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

Diff of /code/trunk/ChangeLog

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

revision 155 by ph10, Tue Apr 24 13:36:11 2007 UTC revision 1354 by ph10, Mon Aug 12 14:31:07 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    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    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
63    
64    12. The source of pcregrep now includes z/OS-specific code so that it can be
65        compiled for z/OS as part of the special z/OS distribution.
66    
67    
68    Version 8.33 28-May-2013
69    ------------------------
70    
71    1.  Added 'U' to some constants that are compared to unsigned integers, to
72        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
73        variables that are added to signed variables, to ensure the result is
74        signed and can be negated.
75    
76    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
77        CMake config files.
78    
79    3.  Revise the creation of config.h.generic so that all boolean macros are
80        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
81        overriding via -D on the command line possible.
82    
83    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
84        to unsigned int is reported to make a quite noticeable speed difference in
85        a specific Windows environment. Testing on Linux did also appear to show
86        some benefit (and it is clearly not harmful). Also fixed the definition of
87        Xop which should be unsigned.
88    
89    5.  Related to (4), changing the definition of the intermediate variable cc
90        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
91        improvements.
92    
93    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
94        unnecessary spaces.
95    
96    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
97        and later.
98    
99    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
100    
101    9.  Optimizing fast_forward_start_bits in JIT.
102    
103    10. Adding support for callouts in JIT, and fixing some issues revealed
104        during this work. Namely:
105    
106        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
107    
108        (b) Minimum length was not checked before the matching is started.
109    
110    11. The value of capture_last that is passed to callouts was incorrect in some
111        cases when there was a capture on one path that was subsequently abandoned
112        after a backtrack. Also, the capture_last value is now reset after a
113        recursion, since all captures are also reset in this case.
114    
115    12. The interpreter no longer returns the "too many substrings" error in the
116        case when an overflowing capture is in a branch that is subsequently
117        abandoned after a backtrack.
118    
119    13. In the pathological case when an offset vector of size 2 is used, pcretest
120        now prints out the matched string after a yield of 0 or 1.
121    
122    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
123        Only supported by the JIT compiler at the moment.
124    
125    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
126    
127    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
128        offset of the starting point of the matching process, provided the offsets
129        vector is large enough.
130    
131    17. The \A escape now records a lookbehind value of 1, though its execution
132        does not actually inspect the previous character. This is to ensure that,
133        in partial multi-segment matching, at least one character from the old
134        segment is retained when a new segment is processed. Otherwise, if there
135        are no lookbehinds in the pattern, \A might match incorrectly at the start
136        of a new segment.
137    
138    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
139    
140    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
141        gives some modest performance improvement in 8-bit mode.
142    
143    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
144        be expressed in certain programming languages using Universal Character
145        Names.
146    
147    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
148        which points out that "non characters" are not "characters that may not
149        appear in Unicode strings" but rather "characters that are reserved for
150        internal use and have only local meaning".
151    
152    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
153        there was a conditional group that depended on an assertion, if the
154        assertion was false, the callout that immediately followed the alternation
155        in the condition was skipped when pcre_exec() was used for matching.
156    
157    23. Allow an explicit callout to be inserted before an assertion that is the
158        condition for a conditional group, for compatibility with automatic
159        callouts, which always insert a callout at this point.
160    
161    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
162        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
163    
164    25. (*PRUNE) is now supported by the JIT compiler.
165    
166    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
167    
168    27. Fix the case where there are two or more SKIPs with arguments that may be
169        ignored.
170    
171    28. (*SKIP) is now supported by the JIT compiler.
172    
173    29. (*THEN) is now supported by the JIT compiler.
174    
175    30. Update RunTest with additional test selector options.
176    
177    31. The way PCRE handles backtracking verbs has been changed in two ways.
178    
179        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
180        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
181        backtracking. In some cases this makes it more Perl-compatible, but Perl's
182        rather obscure rules do not always do the same thing.
183    
184        (2) Previously, backtracking verbs were confined within assertions. This is
185        no longer the case for positive assertions, except for (*ACCEPT). Again,
186        this sometimes improves Perl compatibility, and sometimes does not.
187    
188    32. A number of tests that were in test 2 because Perl did things differently
189        have been moved to test 1, because either Perl or PCRE has changed, and
190        these tests are now compatible.
191    
192    32. Backtracking control verbs are now handled in the same way in JIT and
193        interpreter.
194    
195    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
196        contained a forward subroutine reference caused a compile error.
197    
198    34. Auto-detect and optimize limited repetitions in JIT.
199    
200    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
201        blocking (*UTF) etc.
202    
203    36. In the interpreter, maximizing pattern repetitions for characters and
204        character types now use tail recursion, which reduces stack usage.
205    
206    37. The value of the max lookbehind was not correctly preserved if a compiled
207        and saved regex was reloaded on a host of different endianness.
208    
209    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
210        of the compiled pattern block, expand the flags field from 16 to 32 bits
211        because it was almost full.
212    
213    39. Try madvise first before posix_madvise.
214    
215    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
216        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
217        original change locked out all matches of empty strings. This has been
218        changed so that one match of an empty string per line is recognized.
219        Subsequent searches on the same line (for colouring or for --only-matching,
220        for example) do not recognize empty strings.
221    
222    41. Applied a user patch to fix a number of spelling mistakes in comments.
223    
224    42. Data lines longer than 65536 caused pcretest to crash.
225    
226    43. Clarified the data type for length and startoffset arguments for pcre_exec
227        and pcre_dfa_exec in the function-specific man pages, where they were
228        explicitly stated to be in bytes, never having been updated. I also added
229        some clarification to the pcreapi man page.
230    
231    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
232        a segmentation fault.
233    
234    
235    Version 8.32 30-November-2012
236    -----------------------------
237    
238    1.  Improved JIT compiler optimizations for first character search and single
239        character iterators.
240    
241    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
242        Patch by Daniel Richard G.
243    
244    3.  Single character iterator optimizations in the JIT compiler.
245    
246    4.  Improved JIT compiler optimizations for character ranges.
247    
248    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
249        Reported by Giuseppe D'Angelo.
250    
251    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
252        of a line, was being set incorrectly in cases where .* appeared inside
253        atomic brackets at the start of a pattern, or where there was a subsequent
254        *PRUNE or *SKIP.
255    
256    7.  Improved instruction cache flush for POWER/PowerPC.
257        Patch by Daniel Richard G.
258    
259    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
260        grep:
261    
262        (a) There is now no limit to the number of patterns to be matched.
263    
264        (b) An error is given if a pattern is too long.
265    
266        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
267            are now supported.
268    
269        (d) --exclude-from and --include-from (multiple use) have been added.
270    
271        (e) Exclusions and inclusions now apply to all files and directories, not
272            just to those obtained from scanning a directory recursively.
273    
274        (f) Multiple uses of -f and --file-list are now supported.
275    
276        (g) In a Windows environment, the default for -d has been changed from
277            "read" (the GNU grep default) to "skip", because otherwise the presence
278            of a directory in the file list provokes an error.
279    
280        (h) The documentation has been revised and clarified in places.
281    
282    9.  Improve the matching speed of capturing brackets.
283    
284    10. Changed the meaning of \X so that it now matches a Unicode extended
285        grapheme cluster.
286    
287    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
288        out POSIX threads when JIT support is configured.
289    
290    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
291    
292    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
293        structure could go wrong in environments where size_t is not the same size
294        as int.
295    
296    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
297        set.
298    
299    15. The EBCDIC support had decayed; later updates to the code had included
300        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
301        general tidy up of EBCDIC-related issues, and the documentation was also
302        not quite right. There is now a test that can be run on ASCII systems to
303        check some of the EBCDIC-related things (but is it not a full test).
304    
305    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
306        in a small tidy to the code.
307    
308    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
309    
310    18. If the --only-matching (-o) option in pcregrep is specified multiple
311        times, each one causes appropriate output. For example, -o1 -o2 outputs the
312        substrings matched by the 1st and 2nd capturing parentheses. A separating
313        string can be specified by --om-separator (default empty).
314    
315    19. Improving the first n character searches.
316    
317    20. Turn case lists for horizontal and vertical white space into macros so that
318        they are defined only once.
319    
320    21. This set of changes together give more compatible Unicode case-folding
321        behaviour for characters that have more than one other case when UCP
322        support is available.
323    
324        (a) The Unicode property table now has offsets into a new table of sets of
325            three or more characters that are case-equivalent. The MultiStage2.py
326            script that generates these tables (the pcre_ucd.c file) now scans
327            CaseFolding.txt instead of UnicodeData.txt for character case
328            information.
329    
330        (b) The code for adding characters or ranges of characters to a character
331            class has been abstracted into a generalized function that also handles
332            case-independence. In UTF-mode with UCP support, this uses the new data
333            to handle characters with more than one other case.
334    
335        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
336            whose other case is greater than 256 are now correctly matched
337            caselessly. Previously, the high codepoint matched the low one, but not
338            vice versa.
339    
340        (d) The processing of \h, \H, \v, and \ in character classes now makes use
341            of the new class addition function, using character lists defined as
342            macros alongside the case definitions of 20 above.
343    
344        (e) Caseless back references now work with characters that have more than
345            one other case.
346    
347        (f) General caseless matching of characters with more than one other case
348            is supported.
349    
350    22. Unicode character properties were updated from Unicode 6.2.0
351    
352    23. Improved CMake support under Windows. Patch by Daniel Richard G.
353    
354    24. Add support for 32-bit character strings, and UTF-32
355    
356    25. Major JIT compiler update (code refactoring and bugfixing).
357        Experimental Sparc 32 support is added.
358    
359    26. Applied a modified version of Daniel Richard G's patch to create
360        pcre.h.generic and config.h.generic by "make" instead of in the
361        PrepareRelease script.
362    
363    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
364        pcre_compile.c when checking for a zero character.
365    
366    28. Introducing a native interface for JIT. Through this interface, the compiled
367        machine code can be directly executed. The purpose of this interface is to
368        provide fast pattern matching, so several sanity checks are not performed.
369        However, feature tests are still performed. The new interface provides
370        1.4x speedup compared to the old one.
371    
372    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
373        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
374        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
375    
376    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
377        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
378        the "old" RFC 2279). Instead, it ended up passing a negative length to
379        pcre_exec().
380    
381    31. Add support for GCC's visibility feature to hide internal functions.
382    
383    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
384        "unknown -C option" after outputting 0 or 1.
