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

Diff of /code/trunk/ChangeLog

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

revision 143 by ph10, Mon Apr 2 10:08:14 2007 UTC revision 1353 by ph10, Mon Aug 5 16:24:02 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.1 12-Mar-07  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    
65    Version 8.33 28-May-2013
66    --------------------------
67    
68    1.  Added 'U' to some constants that are compared to unsigned integers, to
69        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
70        variables that are added to signed variables, to ensure the result is
71        signed and can be negated.
72    
73    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
74        CMake config files.
75    
76    3.  Revise the creation of config.h.generic so that all boolean macros are
77        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
78        overriding via -D on the command line possible.
79    
80    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
81        to unsigned int is reported to make a quite noticeable speed difference in
82        a specific Windows environment. Testing on Linux did also appear to show
83        some benefit (and it is clearly not harmful). Also fixed the definition of
84        Xop which should be unsigned.
85    
86    5.  Related to (4), changing the definition of the intermediate variable cc
87        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
88        improvements.
89    
90    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
91        unnecessary spaces.
92    
93    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
94        and later.
95    
96    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
97    
98    9.  Optimizing fast_forward_start_bits in JIT.
99    
100    10. Adding support for callouts in JIT, and fixing some issues revealed
101        during this work. Namely:
102    
103        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
104    
105        (b) Minimum length was not checked before the matching is started.
106    
107    11. The value of capture_last that is passed to callouts was incorrect in some
108        cases when there was a capture on one path that was subsequently abandoned
109        after a backtrack. Also, the capture_last value is now reset after a
110        recursion, since all captures are also reset in this case.
111    
112    12. The interpreter no longer returns the "too many substrings" error in the
113        case when an overflowing capture is in a branch that is subsequently
114        abandoned after a backtrack.
115    
116    13. In the pathological case when an offset vector of size 2 is used, pcretest
117        now prints out the matched string after a yield of 0 or 1.
118    
119    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
120        Only supported by the JIT compiler at the moment.
121    
122    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
123    
124    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
125        offset of the starting point of the matching process, provided the offsets
126        vector is large enough.
127    
128    17. The \A escape now records a lookbehind value of 1, though its execution
129        does not actually inspect the previous character. This is to ensure that,
130        in partial multi-segment matching, at least one character from the old
131        segment is retained when a new segment is processed. Otherwise, if there
132        are no lookbehinds in the pattern, \A might match incorrectly at the start
133        of a new segment.
134    
135    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
136    
137    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
138        gives some modest performance improvement in 8-bit mode.
139    
140    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
141        be expressed in certain programming languages using Universal Character
142        Names.
143    
144    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
145        which points out that "non characters" are not "characters that may not
146        appear in Unicode strings" but rather "characters that are reserved for
147        internal use and have only local meaning".
148    
149    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
150        there was a conditional group that depended on an assertion, if the
151        assertion was false, the callout that immediately followed the alternation
152        in the condition was skipped when pcre_exec() was used for matching.
153    
154    23. Allow an explicit callout to be inserted before an assertion that is the
155        condition for a conditional group, for compatibility with automatic
156        callouts, which always insert a callout at this point.
157    
158    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
159        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
160    
161    25. (*PRUNE) is now supported by the JIT compiler.
162    
163    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
164    
165    27. Fix the case where there are two or more SKIPs with arguments that may be
166        ignored.
167    
168    28. (*SKIP) is now supported by the JIT compiler.
169    
170    29. (*THEN) is now supported by the JIT compiler.
171    
172    30. Update RunTest with additional test selector options.
173    
174    31. The way PCRE handles backtracking verbs has been changed in two ways.
175    
176        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
177        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
178        backtracking. In some cases this makes it more Perl-compatible, but Perl's
179        rather obscure rules do not always do the same thing.
180    
181        (2) Previously, backtracking verbs were confined within assertions. This is
182        no longer the case for positive assertions, except for (*ACCEPT). Again,
183        this sometimes improves Perl compatibility, and sometimes does not.
184    
185    32. A number of tests that were in test 2 because Perl did things differently
186        have been moved to test 1, because either Perl or PCRE has changed, and
187        these tests are now compatible.
188    
189    32. Backtracking control verbs are now handled in the same way in JIT and
190        interpreter.
191    
192    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
193        contained a forward subroutine reference caused a compile error.
194    
195    34. Auto-detect and optimize limited repetitions in JIT.
196    
197    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
198        blocking (*UTF) etc.
199    
200    36. In the interpreter, maximizing pattern repetitions for characters and
201        character types now use tail recursion, which reduces stack usage.
202    
203    37. The value of the max lookbehind was not correctly preserved if a compiled
204        and saved regex was reloaded on a host of different endianness.
205    
206    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
207        of the compiled pattern block, expand the flags field from 16 to 32 bits
208        because it was almost full.
209    
210    39. Try madvise first before posix_madvise.
211    
212    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
213        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
214        original change locked out all matches of empty strings. This has been
215        changed so that one match of an empty string per line is recognized.
216        Subsequent searches on the same line (for colouring or for --only-matching,
217        for example) do not recognize empty strings.
218    
219    41. Applied a user patch to fix a number of spelling mistakes in comments.
220    
221    42. Data lines longer than 65536 caused pcretest to crash.
222    
223    43. Clarified the data type for length and startoffset arguments for pcre_exec
224        and pcre_dfa_exec in the function-specific man pages, where they were
225        explicitly stated to be in bytes, never having been updated. I also added
226        some clarification to the pcreapi man page.
227    
228    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
229        a segmentation fault.
230    
231    
232    Version 8.32 30-November-2012
233    -----------------------------
234    
235    1.  Improved JIT compiler optimizations for first character search and single
236        character iterators.
237    
238    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
239        Patch by Daniel Richard G.
240    
241    3.  Single character iterator optimizations in the JIT compiler.
242    
243    4.  Improved JIT compiler optimizations for character ranges.
244    
245    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
246        Reported by Giuseppe D'Angelo.
247    
248    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
249        of a line, was being set incorrectly in cases where .* appeared inside
250        atomic brackets at the start of a pattern, or where there was a subsequent
251        *PRUNE or *SKIP.
252    
253    7.  Improved instruction cache flush for POWER/PowerPC.
254        Patch by Daniel Richard G.
255    
256    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
257        grep:
258    
259        (a) There is now no limit to the number of patterns to be matched.
260    
261        (b) An error is given if a pattern is too long.
262    
263        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
264            are now supported.
265    
266        (d) --exclude-from and --include-from (multiple use) have been added.
267    
268        (e) Exclusions and inclusions now apply to all files and directories, not
269            just to those obtained from scanning a directory recursively.
270    
271        (f) Multiple uses of -f and --file-list are now supported.
272    
273        (g) In a Windows environment, the default for -d has been changed from
274            "read" (the GNU grep default) to "skip", because otherwise the presence
275            of a directory in the file list provokes an error.
276    
277        (h) The documentation has been revised and clarified in places.
278    
279    9.  Improve the matching speed of capturing brackets.
280    
281    10. Changed the meaning of \X so that it now matches a Unicode extended
282        grapheme cluster.
283    
284    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
285        out POSIX threads when JIT support is configured.
286    
287    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
288    
289    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
290        structure could go wrong in environments where size_t is not the same size
291        as int.
292    
293    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
294        set.
295    
296    15. The EBCDIC support had decayed; later updates to the code had included
297        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
298        general tidy up of EBCDIC-related issues, and the documentation was also
299        not quite right. There is now a test that can be run on ASCII systems to
300        check some of the EBCDIC-related things (but is it not a full test).
301    
302    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
303        in a small tidy to the code.
304    
305    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
306    
307    18. If the --only-matching (-o) option in pcregrep is specified multiple
308        times, each one causes appropriate output. For example, -o1 -o2 outputs the
309        substrings matched by the 1st and 2nd capturing parentheses. A separating
310        string can be specified by --om-separator (default empty).
311    
312    19. Improving the first n character searches.
313    
314    20. Turn case lists for horizontal and vertical white space into macros so that
315        they are defined only once.
316    
317    21. This set of changes together give more compatible Unicode case-folding
318        behaviour for characters that have more than one other case when UCP
319        support is available.
320    
321        (a) The Unicode property table now has offsets into a new table of sets of
322            three or more characters that are case-equivalent. The MultiStage2.py
323            script that generates these tables (the pcre_ucd.c file) now scans
324            CaseFolding.txt instead of UnicodeData.txt for character case
325            information.
326    
327        (b) The code for adding characters or ranges of characters to a character
328            class has been abstracted into a generalized function that also handles
329            case-independence. In UTF-mode with UCP support, this uses the new data
330            to handle characters with more than one other case.
331    
332        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
333            whose other case is greater than 256 are now correctly matched
334            caselessly. Previously, the high codepoint matched the low one, but not
335            vice versa.
336    
337        (d) The processing of \h, \H, \v, and \ in character classes now makes use
338            of the new class addition function, using character lists defined as
339            macros alongside the case definitions of 20 above.
340    
341        (e) Caseless back references now work with characters that have more than
342            one other case.
343    
344        (f) General caseless matching of characters with more than one other case
345            is supported.
346    
347    22. Unicode character properties were updated from Unicode 6.2.0
348    
349    23. Improved CMake support under Windows. Patch by Daniel Richard G.
350    
351    24. Add support for 32-bit character strings, and UTF-32
352    
353    25. Major JIT compiler update (code refactoring and bugfixing).
354        Experimental Sparc 32 support is added.
355    
356    26. Applied a modified version of Daniel Richard G's patch to create
357        pcre.h.generic and config.h.generic by "make" instead of in the
358        PrepareRelease script.
359    
360    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
361        pcre_compile.c when checking for a zero character.
362    
363    28. Introducing a native interface for JIT. Through this interface, the compiled
364        machine code can be directly executed. The purpose of this interface is to
365        provide fast pattern matching, so several sanity checks are not performed.
366        However, feature tests are still performed. The new interface provides
367        1.4x speedup compared to the old one.
368    
369    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
370        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
371        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
372    
373    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
374        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
375        the "old" RFC 2279). Instead, it ended up passing a negative length to
376        pcre_exec().
377    
378    31. Add support for GCC's visibility feature to hide internal functions.
379    
380    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
381        "unknown -C option" after outputting 0 or 1.
382    
383    33. There is now support for generating a code coverage report for the test
384        suite in environments where gcc is the compiler and lcov is installed. This
385        is mainly for the benefit of the developers.
