/[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 1352 by ph10, Mon Jul 29 15:49:21 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    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
42        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43        matched by pcre_exec() without using JIT:
44    
45        (a) If the rest of the pattern did not match after a maximal run of
46            grapheme clusters, the code for backing up to try with fewer of them
47            did not always back up over a full grapheme when characters that do not
48            have the modifier quality were involved, e.g. Hangul syllables.
49    
50        (b) If the match point in a subject started with modifier character, and
51            there was no match, the code could incorrectly back up beyond the match
52            point, and potentially beyond the first character in the subject,
53            leading to a segfault or an incorrect match result.
54    
55    10. A conditional group with an assertion condition could lead to PCRE
56        recording an incorrect first data item for a match if no other firse data
57        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58        first data item, and therefore matched "ca" after "c" instead of at the
59        start.
60    
61    
62    Version 8.33 28-May-2013
63    --------------------------
64    
65    1.  Added 'U' to some constants that are compared to unsigned integers, to
66        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
67        variables that are added to signed variables, to ensure the result is
68        signed and can be negated.
69    
70    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
71        CMake config files.
72    
73    3.  Revise the creation of config.h.generic so that all boolean macros are
74        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
75        overriding via -D on the command line possible.
76    
77    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
78        to unsigned int is reported to make a quite noticeable speed difference in
79        a specific Windows environment. Testing on Linux did also appear to show
80        some benefit (and it is clearly not harmful). Also fixed the definition of
81        Xop which should be unsigned.
82    
83    5.  Related to (4), changing the definition of the intermediate variable cc
84        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
85        improvements.
86    
87    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
88        unnecessary spaces.
89    
90    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
91        and later.
92    
93    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
94    
95    9.  Optimizing fast_forward_start_bits in JIT.
96    
97    10. Adding support for callouts in JIT, and fixing some issues revealed
98        during this work. Namely:
99    
100        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
101    
102        (b) Minimum length was not checked before the matching is started.
103    
104    11. The value of capture_last that is passed to callouts was incorrect in some
105        cases when there was a capture on one path that was subsequently abandoned
106        after a backtrack. Also, the capture_last value is now reset after a
107        recursion, since all captures are also reset in this case.
108    
109    12. The interpreter no longer returns the "too many substrings" error in the
110        case when an overflowing capture is in a branch that is subsequently
111        abandoned after a backtrack.
112    
113    13. In the pathological case when an offset vector of size 2 is used, pcretest
114        now prints out the matched string after a yield of 0 or 1.
115    
116    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
117        Only supported by the JIT compiler at the moment.
118    
119    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
120    
121    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
122        offset of the starting point of the matching process, provided the offsets
123        vector is large enough.
124    
125    17. The \A escape now records a lookbehind value of 1, though its execution
126        does not actually inspect the previous character. This is to ensure that,
127        in partial multi-segment matching, at least one character from the old
128        segment is retained when a new segment is processed. Otherwise, if there
129        are no lookbehinds in the pattern, \A might match incorrectly at the start
130        of a new segment.
131    
132    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
133    
134    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
135        gives some modest performance improvement in 8-bit mode.
136    
137    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
138        be expressed in certain programming languages using Universal Character
139        Names.
140    
141    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
142        which points out that "non characters" are not "characters that may not
143        appear in Unicode strings" but rather "characters that are reserved for
144        internal use and have only local meaning".
145    
146    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
147        there was a conditional group that depended on an assertion, if the
148        assertion was false, the callout that immediately followed the alternation
149        in the condition was skipped when pcre_exec() was used for matching.
150    
151    23. Allow an explicit callout to be inserted before an assertion that is the
152        condition for a conditional group, for compatibility with automatic
153        callouts, which always insert a callout at this point.
154    
155    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
156        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
157    
158    25. (*PRUNE) is now supported by the JIT compiler.
159    
160    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
161    
162    27. Fix the case where there are two or more SKIPs with arguments that may be
163        ignored.
164    
165    28. (*SKIP) is now supported by the JIT compiler.
166    
167    29. (*THEN) is now supported by the JIT compiler.
168    
169    30. Update RunTest with additional test selector options.
170    
171    31. The way PCRE handles backtracking verbs has been changed in two ways.
172    
173        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
174        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
175        backtracking. In some cases this makes it more Perl-compatible, but Perl's
176        rather obscure rules do not always do the same thing.
177    
178        (2) Previously, backtracking verbs were confined within assertions. This is
179        no longer the case for positive assertions, except for (*ACCEPT). Again,
180        this sometimes improves Perl compatibility, and sometimes does not.
181    
182    32. A number of tests that were in test 2 because Perl did things differently
183        have been moved to test 1, because either Perl or PCRE has changed, and
184        these tests are now compatible.
185    
186    32. Backtracking control verbs are now handled in the same way in JIT and
187        interpreter.
188    
189    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
190        contained a forward subroutine reference caused a compile error.
191    
192    34. Auto-detect and optimize limited repetitions in JIT.
193    
194    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
195        blocking (*UTF) etc.
196    
197    36. In the interpreter, maximizing pattern repetitions for characters and
198        character types now use tail recursion, which reduces stack usage.
199    
200    37. The value of the max lookbehind was not correctly preserved if a compiled
201        and saved regex was reloaded on a host of different endianness.
202    
203    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
204        of the compiled pattern block, expand the flags field from 16 to 32 bits
205        because it was almost full.
206    
207    39. Try madvise first before posix_madvise.
208    
209    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
210        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
211        original change locked out all matches of empty strings. This has been
212        changed so that one match of an empty string per line is recognized.
213        Subsequent searches on the same line (for colouring or for --only-matching,
214        for example) do not recognize empty strings.
215    
216    41. Applied a user patch to fix a number of spelling mistakes in comments.
217    
218    42. Data lines longer than 65536 caused pcretest to crash.
219    
220    43. Clarified the data type for length and startoffset arguments for pcre_exec
221        and pcre_dfa_exec in the function-specific man pages, where they were
222        explicitly stated to be in bytes, never having been updated. I also added
223        some clarification to the pcreapi man page.
224    
225    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
226        a segmentation fault.
227    
228    
229    Version 8.32 30-November-2012
230    -----------------------------
231    
232    1.  Improved JIT compiler optimizations for first character search and single
233        character iterators.
234    
235    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
236        Patch by Daniel Richard G.
237    
238    3.  Single character iterator optimizations in the JIT compiler.
239    
240    4.  Improved JIT compiler optimizations for character ranges.
241    
242    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
243        Reported by Giuseppe D'Angelo.
244    
245    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
246        of a line, was being set incorrectly in cases where .* appeared inside
247        atomic brackets at the start of a pattern, or where there was a subsequent
248        *PRUNE or *SKIP.
249    
250    7.  Improved instruction cache flush for POWER/PowerPC.
251        Patch by Daniel Richard G.
252    
253    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
254        grep:
255    
256        (a) There is now no limit to the number of patterns to be matched.
257    
258        (b) An error is given if a pattern is too long.
259    
260        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
261            are now supported.
262    
263        (d) --exclude-from and --include-from (multiple use) have been added.
264    
265        (e) Exclusions and inclusions now apply to all files and directories, not
266            just to those obtained from scanning a directory recursively.
267    
268        (f) Multiple uses of -f and --file-list are now supported.
269    
270        (g) In a Windows environment, the default for -d has been changed from
271            "read" (the GNU grep default) to "skip", because otherwise the presence
272            of a directory in the file list provokes an error.
273    
274        (h) The documentation has been revised and clarified in places.
275    
276    9.  Improve the matching speed of capturing brackets.
277    
278    10. Changed the meaning of \X so that it now matches a Unicode extended
279        grapheme cluster.
280    
281    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
282        out POSIX threads when JIT support is configured.
283    
284    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
285    
286    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
287        structure could go wrong in environments where size_t is not the same size
288        as int.
289    
290    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
291        set.
292    
293    15. The EBCDIC support had decayed; later updates to the code had included
294        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
295        general tidy up of EBCDIC-related issues, and the documentation was also
296        not quite right. There is now a test that can be run on ASCII systems to
297        check some of the EBCDIC-related things (but is it not a full test).
298    
299    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
300        in a small tidy to the code.
301    
302    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
303    
304    18. If the --only-matching (-o) option in pcregrep is specified multiple
305        times, each one causes appropriate output. For example, -o1 -o2 outputs the
306        substrings matched by the 1st and 2nd capturing parentheses. A separating
307        string can be specified by --om-separator (default empty).
308    
309    19. Improving the first n character searches.