385    
386    33. There is now support for generating a code coverage report for the test
387        suite in environments where gcc is the compiler and lcov is installed. This
388        is mainly for the benefit of the developers.
389    
390    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
391        unaddressable using valgrind annotations, allowing valgrind to detect
392        invalid memory accesses. This is mainly for the benefit of the developers.
393    
394    25. (*UTF) can now be used to start a pattern in any of the three libraries.
395    
396    26. Give configure error if --enable-cpp but no C++ compiler found.
397    
398    
399    Version 8.31 06-July-2012
400    -------------------------
401    
402    1.  Fixing a wrong JIT test case and some compiler warnings.
403    
404    2.  Removed a bashism from the RunTest script.
405    
406    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
407        to unsigned type, result still unsigned" that was given by an MS compiler
408        on encountering the code "-sizeof(xxx)".
409    
410    4.  Partial matching support is added to the JIT compiler.
411    
412    5.  Fixed several bugs concerned with partial matching of items that consist
413        of more than one character:
414    
415        (a) /^(..)\1/ did not partially match "aba" because checking references was
416            done on an "all or nothing" basis. This also applied to repeated
417            references.
418    
419        (b) \R did not give a hard partial match if \r was found at the end of the
420            subject.
421    
422        (c) \X did not give a hard partial match after matching one or more
423            characters at the end of the subject.
424    
425        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
426            a partial match for the string "\r".
427    
428        (e) When newline was set to CRLF, the metacharacter "." did not recognize
429            a partial match for a CR character at the end of the subject string.
430    
431    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
432        running pcretest, the text "(JIT)" added to the output whenever JIT is
433        actually used to run the match.
434    
435    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
436        or /S+[+] with a digit between 1 and 7.
437    
438    8.  OP_NOT now supports any UTF character not just single-byte ones.
439    
440    9.  (*MARK) control verb is now supported by the JIT compiler.
441    
442    10. The command "./RunTest list" lists the available tests without actually
443        running any of them. (Because I keep forgetting what they all are.)
444    
445    11. Add PCRE_INFO_MAXLOOKBEHIND.
446    
447    12. Applied a (slightly modified) user-supplied patch that improves performance
448        when the heap is used for recursion (compiled with --disable-stack-for-
449        recursion). Instead of malloc and free for each heap frame each time a
450        logical recursion happens, frames are retained on a chain and re-used where
451        possible. This sometimes gives as much as 30% improvement.
452    
453    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
454        call.
455    
456    14. As documented, (*COMMIT) is now confined to within a positive assertion.
457    
458    15. It is now possible to link pcretest with libedit as an alternative to
459        libreadline.
460    
461    16. (*COMMIT) control verb is now supported by the JIT compiler.
462    
463    17. The Unicode data tables have been updated to Unicode 6.1.0.
464    
465    18. Added --file-list option to pcregrep.
466    
467    19. Added binary file support to pcregrep, including the -a, --binary-files,
468        -I, and --text options.
469    
470    20. The madvise function is renamed for posix_madvise for QNX compatibility
471        reasons. Fixed by Giuseppe D'Angelo.
472    
473    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
474    
475    22. Changed the option for creating symbolic links for 16-bit man pages from
476        -s to -sf so that re-installing does not cause issues.
477    
478    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
479    
480    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
481        matches in certain environments (the workspace was not being correctly
482        retained). Also added to pcre_dfa_exec() a simple plausibility check on
483        some of the workspace data at the beginning of a restart.
484    
485    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
486        was not doing so when it should - probably a typo introduced by SVN 528
487        (change 8.10/14).
488    
489    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
490        \w+ when the character tables indicated that \x{c4} was a word character.
491        There were several related cases, all because the tests for doing a table
492        lookup were testing for characters less than 127 instead of 255.
493    
494    27. If a pattern contains capturing parentheses that are not used in a match,
495        their slots in the ovector are set to -1. For those that are higher than
496        any matched groups, this happens at the end of processing. In the case when
497        there were back references that the ovector was too small to contain
498        (causing temporary malloc'd memory to be used during matching), and the
499        highest capturing number was not used, memory off the end of the ovector
500        was incorrectly being set to -1. (It was using the size of the temporary
501        memory instead of the true size.)
502    
503    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
504        ovector size, it uses memory at the end of the block that it has got.
505    
506    29. Check for an overlong MARK name and give an error at compile time. The
507        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
508    
509    30. JIT compiler update.
510    
511    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
512        Rill for the patch.
513    
514    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
515    
516    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
517    
518    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
519        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
520        was enabled.
521    
522    35. Improve JIT code generation for greedy plus quantifier.
523    
524    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
525        1 to "aa" instead of to an empty string. The bug affected repeated groups
526        that could potentially match an empty string.
527    
528    37. Optimizing single character iterators in JIT.
529    
530    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
531        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
532        codepoints that are too big for the mode are faulted, and in a UTF mode,
533        disallowed codepoints are also faulted.
534    
535    39. If PCRE was compiled with UTF support, in three places in the DFA
536        matcher there was code that should only have been obeyed in UTF mode, but
537        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
538        processing when bytes with values greater than 127 were present. In 16-bit
539        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
540        both cases the values are those that cannot be the first data item in a UTF
541        character. The three items that might have provoked this were recursions,
542        possessively repeated groups, and atomic groups.
543    
544    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
545        and pcregrep, because some OS require shared objects to be explicitly
546        passed to ld, causing the link step to fail if they are not.
547    
548    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
549        mode, patterns that started with \h* or \R* might be incorrectly matched.
550    
551    
552    Version 8.30 04-February-2012
553    -----------------------------
554    
555    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
556        name is defined in ctype.h.
557    
558    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
559        only in quite long subpatterns.
560    
561    3.  Removed the function pcre_info(), which has been obsolete and deprecated
562        since it was replaced by pcre_fullinfo() in February 2000.
563    
564    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
565        match a (*MARK), and the match failed at the start of the subject, a
566        reference to memory before the start of the subject could occur. This bug
567        was introduced by fix 17 of release 8.21.
568    
569    5.  A reference to an unset group with zero minimum repetition was giving
570        totally wrong answers (in non-JavaScript-compatibility mode). For example,
571        /(another)?(\1?)test/ matched against "hello world test". This bug was
572        introduced in release 8.13.
573    
574    6.  Add support for 16-bit character strings (a large amount of work involving
575        many changes and refactorings).
576    
577    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
578        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
579        from a file.
580    
581    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
582        rounding is not applied in this particular case).
583    
584    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
585        if they appear, or are escaped, in patterns.
586    
587    10. Get rid of a number of -Wunused-but-set-variable warnings.
588    
589    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
590        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
591        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
592        also returns the mark "x". This bug applied to capturing parentheses,
593        non-capturing parentheses, and atomic parentheses. It also applied to some
594        assertions.
595    
596    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
597        information out of configure.ac instead of relying on pcre.h.generic, which
598        is not stored in the repository.
599    
600    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
601        -lreadline.
602    
603    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
604    
605    15. Applied Graycode's patch to put the top-level frame on the stack rather
606        than the heap when not using the stack for recursion. This gives a
607        performance improvement in many cases when recursion is not deep.
608    
609    16. Experimental code added to "pcretest -C" to output the stack frame size.
610    
611    
612    Version 8.21 12-Dec-2011
613    ------------------------
614    
615    1.  Updating the JIT compiler.
616    
617    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
618        are added as well.
619    
620    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
621        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
622        calling _pcre_jit_exec. Some extra comments are added.
623    
624    4.  (*MARK) settings inside atomic groups that do not contain any capturing
625        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
626        was introduced by change 18 for 8.20.
627    
628    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
629        ECMA-262 standard.
630    
631    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
632        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
633        This bug was probably introduced by change 9 of 8.13.
634    
635    7.  While fixing 6 above, I noticed that a number of other items were being
636        incorrectly rejected as "not fixed length". This arose partly because newer
637        opcodes had not been added to the fixed-length checking code. I have (a)
638        corrected the bug and added tests for these items, and (b) arranged for an
639        error to occur if an unknown opcode is encountered while checking for fixed
640        length instead of just assuming "not fixed length". The items that were
641        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
642        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
643        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
644    
645    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
646        being incorrectly compiled and would have given unpredicatble results.
647    
648    9.  A possessively repeated subpattern with minimum repeat count greater than
649        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
650        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
651        the first (A) could occur when it should not.
652    
653    10. Add a cast and remove a redundant test from the code.
654    
655    11. JIT should use pcre_malloc/pcre_free for allocation.
656    
657    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
658        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
659        is anything other than /usr, -L is still shown).
660    
661    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
662    
663    14. Perl does not support \N without a following name in a [] class; PCRE now
664        also gives an error.
665    
666    15. If a forward reference was repeated with an upper limit of around 2000,
667        it caused the error "internal error: overran compiling workspace". The
668        maximum number of forward references (including repeats) was limited by the
669        internal workspace, and dependent on the LINK_SIZE. The code has been
670        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
671        the default depends on LINK_SIZE. There is a new upper limit (for safety)
672        of around 200,000 forward references. While doing this, I also speeded up
673        the filling in of repeated forward references.
674    
675    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
676        incorrectly expecting the subject to contain another "a" after the start.
677    
678    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
679        in the match, the SKIP should be ignored. This was not happening; instead
680        the SKIP was being treated as NOMATCH. For patterns such as
681        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
682        tested.
683    
684    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
685        now much more compatible with Perl, in particular in cases where the result
686        is a non-match for a non-anchored pattern. For example, if
687        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
688        "m", where previously it did not return a name. A side effect of this
689        change is that for partial matches, the last encountered mark name is
690        returned, as for non matches. A number of tests that were previously not
691        Perl-compatible have been moved into the Perl-compatible test files. The
692        refactoring has had the pleasing side effect of removing one argument from
693        the match() function, thus reducing its stack requirements.
694    
695    19. If the /S+ option was used in pcretest to study a pattern using JIT,
696        subsequent uses of /S (without +) incorrectly behaved like /S+.
697    
698    21. Retrieve executable code size support for the JIT compiler and fixing
699        some warnings.
700    
701    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
702        not work when the shorter character appeared right at the end of the
703        subject string.
704    
705    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
706        systems.
707    
708    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
709        output it when the /M option is used in pcretest.
710    
711    25. The CheckMan script was not being included in the distribution. Also, added
712        an explicit "perl" to run Perl scripts from the PrepareRelease script
713        because this is reportedly needed in Windows.
