regex-tdfa 1.2.1 → 1.2.2

raw patch · 11 files changed

+232/−105 lines, 11 filesdep ~basenew-uploaderPVP: major bump suggested

API removals or changes: PVP suggests a major version bump

Dependency ranges changed: base

API changes (from Hackage documentation)

Files

• CHANGELOG.md +157/−0
• Text/Regex/TDFA.hs +8/−13
• Text/Regex/TDFA/Common.hs +21/−10
• Text/Regex/TDFA/CorePattern.hs +3/−3
• Text/Regex/TDFA/NewDFA/Engine.hs +6/−6
• Text/Regex/TDFA/NewDFA/Engine_FA.hs +6/−6
• Text/Regex/TDFA/Pattern.hs +1/−1
• Text/Regex/TDFA/ReadRegex.hs +2/−2
• Text/Regex/TDFA/TDFA.hs +0/−1
• Text/Regex/TDFA/TNFA.hs +2/−2
• regex-tdfa.cabal +26/−61

@@ -0,0 +1,157 @@+# 1.2.2++* New maintainer.+* Now we don't reexport the problematic `Show` instance for functions.++# 1.2.1++* Updated dependency versions.++# 1.2.0++"Almost ghc-7.8" with the array 0.4 changes for `Data.Array.Unsafe`++# 1.1.8++Make ghc-7.0.2 on platorm 2011.2.0.0.0 happy++# 1.1.7++fix url below++# 1.1.6++Fix bug preventing `[]] [-] [^]] [^-]` (thanks to Maxime Henrion)++# 1.1.5++try `needUniqTags` in `POr` in CorePattern.hs, try `toAdvice b` for `PStar child`++# 1.1.4++fixed ++# 1.1.3++BROKEN after 100 characters the `compressOrbit` dies!++# 1.1.2++worked++# 1.1.1++add gnu escapes++# 1.1.0++NewDFA code working++# 1.0.7++make NewDFA directory and String_NC++# 1.0.6++try NewDFATest_SBS with `uncons`++# 1.0.5++use `uncons` on SBS++# 1.0.4++try repaired NewDFATest_SBS++* np13: try to improve readability with the `mm` combinator? Yes!+* np12: expand `o` in the case where `t` lookup get `Nothing`? Yes – this is the fix!?+* np11: break multi to not look at `o` and just return `True`? Yes !!!!+* np10: Peel off `CharMap`/`IntMap` and DFA/DT with pattern matching? No+* np9:  `INLINE` `endOf`? No+* np8:  np6 and `NOINLINE` `endOff`? No+* np7:  just return `True`? Fast+* np6:  comment out ans check? No+* np5:  comment out all `Multi0` code? No+* np4:  comment out all `Single0` and `Single` code? No+* np3:  `!off` the multi? No+* np2:  comment out all Testing code? No++# 1.0.3++try to alter `matchTest` to not have the `Bool` args? No++# 1.0.2++arg, the prof is fast and the normal slow!++# 1.0.1++add NewDFATest.hs++# 0.99.20++go to many vs single?++# 0.99.19++try for pre-comparison of orbit-logs!++# 0.99.18++try alternate lazy/strict strategy in NewDFA. Fix offset laziness.++# 0.99.17++radical removal of flag array and adding of `SetVal` to handle groups++# 0.99.16++performance? up to v15++# 0.99.15++get string with NewDFA testing, unit tests and 1000 random regex pass++# 0.99.14++start changing to the new real DFA++# 0.99.13++more cleanup++# 0.99.12++try to debug 0.99.11: fixed `updateWinner`++# 0.99.11++improve above fix and make stuff work better – HAS BUG, along with old TDFA!++# 0.99.10++fixed `((.?)*)*` patterns by changing `PStar nullView` when `mayFirstBeNull`++# 0.99.9++testing changing `bestTrans`/`chooseWith`/`choose` to include `enterOrbit`/`newFlags`/`(_,True)` info++# 0.99.8++testing changing `Maximize` to `Minimize` for `Tag`s, decide `(a*)*` is canonical problem++# 0.99.7++Use `(PGroup Nothing)` in `Pattern` to decompose `PBound`++# 0.99.6++change to nested `nonEmpty` calls for `PBound`++# 0.99.5++remove `PNonEmpty` constructor++# 0.99.4++tests `pnonempty' = \ p -> POr [ PEmpty, p ]` instead of `PNonEmpty`

