/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Diff of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 327 by ph10, Sat Mar 8 19:38:30 2008 UTC revision 1348 by ph10, Fri Jul 5 10:38:37 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.7 05-Mar-08  Version 8.34 xx-xxxx-201x
5    --------------------------
6    
7    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8        executable memory. Patch inspired by Carsten Klein.
9    
10    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11        this macro is never tested and has no effect, because the work to support
12        coverage involves only compiling and linking options and special targets in
13        the Makefile. The comment in config.h implied that defining the macro would
14        enable coverage support, which is totally false. There was also support for
15        setting this macro in the CMake files (my fault, I just copied it from
16        configure). SUPPORT_GCOV has now been removed.
17    
18    3.  Make a small performance improvement in strlen16() and strlen32() in
19        pcretest.
20    
21    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23        these cases). There was also one in pcretest.c.
24    
25    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26    
27    6.  In UTF mode, the code for checking whether a group could match an empty
28        string (which is used for indefinitely repeated groups to allow for
29        breaking an infinite loop) was broken when the group contained a repeated
30        negated single-character class with a character that occupied more than one
31        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32        UTF-8 mode). The effect was undefined: the group might or might not be
33        deemed as matching an empty string, or the program might have crashed.
34    
35    7.  The code for checking whether a group could match an empty string was not
36        recognizing that \h, \H, \v, \V, and \R must match a character.
37    
38    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39        an empty string. If it can, pcretest shows this in its information output.
40    
41    
42    Version 8.33 28-May-2013
43    --------------------------
44    
45    1.  Added 'U' to some constants that are compared to unsigned integers, to
46        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
47        variables that are added to signed variables, to ensure the result is
48        signed and can be negated.
49    
50    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
51        CMake config files.
52    
53    3.  Revise the creation of config.h.generic so that all boolean macros are
54        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
55        overriding via -D on the command line possible.
56    
57    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
58        to unsigned int is reported to make a quite noticeable speed difference in
59        a specific Windows environment. Testing on Linux did also appear to show
60        some benefit (and it is clearly not harmful). Also fixed the definition of
61        Xop which should be unsigned.
62    
63    5.  Related to (4), changing the definition of the intermediate variable cc
64        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
65        improvements.
66    
67    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
68        unnecessary spaces.
69    
70    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
71        and later.
72    
73    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
74    
75    9.  Optimizing fast_forward_start_bits in JIT.
76    
77    10. Adding support for callouts in JIT, and fixing some issues revealed
78        during this work. Namely:
79    
80        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
81    
82        (b) Minimum length was not checked before the matching is started.
83    
84    11. The value of capture_last that is passed to callouts was incorrect in some
85        cases when there was a capture on one path that was subsequently abandoned
86        after a backtrack. Also, the capture_last value is now reset after a
87        recursion, since all captures are also reset in this case.
88    
89    12. The interpreter no longer returns the "too many substrings" error in the
90        case when an overflowing capture is in a branch that is subsequently
91        abandoned after a backtrack.
92    
93    13. In the pathological case when an offset vector of size 2 is used, pcretest
94        now prints out the matched string after a yield of 0 or 1.
95    
96    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
97        Only supported by the JIT compiler at the moment.
98    
99    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
100    
101    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
102        offset of the starting point of the matching process, provided the offsets
103        vector is large enough.
104    
105    17. The \A escape now records a lookbehind value of 1, though its execution
106        does not actually inspect the previous character. This is to ensure that,
107        in partial multi-segment matching, at least one character from the old
108        segment is retained when a new segment is processed. Otherwise, if there
109        are no lookbehinds in the pattern, \A might match incorrectly at the start
110        of a new segment.
111    
112    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
113    
114    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
115        gives some modest performance improvement in 8-bit mode.
116    
117    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
118        be expressed in certain programming languages using Universal Character
119        Names.
120    
121    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
122        which points out that "non characters" are not "characters that may not
123        appear in Unicode strings" but rather "characters that are reserved for
124        internal use and have only local meaning".
125    
126    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
127        there was a conditional group that depended on an assertion, if the
128        assertion was false, the callout that immediately followed the alternation
129        in the condition was skipped when pcre_exec() was used for matching.
130    
131    23. Allow an explicit callout to be inserted before an assertion that is the
132        condition for a conditional group, for compatibility with automatic
133        callouts, which always insert a callout at this point.
134    
135    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
136        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
137    
138    25. (*PRUNE) is now supported by the JIT compiler.
139    
140    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
141    
142    27. Fix the case where there are two or more SKIPs with arguments that may be
143        ignored.
144    
145    28. (*SKIP) is now supported by the JIT compiler.
146    
147    29. (*THEN) is now supported by the JIT compiler.
148    
149    30. Update RunTest with additional test selector options.
150    
151    31. The way PCRE handles backtracking verbs has been changed in two ways.
152    
153        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
154        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
155        backtracking. In some cases this makes it more Perl-compatible, but Perl's
156        rather obscure rules do not always do the same thing.
157    
158        (2) Previously, backtracking verbs were confined within assertions. This is
159        no longer the case for positive assertions, except for (*ACCEPT). Again,
160        this sometimes improves Perl compatibility, and sometimes does not.
161    
162    32. A number of tests that were in test 2 because Perl did things differently
163        have been moved to test 1, because either Perl or PCRE has changed, and
164        these tests are now compatible.
165    
166    32. Backtracking control verbs are now handled in the same way in JIT and
167        interpreter.
168    
169    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
170        contained a forward subroutine reference caused a compile error.
171    
172    34. Auto-detect and optimize limited repetitions in JIT.
173    
174    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
175        blocking (*UTF) etc.
176    
177    36. In the interpreter, maximizing pattern repetitions for characters and
178        character types now use tail recursion, which reduces stack usage.
179    
180    37. The value of the max lookbehind was not correctly preserved if a compiled
181        and saved regex was reloaded on a host of different endianness.
182    
183    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
184        of the compiled pattern block, expand the flags field from 16 to 32 bits
185        because it was almost full.
186    
187    39. Try madvise first before posix_madvise.
188    
189    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
190        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
191        original change locked out all matches of empty strings. This has been
192        changed so that one match of an empty string per line is recognized.
193        Subsequent searches on the same line (for colouring or for --only-matching,
194        for example) do not recognize empty strings.
195    
196    41. Applied a user patch to fix a number of spelling mistakes in comments.
197    
198    42. Data lines longer than 65536 caused pcretest to crash.
199    
200    43. Clarified the data type for length and startoffset arguments for pcre_exec
201        and pcre_dfa_exec in the function-specific man pages, where they were
202        explicitly stated to be in bytes, never having been updated. I also added
203        some clarification to the pcreapi man page.
204    
205    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
206        a segmentation fault.
207    
208    
209    Version 8.32 30-November-2012
210    -----------------------------
211    
212    1.  Improved JIT compiler optimizations for first character search and single
213        character iterators.
214    
215    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
216        Patch by Daniel Richard G.
217    
218    3.  Single character iterator optimizations in the JIT compiler.
219    
220    4.  Improved JIT compiler optimizations for character ranges.
221    
222    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
223        Reported by Giuseppe D'Angelo.
224    
225    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
226        of a line, was being set incorrectly in cases where .* appeared inside
227        atomic brackets at the start of a pattern, or where there was a subsequent
228        *PRUNE or *SKIP.
229    
230    7.  Improved instruction cache flush for POWER/PowerPC.
231        Patch by Daniel Richard G.
232    
233    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
234        grep:
235    
236        (a) There is now no limit to the number of patterns to be matched.
237    
238        (b) An error is given if a pattern is too long.
239    
240        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
241            are now supported.
242    
243        (d) --exclude-from and --include-from (multiple use) have been added.
244    
245        (e) Exclusions and inclusions now apply to all files and directories, not
246            just to those obtained from scanning a directory recursively.
247    
248        (f) Multiple uses of -f and --file-list are now supported.
249    
250        (g) In a Windows environment, the default for -d has been changed from
251            "read" (the GNU grep default) to "skip", because otherwise the presence
252            of a directory in the file list provokes an error.
253    
254        (h) The documentation has been revised and clarified in places.
255    
256    9.  Improve the matching speed of capturing brackets.
257    
258    10. Changed the meaning of \X so that it now matches a Unicode extended
259        grapheme cluster.
260    
261    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
262        out POSIX threads when JIT support is configured.
263    
264    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
265    
266    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
267        structure could go wrong in environments where size_t is not the same size
268        as int.
269    
270    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
271        set.
272    
273    15. The EBCDIC support had decayed; later updates to the code had included
274        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
275        general tidy up of EBCDIC-related issues, and the documentation was also
276        not quite right. There is now a test that can be run on ASCII systems to
277        check some of the EBCDIC-related things (but is it not a full test).