386    
387    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
388        unaddressable using valgrind annotations, allowing valgrind to detect
389        invalid memory accesses. This is mainly for the benefit of the developers.
390    
391    25. (*UTF) can now be used to start a pattern in any of the three libraries.
392    
393    26. Give configure error if --enable-cpp but no C++ compiler found.
394    
395    
396    Version 8.31 06-July-2012
397    -------------------------
398    
399    1.  Fixing a wrong JIT test case and some compiler warnings.
400    
401    2.  Removed a bashism from the RunTest script.
402    
403    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
404        to unsigned type, result still unsigned" that was given by an MS compiler
405        on encountering the code "-sizeof(xxx)".
406    
407    4.  Partial matching support is added to the JIT compiler.
408    
409    5.  Fixed several bugs concerned with partial matching of items that consist
410        of more than one character:
411    
412        (a) /^(..)\1/ did not partially match "aba" because checking references was
413            done on an "all or nothing" basis. This also applied to repeated
414            references.
415    
416        (b) \R did not give a hard partial match if \r was found at the end of the
417            subject.
418    
419        (c) \X did not give a hard partial match after matching one or more
420            characters at the end of the subject.
421    
422        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
423            a partial match for the string "\r".
424    
425        (e) When newline was set to CRLF, the metacharacter "." did not recognize
426            a partial match for a CR character at the end of the subject string.
427    
428    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
429        running pcretest, the text "(JIT)" added to the output whenever JIT is
430        actually used to run the match.
431    
432    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
433        or /S+[+] with a digit between 1 and 7.
434    
435    8.  OP_NOT now supports any UTF character not just single-byte ones.
436    
437    9.  (*MARK) control verb is now supported by the JIT compiler.
438    
439    10. The command "./RunTest list" lists the available tests without actually
440        running any of them. (Because I keep forgetting what they all are.)
441    
442    11. Add PCRE_INFO_MAXLOOKBEHIND.
443    
444    12. Applied a (slightly modified) user-supplied patch that improves performance
445        when the heap is used for recursion (compiled with --disable-stack-for-
446        recursion). Instead of malloc and free for each heap frame each time a
447        logical recursion happens, frames are retained on a chain and re-used where
448        possible. This sometimes gives as much as 30% improvement.
449    
450    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
451        call.
452    
453    14. As documented, (*COMMIT) is now confined to within a positive assertion.
454    
455    15. It is now possible to link pcretest with libedit as an alternative to
456        libreadline.
457    
458    16. (*COMMIT) control verb is now supported by the JIT compiler.
459    
460    17. The Unicode data tables have been updated to Unicode 6.1.0.
461    
462    18. Added --file-list option to pcregrep.
463    
464    19. Added binary file support to pcregrep, including the -a, --binary-files,
465        -I, and --text options.
466    
467    20. The madvise function is renamed for posix_madvise for QNX compatibility
468        reasons. Fixed by Giuseppe D'Angelo.
469    
470    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
471    
472    22. Changed the option for creating symbolic links for 16-bit man pages from
473        -s to -sf so that re-installing does not cause issues.
474    
475    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
476    
477    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
478        matches in certain environments (the workspace was not being correctly
479        retained). Also added to pcre_dfa_exec() a simple plausibility check on
480        some of the workspace data at the beginning of a restart.
481    
482    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
483        was not doing so when it should - probably a typo introduced by SVN 528
484        (change 8.10/14).
485    
486    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
487        \w+ when the character tables indicated that \x{c4} was a word character.
488        There were several related cases, all because the tests for doing a table
489        lookup were testing for characters less than 127 instead of 255.
490    
491    27. If a pattern contains capturing parentheses that are not used in a match,
492        their slots in the ovector are set to -1. For those that are higher than
493        any matched groups, this happens at the end of processing. In the case when
494        there were back references that the ovector was too small to contain
495        (causing temporary malloc'd memory to be used during matching), and the
496        highest capturing number was not used, memory off the end of the ovector
497        was incorrectly being set to -1. (It was using the size of the temporary
498        memory instead of the true size.)
499    
500    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
501        ovector size, it uses memory at the end of the block that it has got.
502    
503    29. Check for an overlong MARK name and give an error at compile time. The
504        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
505    
506    30. JIT compiler update.
507    
508    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
509        Rill for the patch.
510    
511    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
512    
513    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
514    
515    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
516        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
517        was enabled.
518    
519    35. Improve JIT code generation for greedy plus quantifier.
520    
521    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
522        1 to "aa" instead of to an empty string. The bug affected repeated groups
523        that could potentially match an empty string.
524    
525    37. Optimizing single character iterators in JIT.
526    
527    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
528        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
529        codepoints that are too big for the mode are faulted, and in a UTF mode,
530        disallowed codepoints are also faulted.
531    
532    39. If PCRE was compiled with UTF support, in three places in the DFA
533        matcher there was code that should only have been obeyed in UTF mode, but
534        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
535        processing when bytes with values greater than 127 were present. In 16-bit
536        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
537        both cases the values are those that cannot be the first data item in a UTF
538        character. The three items that might have provoked this were recursions,
539        possessively repeated groups, and atomic groups.
540    
541    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
542        and pcregrep, because some OS require shared objects to be explicitly
543        passed to ld, causing the link step to fail if they are not.
544    
545    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
546        mode, patterns that started with \h* or \R* might be incorrectly matched.
547    
548    
549    Version 8.30 04-February-2012
550    -----------------------------
551    
552    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
553        name is defined in ctype.h.
554    
555    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
556        only in quite long subpatterns.
557    
558    3.  Removed the function pcre_info(), which has been obsolete and deprecated
559        since it was replaced by pcre_fullinfo() in February 2000.
560    
561    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
562        match a (*MARK), and the match failed at the start of the subject, a
563        reference to memory before the start of the subject could occur. This bug
564        was introduced by fix 17 of release 8.21.
565    
566    5.  A reference to an unset group with zero minimum repetition was giving
567        totally wrong answers (in non-JavaScript-compatibility mode). For example,
568        /(another)?(\1?)test/ matched against "hello world test". This bug was
569        introduced in release 8.13.
570    
571    6.  Add support for 16-bit character strings (a large amount of work involving
572        many changes and refactorings).
573    
574    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
575        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
576        from a file.
577    
578    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
579        rounding is not applied in this particular case).
580    
581    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
582        if they appear, or are escaped, in patterns.
583    
584    10. Get rid of a number of -Wunused-but-set-variable warnings.
585    
586    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
587        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
588        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
589        also returns the mark "x". This bug applied to capturing parentheses,
590        non-capturing parentheses, and atomic parentheses. It also applied to some
591        assertions.
592    
593    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
594        information out of configure.ac instead of relying on pcre.h.generic, which
595        is not stored in the repository.
596    
597    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
598        -lreadline.
599    
600    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
601    
602    15. Applied Graycode's patch to put the top-level frame on the stack rather
603        than the heap when not using the stack for recursion. This gives a
604        performance improvement in many cases when recursion is not deep.
605    
606    16. Experimental code added to "pcretest -C" to output the stack frame size.
607    
608    
609    Version 8.21 12-Dec-2011
610    ------------------------
611    
612    1.  Updating the JIT compiler.
613    
614    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
615        are added as well.
616    
617    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
618        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
619        calling _pcre_jit_exec. Some extra comments are added.
620    
621    4.  (*MARK) settings inside atomic groups that do not contain any capturing
622        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
623        was introduced by change 18 for 8.20.
624    
625    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
626        ECMA-262 standard.
627    
628    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
629        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
630        This bug was probably introduced by change 9 of 8.13.
631    
632    7.  While fixing 6 above, I noticed that a number of other items were being
633        incorrectly rejected as "not fixed length". This arose partly because newer
634        opcodes had not been added to the fixed-length checking code. I have (a)
635        corrected the bug and added tests for these items, and (b) arranged for an
636        error to occur if an unknown opcode is encountered while checking for fixed
637        length instead of just assuming "not fixed length". The items that were
638        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
639        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
640        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
641    
642    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
643        being incorrectly compiled and would have given unpredicatble results.
644    
645    9.  A possessively repeated subpattern with minimum repeat count greater than
646        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
647        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
648        the first (A) could occur when it should not.
649    
650    10. Add a cast and remove a redundant test from the code.
651    
652    11. JIT should use pcre_malloc/pcre_free for allocation.
653    
654    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
655        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
656        is anything other than /usr, -L is still shown).
657    
658    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
659    
660    14. Perl does not support \N without a following name in a [] class; PCRE now
661        also gives an error.
662    
663    15. If a forward reference was repeated with an upper limit of around 2000,
664        it caused the error "internal error: overran compiling workspace". The
665        maximum number of forward references (including repeats) was limited by the
666        internal workspace, and dependent on the LINK_SIZE. The code has been
667        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
668        the default depends on LINK_SIZE. There is a new upper limit (for safety)
669        of around 200,000 forward references. While doing this, I also speeded up
670        the filling in of repeated forward references.
671    
672    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
673        incorrectly expecting the subject to contain another "a" after the start.
674    
675    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
676        in the match, the SKIP should be ignored. This was not happening; instead
677        the SKIP was being treated as NOMATCH. For patterns such as
678        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
679        tested.
680    
681    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
682        now much more compatible with Perl, in particular in cases where the result
683        is a non-match for a non-anchored pattern. For example, if
684        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
685        "m", where previously it did not return a name. A side effect of this
686        change is that for partial matches, the last encountered mark name is
687        returned, as for non matches. A number of tests that were previously not
688        Perl-compatible have been moved into the Perl-compatible test files. The
689        refactoring has had the pleasing side effect of removing one argument from
690        the match() function, thus reducing its stack requirements.
691    
692    19. If the /S+ option was used in pcretest to study a pattern using JIT,
693        subsequent uses of /S (without +) incorrectly behaved like /S+.
694    
695    21. Retrieve executable code size support for the JIT compiler and fixing
696        some warnings.
697    
698    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
699        not work when the shorter character appeared right at the end of the
700        subject string.
701    
702    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
703        systems.
704    
705    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
706        output it when the /M option is used in pcretest.
707    
708    25. The CheckMan script was not being included in the distribution. Also, added
709        an explicit "perl" to run Perl scripts from the PrepareRelease script
710        because this is reportedly needed in Windows.
711    
712    26. If study data was being save in a file and studying had not found a set of
713        "starts with" bytes for the pattern, the data written to the file (though
714        never used) was taken from uninitialized memory and so caused valgrind to
715        complain.
716    
717    27. Updated RunTest.bat as provided by Sheri Pierce.