714    
715    26. If study data was being save in a file and studying had not found a set of
716        "starts with" bytes for the pattern, the data written to the file (though
717        never used) was taken from uninitialized memory and so caused valgrind to
718        complain.
719    
720    27. Updated RunTest.bat as provided by Sheri Pierce.
721    
722    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
723    
724    29. Computation of memory usage for the table of capturing group names was
725        giving an unnecessarily large value.
726    
727    
728    Version 8.20 21-Oct-2011
729    ------------------------
730    
731    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
732        a POSIX class. After further experiments with Perl, which convinced me that
733        Perl has bugs and confusions, a closing square bracket is no longer allowed
734        in a POSIX name. This bug also affected patterns with classes that started
735        with full stops.
736    
737    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
738        captured substring, but while checking the failing first alternative,
739        substring 1 is temporarily captured. If the output vector supplied to
740        pcre_exec() was not big enough for this capture, the yield of the function
741        was still zero ("insufficient space for captured substrings"). This cannot
742        be totally fixed without adding another stack variable, which seems a lot
743        of expense for a edge case. However, I have improved the situation in cases
744        such as /(a)(b)x|abc/ matched against "abc", where the return code
745        indicates that fewer than the maximum number of slots in the ovector have
746        been set.
747    
748    3.  Related to (2) above: when there are more back references in a pattern than
749        slots in the output vector, pcre_exec() uses temporary memory during
750        matching, and copies in the captures as far as possible afterwards. It was
751        using the entire output vector, but this conflicts with the specification
752        that only 2/3 is used for passing back captured substrings. Now it uses
753        only the first 2/3, for compatibility. This is, of course, another edge
754        case.
755    
756    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
757        main code base, and can be used by building with --enable-jit. When this is
758        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
759        runtime --no-jit option is given.
760    
761    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
762        ovector, the return from the function was zero, implying that there were
763        other matches that did not fit. The correct "exactly full" value is now
764        returned.
765    
766    6.  If a subpattern that was called recursively or as a subroutine contained
767        (*PRUNE) or any other control that caused it to give a non-standard return,
768        invalid errors such as "Error -26 (nested recursion at the same subject
769        position)" or even infinite loops could occur.
770    
771    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
772        computing the minimum length on reaching *ACCEPT, and so ended up with the
773        wrong value of 1 rather than 0. Further investigation indicates that
774        computing a minimum subject length in the presence of *ACCEPT is difficult
775        (think back references, subroutine calls), and so I have changed the code
776        so that no minimum is registered for a pattern that contains *ACCEPT.
777    
778    8.  If (*THEN) was present in the first (true) branch of a conditional group,
779        it was not handled as intended. [But see 16 below.]
780    
781    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
782        Sheri Pierce.
783    
784    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
785        the first byte in a match must be "a".
786    
787    11. Change 17 for 8.13 increased the recursion depth for patterns like
788        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
789        pattern contains any instances of (*THEN). If it does not, the old
790        optimizations are restored. It would be nice to do this on a per-group
791        basis, but at the moment that is not feasible.
792    
793    12. In some environments, the output of pcretest -C is CRLF terminated. This
794        broke RunTest's code that checks for the link size. A single white space
795        character after the value is now allowed for.
796    
797    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
798        For "fr", it uses the Windows-specific input and output files.
799    
800    14. If (*THEN) appeared in a group that was called recursively or as a
801        subroutine, it did not work as intended. [But see next item.]
802    
803    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
804        pattern fragments (but not containing any | characters). If A and B are
805        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
806        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
807        D. In other words, Perl considers parentheses that do not contain any |
808        characters to be part of a surrounding alternative, whereas PCRE was
809        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
810        differently. PCRE now behaves in the same way as Perl, except in the case
811        of subroutine/recursion calls such as (?1) which have in any case always
812        been different (but PCRE had them first :-).
813    
814    16. Related to 15 above: Perl does not treat the | in a conditional group as
815        creating alternatives. Such a group is treated in the same way as an
816        ordinary group without any | characters when processing (*THEN). PCRE has
817        been changed to match Perl's behaviour.
818    
819    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
820        RunGrepTest script failed.
821    
822    18. Change 22 for version 13 caused atomic groups to use more stack. This is
823        inevitable for groups that contain captures, but it can lead to a lot of
824        stack use in large patterns. The old behaviour has been restored for atomic
825        groups that do not contain any capturing parentheses.
826    
827    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
828        suppress the check for a minimum subject length at run time. (If it was
829        given to pcre_exec() or pcre_dfa_exec() it did work.)
830    
831    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
832        fail to work when decoding hex characters in data strings in EBCDIC
833        environments.
834    
835    21. It appears that in at least one Mac OS environment, the isxdigit() function
836        is implemented as a macro that evaluates to its argument more than once,
837        contravening the C 90 Standard (I haven't checked a later standard). There
838        was an instance in pcretest which caused it to go wrong when processing
839        \x{...} escapes in subject strings. The has been rewritten to avoid using
840        things like p++ in the argument of isxdigit().
841    
842    
843    Version 8.13 16-Aug-2011
844    ------------------------
845    
846    1.  The Unicode data tables have been updated to Unicode 6.0.0.
847    
848    2.  Two minor typos in pcre_internal.h have been fixed.
849    
850    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
851        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
852        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
853    
854    4.  There were a number of related bugs in the code for matching backrefences
855        caselessly in UTF-8 mode when codes for the characters concerned were
856        different numbers of bytes. For example, U+023A and U+2C65 are an upper
857        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
858        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
859        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
860        2-byte code at the end of the subject (it thought there wasn't enough data
861        left).
862    
863    5.  Comprehensive information about what went wrong is now returned by
864        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
865        as the output vector has at least 2 elements. The offset of the start of
866        the failing character and a reason code are placed in the vector.
867    
868    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
869        now returned is for the first byte of the failing character, instead of the
870        last byte inspected. This is an incompatible change, but I hope it is small
871        enough not to be a problem. It makes the returned offset consistent with
872        pcre_exec() and pcre_dfa_exec().
873    
874    7.  pcretest now gives a text phrase as well as the error number when
875        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
876        failure, the offset and reason code are output.
877    
878    8.  When \R was used with a maximizing quantifier it failed to skip backwards
879        over a \r\n pair if the subsequent match failed. Instead, it just skipped
880        back over a single character (\n). This seems wrong (because it treated the
881        two characters as a single entity when going forwards), conflicts with the
882        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
883        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
884        has been changed.
885    
886    9.  Some internal refactoring has changed the processing so that the handling
887        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
888        time (the PCRE_DOTALL option was changed this way some time ago: version
889        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
890        which was always a bit of a fudge. It also means that there is one less
891        argument for the match() function, which reduces its stack requirements
892        slightly. This change also fixes an incompatibility with Perl: the pattern
893        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
894    
895    10. More internal refactoring has drastically reduced the number of recursive
896        calls to match() for possessively repeated groups such as (abc)++ when
897        using pcre_exec().
898    
899    11. While implementing 10, a number of bugs in the handling of groups were
900        discovered and fixed:
901    
902        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
903        (a|)*(?1) gave a compile-time internal error.
904        ((a|)+)+  did not notice that the outer group could match an empty string.
905        (^a|^)+   was not marked as anchored.
906        (.*a|.*)+ was not marked as matching at start or after a newline.
907    
908    12. Yet more internal refactoring has removed another argument from the match()
909        function. Special calls to this function are now indicated by setting a
910        value in a variable in the "match data" data block.
911    
912    13. Be more explicit in pcre_study() instead of relying on "default" for
913        opcodes that mean there is no starting character; this means that when new
914        ones are added and accidentally left out of pcre_study(), testing should
915        pick them up.
916    
917    14. The -s option of pcretest has been documented for ages as being an old
918        synonym of -m (show memory usage). I have changed it to mean "force study
919        for every regex", that is, assume /S for every regex. This is similar to -i
920        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
921        using it. It makes it easier to run collections of tests with and without
922        study enabled, and thereby test pcre_study() more easily. All the standard
923        tests are now run with and without -s (but some patterns can be marked as
924        "never study" - see 20 below).
925    
926    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
927        restoration of the capturing data to the outer values was not happening
928        correctly.
929    
930    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
931        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
932        pattern had matched an empty string, and so incorrectly returned a no
933        match.
934    
935    17. There was optimizing code for the last branch of non-capturing parentheses,
936        and also for the obeyed branch of a conditional subexpression, which used
937        tail recursion to cut down on stack usage. Unfortunately, now that there is
938        the possibility of (*THEN) occurring in these branches, tail recursion is
939        no longer possible because the return has to be checked for (*THEN). These
940        two optimizations have therefore been removed. [But see 8.20/11 above.]
941    
942    18. If a pattern containing \R was studied, it was assumed that \R always
943        matched two bytes, thus causing the minimum subject length to be
944        incorrectly computed because \R can also match just one byte.
945    
946    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
947        was incorrectly computed.
948    
949    20. If /S is present twice on a test pattern in pcretest input, it now
950        *disables* studying, thereby overriding the use of -s on the command line
951        (see 14 above). This is necessary for one or two tests to keep the output
952        identical in both cases.
953    
954    21. When (*ACCEPT) was used in an assertion that matched an empty string and
955        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
956    
957    22. When an atomic group that contained a capturing parenthesis was
958        successfully matched, but the branch in which it appeared failed, the
959        capturing was not being forgotten if a higher numbered group was later
960        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
961        group 1 to "a", when in fact it should be unset. This applied to multi-
962        branched capturing and non-capturing groups, repeated or not, and also to
963        positive assertions (capturing in negative assertions does not happen
964        in PCRE) and also to nested atomic groups.
965    
966    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
967        subject after a captured substring, to make it easier to tell which of a
968        number of identical substrings has been captured.
969    
970    24. The way atomic groups are processed by pcre_exec() has been changed so that
971        if they are repeated, backtracking one repetition now resets captured
972        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
973        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
974        "aaa". Previously, it would have been "a". As part of this code
975        refactoring, the way recursive calls are handled has also been changed.
976    
977    25. If an assertion condition captured any substrings, they were not passed
978        back unless some other capturing happened later. For example, if
979        (?(?=(a))a) was matched against "a", no capturing was returned.
980    
981    26. When studying a pattern that contained subroutine calls or assertions,
982        the code for finding the minimum length of a possible match was handling
983        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
984        group 1 called group 2 while simultaneously a separate group 2 called group
985        1). A stack overflow occurred in this case. I have fixed this by limiting
986        the recursion depth to 10.