310    
311    20. Turn case lists for horizontal and vertical white space into macros so that
312        they are defined only once.
313    
314    21. This set of changes together give more compatible Unicode case-folding
315        behaviour for characters that have more than one other case when UCP
316        support is available.
317    
318        (a) The Unicode property table now has offsets into a new table of sets of
319            three or more characters that are case-equivalent. The MultiStage2.py
320            script that generates these tables (the pcre_ucd.c file) now scans
321            CaseFolding.txt instead of UnicodeData.txt for character case
322            information.
323    
324        (b) The code for adding characters or ranges of characters to a character
325            class has been abstracted into a generalized function that also handles
326            case-independence. In UTF-mode with UCP support, this uses the new data
327            to handle characters with more than one other case.
328    
329        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
330            whose other case is greater than 256 are now correctly matched
331            caselessly. Previously, the high codepoint matched the low one, but not
332            vice versa.
333    
334        (d) The processing of \h, \H, \v, and \ in character classes now makes use
335            of the new class addition function, using character lists defined as
336            macros alongside the case definitions of 20 above.
337    
338        (e) Caseless back references now work with characters that have more than
339            one other case.
340    
341        (f) General caseless matching of characters with more than one other case
342            is supported.
343    
344    22. Unicode character properties were updated from Unicode 6.2.0
345    
346    23. Improved CMake support under Windows. Patch by Daniel Richard G.
347    
348    24. Add support for 32-bit character strings, and UTF-32
349    
350    25. Major JIT compiler update (code refactoring and bugfixing).
351        Experimental Sparc 32 support is added.
352    
353    26. Applied a modified version of Daniel Richard G's patch to create
354        pcre.h.generic and config.h.generic by "make" instead of in the
355        PrepareRelease script.
356    
357    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
358        pcre_compile.c when checking for a zero character.
359    
360    28. Introducing a native interface for JIT. Through this interface, the compiled
361        machine code can be directly executed. The purpose of this interface is to
362        provide fast pattern matching, so several sanity checks are not performed.
363        However, feature tests are still performed. The new interface provides
364        1.4x speedup compared to the old one.
365    
366    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
367        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
368        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
369    
370    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
371        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
372        the "old" RFC 2279). Instead, it ended up passing a negative length to
373        pcre_exec().
374    
375    31. Add support for GCC's visibility feature to hide internal functions.
376    
377    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
378        "unknown -C option" after outputting 0 or 1.
379    
380    33. There is now support for generating a code coverage report for the test
381        suite in environments where gcc is the compiler and lcov is installed. This
382        is mainly for the benefit of the developers.
383    
384    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
385        unaddressable using valgrind annotations, allowing valgrind to detect
386        invalid memory accesses. This is mainly for the benefit of the developers.
387    
388    25. (*UTF) can now be used to start a pattern in any of the three libraries.
389    
390    26. Give configure error if --enable-cpp but no C++ compiler found.
391    
392    
393    Version 8.31 06-July-2012
394    -------------------------
395    
396    1.  Fixing a wrong JIT test case and some compiler warnings.
397    
398    2.  Removed a bashism from the RunTest script.
399    
400    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
401        to unsigned type, result still unsigned" that was given by an MS compiler
402        on encountering the code "-sizeof(xxx)".
403    
404    4.  Partial matching support is added to the JIT compiler.
405    
406    5.  Fixed several bugs concerned with partial matching of items that consist
407        of more than one character:
408    
409        (a) /^(..)\1/ did not partially match "aba" because checking references was
410            done on an "all or nothing" basis. This also applied to repeated
411            references.
412    
413        (b) \R did not give a hard partial match if \r was found at the end of the
414            subject.
415    
416        (c) \X did not give a hard partial match after matching one or more
417            characters at the end of the subject.
418    
419        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
420            a partial match for the string "\r".
421    
422        (e) When newline was set to CRLF, the metacharacter "." did not recognize
423            a partial match for a CR character at the end of the subject string.
424    
425    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
426        running pcretest, the text "(JIT)" added to the output whenever JIT is
427        actually used to run the match.
428    
429    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
430        or /S+[+] with a digit between 1 and 7.
431    
432    8.  OP_NOT now supports any UTF character not just single-byte ones.
433    
434    9.  (*MARK) control verb is now supported by the JIT compiler.
435    
436    10. The command "./RunTest list" lists the available tests without actually
437        running any of them. (Because I keep forgetting what they all are.)
438    
439    11. Add PCRE_INFO_MAXLOOKBEHIND.
440    
441    12. Applied a (slightly modified) user-supplied patch that improves performance
442        when the heap is used for recursion (compiled with --disable-stack-for-
443        recursion). Instead of malloc and free for each heap frame each time a
444        logical recursion happens, frames are retained on a chain and re-used where
445        possible. This sometimes gives as much as 30% improvement.
446    
447    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
448        call.
449    
450    14. As documented, (*COMMIT) is now confined to within a positive assertion.
451    
452    15. It is now possible to link pcretest with libedit as an alternative to
453        libreadline.
454    
455    16. (*COMMIT) control verb is now supported by the JIT compiler.
456    
457    17. The Unicode data tables have been updated to Unicode 6.1.0.
458    
459    18. Added --file-list option to pcregrep.
460    
461    19. Added binary file support to pcregrep, including the -a, --binary-files,
462        -I, and --text options.
463    
464    20. The madvise function is renamed for posix_madvise for QNX compatibility
465        reasons. Fixed by Giuseppe D'Angelo.
466    
467    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
468    
469    22. Changed the option for creating symbolic links for 16-bit man pages from
470        -s to -sf so that re-installing does not cause issues.
471    
472    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
473    
474    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
475        matches in certain environments (the workspace was not being correctly
476        retained). Also added to pcre_dfa_exec() a simple plausibility check on
477        some of the workspace data at the beginning of a restart.
478    
479    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
480        was not doing so when it should - probably a typo introduced by SVN 528
481        (change 8.10/14).
482    
483    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
484        \w+ when the character tables indicated that \x{c4} was a word character.
485        There were several related cases, all because the tests for doing a table
486        lookup were testing for characters less than 127 instead of 255.
487    
488    27. If a pattern contains capturing parentheses that are not used in a match,
489        their slots in the ovector are set to -1. For those that are higher than
490        any matched groups, this happens at the end of processing. In the case when
491        there were back references that the ovector was too small to contain
492        (causing temporary malloc'd memory to be used during matching), and the
493        highest capturing number was not used, memory off the end of the ovector
494        was incorrectly being set to -1. (It was using the size of the temporary
495        memory instead of the true size.)
496    
497    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
498        ovector size, it uses memory at the end of the block that it has got.
499    
500    29. Check for an overlong MARK name and give an error at compile time. The
501        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
502    
503    30. JIT compiler update.
504    
505    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
506        Rill for the patch.
507    
508    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
509    
510    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
511    
512    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
513        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
514        was enabled.
515    
516    35. Improve JIT code generation for greedy plus quantifier.
517    
518    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
519        1 to "aa" instead of to an empty string. The bug affected repeated groups
520        that could potentially match an empty string.
521    
522    37. Optimizing single character iterators in JIT.
523    
524    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
525        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
526        codepoints that are too big for the mode are faulted, and in a UTF mode,
527        disallowed codepoints are also faulted.
528    
529    39. If PCRE was compiled with UTF support, in three places in the DFA
530        matcher there was code that should only have been obeyed in UTF mode, but
531        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
532        processing when bytes with values greater than 127 were present. In 16-bit
533        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
534        both cases the values are those that cannot be the first data item in a UTF
535        character. The three items that might have provoked this were recursions,
536        possessively repeated groups, and atomic groups.
537    
538    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
539        and pcregrep, because some OS require shared objects to be explicitly
540        passed to ld, causing the link step to fail if they are not.
541    
542    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
543        mode, patterns that started with \h* or \R* might be incorrectly matched.
544    
545    
546    Version 8.30 04-February-2012
547    -----------------------------
548    
549    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
550        name is defined in ctype.h.
551    
552    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
553        only in quite long subpatterns.
554    
555    3.  Removed the function pcre_info(), which has been obsolete and deprecated
556        since it was replaced by pcre_fullinfo() in February 2000.
557    
558    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
559        match a (*MARK), and the match failed at the start of the subject, a
560        reference to memory before the start of the subject could occur. This bug
561        was introduced by fix 17 of release 8.21.
562    
563    5.  A reference to an unset group with zero minimum repetition was giving
564        totally wrong answers (in non-JavaScript-compatibility mode). For example,
565        /(another)?(\1?)test/ matched against "hello world test". This bug was
566        introduced in release 8.13.