278    
279    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
280        in a small tidy to the code.
281    
282    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
283    
284    18. If the --only-matching (-o) option in pcregrep is specified multiple
285        times, each one causes appropriate output. For example, -o1 -o2 outputs the
286        substrings matched by the 1st and 2nd capturing parentheses. A separating
287        string can be specified by --om-separator (default empty).
288    
289    19. Improving the first n character searches.
290    
291    20. Turn case lists for horizontal and vertical white space into macros so that
292        they are defined only once.
293    
294    21. This set of changes together give more compatible Unicode case-folding
295        behaviour for characters that have more than one other case when UCP
296        support is available.
297    
298        (a) The Unicode property table now has offsets into a new table of sets of
299            three or more characters that are case-equivalent. The MultiStage2.py
300            script that generates these tables (the pcre_ucd.c file) now scans
301            CaseFolding.txt instead of UnicodeData.txt for character case
302            information.
303    
304        (b) The code for adding characters or ranges of characters to a character
305            class has been abstracted into a generalized function that also handles
306            case-independence. In UTF-mode with UCP support, this uses the new data
307            to handle characters with more than one other case.
308    
309        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
310            whose other case is greater than 256 are now correctly matched
311            caselessly. Previously, the high codepoint matched the low one, but not
312            vice versa.
313    
314        (d) The processing of \h, \H, \v, and \ in character classes now makes use
315            of the new class addition function, using character lists defined as
316            macros alongside the case definitions of 20 above.
317    
318        (e) Caseless back references now work with characters that have more than
319            one other case.
320    
321        (f) General caseless matching of characters with more than one other case
322            is supported.
323    
324    22. Unicode character properties were updated from Unicode 6.2.0
325    
326    23. Improved CMake support under Windows. Patch by Daniel Richard G.
327    
328    24. Add support for 32-bit character strings, and UTF-32
329    
330    25. Major JIT compiler update (code refactoring and bugfixing).
331        Experimental Sparc 32 support is added.
332    
333    26. Applied a modified version of Daniel Richard G's patch to create
334        pcre.h.generic and config.h.generic by "make" instead of in the
335        PrepareRelease script.
336    
337    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
338        pcre_compile.c when checking for a zero character.
339    
340    28. Introducing a native interface for JIT. Through this interface, the compiled
341        machine code can be directly executed. The purpose of this interface is to
342        provide fast pattern matching, so several sanity checks are not performed.
343        However, feature tests are still performed. The new interface provides
344        1.4x speedup compared to the old one.
345    
346    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
347        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
348        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
349    
350    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
351        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
352        the "old" RFC 2279). Instead, it ended up passing a negative length to
353        pcre_exec().
354    
355    31. Add support for GCC's visibility feature to hide internal functions.
356    
357    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
358        "unknown -C option" after outputting 0 or 1.
359    
360    33. There is now support for generating a code coverage report for the test
361        suite in environments where gcc is the compiler and lcov is installed. This
362        is mainly for the benefit of the developers.
363    
364    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
365        unaddressable using valgrind annotations, allowing valgrind to detect
366        invalid memory accesses. This is mainly for the benefit of the developers.
367    
368    25. (*UTF) can now be used to start a pattern in any of the three libraries.
369    
370    26. Give configure error if --enable-cpp but no C++ compiler found.
371    
372    
373    Version 8.31 06-July-2012
374    -------------------------
375    
376    1.  Fixing a wrong JIT test case and some compiler warnings.
377    
378    2.  Removed a bashism from the RunTest script.
379    
380    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
381        to unsigned type, result still unsigned" that was given by an MS compiler
382        on encountering the code "-sizeof(xxx)".
383    
384    4.  Partial matching support is added to the JIT compiler.
385    
386    5.  Fixed several bugs concerned with partial matching of items that consist
387        of more than one character:
388    
389        (a) /^(..)\1/ did not partially match "aba" because checking references was
390            done on an "all or nothing" basis. This also applied to repeated
391            references.
392    
393        (b) \R did not give a hard partial match if \r was found at the end of the
394            subject.
395    
396        (c) \X did not give a hard partial match after matching one or more
397            characters at the end of the subject.
398    
399        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
400            a partial match for the string "\r".
401    
402        (e) When newline was set to CRLF, the metacharacter "." did not recognize
403            a partial match for a CR character at the end of the subject string.
404    
405    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
406        running pcretest, the text "(JIT)" added to the output whenever JIT is
407        actually used to run the match.
408    
409    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
410        or /S+[+] with a digit between 1 and 7.
411    
412    8.  OP_NOT now supports any UTF character not just single-byte ones.
413    
414    9.  (*MARK) control verb is now supported by the JIT compiler.
415    
416    10. The command "./RunTest list" lists the available tests without actually
417        running any of them. (Because I keep forgetting what they all are.)
418    
419    11. Add PCRE_INFO_MAXLOOKBEHIND.
420    
421    12. Applied a (slightly modified) user-supplied patch that improves performance
422        when the heap is used for recursion (compiled with --disable-stack-for-
423        recursion). Instead of malloc and free for each heap frame each time a
424        logical recursion happens, frames are retained on a chain and re-used where
425        possible. This sometimes gives as much as 30% improvement.
426    
427    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
428        call.
429    
430    14. As documented, (*COMMIT) is now confined to within a positive assertion.
431    
432    15. It is now possible to link pcretest with libedit as an alternative to
433        libreadline.
434    
435    16. (*COMMIT) control verb is now supported by the JIT compiler.
436    
437    17. The Unicode data tables have been updated to Unicode 6.1.0.
438    
439    18. Added --file-list option to pcregrep.
440    
441    19. Added binary file support to pcregrep, including the -a, --binary-files,
442        -I, and --text options.
443    
444    20. The madvise function is renamed for posix_madvise for QNX compatibility
445        reasons. Fixed by Giuseppe D'Angelo.
446    
447    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
448    
449    22. Changed the option for creating symbolic links for 16-bit man pages from
450        -s to -sf so that re-installing does not cause issues.
451    
452    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
453    
454    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
455        matches in certain environments (the workspace was not being correctly
456        retained). Also added to pcre_dfa_exec() a simple plausibility check on
457        some of the workspace data at the beginning of a restart.
458    
459    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
460        was not doing so when it should - probably a typo introduced by SVN 528
461        (change 8.10/14).
462    
463    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
464        \w+ when the character tables indicated that \x{c4} was a word character.
465        There were several related cases, all because the tests for doing a table
466        lookup were testing for characters less than 127 instead of 255.
467    
468    27. If a pattern contains capturing parentheses that are not used in a match,
469        their slots in the ovector are set to -1. For those that are higher than
470        any matched groups, this happens at the end of processing. In the case when
471        there were back references that the ovector was too small to contain
472        (causing temporary malloc'd memory to be used during matching), and the
473        highest capturing number was not used, memory off the end of the ovector
474        was incorrectly being set to -1. (It was using the size of the temporary
475        memory instead of the true size.)
476    
477    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
478        ovector size, it uses memory at the end of the block that it has got.
479    
480    29. Check for an overlong MARK name and give an error at compile time. The
481        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
482    
483    30. JIT compiler update.
484    
485    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
486        Rill for the patch.
487    
488    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
489    
490    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
491    
492    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
493        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
494        was enabled.
495    
496    35. Improve JIT code generation for greedy plus quantifier.
497    
498    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
499        1 to "aa" instead of to an empty string. The bug affected repeated groups
500        that could potentially match an empty string.
501    
502    37. Optimizing single character iterators in JIT.
503    
504    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
505        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
506        codepoints that are too big for the mode are faulted, and in a UTF mode,
507        disallowed codepoints are also faulted.
508    
509    39. If PCRE was compiled with UTF support, in three places in the DFA
510        matcher there was code that should only have been obeyed in UTF mode, but
511        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
512        processing when bytes with values greater than 127 were present. In 16-bit
513        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
514        both cases the values are those that cannot be the first data item in a UTF
515        character. The three items that might have provoked this were recursions,
516        possessively repeated groups, and atomic groups.
517    
518    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
519        and pcregrep, because some OS require shared objects to be explicitly
520        passed to ld, causing the link step to fail if they are not.
521    
522    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
523        mode, patterns that started with \h* or \R* might be incorrectly matched.
524    
525    
526    Version 8.30 04-February-2012
527    -----------------------------
528    
529    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
530        name is defined in ctype.h.
531    
532    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
533        only in quite long subpatterns.
534    
535    3.  Removed the function pcre_info(), which has been obsolete and deprecated
536        since it was replaced by pcre_fullinfo() in February 2000.
537    
538    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
539        match a (*MARK), and the match failed at the start of the subject, a
540        reference to memory before the start of the subject could occur. This bug
541        was introduced by fix 17 of release 8.21.