987    
988    27. Updated RunTest.bat in the distribution to the version supplied by Tom
989        Fortmann. This supports explicit test numbers on the command line, and has
990        argument validation and error reporting.
991    
992    28. An instance of \X with an unlimited repeat could fail if at any point the
993        first character it looked at was a mark character.
994    
995    29. Some minor code refactoring concerning Unicode properties and scripts
996        should reduce the stack requirement of match() slightly.
997    
998    30. Added the '=' option to pcretest to check the setting of unused capturing
999        slots at the end of the pattern, which are documented as being -1, but are
1000        not included in the return count.
1001    
1002    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1003        compiled something random. Now it gives a compile-time error (as does
1004        Perl).
1005    
1006    32. A *MARK encountered during the processing of a positive assertion is now
1007        recorded and passed back (compatible with Perl).
1008    
1009    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1010        had alternative anchored branches, the search for a second match in a line
1011        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1012        matched the line "0102" twice. The same bug affected patterns that started
1013        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1014        twice.
1015    
1016    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1017        does, and because of capturing effects, quantifying parenthesized
1018        assertions may at times be useful. Quantifiers are now allowed for
1019        parenthesized assertions.
1020    
1021    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1022    
1023    36. \g was being checked for fancy things in a character class, when it should
1024        just be a literal "g".
1025    
1026    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1027        appearance of a nested POSIX class supersedes an apparent external class.
1028        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1029        unescaped square brackets may also appear as part of class names. For
1030        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1031        more like Perl. (But see 8.20/1 above.)
1032    
1033    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1034        was because it thought it was \N{name}, which is not supported).
1035    
1036    39. Add minix to OS list not supporting the -S option in pcretest.
1037    
1038    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1039        cannot analyze patterns in sufficient detail to catch mutual recursions
1040        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1041        subgroup is called recursively as a subpattern for a second time at the
1042        same position in the subject string. In previous releases this might have
1043        been caught by the recursion limit, or it might have run out of stack.
1044    
1045    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1046        happen only once. PCRE was, however incorrectly giving a compile time error
1047        "recursive call could loop indefinitely" because it cannot analyze the
1048        pattern in sufficient detail. The compile time test no longer happens when
1049        PCRE is compiling a conditional subpattern, but actual runaway loops are
1050        now caught at runtime (see 40 above).
1051    
1052    42. It seems that Perl allows any characters other than a closing parenthesis
1053        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1054        has been changed to be the same.
1055    
1056    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1057        as not to get warnings when autogen.sh is called. Also changed
1058        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1059    
1060    44. To help people who use pcregrep to scan files containing exceedingly long
1061        lines, the following changes have been made:
1062    
1063        (a) The default value of the buffer size parameter has been increased from
1064            8K to 20K. (The actual buffer used is three times this size.)
1065    
1066        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1067            PCRE is built.
1068    
1069        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1070            to be set at run time.
1071    
1072        (d) Numerical values in pcregrep options can be followed by K or M, for
1073            example --buffer-size=50K.
1074    
1075        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1076            given and the return code is set to 2.
1077    
1078    45. Add a pointer to the latest mark to the callout data block.
1079    
1080    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1081        partial match of an empty string instead of no match. This was specific to
1082        the use of ".".
1083    
1084    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1085        complete match instead of a partial match. This bug was dependent on both
1086        the PCRE_UTF8 and PCRE_DOTALL options being set.
1087    
1088    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1089        starting byte set, because \b was not being ignored.
1090    
1091    
1092    Version 8.12 15-Jan-2011
1093    ------------------------
1094    
1095    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1096        checks for such things as part of the documentation building process.
1097    
1098    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1099        --match-limit and --recursion-limit options (added for 8.11). In
1100        particular, this made one of the standard tests fail. (The integer value
1101        went into the wrong half of a long int.)
1102    
1103    3.  If the --colour option was given to pcregrep with -v (invert match), it
1104        did strange things, either producing crazy output, or crashing. It should,
1105        of course, ignore a request for colour when reporting lines that do not
1106        match.
1107    
1108    4.  Another pcregrep bug caused similar problems if --colour was specified with
1109        -M (multiline) and the pattern match finished with a line ending.
1110    
1111    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1112        matched in multiline mode, the following line was shown as part of the
1113        match. This seems wrong, so I have changed it.
1114    
1115    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1116        the check for further matches in the same line (so they could be coloured)
1117        to overrun the end of the current line. If another match was found, it was
1118        incorrectly shown (and then shown again when found in the next line).
1119    
1120    7.  If pcregrep was compiled under Windows, there was a reference to the
1121        function pcregrep_exit() before it was defined. I am assuming this was
1122        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1123        reported by a user. I've moved the definition above the reference.
1124    
1125    
1126    Version 8.11 10-Dec-2010
1127    ------------------------
1128    
1129    1.  (*THEN) was not working properly if there were untried alternatives prior
1130        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1131        backtracked to try for "b" instead of moving to the next alternative branch
1132        at the same level (in this case, to look for "c"). The Perl documentation
1133        is clear that when (*THEN) is backtracked onto, it goes to the "next
1134        alternative in the innermost enclosing group".
1135    
1136    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1137        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1138        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1139        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1140        (*THEN).
1141    
1142    3.  If \s appeared in a character class, it removed the VT character from
1143        the class, even if it had been included by some previous item, for example
1144        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1145        of \s, but is part of the POSIX "space" class.)
1146    
1147    4.  A partial match never returns an empty string (because you can always
1148        match an empty string at the end of the subject); however the checking for
1149        an empty string was starting at the "start of match" point. This has been
1150        changed to the "earliest inspected character" point, because the returned
1151        data for a partial match starts at this character. This means that, for
1152        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1153        (previously it gave "no match").
1154    
1155    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1156        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1157        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1158        has an implication that the given string is incomplete (because a partial
1159        match is preferred over a full match). For this reason, these items now
1160        give a partial match in this situation. [Aside: previously, the one case
1161        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1162        match rather than a full match, which was wrong by the old rules, but is
1163        now correct.]
1164    
1165    6.  There was a bug in the handling of #-introduced comments, recognized when
1166        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1167        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1168        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1169        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1170        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1171        places in pcre_compile().
1172    
1173    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1174        comments when looking ahead for named forward references to subpatterns,
1175        the only newline sequence it recognized was NL. It now handles newlines
1176        according to the set newline convention.
1177    
1178    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1179        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1180        cater for a lack of strerror(). These oversights have been fixed.
1181    
1182    9.  Added --match-limit and --recursion-limit to pcregrep.
1183    
1184    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1185    
1186    11. When the -o option was used, pcregrep was setting a return code of 1, even
1187        when matches were found, and --line-buffered was not being honoured.
1188    
1189    12. Added an optional parentheses number to the -o and --only-matching options
1190        of pcregrep.
1191    
1192    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1193        can match an empty string. The code to do it in pcretest and pcredemo
1194        needed fixing:
1195    
1196        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1197            only one byte after an empty string match just before CRLF (this case
1198            just got forgotten; "any" and "anycrlf" were OK).
1199    
1200        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1201            mode when an empty string match preceded an ASCII character followed by
1202            a non-ASCII character. (The code for advancing by one character rather
1203            than one byte was nonsense.)
1204    
1205        (c) The pcredemo.c sample program did not have any code at all to handle
1206            the cases when CRLF is a valid newline sequence.
1207    
1208    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1209        as a starting offset was within the subject string. There is now a new
1210        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1211        negative or greater than the length of the string. In order to test this,
1212        pcretest is extended to allow the setting of negative starting offsets.
1213    
1214    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1215        starting offset points to the beginning of a UTF-8 character was
1216        unnecessarily clumsy. I tidied it up.
1217    
1218    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1219        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1220    
1221    17. Nobody had reported that the --include_dir option, which was added in
1222        release 7.7 should have been called --include-dir (hyphen, not underscore)
1223        for compatibility with GNU grep. I have changed it to --include-dir, but
1224        left --include_dir as an undocumented synonym, and the same for
1225        --exclude-dir, though that is not available in GNU grep, at least as of
1226        release 2.5.4.
1227    
1228    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1229        characters from a string of bytes) have been redefined so as not to use
1230        loops, in order to improve performance in some environments. At the same
1231        time, I abstracted some of the common code into auxiliary macros to save
1232        repetition (this should not affect the compiled code).
1233    
1234    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1235        compile-time error is now given if \c is not followed by an ASCII
1236        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1237        different, and any byte value is allowed.)
1238    
1239    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1240        START_OPTIMIZE option, which is now allowed at compile time - but just
1241        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1242        to pcregrep and other applications that have no direct access to PCRE
1243        options. The new /Y option in pcretest sets this option when calling
1244        pcre_compile().
1245    
1246    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1247        back references. Groups containing recursive back references were forced to
1248        be atomic by that change, but in the case of named groups, the amount of
1249        memory required was incorrectly computed, leading to "Failed: internal
1250        error: code overflow". This has been fixed.
1251    
1252    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1253        pcretest.c, to avoid build problems in some Borland environments.
1254    
1255    
1256    Version 8.10 25-Jun-2010
1257    ------------------------
1258    
1259    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1260        THEN.
1261    
1262    2.  (*ACCEPT) was not working when inside an atomic group.
1263    
1264    3.  Inside a character class, \B is treated as a literal by default, but
1265        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1266        causes the error). The code is unchanged, but I tidied the documentation.
1267    
1268    4.  Inside a character class, PCRE always treated \R and \X as literals,
1269        whereas Perl faults them if its -w option is set. I have changed PCRE so
1270        that it faults them when PCRE_EXTRA is set.
1271    
1272    5.  Added support for \N, which always matches any character other than
1273        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1274    
1275    6.  When compiling pcregrep with newer versions of gcc which may have
1276        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1277        declared with attribute warn_unused_result" were given. Just casting the
1278        result to (void) does not stop the warnings; a more elaborate fudge is
1279        needed. I've used a macro to implement this.
1280    
1281    7.  Minor change to pcretest.c to avoid a compiler warning.
1282    
1283    8.  Added four artifical Unicode properties to help with an option to make
1284        \s etc use properties (see next item). The new properties are: Xan
1285        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1286    
1287    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1288        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1289        this option. Modified pcretest to add /W to test this facility. Added
1290        REG_UCP to make it available via the POSIX interface.
1291    
1292    10. Added --line-buffered to pcregrep.
