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

Diff of /code/trunk/ChangeLog

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

revision 257 by ph10, Wed Sep 19 09:11:19 2007 UTC revision 1024 by ph10, Wed Aug 29 11:08:24 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.4 10-Sep-07  Version 8.32
5    ------------
6    
7    1.  Improved JIT compiler optimizations for first character search and single
8        character iterators.
9    
10    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11        Patch by Daniel Richard G.
12    
13    3.  Single character iterator optimizations in the JIT compiler.
14    
15    4.  Improved JIT compiler optimizations for character ranges.
16    
17    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18        Reported by Giuseppe D'Angelo.
19    
20    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21        of a line, was being set incorrectly in cases where .* appeared inside
22        atomic brackets at the start of a pattern, or where there was a subsequent
23        *PRUNE or *SKIP.
24    
25    7.  Improved instruction cache flush for POWER/PowerPC.
26        Patch by Daniel Richard G.
27    
28    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
29        grep:
30    
31        (a) There is now no limit to the number of patterns to be matched.
32    
33        (b) An error is given if a pattern is too long.
34    
35        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
36            are now supported.
37    
38        (d) --exclude-from and --include-from (multiple use) have been added.
39    
40        (e) Exclusions and inclusions now apply to all files and directories, not
41            just to those obtained from scanning a directory recursively.
42    
43        (f) Multiple uses of -f and --file-list are now supported.
44    
45        (g) In a Windows environment, the default for -d has been changed from
46            "read" (the GNU grep default) to "skip", because otherwise the presence
47            of a directory in the file list provokes an error.
48    
49        (h) The documentation has been revised and clarified in places.
50    
51    9.  Improve the matching speed of capturing brackets.
52    
53    10. Changed the meaning of \X so that it now matches a Unicode extended
54        grapheme cluster.
55    
56    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
57        out POSIX threads when JIT support is configured.
58    
59    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
60    
61    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
62        structure could go wrong in environments where size_t is not the same size
63        as int.
64    
65    
66    
67    Version 8.31 06-July-2012
68    -------------------------
69    
70    1.  Fixing a wrong JIT test case and some compiler warnings.
71    
72    2.  Removed a bashism from the RunTest script.
73    
74    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
75        to unsigned type, result still unsigned" that was given by an MS compiler
76        on encountering the code "-sizeof(xxx)".
77    
78    4.  Partial matching support is added to the JIT compiler.
79    
80    5.  Fixed several bugs concerned with partial matching of items that consist
81        of more than one character:
82    
83        (a) /^(..)\1/ did not partially match "aba" because checking references was
84            done on an "all or nothing" basis. This also applied to repeated
85            references.
86    
87        (b) \R did not give a hard partial match if \r was found at the end of the
88            subject.
89    
90        (c) \X did not give a hard partial match after matching one or more
91            characters at the end of the subject.
92    
93        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
94            a partial match for the string "\r".
95    
96        (e) When newline was set to CRLF, the metacharacter "." did not recognize
97            a partial match for a CR character at the end of the subject string.
98    
99    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
100        running pcretest, the text "(JIT)" added to the output whenever JIT is
101        actually used to run the match.
102    
103    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
104        or /S+[+] with a digit between 1 and 7.
105    
106    8.  OP_NOT now supports any UTF character not just single-byte ones.
107    
108    9.  (*MARK) control verb is now supported by the JIT compiler.
109    
110    10. The command "./RunTest list" lists the available tests without actually
111        running any of them. (Because I keep forgetting what they all are.)
112    
113    11. Add PCRE_INFO_MAXLOOKBEHIND.
114    
115    12. Applied a (slightly modified) user-supplied patch that improves performance
116        when the heap is used for recursion (compiled with --disable-stack-for-
117        recursion). Instead of malloc and free for each heap frame each time a
118        logical recursion happens, frames are retained on a chain and re-used where
119        possible. This sometimes gives as much as 30% improvement.
120    
121    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
122        call.
123    
124    14. As documented, (*COMMIT) is now confined to within a positive assertion.
125    
126    15. It is now possible to link pcretest with libedit as an alternative to
127        libreadline.
128    
129    16. (*COMMIT) control verb is now supported by the JIT compiler.
130    
131    17. The Unicode data tables have been updated to Unicode 6.1.0.
132    
133    18. Added --file-list option to pcregrep.
134    
135    19. Added binary file support to pcregrep, including the -a, --binary-files,
136        -I, and --text options.
137    
138    20. The madvise function is renamed for posix_madvise for QNX compatibility
139        reasons. Fixed by Giuseppe D'Angelo.
140    
141    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
142    
143    22. Changed the option for creating symbolic links for 16-bit man pages from
144        -s to -sf so that re-installing does not cause issues.
145    
146    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
147    
148    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
149        matches in certain environments (the workspace was not being correctly
150        retained). Also added to pcre_dfa_exec() a simple plausibility check on
151        some of the workspace data at the beginning of a restart.
152    
153    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
154        was not doing so when it should - probably a typo introduced by SVN 528
155        (change 8.10/14).
156    
157    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
158        \w+ when the character tables indicated that \x{c4} was a word character.
159        There were several related cases, all because the tests for doing a table
160        lookup were testing for characters less than 127 instead of 255.
161    
162    27. If a pattern contains capturing parentheses that are not used in a match,
163        their slots in the ovector are set to -1. For those that are higher than
164        any matched groups, this happens at the end of processing. In the case when
165        there were back references that the ovector was too small to contain
166        (causing temporary malloc'd memory to be used during matching), and the
167        highest capturing number was not used, memory off the end of the ovector
168        was incorrectly being set to -1. (It was using the size of the temporary
169        memory instead of the true size.)