718    
719    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
720    
721    29. Computation of memory usage for the table of capturing group names was
722        giving an unnecessarily large value.
723    
724    
725    Version 8.20 21-Oct-2011
726    ------------------------
727    
728    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
729        a POSIX class. After further experiments with Perl, which convinced me that
730        Perl has bugs and confusions, a closing square bracket is no longer allowed
731        in a POSIX name. This bug also affected patterns with classes that started
732        with full stops.
733    
734    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
735        captured substring, but while checking the failing first alternative,
736        substring 1 is temporarily captured. If the output vector supplied to
737        pcre_exec() was not big enough for this capture, the yield of the function
738        was still zero ("insufficient space for captured substrings"). This cannot
739        be totally fixed without adding another stack variable, which seems a lot
740        of expense for a edge case. However, I have improved the situation in cases
741        such as /(a)(b)x|abc/ matched against "abc", where the return code
742        indicates that fewer than the maximum number of slots in the ovector have
743        been set.
744    
745    3.  Related to (2) above: when there are more back references in a pattern than
746        slots in the output vector, pcre_exec() uses temporary memory during
747        matching, and copies in the captures as far as possible afterwards. It was
748        using the entire output vector, but this conflicts with the specification
749        that only 2/3 is used for passing back captured substrings. Now it uses
750        only the first 2/3, for compatibility. This is, of course, another edge
751        case.
752    
753    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
754        main code base, and can be used by building with --enable-jit. When this is
755        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
756        runtime --no-jit option is given.
757    
758    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
759        ovector, the return from the function was zero, implying that there were
760        other matches that did not fit. The correct "exactly full" value is now
761        returned.
762    
763    6.  If a subpattern that was called recursively or as a subroutine contained
764        (*PRUNE) or any other control that caused it to give a non-standard return,
765        invalid errors such as "Error -26 (nested recursion at the same subject
766        position)" or even infinite loops could occur.
767    
768    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
769        computing the minimum length on reaching *ACCEPT, and so ended up with the
770        wrong value of 1 rather than 0. Further investigation indicates that
771        computing a minimum subject length in the presence of *ACCEPT is difficult
772        (think back references, subroutine calls), and so I have changed the code
773        so that no minimum is registered for a pattern that contains *ACCEPT.
774    
775    8.  If (*THEN) was present in the first (true) branch of a conditional group,
776        it was not handled as intended. [But see 16 below.]
777    
778    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
779        Sheri Pierce.
780    
781    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
782        the first byte in a match must be "a".
783    
784    11. Change 17 for 8.13 increased the recursion depth for patterns like
785        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
786        pattern contains any instances of (*THEN). If it does not, the old
787        optimizations are restored. It would be nice to do this on a per-group
788        basis, but at the moment that is not feasible.
789    
790    12. In some environments, the output of pcretest -C is CRLF terminated. This
791        broke RunTest's code that checks for the link size. A single white space
792        character after the value is now allowed for.
793    
794    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
795        For "fr", it uses the Windows-specific input and output files.
796    
797    14. If (*THEN) appeared in a group that was called recursively or as a
798        subroutine, it did not work as intended. [But see next item.]
799    
800    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
801        pattern fragments (but not containing any | characters). If A and B are
802        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
803        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
804        D. In other words, Perl considers parentheses that do not contain any |
805        characters to be part of a surrounding alternative, whereas PCRE was
806        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
807        differently. PCRE now behaves in the same way as Perl, except in the case
808        of subroutine/recursion calls such as (?1) which have in any case always
809        been different (but PCRE had them first :-).
810    
811    16. Related to 15 above: Perl does not treat the | in a conditional group as
812        creating alternatives. Such a group is treated in the same way as an
813        ordinary group without any | characters when processing (*THEN). PCRE has
814        been changed to match Perl's behaviour.
815    
816    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
817        RunGrepTest script failed.
818    
819    18. Change 22 for version 13 caused atomic groups to use more stack. This is
820        inevitable for groups that contain captures, but it can lead to a lot of
821        stack use in large patterns. The old behaviour has been restored for atomic
822        groups that do not contain any capturing parentheses.
823    
824    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
825        suppress the check for a minimum subject length at run time. (If it was
826        given to pcre_exec() or pcre_dfa_exec() it did work.)
827    
828    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
829        fail to work when decoding hex characters in data strings in EBCDIC
830        environments.
831    
832    21. It appears that in at least one Mac OS environment, the isxdigit() function
833        is implemented as a macro that evaluates to its argument more than once,
834        contravening the C 90 Standard (I haven't checked a later standard). There
835        was an instance in pcretest which caused it to go wrong when processing
836        \x{...} escapes in subject strings. The has been rewritten to avoid using
837        things like p++ in the argument of isxdigit().
838    
839    
840    Version 8.13 16-Aug-2011
841    ------------------------
842    
843    1.  The Unicode data tables have been updated to Unicode 6.0.0.
844    
845    2.  Two minor typos in pcre_internal.h have been fixed.
846    
847    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
848        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
849        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
850    
851    4.  There were a number of related bugs in the code for matching backrefences
852        caselessly in UTF-8 mode when codes for the characters concerned were
853        different numbers of bytes. For example, U+023A and U+2C65 are an upper
854        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
855        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
856        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
857        2-byte code at the end of the subject (it thought there wasn't enough data
858        left).
859    
860    5.  Comprehensive information about what went wrong is now returned by
861        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
862        as the output vector has at least 2 elements. The offset of the start of
863        the failing character and a reason code are placed in the vector.
864    
865    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
866        now returned is for the first byte of the failing character, instead of the
867        last byte inspected. This is an incompatible change, but I hope it is small
868        enough not to be a problem. It makes the returned offset consistent with
869        pcre_exec() and pcre_dfa_exec().
870    
871    7.  pcretest now gives a text phrase as well as the error number when
872        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
873        failure, the offset and reason code are output.
874    
875    8.  When \R was used with a maximizing quantifier it failed to skip backwards
876        over a \r\n pair if the subsequent match failed. Instead, it just skipped
877        back over a single character (\n). This seems wrong (because it treated the
878        two characters as a single entity when going forwards), conflicts with the
879        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
880        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
881        has been changed.
882    
883    9.  Some internal refactoring has changed the processing so that the handling
884        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
885        time (the PCRE_DOTALL option was changed this way some time ago: version
886        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
887        which was always a bit of a fudge. It also means that there is one less
888        argument for the match() function, which reduces its stack requirements
889        slightly. This change also fixes an incompatibility with Perl: the pattern
890        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
891    
892    10. More internal refactoring has drastically reduced the number of recursive
893        calls to match() for possessively repeated groups such as (abc)++ when
894        using pcre_exec().
895    
896    11. While implementing 10, a number of bugs in the handling of groups were
897        discovered and fixed:
898    
899        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
900        (a|)*(?1) gave a compile-time internal error.
901        ((a|)+)+  did not notice that the outer group could match an empty string.
902        (^a|^)+   was not marked as anchored.
903        (.*a|.*)+ was not marked as matching at start or after a newline.
904    
905    12. Yet more internal refactoring has removed another argument from the match()
906        function. Special calls to this function are now indicated by setting a
907        value in a variable in the "match data" data block.
908    
909    13. Be more explicit in pcre_study() instead of relying on "default" for
910        opcodes that mean there is no starting character; this means that when new
911        ones are added and accidentally left out of pcre_study(), testing should
912        pick them up.
913    
914    14. The -s option of pcretest has been documented for ages as being an old
915        synonym of -m (show memory usage). I have changed it to mean "force study
916        for every regex", that is, assume /S for every regex. This is similar to -i
917        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
918        using it. It makes it easier to run collections of tests with and without
919        study enabled, and thereby test pcre_study() more easily. All the standard
920        tests are now run with and without -s (but some patterns can be marked as
921        "never study" - see 20 below).
922    
923    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
924        restoration of the capturing data to the outer values was not happening
925        correctly.
926    
927    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
928        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
929        pattern had matched an empty string, and so incorrectly returned a no
930        match.
931    
932    17. There was optimizing code for the last branch of non-capturing parentheses,
933        and also for the obeyed branch of a conditional subexpression, which used
934        tail recursion to cut down on stack usage. Unfortunately, now that there is
935        the possibility of (*THEN) occurring in these branches, tail recursion is
936        no longer possible because the return has to be checked for (*THEN). These
937        two optimizations have therefore been removed. [But see 8.20/11 above.]
938    
939    18. If a pattern containing \R was studied, it was assumed that \R always
940        matched two bytes, thus causing the minimum subject length to be
941        incorrectly computed because \R can also match just one byte.
942    
943    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
944        was incorrectly computed.
945    
946    20. If /S is present twice on a test pattern in pcretest input, it now
947        *disables* studying, thereby overriding the use of -s on the command line
948        (see 14 above). This is necessary for one or two tests to keep the output
949        identical in both cases.
950    
951    21. When (*ACCEPT) was used in an assertion that matched an empty string and
952        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
953    
954    22. When an atomic group that contained a capturing parenthesis was
955        successfully matched, but the branch in which it appeared failed, the
956        capturing was not being forgotten if a higher numbered group was later
957        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
958        group 1 to "a", when in fact it should be unset. This applied to multi-
959        branched capturing and non-capturing groups, repeated or not, and also to
960        positive assertions (capturing in negative assertions does not happen
961        in PCRE) and also to nested atomic groups.
962    
963    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
964        subject after a captured substring, to make it easier to tell which of a
965        number of identical substrings has been captured.
966    
967    24. The way atomic groups are processed by pcre_exec() has been changed so that
968        if they are repeated, backtracking one repetition now resets captured
969        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
970        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
971        "aaa". Previously, it would have been "a". As part of this code
972        refactoring, the way recursive calls are handled has also been changed.
973    
974    25. If an assertion condition captured any substrings, they were not passed
975        back unless some other capturing happened later. For example, if
976        (?(?=(a))a) was matched against "a", no capturing was returned.
977    
978    26. When studying a pattern that contained subroutine calls or assertions,
979        the code for finding the minimum length of a possible match was handling
980        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
981        group 1 called group 2 while simultaneously a separate group 2 called group
982        1). A stack overflow occurred in this case. I have fixed this by limiting
983        the recursion depth to 10.
984    
985    27. Updated RunTest.bat in the distribution to the version supplied by Tom
986        Fortmann. This supports explicit test numbers on the command line, and has
987        argument validation and error reporting.
988    
989    28. An instance of \X with an unlimited repeat could fail if at any point the
990        first character it looked at was a mark character.