1293    
1294    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1295        studied, and the match started with a letter with a code point greater than
1296        127 whose first byte was different to the first byte of the other case of
1297        the letter, the other case of this starting letter was not recognized
1298        (#976).
1299    
1300    12. If a pattern that was studied started with a repeated Unicode property
1301        test, for example, \p{Nd}+, there was the theoretical possibility of
1302        setting up an incorrect bitmap of starting bytes, but fortunately it could
1303        not have actually happened in practice until change 8 above was made (it
1304        added property types that matched character-matching opcodes).
1305    
1306    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1307        possible starting bytes for non-anchored patterns.
1308    
1309    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1310        \R, and also a number of cases that involve Unicode properties, both
1311        explicit and implicit when PCRE_UCP is set.
1312    
1313    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1314        input, it could crash or give wrong results if characters with values
1315        greater than 0xc0 were present in the subject string. (Detail: it assumed
1316        UTF-8 input when processing these items.)
1317    
1318    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1319        size_t is 64-bit (#991).
1320    
1321    17. Added a check for running out of memory when PCRE is compiled with
1322        --disable-stack-for-recursion (#990).
1323    
1324    18. If the last data line in a file for pcretest does not have a newline on
1325        the end, a newline was missing in the output.
1326    
1327    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1328        less than 128) in its various bitmaps. However, there is a facility for
1329        generating tables according to the current locale when PCRE is compiled. It
1330        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1331        characters, are recognized by isspace() and therefore were getting set in
1332        these tables, and indeed these tables seem to approximate to ISO 8859. This
1333        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1334        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1335        which of course cannot start UTF-8 characters. I have changed the code so
1336        that only real ASCII characters (less than 128) and the correct starting
1337        bytes for UTF-8 encodings are set for characters greater than 127 when in
1338        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1339        altogether.)
1340    
1341    20. Added the /T option to pcretest so as to be able to run tests with non-
1342        standard character tables, thus making it possible to include the tests
1343        used for 19 above in the standard set of tests.
1344    
1345    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1346        reference to a subpattern the other side of a comment that contains an
1347        opening parenthesis caused either an internal compiling error, or a
1348        reference to the wrong subpattern.
1349    
1350    
1351    Version 8.02 19-Mar-2010
1352    ------------------------
1353    
1354    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1355    
1356    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1357        configured.
1358    
1359    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1360        original author of that file, following a query about its status.
1361    
1362    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1363        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1364    
1365    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1366        quantifier applied to a forward-referencing subroutine call, could compile
1367        incorrect code or give the error "internal error: previously-checked
1368        referenced subpattern not found".
1369    
1370    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1371        variables to point to external functions. For these systems, therefore,
1372        pcre_malloc etc. are now initialized to local functions that call the
1373        relevant global functions.
1374    
1375    7.  There were two entries missing in the vectors called coptable and poptable
1376        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1377        I've fixed the data, and added a kludgy way of testing at compile time that
1378        the lengths are correct (equal to the number of opcodes).
1379    
1380    8.  Following on from 7, I added a similar kludge to check the length of the
1381        eint vector in pcreposix.c.
1382    
1383    9.  Error texts for pcre_compile() are held as one long string to avoid too
1384        much relocation at load time. To find a text, the string is searched,
1385        counting zeros. There was no check for running off the end of the string,
1386        which could happen if a new error number was added without updating the
1387        string.
1388    
1389    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1390    
1391    11. \K was not working if it appeared in an atomic group or in a group that
1392        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1393        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1394        the assertion is positive, but not if it is negative.
1395    
1396    12. Change 11 fortuitously reduced the size of the stack frame used in the
1397        "match()" function of pcre_exec.c by one pointer. Forthcoming
1398        implementation of support for (*MARK) will need an extra pointer on the
1399        stack; I have reserved it now, so that the stack frame size does not
1400        decrease.
1401    
1402    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1403        item in branch that calls a recursion is a subroutine call - as in the
1404        second branch in the above example - was incorrectly given the compile-
1405        time error "recursive call could loop indefinitely" because pcre_compile()
1406        was not correctly checking the subroutine for matching a non-empty string.
1407    
1408    14. The checks for overrunning compiling workspace could trigger after an
1409        overrun had occurred. This is a "should never occur" error, but it can be
1410        triggered by pathological patterns such as hundreds of nested parentheses.
1411        The checks now trigger 100 bytes before the end of the workspace.
1412    
1413    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1414    
1415    
1416    Version 8.01 19-Jan-2010
1417    ------------------------
1418    
1419    1.  If a pattern contained a conditional subpattern with only one branch (in
1420        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1421        computed the wrong minimum data length (which is of course zero for such
1422        subpatterns). This could cause incorrect "no match" results.
1423    
1424    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1425        the pattern is reset in the first branch, pcre_compile() failed with
1426        "internal error: code overflow at offset...". This happened only when
1427        the reset was to the original external option setting. (An optimization
1428        abstracts leading options settings into an external setting, which was the
1429        cause of this.)
1430    
1431    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1432        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1433        assertion pattern did not match (meaning that the assertion was true), it
1434        was incorrectly treated as false if the SKIP had been reached during the
1435        matching. This also applied to assertions used as conditions.
1436    
1437    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1438        assertion subpattern, including such a pattern used as a condition,
1439        unpredictable results occurred, instead of the error return
1440        PCRE_ERROR_DFA_UITEM.
1441    
1442    5.  The C++ GlobalReplace function was not working like Perl for the special
1443        situation when an empty string is matched. It now does the fancy magic
1444        stuff that is necessary.
1445    
1446    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1447        removed. (These were left over from very, very early versions of PCRE.)
1448    
1449    7.  Some cosmetic changes to the code to make life easier when compiling it
1450        as part of something else:
1451    
1452        (a) Change DEBUG to PCRE_DEBUG.
1453    
1454        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1455            called "current" as "current_branch", to prevent a collision with the
1456            Linux macro when compiled as a kernel module.
1457    
1458        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1459            prevent a collision with the Linux macro when compiled as a kernel
1460            module.
1461    
1462    8.  In pcre_compile() there are some checks for integer overflows that used to
1463        cast potentially large values to (double). This has been changed to that
1464        when building, a check for int64_t is made, and if it is found, it is used
1465        instead, thus avoiding the use of floating point arithmetic. (There is no
1466        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1467        double.
1468    
1469    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1470        2005 (difference between two addresses compared to an unsigned value).
1471    
1472    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1473        custom one, because of the following reported problem in Windows:
1474    
1475          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1476              under Win32.
1477          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1478              therefore missing the function definition.
1479          - The compiler thus generates a "C" signature for the test function.
1480          - The linker fails to find the "C" function.
1481          - PCRE fails to configure if asked to do so against libbz2.
1482    
1483    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1484        messages were output:
1485    
1486          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1487          rerunning libtoolize, to keep the correct libtool macros in-tree.
1488          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1489    
1490        I have done both of these things.
1491    
1492    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1493        most of the time, it *can* run out if it is given a pattern that contains a
1494        runaway infinite recursion. I updated the discussion in the pcrestack man
1495        page.
1496    
1497    13. Now that we have gone to the x.xx style of version numbers, the minor
1498        version may start with zero. Using 08 or 09 is a bad idea because users
1499        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1500        interpreted as invalid octal numbers. I've updated the previous comment in
1501        configure.ac, and also added a check that gives an error if 08 or 09 are
1502        used.
1503    
1504    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1505        causing partial matching to fail when the end of the subject matched \W
1506        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1507    
1508    15. There were some discrepancies between the declarations in pcre_internal.h
1509        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1510        their definitions. The declarations used "const uschar *" and the
1511        definitions used USPTR. Even though USPTR is normally defined as "const
1512        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1513        reported that: "This difference in casting confuses some C++ compilers, for
1514        example, SunCC recognizes above declarations as different functions and
1515        generates broken code for hbpcre." I have changed the declarations to use
1516        USPTR.
1517    
1518    16. GNU libtool is named differently on some systems. The autogen.sh script now
1519        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1520        (FreeBSD).
1521    
1522    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1523        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1524        comment: "Figure out how to create a longlong from a string: strtoll and
1525        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1526        instance, but it only takes 2 args instead of 3!"
1527    
1528    18. A subtle bug concerned with back references has been fixed by a change of
1529        specification, with a corresponding code fix. A pattern such as
1530        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1531        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1532        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1533        same bug. Such groups have to be quantified to be useful, or contained
1534        inside another quantified group. (If there's no repetition, the reference
1535        can never match.) The problem arises because, having left the group and
1536        moved on to the rest of the pattern, a later failure that backtracks into
1537        the group uses the captured value from the final iteration of the group
1538        rather than the correct earlier one. I have fixed this in PCRE by forcing
1539        any group that contains a reference to itself to be an atomic group; that
1540        is, there cannot be any backtracking into it once it has completed. This is
1541        similar to recursive and subroutine calls.
1542    
1543    
1544    Version 8.00 19-Oct-09
1545    ----------------------
1546    
1547    1.  The table for translating pcre_compile() error codes into POSIX error codes
1548        was out-of-date, and there was no check on the pcre_compile() error code
1549        being within the table. This could lead to an OK return being given in
1550        error.
1551    
1552    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1553        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1554        in a Windows environment.
1555    
1556    3.  The pcregrep --count option prints the count for each file even when it is
1557        zero, as does GNU grep. However, pcregrep was also printing all files when
1558        --files-with-matches was added. Now, when both options are given, it prints
1559        counts only for those files that have at least one match. (GNU grep just
1560        prints the file name in this circumstance, but including the count seems
1561        more useful - otherwise, why use --count?) Also ensured that the
1562        combination -clh just lists non-zero counts, with no names.
1563    
1564    4.  The long form of the pcregrep -F option was incorrectly implemented as
1565        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1566        but it seems right to fix it, and I didn't think it was worth preserving
1567        the old behaviour.
1568    
1569    5.  The command line items --regex=pattern and --regexp=pattern were not
1570        recognized by pcregrep, which required --regex pattern or --regexp pattern
1571        (with a space rather than an '='). The man page documented the '=' forms,
1572        which are compatible with GNU grep; these now work.
1573    
1574    6.  No libpcreposix.pc file was created for pkg-config; there was just
1575        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1576    
1577    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1578        when UCP support is not needed, by modifying the Python script that
1579        generates it from Unicode data files. This should not matter if the module
1580        is correctly used as a library, but I received one complaint about 50K of
1581        unwanted data. My guess is that the person linked everything into his
1582        program rather than using a library. Anyway, it does no harm.