170    
171    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
172        ovector size, it uses memory at the end of the block that it has got.
173    
174    29. Check for an overlong MARK name and give an error at compile time. The
175        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
176    
177    30. JIT compiler update.
178    
179    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
180        Rill for the patch.
181    
182    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
183    
184    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
185    
186    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
187        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
188        was enabled.
189    
190    35. Improve JIT code generation for greedy plus quantifier.
191    
192    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
193        1 to "aa" instead of to an empty string. The bug affected repeated groups
194        that could potentially match an empty string.
195    
196    37. Optimizing single character iterators in JIT.
197    
198    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
199        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
200        codepoints that are too big for the mode are faulted, and in a UTF mode,
201        disallowed codepoints are also faulted.
202    
203    39. If PCRE was compiled with UTF support, in three places in the DFA
204        matcher there was code that should only have been obeyed in UTF mode, but
205        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
206        processing when bytes with values greater than 127 were present. In 16-bit
207        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
208        both cases the values are those that cannot be the first data item in a UTF
209        character. The three items that might have provoked this were recursions,
210        possessively repeated groups, and atomic groups.
211    
212    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
213        and pcregrep, because some OS require shared objects to be explicitly
214        passed to ld, causing the link step to fail if they are not.
215    
216    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
217        mode, patterns that started with \h* or \R* might be incorrectly matched.
218    
219    
220    Version 8.30 04-February-2012
221    -----------------------------
222    
223    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
224        name is defined in ctype.h.
225    
226    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
227        only in quite long subpatterns.
228    
229    3.  Removed the function pcre_info(), which has been obsolete and deprecated
230        since it was replaced by pcre_fullinfo() in February 2000.
231    
232    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
233        match a (*MARK), and the match failed at the start of the subject, a
234        reference to memory before the start of the subject could occur. This bug
235        was introduced by fix 17 of release 8.21.
236    
237    5.  A reference to an unset group with zero minimum repetition was giving
238        totally wrong answers (in non-JavaScript-compatibility mode). For example,
239        /(another)?(\1?)test/ matched against "hello world test". This bug was
240        introduced in release 8.13.
241    
242    6.  Add support for 16-bit character strings (a large amount of work involving
243        many changes and refactorings).
244    
245    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
246        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
247        from a file.
248    
249    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
250        rounding is not applied in this particular case).
251    
252    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
253        if they appear, or are escaped, in patterns.
254    
255    10. Get rid of a number of -Wunused-but-set-variable warnings.
256    
257    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
258        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
259        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
260        also returns the mark "x". This bug applied to capturing parentheses,
261        non-capturing parentheses, and atomic parentheses. It also applied to some
262        assertions.
263    
264    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
265        information out of configure.ac instead of relying on pcre.h.generic, which
266        is not stored in the repository.
267    
268    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
269        -lreadline.
270    
271    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
272    
273    15. Applied Graycode's patch to put the top-level frame on the stack rather
274        than the heap when not using the stack for recursion. This gives a
275        performance improvement in many cases when recursion is not deep.
276    
277    16. Experimental code added to "pcretest -C" to output the stack frame size.
278    
279    
280    Version 8.21 12-Dec-2011
281    ------------------------
282    
283    1.  Updating the JIT compiler.
284    
285    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
286        are added as well.
287    
288    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
289        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
290        calling _pcre_jit_exec. Some extra comments are added.
291    
292    4.  (*MARK) settings inside atomic groups that do not contain any capturing
293        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
294        was introduced by change 18 for 8.20.
295    
296    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
297        ECMA-262 standard.
298    
299    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
300        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
301        This bug was probably introduced by change 9 of 8.13.
302    
303    7.  While fixing 6 above, I noticed that a number of other items were being
304        incorrectly rejected as "not fixed length". This arose partly because newer
305        opcodes had not been added to the fixed-length checking code. I have (a)
306        corrected the bug and added tests for these items, and (b) arranged for an
307        error to occur if an unknown opcode is encountered while checking for fixed
308        length instead of just assuming "not fixed length". The items that were
309        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
310        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
311        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
312    
313    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
314        being incorrectly compiled and would have given unpredicatble results.
315    
316    9.  A possessively repeated subpattern with minimum repeat count greater than
317        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
318        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
319        the first (A) could occur when it should not.
320    
321    10. Add a cast and remove a redundant test from the code.
322    
323    11. JIT should use pcre_malloc/pcre_free for allocation.
324    
325    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
326        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
327        is anything other than /usr, -L is still shown).
328    
329    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
330    
331    14. Perl does not support \N without a following name in a [] class; PCRE now
332        also gives an error.
333    
334    15. If a forward reference was repeated with an upper limit of around 2000,
335        it caused the error "internal error: overran compiling workspace". The
336        maximum number of forward references (including repeats) was limited by the
337        internal workspace, and dependent on the LINK_SIZE. The code has been
338        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
339        the default depends on LINK_SIZE. There is a new upper limit (for safety)
340        of around 200,000 forward references. While doing this, I also speeded up
341        the filling in of repeated forward references.
342    
343    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
344        incorrectly expecting the subject to contain another "a" after the start.
345    
346    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
347        in the match, the SKIP should be ignored. This was not happening; instead
348        the SKIP was being treated as NOMATCH. For patterns such as
349        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
350        tested.
351    
352    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
353        now much more compatible with Perl, in particular in cases where the result
354        is a non-match for a non-anchored pattern. For example, if
355        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
356        "m", where previously it did not return a name. A side effect of this
357        change is that for partial matches, the last encountered mark name is
358        returned, as for non matches. A number of tests that were previously not
359        Perl-compatible have been moved into the Perl-compatible test files. The
360        refactoring has had the pleasing side effect of removing one argument from
361        the match() function, thus reducing its stack requirements.