991    
992    29. Some minor code refactoring concerning Unicode properties and scripts
993        should reduce the stack requirement of match() slightly.
994    
995    30. Added the '=' option to pcretest to check the setting of unused capturing
996        slots at the end of the pattern, which are documented as being -1, but are
997        not included in the return count.
998    
999    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1000        compiled something random. Now it gives a compile-time error (as does
1001        Perl).
1002    
1003    32. A *MARK encountered during the processing of a positive assertion is now
1004        recorded and passed back (compatible with Perl).
1005    
1006    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1007        had alternative anchored branches, the search for a second match in a line
1008        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1009        matched the line "0102" twice. The same bug affected patterns that started
1010        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1011        twice.
1012    
1013    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1014        does, and because of capturing effects, quantifying parenthesized
1015        assertions may at times be useful. Quantifiers are now allowed for
1016        parenthesized assertions.
1017    
1018    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1019    
1020    36. \g was being checked for fancy things in a character class, when it should
1021        just be a literal "g".
1022    
1023    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1024        appearance of a nested POSIX class supersedes an apparent external class.
1025        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1026        unescaped square brackets may also appear as part of class names. For
1027        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1028        more like Perl. (But see 8.20/1 above.)
1029    
1030    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1031        was because it thought it was \N{name}, which is not supported).
1032    
1033    39. Add minix to OS list not supporting the -S option in pcretest.
1034    
1035    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1036        cannot analyze patterns in sufficient detail to catch mutual recursions
1037        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1038        subgroup is called recursively as a subpattern for a second time at the
1039        same position in the subject string. In previous releases this might have
1040        been caught by the recursion limit, or it might have run out of stack.
1041    
1042    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1043        happen only once. PCRE was, however incorrectly giving a compile time error
1044        "recursive call could loop indefinitely" because it cannot analyze the
1045        pattern in sufficient detail. The compile time test no longer happens when
1046        PCRE is compiling a conditional subpattern, but actual runaway loops are
1047        now caught at runtime (see 40 above).
1048    
1049    42. It seems that Perl allows any characters other than a closing parenthesis
1050        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1051        has been changed to be the same.
1052    
1053    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1054        as not to get warnings when autogen.sh is called. Also changed
1055        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1056    
1057    44. To help people who use pcregrep to scan files containing exceedingly long
1058        lines, the following changes have been made:
1059    
1060        (a) The default value of the buffer size parameter has been increased from
1061            8K to 20K. (The actual buffer used is three times this size.)
1062    
1063        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1064            PCRE is built.
1065    
1066        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1067            to be set at run time.
1068    
1069        (d) Numerical values in pcregrep options can be followed by K or M, for
1070            example --buffer-size=50K.
1071    
1072        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1073            given and the return code is set to 2.
1074    
1075    45. Add a pointer to the latest mark to the callout data block.
1076    
1077    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1078        partial match of an empty string instead of no match. This was specific to
1079        the use of ".".
1080    
1081    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1082        complete match instead of a partial match. This bug was dependent on both
1083        the PCRE_UTF8 and PCRE_DOTALL options being set.
1084    
1085    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1086        starting byte set, because \b was not being ignored.
1087    
1088    
1089    Version 8.12 15-Jan-2011
1090    ------------------------
1091    
1092    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1093        checks for such things as part of the documentation building process.
1094    
1095    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1096        --match-limit and --recursion-limit options (added for 8.11). In
1097        particular, this made one of the standard tests fail. (The integer value
1098        went into the wrong half of a long int.)
1099    
1100    3.  If the --colour option was given to pcregrep with -v (invert match), it
1101        did strange things, either producing crazy output, or crashing. It should,
1102        of course, ignore a request for colour when reporting lines that do not
1103        match.
1104    
1105    4.  Another pcregrep bug caused similar problems if --colour was specified with
1106        -M (multiline) and the pattern match finished with a line ending.
1107    
1108    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1109        matched in multiline mode, the following line was shown as part of the
1110        match. This seems wrong, so I have changed it.
1111    
1112    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1113        the check for further matches in the same line (so they could be coloured)
1114        to overrun the end of the current line. If another match was found, it was
1115        incorrectly shown (and then shown again when found in the next line).
1116    
1117    7.  If pcregrep was compiled under Windows, there was a reference to the
1118        function pcregrep_exit() before it was defined. I am assuming this was
1119        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1120        reported by a user. I've moved the definition above the reference.
1121    
1122    
1123    Version 8.11 10-Dec-2010
1124    ------------------------
1125    
1126    1.  (*THEN) was not working properly if there were untried alternatives prior
1127        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1128        backtracked to try for "b" instead of moving to the next alternative branch
1129        at the same level (in this case, to look for "c"). The Perl documentation
1130        is clear that when (*THEN) is backtracked onto, it goes to the "next
1131        alternative in the innermost enclosing group".
1132    
1133    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1134        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1135        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1136        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1137        (*THEN).
1138    
1139    3.  If \s appeared in a character class, it removed the VT character from
1140        the class, even if it had been included by some previous item, for example
1141        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1142        of \s, but is part of the POSIX "space" class.)
1143    
1144    4.  A partial match never returns an empty string (because you can always
1145        match an empty string at the end of the subject); however the checking for
1146        an empty string was starting at the "start of match" point. This has been
1147        changed to the "earliest inspected character" point, because the returned
1148        data for a partial match starts at this character. This means that, for
1149        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1150        (previously it gave "no match").
1151    
1152    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1153        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1154        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1155        has an implication that the given string is incomplete (because a partial
1156        match is preferred over a full match). For this reason, these items now
1157        give a partial match in this situation. [Aside: previously, the one case
1158        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1159        match rather than a full match, which was wrong by the old rules, but is
1160        now correct.]
1161    
1162    6.  There was a bug in the handling of #-introduced comments, recognized when
1163        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1164        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1165        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1166        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1167        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1168        places in pcre_compile().
1169    
1170    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1171        comments when looking ahead for named forward references to subpatterns,
1172        the only newline sequence it recognized was NL. It now handles newlines
1173        according to the set newline convention.
1174    
1175    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1176        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1177        cater for a lack of strerror(). These oversights have been fixed.
1178    
1179    9.  Added --match-limit and --recursion-limit to pcregrep.
1180    
1181    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1182    
1183    11. When the -o option was used, pcregrep was setting a return code of 1, even
1184        when matches were found, and --line-buffered was not being honoured.
1185    
1186    12. Added an optional parentheses number to the -o and --only-matching options
1187        of pcregrep.
1188    
1189    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1190        can match an empty string. The code to do it in pcretest and pcredemo
1191        needed fixing:
1192    
1193        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1194            only one byte after an empty string match just before CRLF (this case
1195            just got forgotten; "any" and "anycrlf" were OK).
1196    
1197        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1198            mode when an empty string match preceded an ASCII character followed by
1199            a non-ASCII character. (The code for advancing by one character rather
1200            than one byte was nonsense.)
1201    
1202        (c) The pcredemo.c sample program did not have any code at all to handle
1203            the cases when CRLF is a valid newline sequence.
1204    
1205    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1206        as a starting offset was within the subject string. There is now a new
1207        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1208        negative or greater than the length of the string. In order to test this,
1209        pcretest is extended to allow the setting of negative starting offsets.
1210    
1211    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1212        starting offset points to the beginning of a UTF-8 character was
1213        unnecessarily clumsy. I tidied it up.
1214    
1215    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1216        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1217    
1218    17. Nobody had reported that the --include_dir option, which was added in
1219        release 7.7 should have been called --include-dir (hyphen, not underscore)
1220        for compatibility with GNU grep. I have changed it to --include-dir, but
1221        left --include_dir as an undocumented synonym, and the same for
1222        --exclude-dir, though that is not available in GNU grep, at least as of
1223        release 2.5.4.
1224    
1225    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1226        characters from a string of bytes) have been redefined so as not to use
1227        loops, in order to improve performance in some environments. At the same
1228        time, I abstracted some of the common code into auxiliary macros to save
1229        repetition (this should not affect the compiled code).
1230    
1231    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1232        compile-time error is now given if \c is not followed by an ASCII
1233        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1234        different, and any byte value is allowed.)
1235    
1236    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1237        START_OPTIMIZE option, which is now allowed at compile time - but just
1238        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1239        to pcregrep and other applications that have no direct access to PCRE
1240        options. The new /Y option in pcretest sets this option when calling
1241        pcre_compile().
1242    
1243    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1244        back references. Groups containing recursive back references were forced to
1245        be atomic by that change, but in the case of named groups, the amount of
1246        memory required was incorrectly computed, leading to "Failed: internal
1247        error: code overflow". This has been fixed.
1248    
1249    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1250        pcretest.c, to avoid build problems in some Borland environments.
1251    
1252    
1253    Version 8.10 25-Jun-2010
1254    ------------------------
1255    
1256    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1257        THEN.
1258    
1259    2.  (*ACCEPT) was not working when inside an atomic group.
1260    
1261    3.  Inside a character class, \B is treated as a literal by default, but
1262        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1263        causes the error). The code is unchanged, but I tidied the documentation.
1264    
1265    4.  Inside a character class, PCRE always treated \R and \X as literals,
1266        whereas Perl faults them if its -w option is set. I have changed PCRE so
1267        that it faults them when PCRE_EXTRA is set.
1268    
1269    5.  Added support for \N, which always matches any character other than
1270        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1271    
1272    6.  When compiling pcregrep with newer versions of gcc which may have
1273        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1274        declared with attribute warn_unused_result" were given. Just casting the
1275        result to (void) does not stop the warnings; a more elaborate fudge is
1276        needed. I've used a macro to implement this.
1277    
1278    7.  Minor change to pcretest.c to avoid a compiler warning.
1279    
1280    8.  Added four artifical Unicode properties to help with an option to make
1281        \s etc use properties (see next item). The new properties are: Xan
1282        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1283    
1284    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1285        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1286        this option. Modified pcretest to add /W to test this facility. Added
1287        REG_UCP to make it available via the POSIX interface.
1288    
1289    10. Added --line-buffered to pcregrep.
1290    
1291    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1292        studied, and the match started with a letter with a code point greater than
1293        127 whose first byte was different to the first byte of the other case of
1294        the letter, the other case of this starting letter was not recognized
1295        (#976).
1296    
1297    12. If a pattern that was studied started with a repeated Unicode property
1298        test, for example, \p{Nd}+, there was the theoretical possibility of
1299        setting up an incorrect bitmap of starting bytes, but fortunately it could
1300        not have actually happened in practice until change 8 above was made (it
1301        added property types that matched character-matching opcodes).