567    
568    6.  Add support for 16-bit character strings (a large amount of work involving
569        many changes and refactorings).
570    
571    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
572        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
573        from a file.
574    
575    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
576        rounding is not applied in this particular case).
577    
578    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
579        if they appear, or are escaped, in patterns.
580    
581    10. Get rid of a number of -Wunused-but-set-variable warnings.
582    
583    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
584        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
585        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
586        also returns the mark "x". This bug applied to capturing parentheses,
587        non-capturing parentheses, and atomic parentheses. It also applied to some
588        assertions.
589    
590    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
591        information out of configure.ac instead of relying on pcre.h.generic, which
592        is not stored in the repository.
593    
594    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
595        -lreadline.
596    
597    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
598    
599    15. Applied Graycode's patch to put the top-level frame on the stack rather
600        than the heap when not using the stack for recursion. This gives a
601        performance improvement in many cases when recursion is not deep.
602    
603    16. Experimental code added to "pcretest -C" to output the stack frame size.
604    
605    
606    Version 8.21 12-Dec-2011
607    ------------------------
608    
609    1.  Updating the JIT compiler.
610    
611    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
612        are added as well.
613    
614    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
615        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
616        calling _pcre_jit_exec. Some extra comments are added.
617    
618    4.  (*MARK) settings inside atomic groups that do not contain any capturing
619        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
620        was introduced by change 18 for 8.20.
621    
622    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
623        ECMA-262 standard.
624    
625    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
626        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
627        This bug was probably introduced by change 9 of 8.13.
628    
629    7.  While fixing 6 above, I noticed that a number of other items were being
630        incorrectly rejected as "not fixed length". This arose partly because newer
631        opcodes had not been added to the fixed-length checking code. I have (a)
632        corrected the bug and added tests for these items, and (b) arranged for an
633        error to occur if an unknown opcode is encountered while checking for fixed
634        length instead of just assuming "not fixed length". The items that were
635        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
636        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
637        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
638    
639    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
640        being incorrectly compiled and would have given unpredicatble results.
641    
642    9.  A possessively repeated subpattern with minimum repeat count greater than
643        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
644        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
645        the first (A) could occur when it should not.
646    
647    10. Add a cast and remove a redundant test from the code.
648    
649    11. JIT should use pcre_malloc/pcre_free for allocation.
650    
651    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
652        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
653        is anything other than /usr, -L is still shown).
654    
655    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
656    
657    14. Perl does not support \N without a following name in a [] class; PCRE now
658        also gives an error.
659    
660    15. If a forward reference was repeated with an upper limit of around 2000,
661        it caused the error "internal error: overran compiling workspace". The
662        maximum number of forward references (including repeats) was limited by the
663        internal workspace, and dependent on the LINK_SIZE. The code has been
664        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
665        the default depends on LINK_SIZE. There is a new upper limit (for safety)
666        of around 200,000 forward references. While doing this, I also speeded up
667        the filling in of repeated forward references.
668    
669    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
670        incorrectly expecting the subject to contain another "a" after the start.
671    
672    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
673        in the match, the SKIP should be ignored. This was not happening; instead
674        the SKIP was being treated as NOMATCH. For patterns such as
675        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
676        tested.
677    
678    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
679        now much more compatible with Perl, in particular in cases where the result
680        is a non-match for a non-anchored pattern. For example, if
681        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
682        "m", where previously it did not return a name. A side effect of this
683        change is that for partial matches, the last encountered mark name is
684        returned, as for non matches. A number of tests that were previously not
685        Perl-compatible have been moved into the Perl-compatible test files. The
686        refactoring has had the pleasing side effect of removing one argument from
687        the match() function, thus reducing its stack requirements.
688    
689    19. If the /S+ option was used in pcretest to study a pattern using JIT,
690        subsequent uses of /S (without +) incorrectly behaved like /S+.
691    
692    21. Retrieve executable code size support for the JIT compiler and fixing
693        some warnings.
694    
695    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
696        not work when the shorter character appeared right at the end of the
697        subject string.
698    
699    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
700        systems.
701    
702    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
703        output it when the /M option is used in pcretest.
704    
705    25. The CheckMan script was not being included in the distribution. Also, added
706        an explicit "perl" to run Perl scripts from the PrepareRelease script
707        because this is reportedly needed in Windows.
708    
709    26. If study data was being save in a file and studying had not found a set of
710        "starts with" bytes for the pattern, the data written to the file (though
711        never used) was taken from uninitialized memory and so caused valgrind to
712        complain.
713    
714    27. Updated RunTest.bat as provided by Sheri Pierce.
715    
716    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
717    
718    29. Computation of memory usage for the table of capturing group names was
719        giving an unnecessarily large value.
720    
721    
722    Version 8.20 21-Oct-2011
723    ------------------------
724    
725    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
726        a POSIX class. After further experiments with Perl, which convinced me that
727        Perl has bugs and confusions, a closing square bracket is no longer allowed
728        in a POSIX name. This bug also affected patterns with classes that started
729        with full stops.
730    
731    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
732        captured substring, but while checking the failing first alternative,
733        substring 1 is temporarily captured. If the output vector supplied to
734        pcre_exec() was not big enough for this capture, the yield of the function
735        was still zero ("insufficient space for captured substrings"). This cannot
736        be totally fixed without adding another stack variable, which seems a lot
737        of expense for a edge case. However, I have improved the situation in cases
738        such as /(a)(b)x|abc/ matched against "abc", where the return code
739        indicates that fewer than the maximum number of slots in the ovector have
740        been set.
741    
742    3.  Related to (2) above: when there are more back references in a pattern than
743        slots in the output vector, pcre_exec() uses temporary memory during
744        matching, and copies in the captures as far as possible afterwards. It was
745        using the entire output vector, but this conflicts with the specification
746        that only 2/3 is used for passing back captured substrings. Now it uses
747        only the first 2/3, for compatibility. This is, of course, another edge
748        case.
749    
750    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
751        main code base, and can be used by building with --enable-jit. When this is
752        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
753        runtime --no-jit option is given.
754    
755    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
756        ovector, the return from the function was zero, implying that there were
757        other matches that did not fit. The correct "exactly full" value is now
758        returned.
759    
760    6.  If a subpattern that was called recursively or as a subroutine contained
761        (*PRUNE) or any other control that caused it to give a non-standard return,
762        invalid errors such as "Error -26 (nested recursion at the same subject
763        position)" or even infinite loops could occur.
764    
765    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
766        computing the minimum length on reaching *ACCEPT, and so ended up with the
767        wrong value of 1 rather than 0. Further investigation indicates that
768        computing a minimum subject length in the presence of *ACCEPT is difficult
769        (think back references, subroutine calls), and so I have changed the code
770        so that no minimum is registered for a pattern that contains *ACCEPT.
771    
772    8.  If (*THEN) was present in the first (true) branch of a conditional group,
773        it was not handled as intended. [But see 16 below.]
774    
775    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
776        Sheri Pierce.
777    
778    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
779        the first byte in a match must be "a".
780    
781    11. Change 17 for 8.13 increased the recursion depth for patterns like
782        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
783        pattern contains any instances of (*THEN). If it does not, the old
784        optimizations are restored. It would be nice to do this on a per-group
785        basis, but at the moment that is not feasible.
786    
787    12. In some environments, the output of pcretest -C is CRLF terminated. This
788        broke RunTest's code that checks for the link size. A single white space
789        character after the value is now allowed for.
790    
791    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
792        For "fr", it uses the Windows-specific input and output files.
793    
794    14. If (*THEN) appeared in a group that was called recursively or as a
795        subroutine, it did not work as intended. [But see next item.]
796    
797    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
798        pattern fragments (but not containing any | characters). If A and B are
799        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
800        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
801        D. In other words, Perl considers parentheses that do not contain any |
802        characters to be part of a surrounding alternative, whereas PCRE was
803        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
804        differently. PCRE now behaves in the same way as Perl, except in the case
805        of subroutine/recursion calls such as (?1) which have in any case always
806        been different (but PCRE had them first :-).
807    
808    16. Related to 15 above: Perl does not treat the | in a conditional group as
809        creating alternatives. Such a group is treated in the same way as an
810        ordinary group without any | characters when processing (*THEN). PCRE has
811        been changed to match Perl's behaviour.
812    
813    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
814        RunGrepTest script failed.
815    
816    18. Change 22 for version 13 caused atomic groups to use more stack. This is
817        inevitable for groups that contain captures, but it can lead to a lot of
818        stack use in large patterns. The old behaviour has been restored for atomic
819        groups that do not contain any capturing parentheses.