362    
363    19. If the /S+ option was used in pcretest to study a pattern using JIT,
364        subsequent uses of /S (without +) incorrectly behaved like /S+.
365    
366    21. Retrieve executable code size support for the JIT compiler and fixing
367        some warnings.
368    
369    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
370        not work when the shorter character appeared right at the end of the
371        subject string.
372    
373    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
374        systems.
375    
376    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
377        output it when the /M option is used in pcretest.
378    
379    25. The CheckMan script was not being included in the distribution. Also, added
380        an explicit "perl" to run Perl scripts from the PrepareRelease script
381        because this is reportedly needed in Windows.
382    
383    26. If study data was being save in a file and studying had not found a set of
384        "starts with" bytes for the pattern, the data written to the file (though
385        never used) was taken from uninitialized memory and so caused valgrind to
386        complain.
387    
388    27. Updated RunTest.bat as provided by Sheri Pierce.
389    
390    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
391    
392    29. Computation of memory usage for the table of capturing group names was
393        giving an unnecessarily large value.
394    
395    
396    Version 8.20 21-Oct-2011
397    ------------------------
398    
399    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
400        a POSIX class. After further experiments with Perl, which convinced me that
401        Perl has bugs and confusions, a closing square bracket is no longer allowed
402        in a POSIX name. This bug also affected patterns with classes that started
403        with full stops.
404    
405    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
406        captured substring, but while checking the failing first alternative,
407        substring 1 is temporarily captured. If the output vector supplied to
408        pcre_exec() was not big enough for this capture, the yield of the function
409        was still zero ("insufficient space for captured substrings"). This cannot
410        be totally fixed without adding another stack variable, which seems a lot
411        of expense for a edge case. However, I have improved the situation in cases
412        such as /(a)(b)x|abc/ matched against "abc", where the return code
413        indicates that fewer than the maximum number of slots in the ovector have
414        been set.
415    
416    3.  Related to (2) above: when there are more back references in a pattern than
417        slots in the output vector, pcre_exec() uses temporary memory during
418        matching, and copies in the captures as far as possible afterwards. It was
419        using the entire output vector, but this conflicts with the specification
420        that only 2/3 is used for passing back captured substrings. Now it uses
421        only the first 2/3, for compatibility. This is, of course, another edge
422        case.
423    
424    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
425        main code base, and can be used by building with --enable-jit. When this is
426        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
427        runtime --no-jit option is given.
428    
429    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
430        ovector, the return from the function was zero, implying that there were
431        other matches that did not fit. The correct "exactly full" value is now
432        returned.
433    
434    6.  If a subpattern that was called recursively or as a subroutine contained
435        (*PRUNE) or any other control that caused it to give a non-standard return,
436        invalid errors such as "Error -26 (nested recursion at the same subject
437        position)" or even infinite loops could occur.
438    
439    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
440        computing the minimum length on reaching *ACCEPT, and so ended up with the
441        wrong value of 1 rather than 0. Further investigation indicates that
442        computing a minimum subject length in the presence of *ACCEPT is difficult
443        (think back references, subroutine calls), and so I have changed the code
444        so that no minimum is registered for a pattern that contains *ACCEPT.
445    
446    8.  If (*THEN) was present in the first (true) branch of a conditional group,
447        it was not handled as intended. [But see 16 below.]
448    
449    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
450        Sheri Pierce.
451    
452    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
453        the first byte in a match must be "a".
454    
455    11. Change 17 for 8.13 increased the recursion depth for patterns like
456        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
457        pattern contains any instances of (*THEN). If it does not, the old
458        optimizations are restored. It would be nice to do this on a per-group
459        basis, but at the moment that is not feasible.
460    
461    12. In some environments, the output of pcretest -C is CRLF terminated. This
462        broke RunTest's code that checks for the link size. A single white space
463        character after the value is now allowed for.
464    
465    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
466        For "fr", it uses the Windows-specific input and output files.
467    
468    14. If (*THEN) appeared in a group that was called recursively or as a
469        subroutine, it did not work as intended. [But see next item.]
470    
471    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
472        pattern fragments (but not containing any | characters). If A and B are
473        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
474        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
475        D. In other words, Perl considers parentheses that do not contain any |
476        characters to be part of a surrounding alternative, whereas PCRE was
477        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
478        differently. PCRE now behaves in the same way as Perl, except in the case
479        of subroutine/recursion calls such as (?1) which have in any case always
480        been different (but PCRE had them first :-).
481    
482    16. Related to 15 above: Perl does not treat the | in a conditional group as
483        creating alternatives. Such a group is treated in the same way as an
484        ordinary group without any | characters when processing (*THEN). PCRE has
485        been changed to match Perl's behaviour.
486    
487    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
488        RunGrepTest script failed.
489    
490    18. Change 22 for version 13 caused atomic groups to use more stack. This is
491        inevitable for groups that contain captures, but it can lead to a lot of
492        stack use in large patterns. The old behaviour has been restored for atomic
493        groups that do not contain any capturing parentheses.
494    
495    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
496        suppress the check for a minimum subject length at run time. (If it was
497        given to pcre_exec() or pcre_dfa_exec() it did work.)
498    
499    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
500        fail to work when decoding hex characters in data strings in EBCDIC
501        environments.
502    
503    21. It appears that in at least one Mac OS environment, the isxdigit() function
504        is implemented as a macro that evaluates to its argument more than once,
505        contravening the C 90 Standard (I haven't checked a later standard). There
506        was an instance in pcretest which caused it to go wrong when processing
507        \x{...} escapes in subject strings. The has been rewritten to avoid using
508        things like p++ in the argument of isxdigit().