1302    
1303    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1304        possible starting bytes for non-anchored patterns.
1305    
1306    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1307        \R, and also a number of cases that involve Unicode properties, both
1308        explicit and implicit when PCRE_UCP is set.
1309    
1310    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1311        input, it could crash or give wrong results if characters with values
1312        greater than 0xc0 were present in the subject string. (Detail: it assumed
1313        UTF-8 input when processing these items.)
1314    
1315    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1316        size_t is 64-bit (#991).
1317    
1318    17. Added a check for running out of memory when PCRE is compiled with
1319        --disable-stack-for-recursion (#990).
1320    
1321    18. If the last data line in a file for pcretest does not have a newline on
1322        the end, a newline was missing in the output.
1323    
1324    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1325        less than 128) in its various bitmaps. However, there is a facility for
1326        generating tables according to the current locale when PCRE is compiled. It
1327        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1328        characters, are recognized by isspace() and therefore were getting set in
1329        these tables, and indeed these tables seem to approximate to ISO 8859. This
1330        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1331        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1332        which of course cannot start UTF-8 characters. I have changed the code so
1333        that only real ASCII characters (less than 128) and the correct starting
1334        bytes for UTF-8 encodings are set for characters greater than 127 when in
1335        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1336        altogether.)
1337    
1338    20. Added the /T option to pcretest so as to be able to run tests with non-
1339        standard character tables, thus making it possible to include the tests
1340        used for 19 above in the standard set of tests.
1341    
1342    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1343        reference to a subpattern the other side of a comment that contains an
1344        opening parenthesis caused either an internal compiling error, or a
1345        reference to the wrong subpattern.
1346    
1347    
1348    Version 8.02 19-Mar-2010
1349    ------------------------
1350    
1351    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1352    
1353    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1354        configured.
1355    
1356    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1357        original author of that file, following a query about its status.
1358    
1359    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1360        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1361    
1362    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1363        quantifier applied to a forward-referencing subroutine call, could compile
1364        incorrect code or give the error "internal error: previously-checked
1365        referenced subpattern not found".
1366    
1367    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1368        variables to point to external functions. For these systems, therefore,
1369        pcre_malloc etc. are now initialized to local functions that call the
1370        relevant global functions.
1371    
1372    7.  There were two entries missing in the vectors called coptable and poptable
1373        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1374        I've fixed the data, and added a kludgy way of testing at compile time that
1375        the lengths are correct (equal to the number of opcodes).
1376    
1377    8.  Following on from 7, I added a similar kludge to check the length of the
1378        eint vector in pcreposix.c.
1379    
1380    9.  Error texts for pcre_compile() are held as one long string to avoid too
1381        much relocation at load time. To find a text, the string is searched,
1382        counting zeros. There was no check for running off the end of the string,
1383        which could happen if a new error number was added without updating the
1384        string.
1385    
1386    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1387    
1388    11. \K was not working if it appeared in an atomic group or in a group that
1389        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1390        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1391        the assertion is positive, but not if it is negative.
1392    
1393    12. Change 11 fortuitously reduced the size of the stack frame used in the
1394        "match()" function of pcre_exec.c by one pointer. Forthcoming
1395        implementation of support for (*MARK) will need an extra pointer on the
1396        stack; I have reserved it now, so that the stack frame size does not
1397        decrease.
1398    
1399    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1400        item in branch that calls a recursion is a subroutine call - as in the
1401        second branch in the above example - was incorrectly given the compile-
1402        time error "recursive call could loop indefinitely" because pcre_compile()
1403        was not correctly checking the subroutine for matching a non-empty string.
1404    
1405    14. The checks for overrunning compiling workspace could trigger after an
1406        overrun had occurred. This is a "should never occur" error, but it can be
1407        triggered by pathological patterns such as hundreds of nested parentheses.
1408        The checks now trigger 100 bytes before the end of the workspace.
1409    
1410    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1411    
1412    
1413    Version 8.01 19-Jan-2010
1414    ------------------------
1415    
1416    1.  If a pattern contained a conditional subpattern with only one branch (in
1417        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1418        computed the wrong minimum data length (which is of course zero for such
1419        subpatterns). This could cause incorrect "no match" results.
1420    
1421    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1422        the pattern is reset in the first branch, pcre_compile() failed with
1423        "internal error: code overflow at offset...". This happened only when
1424        the reset was to the original external option setting. (An optimization
1425        abstracts leading options settings into an external setting, which was the
1426        cause of this.)
1427    
1428    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1429        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1430        assertion pattern did not match (meaning that the assertion was true), it
1431        was incorrectly treated as false if the SKIP had been reached during the
1432        matching. This also applied to assertions used as conditions.
1433    
1434    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1435        assertion subpattern, including such a pattern used as a condition,
1436        unpredictable results occurred, instead of the error return
1437        PCRE_ERROR_DFA_UITEM.
1438    
1439    5.  The C++ GlobalReplace function was not working like Perl for the special
1440        situation when an empty string is matched. It now does the fancy magic
1441        stuff that is necessary.
1442    
1443    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1444        removed. (These were left over from very, very early versions of PCRE.)
1445    
1446    7.  Some cosmetic changes to the code to make life easier when compiling it
1447        as part of something else:
1448    
1449        (a) Change DEBUG to PCRE_DEBUG.
1450    
1451        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1452            called "current" as "current_branch", to prevent a collision with the
1453            Linux macro when compiled as a kernel module.
1454    
1455        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1456            prevent a collision with the Linux macro when compiled as a kernel
1457            module.
1458    
1459    8.  In pcre_compile() there are some checks for integer overflows that used to
1460        cast potentially large values to (double). This has been changed to that
1461        when building, a check for int64_t is made, and if it is found, it is used
1462        instead, thus avoiding the use of floating point arithmetic. (There is no
1463        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1464        double.
1465    
1466    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1467        2005 (difference between two addresses compared to an unsigned value).
1468    
1469    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1470        custom one, because of the following reported problem in Windows:
1471    
1472          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1473              under Win32.
1474          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1475              therefore missing the function definition.
1476          - The compiler thus generates a "C" signature for the test function.
1477          - The linker fails to find the "C" function.
1478          - PCRE fails to configure if asked to do so against libbz2.
1479    
1480    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1481        messages were output:
1482    
1483          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1484          rerunning libtoolize, to keep the correct libtool macros in-tree.
1485          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1486    
1487        I have done both of these things.
1488    
1489    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1490        most of the time, it *can* run out if it is given a pattern that contains a
1491        runaway infinite recursion. I updated the discussion in the pcrestack man
1492        page.
1493    
1494    13. Now that we have gone to the x.xx style of version numbers, the minor
1495        version may start with zero. Using 08 or 09 is a bad idea because users
1496        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1497        interpreted as invalid octal numbers. I've updated the previous comment in
1498        configure.ac, and also added a check that gives an error if 08 or 09 are
1499        used.
1500    
1501    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1502        causing partial matching to fail when the end of the subject matched \W
1503        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1504    
1505    15. There were some discrepancies between the declarations in pcre_internal.h
1506        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1507        their definitions. The declarations used "const uschar *" and the
1508        definitions used USPTR. Even though USPTR is normally defined as "const
1509        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1510        reported that: "This difference in casting confuses some C++ compilers, for
1511        example, SunCC recognizes above declarations as different functions and
1512        generates broken code for hbpcre." I have changed the declarations to use
1513        USPTR.
1514    
1515    16. GNU libtool is named differently on some systems. The autogen.sh script now
1516        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1517        (FreeBSD).
1518    
1519    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1520        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1521        comment: "Figure out how to create a longlong from a string: strtoll and
1522        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1523        instance, but it only takes 2 args instead of 3!"
1524    
1525    18. A subtle bug concerned with back references has been fixed by a change of
1526        specification, with a corresponding code fix. A pattern such as
1527        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1528        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1529        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1530        same bug. Such groups have to be quantified to be useful, or contained
1531        inside another quantified group. (If there's no repetition, the reference
1532        can never match.) The problem arises because, having left the group and
1533        moved on to the rest of the pattern, a later failure that backtracks into
1534        the group uses the captured value from the final iteration of the group
1535        rather than the correct earlier one. I have fixed this in PCRE by forcing
1536        any group that contains a reference to itself to be an atomic group; that
1537        is, there cannot be any backtracking into it once it has completed. This is
1538        similar to recursive and subroutine calls.
1539    
1540    
1541    Version 8.00 19-Oct-09
1542    ----------------------
1543    
1544    1.  The table for translating pcre_compile() error codes into POSIX error codes
1545        was out-of-date, and there was no check on the pcre_compile() error code
1546        being within the table. This could lead to an OK return being given in
1547        error.
1548    
1549    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1550        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1551        in a Windows environment.
1552    
1553    3.  The pcregrep --count option prints the count for each file even when it is
1554        zero, as does GNU grep. However, pcregrep was also printing all files when
1555        --files-with-matches was added. Now, when both options are given, it prints
1556        counts only for those files that have at least one match. (GNU grep just
1557        prints the file name in this circumstance, but including the count seems
1558        more useful - otherwise, why use --count?) Also ensured that the
1559        combination -clh just lists non-zero counts, with no names.
1560    
1561    4.  The long form of the pcregrep -F option was incorrectly implemented as
1562        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1563        but it seems right to fix it, and I didn't think it was worth preserving
1564        the old behaviour.
1565    
1566    5.  The command line items --regex=pattern and --regexp=pattern were not
1567        recognized by pcregrep, which required --regex pattern or --regexp pattern
1568        (with a space rather than an '='). The man page documented the '=' forms,
1569        which are compatible with GNU grep; these now work.
1570    
1571    6.  No libpcreposix.pc file was created for pkg-config; there was just
1572        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1573    
1574    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1575        when UCP support is not needed, by modifying the Python script that
1576        generates it from Unicode data files. This should not matter if the module
1577        is correctly used as a library, but I received one complaint about 50K of
1578        unwanted data. My guess is that the person linked everything into his
1579        program rather than using a library. Anyway, it does no harm.
1580    
1581    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1582        was a minimum greater than 1 for a wide character in a possessive
1583        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1584        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1585        character. Chaos in the form of incorrect output or a compiling loop could
1586        result.