542    
543    5.  A reference to an unset group with zero minimum repetition was giving
544        totally wrong answers (in non-JavaScript-compatibility mode). For example,
545        /(another)?(\1?)test/ matched against "hello world test". This bug was
546        introduced in release 8.13.
547    
548    6.  Add support for 16-bit character strings (a large amount of work involving
549        many changes and refactorings).
550    
551    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
552        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
553        from a file.
554    
555    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
556        rounding is not applied in this particular case).
557    
558    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
559        if they appear, or are escaped, in patterns.
560    
561    10. Get rid of a number of -Wunused-but-set-variable warnings.
562    
563    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
564        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
565        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
566        also returns the mark "x". This bug applied to capturing parentheses,
567        non-capturing parentheses, and atomic parentheses. It also applied to some
568        assertions.
569    
570    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
571        information out of configure.ac instead of relying on pcre.h.generic, which
572        is not stored in the repository.
573    
574    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
575        -lreadline.
576    
577    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
578    
579    15. Applied Graycode's patch to put the top-level frame on the stack rather
580        than the heap when not using the stack for recursion. This gives a
581        performance improvement in many cases when recursion is not deep.
582    
583    16. Experimental code added to "pcretest -C" to output the stack frame size.
584    
585    
586    Version 8.21 12-Dec-2011
587    ------------------------
588    
589    1.  Updating the JIT compiler.
590    
591    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
592        are added as well.
593    
594    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
595        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
596        calling _pcre_jit_exec. Some extra comments are added.
597    
598    4.  (*MARK) settings inside atomic groups that do not contain any capturing
599        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
600        was introduced by change 18 for 8.20.
601    
602    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
603        ECMA-262 standard.
604    
605    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
606        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
607        This bug was probably introduced by change 9 of 8.13.
608    
609    7.  While fixing 6 above, I noticed that a number of other items were being
610        incorrectly rejected as "not fixed length". This arose partly because newer
611        opcodes had not been added to the fixed-length checking code. I have (a)
612        corrected the bug and added tests for these items, and (b) arranged for an
613        error to occur if an unknown opcode is encountered while checking for fixed
614        length instead of just assuming "not fixed length". The items that were
615        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
616        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
617        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
618    
619    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
620        being incorrectly compiled and would have given unpredicatble results.
621    
622    9.  A possessively repeated subpattern with minimum repeat count greater than
623        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
624        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
625        the first (A) could occur when it should not.
626    
627    10. Add a cast and remove a redundant test from the code.
628    
629    11. JIT should use pcre_malloc/pcre_free for allocation.
630    
631    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
632        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
633        is anything other than /usr, -L is still shown).
634    
635    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
636    
637    14. Perl does not support \N without a following name in a [] class; PCRE now
638        also gives an error.
639    
640    15. If a forward reference was repeated with an upper limit of around 2000,
641        it caused the error "internal error: overran compiling workspace". The
642        maximum number of forward references (including repeats) was limited by the
643        internal workspace, and dependent on the LINK_SIZE. The code has been
644        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
645        the default depends on LINK_SIZE. There is a new upper limit (for safety)
646        of around 200,000 forward references. While doing this, I also speeded up
647        the filling in of repeated forward references.
648    
649    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
650        incorrectly expecting the subject to contain another "a" after the start.
651    
652    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
653        in the match, the SKIP should be ignored. This was not happening; instead
654        the SKIP was being treated as NOMATCH. For patterns such as
655        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
656        tested.
657    
658    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
659        now much more compatible with Perl, in particular in cases where the result
660        is a non-match for a non-anchored pattern. For example, if
661        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
662        "m", where previously it did not return a name. A side effect of this
663        change is that for partial matches, the last encountered mark name is
664        returned, as for non matches. A number of tests that were previously not
665        Perl-compatible have been moved into the Perl-compatible test files. The
666        refactoring has had the pleasing side effect of removing one argument from
667        the match() function, thus reducing its stack requirements.
668    
669    19. If the /S+ option was used in pcretest to study a pattern using JIT,
670        subsequent uses of /S (without +) incorrectly behaved like /S+.
671    
672    21. Retrieve executable code size support for the JIT compiler and fixing
673        some warnings.
674    
675    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
676        not work when the shorter character appeared right at the end of the
677        subject string.
678    
679    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
680        systems.
681    
682    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
683        output it when the /M option is used in pcretest.
684    
685    25. The CheckMan script was not being included in the distribution. Also, added
686        an explicit "perl" to run Perl scripts from the PrepareRelease script
687        because this is reportedly needed in Windows.
688    
689    26. If study data was being save in a file and studying had not found a set of
690        "starts with" bytes for the pattern, the data written to the file (though
691        never used) was taken from uninitialized memory and so caused valgrind to
692        complain.
693    
694    27. Updated RunTest.bat as provided by Sheri Pierce.
695    
696    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
697    
698    29. Computation of memory usage for the table of capturing group names was
699        giving an unnecessarily large value.
700    
701    
702    Version 8.20 21-Oct-2011
703    ------------------------
704    
705    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
706        a POSIX class. After further experiments with Perl, which convinced me that
707        Perl has bugs and confusions, a closing square bracket is no longer allowed
708        in a POSIX name. This bug also affected patterns with classes that started
709        with full stops.
710    
711    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
712        captured substring, but while checking the failing first alternative,
713        substring 1 is temporarily captured. If the output vector supplied to
714        pcre_exec() was not big enough for this capture, the yield of the function
715        was still zero ("insufficient space for captured substrings"). This cannot
716        be totally fixed without adding another stack variable, which seems a lot
717        of expense for a edge case. However, I have improved the situation in cases
718        such as /(a)(b)x|abc/ matched against "abc", where the return code
719        indicates that fewer than the maximum number of slots in the ovector have
720        been set.
721    
722    3.  Related to (2) above: when there are more back references in a pattern than
723        slots in the output vector, pcre_exec() uses temporary memory during
724        matching, and copies in the captures as far as possible afterwards. It was
725        using the entire output vector, but this conflicts with the specification
726        that only 2/3 is used for passing back captured substrings. Now it uses
727        only the first 2/3, for compatibility. This is, of course, another edge
728        case.
729    
730    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
731        main code base, and can be used by building with --enable-jit. When this is
732        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
733        runtime --no-jit option is given.
734    
735    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
736        ovector, the return from the function was zero, implying that there were
737        other matches that did not fit. The correct "exactly full" value is now
738        returned.
739    
740    6.  If a subpattern that was called recursively or as a subroutine contained
741        (*PRUNE) or any other control that caused it to give a non-standard return,
742        invalid errors such as "Error -26 (nested recursion at the same subject
743        position)" or even infinite loops could occur.
744    
745    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
746        computing the minimum length on reaching *ACCEPT, and so ended up with the
747        wrong value of 1 rather than 0. Further investigation indicates that
748        computing a minimum subject length in the presence of *ACCEPT is difficult
749        (think back references, subroutine calls), and so I have changed the code
750        so that no minimum is registered for a pattern that contains *ACCEPT.
751    
752    8.  If (*THEN) was present in the first (true) branch of a conditional group,
753        it was not handled as intended. [But see 16 below.]
754    
755    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
756        Sheri Pierce.
757    
758    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
759        the first byte in a match must be "a".
760    
761    11. Change 17 for 8.13 increased the recursion depth for patterns like
762        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
763        pattern contains any instances of (*THEN). If it does not, the old
764        optimizations are restored. It would be nice to do this on a per-group
765        basis, but at the moment that is not feasible.
766    
767    12. In some environments, the output of pcretest -C is CRLF terminated. This
768        broke RunTest's code that checks for the link size. A single white space
769        character after the value is now allowed for.
770    
771    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
772        For "fr", it uses the Windows-specific input and output files.
773    
774    14. If (*THEN) appeared in a group that was called recursively or as a
775        subroutine, it did not work as intended. [But see next item.]
776    
777    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
778        pattern fragments (but not containing any | characters). If A and B are
779        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
780        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
781        D. In other words, Perl considers parentheses that do not contain any |
782        characters to be part of a surrounding alternative, whereas PCRE was
783        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
784        differently. PCRE now behaves in the same way as Perl, except in the case
785        of subroutine/recursion calls such as (?1) which have in any case always
786        been different (but PCRE had them first :-).
787    
788    16. Related to 15 above: Perl does not treat the | in a conditional group as
789        creating alternatives. Such a group is treated in the same way as an
790        ordinary group without any | characters when processing (*THEN). PCRE has
791        been changed to match Perl's behaviour.
792    
793    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
794        RunGrepTest script failed.
795    
796    18. Change 22 for version 13 caused atomic groups to use more stack. This is
797        inevitable for groups that contain captures, but it can lead to a lot of
798        stack use in large patterns. The old behaviour has been restored for atomic
799        groups that do not contain any capturing parentheses.