1583    
1584    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1585        was a minimum greater than 1 for a wide character in a possessive
1586        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1587        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1588        character. Chaos in the form of incorrect output or a compiling loop could
1589        result.
1590    
1591    9.  The restrictions on what a pattern can contain when partial matching is
1592        requested for pcre_exec() have been removed. All patterns can now be
1593        partially matched by this function. In addition, if there are at least two
1594        slots in the offset vector, the offset of the earliest inspected character
1595        for the match and the offset of the end of the subject are set in them when
1596        PCRE_ERROR_PARTIAL is returned.
1597    
1598    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1599        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1600        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1601        and may be more useful for multi-segment matching.
1602    
1603    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1604        used to be given if ever the end of the subject was reached; now it is
1605        given only if matching could not proceed because another character was
1606        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1607        string "Z", which now yields "no match" instead of "partial match". In the
1608        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1609        final character ended with (*FAIL).
1610    
1611    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1612        if the pattern had a "must contain" character that was already found in the
1613        earlier partial match, unless partial matching was again requested. For
1614        example, with the pattern /dog.(body)?/, the "must contain" character is
1615        "g". If the first part-match was for the string "dog", restarting with
1616        "sbody" failed. This bug has been fixed.
1617    
1618    13. The string returned by pcre_dfa_exec() after a partial match has been
1619        changed so that it starts at the first inspected character rather than the
1620        first character of the match. This makes a difference only if the pattern
1621        starts with a lookbehind assertion or \b or \B (\K is not supported by
1622        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1623        matching functions compatible, and I think it's the right thing to do.
1624    
1625    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1626        so that the demonstration program is easily available in environments where
1627        PCRE has not been installed from source.
1628    
1629    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1630        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1631        library.
1632    
1633    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1634        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1635        is not the first non-POSIX option to be added. Clearly some people find
1636        these options useful.
1637    
1638    17. If a caller to the POSIX matching function regexec() passes a non-zero
1639        value for nmatch with a NULL value for pmatch, the value of
1640        nmatch is forced to zero.
1641    
1642    18. RunGrepTest did not have a test for the availability of the -u option of
1643        the diff command, as RunTest does. It now checks in the same way as
1644        RunTest, and also checks for the -b option.
1645    
1646    19. If an odd number of negated classes containing just a single character
1647        interposed, within parentheses, between a forward reference to a named
1648        subpattern and the definition of the subpattern, compilation crashed with
1649        an internal error, complaining that it could not find the referenced
1650        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1651        [The bug was that it was starting one character too far in when skipping
1652        over the character class, thus treating the ] as data rather than
1653        terminating the class. This meant it could skip too much.]
1654    
1655    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1656        /g option in pcretest when the pattern contains \K, which makes it possible
1657        to have an empty string match not at the start, even when the pattern is
1658        anchored. Updated pcretest and pcredemo to use this option.
1659    
1660    21. If the maximum number of capturing subpatterns in a recursion was greater
1661        than the maximum at the outer level, the higher number was returned, but
1662        with unset values at the outer level. The correct (outer level) value is
1663        now given.
1664    
1665    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1666        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1667        make it do so. The string so far is captured, making this feature
1668        compatible with Perl.
1669    
1670    23. The tests have been re-organized, adding tests 11 and 12, to make it
1671        possible to check the Perl 5.10 features against Perl 5.10.
1672    
1673    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1674        pattern matches a fixed length string. PCRE did not allow this; now it
1675        does. Neither allows recursion.
1676    
1677    25. I finally figured out how to implement a request to provide the minimum
1678        length of subject string that was needed in order to match a given pattern.
1679        (It was back references and recursion that I had previously got hung up
1680        on.) This code has now been added to pcre_study(); it finds a lower bound
1681        to the length of subject needed. It is not necessarily the greatest lower
1682        bound, but using it to avoid searching strings that are too short does give
1683        some useful speed-ups. The value is available to calling programs via
1684        pcre_fullinfo().
1685    
1686    26. While implementing 25, I discovered to my embarrassment that pcretest had
1687        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1688        study optimizations had never been tested with that matching function.
1689        Oops. What is worse, even when it was passed study data, there was a bug in
1690        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1691        were also very few tests of studied patterns with pcre_dfa_exec().
1692    
1693    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1694        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1695        on the other side of the coin, they are no longer allowed to have different
1696        names, because these cannot be distinguished in PCRE, and this has caused
1697        confusion. (This is a difference from Perl.)
1698    
1699    28. When duplicate subpattern names are present (necessarily with different
1700        numbers, as required by 27 above), and a test is made by name in a
1701        conditional pattern, either for a subpattern having been matched, or for
1702        recursion in such a pattern, all the associated numbered subpatterns are
1703        tested, and the overall condition is true if the condition is true for any
1704        one of them. This is the way Perl works, and is also more like the way
1705        testing by number works.
1706    
1707    
1708    Version 7.9 11-Apr-09
1709    ---------------------
1710    
1711    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1712        (pcretest), all targets were linked against these libraries. This included
1713        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1714        libraries. This caused unwanted dependencies to be created. This problem
1715        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1716        pcretest is linked with readline.
1717    
1718    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1719        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1720        moved outside it again, because FALSE and TRUE are already defined in AIX,
1721        but BOOL is not.
1722    
1723    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1724        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1725    
1726    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1727        hyphens) following file names and line numbers when outputting matching
1728        lines. This is not true; no spaces are inserted. I have also clarified the
1729        wording for the --colour (or --color) option.
1730    
1731    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1732        was not coloured; this is different to GNU grep, so I have changed it to be
1733        the same.
1734    
1735    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1736        each matching line was coloured. Now it goes on to look for further matches
1737        of any of the test patterns, which is the same behaviour as GNU grep.
1738    
1739    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1740        doesn't make sense to accept an empty string match in pcregrep, so I have
1741        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1742        seems to be how GNU grep behaves. [But see later change 40 for release
1743        8.33.]
1744    
1745    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1746        start or after a newline", because the conditional assertion was not being
1747        correctly handled. The rule now is that both the assertion and what follows
1748        in the first alternative must satisfy the test.
1749    
1750    9.  If auto-callout was enabled in a pattern with a conditional group whose
1751        condition was an assertion, PCRE could crash during matching, both with
1752        pcre_exec() and pcre_dfa_exec().
1753    
1754    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1755        used for matching.
1756    
1757    11. Unicode property support in character classes was not working for
1758        characters (bytes) greater than 127 when not in UTF-8 mode.
1759    
1760    12. Added the -M command line option to pcretest.
1761    
1762    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1763    
1764    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1765    
1766    16. Added comments and documentation about mis-use of no_arg in the C++
1767        wrapper.
1768    
1769    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1770        from Martin Jerabek that uses macro names for all relevant character and
1771        string constants.
1772    
1773    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1774        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1775        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1776        these, but not everybody uses configure.
1777    
1778    19. A conditional group that had only one branch was not being correctly
1779        recognized as an item that could match an empty string. This meant that an
1780        enclosing group might also not be so recognized, causing infinite looping
1781        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1782        with the subject "ab", where knowledge that the repeated group can match
1783        nothing is needed in order to break the loop.
1784    
1785    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1786        exec(), but without supplying a callout function, matching went wrong.
1787    
1788    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1789        leak if the size of the offset vector was greater than 30. When the vector
1790        is smaller, the saved offsets during recursion go onto a local stack
1791        vector, but for larger vectors malloc() is used. It was failing to free
1792        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1793        error, in fact).
1794    
1795    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1796        heapframe that is used only when UTF-8 support is enabled. This caused no
1797        problem, but was untidy.
1798    
1799    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1800        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1801        included within another project.
1802    
1803    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1804        slightly modified by me:
1805    
1806          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1807              not building pcregrep.
1808    
1809          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1810              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1811    
1812    25. Forward references, both numeric and by name, in patterns that made use of
1813        duplicate group numbers, could behave incorrectly or give incorrect errors,
1814        because when scanning forward to find the reference group, PCRE was not
1815        taking into account the duplicate group numbers. A pattern such as
1816        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1817    
1818    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1819        the feature of a custom pointer more persuasive (as requested by a user).
1820    
1821    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1822        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1823        pre-defined.
1824    
1825    28. Added support for (*UTF8) at the start of a pattern.
1826    
1827    29. Arrange for flags added by the "release type" setting in CMake to be shown
1828        in the configuration summary.
1829    
1830    
1831    Version 7.8 05-Sep-08
1832    ---------------------
1833    
1834    1.  Replaced UCP searching code with optimized version as implemented for Ad
1835        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1836        stage table and inline lookup instead of a function, giving speed ups of 2
1837        to 5 times on some simple patterns that I tested. Permission was given to
1838        distribute the MultiStage2.py script that generates the tables (it's not in
1839        the tarball, but is in the Subversion repository).
1840    
1841    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1842        scripts.
1843    
1844    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1845        a group with a zero qualifier. The result of the study could be incorrect,
1846        or the function might crash, depending on the pattern.
1847    
1848    4.  Caseless matching was not working for non-ASCII characters in back
1849        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1850        It now works when Unicode Property Support is available.
1851    
1852    5.  In pcretest, an escape such as \x{de} in the data was always generating
1853        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1854        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1855        truncation.
1856    
1857    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1858    
1859    7.  Added two (int) casts to pcregrep when printing the difference of two
1860        pointers, in case they are 64-bit values.
1861    
1862    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1863        test 2 if it fails.
1864    
1865    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1866        and a #define of that name to empty if it is not externally set. This is to
1867        allow users of MSVC to set it if necessary.
1868    
1869    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1870        the convenience functions in the pcre_get.c source file.
1871    
1872    11. An option change at the start of a pattern that had top-level alternatives
1873        could cause overwriting and/or a crash. This command provoked a crash in
1874        some environments:
1875    
1876          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1877    
1878        This potential security problem was recorded as CVE-2008-2371.
1879    
1880    12. For a pattern where the match had to start at the beginning or immediately
1881        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1882        pcre_dfa_exec() could read past the end of the passed subject if there was
1883        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1884        pcretest so that it places the subject at the end of its malloc-ed buffer.
1885    
1886    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1887        exec() might read past the end of the data buffer in UTF-8 mode.