@@ -9,7 +9,7 @@ This regex-tdfa package implements, correctly, POSIX extended regular expressions.  It is highly unlikely that the regex-posix package on your operating system is correct, see-http://www.haskell.org/haskellwiki/Regex_Posix for examples of your+<http://www.haskell.org/haskellwiki/Regex_Posix> for examples of your OS's bugs.  This package does provide captured parenthesized subexpressions.@@ -23,23 +23,18 @@ As of version 1.1.1 the following GNU extensions are recognized, all anchors: -\\\` at beginning of entire text--\\\' at end of entire text--\\< at beginning of word--\\> at end of word--\\b at either beginning or end of word--\\B at neither beginning nor end of word+* \\\` at beginning of entire text+* \\\' at end of entire text+* \\\< at beginning of word+* \\\> at end of word+* \\b at either beginning or end of word+* \\B at neither beginning nor end of word  The above are controlled by the 'newSyntax' Bool in 'CompOption'.  Where the "word" boundaries means between characters that are and are not in the [:word:] character class which contains [a-zA-Z0-9_].  Note-that \< and \b may match before the entire text and \> and \b may+that \\\< and \\b may match before the entire text and \\\> and \\b may match at the end of the entire text.  There is no locale support, so collating elements like [.ch.] are

@@ -7,7 +7,6 @@ import Text.Regex.Base(RegexOptions(..))  {- By Chris Kuklewicz, 2007-2009. BSD License, see the LICENSE file. -}-import Text.Show.Functions() import Data.Array.IArray(Array) import Data.IntSet.EnumSet2(EnumSet) import qualified Data.IntSet.EnumSet2 as Set(toList)@@ -32,14 +31,14 @@ on :: (t1 -> t1 -> t2) -> (t -> t1) -> t -> t -> t2 f `on` g = (\x y -> (g x) `f` (g y)) --- | after sort or sortBy the use of nub/nubBy can be replaced by norep/norepBy+-- | after 'sort' or 'sortBy' the use of 'nub'\/'nubBy' can be replaced by 'norep'\/'norepBy' norep :: (Eq a) => [a]->[a] norep [] = [] norep x@[_] = x norep (a:bs@(c:cs)) | a==c = norep (a:cs)                     | otherwise = a:norep bs --- | after sort or sortBy the use of nub/nubBy can be replaced by norep/norepBy+-- | after 'sort' or 'sortBy' the use of 'nub'\/'nubBy' can be replaced by 'norep'\/'norepBy' norepBy :: (a -> a -> Bool) -> [a] -> [a] norepBy _ [] = [] norepBy _ x@[_] = x@@ -81,7 +80,7 @@   showsPrec p (DoPa {dopaIndex=i}) = ('#':) . showsPrec p i  -- | Control whether the pattern is multiline or case-sensitive like Text.Regex and whether to--- capture the subgroups (\1, \2, etc).  Controls enabling extra anchor syntax.+-- capture the subgroups (\\1, \\2, etc).  Controls enabling extra anchor syntax. data CompOption = CompOption {     caseSensitive :: Bool    -- ^ True in blankCompOpt and defaultCompOpt   , multiline :: Bool {- ^ False in blankCompOpt, True in defaultCompOpt. Compile for@@ -115,7 +114,7 @@ type Position = Int  -- | GroupIndex is for indexing submatches from capturing--- parenthesized groups (PGroup/Group)+-- parenthesized groups (PGroup\/Group) type GroupIndex = Int -- | GroupInfo collects the parent and tag information for an instance  -- of a group@@ -180,9 +179,9 @@ type QTrans = IntMap {- Destination Index -} [TagCommand]  -- | Known predicates, just Beginning of Line (^) and End of Line ($).--- Also support for GNU extensions is being added: \` beginning of--- buffer, \' end of buffer, \< and \> for begin and end of words, \b--- and \B for word boundary and not word boundary.+-- Also support for GNU extensions is being added: \\\` beginning of+-- buffer, \\\' end of buffer, \\\< and \\\> for begin and end of words, \\b+-- and \\B for word boundary and not word boundary. data WhichTest = Test_BOL | Test_EOL -- '^' and '$' (affected by multiline option)                | Test_BOB | Test_EOB -- \` and \' begin and end buffer                | Test_BOW | Test_EOW -- \< and \> begin and end word@@ -233,7 +232,7 @@ -- going to the same destination must be compared to find the best. -- -- A Destination IntMap entry may have an empty Source IntMap if and--- only if the destination is the starting index and the NFA/DFA.+-- only if the destination is the starting index and the NFA\/DFA. -- This instructs the matching engine to spawn a new entry starting at -- the post-update position. type DTrans = IntMap {- Index of Destination -} (IntMap {- Index of Source -} (DoPa,Instructions))@@ -259,7 +258,19 @@ data Instructions = Instructions   { newPos :: ![(Tag,Action)] -- False is preUpdate, True is postUpdate (there are no Orbit tags here) -- 2009 : Change to enum from bool?   , newOrbits :: !(Maybe (Position -> OrbitTransformer))-  } deriving (Show)+  }++instance Show Instructions where+  showsPrec p (Instructions pos _)+    = showParen (p >= 11) $+        showString "Instructions {" .+        showString "newPos = " .+        showsPrec 0 pos .+        showString ", " .+        showString "newOrbits = " .+        showString "<function>" .+        showString "}"+ data Action = SetPre | SetPost | SetVal Int deriving (Show,Eq) type OrbitTransformer = OrbitLog -> OrbitLog type OrbitLog = IntMap Orbits