800    
801    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
802        suppress the check for a minimum subject length at run time. (If it was
803        given to pcre_exec() or pcre_dfa_exec() it did work.)
804    
805    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
806        fail to work when decoding hex characters in data strings in EBCDIC
807        environments.
808    
809    21. It appears that in at least one Mac OS environment, the isxdigit() function
810        is implemented as a macro that evaluates to its argument more than once,
811        contravening the C 90 Standard (I haven't checked a later standard). There
812        was an instance in pcretest which caused it to go wrong when processing
813        \x{...} escapes in subject strings. The has been rewritten to avoid using
814        things like p++ in the argument of isxdigit().
815    
816    
817    Version 8.13 16-Aug-2011
818    ------------------------
819    
820    1.  The Unicode data tables have been updated to Unicode 6.0.0.
821    
822    2.  Two minor typos in pcre_internal.h have been fixed.
823    
824    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
825        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
826        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
827    
828    4.  There were a number of related bugs in the code for matching backrefences
829        caselessly in UTF-8 mode when codes for the characters concerned were
830        different numbers of bytes. For example, U+023A and U+2C65 are an upper
831        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
832        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
833        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
834        2-byte code at the end of the subject (it thought there wasn't enough data
835        left).
836    
837    5.  Comprehensive information about what went wrong is now returned by
838        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
839        as the output vector has at least 2 elements. The offset of the start of
840        the failing character and a reason code are placed in the vector.
841    
842    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
843        now returned is for the first byte of the failing character, instead of the
844        last byte inspected. This is an incompatible change, but I hope it is small
845        enough not to be a problem. It makes the returned offset consistent with
846        pcre_exec() and pcre_dfa_exec().
847    
848    7.  pcretest now gives a text phrase as well as the error number when
849        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
850        failure, the offset and reason code are output.
851    
852    8.  When \R was used with a maximizing quantifier it failed to skip backwards
853        over a \r\n pair if the subsequent match failed. Instead, it just skipped
854        back over a single character (\n). This seems wrong (because it treated the
855        two characters as a single entity when going forwards), conflicts with the
856        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
857        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
858        has been changed.
859    
860    9.  Some internal refactoring has changed the processing so that the handling
861        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
862        time (the PCRE_DOTALL option was changed this way some time ago: version
863        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
864        which was always a bit of a fudge. It also means that there is one less
865        argument for the match() function, which reduces its stack requirements
866        slightly. This change also fixes an incompatibility with Perl: the pattern
867        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
868    
869    10. More internal refactoring has drastically reduced the number of recursive
870        calls to match() for possessively repeated groups such as (abc)++ when
871        using pcre_exec().
872    
873    11. While implementing 10, a number of bugs in the handling of groups were
874        discovered and fixed:
875    
876        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
877        (a|)*(?1) gave a compile-time internal error.
878        ((a|)+)+  did not notice that the outer group could match an empty string.
879        (^a|^)+   was not marked as anchored.
880        (.*a|.*)+ was not marked as matching at start or after a newline.
881    
882    12. Yet more internal refactoring has removed another argument from the match()
883        function. Special calls to this function are now indicated by setting a
884        value in a variable in the "match data" data block.
885    
886    13. Be more explicit in pcre_study() instead of relying on "default" for
887        opcodes that mean there is no starting character; this means that when new
888        ones are added and accidentally left out of pcre_study(), testing should
889        pick them up.
890    
891    14. The -s option of pcretest has been documented for ages as being an old
892        synonym of -m (show memory usage). I have changed it to mean "force study
893        for every regex", that is, assume /S for every regex. This is similar to -i
894        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
895        using it. It makes it easier to run collections of tests with and without
896        study enabled, and thereby test pcre_study() more easily. All the standard
897        tests are now run with and without -s (but some patterns can be marked as
898        "never study" - see 20 below).
899    
900    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
901        restoration of the capturing data to the outer values was not happening
902        correctly.
903    
904    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
905        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
906        pattern had matched an empty string, and so incorrectly returned a no
907        match.
908    
909    17. There was optimizing code for the last branch of non-capturing parentheses,
910        and also for the obeyed branch of a conditional subexpression, which used
911        tail recursion to cut down on stack usage. Unfortunately, now that there is
912        the possibility of (*THEN) occurring in these branches, tail recursion is
913        no longer possible because the return has to be checked for (*THEN). These
914        two optimizations have therefore been removed. [But see 8.20/11 above.]
915    
916    18. If a pattern containing \R was studied, it was assumed that \R always
917        matched two bytes, thus causing the minimum subject length to be
918        incorrectly computed because \R can also match just one byte.
919    
920    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
921        was incorrectly computed.
922    
923    20. If /S is present twice on a test pattern in pcretest input, it now
924        *disables* studying, thereby overriding the use of -s on the command line
925        (see 14 above). This is necessary for one or two tests to keep the output
926        identical in both cases.
927    
928    21. When (*ACCEPT) was used in an assertion that matched an empty string and
929        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
930    
931    22. When an atomic group that contained a capturing parenthesis was
932        successfully matched, but the branch in which it appeared failed, the
933        capturing was not being forgotten if a higher numbered group was later
934        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
935        group 1 to "a", when in fact it should be unset. This applied to multi-
936        branched capturing and non-capturing groups, repeated or not, and also to
937        positive assertions (capturing in negative assertions does not happen
938        in PCRE) and also to nested atomic groups.
939    
940    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
941        subject after a captured substring, to make it easier to tell which of a
942        number of identical substrings has been captured.
943    
944    24. The way atomic groups are processed by pcre_exec() has been changed so that
945        if they are repeated, backtracking one repetition now resets captured
946        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
947        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
948        "aaa". Previously, it would have been "a". As part of this code
949        refactoring, the way recursive calls are handled has also been changed.
950    
951    25. If an assertion condition captured any substrings, they were not passed
952        back unless some other capturing happened later. For example, if
953        (?(?=(a))a) was matched against "a", no capturing was returned.
954    
955    26. When studying a pattern that contained subroutine calls or assertions,
956        the code for finding the minimum length of a possible match was handling
957        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
958        group 1 called group 2 while simultaneously a separate group 2 called group
959        1). A stack overflow occurred in this case. I have fixed this by limiting
960        the recursion depth to 10.
961    
962    27. Updated RunTest.bat in the distribution to the version supplied by Tom
963        Fortmann. This supports explicit test numbers on the command line, and has
964        argument validation and error reporting.
965    
966    28. An instance of \X with an unlimited repeat could fail if at any point the
967        first character it looked at was a mark character.
968    
969    29. Some minor code refactoring concerning Unicode properties and scripts
970        should reduce the stack requirement of match() slightly.
971    
972    30. Added the '=' option to pcretest to check the setting of unused capturing
973        slots at the end of the pattern, which are documented as being -1, but are
974        not included in the return count.
975    
976    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
977        compiled something random. Now it gives a compile-time error (as does
978        Perl).
979    
980    32. A *MARK encountered during the processing of a positive assertion is now
981        recorded and passed back (compatible with Perl).
982    
983    33. If --only-matching or --colour was set on a pcregrep call whose pattern
984        had alternative anchored branches, the search for a second match in a line
985        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
986        matched the line "0102" twice. The same bug affected patterns that started
987        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
988        twice.
989    
990    34. Previously, PCRE did not allow quantification of assertions. However, Perl
991        does, and because of capturing effects, quantifying parenthesized
992        assertions may at times be useful. Quantifiers are now allowed for
993        parenthesized assertions.
994    
995    35. A minor code tidy in pcre_compile() when checking options for \R usage.
996    
997    36. \g was being checked for fancy things in a character class, when it should
998        just be a literal "g".
999    
1000    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1001        appearance of a nested POSIX class supersedes an apparent external class.
1002        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1003        unescaped square brackets may also appear as part of class names. For
1004        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1005        more like Perl. (But see 8.20/1 above.)
1006    
1007    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1008        was because it thought it was \N{name}, which is not supported).
1009    
1010    39. Add minix to OS list not supporting the -S option in pcretest.
1011    
1012    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1013        cannot analyze patterns in sufficient detail to catch mutual recursions
1014        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1015        subgroup is called recursively as a subpattern for a second time at the
1016        same position in the subject string. In previous releases this might have
1017        been caught by the recursion limit, or it might have run out of stack.
1018    
1019    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1020        happen only once. PCRE was, however incorrectly giving a compile time error
1021        "recursive call could loop indefinitely" because it cannot analyze the
1022        pattern in sufficient detail. The compile time test no longer happens when
1023        PCRE is compiling a conditional subpattern, but actual runaway loops are
1024        now caught at runtime (see 40 above).