1888    
1889    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1890        the data contained the byte 0x85 as part of a UTF-8 character within its
1891        first line. This applied both to normal and DFA matching.
1892    
1893    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1894        /^[^d]*?$/8 failed to match "abc".
1895    
1896    16. Added a missing copyright notice to pcrecpp_internal.h.
1897    
1898    17. Make it more clear in the documentation that values returned from
1899        pcre_exec() in ovector are byte offsets, not character counts.
1900    
1901    18. Tidied a few places to stop certain compilers from issuing warnings.
1902    
1903    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1904        supplied by Stefan Weber. I made a further small update for 7.8 because
1905        there is a change of source arrangements: the pcre_searchfuncs.c module is
1906        replaced by pcre_ucd.c.
1907    
1908    
1909    Version 7.7 07-May-08
1910    ---------------------
1911    
1912    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1913        a string to a long long, pretend we don't even have a long long." This is
1914        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1915    
1916    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1917        pre-7.6 versions, which defined a global no_arg variable instead of putting
1918        it in the RE class. (See also #8 below.)
1919    
1920    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1921        Lopes.
1922    
1923    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1924    
1925        (1) The include/exclude patterns were being applied to the whole pathnames
1926            of files, instead of just to the final components.
1927    
1928        (2) If there was more than one level of directory, the subdirectories were
1929            skipped unless they satisfied the include/exclude conditions. This is
1930            inconsistent with GNU grep (and could even be seen as contrary to the
1931            pcregrep specification - which I improved to make it absolutely clear).
1932            The action now is always to scan all levels of directory, and just
1933            apply the include/exclude patterns to regular files.
1934    
1935    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1936        --exclude_dir in the tests to avoid scanning .svn directories.
1937    
1938    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1939        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1940        doesn't support NULs in patterns.
1941    
1942    7.  Added some missing "const"s to declarations of static tables in
1943        pcre_compile.c and pcre_dfa_exec.c.
1944    
1945    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1946        caused by fix #2  above. (Subsequently also a second patch to fix the
1947        first patch. And a third patch - this was a messy problem.)
1948    
1949    9.  Applied Craig's patch to remove the use of push_back().
1950    
1951    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1952        matching function regexec().
1953    
1954    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1955        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1956        references. PCRE supports relative numbers with this syntax (I don't think
1957        Oniguruma does).
1958    
1959    12. Previously, a group with a zero repeat such as (...){0} was completely
1960        omitted from the compiled regex. However, this means that if the group
1961        was called as a subroutine from elsewhere in the pattern, things went wrong
1962        (an internal error was given). Such groups are now left in the compiled
1963        pattern, with a new opcode that causes them to be skipped at execution
1964        time.
1965    
1966    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1967        to the way PCRE behaves:
1968    
1969        (a) A lone ] character is dis-allowed (Perl treats it as data).
1970    
1971        (b) A back reference to an unmatched subpattern matches an empty string
1972            (Perl fails the current match path).
1973    
1974        (c) A data ] in a character class must be notated as \] because if the
1975            first data character in a class is ], it defines an empty class. (In
1976            Perl it is not possible to have an empty class.) The empty class []
1977            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1978            The negative empty class [^] matches any one character, independently
1979            of the DOTALL setting.
1980    
1981    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1982        non-existent subpattern following a character class starting with ']' and
1983        containing () gave an internal compiling error instead of "reference to
1984        non-existent subpattern". Fortunately, when the pattern did exist, the
1985        compiled code was correct. (When scanning forwards to check for the
1986        existence of the subpattern, it was treating the data ']' as terminating
1987        the class, so got the count wrong. When actually compiling, the reference
1988        was subsequently set up correctly.)
1989    
1990    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1991        it was being rejected as not supported by pcre_dfa_exec(), even though
1992        other assertions are supported. I have made pcre_dfa_exec() support
1993        (*FAIL).
1994    
1995    16. The implementation of 13c above involved the invention of a new opcode,
1996        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1997        cannot be changed at match time, I realized I could make a small
1998        improvement to matching performance by compiling OP_ALLANY instead of
1999        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2000        on the OP_ANY path.
2001    
2002    17. Compiling pcretest on Windows with readline support failed without the
2003        following two fixes: (1) Make the unistd.h include conditional on
2004        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2005    
2006    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2007        ncurses library to be included for pcretest when ReadLine support is
2008        requested, but also to allow for it to be overridden. This patch came from
2009        Daniel Bergström.
2010    
2011    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2012        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2013        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2014        spotting this.
2015    
2016    
2017    Version 7.6 28-Jan-08
2018    ---------------------
2019    
2020    1.  A character class containing a very large number of characters with
2021        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2022        overflow.
2023    
2024    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2025        HAVE_LONG_LONG is not defined.
2026    
2027    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2028        bring them up to date and include new features. This patch includes:
2029    
2030        - Fixed PH's badly added libz and libbz2 support.
2031        - Fixed a problem with static linking.
2032        - Added pcredemo. [But later removed - see 7 below.]
2033        - Fixed dftables problem and added an option.
2034        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2035            HAVE_LONG_LONG.
2036        - Added readline support for pcretest.
2037        - Added an listing of the option settings after cmake has run.
2038    
2039    4.  A user submitted a patch to Makefile that makes it easy to create
2040        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2041        Makefile.am that cause it to include this special target, without
2042        affecting anything else. Note that the same mingw target plus all
2043        the other distribution libraries and programs are now supported
2044        when configuring with CMake (see 6 below) instead of with
2045        Configure/Make.
2046    
2047    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2048        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2049        exported in the Windows port". It has not yet been confirmed that the patch
2050        solves the problem, but it does no harm.
2051    
2052    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2053        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2054        with CMake, and also correct the comment about stack recursion.
2055    
2056    7.  Remove the automatic building of pcredemo from the ./configure system and
2057        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2058        of a program that users should build themselves after PCRE is installed, so
2059        building it automatically is not really right. What is more, it gave
2060        trouble in some build environments.
2061    
2062    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2063    
2064    
2065    Version 7.5 10-Jan-08
2066    ---------------------
2067    
2068    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2069        values in parens when parsing an RE using the C++ wrapper."
2070    
2071    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2072        Characters greater than 255 were excluded from the class instead of being
2073        included.
2074    
2075    3.  The same bug as (2) above applied to negated POSIX classes such as
2076        [:^space:].
2077    
2078    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2079        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2080        I have changed it.
2081    
2082    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2083        first named subpattern) and a construct such as (?&a) would reference the
2084        first named subpattern whose name started with "a" (in other words, the
2085        length check was missing). Both these problems are fixed. "Subpattern name
2086        expected" is now given for (?&) (a zero-length name), and this patch also
2087        makes it give the same error for \k'' (previously it complained that that
2088        was a reference to a non-existent subpattern).
2089    
2090    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2091        this is right because (?- can be followed by option settings as well as by
2092        digits. I have, however, made the messages clearer.
2093    
2094    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2095        than the number used in the conditional) now cause a compile-time error.
2096        This is actually not compatible with Perl, which accepts such patterns, but
2097        treats the conditional as always being FALSE (as PCRE used to), but it
2098        seems to me that giving a diagnostic is better.
2099    
2100    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2101        and messages.
2102    
2103    9.  Fix two occurrences of "backslash" in comments that should have been
2104        "backspace".
2105    
2106    10. Remove two redundant lines of code that can never be obeyed (their function
2107        was moved elsewhere).
2108    
2109    11. The program that makes PCRE's Unicode character property table had a bug
2110        which caused it to generate incorrect table entries for sequences of
2111        characters that have the same character type, but are in different scripts.
2112        It amalgamated them into a single range, with the script of the first of
2113        them. In other words, some characters were in the wrong script. There were
2114        thirteen such cases, affecting characters in the following ranges:
2115    
2116          U+002b0 - U+002c1
2117          U+0060c - U+0060d
2118          U+0061e - U+00612
2119          U+0064b - U+0065e
2120          U+0074d - U+0076d
2121          U+01800 - U+01805
2122          U+01d00 - U+01d77
2123          U+01d9b - U+01dbf
2124          U+0200b - U+0200f
2125          U+030fc - U+030fe
2126          U+03260 - U+0327f
2127          U+0fb46 - U+0fbb1
2128          U+10450 - U+1049d
2129    
2130    12. The -o option (show only the matching part of a line) for pcregrep was not
2131        compatible with GNU grep in that, if there was more than one match in a
2132        line, it showed only the first of them. It now behaves in the same way as
2133        GNU grep.
2134    
2135    13. If the -o and -v options were combined for pcregrep, it printed a blank
2136        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2137        does the same. The return code can be used to tell if there were any
2138        non-matching lines.
2139    
2140    14. Added --file-offsets and --line-offsets to pcregrep.
2141    
2142    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2143        infinitely looping recursion. The bug was that positive lookaheads were not
2144        being skipped when checking for a possible empty match (negative lookaheads
2145        and both kinds of lookbehind were skipped).
2146    
2147    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2148        inclusion of <windows.h> to before rather than after the definition of
2149        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2150    
2151    17. Specifying a possessive quantifier with a specific limit for a Unicode
2152        character property caused pcre_compile() to compile bad code, which led at
2153        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2154        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2155        caused the error; without that there was no problem.
2156    
2157    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2158    
2159    19. Added --enable-pcretest-libreadline.
2160    
2161    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2162        RE::GlobalReplace(). As a result, the number of replacements returned was
2163        double what it should be. I removed one of the increments, but Craig sent a
2164        later patch that removed the other one (the right fix) and added unit tests
2165        that check the return values (which was not done before).
2166    
2167    21. Several CMake things:
2168    
2169        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2170            the names libpcre and libpcreposix, not just pcre and pcreposix.
2171    
2172        (2) The above change means that pcretest and pcregrep are now correctly
2173            linked with the newly-built libraries, not previously installed ones.
2174    
2175        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2176    
2177    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2178        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2179        UTF-8 newline character). The key issue is that the pattern starts .*;
2180        this means that the match must be either at the beginning, or after a
2181        newline. The bug was in the code for advancing after a failed match and
2182        checking that the new position followed a newline. It was not taking
2183        account of UTF-8 characters correctly.
2184    
2185    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2186        character classes. PCRE was not treating the sequence [:...:] as a
2187        character class unless the ... were all letters. Perl, however, seems to
2188        allow any characters between [: and :], though of course it rejects as
2189        unknown any "names" that contain non-letters, because all the known class
2190        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2191        for example, whereas PCRE did not - it did not recognize a POSIX character
2192        class. This seemed a bit dangerous, so the code has been changed to be
2193        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2194        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2195        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2196        Perl does, and where it didn't before.