820    
821    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
822        suppress the check for a minimum subject length at run time. (If it was
823        given to pcre_exec() or pcre_dfa_exec() it did work.)
824    
825    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
826        fail to work when decoding hex characters in data strings in EBCDIC
827        environments.
828    
829    21. It appears that in at least one Mac OS environment, the isxdigit() function
830        is implemented as a macro that evaluates to its argument more than once,
831        contravening the C 90 Standard (I haven't checked a later standard). There
832        was an instance in pcretest which caused it to go wrong when processing
833        \x{...} escapes in subject strings. The has been rewritten to avoid using
834        things like p++ in the argument of isxdigit().
835    
836    
837    Version 8.13 16-Aug-2011
838    ------------------------
839    
840    1.  The Unicode data tables have been updated to Unicode 6.0.0.
841    
842    2.  Two minor typos in pcre_internal.h have been fixed.
843    
844    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
845        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
846        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
847    
848    4.  There were a number of related bugs in the code for matching backrefences
849        caselessly in UTF-8 mode when codes for the characters concerned were
850        different numbers of bytes. For example, U+023A and U+2C65 are an upper
851        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
852        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
853        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
854        2-byte code at the end of the subject (it thought there wasn't enough data
855        left).
856    
857    5.  Comprehensive information about what went wrong is now returned by
858        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
859        as the output vector has at least 2 elements. The offset of the start of
860        the failing character and a reason code are placed in the vector.
861    
862    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
863        now returned is for the first byte of the failing character, instead of the
864        last byte inspected. This is an incompatible change, but I hope it is small
865        enough not to be a problem. It makes the returned offset consistent with
866        pcre_exec() and pcre_dfa_exec().
867    
868    7.  pcretest now gives a text phrase as well as the error number when
869        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
870        failure, the offset and reason code are output.
871    
872    8.  When \R was used with a maximizing quantifier it failed to skip backwards
873        over a \r\n pair if the subsequent match failed. Instead, it just skipped
874        back over a single character (\n). This seems wrong (because it treated the
875        two characters as a single entity when going forwards), conflicts with the
876        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
877        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
878        has been changed.
879    
880    9.  Some internal refactoring has changed the processing so that the handling
881        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
882        time (the PCRE_DOTALL option was changed this way some time ago: version
883        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
884        which was always a bit of a fudge. It also means that there is one less
885        argument for the match() function, which reduces its stack requirements
886        slightly. This change also fixes an incompatibility with Perl: the pattern
887        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
888    
889    10. More internal refactoring has drastically reduced the number of recursive
890        calls to match() for possessively repeated groups such as (abc)++ when
891        using pcre_exec().
892    
893    11. While implementing 10, a number of bugs in the handling of groups were
894        discovered and fixed:
895    
896        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
897        (a|)*(?1) gave a compile-time internal error.
898        ((a|)+)+  did not notice that the outer group could match an empty string.
899        (^a|^)+   was not marked as anchored.
900        (.*a|.*)+ was not marked as matching at start or after a newline.
901    
902    12. Yet more internal refactoring has removed another argument from the match()
903        function. Special calls to this function are now indicated by setting a
904        value in a variable in the "match data" data block.
905    
906    13. Be more explicit in pcre_study() instead of relying on "default" for
907        opcodes that mean there is no starting character; this means that when new
908        ones are added and accidentally left out of pcre_study(), testing should
909        pick them up.
910    
911    14. The -s option of pcretest has been documented for ages as being an old
912        synonym of -m (show memory usage). I have changed it to mean "force study
913        for every regex", that is, assume /S for every regex. This is similar to -i
914        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
915        using it. It makes it easier to run collections of tests with and without
916        study enabled, and thereby test pcre_study() more easily. All the standard
917        tests are now run with and without -s (but some patterns can be marked as
918        "never study" - see 20 below).
919    
920    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
921        restoration of the capturing data to the outer values was not happening
922        correctly.
923    
924    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
925        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
926        pattern had matched an empty string, and so incorrectly returned a no
927        match.
928    
929    17. There was optimizing code for the last branch of non-capturing parentheses,
930        and also for the obeyed branch of a conditional subexpression, which used
931        tail recursion to cut down on stack usage. Unfortunately, now that there is
932        the possibility of (*THEN) occurring in these branches, tail recursion is
933        no longer possible because the return has to be checked for (*THEN). These
934        two optimizations have therefore been removed. [But see 8.20/11 above.]
935    
936    18. If a pattern containing \R was studied, it was assumed that \R always
937        matched two bytes, thus causing the minimum subject length to be
938        incorrectly computed because \R can also match just one byte.
939    
940    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
941        was incorrectly computed.
942    
943    20. If /S is present twice on a test pattern in pcretest input, it now
944        *disables* studying, thereby overriding the use of -s on the command line
945        (see 14 above). This is necessary for one or two tests to keep the output
946        identical in both cases.
947    
948    21. When (*ACCEPT) was used in an assertion that matched an empty string and
949        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
950    
951    22. When an atomic group that contained a capturing parenthesis was
952        successfully matched, but the branch in which it appeared failed, the
953        capturing was not being forgotten if a higher numbered group was later
954        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
955        group 1 to "a", when in fact it should be unset. This applied to multi-
956        branched capturing and non-capturing groups, repeated or not, and also to
957        positive assertions (capturing in negative assertions does not happen
958        in PCRE) and also to nested atomic groups.
959    
960    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
961        subject after a captured substring, to make it easier to tell which of a
962        number of identical substrings has been captured.
963    
964    24. The way atomic groups are processed by pcre_exec() has been changed so that
965        if they are repeated, backtracking one repetition now resets captured
966        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
967        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
968        "aaa". Previously, it would have been "a". As part of this code
969        refactoring, the way recursive calls are handled has also been changed.
970    
971    25. If an assertion condition captured any substrings, they were not passed
972        back unless some other capturing happened later. For example, if
973        (?(?=(a))a) was matched against "a", no capturing was returned.
974    
975    26. When studying a pattern that contained subroutine calls or assertions,
976        the code for finding the minimum length of a possible match was handling
977        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
978        group 1 called group 2 while simultaneously a separate group 2 called group
979        1). A stack overflow occurred in this case. I have fixed this by limiting
980        the recursion depth to 10.
981    
982    27. Updated RunTest.bat in the distribution to the version supplied by Tom
983        Fortmann. This supports explicit test numbers on the command line, and has
984        argument validation and error reporting.
985    
986    28. An instance of \X with an unlimited repeat could fail if at any point the
987        first character it looked at was a mark character.
988    
989    29. Some minor code refactoring concerning Unicode properties and scripts
990        should reduce the stack requirement of match() slightly.
991    
992    30. Added the '=' option to pcretest to check the setting of unused capturing
993        slots at the end of the pattern, which are documented as being -1, but are
994        not included in the return count.
995    
996    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
997        compiled something random. Now it gives a compile-time error (as does
998        Perl).
999    
1000    32. A *MARK encountered during the processing of a positive assertion is now
1001        recorded and passed back (compatible with Perl).
1002    
1003    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1004        had alternative anchored branches, the search for a second match in a line
1005        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1006        matched the line "0102" twice. The same bug affected patterns that started
1007        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1008        twice.
1009    
1010    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1011        does, and because of capturing effects, quantifying parenthesized
1012        assertions may at times be useful. Quantifiers are now allowed for
1013        parenthesized assertions.
1014    
1015    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1016    
1017    36. \g was being checked for fancy things in a character class, when it should
1018        just be a literal "g".
1019    
1020    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1021        appearance of a nested POSIX class supersedes an apparent external class.
1022        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1023        unescaped square brackets may also appear as part of class names. For
1024        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1025        more like Perl. (But see 8.20/1 above.)
1026    
1027    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1028        was because it thought it was \N{name}, which is not supported).
1029    
1030    39. Add minix to OS list not supporting the -S option in pcretest.
1031    
1032    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1033        cannot analyze patterns in sufficient detail to catch mutual recursions
1034        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1035        subgroup is called recursively as a subpattern for a second time at the
1036        same position in the subject string. In previous releases this might have
1037        been caught by the recursion limit, or it might have run out of stack.
1038    
1039    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1040        happen only once. PCRE was, however incorrectly giving a compile time error
1041        "recursive call could loop indefinitely" because it cannot analyze the
1042        pattern in sufficient detail. The compile time test no longer happens when
1043        PCRE is compiling a conditional subpattern, but actual runaway loops are
1044        now caught at runtime (see 40 above).