1025    
1026    42. It seems that Perl allows any characters other than a closing parenthesis
1027        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1028        has been changed to be the same.
1029    
1030    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1031        as not to get warnings when autogen.sh is called. Also changed
1032        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1033    
1034    44. To help people who use pcregrep to scan files containing exceedingly long
1035        lines, the following changes have been made:
1036    
1037        (a) The default value of the buffer size parameter has been increased from
1038            8K to 20K. (The actual buffer used is three times this size.)
1039    
1040        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1041            PCRE is built.
1042    
1043        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1044            to be set at run time.
1045    
1046        (d) Numerical values in pcregrep options can be followed by K or M, for
1047            example --buffer-size=50K.
1048    
1049        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1050            given and the return code is set to 2.
1051    
1052    45. Add a pointer to the latest mark to the callout data block.
1053    
1054    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1055        partial match of an empty string instead of no match. This was specific to
1056        the use of ".".
1057    
1058    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1059        complete match instead of a partial match. This bug was dependent on both
1060        the PCRE_UTF8 and PCRE_DOTALL options being set.
1061    
1062    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1063        starting byte set, because \b was not being ignored.
1064    
1065    
1066    Version 8.12 15-Jan-2011
1067    ------------------------
1068    
1069    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1070        checks for such things as part of the documentation building process.
1071    
1072    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1073        --match-limit and --recursion-limit options (added for 8.11). In
1074        particular, this made one of the standard tests fail. (The integer value
1075        went into the wrong half of a long int.)
1076    
1077    3.  If the --colour option was given to pcregrep with -v (invert match), it
1078        did strange things, either producing crazy output, or crashing. It should,
1079        of course, ignore a request for colour when reporting lines that do not
1080        match.
1081    
1082    4.  Another pcregrep bug caused similar problems if --colour was specified with
1083        -M (multiline) and the pattern match finished with a line ending.
1084    
1085    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1086        matched in multiline mode, the following line was shown as part of the
1087        match. This seems wrong, so I have changed it.
1088    
1089    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1090        the check for further matches in the same line (so they could be coloured)
1091        to overrun the end of the current line. If another match was found, it was
1092        incorrectly shown (and then shown again when found in the next line).
1093    
1094    7.  If pcregrep was compiled under Windows, there was a reference to the
1095        function pcregrep_exit() before it was defined. I am assuming this was
1096        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1097        reported by a user. I've moved the definition above the reference.
1098    
1099    
1100    Version 8.11 10-Dec-2010
1101    ------------------------
1102    
1103    1.  (*THEN) was not working properly if there were untried alternatives prior
1104        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1105        backtracked to try for "b" instead of moving to the next alternative branch
1106        at the same level (in this case, to look for "c"). The Perl documentation
1107        is clear that when (*THEN) is backtracked onto, it goes to the "next
1108        alternative in the innermost enclosing group".
1109    
1110    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1111        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1112        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1113        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1114        (*THEN).
1115    
1116    3.  If \s appeared in a character class, it removed the VT character from
1117        the class, even if it had been included by some previous item, for example
1118        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1119        of \s, but is part of the POSIX "space" class.)
1120    
1121    4.  A partial match never returns an empty string (because you can always
1122        match an empty string at the end of the subject); however the checking for
1123        an empty string was starting at the "start of match" point. This has been
1124        changed to the "earliest inspected character" point, because the returned
1125        data for a partial match starts at this character. This means that, for
1126        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1127        (previously it gave "no match").
1128    
1129    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1130        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1131        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1132        has an implication that the given string is incomplete (because a partial
1133        match is preferred over a full match). For this reason, these items now
1134        give a partial match in this situation. [Aside: previously, the one case
1135        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1136        match rather than a full match, which was wrong by the old rules, but is
1137        now correct.]
1138    
1139    6.  There was a bug in the handling of #-introduced comments, recognized when
1140        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1141        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1142        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1143        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1144        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1145        places in pcre_compile().
1146    
1147    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1148        comments when looking ahead for named forward references to subpatterns,
1149        the only newline sequence it recognized was NL. It now handles newlines
1150        according to the set newline convention.
1151    
1152    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1153        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1154        cater for a lack of strerror(). These oversights have been fixed.
1155    
1156    9.  Added --match-limit and --recursion-limit to pcregrep.
1157    
1158    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1159    
1160    11. When the -o option was used, pcregrep was setting a return code of 1, even
1161        when matches were found, and --line-buffered was not being honoured.
1162    
1163    12. Added an optional parentheses number to the -o and --only-matching options
1164        of pcregrep.
1165    
1166    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1167        can match an empty string. The code to do it in pcretest and pcredemo
1168        needed fixing:
1169    
1170        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1171            only one byte after an empty string match just before CRLF (this case
1172            just got forgotten; "any" and "anycrlf" were OK).
1173    
1174        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1175            mode when an empty string match preceded an ASCII character followed by
1176            a non-ASCII character. (The code for advancing by one character rather
1177            than one byte was nonsense.)
1178    
1179        (c) The pcredemo.c sample program did not have any code at all to handle
1180            the cases when CRLF is a valid newline sequence.
1181    
1182    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1183        as a starting offset was within the subject string. There is now a new
1184        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1185        negative or greater than the length of the string. In order to test this,
1186        pcretest is extended to allow the setting of negative starting offsets.
1187    
1188    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1189        starting offset points to the beginning of a UTF-8 character was
1190        unnecessarily clumsy. I tidied it up.
1191    
1192    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1193        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1194    
1195    17. Nobody had reported that the --include_dir option, which was added in
1196        release 7.7 should have been called --include-dir (hyphen, not underscore)
1197        for compatibility with GNU grep. I have changed it to --include-dir, but
1198        left --include_dir as an undocumented synonym, and the same for
1199        --exclude-dir, though that is not available in GNU grep, at least as of
1200        release 2.5.4.
1201    
1202    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1203        characters from a string of bytes) have been redefined so as not to use
1204        loops, in order to improve performance in some environments. At the same
1205        time, I abstracted some of the common code into auxiliary macros to save
1206        repetition (this should not affect the compiled code).
1207    
1208    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1209        compile-time error is now given if \c is not followed by an ASCII
1210        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1211        different, and any byte value is allowed.)
1212    
1213    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1214        START_OPTIMIZE option, which is now allowed at compile time - but just
1215        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1216        to pcregrep and other applications that have no direct access to PCRE
1217        options. The new /Y option in pcretest sets this option when calling
1218        pcre_compile().
1219    
1220    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1221        back references. Groups containing recursive back references were forced to
1222        be atomic by that change, but in the case of named groups, the amount of
1223        memory required was incorrectly computed, leading to "Failed: internal
1224        error: code overflow". This has been fixed.
1225    
1226    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1227        pcretest.c, to avoid build problems in some Borland environments.
1228    
1229    
1230    Version 8.10 25-Jun-2010
1231    ------------------------
1232    
1233    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1234        THEN.
1235    
1236    2.  (*ACCEPT) was not working when inside an atomic group.
1237    
1238    3.  Inside a character class, \B is treated as a literal by default, but
1239        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1240        causes the error). The code is unchanged, but I tidied the documentation.
1241    
1242    4.  Inside a character class, PCRE always treated \R and \X as literals,
1243        whereas Perl faults them if its -w option is set. I have changed PCRE so
1244        that it faults them when PCRE_EXTRA is set.
1245    
1246    5.  Added support for \N, which always matches any character other than
1247        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1248    
1249    6.  When compiling pcregrep with newer versions of gcc which may have
1250        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1251        declared with attribute warn_unused_result" were given. Just casting the
1252        result to (void) does not stop the warnings; a more elaborate fudge is
1253        needed. I've used a macro to implement this.
1254    
1255    7.  Minor change to pcretest.c to avoid a compiler warning.
1256    
1257    8.  Added four artifical Unicode properties to help with an option to make
1258        \s etc use properties (see next item). The new properties are: Xan
1259        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1260    
1261    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1262        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1263        this option. Modified pcretest to add /W to test this facility. Added
1264        REG_UCP to make it available via the POSIX interface.
1265    
1266    10. Added --line-buffered to pcregrep.
1267    
1268    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1269        studied, and the match started with a letter with a code point greater than
1270        127 whose first byte was different to the first byte of the other case of
1271        the letter, the other case of this starting letter was not recognized
1272        (#976).
1273    
1274    12. If a pattern that was studied started with a repeated Unicode property
1275        test, for example, \p{Nd}+, there was the theoretical possibility of
1276        setting up an incorrect bitmap of starting bytes, but fortunately it could
1277        not have actually happened in practice until change 8 above was made (it
1278        added property types that matched character-matching opcodes).
1279    
1280    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1281        possible starting bytes for non-anchored patterns.