@@ -1,16 +1,16 @@ -- | The CorePattern module deconstructs the Pattern tree created by--- ReadRegex.parseRegex and returns a simpler Q/P tree with+-- ReadRegex.parseRegex and returns a simpler Q\/P tree with -- annotations at each Q node.  This will be converted by the TNFA -- module into a QNFA finite automata. ----- Of particular note, this Pattern to Q/P conversion creates and+-- Of particular note, this Pattern to Q\/P conversion creates and -- assigns all the internal Tags that will be used during the matching -- process, and associates the captures groups with the tags that -- represent their starting and ending locations and with their -- immediate parent group. -- -- Each Maximize and Minimize tag is held as either a preTag or a--- postTag by one and only one location in the Q/P tree.  The Orbit+-- postTag by one and only one location in the Q\/P tree.  The Orbit -- tags are each held by one and only one Star node.  Tags that stop a -- Group are also held in perhaps numerous preReset lists. --

@@ -529,11 +529,11 @@  newtype F s = F ([F s] -> C s) type C s = Position-	  -> ((Int, Instructions), STUArray s Tag Position, IntMap Orbits)-	  -> [(Int, Action)]-	  -> ((Int, Instructions), STUArray s Tag Position, IntMap Orbits)-	  -> [(Int, Action)]-	  -> ST s Ordering+        -> ((Int, Instructions), STUArray s Tag Position, IntMap Orbits)+        -> [(Int, Action)]+        -> ((Int, Instructions), STUArray s Tag Position, IntMap Orbits)+        -> [(Int, Action)]+        -> ST s Ordering  {-# INLINE orderOf #-} orderOf :: Action -> Action -> Ordering@@ -703,7 +703,7 @@ Prelude Data.Array.Base> :i STUArray data STUArray s i e   = STUArray !i !i !Int (GHC.Prim.MutableByteArray# s)-  	-- Defined in Data.Array.Base+  -- Defined in Data.Array.Base -} -- This has been updated for ghc 6.8.3 and still works with ghc 6.10.1 {-# INLINE copySTU #-}

@@ -388,11 +388,11 @@  newtype F s = F ([F s] -> C s) type C s = Position-	  -> ((Int, Instructions), STUArray s Tag Position, IntMap Orbits)-	  -> [(Int, Action)]-	  -> ((Int, Instructions), STUArray s Tag Position, IntMap Orbits)-	  -> [(Int, Action)]-	  -> ST s Ordering+        -> ((Int, Instructions), STUArray s Tag Position, IntMap Orbits)+        -> [(Int, Action)]+        -> ((Int, Instructions), STUArray s Tag Position, IntMap Orbits)+        -> [(Int, Action)]+        -> ST s Ordering  {-# INLINE orderOf #-} orderOf :: Action -> Action -> Ordering@@ -561,7 +561,7 @@ Prelude Data.Array.Base> :i STUArray data STUArray s i e   = STUArray !i !i !Int (GHC.Prim.MutableByteArray# s)-  	-- Defined in Data.Array.Base+  -- Defined in Data.Array.Base -} -- This has been updated for ghc 6.8.3 and still works with ghc 6.10.1 {-# INLINE copySTU #-}

@@ -132,7 +132,7 @@  -- | Do the transformation and simplification in a single traversal. -- This removes the PPlus, PQuest, and PBound values, changing to POr--- and PEmpty and PStar True/False.  For some PBound values it adds+-- and PEmpty and PStar True\/False.  For some PBound values it adds -- PNonEmpty and PNonCapture semantic marker.  It also simplifies to -- flatten out nested POr and PConcat instances and eliminate some -- unneeded PEmpty values.