1045    
1046    42. It seems that Perl allows any characters other than a closing parenthesis
1047        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1048        has been changed to be the same.
1049    
1050    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1051        as not to get warnings when autogen.sh is called. Also changed
1052        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1053    
1054    44. To help people who use pcregrep to scan files containing exceedingly long
1055        lines, the following changes have been made:
1056    
1057        (a) The default value of the buffer size parameter has been increased from
1058            8K to 20K. (The actual buffer used is three times this size.)
1059    
1060        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1061            PCRE is built.
1062    
1063        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1064            to be set at run time.
1065    
1066        (d) Numerical values in pcregrep options can be followed by K or M, for
1067            example --buffer-size=50K.
1068    
1069        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1070            given and the return code is set to 2.
1071    
1072    45. Add a pointer to the latest mark to the callout data block.
1073    
1074    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1075        partial match of an empty string instead of no match. This was specific to
1076        the use of ".".
1077    
1078    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1079        complete match instead of a partial match. This bug was dependent on both
1080        the PCRE_UTF8 and PCRE_DOTALL options being set.
1081    
1082    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1083        starting byte set, because \b was not being ignored.
1084    
1085    
1086    Version 8.12 15-Jan-2011
1087    ------------------------
1088    
1089    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1090        checks for such things as part of the documentation building process.
1091    
1092    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1093        --match-limit and --recursion-limit options (added for 8.11). In
1094        particular, this made one of the standard tests fail. (The integer value
1095        went into the wrong half of a long int.)
1096    
1097    3.  If the --colour option was given to pcregrep with -v (invert match), it
1098        did strange things, either producing crazy output, or crashing. It should,
1099        of course, ignore a request for colour when reporting lines that do not
1100        match.
1101    
1102    4.  Another pcregrep bug caused similar problems if --colour was specified with
1103        -M (multiline) and the pattern match finished with a line ending.
1104    
1105    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1106        matched in multiline mode, the following line was shown as part of the
1107        match. This seems wrong, so I have changed it.
1108    
1109    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1110        the check for further matches in the same line (so they could be coloured)
1111        to overrun the end of the current line. If another match was found, it was
1112        incorrectly shown (and then shown again when found in the next line).
1113    
1114    7.  If pcregrep was compiled under Windows, there was a reference to the
1115        function pcregrep_exit() before it was defined. I am assuming this was
1116        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1117        reported by a user. I've moved the definition above the reference.
1118    
1119    
1120    Version 8.11 10-Dec-2010
1121    ------------------------
1122    
1123    1.  (*THEN) was not working properly if there were untried alternatives prior
1124        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1125        backtracked to try for "b" instead of moving to the next alternative branch
1126        at the same level (in this case, to look for "c"). The Perl documentation
1127        is clear that when (*THEN) is backtracked onto, it goes to the "next
1128        alternative in the innermost enclosing group".
1129    
1130    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1131        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1132        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1133        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1134        (*THEN).
1135    
1136    3.  If \s appeared in a character class, it removed the VT character from
1137        the class, even if it had been included by some previous item, for example
1138        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1139        of \s, but is part of the POSIX "space" class.)
1140    
1141    4.  A partial match never returns an empty string (because you can always
1142        match an empty string at the end of the subject); however the checking for
1143        an empty string was starting at the "start of match" point. This has been
1144        changed to the "earliest inspected character" point, because the returned
1145        data for a partial match starts at this character. This means that, for
1146        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1147        (previously it gave "no match").
1148    
1149    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1150        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1151        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1152        has an implication that the given string is incomplete (because a partial
1153        match is preferred over a full match). For this reason, these items now
1154        give a partial match in this situation. [Aside: previously, the one case
1155        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1156        match rather than a full match, which was wrong by the old rules, but is
1157        now correct.]
1158    
1159    6.  There was a bug in the handling of #-introduced comments, recognized when
1160        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1161        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1162        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1163        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1164        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1165        places in pcre_compile().
1166    
1167    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1168        comments when looking ahead for named forward references to subpatterns,
1169        the only newline sequence it recognized was NL. It now handles newlines
1170        according to the set newline convention.
1171    
1172    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1173        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1174        cater for a lack of strerror(). These oversights have been fixed.
1175    
1176    9.  Added --match-limit and --recursion-limit to pcregrep.
1177    
1178    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1179    
1180    11. When the -o option was used, pcregrep was setting a return code of 1, even
1181        when matches were found, and --line-buffered was not being honoured.
1182    
1183    12. Added an optional parentheses number to the -o and --only-matching options
1184        of pcregrep.
1185    
1186    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1187        can match an empty string. The code to do it in pcretest and pcredemo
1188        needed fixing:
1189    
1190        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1191            only one byte after an empty string match just before CRLF (this case
1192            just got forgotten; "any" and "anycrlf" were OK).
1193    
1194        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1195            mode when an empty string match preceded an ASCII character followed by
1196            a non-ASCII character. (The code for advancing by one character rather
1197            than one byte was nonsense.)
1198    
1199        (c) The pcredemo.c sample program did not have any code at all to handle
1200            the cases when CRLF is a valid newline sequence.
1201    
1202    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1203        as a starting offset was within the subject string. There is now a new
1204        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1205        negative or greater than the length of the string. In order to test this,
1206        pcretest is extended to allow the setting of negative starting offsets.
1207    
1208    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1209        starting offset points to the beginning of a UTF-8 character was
1210        unnecessarily clumsy. I tidied it up.
1211    
1212    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1213        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1214    
1215    17. Nobody had reported that the --include_dir option, which was added in
1216        release 7.7 should have been called --include-dir (hyphen, not underscore)
1217        for compatibility with GNU grep. I have changed it to --include-dir, but
1218        left --include_dir as an undocumented synonym, and the same for
1219        --exclude-dir, though that is not available in GNU grep, at least as of
1220        release 2.5.4.
1221    
1222    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1223        characters from a string of bytes) have been redefined so as not to use
1224        loops, in order to improve performance in some environments. At the same
1225        time, I abstracted some of the common code into auxiliary macros to save
1226        repetition (this should not affect the compiled code).
1227    
1228    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1229        compile-time error is now given if \c is not followed by an ASCII
1230        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1231        different, and any byte value is allowed.)
1232    
1233    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1234        START_OPTIMIZE option, which is now allowed at compile time - but just
1235        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1236        to pcregrep and other applications that have no direct access to PCRE
1237        options. The new /Y option in pcretest sets this option when calling
1238        pcre_compile().
1239    
1240    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1241        back references. Groups containing recursive back references were forced to
1242        be atomic by that change, but in the case of named groups, the amount of
1243        memory required was incorrectly computed, leading to "Failed: internal
1244        error: code overflow". This has been fixed.
1245    
1246    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1247        pcretest.c, to avoid build problems in some Borland environments.
1248    
1249    
1250    Version 8.10 25-Jun-2010
1251    ------------------------
1252    
1253    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1254        THEN.
1255    
1256    2.  (*ACCEPT) was not working when inside an atomic group.
1257    
1258    3.  Inside a character class, \B is treated as a literal by default, but
1259        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1260        causes the error). The code is unchanged, but I tidied the documentation.
1261    
1262    4.  Inside a character class, PCRE always treated \R and \X as literals,
1263        whereas Perl faults them if its -w option is set. I have changed PCRE so
1264        that it faults them when PCRE_EXTRA is set.
1265    
1266    5.  Added support for \N, which always matches any character other than
1267        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1268    
1269    6.  When compiling pcregrep with newer versions of gcc which may have
1270        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1271        declared with attribute warn_unused_result" were given. Just casting the
1272        result to (void) does not stop the warnings; a more elaborate fudge is
1273        needed. I've used a macro to implement this.
1274    
1275    7.  Minor change to pcretest.c to avoid a compiler warning.
1276    
1277    8.  Added four artifical Unicode properties to help with an option to make
1278        \s etc use properties (see next item). The new properties are: Xan
1279        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1280    
1281    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1282        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1283        this option. Modified pcretest to add /W to test this facility. Added
1284        REG_UCP to make it available via the POSIX interface.
1285    
1286    10. Added --line-buffered to pcregrep.
1287    
1288    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1289        studied, and the match started with a letter with a code point greater than
1290        127 whose first byte was different to the first byte of the other case of
1291        the letter, the other case of this starting letter was not recognized
1292        (#976).