1587    
1588    9.  The restrictions on what a pattern can contain when partial matching is
1589        requested for pcre_exec() have been removed. All patterns can now be
1590        partially matched by this function. In addition, if there are at least two
1591        slots in the offset vector, the offset of the earliest inspected character
1592        for the match and the offset of the end of the subject are set in them when
1593        PCRE_ERROR_PARTIAL is returned.
1594    
1595    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1596        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1597        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1598        and may be more useful for multi-segment matching.
1599    
1600    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1601        used to be given if ever the end of the subject was reached; now it is
1602        given only if matching could not proceed because another character was
1603        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1604        string "Z", which now yields "no match" instead of "partial match". In the
1605        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1606        final character ended with (*FAIL).
1607    
1608    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1609        if the pattern had a "must contain" character that was already found in the
1610        earlier partial match, unless partial matching was again requested. For
1611        example, with the pattern /dog.(body)?/, the "must contain" character is
1612        "g". If the first part-match was for the string "dog", restarting with
1613        "sbody" failed. This bug has been fixed.
1614    
1615    13. The string returned by pcre_dfa_exec() after a partial match has been
1616        changed so that it starts at the first inspected character rather than the
1617        first character of the match. This makes a difference only if the pattern
1618        starts with a lookbehind assertion or \b or \B (\K is not supported by
1619        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1620        matching functions compatible, and I think it's the right thing to do.
1621    
1622    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1623        so that the demonstration program is easily available in environments where
1624        PCRE has not been installed from source.
1625    
1626    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1627        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1628        library.
1629    
1630    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1631        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1632        is not the first non-POSIX option to be added. Clearly some people find
1633        these options useful.
1634    
1635    17. If a caller to the POSIX matching function regexec() passes a non-zero
1636        value for nmatch with a NULL value for pmatch, the value of
1637        nmatch is forced to zero.
1638    
1639    18. RunGrepTest did not have a test for the availability of the -u option of
1640        the diff command, as RunTest does. It now checks in the same way as
1641        RunTest, and also checks for the -b option.
1642    
1643    19. If an odd number of negated classes containing just a single character
1644        interposed, within parentheses, between a forward reference to a named
1645        subpattern and the definition of the subpattern, compilation crashed with
1646        an internal error, complaining that it could not find the referenced
1647        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1648        [The bug was that it was starting one character too far in when skipping
1649        over the character class, thus treating the ] as data rather than
1650        terminating the class. This meant it could skip too much.]
1651    
1652    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1653        /g option in pcretest when the pattern contains \K, which makes it possible
1654        to have an empty string match not at the start, even when the pattern is
1655        anchored. Updated pcretest and pcredemo to use this option.
1656    
1657    21. If the maximum number of capturing subpatterns in a recursion was greater
1658        than the maximum at the outer level, the higher number was returned, but
1659        with unset values at the outer level. The correct (outer level) value is
1660        now given.
1661    
1662    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1663        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1664        make it do so. The string so far is captured, making this feature
1665        compatible with Perl.
1666    
1667    23. The tests have been re-organized, adding tests 11 and 12, to make it
1668        possible to check the Perl 5.10 features against Perl 5.10.
1669    
1670    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1671        pattern matches a fixed length string. PCRE did not allow this; now it
1672        does. Neither allows recursion.
1673    
1674    25. I finally figured out how to implement a request to provide the minimum
1675        length of subject string that was needed in order to match a given pattern.
1676        (It was back references and recursion that I had previously got hung up
1677        on.) This code has now been added to pcre_study(); it finds a lower bound
1678        to the length of subject needed. It is not necessarily the greatest lower
1679        bound, but using it to avoid searching strings that are too short does give
1680        some useful speed-ups. The value is available to calling programs via
1681        pcre_fullinfo().
1682    
1683    26. While implementing 25, I discovered to my embarrassment that pcretest had
1684        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1685        study optimizations had never been tested with that matching function.
1686        Oops. What is worse, even when it was passed study data, there was a bug in
1687        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1688        were also very few tests of studied patterns with pcre_dfa_exec().
1689    
1690    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1691        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1692        on the other side of the coin, they are no longer allowed to have different
1693        names, because these cannot be distinguished in PCRE, and this has caused
1694        confusion. (This is a difference from Perl.)
1695    
1696    28. When duplicate subpattern names are present (necessarily with different
1697        numbers, as required by 27 above), and a test is made by name in a
1698        conditional pattern, either for a subpattern having been matched, or for
1699        recursion in such a pattern, all the associated numbered subpatterns are
1700        tested, and the overall condition is true if the condition is true for any
1701        one of them. This is the way Perl works, and is also more like the way
1702        testing by number works.
1703    
1704    
1705    Version 7.9 11-Apr-09
1706    ---------------------
1707    
1708    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1709        (pcretest), all targets were linked against these libraries. This included
1710        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1711        libraries. This caused unwanted dependencies to be created. This problem
1712        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1713        pcretest is linked with readline.
1714    
1715    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1716        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1717        moved outside it again, because FALSE and TRUE are already defined in AIX,
1718        but BOOL is not.
1719    
1720    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1721        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1722    
1723    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1724        hyphens) following file names and line numbers when outputting matching
1725        lines. This is not true; no spaces are inserted. I have also clarified the
1726        wording for the --colour (or --color) option.
1727    
1728    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1729        was not coloured; this is different to GNU grep, so I have changed it to be
1730        the same.
1731    
1732    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1733        each matching line was coloured. Now it goes on to look for further matches
1734        of any of the test patterns, which is the same behaviour as GNU grep.
1735    
1736    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1737        doesn't make sense to accept an empty string match in pcregrep, so I have
1738        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1739        seems to be how GNU grep behaves. [But see later change 40 for release
1740        8.33.]
1741    
1742    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1743        start or after a newline", because the conditional assertion was not being
1744        correctly handled. The rule now is that both the assertion and what follows
1745        in the first alternative must satisfy the test.
1746    
1747    9.  If auto-callout was enabled in a pattern with a conditional group whose
1748        condition was an assertion, PCRE could crash during matching, both with
1749        pcre_exec() and pcre_dfa_exec().
1750    
1751    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1752        used for matching.
1753    
1754    11. Unicode property support in character classes was not working for
1755        characters (bytes) greater than 127 when not in UTF-8 mode.
1756    
1757    12. Added the -M command line option to pcretest.
1758    
1759    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1760    
1761    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1762    
1763    16. Added comments and documentation about mis-use of no_arg in the C++
1764        wrapper.
1765    
1766    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1767        from Martin Jerabek that uses macro names for all relevant character and
1768        string constants.
1769    
1770    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1771        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1772        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1773        these, but not everybody uses configure.
1774    
1775    19. A conditional group that had only one branch was not being correctly
1776        recognized as an item that could match an empty string. This meant that an
1777        enclosing group might also not be so recognized, causing infinite looping
1778        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1779        with the subject "ab", where knowledge that the repeated group can match
1780        nothing is needed in order to break the loop.
1781    
1782    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1783        exec(), but without supplying a callout function, matching went wrong.
1784    
1785    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1786        leak if the size of the offset vector was greater than 30. When the vector
1787        is smaller, the saved offsets during recursion go onto a local stack
1788        vector, but for larger vectors malloc() is used. It was failing to free
1789        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1790        error, in fact).
1791    
1792    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1793        heapframe that is used only when UTF-8 support is enabled. This caused no
1794        problem, but was untidy.
1795    
1796    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1797        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1798        included within another project.
1799    
1800    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1801        slightly modified by me:
1802    
1803          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1804              not building pcregrep.
1805    
1806          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1807              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1808    
1809    25. Forward references, both numeric and by name, in patterns that made use of
1810        duplicate group numbers, could behave incorrectly or give incorrect errors,
1811        because when scanning forward to find the reference group, PCRE was not
1812        taking into account the duplicate group numbers. A pattern such as
1813        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1814    
1815    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1816        the feature of a custom pointer more persuasive (as requested by a user).
1817    
1818    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1819        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1820        pre-defined.
1821    
1822    28. Added support for (*UTF8) at the start of a pattern.
1823    
1824    29. Arrange for flags added by the "release type" setting in CMake to be shown
1825        in the configuration summary.
1826    
1827    
1828    Version 7.8 05-Sep-08
1829    ---------------------
1830    
1831    1.  Replaced UCP searching code with optimized version as implemented for Ad
1832        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1833        stage table and inline lookup instead of a function, giving speed ups of 2
1834        to 5 times on some simple patterns that I tested. Permission was given to
1835        distribute the MultiStage2.py script that generates the tables (it's not in
1836        the tarball, but is in the Subversion repository).
1837    
1838    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1839        scripts.
1840    
1841    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1842        a group with a zero qualifier. The result of the study could be incorrect,
1843        or the function might crash, depending on the pattern.
1844    
1845    4.  Caseless matching was not working for non-ASCII characters in back
1846        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1847        It now works when Unicode Property Support is available.
1848    
1849    5.  In pcretest, an escape such as \x{de} in the data was always generating
1850        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1851        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1852        truncation.
1853    
1854    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1855    
1856    7.  Added two (int) casts to pcregrep when printing the difference of two
1857        pointers, in case they are 64-bit values.
1858    
1859    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1860        test 2 if it fails.
1861    
1862    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1863        and a #define of that name to empty if it is not externally set. This is to
1864        allow users of MSVC to set it if necessary.
1865    
1866    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1867        the convenience functions in the pcre_get.c source file.
1868    
1869    11. An option change at the start of a pattern that had top-level alternatives
1870        could cause overwriting and/or a crash. This command provoked a crash in
1871        some environments:
1872    
1873          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1874    
1875        This potential security problem was recorded as CVE-2008-2371.
1876    
1877    12. For a pattern where the match had to start at the beginning or immediately
1878        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1879        pcre_dfa_exec() could read past the end of the passed subject if there was
1880        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1881        pcretest so that it places the subject at the end of its malloc-ed buffer.
1882    
1883    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1884        exec() might read past the end of the data buffer in UTF-8 mode.
1885    
1886    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1887        the data contained the byte 0x85 as part of a UTF-8 character within its
1888        first line. This applied both to normal and DFA matching.
1889    
1890    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1891        /^[^d]*?$/8 failed to match "abc".
1892    
1893    16. Added a missing copyright notice to pcrecpp_internal.h.
1894    
1895    17. Make it more clear in the documentation that values returned from
1896        pcre_exec() in ovector are byte offsets, not character counts.
1897    
1898    18. Tidied a few places to stop certain compilers from issuing warnings.
1899    
1900    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1901        supplied by Stefan Weber. I made a further small update for 7.8 because
1902        there is a change of source arrangements: the pcre_searchfuncs.c module is
1903        replaced by pcre_ucd.c.