509    
510    
511    Version 8.13 16-Aug-2011
512    ------------------------
513    
514    1.  The Unicode data tables have been updated to Unicode 6.0.0.
515    
516    2.  Two minor typos in pcre_internal.h have been fixed.
517    
518    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
519        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
520        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
521    
522    4.  There were a number of related bugs in the code for matching backrefences
523        caselessly in UTF-8 mode when codes for the characters concerned were
524        different numbers of bytes. For example, U+023A and U+2C65 are an upper
525        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
526        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
527        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
528        2-byte code at the end of the subject (it thought there wasn't enough data
529        left).
530    
531    5.  Comprehensive information about what went wrong is now returned by
532        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
533        as the output vector has at least 2 elements. The offset of the start of
534        the failing character and a reason code are placed in the vector.
535    
536    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
537        now returned is for the first byte of the failing character, instead of the
538        last byte inspected. This is an incompatible change, but I hope it is small
539        enough not to be a problem. It makes the returned offset consistent with
540        pcre_exec() and pcre_dfa_exec().
541    
542    7.  pcretest now gives a text phrase as well as the error number when
543        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
544        failure, the offset and reason code are output.
545    
546    8.  When \R was used with a maximizing quantifier it failed to skip backwards
547        over a \r\n pair if the subsequent match failed. Instead, it just skipped
548        back over a single character (\n). This seems wrong (because it treated the
549        two characters as a single entity when going forwards), conflicts with the
550        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
551        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
552        has been changed.
553    
554    9.  Some internal refactoring has changed the processing so that the handling
555        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
556        time (the PCRE_DOTALL option was changed this way some time ago: version
557        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
558        which was always a bit of a fudge. It also means that there is one less
559        argument for the match() function, which reduces its stack requirements
560        slightly. This change also fixes an incompatibility with Perl: the pattern
561        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
562    
563    10. More internal refactoring has drastically reduced the number of recursive
564        calls to match() for possessively repeated groups such as (abc)++ when
565        using pcre_exec().
566    
567    11. While implementing 10, a number of bugs in the handling of groups were
568        discovered and fixed:
569    
570        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
571        (a|)*(?1) gave a compile-time internal error.
572        ((a|)+)+  did not notice that the outer group could match an empty string.
573        (^a|^)+   was not marked as anchored.
574        (.*a|.*)+ was not marked as matching at start or after a newline.
575    
576    12. Yet more internal refactoring has removed another argument from the match()
577        function. Special calls to this function are now indicated by setting a
578        value in a variable in the "match data" data block.
579    
580    13. Be more explicit in pcre_study() instead of relying on "default" for
581        opcodes that mean there is no starting character; this means that when new
582        ones are added and accidentally left out of pcre_study(), testing should
583        pick them up.
584    
585    14. The -s option of pcretest has been documented for ages as being an old
586        synonym of -m (show memory usage). I have changed it to mean "force study
587        for every regex", that is, assume /S for every regex. This is similar to -i
588        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
589        using it. It makes it easier to run collections of tests with and without
590        study enabled, and thereby test pcre_study() more easily. All the standard
591        tests are now run with and without -s (but some patterns can be marked as
592        "never study" - see 20 below).
593    
594    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
595        restoration of the capturing data to the outer values was not happening
596        correctly.
597    
598    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
599        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
600        pattern had matched an empty string, and so incorrectly returned a no
601        match.
602    
603    17. There was optimizing code for the last branch of non-capturing parentheses,
604        and also for the obeyed branch of a conditional subexpression, which used
605        tail recursion to cut down on stack usage. Unfortunately, now that there is
606        the possibility of (*THEN) occurring in these branches, tail recursion is
607        no longer possible because the return has to be checked for (*THEN). These
608        two optimizations have therefore been removed. [But see 8.20/11 above.]
609    
610    18. If a pattern containing \R was studied, it was assumed that \R always
611        matched two bytes, thus causing the minimum subject length to be
612        incorrectly computed because \R can also match just one byte.
613    
614    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
615        was incorrectly computed.
616    
617    20. If /S is present twice on a test pattern in pcretest input, it now
618        *disables* studying, thereby overriding the use of -s on the command line
619        (see 14 above). This is necessary for one or two tests to keep the output
620        identical in both cases.
621    
622    21. When (*ACCEPT) was used in an assertion that matched an empty string and
623        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
624    
625    22. When an atomic group that contained a capturing parenthesis was
626        successfully matched, but the branch in which it appeared failed, the
627        capturing was not being forgotten if a higher numbered group was later
628        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
629        group 1 to "a", when in fact it should be unset. This applied to multi-
630        branched capturing and non-capturing groups, repeated or not, and also to
631        positive assertions (capturing in negative assertions does not happen
632        in PCRE) and also to nested atomic groups.
633    
634    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
635        subject after a captured substring, to make it easier to tell which of a
636        number of identical substrings has been captured.
637    
638    24. The way atomic groups are processed by pcre_exec() has been changed so that
639        if they are repeated, backtracking one repetition now resets captured
640        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
641        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
642        "aaa". Previously, it would have been "a". As part of this code
643        refactoring, the way recursive calls are handled has also been changed.
644    
645    25. If an assertion condition captured any substrings, they were not passed
646        back unless some other capturing happened later. For example, if
647        (?(?=(a))a) was matched against "a", no capturing was returned.
648    
649    26. When studying a pattern that contained subroutine calls or assertions,
650        the code for finding the minimum length of a possible match was handling
651        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
652        group 1 called group 2 while simultaneously a separate group 2 called group
653        1). A stack overflow occurred in this case. I have fixed this by limiting
654        the recursion depth to 10.
655    
656    27. Updated RunTest.bat in the distribution to the version supplied by Tom
657        Fortmann. This supports explicit test numbers on the command line, and has
658        argument validation and error reporting.