1293    
1294    12. If a pattern that was studied started with a repeated Unicode property
1295        test, for example, \p{Nd}+, there was the theoretical possibility of
1296        setting up an incorrect bitmap of starting bytes, but fortunately it could
1297        not have actually happened in practice until change 8 above was made (it
1298        added property types that matched character-matching opcodes).
1299    
1300    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1301        possible starting bytes for non-anchored patterns.
1302    
1303    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1304        \R, and also a number of cases that involve Unicode properties, both
1305        explicit and implicit when PCRE_UCP is set.
1306    
1307    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1308        input, it could crash or give wrong results if characters with values
1309        greater than 0xc0 were present in the subject string. (Detail: it assumed
1310        UTF-8 input when processing these items.)
1311    
1312    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1313        size_t is 64-bit (#991).
1314    
1315    17. Added a check for running out of memory when PCRE is compiled with
1316        --disable-stack-for-recursion (#990).
1317    
1318    18. If the last data line in a file for pcretest does not have a newline on
1319        the end, a newline was missing in the output.
1320    
1321    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1322        less than 128) in its various bitmaps. However, there is a facility for
1323        generating tables according to the current locale when PCRE is compiled. It
1324        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1325        characters, are recognized by isspace() and therefore were getting set in
1326        these tables, and indeed these tables seem to approximate to ISO 8859. This
1327        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1328        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1329        which of course cannot start UTF-8 characters. I have changed the code so
1330        that only real ASCII characters (less than 128) and the correct starting
1331        bytes for UTF-8 encodings are set for characters greater than 127 when in
1332        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1333        altogether.)
1334    
1335    20. Added the /T option to pcretest so as to be able to run tests with non-
1336        standard character tables, thus making it possible to include the tests
1337        used for 19 above in the standard set of tests.
1338    
1339    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1340        reference to a subpattern the other side of a comment that contains an
1341        opening parenthesis caused either an internal compiling error, or a
1342        reference to the wrong subpattern.
1343    
1344    
1345    Version 8.02 19-Mar-2010
1346    ------------------------
1347    
1348    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1349    
1350    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1351        configured.
1352    
1353    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1354        original author of that file, following a query about its status.
1355    
1356    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1357        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1358    
1359    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1360        quantifier applied to a forward-referencing subroutine call, could compile
1361        incorrect code or give the error "internal error: previously-checked
1362        referenced subpattern not found".
1363    
1364    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1365        variables to point to external functions. For these systems, therefore,
1366        pcre_malloc etc. are now initialized to local functions that call the
1367        relevant global functions.
1368    
1369    7.  There were two entries missing in the vectors called coptable and poptable
1370        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1371        I've fixed the data, and added a kludgy way of testing at compile time that
1372        the lengths are correct (equal to the number of opcodes).
1373    
1374    8.  Following on from 7, I added a similar kludge to check the length of the
1375        eint vector in pcreposix.c.
1376    
1377    9.  Error texts for pcre_compile() are held as one long string to avoid too
1378        much relocation at load time. To find a text, the string is searched,
1379        counting zeros. There was no check for running off the end of the string,
1380        which could happen if a new error number was added without updating the
1381        string.
1382    
1383    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1384    
1385    11. \K was not working if it appeared in an atomic group or in a group that
1386        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1387        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1388        the assertion is positive, but not if it is negative.
1389    
1390    12. Change 11 fortuitously reduced the size of the stack frame used in the
1391        "match()" function of pcre_exec.c by one pointer. Forthcoming
1392        implementation of support for (*MARK) will need an extra pointer on the
1393        stack; I have reserved it now, so that the stack frame size does not
1394        decrease.
1395    
1396    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1397        item in branch that calls a recursion is a subroutine call - as in the
1398        second branch in the above example - was incorrectly given the compile-
1399        time error "recursive call could loop indefinitely" because pcre_compile()
1400        was not correctly checking the subroutine for matching a non-empty string.
1401    
1402    14. The checks for overrunning compiling workspace could trigger after an
1403        overrun had occurred. This is a "should never occur" error, but it can be
1404        triggered by pathological patterns such as hundreds of nested parentheses.
1405        The checks now trigger 100 bytes before the end of the workspace.
1406    
1407    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1408    
1409    
1410    Version 8.01 19-Jan-2010
1411    ------------------------
1412    
1413    1.  If a pattern contained a conditional subpattern with only one branch (in
1414        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1415        computed the wrong minimum data length (which is of course zero for such
1416        subpatterns). This could cause incorrect "no match" results.
1417    
1418    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1419        the pattern is reset in the first branch, pcre_compile() failed with
1420        "internal error: code overflow at offset...". This happened only when
1421        the reset was to the original external option setting. (An optimization
1422        abstracts leading options settings into an external setting, which was the
1423        cause of this.)
1424    
1425    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1426        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1427        assertion pattern did not match (meaning that the assertion was true), it
1428        was incorrectly treated as false if the SKIP had been reached during the
1429        matching. This also applied to assertions used as conditions.
1430    
1431    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1432        assertion subpattern, including such a pattern used as a condition,
1433        unpredictable results occurred, instead of the error return
1434        PCRE_ERROR_DFA_UITEM.
1435    
1436    5.  The C++ GlobalReplace function was not working like Perl for the special
1437        situation when an empty string is matched. It now does the fancy magic
1438        stuff that is necessary.
1439    
1440    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1441        removed. (These were left over from very, very early versions of PCRE.)
1442    
1443    7.  Some cosmetic changes to the code to make life easier when compiling it
1444        as part of something else:
1445    
1446        (a) Change DEBUG to PCRE_DEBUG.
1447    
1448        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1449            called "current" as "current_branch", to prevent a collision with the
1450            Linux macro when compiled as a kernel module.
1451    
1452        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1453            prevent a collision with the Linux macro when compiled as a kernel
1454            module.
1455    
1456    8.  In pcre_compile() there are some checks for integer overflows that used to
1457        cast potentially large values to (double). This has been changed to that
1458        when building, a check for int64_t is made, and if it is found, it is used
1459        instead, thus avoiding the use of floating point arithmetic. (There is no
1460        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1461        double.
1462    
1463    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1464        2005 (difference between two addresses compared to an unsigned value).
1465    
1466    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1467        custom one, because of the following reported problem in Windows:
1468    
1469          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1470              under Win32.
1471          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1472              therefore missing the function definition.
1473          - The compiler thus generates a "C" signature for the test function.
1474          - The linker fails to find the "C" function.
1475          - PCRE fails to configure if asked to do so against libbz2.
1476    
1477    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1478        messages were output:
1479    
1480          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1481          rerunning libtoolize, to keep the correct libtool macros in-tree.
1482          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1483    
1484        I have done both of these things.
1485    
1486    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1487        most of the time, it *can* run out if it is given a pattern that contains a
1488        runaway infinite recursion. I updated the discussion in the pcrestack man
1489        page.
1490    
1491    13. Now that we have gone to the x.xx style of version numbers, the minor
1492        version may start with zero. Using 08 or 09 is a bad idea because users
1493        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1494        interpreted as invalid octal numbers. I've updated the previous comment in
1495        configure.ac, and also added a check that gives an error if 08 or 09 are
1496        used.
1497    
1498    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1499        causing partial matching to fail when the end of the subject matched \W
1500        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1501    
1502    15. There were some discrepancies between the declarations in pcre_internal.h
1503        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1504        their definitions. The declarations used "const uschar *" and the
1505        definitions used USPTR. Even though USPTR is normally defined as "const
1506        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1507        reported that: "This difference in casting confuses some C++ compilers, for
1508        example, SunCC recognizes above declarations as different functions and
1509        generates broken code for hbpcre." I have changed the declarations to use
1510        USPTR.
1511    
1512    16. GNU libtool is named differently on some systems. The autogen.sh script now
1513        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1514        (FreeBSD).
1515    
1516    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1517        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1518        comment: "Figure out how to create a longlong from a string: strtoll and
1519        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1520        instance, but it only takes 2 args instead of 3!"
1521    
1522    18. A subtle bug concerned with back references has been fixed by a change of
1523        specification, with a corresponding code fix. A pattern such as
1524        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1525        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1526        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1527        same bug. Such groups have to be quantified to be useful, or contained
1528        inside another quantified group. (If there's no repetition, the reference
1529        can never match.) The problem arises because, having left the group and
1530        moved on to the rest of the pattern, a later failure that backtracks into
1531        the group uses the captured value from the final iteration of the group
1532        rather than the correct earlier one. I have fixed this in PCRE by forcing
1533        any group that contains a reference to itself to be an atomic group; that
1534        is, there cannot be any backtracking into it once it has completed. This is
1535        similar to recursive and subroutine calls.
