPOSIX_EXP_DECL +# ifdef __cplusplus +# define PCREPOSIX_EXP_DECL extern "C" +# define PCREPOSIX_EXP_DEFN extern "C" +# else +# define PCREPOSIX_EXP_DECL extern +# define PCREPOSIX_EXP_DEFN extern +# endif +#endif + +/* The functions */ + +PCREPOSIX_EXP_DECL int regcomp(regex_t *, const char *, int); +PCREPOSIX_EXP_DECL int regexec(const regex_t *, const char *, size_t, + regmatch_t *, int); +PCREPOSIX_EXP_DECL size_t regerror(int, const regex_t *, char *, size_t); +PCREPOSIX_EXP_DECL void regfree(regex_t *); + +#ifdef __cplusplus +} /* extern "C" */ +#endif + +#endif /* End of pcreposix.h */ diff --git a/pcre/pcretest.c b/pcre/pcretest.c new file mode 100644 --- /dev/null +++ b/pcre/pcretest.c @@ -0,0 +1,2448 @@ +/************************************************* +* 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 new file mode 100644 --- /dev/null +++ b/pcre/ucp.h @@ -0,0 +1,133 @@ +/************************************************* +* Unicode Property Table handler * +*************************************************/ + +#ifndef _UCP_H +#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. */ + +/* These are the general character categories. */ + +enum { + ucp_C, /* Other */ + ucp_L, /* Letter */ + ucp_M, /* Mark */ + ucp_N, /* Number */ + ucp_P, /* Punctuation */ + ucp_S, /* Symbol */ + ucp_Z /* Separator */ +}; + +/* These are the particular character types. */ + +enum { + ucp_Cc, /* Control */ + ucp_Cf, /* Format */ + ucp_Cn, /* Unassigned */ + ucp_Co, /* Private use */ + ucp_Cs, /* Surrogate */ + ucp_Ll, /* Lower case letter */ + ucp_Lm, /* Modifier letter */ + ucp_Lo, /* Other letter */ + ucp_Lt, /* Title case letter */ + ucp_Lu, /* Upper case letter */ + ucp_Mc, /* Spacing mark */ + ucp_Me, /* Enclosing mark */ + ucp_Mn, /* Non-spacing mark */ + ucp_Nd, /* Decimal number */ + ucp_Nl, /* Letter number */ + ucp_No, /* Other number */ + ucp_Pc, /* Connector punctuation */ + ucp_Pd, /* Dash punctuation */ + ucp_Pe, /* Close punctuation */ + ucp_Pf, /* Final punctuation */ + ucp_Pi, /* Initial punctuation */ + ucp_Po, /* Other punctuation */ + ucp_Ps, /* Open punctuation */ + ucp_Sc, /* Currency symbol */ + ucp_Sk, /* Modifier symbol */ + ucp_Sm, /* Mathematical symbol */ + ucp_So, /* Other symbol */ + ucp_Zl, /* Line separator */ + ucp_Zp, /* Paragraph separator */ + ucp_Zs /* Space separator */ +}; + +/* These are the script identifications. */ + +enum { + ucp_Arabic, + ucp_Armenian, + ucp_Bengali, + ucp_Bopomofo, + ucp_Braille, + ucp_Buginese, + ucp_Buhid, + ucp_Canadian_Aboriginal, + ucp_Cherokee, + ucp_Common, + ucp_Coptic, + ucp_Cypriot, + ucp_Cyrillic, + ucp_Deseret, + ucp_Devanagari, + ucp_Ethiopic, + ucp_Georgian, + ucp_Glagolitic, + ucp_Gothic, + ucp_Greek, + ucp_Gujarati, + ucp_Gurmukhi, + ucp_Han, + ucp_Hangul, + ucp_Hanunoo, + ucp_Hebrew, + ucp_Hiragana, + ucp_Inherited, + ucp_Kannada, + ucp_Katakana, + ucp_Kharoshthi, + ucp_Khmer, + ucp_Lao, + ucp_Latin, + ucp_Limbu, + ucp_Linear_B, + ucp_Malayalam, + ucp_Mongolian, + ucp_Myanmar, + ucp_New_Tai_Lue, + ucp_Ogham, + ucp_Old_Italic, + ucp_Old_Persian, + ucp_Oriya, + ucp_Osmanya, + ucp_Runic, + ucp_Shavian, + ucp_Sinhala, + ucp_Syloti_Nagri, + ucp_Syriac, + ucp_Tagalog, + ucp_Tagbanwa, + ucp_Tai_Le, + ucp_Tamil, + ucp_Telugu, + ucp_Thaana, + ucp_Thai, + ucp_Tibetan, + 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 */ +}; + +#endif + +/* End of ucp.h */ diff --git a/pcre/ucpinternal.h b/pcre/ucpinternal.h new file mode 100644 --- /dev/null +++ b/pcre/ucpinternal.h @@ -0,0 +1,92 @@ +/************************************************* +* 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 new file mode 100644 --- /dev/null +++ b/pcre/ucptable.h @@ -0,0 +1,3088 @@ +/* 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 new file mode 100644 --- /dev/null +++ b/regex-pcre-builtin.cabal @@ -0,0 +1,83 @@ +Name: regex-pcre-builtin +-- Keep the Version below in sync with ./Text/Regex/PCRE.hs value getVersion_Text_Regex_PCRE :: Version +Version: 0.94.1.0.7.7 +Cabal-Version: >=1.2.3 +License: BSD3 +License-File: LICENSE +Copyright: Copyright (c) 2006, Christopher Kuklewicz +Author: Christopher Kuklewicz +Maintainer: TextRegexLazy@personal.mightyreason.com +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/ +Synopsis: Replaces/Enhances Text.Regex +Description: The PCRE backend to accompany regex-base, see www.pcre.org +Category: Text, Pugs +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 + +flag splitBase + description: Choose the new smaller, split-up base package. +library + 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, PatternSignatures, GeneralizedNewtypeDeriving, FlexibleContexts, TypeSynonymInstances, FlexibleInstances + else + Build-Depends: base < 3.0, regex-base >= 0.93 + Extensions: MultiParamTypeClasses, FunctionalDependencies, CPP + -- Data-Files: + -- Extra-Source-Files: + -- Extra-Tmp-Files: + Exposed-Modules: Text.Regex.PCRE + Text.Regex.PCRE.Wrap + Text.Regex.PCRE.String + Text.Regex.PCRE.Sequence + Text.Regex.PCRE.ByteString + Text.Regex.PCRE.ByteString.Lazy + Buildable: True + -- Other-Modules: + -- HS-Source-Dirs: "." + -- 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 -ddump-minimal-imports + -- GHC-Prof-Options: + -- Hugs-Options: + -- NHC-Options: + -- C-Sources: + -- Includes: + -- LD-Options: + -- Frameworks: + -- The only reason to NOT define -DHAVE_PCRE_H is if you are on + -- 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 + --Include-Dirs: /opt/local/include + --Extra-Lib-Dirs: /opt/local/lib + Include-dirs: pcre + Includes: pcre + 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 +