659    
660    28. An instance of \X with an unlimited repeat could fail if at any point the
661        first character it looked at was a mark character.
662    
663    29. Some minor code refactoring concerning Unicode properties and scripts
664        should reduce the stack requirement of match() slightly.
665    
666    30. Added the '=' option to pcretest to check the setting of unused capturing
667        slots at the end of the pattern, which are documented as being -1, but are
668        not included in the return count.
669    
670    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
671        compiled something random. Now it gives a compile-time error (as does
672        Perl).
673    
674    32. A *MARK encountered during the processing of a positive assertion is now
675        recorded and passed back (compatible with Perl).
676    
677    33. If --only-matching or --colour was set on a pcregrep call whose pattern
678        had alternative anchored branches, the search for a second match in a line
679        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
680        matched the line "0102" twice. The same bug affected patterns that started
681        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
682        twice.
683    
684    34. Previously, PCRE did not allow quantification of assertions. However, Perl
685        does, and because of capturing effects, quantifying parenthesized
686        assertions may at times be useful. Quantifiers are now allowed for
687        parenthesized assertions.
688    
689    35. A minor code tidy in pcre_compile() when checking options for \R usage.
690    
691    36. \g was being checked for fancy things in a character class, when it should
692        just be a literal "g".
693    
694    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
695        appearance of a nested POSIX class supersedes an apparent external class.
696        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
697        unescaped square brackets may also appear as part of class names. For
698        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
699        more like Perl. (But see 8.20/1 above.)
700    
701    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
702        was because it thought it was \N{name}, which is not supported).
703    
704    39. Add minix to OS list not supporting the -S option in pcretest.
705    
706    40. PCRE tries to detect cases of infinite recursion at compile time, but it
707        cannot analyze patterns in sufficient detail to catch mutual recursions
708        such as ((?1))((?2)). There is now a runtime test that gives an error if a
709        subgroup is called recursively as a subpattern for a second time at the
710        same position in the subject string. In previous releases this might have
711        been caught by the recursion limit, or it might have run out of stack.
712    
713    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
714        happen only once. PCRE was, however incorrectly giving a compile time error
715        "recursive call could loop indefinitely" because it cannot analyze the
716        pattern in sufficient detail. The compile time test no longer happens when
717        PCRE is compiling a conditional subpattern, but actual runaway loops are
718        now caught at runtime (see 40 above).
719    
720    42. It seems that Perl allows any characters other than a closing parenthesis
721        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
722        has been changed to be the same.
723    
724    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
725        as not to get warnings when autogen.sh is called. Also changed
726        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
727    
728    44. To help people who use pcregrep to scan files containing exceedingly long
729        lines, the following changes have been made:
730    
731        (a) The default value of the buffer size parameter has been increased from
732            8K to 20K. (The actual buffer used is three times this size.)
733    
734        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
735            PCRE is built.
736    
737        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
738            to be set at run time.
739    
740        (d) Numerical values in pcregrep options can be followed by K or M, for
741            example --buffer-size=50K.
742    
743        (e) If a line being scanned overflows pcregrep's buffer, an error is now
744            given and the return code is set to 2.
745    
746    45. Add a pointer to the latest mark to the callout data block.
747    
748    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
749        partial match of an empty string instead of no match. This was specific to
750        the use of ".".
751    
752    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
753        complete match instead of a partial match. This bug was dependent on both
754        the PCRE_UTF8 and PCRE_DOTALL options being set.
755    
756    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
757        starting byte set, because \b was not being ignored.
758    
759    
760    Version 8.12 15-Jan-2011
761    ------------------------
762    
763    1.  Fixed some typos in the markup of the man pages, and wrote a script that
764        checks for such things as part of the documentation building process.
765    
766    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
767        --match-limit and --recursion-limit options (added for 8.11). In
768        particular, this made one of the standard tests fail. (The integer value
769        went into the wrong half of a long int.)
770    
771    3.  If the --colour option was given to pcregrep with -v (invert match), it
772        did strange things, either producing crazy output, or crashing. It should,
773        of course, ignore a request for colour when reporting lines that do not
774        match.
775    
776    4.  Another pcregrep bug caused similar problems if --colour was specified with
777        -M (multiline) and the pattern match finished with a line ending.
778    
779    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
780        matched in multiline mode, the following line was shown as part of the
781        match. This seems wrong, so I have changed it.
782    
783    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
784        the check for further matches in the same line (so they could be coloured)
785        to overrun the end of the current line. If another match was found, it was
786        incorrectly shown (and then shown again when found in the next line).
787    
788    7.  If pcregrep was compiled under Windows, there was a reference to the
789        function pcregrep_exit() before it was defined. I am assuming this was
790        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
791        reported by a user. I've moved the definition above the reference.
792    
793    
794    Version 8.11 10-Dec-2010
795    ------------------------
796    
797    1.  (*THEN) was not working properly if there were untried alternatives prior
798        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
799        backtracked to try for "b" instead of moving to the next alternative branch
800        at the same level (in this case, to look for "c"). The Perl documentation
801        is clear that when (*THEN) is backtracked onto, it goes to the "next
802        alternative in the innermost enclosing group".
803    
804    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
805        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
806        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
807        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
808        (*THEN).
809    
810    3.  If \s appeared in a character class, it removed the VT character from
811        the class, even if it had been included by some previous item, for example
812        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
813        of \s, but is part of the POSIX "space" class.)
814    
815    4.  A partial match never returns an empty string (because you can always
816        match an empty string at the end of the subject); however the checking for
817        an empty string was starting at the "start of match" point. This has been
818        changed to the "earliest inspected character" point, because the returned
819        data for a partial match starts at this character. This means that, for
820        example, /(?<=abc)def/ gives a partial match for the subject "abc"
821        (previously it gave "no match").