1536    
1537    
1538    Version 8.00 19-Oct-09
1539    ----------------------
1540    
1541    1.  The table for translating pcre_compile() error codes into POSIX error codes
1542        was out-of-date, and there was no check on the pcre_compile() error code
1543        being within the table. This could lead to an OK return being given in
1544        error.
1545    
1546    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1547        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1548        in a Windows environment.
1549    
1550    3.  The pcregrep --count option prints the count for each file even when it is
1551        zero, as does GNU grep. However, pcregrep was also printing all files when
1552        --files-with-matches was added. Now, when both options are given, it prints
1553        counts only for those files that have at least one match. (GNU grep just
1554        prints the file name in this circumstance, but including the count seems
1555        more useful - otherwise, why use --count?) Also ensured that the
1556        combination -clh just lists non-zero counts, with no names.
1557    
1558    4.  The long form of the pcregrep -F option was incorrectly implemented as
1559        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1560        but it seems right to fix it, and I didn't think it was worth preserving
1561        the old behaviour.
1562    
1563    5.  The command line items --regex=pattern and --regexp=pattern were not
1564        recognized by pcregrep, which required --regex pattern or --regexp pattern
1565        (with a space rather than an '='). The man page documented the '=' forms,
1566        which are compatible with GNU grep; these now work.
1567    
1568    6.  No libpcreposix.pc file was created for pkg-config; there was just
1569        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1570    
1571    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1572        when UCP support is not needed, by modifying the Python script that
1573        generates it from Unicode data files. This should not matter if the module
1574        is correctly used as a library, but I received one complaint about 50K of
1575        unwanted data. My guess is that the person linked everything into his
1576        program rather than using a library. Anyway, it does no harm.
1577    
1578    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1579        was a minimum greater than 1 for a wide character in a possessive
1580        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1581        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1582        character. Chaos in the form of incorrect output or a compiling loop could
1583        result.
1584    
1585    9.  The restrictions on what a pattern can contain when partial matching is
1586        requested for pcre_exec() have been removed. All patterns can now be
1587        partially matched by this function. In addition, if there are at least two
1588        slots in the offset vector, the offset of the earliest inspected character
1589        for the match and the offset of the end of the subject are set in them when
1590        PCRE_ERROR_PARTIAL is returned.
1591    
1592    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1593        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1594        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1595        and may be more useful for multi-segment matching.
1596    
1597    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1598        used to be given if ever the end of the subject was reached; now it is
1599        given only if matching could not proceed because another character was
1600        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1601        string "Z", which now yields "no match" instead of "partial match". In the
1602        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1603        final character ended with (*FAIL).
1604    
1605    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1606        if the pattern had a "must contain" character that was already found in the
1607        earlier partial match, unless partial matching was again requested. For
1608        example, with the pattern /dog.(body)?/, the "must contain" character is
1609        "g". If the first part-match was for the string "dog", restarting with
1610        "sbody" failed. This bug has been fixed.
1611    
1612    13. The string returned by pcre_dfa_exec() after a partial match has been
1613        changed so that it starts at the first inspected character rather than the
1614        first character of the match. This makes a difference only if the pattern
1615        starts with a lookbehind assertion or \b or \B (\K is not supported by
1616        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1617        matching functions compatible, and I think it's the right thing to do.
1618    
1619    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1620        so that the demonstration program is easily available in environments where
1621        PCRE has not been installed from source.
1622    
1623    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1624        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1625        library.
1626    
1627    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1628        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1629        is not the first non-POSIX option to be added. Clearly some people find
1630        these options useful.
1631    
1632    17. If a caller to the POSIX matching function regexec() passes a non-zero
1633        value for nmatch with a NULL value for pmatch, the value of
1634        nmatch is forced to zero.
1635    
1636    18. RunGrepTest did not have a test for the availability of the -u option of
1637        the diff command, as RunTest does. It now checks in the same way as
1638        RunTest, and also checks for the -b option.
1639    
1640    19. If an odd number of negated classes containing just a single character
1641        interposed, within parentheses, between a forward reference to a named
1642        subpattern and the definition of the subpattern, compilation crashed with
1643        an internal error, complaining that it could not find the referenced
1644        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1645        [The bug was that it was starting one character too far in when skipping
1646        over the character class, thus treating the ] as data rather than
1647        terminating the class. This meant it could skip too much.]
1648    
1649    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1650        /g option in pcretest when the pattern contains \K, which makes it possible
1651        to have an empty string match not at the start, even when the pattern is
1652        anchored. Updated pcretest and pcredemo to use this option.
1653    
1654    21. If the maximum number of capturing subpatterns in a recursion was greater
1655        than the maximum at the outer level, the higher number was returned, but
1656        with unset values at the outer level. The correct (outer level) value is
1657        now given.
1658    
1659    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1660        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1661        make it do so. The string so far is captured, making this feature
1662        compatible with Perl.
1663    
1664    23. The tests have been re-organized, adding tests 11 and 12, to make it
1665        possible to check the Perl 5.10 features against Perl 5.10.
1666    
1667    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1668        pattern matches a fixed length string. PCRE did not allow this; now it
1669        does. Neither allows recursion.
1670    
1671    25. I finally figured out how to implement a request to provide the minimum
1672        length of subject string that was needed in order to match a given pattern.
1673        (It was back references and recursion that I had previously got hung up
1674        on.) This code has now been added to pcre_study(); it finds a lower bound
1675        to the length of subject needed. It is not necessarily the greatest lower
1676        bound, but using it to avoid searching strings that are too short does give
1677        some useful speed-ups. The value is available to calling programs via
1678        pcre_fullinfo().
1679    
1680    26. While implementing 25, I discovered to my embarrassment that pcretest had
1681        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1682        study optimizations had never been tested with that matching function.
1683        Oops. What is worse, even when it was passed study data, there was a bug in
1684        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1685        were also very few tests of studied patterns with pcre_dfa_exec().
1686    
1687    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1688        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1689        on the other side of the coin, they are no longer allowed to have different
1690        names, because these cannot be distinguished in PCRE, and this has caused
1691        confusion. (This is a difference from Perl.)
1692    
1693    28. When duplicate subpattern names are present (necessarily with different
1694        numbers, as required by 27 above), and a test is made by name in a
1695        conditional pattern, either for a subpattern having been matched, or for
1696        recursion in such a pattern, all the associated numbered subpatterns are
1697        tested, and the overall condition is true if the condition is true for any
1698        one of them. This is the way Perl works, and is also more like the way
1699        testing by number works.
1700    
1701    
1702    Version 7.9 11-Apr-09
1703    ---------------------
1704    
1705    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1706        (pcretest), all targets were linked against these libraries. This included
1707        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1708        libraries. This caused unwanted dependencies to be created. This problem
1709        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1710        pcretest is linked with readline.
1711    
1712    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1713        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1714        moved outside it again, because FALSE and TRUE are already defined in AIX,
1715        but BOOL is not.
1716    
1717    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1718        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1719    
1720    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1721        hyphens) following file names and line numbers when outputting matching
1722        lines. This is not true; no spaces are inserted. I have also clarified the
1723        wording for the --colour (or --color) option.
1724    
1725    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1726        was not coloured; this is different to GNU grep, so I have changed it to be
1727        the same.
1728    
1729    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1730        each matching line was coloured. Now it goes on to look for further matches
1731        of any of the test patterns, which is the same behaviour as GNU grep.
1732    
1733    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1734        doesn't make sense to accept an empty string match in pcregrep, so I have
1735        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1736        seems to be how GNU grep behaves. [But see later change 40 for release
1737        8.33.]
1738    
1739    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1740        start or after a newline", because the conditional assertion was not being
1741        correctly handled. The rule now is that both the assertion and what follows
1742        in the first alternative must satisfy the test.
1743    
1744    9.  If auto-callout was enabled in a pattern with a conditional group whose
1745        condition was an assertion, PCRE could crash during matching, both with
1746        pcre_exec() and pcre_dfa_exec().
1747    
1748    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1749        used for matching.
1750    
1751    11. Unicode property support in character classes was not working for
1752        characters (bytes) greater than 127 when not in UTF-8 mode.
1753    
1754    12. Added the -M command line option to pcretest.
1755    
1756    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1757    
1758    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1759    
1760    16. Added comments and documentation about mis-use of no_arg in the C++
1761        wrapper.
1762    
1763    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1764        from Martin Jerabek that uses macro names for all relevant character and
1765        string constants.
1766    
1767    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1768        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1769        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1770        these, but not everybody uses configure.