1282    
1283    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1284        \R, and also a number of cases that involve Unicode properties, both
1285        explicit and implicit when PCRE_UCP is set.
1286    
1287    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1288        input, it could crash or give wrong results if characters with values
1289        greater than 0xc0 were present in the subject string. (Detail: it assumed
1290        UTF-8 input when processing these items.)
1291    
1292    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1293        size_t is 64-bit (#991).
1294    
1295    17. Added a check for running out of memory when PCRE is compiled with
1296        --disable-stack-for-recursion (#990).
1297    
1298    18. If the last data line in a file for pcretest does not have a newline on
1299        the end, a newline was missing in the output.
1300    
1301    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1302        less than 128) in its various bitmaps. However, there is a facility for
1303        generating tables according to the current locale when PCRE is compiled. It
1304        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1305        characters, are recognized by isspace() and therefore were getting set in
1306        these tables, and indeed these tables seem to approximate to ISO 8859. This
1307        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1308        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1309        which of course cannot start UTF-8 characters. I have changed the code so
1310        that only real ASCII characters (less than 128) and the correct starting
1311        bytes for UTF-8 encodings are set for characters greater than 127 when in
1312        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1313        altogether.)
1314    
1315    20. Added the /T option to pcretest so as to be able to run tests with non-
1316        standard character tables, thus making it possible to include the tests
1317        used for 19 above in the standard set of tests.
1318    
1319    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1320        reference to a subpattern the other side of a comment that contains an
1321        opening parenthesis caused either an internal compiling error, or a
1322        reference to the wrong subpattern.
1323    
1324    
1325    Version 8.02 19-Mar-2010
1326    ------------------------
1327    
1328    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1329    
1330    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1331        configured.
1332    
1333    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1334        original author of that file, following a query about its status.
1335    
1336    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1337        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1338    
1339    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1340        quantifier applied to a forward-referencing subroutine call, could compile
1341        incorrect code or give the error "internal error: previously-checked
1342        referenced subpattern not found".
1343    
1344    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1345        variables to point to external functions. For these systems, therefore,
1346        pcre_malloc etc. are now initialized to local functions that call the
1347        relevant global functions.
1348    
1349    7.  There were two entries missing in the vectors called coptable and poptable
1350        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1351        I've fixed the data, and added a kludgy way of testing at compile time that
1352        the lengths are correct (equal to the number of opcodes).
1353    
1354    8.  Following on from 7, I added a similar kludge to check the length of the
1355        eint vector in pcreposix.c.
1356    
1357    9.  Error texts for pcre_compile() are held as one long string to avoid too
1358        much relocation at load time. To find a text, the string is searched,
1359        counting zeros. There was no check for running off the end of the string,
1360        which could happen if a new error number was added without updating the
1361        string.
1362    
1363    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1364    
1365    11. \K was not working if it appeared in an atomic group or in a group that
1366        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1367        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1368        the assertion is positive, but not if it is negative.
1369    
1370    12. Change 11 fortuitously reduced the size of the stack frame used in the
1371        "match()" function of pcre_exec.c by one pointer. Forthcoming
1372        implementation of support for (*MARK) will need an extra pointer on the
1373        stack; I have reserved it now, so that the stack frame size does not
1374        decrease.
1375    
1376    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1377        item in branch that calls a recursion is a subroutine call - as in the
1378        second branch in the above example - was incorrectly given the compile-
1379        time error "recursive call could loop indefinitely" because pcre_compile()
1380        was not correctly checking the subroutine for matching a non-empty string.
1381    
1382    14. The checks for overrunning compiling workspace could trigger after an
1383        overrun had occurred. This is a "should never occur" error, but it can be
1384        triggered by pathological patterns such as hundreds of nested parentheses.
1385        The checks now trigger 100 bytes before the end of the workspace.
1386    
1387    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1388    
1389    
1390    Version 8.01 19-Jan-2010
1391    ------------------------
1392    
1393    1.  If a pattern contained a conditional subpattern with only one branch (in
1394        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1395        computed the wrong minimum data length (which is of course zero for such
1396        subpatterns). This could cause incorrect "no match" results.
1397    
1398    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1399        the pattern is reset in the first branch, pcre_compile() failed with
1400        "internal error: code overflow at offset...". This happened only when
1401        the reset was to the original external option setting. (An optimization
1402        abstracts leading options settings into an external setting, which was the
1403        cause of this.)
1404    
1405    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1406        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1407        assertion pattern did not match (meaning that the assertion was true), it
1408        was incorrectly treated as false if the SKIP had been reached during the
1409        matching. This also applied to assertions used as conditions.
1410    
1411    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1412        assertion subpattern, including such a pattern used as a condition,
1413        unpredictable results occurred, instead of the error return
1414        PCRE_ERROR_DFA_UITEM.
1415    
1416    5.  The C++ GlobalReplace function was not working like Perl for the special
1417        situation when an empty string is matched. It now does the fancy magic
1418        stuff that is necessary.
1419    
1420    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1421        removed. (These were left over from very, very early versions of PCRE.)
1422    
1423    7.  Some cosmetic changes to the code to make life easier when compiling it
1424        as part of something else:
1425    
1426        (a) Change DEBUG to PCRE_DEBUG.
1427    
1428        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1429            called "current" as "current_branch", to prevent a collision with the
1430            Linux macro when compiled as a kernel module.
1431    
1432        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1433            prevent a collision with the Linux macro when compiled as a kernel
1434            module.
1435    
1436    8.  In pcre_compile() there are some checks for integer overflows that used to
1437        cast potentially large values to (double). This has been changed to that
1438        when building, a check for int64_t is made, and if it is found, it is used
1439        instead, thus avoiding the use of floating point arithmetic. (There is no
1440        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1441        double.
1442    
1443    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1444        2005 (difference between two addresses compared to an unsigned value).
1445    
1446    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1447        custom one, because of the following reported problem in Windows:
1448    
1449          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1450              under Win32.
1451          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1452              therefore missing the function definition.
1453          - The compiler thus generates a "C" signature for the test function.
1454          - The linker fails to find the "C" function.
1455          - PCRE fails to configure if asked to do so against libbz2.
1456    
1457    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1458        messages were output:
1459    
1460          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1461          rerunning libtoolize, to keep the correct libtool macros in-tree.
1462          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1463    
1464        I have done both of these things.
1465    
1466    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1467        most of the time, it *can* run out if it is given a pattern that contains a
1468        runaway infinite recursion. I updated the discussion in the pcrestack man
1469        page.
1470    
1471    13. Now that we have gone to the x.xx style of version numbers, the minor
1472        version may start with zero. Using 08 or 09 is a bad idea because users
1473        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1474        interpreted as invalid octal numbers. I've updated the previous comment in
1475        configure.ac, and also added a check that gives an error if 08 or 09 are
1476        used.
1477    
1478    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1479        causing partial matching to fail when the end of the subject matched \W
1480        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1481    
1482    15. There were some discrepancies between the declarations in pcre_internal.h
1483        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1484        their definitions. The declarations used "const uschar *" and the
1485        definitions used USPTR. Even though USPTR is normally defined as "const
1486        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1487        reported that: "This difference in casting confuses some C++ compilers, for
1488        example, SunCC recognizes above declarations as different functions and
1489        generates broken code for hbpcre." I have changed the declarations to use
1490        USPTR.
1491    
1492    16. GNU libtool is named differently on some systems. The autogen.sh script now
1493        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1494        (FreeBSD).
1495    
1496    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1497        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1498        comment: "Figure out how to create a longlong from a string: strtoll and
1499        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1500        instance, but it only takes 2 args instead of 3!"
1501    
1502    18. A subtle bug concerned with back references has been fixed by a change of
1503        specification, with a corresponding code fix. A pattern such as
1504        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1505        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1506        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1507        same bug. Such groups have to be quantified to be useful, or contained
1508        inside another quantified group. (If there's no repetition, the reference
1509        can never match.) The problem arises because, having left the group and
1510        moved on to the rest of the pattern, a later failure that backtracks into
1511        the group uses the captured value from the final iteration of the group
1512        rather than the correct earlier one. I have fixed this in PCRE by forcing
1513        any group that contains a reference to itself to be an atomic group; that
1514        is, there cannot be any backtracking into it once it has completed. This is
1515        similar to recursive and subroutine calls.
1516    
1517    
1518    Version 8.00 19-Oct-09
1519    ----------------------
1520    
1521    1.  The table for translating pcre_compile() error codes into POSIX error codes
1522        was out-of-date, and there was no check on the pcre_compile() error code
1523        being within the table. This could lead to an OK return being given in
1524        error.
1525    
1526    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1527        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1528        in a Windows environment.