1904    
1905    
1906    Version 7.7 07-May-08
1907    ---------------------
1908    
1909    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1910        a string to a long long, pretend we don't even have a long long." This is
1911        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1912    
1913    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1914        pre-7.6 versions, which defined a global no_arg variable instead of putting
1915        it in the RE class. (See also #8 below.)
1916    
1917    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1918        Lopes.
1919    
1920    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1921    
1922        (1) The include/exclude patterns were being applied to the whole pathnames
1923            of files, instead of just to the final components.
1924    
1925        (2) If there was more than one level of directory, the subdirectories were
1926            skipped unless they satisfied the include/exclude conditions. This is
1927            inconsistent with GNU grep (and could even be seen as contrary to the
1928            pcregrep specification - which I improved to make it absolutely clear).
1929            The action now is always to scan all levels of directory, and just
1930            apply the include/exclude patterns to regular files.
1931    
1932    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1933        --exclude_dir in the tests to avoid scanning .svn directories.
1934    
1935    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1936        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1937        doesn't support NULs in patterns.
1938    
1939    7.  Added some missing "const"s to declarations of static tables in
1940        pcre_compile.c and pcre_dfa_exec.c.
1941    
1942    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1943        caused by fix #2  above. (Subsequently also a second patch to fix the
1944        first patch. And a third patch - this was a messy problem.)
1945    
1946    9.  Applied Craig's patch to remove the use of push_back().
1947    
1948    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1949        matching function regexec().
1950    
1951    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1952        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1953        references. PCRE supports relative numbers with this syntax (I don't think
1954        Oniguruma does).
1955    
1956    12. Previously, a group with a zero repeat such as (...){0} was completely
1957        omitted from the compiled regex. However, this means that if the group
1958        was called as a subroutine from elsewhere in the pattern, things went wrong
1959        (an internal error was given). Such groups are now left in the compiled
1960        pattern, with a new opcode that causes them to be skipped at execution
1961        time.
1962    
1963    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1964        to the way PCRE behaves:
1965    
1966        (a) A lone ] character is dis-allowed (Perl treats it as data).
1967    
1968        (b) A back reference to an unmatched subpattern matches an empty string
1969            (Perl fails the current match path).
1970    
1971        (c) A data ] in a character class must be notated as \] because if the
1972            first data character in a class is ], it defines an empty class. (In
1973            Perl it is not possible to have an empty class.) The empty class []
1974            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1975            The negative empty class [^] matches any one character, independently
1976            of the DOTALL setting.
1977    
1978    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1979        non-existent subpattern following a character class starting with ']' and
1980        containing () gave an internal compiling error instead of "reference to
1981        non-existent subpattern". Fortunately, when the pattern did exist, the
1982        compiled code was correct. (When scanning forwards to check for the
1983        existence of the subpattern, it was treating the data ']' as terminating
1984        the class, so got the count wrong. When actually compiling, the reference
1985        was subsequently set up correctly.)
1986    
1987    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1988        it was being rejected as not supported by pcre_dfa_exec(), even though
1989        other assertions are supported. I have made pcre_dfa_exec() support
1990        (*FAIL).
1991    
1992    16. The implementation of 13c above involved the invention of a new opcode,
1993        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1994        cannot be changed at match time, I realized I could make a small
1995        improvement to matching performance by compiling OP_ALLANY instead of
1996        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1997        on the OP_ANY path.
1998    
1999    17. Compiling pcretest on Windows with readline support failed without the
2000        following two fixes: (1) Make the unistd.h include conditional on
2001        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2002    
2003    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2004        ncurses library to be included for pcretest when ReadLine support is
2005        requested, but also to allow for it to be overridden. This patch came from
2006        Daniel Bergström.
2007    
2008    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2009        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2010        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2011        spotting this.
2012    
2013    
2014    Version 7.6 28-Jan-08
2015    ---------------------
2016    
2017    1.  A character class containing a very large number of characters with
2018        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2019        overflow.
2020    
2021    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2022        HAVE_LONG_LONG is not defined.
2023    
2024    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2025        bring them up to date and include new features. This patch includes:
2026    
2027        - Fixed PH's badly added libz and libbz2 support.
2028        - Fixed a problem with static linking.
2029        - Added pcredemo. [But later removed - see 7 below.]
2030        - Fixed dftables problem and added an option.
2031        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2032            HAVE_LONG_LONG.
2033        - Added readline support for pcretest.
2034        - Added an listing of the option settings after cmake has run.
2035    
2036    4.  A user submitted a patch to Makefile that makes it easy to create
2037        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2038        Makefile.am that cause it to include this special target, without
2039        affecting anything else. Note that the same mingw target plus all
2040        the other distribution libraries and programs are now supported
2041        when configuring with CMake (see 6 below) instead of with
2042        Configure/Make.
2043    
2044    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2045        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2046        exported in the Windows port". It has not yet been confirmed that the patch
2047        solves the problem, but it does no harm.
2048    
2049    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2050        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2051        with CMake, and also correct the comment about stack recursion.
2052    
2053    7.  Remove the automatic building of pcredemo from the ./configure system and
2054        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2055        of a program that users should build themselves after PCRE is installed, so
2056        building it automatically is not really right. What is more, it gave
2057        trouble in some build environments.
2058    
2059    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2060    
2061    
2062    Version 7.5 10-Jan-08
2063    ---------------------
2064    
2065    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2066        values in parens when parsing an RE using the C++ wrapper."
2067    
2068    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2069        Characters greater than 255 were excluded from the class instead of being
2070        included.
2071    
2072    3.  The same bug as (2) above applied to negated POSIX classes such as
2073        [:^space:].
2074    
2075    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2076        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2077        I have changed it.
2078    
2079    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2080        first named subpattern) and a construct such as (?&a) would reference the
2081        first named subpattern whose name started with "a" (in other words, the
2082        length check was missing). Both these problems are fixed. "Subpattern name
2083        expected" is now given for (?&) (a zero-length name), and this patch also
2084        makes it give the same error for \k'' (previously it complained that that
2085        was a reference to a non-existent subpattern).
2086    
2087    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2088        this is right because (?- can be followed by option settings as well as by
2089        digits. I have, however, made the messages clearer.
2090    
2091    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2092        than the number used in the conditional) now cause a compile-time error.
2093        This is actually not compatible with Perl, which accepts such patterns, but
2094        treats the conditional as always being FALSE (as PCRE used to), but it
2095        seems to me that giving a diagnostic is better.
2096    
2097    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2098        and messages.
2099    
2100    9.  Fix two occurrences of "backslash" in comments that should have been
2101        "backspace".
2102    
2103    10. Remove two redundant lines of code that can never be obeyed (their function
2104        was moved elsewhere).
2105    
2106    11. The program that makes PCRE's Unicode character property table had a bug
2107        which caused it to generate incorrect table entries for sequences of
2108        characters that have the same character type, but are in different scripts.
2109        It amalgamated them into a single range, with the script of the first of
2110        them. In other words, some characters were in the wrong script. There were
2111        thirteen such cases, affecting characters in the following ranges:
2112    
2113          U+002b0 - U+002c1
2114          U+0060c - U+0060d
2115          U+0061e - U+00612
2116          U+0064b - U+0065e
2117          U+0074d - U+0076d
2118          U+01800 - U+01805
2119          U+01d00 - U+01d77
2120          U+01d9b - U+01dbf
2121          U+0200b - U+0200f
2122          U+030fc - U+030fe
2123          U+03260 - U+0327f
2124          U+0fb46 - U+0fbb1
2125          U+10450 - U+1049d
2126    
2127    12. The -o option (show only the matching part of a line) for pcregrep was not
2128        compatible with GNU grep in that, if there was more than one match in a
2129        line, it showed only the first of them. It now behaves in the same way as
2130        GNU grep.
2131    
2132    13. If the -o and -v options were combined for pcregrep, it printed a blank
2133        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2134        does the same. The return code can be used to tell if there were any
2135        non-matching lines.
2136    
2137    14. Added --file-offsets and --line-offsets to pcregrep.
2138    
2139    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2140        infinitely looping recursion. The bug was that positive lookaheads were not
2141        being skipped when checking for a possible empty match (negative lookaheads
2142        and both kinds of lookbehind were skipped).
2143    
2144    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2145        inclusion of <windows.h> to before rather than after the definition of
2146        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2147    
2148    17. Specifying a possessive quantifier with a specific limit for a Unicode
2149        character property caused pcre_compile() to compile bad code, which led at
2150        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2151        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2152        caused the error; without that there was no problem.
2153    
2154    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2155    
2156    19. Added --enable-pcretest-libreadline.
2157    
2158    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2159        RE::GlobalReplace(). As a result, the number of replacements returned was
2160        double what it should be. I removed one of the increments, but Craig sent a
2161        later patch that removed the other one (the right fix) and added unit tests
2162        that check the return values (which was not done before).
2163    
2164    21. Several CMake things:
2165    
2166        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2167            the names libpcre and libpcreposix, not just pcre and pcreposix.
2168    
2169        (2) The above change means that pcretest and pcregrep are now correctly
2170            linked with the newly-built libraries, not previously installed ones.
2171    
2172        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2173    
2174    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2175        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2176        UTF-8 newline character). The key issue is that the pattern starts .*;
2177        this means that the match must be either at the beginning, or after a
2178        newline. The bug was in the code for advancing after a failed match and
2179        checking that the new position followed a newline. It was not taking
2180        account of UTF-8 characters correctly.
2181    
2182    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2183        character classes. PCRE was not treating the sequence [:...:] as a
2184        character class unless the ... were all letters. Perl, however, seems to
2185        allow any characters between [: and :], though of course it rejects as
2186        unknown any "names" that contain non-letters, because all the known class
2187        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2188        for example, whereas PCRE did not - it did not recognize a POSIX character
2189        class. This seemed a bit dangerous, so the code has been changed to be
2190        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2191        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2192        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2193        Perl does, and where it didn't before.
2194    
2195    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2196        Windows environments %n is disabled by default.
2197    
2198    
2199    Version 7.4 21-Sep-07
2200    ---------------------
2201    
2202    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2203        means that a class such as [\s] counted as "explicit reference to CR or
2204        LF". That isn't really right - the whole point of the change was to try to
2205        help when there was an actual mention of one of the two characters. So now
2206        the change happens only if \r or \n (or a literal CR or LF) character is
2207        encountered.
2208    
2209    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2210        of both had grown to the point where there were only 3 bits left.