822    
823    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
824        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
825        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
826        has an implication that the given string is incomplete (because a partial
827        match is preferred over a full match). For this reason, these items now
828        give a partial match in this situation. [Aside: previously, the one case
829        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
830        match rather than a full match, which was wrong by the old rules, but is
831        now correct.]
832    
833    6.  There was a bug in the handling of #-introduced comments, recognized when
834        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
835        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
836        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
837        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
838        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
839        places in pcre_compile().
840    
841    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
842        comments when looking ahead for named forward references to subpatterns,
843        the only newline sequence it recognized was NL. It now handles newlines
844        according to the set newline convention.
845    
846    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
847        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
848        cater for a lack of strerror(). These oversights have been fixed.
849    
850    9.  Added --match-limit and --recursion-limit to pcregrep.
851    
852    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
853    
854    11. When the -o option was used, pcregrep was setting a return code of 1, even
855        when matches were found, and --line-buffered was not being honoured.
856    
857    12. Added an optional parentheses number to the -o and --only-matching options
858        of pcregrep.
859    
860    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
861        can match an empty string. The code to do it in pcretest and pcredemo
862        needed fixing:
863    
864        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
865            only one byte after an empty string match just before CRLF (this case
866            just got forgotten; "any" and "anycrlf" were OK).
867    
868        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
869            mode when an empty string match preceded an ASCII character followed by
870            a non-ASCII character. (The code for advancing by one character rather
871            than one byte was nonsense.)
872    
873        (c) The pcredemo.c sample program did not have any code at all to handle
874            the cases when CRLF is a valid newline sequence.
875    
876    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
877        as a starting offset was within the subject string. There is now a new
878        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
879        negative or greater than the length of the string. In order to test this,
880        pcretest is extended to allow the setting of negative starting offsets.
881    
882    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
883        starting offset points to the beginning of a UTF-8 character was
884        unnecessarily clumsy. I tidied it up.
885    
886    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
887        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
888    
889    17. Nobody had reported that the --include_dir option, which was added in
890        release 7.7 should have been called --include-dir (hyphen, not underscore)
891        for compatibility with GNU grep. I have changed it to --include-dir, but
892        left --include_dir as an undocumented synonym, and the same for
893        --exclude-dir, though that is not available in GNU grep, at least as of
894        release 2.5.4.
895    
896    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
897        characters from a string of bytes) have been redefined so as not to use
898        loops, in order to improve performance in some environments. At the same
899        time, I abstracted some of the common code into auxiliary macros to save
900        repetition (this should not affect the compiled code).
901    
902    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
903        compile-time error is now given if \c is not followed by an ASCII
904        character, that is, a byte less than 128. (In EBCDIC mode, the code is
905        different, and any byte value is allowed.)
906    
907    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
908        START_OPTIMIZE option, which is now allowed at compile time - but just
909        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
910        to pcregrep and other applications that have no direct access to PCRE
911        options. The new /Y option in pcretest sets this option when calling
912        pcre_compile().
913    
914    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
915        back references. Groups containing recursive back references were forced to
916        be atomic by that change, but in the case of named groups, the amount of
917        memory required was incorrectly computed, leading to "Failed: internal
918        error: code overflow". This has been fixed.
919    
920    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
921        pcretest.c, to avoid build problems in some Borland environments.
922    
923    
924    Version 8.10 25-Jun-2010
925    ------------------------
926    
927    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
928        THEN.
929    
930    2.  (*ACCEPT) was not working when inside an atomic group.
931    
932    3.  Inside a character class, \B is treated as a literal by default, but
933        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
934        causes the error). The code is unchanged, but I tidied the documentation.
935    
936    4.  Inside a character class, PCRE always treated \R and \X as literals,
937        whereas Perl faults them if its -w option is set. I have changed PCRE so
938        that it faults them when PCRE_EXTRA is set.
939    
940    5.  Added support for \N, which always matches any character other than
941        newline. (It is the same as "." when PCRE_DOTALL is not set.)
942    
943    6.  When compiling pcregrep with newer versions of gcc which may have
944        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
945        declared with attribute warn_unused_result" were given. Just casting the
946        result to (void) does not stop the warnings; a more elaborate fudge is
947        needed. I've used a macro to implement this.
948    
949    7.  Minor change to pcretest.c to avoid a compiler warning.
950    
951    8.  Added four artifical Unicode properties to help with an option to make
952        \s etc use properties (see next item). The new properties are: Xan
953        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
954    
955    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
956        use Unicode properties. (*UCP) at the start of a pattern can be used to set
957        this option. Modified pcretest to add /W to test this facility. Added
958        REG_UCP to make it available via the POSIX interface.
959    
960    10. Added --line-buffered to pcregrep.
961    
962    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
963        studied, and the match started with a letter with a code point greater than
964        127 whose first byte was different to the first byte of the other case of
965        the letter, the other case of this starting letter was not recognized
966        (#976).
967    
968    12. If a pattern that was studied started with a repeated Unicode property
969        test, for example, \p{Nd}+, there was the theoretical possibility of
970        setting up an incorrect bitmap of starting bytes, but fortunately it could
971        not have actually happened in practice until change 8 above was made (it
972        added property types that matched character-matching opcodes).
973    
974    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
975        possible starting bytes for non-anchored patterns.
976    
977    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
978        \R, and also a number of cases that involve Unicode properties, both
979        explicit and implicit when PCRE_UCP is set.
980    
981    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
982        input, it could crash or give wrong results if characters with values
983        greater than 0xc0 were present in the subject string. (Detail: it assumed
984        UTF-8 input when processing these items.)
985    
986    16. Added a lot of (int) casts to avoid compiler warnings in systems where
987        size_t is 64-bit (#991).