1529    
1530    3.  The pcregrep --count option prints the count for each file even when it is
1531        zero, as does GNU grep. However, pcregrep was also printing all files when
1532        --files-with-matches was added. Now, when both options are given, it prints
1533        counts only for those files that have at least one match. (GNU grep just
1534        prints the file name in this circumstance, but including the count seems
1535        more useful - otherwise, why use --count?) Also ensured that the
1536        combination -clh just lists non-zero counts, with no names.
1537    
1538    4.  The long form of the pcregrep -F option was incorrectly implemented as
1539        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1540        but it seems right to fix it, and I didn't think it was worth preserving
1541        the old behaviour.
1542    
1543    5.  The command line items --regex=pattern and --regexp=pattern were not
1544        recognized by pcregrep, which required --regex pattern or --regexp pattern
1545        (with a space rather than an '='). The man page documented the '=' forms,
1546        which are compatible with GNU grep; these now work.
1547    
1548    6.  No libpcreposix.pc file was created for pkg-config; there was just
1549        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1550    
1551    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1552        when UCP support is not needed, by modifying the Python script that
1553        generates it from Unicode data files. This should not matter if the module
1554        is correctly used as a library, but I received one complaint about 50K of
1555        unwanted data. My guess is that the person linked everything into his
1556        program rather than using a library. Anyway, it does no harm.
1557    
1558    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1559        was a minimum greater than 1 for a wide character in a possessive
1560        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1561        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1562        character. Chaos in the form of incorrect output or a compiling loop could
1563        result.
1564    
1565    9.  The restrictions on what a pattern can contain when partial matching is
1566        requested for pcre_exec() have been removed. All patterns can now be
1567        partially matched by this function. In addition, if there are at least two
1568        slots in the offset vector, the offset of the earliest inspected character
1569        for the match and the offset of the end of the subject are set in them when
1570        PCRE_ERROR_PARTIAL is returned.
1571    
1572    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1573        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1574        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1575        and may be more useful for multi-segment matching.
1576    
1577    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1578        used to be given if ever the end of the subject was reached; now it is
1579        given only if matching could not proceed because another character was
1580        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1581        string "Z", which now yields "no match" instead of "partial match". In the
1582        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1583        final character ended with (*FAIL).
1584    
1585    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1586        if the pattern had a "must contain" character that was already found in the
1587        earlier partial match, unless partial matching was again requested. For
1588        example, with the pattern /dog.(body)?/, the "must contain" character is
1589        "g". If the first part-match was for the string "dog", restarting with
1590        "sbody" failed. This bug has been fixed.
1591    
1592    13. The string returned by pcre_dfa_exec() after a partial match has been
1593        changed so that it starts at the first inspected character rather than the
1594        first character of the match. This makes a difference only if the pattern
1595        starts with a lookbehind assertion or \b or \B (\K is not supported by
1596        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1597        matching functions compatible, and I think it's the right thing to do.
1598    
1599    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1600        so that the demonstration program is easily available in environments where
1601        PCRE has not been installed from source.
1602    
1603    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1604        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1605        library.
1606    
1607    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1608        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1609        is not the first non-POSIX option to be added. Clearly some people find
1610        these options useful.
1611    
1612    17. If a caller to the POSIX matching function regexec() passes a non-zero
1613        value for nmatch with a NULL value for pmatch, the value of
1614        nmatch is forced to zero.
1615    
1616    18. RunGrepTest did not have a test for the availability of the -u option of
1617        the diff command, as RunTest does. It now checks in the same way as
1618        RunTest, and also checks for the -b option.
1619    
1620    19. If an odd number of negated classes containing just a single character
1621        interposed, within parentheses, between a forward reference to a named
1622        subpattern and the definition of the subpattern, compilation crashed with
1623        an internal error, complaining that it could not find the referenced
1624        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1625        [The bug was that it was starting one character too far in when skipping
1626        over the character class, thus treating the ] as data rather than
1627        terminating the class. This meant it could skip too much.]
1628    
1629    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1630        /g option in pcretest when the pattern contains \K, which makes it possible
1631        to have an empty string match not at the start, even when the pattern is
1632        anchored. Updated pcretest and pcredemo to use this option.
1633    
1634    21. If the maximum number of capturing subpatterns in a recursion was greater
1635        than the maximum at the outer level, the higher number was returned, but
1636        with unset values at the outer level. The correct (outer level) value is
1637        now given.
1638    
1639    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1640        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1641        make it do so. The string so far is captured, making this feature
1642        compatible with Perl.
1643    
1644    23. The tests have been re-organized, adding tests 11 and 12, to make it
1645        possible to check the Perl 5.10 features against Perl 5.10.
1646    
1647    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1648        pattern matches a fixed length string. PCRE did not allow this; now it
1649        does. Neither allows recursion.
1650    
1651    25. I finally figured out how to implement a request to provide the minimum
1652        length of subject string that was needed in order to match a given pattern.
1653        (It was back references and recursion that I had previously got hung up
1654        on.) This code has now been added to pcre_study(); it finds a lower bound
1655        to the length of subject needed. It is not necessarily the greatest lower
1656        bound, but using it to avoid searching strings that are too short does give
1657        some useful speed-ups. The value is available to calling programs via
1658        pcre_fullinfo().
1659    
1660    26. While implementing 25, I discovered to my embarrassment that pcretest had
1661        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1662        study optimizations had never been tested with that matching function.
1663        Oops. What is worse, even when it was passed study data, there was a bug in
1664        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1665        were also very few tests of studied patterns with pcre_dfa_exec().
1666    
1667    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1668        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1669        on the other side of the coin, they are no longer allowed to have different
1670        names, because these cannot be distinguished in PCRE, and this has caused
1671        confusion. (This is a difference from Perl.)
1672    
1673    28. When duplicate subpattern names are present (necessarily with different
1674        numbers, as required by 27 above), and a test is made by name in a
1675        conditional pattern, either for a subpattern having been matched, or for
1676        recursion in such a pattern, all the associated numbered subpatterns are
1677        tested, and the overall condition is true if the condition is true for any
1678        one of them. This is the way Perl works, and is also more like the way
1679        testing by number works.
1680    
1681    
1682    Version 7.9 11-Apr-09
1683    ---------------------
1684    
1685    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1686        (pcretest), all targets were linked against these libraries. This included
1687        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1688        libraries. This caused unwanted dependencies to be created. This problem
1689        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1690        pcretest is linked with readline.
1691    
1692    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1693        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1694        moved outside it again, because FALSE and TRUE are already defined in AIX,
1695        but BOOL is not.
1696    
1697    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1698        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1699    
1700    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1701        hyphens) following file names and line numbers when outputting matching
1702        lines. This is not true; no spaces are inserted. I have also clarified the
1703        wording for the --colour (or --color) option.
1704    
1705    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1706        was not coloured; this is different to GNU grep, so I have changed it to be
1707        the same.
1708    
1709    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1710        each matching line was coloured. Now it goes on to look for further matches
1711        of any of the test patterns, which is the same behaviour as GNU grep.
1712    
1713    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1714        doesn't make sense to accept an empty string match in pcregrep, so I have
1715        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1716        seems to be how GNU grep behaves. [But see later change 40 for release
1717        8.33.]
1718    
1719    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1720        start or after a newline", because the conditional assertion was not being
1721        correctly handled. The rule now is that both the assertion and what follows
1722        in the first alternative must satisfy the test.
1723    
1724    9.  If auto-callout was enabled in a pattern with a conditional group whose
1725        condition was an assertion, PCRE could crash during matching, both with
1726        pcre_exec() and pcre_dfa_exec().
1727    
1728    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1729        used for matching.
1730    
1731    11. Unicode property support in character classes was not working for
1732        characters (bytes) greater than 127 when not in UTF-8 mode.
1733    
1734    12. Added the -M command line option to pcretest.
1735    
1736    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1737    
1738    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1739    
1740    16. Added comments and documentation about mis-use of no_arg in the C++
1741        wrapper.
1742    
1743    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1744        from Martin Jerabek that uses macro names for all relevant character and
1745        string constants.
1746    
1747    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1748        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1749        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1750        these, but not everybody uses configure.
1751    
1752    19. A conditional group that had only one branch was not being correctly
1753        recognized as an item that could match an empty string. This meant that an
1754        enclosing group might also not be so recognized, causing infinite looping
1755        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1756        with the subject "ab", where knowledge that the repeated group can match
1757        nothing is needed in order to break the loop.
1758    
1759    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1760        exec(), but without supplying a callout function, matching went wrong.
1761    
1762    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1763        leak if the size of the offset vector was greater than 30. When the vector
1764        is smaller, the saved offsets during recursion go onto a local stack
1765        vector, but for larger vectors malloc() is used. It was failing to free
1766        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1767        error, in fact).