2211        Fortunately, there was spare space in the data structure, and so I have
2212        moved the internal flags into a new 16-bit field to free up more option
2213        bits.
2214    
2215    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2216        but did not set the internal JCHANGED flag - either of these is enough to
2217        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2218        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2219        start sets both bits.
2220    
2221    4.  Added options (at build time, compile time, exec time) to change \R from
2222        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2223    
2224    5.  doc/pcresyntax.html was missing from the distribution.
2225    
2226    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2227        compatibility, even though it is no longer used.
2228    
2229    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2230        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2231        windows.h file is present (where different names are used). [This was
2232        reversed later after testing - see 16 below.]
2233    
2234    8.  Changed all #include <config.h> to #include "config.h". There were also
2235        some further <pcre.h> cases that I changed to "pcre.h".
2236    
2237    9.  When pcregrep was used with the --colour option, it missed the line ending
2238        sequence off the lines that it output.
2239    
2240    10. It was pointed out to me that arrays of string pointers cause lots of
2241        relocations when a shared library is dynamically loaded. A technique of
2242        using a single long string with a table of offsets can drastically reduce
2243        these. I have refactored PCRE in four places to do this. The result is
2244        dramatic:
2245    
2246          Originally:                          290
2247          After changing UCP table:            187
2248          After changing error message table:   43
2249          After changing table of "verbs"       36
2250          After changing table of Posix names   22
2251    
2252        Thanks to the folks working on Gregex for glib for this insight.
2253    
2254    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2255        unicode-properties was also set.
2256    
2257    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2258    
2259    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2260        checked only for CRLF.
2261    
2262    14. Added casts to pcretest.c to avoid compiler warnings.
2263    
2264    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2265    
2266    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2267        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2268        entirely. This removes changes made in 7 above.
2269    
2270    17. The CMake files have been updated, and there is now more information about
2271        building with CMake in the NON-UNIX-USE document.
2272    
2273    
2274    Version 7.3 28-Aug-07
2275    ---------------------
2276    
2277     1. In the rejigging of the build system that eventually resulted in 7.1, the
2278        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2279        brackets there is not right, since it causes compilers to look for an
2280        installed pcre.h, not the version that is in the source that is being
2281        compiled (which of course may be different). I have changed it back to:
2282    
2283          #include "pcre.h"
2284    
2285        I have a vague recollection that the change was concerned with compiling in
2286        different directories, but in the new build system, that is taken care of
2287        by the VPATH setting the Makefile.
2288    
2289     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2290        when the subject happened to end in the byte 0x85 (e.g. if the last
2291        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2292        characters but of course it shouldn't be taken as a newline when it is part
2293        of another character. The bug was that, for an unlimited repeat of . in
2294        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2295        characters when looking for a newline.
2296    
2297     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2298    
2299     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2300        in debug output.
2301    
2302     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2303        long printing in the pcrecpp unittest when running under MinGW.
2304    
2305     6. ESC_K was left out of the EBCDIC table.
2306    
2307     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2308        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2309        limit also applies to "virtual nesting" when a pattern is recursive, and in
2310        this case 1000 isn't so big. I have been able to remove this limit at the
2311        expense of backing off one optimization in certain circumstances. Normally,
2312        when pcre_exec() would call its internal match() function recursively and
2313        immediately return the result unconditionally, it uses a "tail recursion"
2314        feature to save stack. However, when a subpattern that can match an empty
2315        string has an unlimited repetition quantifier, it no longer makes this
2316        optimization. That gives it a stack frame in which to save the data for
2317        checking that an empty string has been matched. Previously this was taken
2318        from the 1000-entry workspace that had been reserved. So now there is no
2319        explicit limit, but more stack is used.
2320    
2321     8. Applied Daniel's patches to solve problems with the import/export magic
2322        syntax that is required for Windows, and which was going wrong for the
2323        pcreposix and pcrecpp parts of the library. These were overlooked when this
2324        problem was solved for the main library.
2325    
2326     9. There were some crude static tests to avoid integer overflow when computing
2327        the size of patterns that contain repeated groups with explicit upper
2328        limits. As the maximum quantifier is 65535, the maximum group length was
2329        set at 30,000 so that the product of these two numbers did not overflow a
2330        32-bit integer. However, it turns out that people want to use groups that
2331        are longer than 30,000 bytes (though not repeat them that many times).
2332        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2333        made it possible to implement the integer overflow checks in a much more
2334        dynamic way, which I have now done. The artificial limitation on group
2335        length has been removed - we now have only the limit on the total length of
2336        the compiled pattern, which depends on the LINK_SIZE setting.
2337    
2338    10. Fixed a bug in the documentation for get/copy named substring when
2339        duplicate names are permitted. If none of the named substrings are set, the
2340        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2341        empty string.
2342    
2343    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2344        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2345        because the ] is interpreted as the first data character and the
2346        terminating ] is not found. PCRE has been made compatible with Perl in this
2347        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2348        cause memory overwriting.
2349    
2350    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2351        string has been matched (to stop an infinite loop). It was not recognizing
2352        a conditional subpattern that could match an empty string if that
2353        subpattern was within another subpattern. For example, it looped when
2354        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2355        condition was not nested. This bug has been fixed.
2356    
2357    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2358        past the start of the subject in the presence of bytes with the top bit
2359        set, for example "\x8aBCD".
2360    
2361    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2362        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2363    
2364    14. Optimized (?!) to (*FAIL).
2365    
2366    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2367        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2368        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2369        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2370        does: it's just the validity check that is more restrictive.
2371    
2372    16. Inserted checks for integer overflows during escape sequence (backslash)
2373        processing, and also fixed erroneous offset values for syntax errors during
2374        backslash processing.
2375    
2376    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2377        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2378    
2379    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2380        caused an overrun.
2381    
2382    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2383        something other than just ASCII characters) inside a group that had an
2384        unlimited repeat caused a loop at compile time (while checking to see
2385        whether the group could match an empty string).
2386    
2387    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2388        [\P{Any}] did so. (Error in the code for printing property names.)
2389    
2390    21. An orphan \E inside a character class could cause a crash.
2391    
2392    22. A repeated capturing bracket such as (A)? could cause a wild memory
2393        reference during compilation.
2394    
2395    23. There are several functions in pcre_compile() that scan along a compiled
2396        expression for various reasons (e.g. to see if it's fixed length for look
2397        behind). There were bugs in these functions when a repeated \p or \P was
2398        present in the pattern. These operators have additional parameters compared
2399        with \d, etc, and these were not being taken into account when moving along
2400        the compiled data. Specifically:
2401    
2402        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2403            length.
2404    
2405        (b) An item such as \pL+ within a repeated group could cause crashes or
2406            loops.
2407    
2408        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2409            "reference to non-existent subpattern" error.
2410    
2411        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2412    
2413    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2414        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2415    
2416    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2417    
2418    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2419        character were causing crashes (broken optimization).
2420    
2421    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2422        \p or \P) caused a compile-time loop.
2423    
2424    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2425        break. For example, the unstudied pattern [\r\n]A does not match the string
2426        "\r\nA" because change 7.0/46 below moves the current point on by two
2427        characters after failing to match at the start. However, the pattern \nA
2428        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2429        the same is true. There doesn't seem any very clean way out of this, but
2430        what I have chosen to do makes the common cases work: PCRE now takes note
2431        of whether there can be an explicit match for \r or \n anywhere in the
2432        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2433        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2434        pattern has explicit CR or LF references.
2435    
2436    29. Added (*CR) etc for changing newline setting at start of pattern.
2437    
2438    
2439    Version 7.2 19-Jun-07
2440    ---------------------
2441    
2442     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2443        which is apparently normally available under Windows.
2444    
2445     2. Re-jig the pcregrep tests with different newline settings in an attempt
2446        to make them independent of the local environment's newline setting.
2447    
2448     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2449    
2450     4. Some of the "internals" tests were previously cut out when the link size
2451        was not 2, because the output contained actual offsets. The recent new
2452        "Z" feature of pcretest means that these can be cut out, making the tests
2453        usable with all link sizes.
2454    
2455     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2456        stack recursion. This gives a massive performance boost under BSD, but just
2457        a small improvement under Linux. However, it saves one field in the frame
2458        in all cases.
2459    
2460     6. Added more features from the forthcoming Perl 5.10:
2461    
2462        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2463            recursion call. It refers to the nth most recently opened parentheses.
2464    
2465        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2466            to be opened parentheses.
2467    
2468        (c) Conditions that refer to capturing parentheses can be specified
2469            relatively, for example, (?(-2)... or (?(+3)...
2470    
2471        (d) \K resets the start of the current match so that everything before
2472            is not part of it.
2473    
2474        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2475    
2476        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2477            reference syntax.
2478    
2479        (g) (?| introduces a group in which the numbering of parentheses in each
2480            alternative starts with the same number.
2481    
2482        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2483    
2484     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2485        PCRE_INFO_JCHANGED.
2486    
2487     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2488        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2489        for detecting groups that can match an empty string.
2490    
2491     9. A pattern with a very large number of alternatives (more than several
2492        hundred) was running out of internal workspace during the pre-compile
2493        phase, where pcre_compile() figures out how much memory will be needed. A
2494        bit of new cunning has reduced the workspace needed for groups with
2495        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2496        workspace instead of running out of the 4096 that are available.
2497    
2498    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2499    
2500    11. Applied patch from Google to remove an optimization that didn't quite work.
2501        The report of the bug said:
2502    
2503          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2504          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2505          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2506    
2507    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2508        it matched the wrong number of bytes.
2509    
2510    
2511    Version 7.1 24-Apr-07
2512  ---------------------  ---------------------
2513    
2514   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
# Line 136  Version 7.1 12-Mar-07 Line 2643  Version 7.1 12-Mar-07
2643      string at the end of a line did not allow for this case. They now check for      string at the end of a line did not allow for this case. They now check for
2644      it specially.      it specially.
2645    
2646    20. pcretest is supposed to handle patterns and data of any length, by
2647        extending its buffers when necessary. It was getting this wrong when the
2648        buffer for a data line had to be extended.
2649    
2650    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2651        CRLF as a newline sequence.
2652    
2653    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2654        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2655        I have nevertheless tidied it up.
2656    
2657    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2658    
2659    24. Added a man page for pcre-config.
2660    
2661    
2662  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06
2663  ---------------------  ---------------------

Legend:
Removed from v.143  
changed lines
  Added in v.1353

  ViewVC Help
Powered by ViewVC 1.1.5