2197    
2198    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2199        Windows environments %n is disabled by default.
2200    
2201    
2202    Version 7.4 21-Sep-07
2203    ---------------------
2204    
2205    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2206        means that a class such as [\s] counted as "explicit reference to CR or
2207        LF". That isn't really right - the whole point of the change was to try to
2208        help when there was an actual mention of one of the two characters. So now
2209        the change happens only if \r or \n (or a literal CR or LF) character is
2210        encountered.
2211    
2212    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2213        of both had grown to the point where there were only 3 bits left.
2214        Fortunately, there was spare space in the data structure, and so I have
2215        moved the internal flags into a new 16-bit field to free up more option
2216        bits.
2217    
2218    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2219        but did not set the internal JCHANGED flag - either of these is enough to
2220        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2221        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2222        start sets both bits.
2223    
2224    4.  Added options (at build time, compile time, exec time) to change \R from
2225        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2226    
2227    5.  doc/pcresyntax.html was missing from the distribution.
2228    
2229    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2230        compatibility, even though it is no longer used.
2231    
2232    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2233        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2234        windows.h file is present (where different names are used). [This was
2235        reversed later after testing - see 16 below.]
2236    
2237    8.  Changed all #include <config.h> to #include "config.h". There were also
2238        some further <pcre.h> cases that I changed to "pcre.h".
2239    
2240    9.  When pcregrep was used with the --colour option, it missed the line ending
2241        sequence off the lines that it output.
2242    
2243    10. It was pointed out to me that arrays of string pointers cause lots of
2244        relocations when a shared library is dynamically loaded. A technique of
2245        using a single long string with a table of offsets can drastically reduce
2246        these. I have refactored PCRE in four places to do this. The result is
2247        dramatic:
2248    
2249          Originally:                          290
2250          After changing UCP table:            187
2251          After changing error message table:   43
2252          After changing table of "verbs"       36
2253          After changing table of Posix names   22
2254    
2255        Thanks to the folks working on Gregex for glib for this insight.
2256    
2257    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2258        unicode-properties was also set.
2259    
2260    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2261    
2262    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2263        checked only for CRLF.
2264    
2265    14. Added casts to pcretest.c to avoid compiler warnings.
2266    
2267    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2268    
2269    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2270        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2271        entirely. This removes changes made in 7 above.
2272    
2273    17. The CMake files have been updated, and there is now more information about
2274        building with CMake in the NON-UNIX-USE document.
2275    
2276    
2277    Version 7.3 28-Aug-07
2278    ---------------------
2279    
2280     1. In the rejigging of the build system that eventually resulted in 7.1, the
2281        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2282        brackets there is not right, since it causes compilers to look for an
2283        installed pcre.h, not the version that is in the source that is being
2284        compiled (which of course may be different). I have changed it back to:
2285    
2286          #include "pcre.h"
2287    
2288        I have a vague recollection that the change was concerned with compiling in
2289        different directories, but in the new build system, that is taken care of
2290        by the VPATH setting the Makefile.
2291    
2292     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2293        when the subject happened to end in the byte 0x85 (e.g. if the last
2294        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2295        characters but of course it shouldn't be taken as a newline when it is part
2296        of another character. The bug was that, for an unlimited repeat of . in
2297        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2298        characters when looking for a newline.
2299    
2300     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2301    
2302     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2303        in debug output.
2304    
2305     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2306        long printing in the pcrecpp unittest when running under MinGW.
2307    
2308     6. ESC_K was left out of the EBCDIC table.
2309    
2310     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2311        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2312        limit also applies to "virtual nesting" when a pattern is recursive, and in
2313        this case 1000 isn't so big. I have been able to remove this limit at the
2314        expense of backing off one optimization in certain circumstances. Normally,
2315        when pcre_exec() would call its internal match() function recursively and
2316        immediately return the result unconditionally, it uses a "tail recursion"
2317        feature to save stack. However, when a subpattern that can match an empty
2318        string has an unlimited repetition quantifier, it no longer makes this
2319        optimization. That gives it a stack frame in which to save the data for
2320        checking that an empty string has been matched. Previously this was taken
2321        from the 1000-entry workspace that had been reserved. So now there is no
2322        explicit limit, but more stack is used.
2323    
2324     8. Applied Daniel's patches to solve problems with the import/export magic
2325        syntax that is required for Windows, and which was going wrong for the
2326        pcreposix and pcrecpp parts of the library. These were overlooked when this
2327        problem was solved for the main library.
2328    
2329     9. There were some crude static tests to avoid integer overflow when computing
2330        the size of patterns that contain repeated groups with explicit upper
2331        limits. As the maximum quantifier is 65535, the maximum group length was
2332        set at 30,000 so that the product of these two numbers did not overflow a
2333        32-bit integer. However, it turns out that people want to use groups that
2334        are longer than 30,000 bytes (though not repeat them that many times).
2335        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2336        made it possible to implement the integer overflow checks in a much more
2337        dynamic way, which I have now done. The artificial limitation on group
2338        length has been removed - we now have only the limit on the total length of
2339        the compiled pattern, which depends on the LINK_SIZE setting.
2340    
2341    10. Fixed a bug in the documentation for get/copy named substring when
2342        duplicate names are permitted. If none of the named substrings are set, the
2343        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2344        empty string.
2345    
2346    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2347        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2348        because the ] is interpreted as the first data character and the
2349        terminating ] is not found. PCRE has been made compatible with Perl in this
2350        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2351        cause memory overwriting.
2352    
2353    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2354        string has been matched (to stop an infinite loop). It was not recognizing
2355        a conditional subpattern that could match an empty string if that
2356        subpattern was within another subpattern. For example, it looped when
2357        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2358        condition was not nested. This bug has been fixed.
2359    
2360    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2361        past the start of the subject in the presence of bytes with the top bit
2362        set, for example "\x8aBCD".
2363    
2364    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2365        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2366    
2367    14. Optimized (?!) to (*FAIL).
2368    
2369    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2370        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2371        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2372        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2373        does: it's just the validity check that is more restrictive.
2374    
2375    16. Inserted checks for integer overflows during escape sequence (backslash)
2376        processing, and also fixed erroneous offset values for syntax errors during
2377        backslash processing.
2378    
2379    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2380        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2381    
2382    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2383        caused an overrun.
2384    
2385    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2386        something other than just ASCII characters) inside a group that had an
2387        unlimited repeat caused a loop at compile time (while checking to see
2388        whether the group could match an empty string).
2389    
2390    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2391        [\P{Any}] did so. (Error in the code for printing property names.)
2392    
2393    21. An orphan \E inside a character class could cause a crash.
2394    
2395    22. A repeated capturing bracket such as (A)? could cause a wild memory
2396        reference during compilation.
2397    
2398    23. There are several functions in pcre_compile() that scan along a compiled
2399        expression for various reasons (e.g. to see if it's fixed length for look
2400        behind). There were bugs in these functions when a repeated \p or \P was
2401        present in the pattern. These operators have additional parameters compared
2402        with \d, etc, and these were not being taken into account when moving along
2403        the compiled data. Specifically:
2404    
2405        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2406            length.
2407    
2408        (b) An item such as \pL+ within a repeated group could cause crashes or
2409            loops.
2410    
2411        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2412            "reference to non-existent subpattern" error.
2413    
2414        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2415    
2416    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2417        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2418    
2419    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2420    
2421    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2422        character were causing crashes (broken optimization).
2423    
2424    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2425        \p or \P) caused a compile-time loop.
2426    
2427    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2428        break. For example, the unstudied pattern [\r\n]A does not match the string
2429        "\r\nA" because change 7.0/46 below moves the current point on by two
2430        characters after failing to match at the start. However, the pattern \nA
2431        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2432        the same is true. There doesn't seem any very clean way out of this, but
2433        what I have chosen to do makes the common cases work: PCRE now takes note
2434        of whether there can be an explicit match for \r or \n anywhere in the
2435        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2436        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2437        pattern has explicit CR or LF references.
2438    
2439    29. Added (*CR) etc for changing newline setting at start of pattern.
2440    
2441    
2442    Version 7.2 19-Jun-07
2443    ---------------------
2444    
2445     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2446        which is apparently normally available under Windows.
2447    
2448     2. Re-jig the pcregrep tests with different newline settings in an attempt
2449        to make them independent of the local environment's newline setting.
2450    
2451     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2452    
2453     4. Some of the "internals" tests were previously cut out when the link size
2454        was not 2, because the output contained actual offsets. The recent new
2455        "Z" feature of pcretest means that these can be cut out, making the tests
2456        usable with all link sizes.
2457    
2458     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2459        stack recursion. This gives a massive performance boost under BSD, but just
2460        a small improvement under Linux. However, it saves one field in the frame
2461        in all cases.
2462    
2463     6. Added more features from the forthcoming Perl 5.10:
2464    
2465        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2466            recursion call. It refers to the nth most recently opened parentheses.
2467    
2468        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2469            to be opened parentheses.
2470    
2471        (c) Conditions that refer to capturing parentheses can be specified
2472            relatively, for example, (?(-2)... or (?(+3)...
2473    
2474        (d) \K resets the start of the current match so that everything before
2475            is not part of it.
2476    
2477        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2478    
2479        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2480            reference syntax.
2481    
2482        (g) (?| introduces a group in which the numbering of parentheses in each
2483            alternative starts with the same number.
2484    
2485        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2486    
2487     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2488        PCRE_INFO_JCHANGED.
2489    
2490     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2491        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2492        for detecting groups that can match an empty string.
2493    
2494     9. A pattern with a very large number of alternatives (more than several
2495        hundred) was running out of internal workspace during the pre-compile
2496        phase, where pcre_compile() figures out how much memory will be needed. A
2497        bit of new cunning has reduced the workspace needed for groups with
2498        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2499        workspace instead of running out of the 4096 that are available.
2500    
2501    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2502    
2503    11. Applied patch from Google to remove an optimization that didn't quite work.
2504        The report of the bug said:
2505    
2506          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2507          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2508          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2509    
2510    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2511        it matched the wrong number of bytes.
2512    
2513    
2514  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
2515  ---------------------  ---------------------
2516    

Legend:
Removed from v.155  
changed lines
  Added in v.1354

  ViewVC Help
Powered by ViewVC 1.1.5