988    
989    17. Added a check for running out of memory when PCRE is compiled with
990        --disable-stack-for-recursion (#990).
991    
992    18. If the last data line in a file for pcretest does not have a newline on
993        the end, a newline was missing in the output.
994    
995    19. The default pcre_chartables.c file recognizes only ASCII characters (values
996        less than 128) in its various bitmaps. However, there is a facility for
997        generating tables according to the current locale when PCRE is compiled. It
998        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
999        characters, are recognized by isspace() and therefore were getting set in
1000        these tables, and indeed these tables seem to approximate to ISO 8859. This
1001        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1002        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1003        which of course cannot start UTF-8 characters. I have changed the code so
1004        that only real ASCII characters (less than 128) and the correct starting
1005        bytes for UTF-8 encodings are set for characters greater than 127 when in
1006        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1007        altogether.)
1008    
1009    20. Added the /T option to pcretest so as to be able to run tests with non-
1010        standard character tables, thus making it possible to include the tests
1011        used for 19 above in the standard set of tests.
1012    
1013    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1014        reference to a subpattern the other side of a comment that contains an
1015        opening parenthesis caused either an internal compiling error, or a
1016        reference to the wrong subpattern.
1017    
1018    
1019    Version 8.02 19-Mar-2010
1020    ------------------------
1021    
1022    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1023    
1024    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1025        configured.
1026    
1027    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1028        original author of that file, following a query about its status.
1029    
1030    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1031        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1032    
1033    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1034        quantifier applied to a forward-referencing subroutine call, could compile
1035        incorrect code or give the error "internal error: previously-checked
1036        referenced subpattern not found".
1037    
1038    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1039        variables to point to external functions. For these systems, therefore,
1040        pcre_malloc etc. are now initialized to local functions that call the
1041        relevant global functions.
1042    
1043    7.  There were two entries missing in the vectors called coptable and poptable
1044        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1045        I've fixed the data, and added a kludgy way of testing at compile time that
1046        the lengths are correct (equal to the number of opcodes).
1047    
1048    8.  Following on from 7, I added a similar kludge to check the length of the
1049        eint vector in pcreposix.c.
1050    
1051    9.  Error texts for pcre_compile() are held as one long string to avoid too
1052        much relocation at load time. To find a text, the string is searched,
1053        counting zeros. There was no check for running off the end of the string,
1054        which could happen if a new error number was added without updating the
1055        string.
1056    
1057    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1058    
1059    11. \K was not working if it appeared in an atomic group or in a group that
1060        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1061        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1062        the assertion is positive, but not if it is negative.
1063    
1064    12. Change 11 fortuitously reduced the size of the stack frame used in the
1065        "match()" function of pcre_exec.c by one pointer. Forthcoming
1066        implementation of support for (*MARK) will need an extra pointer on the
1067        stack; I have reserved it now, so that the stack frame size does not
1068        decrease.
1069    
1070    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1071        item in branch that calls a recursion is a subroutine call - as in the
1072        second branch in the above example - was incorrectly given the compile-
1073        time error "recursive call could loop indefinitely" because pcre_compile()
1074        was not correctly checking the subroutine for matching a non-empty string.
1075    
1076    14. The checks for overrunning compiling workspace could trigger after an
1077        overrun had occurred. This is a "should never occur" error, but it can be
1078        triggered by pathological patterns such as hundreds of nested parentheses.
1079        The checks now trigger 100 bytes before the end of the workspace.
1080    
1081    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1082    
1083    
1084    Version 8.01 19-Jan-2010
1085    ------------------------
1086    
1087    1.  If a pattern contained a conditional subpattern with only one branch (in
1088        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1089        computed the wrong minimum data length (which is of course zero for such
1090        subpatterns). This could cause incorrect "no match" results.
1091    
1092    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1093        the pattern is reset in the first branch, pcre_compile() failed with
1094        "internal error: code overflow at offset...". This happened only when
1095        the reset was to the original external option setting. (An optimization
1096        abstracts leading options settings into an external setting, which was the
1097        cause of this.)
1098    
1099    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1100        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1101        assertion pattern did not match (meaning that the assertion was true), it
1102        was incorrectly treated as false if the SKIP had been reached during the
1103        matching. This also applied to assertions used as conditions.
1104    
1105    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1106        assertion subpattern, including such a pattern used as a condition,
1107        unpredictable results occurred, instead of the error return
1108        PCRE_ERROR_DFA_UITEM.
1109    
1110    5.  The C++ GlobalReplace function was not working like Perl for the special
1111        situation when an empty string is matched. It now does the fancy magic
1112        stuff that is necessary.
1113    
1114    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1115        removed. (These were left over from very, very early versions of PCRE.)
1116    
1117    7.  Some cosmetic changes to the code to make life easier when compiling it
1118        as part of something else:
1119    
1120        (a) Change DEBUG to PCRE_DEBUG.
1121    
1122        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1123            called "current" as "current_branch", to prevent a collision with the
1124            Linux macro when compiled as a kernel module.
1125    
1126        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1127            prevent a collision with the Linux macro when compiled as a kernel
1128            module.
1129    
1130    8.  In pcre_compile() there are some checks for integer overflows that used to
1131        cast potentially large values to (double). This has been changed to that
1132        when building, a check for int64_t is made, and if it is found, it is used
1133        instead, thus avoiding the use of floating point arithmetic. (There is no
1134        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1135        double.
1136    
1137    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1138        2005 (difference between two addresses compared to an unsigned value).
1139    
1140    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1141        custom one, because of the following reported problem in Windows:
1142    
1143          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1144              under Win32.