1771    
1772    19. A conditional group that had only one branch was not being correctly
1773        recognized as an item that could match an empty string. This meant that an
1774        enclosing group might also not be so recognized, causing infinite looping
1775        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1776        with the subject "ab", where knowledge that the repeated group can match
1777        nothing is needed in order to break the loop.
1778    
1779    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1780        exec(), but without supplying a callout function, matching went wrong.
1781    
1782    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1783        leak if the size of the offset vector was greater than 30. When the vector
1784        is smaller, the saved offsets during recursion go onto a local stack
1785        vector, but for larger vectors malloc() is used. It was failing to free
1786        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1787        error, in fact).
1788    
1789    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1790        heapframe that is used only when UTF-8 support is enabled. This caused no
1791        problem, but was untidy.
1792    
1793    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1794        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1795        included within another project.
1796    
1797    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1798        slightly modified by me:
1799    
1800          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1801              not building pcregrep.
1802    
1803          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1804              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1805    
1806    25. Forward references, both numeric and by name, in patterns that made use of
1807        duplicate group numbers, could behave incorrectly or give incorrect errors,
1808        because when scanning forward to find the reference group, PCRE was not
1809        taking into account the duplicate group numbers. A pattern such as
1810        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1811    
1812    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1813        the feature of a custom pointer more persuasive (as requested by a user).
1814    
1815    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1816        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1817        pre-defined.
1818    
1819    28. Added support for (*UTF8) at the start of a pattern.
1820    
1821    29. Arrange for flags added by the "release type" setting in CMake to be shown
1822        in the configuration summary.
1823    
1824    
1825    Version 7.8 05-Sep-08
1826    ---------------------
1827    
1828    1.  Replaced UCP searching code with optimized version as implemented for Ad
1829        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1830        stage table and inline lookup instead of a function, giving speed ups of 2
1831        to 5 times on some simple patterns that I tested. Permission was given to
1832        distribute the MultiStage2.py script that generates the tables (it's not in
1833        the tarball, but is in the Subversion repository).
1834    
1835    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1836        scripts.
1837    
1838    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1839        a group with a zero qualifier. The result of the study could be incorrect,
1840        or the function might crash, depending on the pattern.
1841    
1842    4.  Caseless matching was not working for non-ASCII characters in back
1843        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1844        It now works when Unicode Property Support is available.
1845    
1846    5.  In pcretest, an escape such as \x{de} in the data was always generating
1847        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1848        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1849        truncation.
1850    
1851    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1852    
1853    7.  Added two (int) casts to pcregrep when printing the difference of two
1854        pointers, in case they are 64-bit values.
1855    
1856    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1857        test 2 if it fails.
1858    
1859    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1860        and a #define of that name to empty if it is not externally set. This is to
1861        allow users of MSVC to set it if necessary.
1862    
1863    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1864        the convenience functions in the pcre_get.c source file.
1865    
1866    11. An option change at the start of a pattern that had top-level alternatives
1867        could cause overwriting and/or a crash. This command provoked a crash in
1868        some environments:
1869    
1870          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1871    
1872        This potential security problem was recorded as CVE-2008-2371.
1873    
1874    12. For a pattern where the match had to start at the beginning or immediately
1875        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1876        pcre_dfa_exec() could read past the end of the passed subject if there was
1877        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1878        pcretest so that it places the subject at the end of its malloc-ed buffer.
1879    
1880    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1881        exec() might read past the end of the data buffer in UTF-8 mode.
1882    
1883    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1884        the data contained the byte 0x85 as part of a UTF-8 character within its
1885        first line. This applied both to normal and DFA matching.
1886    
1887    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1888        /^[^d]*?$/8 failed to match "abc".
1889    
1890    16. Added a missing copyright notice to pcrecpp_internal.h.
1891    
1892    17. Make it more clear in the documentation that values returned from
1893        pcre_exec() in ovector are byte offsets, not character counts.
1894    
1895    18. Tidied a few places to stop certain compilers from issuing warnings.
1896    
1897    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1898        supplied by Stefan Weber. I made a further small update for 7.8 because
1899        there is a change of source arrangements: the pcre_searchfuncs.c module is
1900        replaced by pcre_ucd.c.
1901    
1902    
1903    Version 7.7 07-May-08
1904  ---------------------  ---------------------
1905    
1906  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
1907      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
1908      done by checking for the strtoq, strtoll, and _strtoi64 functions.      done by checking for the strtoq, strtoll, and _strtoi64 functions.
1909    
1910  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
1911      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
1912      it in the RE class.      it in the RE class. (See also #8 below.)
1913    
1914  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
1915      Lopes.      Lopes.
1916    
1917  4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:  4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1918    
1919      (1) The include/exclude patterns were being applied to the whole pathnames      (1) The include/exclude patterns were being applied to the whole pathnames
1920          of files, instead of just to the final components.          of files, instead of just to the final components.
1921    
1922      (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
1923          skipped unless they satisfied the include/exclude conditions. This is          skipped unless they satisfied the include/exclude conditions. This is
1924          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
1925          pcregrep specification - which I improved to make it absolutely clear).          pcregrep specification - which I improved to make it absolutely clear).
1926          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
1927          apply the include/exclude patterns to regular files.          apply the include/exclude patterns to regular files.
1928    
1929  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
1930      --exclude_dir in the tests to avoid scanning .svn directories.      --exclude_dir in the tests to avoid scanning .svn directories.
1931    
1932  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
1933      NUL character as backslash + 0 rather than backslash + NUL, because PCRE      NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1934      doesn't support NULs in patterns.      doesn't support NULs in patterns.
1935    
1936  7.  Added some missing "const"s to declarations of static tables in  7.  Added some missing "const"s to declarations of static tables in
1937      pcre_compile.c and pcre_dfa_exec.c.      pcre_compile.c and pcre_dfa_exec.c.
1938    
1939    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1940        caused by fix #2  above. (Subsequently also a second patch to fix the
1941        first patch. And a third patch - this was a messy problem.)
1942    
1943    9.  Applied Craig's patch to remove the use of push_back().
1944    
1945    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1946        matching function regexec().
1947    
1948    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1949        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1950        references. PCRE supports relative numbers with this syntax (I don't think
1951        Oniguruma does).
1952    
1953    12. Previously, a group with a zero repeat such as (...){0} was completely
1954        omitted from the compiled regex. However, this means that if the group
1955        was called as a subroutine from elsewhere in the pattern, things went wrong
1956        (an internal error was given). Such groups are now left in the compiled
1957        pattern, with a new opcode that causes them to be skipped at execution
1958        time.
1959    
1960    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1961        to the way PCRE behaves:
1962    
1963        (a) A lone ] character is dis-allowed (Perl treats it as data).
1964    
1965        (b) A back reference to an unmatched subpattern matches an empty string
1966            (Perl fails the current match path).
1967    
1968        (c) A data ] in a character class must be notated as \] because if the
1969            first data character in a class is ], it defines an empty class. (In
1970            Perl it is not possible to have an empty class.) The empty class []
1971            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1972            The negative empty class [^] matches any one character, independently
1973            of the DOTALL setting.
1974    
1975    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1976        non-existent subpattern following a character class starting with ']' and
1977        containing () gave an internal compiling error instead of "reference to
1978        non-existent subpattern". Fortunately, when the pattern did exist, the
1979        compiled code was correct. (When scanning forwards to check for the
1980        existence of the subpattern, it was treating the data ']' as terminating
1981        the class, so got the count wrong. When actually compiling, the reference
1982        was subsequently set up correctly.)
1983    
1984    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1985        it was being rejected as not supported by pcre_dfa_exec(), even though
1986        other assertions are supported. I have made pcre_dfa_exec() support
1987        (*FAIL).
1988    
1989    16. The implementation of 13c above involved the invention of a new opcode,
1990        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1991        cannot be changed at match time, I realized I could make a small
1992        improvement to matching performance by compiling OP_ALLANY instead of
1993        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1994        on the OP_ANY path.
1995    
1996    17. Compiling pcretest on Windows with readline support failed without the
1997        following two fixes: (1) Make the unistd.h include conditional on
1998        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1999    
2000    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2001        ncurses library to be included for pcretest when ReadLine support is
2002        requested, but also to allow for it to be overridden. This patch came from
2003        Daniel Bergström.
2004    
2005    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2006        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2007        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2008        spotting this.
2009    
2010    
2011  Version 7.6 28-Jan-08  Version 7.6 28-Jan-08

Legend:
Removed from v.327  
changed lines
  Added in v.1352

  ViewVC Help
Powered by ViewVC 1.1.5