@@ -1,6 +1,6 @@ {-# OPTIONS_GHC -fno-warn-missing-signatures #-} -- | This is a POSIX version of parseRegex that allows NUL characters.--- Lazy/Possessive/Backrefs are not recognized.  Anchors ^ and $ are+-- Lazy\/Possessive\/Backrefs are not recognized.  Anchors \^ and \$ are -- recognized. -- -- The PGroup returned always have (Maybe GroupIndex) set to (Just _)@@ -23,7 +23,7 @@ -- | Return either an error message or a tuple of the Pattern and the -- largest group index and the largest DoPa index (both have smallest -- index of 1).  Since the regular expression is supplied as [Char] it--- automatically supports unicode and '\NUL' characters.+-- automatically supports unicode and @\\NUL@ characters. parseRegex :: String -> Either ParseError (Pattern,(GroupIndex,DoPa)) parseRegex x = runParser (do pat <- p_regex                              eof

@@ -6,7 +6,6 @@                             ,examineDFA,nfaToDFA,dfaMap) where  --import Control.Arrow((***))-import Control.Monad.Instances() import Data.Monoid(Monoid(..)) import Control.Monad.State(State,MonadState(..),execState) import Data.Array.IArray(Array,(!),bounds,{-assocs-})

@@ -6,7 +6,7 @@ -- TODO : try rewriting whole qToNFA in terms of "act" -- (That will require re-organizing the continuation data a bit) --- | "Text.Regex.TDFA.TNFA" converts the CorePattern Q/P data (and its+-- | "Text.Regex.TDFA.TNFA" converts the CorePattern Q\/P data (and its -- Pattern leafs) to a QNFA tagged non-deterministic finite automata. --  -- This holds every possible way to follow one state by another, while@@ -17,7 +17,7 @@ -- far, keeping all this information has helped fix bugs in both the -- design and implementation. ----- The QNFA for a Pattern with a starTraned Q/P form with N one+-- The QNFA for a Pattern with a starTraned Q\/P form with N one -- character accepting leaves has at most N+1 nodes.  These nodes -- repesent the future choices after accepting a leaf.  The processing -- of Or nodes often reduces this number by sharing at the end of the