1768    
1769    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1770        heapframe that is used only when UTF-8 support is enabled. This caused no
1771        problem, but was untidy.
1772    
1773    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1774        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1775        included within another project.
1776    
1777    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1778        slightly modified by me:
1779    
1780          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1781              not building pcregrep.
1782    
1783          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1784              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1785    
1786    25. Forward references, both numeric and by name, in patterns that made use of
1787        duplicate group numbers, could behave incorrectly or give incorrect errors,
1788        because when scanning forward to find the reference group, PCRE was not
1789        taking into account the duplicate group numbers. A pattern such as
1790        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1791    
1792    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1793        the feature of a custom pointer more persuasive (as requested by a user).
1794    
1795    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1796        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1797        pre-defined.
1798    
1799    28. Added support for (*UTF8) at the start of a pattern.
1800    
1801    29. Arrange for flags added by the "release type" setting in CMake to be shown
1802        in the configuration summary.
1803    
1804    
1805    Version 7.8 05-Sep-08
1806    ---------------------
1807    
1808    1.  Replaced UCP searching code with optimized version as implemented for Ad
1809        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1810        stage table and inline lookup instead of a function, giving speed ups of 2
1811        to 5 times on some simple patterns that I tested. Permission was given to
1812        distribute the MultiStage2.py script that generates the tables (it's not in
1813        the tarball, but is in the Subversion repository).
1814    
1815    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1816        scripts.
1817    
1818    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1819        a group with a zero qualifier. The result of the study could be incorrect,
1820        or the function might crash, depending on the pattern.
1821    
1822    4.  Caseless matching was not working for non-ASCII characters in back
1823        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1824        It now works when Unicode Property Support is available.
1825    
1826    5.  In pcretest, an escape such as \x{de} in the data was always generating
1827        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1828        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1829        truncation.
1830    
1831    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1832    
1833    7.  Added two (int) casts to pcregrep when printing the difference of two
1834        pointers, in case they are 64-bit values.
1835    
1836    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1837        test 2 if it fails.
1838    
1839    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1840        and a #define of that name to empty if it is not externally set. This is to
1841        allow users of MSVC to set it if necessary.
1842    
1843    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1844        the convenience functions in the pcre_get.c source file.
1845    
1846    11. An option change at the start of a pattern that had top-level alternatives
1847        could cause overwriting and/or a crash. This command provoked a crash in
1848        some environments:
1849    
1850          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1851    
1852        This potential security problem was recorded as CVE-2008-2371.
1853    
1854    12. For a pattern where the match had to start at the beginning or immediately
1855        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1856        pcre_dfa_exec() could read past the end of the passed subject if there was
1857        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1858        pcretest so that it places the subject at the end of its malloc-ed buffer.
1859    
1860    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1861        exec() might read past the end of the data buffer in UTF-8 mode.
1862    
1863    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1864        the data contained the byte 0x85 as part of a UTF-8 character within its
1865        first line. This applied both to normal and DFA matching.
1866    
1867    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1868        /^[^d]*?$/8 failed to match "abc".
1869    
1870    16. Added a missing copyright notice to pcrecpp_internal.h.
1871    
1872    17. Make it more clear in the documentation that values returned from
1873        pcre_exec() in ovector are byte offsets, not character counts.
1874    
1875    18. Tidied a few places to stop certain compilers from issuing warnings.
1876    
1877    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1878        supplied by Stefan Weber. I made a further small update for 7.8 because
1879        there is a change of source arrangements: the pcre_searchfuncs.c module is
1880        replaced by pcre_ucd.c.
1881    
1882    
1883    Version 7.7 07-May-08
1884  ---------------------  ---------------------
1885    
1886  1.  Applied Craig's patch to sort out a long long problem: "If we can't convert  1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1887      a string to a long long, pretend we don't even have a long long." This is      a string to a long long, pretend we don't even have a long long." This is
1888      done by checking for the strtoq, strtoll, and _strtoi64 functions.      done by checking for the strtoq, strtoll, and _strtoi64 functions.
1889    
1890  2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with  2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1891      pre-7.6 versions, which defined a global no_arg variable instead of putting      pre-7.6 versions, which defined a global no_arg variable instead of putting
1892      it in the RE class.      it in the RE class. (See also #8 below.)
1893    
1894  3.  Remove a line of dead code, identified by coverity and reported by Nuno  3.  Remove a line of dead code, identified by coverity and reported by Nuno
1895      Lopes.      Lopes.
1896    
1897  4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:  4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1898    
1899      (1) The include/exclude patterns were being applied to the whole pathnames      (1) The include/exclude patterns were being applied to the whole pathnames
1900          of files, instead of just to the final components.          of files, instead of just to the final components.
1901    
1902      (2) If there was more than one level of directory, the subdirectories were      (2) If there was more than one level of directory, the subdirectories were
1903          skipped unless they satisfied the include/exclude conditions. This is          skipped unless they satisfied the include/exclude conditions. This is
1904          inconsistent with GNU grep (and could even be seen as contrary to the          inconsistent with GNU grep (and could even be seen as contrary to the
1905          pcregrep specification - which I improved to make it absolutely clear).          pcregrep specification - which I improved to make it absolutely clear).
1906          The action now is always to scan all levels of directory, and just          The action now is always to scan all levels of directory, and just
1907          apply the include/exclude patterns to regular files.          apply the include/exclude patterns to regular files.
1908    
1909  5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used  5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1910      --exclude_dir in the tests to avoid scanning .svn directories.      --exclude_dir in the tests to avoid scanning .svn directories.
1911    
1912  6.  Applied Craig's patch to the QuoteMeta function so that it escapes the  6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1913      NUL character as backslash + 0 rather than backslash + NUL, because PCRE      NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1914      doesn't support NULs in patterns.      doesn't support NULs in patterns.
1915    
1916  7.  Added some missing "const"s to declarations of static tables in  7.  Added some missing "const"s to declarations of static tables in
1917      pcre_compile.c and pcre_dfa_exec.c.      pcre_compile.c and pcre_dfa_exec.c.
1918    
1919    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1920        caused by fix #2  above. (Subsequently also a second patch to fix the
1921        first patch. And a third patch - this was a messy problem.)
1922    
1923    9.  Applied Craig's patch to remove the use of push_back().
1924    
1925    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1926        matching function regexec().
1927    
1928    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1929        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1930        references. PCRE supports relative numbers with this syntax (I don't think
1931        Oniguruma does).
1932    
1933    12. Previously, a group with a zero repeat such as (...){0} was completely
1934        omitted from the compiled regex. However, this means that if the group
1935        was called as a subroutine from elsewhere in the pattern, things went wrong
1936        (an internal error was given). Such groups are now left in the compiled
1937        pattern, with a new opcode that causes them to be skipped at execution
1938        time.
1939    
1940    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1941        to the way PCRE behaves:
1942    
1943        (a) A lone ] character is dis-allowed (Perl treats it as data).
1944    
1945        (b) A back reference to an unmatched subpattern matches an empty string
1946            (Perl fails the current match path).
1947    
1948        (c) A data ] in a character class must be notated as \] because if the
1949            first data character in a class is ], it defines an empty class. (In
1950            Perl it is not possible to have an empty class.) The empty class []
1951            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1952            The negative empty class [^] matches any one character, independently
1953            of the DOTALL setting.
1954    
1955    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1956        non-existent subpattern following a character class starting with ']' and
1957        containing () gave an internal compiling error instead of "reference to
1958        non-existent subpattern". Fortunately, when the pattern did exist, the
1959        compiled code was correct. (When scanning forwards to check for the
1960        existence of the subpattern, it was treating the data ']' as terminating
1961        the class, so got the count wrong. When actually compiling, the reference
1962        was subsequently set up correctly.)
1963    
1964    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1965        it was being rejected as not supported by pcre_dfa_exec(), even though
1966        other assertions are supported. I have made pcre_dfa_exec() support
1967        (*FAIL).
1968    
1969    16. The implementation of 13c above involved the invention of a new opcode,
1970        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1971        cannot be changed at match time, I realized I could make a small
1972        improvement to matching performance by compiling OP_ALLANY instead of
1973        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1974        on the OP_ANY path.
1975    
1976    17. Compiling pcretest on Windows with readline support failed without the
1977        following two fixes: (1) Make the unistd.h include conditional on
1978        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1979    
1980    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1981        ncurses library to be included for pcretest when ReadLine support is
1982        requested, but also to allow for it to be overridden. This patch came from
1983        Daniel Bergström.
1984    
1985    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1986        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1987        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1988        spotting this.
1989    
1990    
1991  Version 7.6 28-Jan-08  Version 7.6 28-Jan-08

Legend:
Removed from v.327  
changed lines
  Added in v.1348

  ViewVC Help
Powered by ViewVC 1.1.5