@@ -1,76 +1,40 @@ Name:                   regex-tdfa-Version:                1.2.1--- 0.99.4 tests pnonempty' = \ p -> POr [ PEmpty, p ] instead of PNonEmpty--- 0.99.5 remove PNonEmpty constructor--- 0.99.6 change to nested nonEmpty calls for PBound--- 0.99.7 Use (PGroup Nothing) in Pattern to decompose PBound--- 0.99.8 testing chaning Maximize to Minimize for Tags, decide (a*)* is canonical problem--- 0.99.9 testing changing bestTrans/chooseWith/choose to include enterOrbit/newFlags/(_,True) info--- 0.99.10 fixed ((.?)*)* patterns by changing PStar nullView when mayFirstBeNull--- 0.99.11 improve above fix and make stuff work better -- HAS BUG, along with old TDFA!--- 0.99.12 try to debug 0.99.11 : fixed updateWinner--- 0.99.13 more cleanup--- 0.99.14 start changing to the new real DFA--- 0.99.15 get string with NewDFA testing, unit tests and 1000 random regex pass--- 0.99.16 performance? up to v15--- 0.99.17 radical removal of flag array and adding of SetVal to handle groups--- 0.99.18 try alternate lazy/strict strategy in NewDFA.  Fix offset laziness.--- 0.99.19 try for pre-comparison of orbit-logs!--- 0.99.20 go to many vs single?--- 1.0.0--- 1.0.1 add NewDFATest.hs--- 1.0.2 arg, the prof is fast and the normal slow!--- 1.0.3 try to alter matchTest to not have the Bool args? No---  np2  comment out all Testing code? No---  np3  !off the multi? No---  np4  comment out all Single0 and Single code? No---  np5  comment out all Multi0 code? No---  np6  comment out ans check? No---  np7  just return True? Fast---  np8  np6 and NOINLINE endOff? No---  np9  INLINE endOf? No---  np10 Peel off CharMap/IntMap and DFA/DT with pattern matching? No---  np11 break multi to not look at o and just return True? Yes !!!!---  np12 expand o in the case where t lookup get Nothing? Yes--this is the fix!?---  np13 try to improve readability with the "mm" combinator? Yes!--- 1.0.4 try repaired NewDFATest_SBS--- 1.0.5 use "uncons" on SBS--- 1.0.6 try NewDFATest_SBS with uncons--- 1.0.7 make NewDFA directory and String_NC--- 1.1.0 NewDFA code working--- 1.1.1 add gnu escapes--- 1.1.2 worked--- 1.1.3 BROKEN after 100 characters the compressOrbit dies!--- 1.1.4 fixed --- 1.1.5 try needUniqTags in POr in CorePattern.hs, try (toAdvice b) for PStar child--- 1.1.6 Fix bug preventing []] [-] [^]] [^-] (thanks to Maxime Henrion)--- 1.1.7 fix url below--- 1.1.8 Make ghc-7.0.2 on platorm 2011.2.0.0.0 happy--- 1.2.0 "Almost ghc-7.8" with the array 0.4 changes for Data.Array.Unsafe--- 1.2.1 Get Build-Depends from "Mon Jun 15 14:24:28 UTC 2015 by AdamBergmark to revision 1"+Version:                1.2.2 License:                BSD3 License-File:           LICENSE Copyright:              Copyright (c) 2007, Christopher Kuklewicz Author:                 Christopher Kuklewicz-Maintainer:             TextRegexLazy@personal.mightyreason.com+Maintainer:             Artyom <yom@artyom.me>, Christopher Kuklewicz <TextRegexLazy@personal.mightyreason.com> Stability:              Seems to work, but not POSIX yet-Homepage:               http://hackage.haskell.org/package/regex-tdfa-Package-URL:            http://code.haskell.org/regex-tdfa/+Homepage:               https://github.com/ChrisKuklewicz/regex-tdfa+Bug-Reports:            https://github.com/ChrisKuklewicz/regex-tdfa/issues Synopsis:               Replaces/Enhances Text.Regex Description:            A new all Haskell "tagged" DFA regex engine, inspired by libtre Category:               Text-Tested-With:            GHC+Tested-With:            GHC == 7.6.3, GHC == 7.8.4, GHC == 7.10.3, GHC == 8.0.1 Build-Type:             Simple-Cabal-Version:          >= 1.2.3+extra-source-files:+  CHANGELOG.md+Cabal-Version:          >= 1.6 -flag base4+source-repository head+  type:                git+  location:            git://github.com/ChrisKuklewicz/regex-tdfa.git +flag devel+  description: enable flags that are useful for development+  default: False+  manual: True+ library -  Build-Depends:        regex-base >= 0.93.1, parsec, mtl, containers, array >= 0.4, bytestring-  if flag(base4)-    Build-Depends:      base >= 4 && < 5, base >= 4.4 || < 4.4, ghc-prim-  else-    Build-Depends:      base < 4.0+  Build-Depends:        array >= 0.4+                      , base >= 4 && < 5+                      , bytestring+                      , containers+                      , ghc-prim+                      , mtl+                      , parsec+                      , regex-base >= 0.93.1    other-modules:          Paths_regex_tdfa   Exposed-Modules:        Data.IntMap.CharMap2@@ -98,4 +62,5 @@   Buildable:              True   Extensions:             MultiParamTypeClasses, FunctionalDependencies, BangPatterns, MagicHash, RecursiveDo, NoMonoPatBinds, ForeignFunctionInterface, UnboxedTuples, TypeOperators, FlexibleContexts, ExistentialQuantification, UnliftedFFITypes, TypeSynonymInstances, FlexibleInstances   GHC-Options:            -Wall -funbox-strict-fields -fspec-constr-count=10 -O2 -fno-warn-orphans-  GHC-Prof-Options:       -auto-all+  if flag(devel)+    ghc-prof-options: -auto-all