1145          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1146              therefore missing the function definition.
1147          - The compiler thus generates a "C" signature for the test function.
1148          - The linker fails to find the "C" function.
1149          - PCRE fails to configure if asked to do so against libbz2.
1150    
1151    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1152        messages were output:
1153    
1154          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1155          rerunning libtoolize, to keep the correct libtool macros in-tree.
1156          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1157    
1158        I have done both of these things.
1159    
1160    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1161        most of the time, it *can* run out if it is given a pattern that contains a
1162        runaway infinite recursion. I updated the discussion in the pcrestack man
1163        page.
1164    
1165    13. Now that we have gone to the x.xx style of version numbers, the minor
1166        version may start with zero. Using 08 or 09 is a bad idea because users
1167        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1168        interpreted as invalid octal numbers. I've updated the previous comment in
1169        configure.ac, and also added a check that gives an error if 08 or 09 are
1170        used.
1171    
1172    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1173        causing partial matching to fail when the end of the subject matched \W
1174        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1175    
1176    15. There were some discrepancies between the declarations in pcre_internal.h
1177        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1178        their definitions. The declarations used "const uschar *" and the
1179        definitions used USPTR. Even though USPTR is normally defined as "const
1180        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1181        reported that: "This difference in casting confuses some C++ compilers, for
1182        example, SunCC recognizes above declarations as different functions and
1183        generates broken code for hbpcre." I have changed the declarations to use
1184        USPTR.
1185    
1186    16. GNU libtool is named differently on some systems. The autogen.sh script now
1187        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1188        (FreeBSD).
1189    
1190    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1191        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1192        comment: "Figure out how to create a longlong from a string: strtoll and
1193        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1194        instance, but it only takes 2 args instead of 3!"
1195    
1196    18. A subtle bug concerned with back references has been fixed by a change of
1197        specification, with a corresponding code fix. A pattern such as
1198        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1199        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1200        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1201        same bug. Such groups have to be quantified to be useful, or contained
1202        inside another quantified group. (If there's no repetition, the reference
1203        can never match.) The problem arises because, having left the group and
1204        moved on to the rest of the pattern, a later failure that backtracks into
1205        the group uses the captured value from the final iteration of the group
1206        rather than the correct earlier one. I have fixed this in PCRE by forcing
1207        any group that contains a reference to itself to be an atomic group; that
1208        is, there cannot be any backtracking into it once it has completed. This is
1209        similar to recursive and subroutine calls.
1210    
1211    
1212    Version 8.00 19-Oct-09
1213    ----------------------
1214    
1215    1.  The table for translating pcre_compile() error codes into POSIX error codes
1216        was out-of-date, and there was no check on the pcre_compile() error code
1217        being within the table. This could lead to an OK return being given in
1218        error.
1219    
1220    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1221        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1222        in a Windows environment.
1223    
1224    3.  The pcregrep --count option prints the count for each file even when it is
1225        zero, as does GNU grep. However, pcregrep was also printing all files when
1226        --files-with-matches was added. Now, when both options are given, it prints
1227        counts only for those files that have at least one match. (GNU grep just
1228        prints the file name in this circumstance, but including the count seems
1229        more useful - otherwise, why use --count?) Also ensured that the
1230        combination -clh just lists non-zero counts, with no names.
1231    
1232    4.  The long form of the pcregrep -F option was incorrectly implemented as
1233        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1234        but it seems right to fix it, and I didn't think it was worth preserving
1235        the old behaviour.
1236    
1237    5.  The command line items --regex=pattern and --regexp=pattern were not
1238        recognized by pcregrep, which required --regex pattern or --regexp pattern
1239        (with a space rather than an '='). The man page documented the '=' forms,
1240        which are compatible with GNU grep; these now work.
1241    
1242    6.  No libpcreposix.pc file was created for pkg-config; there was just
1243        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1244    
1245    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1246        when UCP support is not needed, by modifying the Python script that
1247        generates it from Unicode data files. This should not matter if the module
1248        is correctly used as a library, but I received one complaint about 50K of
1249        unwanted data. My guess is that the person linked everything into his
1250        program rather than using a library. Anyway, it does no harm.
1251    
1252    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1253        was a minimum greater than 1 for a wide character in a possessive
1254        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1255        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1256        character. Chaos in the form of incorrect output or a compiling loop could
1257        result.
1258    
1259    9.  The restrictions on what a pattern can contain when partial matching is
1260        requested for pcre_exec() have been removed. All patterns can now be
1261        partially matched by this function. In addition, if there are at least two
1262        slots in the offset vector, the offset of the earliest inspected character
1263        for the match and the offset of the end of the subject are set in them when
1264        PCRE_ERROR_PARTIAL is returned.
1265    
1266    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1267        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1268        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1269        and may be more useful for multi-segment matching.
1270    
1271    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1272        used to be given if ever the end of the subject was reached; now it is
1273        given only if matching could not proceed because another character was
1274        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1275        string "Z", which now yields "no match" instead of "partial match". In the
1276        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1277        final character ended with (*FAIL).
1278    
1279    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1280        if the pattern had a "must contain" character that was already found in the
1281        earlier partial match, unless partial matching was again requested. For
1282        example, with the pattern /dog.(body)?/, the "must contain" character is
1283        "g". If the first part-match was for the string "dog", restarting with
1284        "sbody" failed. This bug has been fixed.
1285    
1286    13. The string returned by pcre_dfa_exec() after a partial match has been
1287        changed so that it starts at the first inspected character rather than the
1288        first character of the match. This makes a difference only if the pattern
1289        starts with a lookbehind assertion or \b or \B (\K is not supported by
1290        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1291        matching functions compatible, and I think it's the right thing to do.
1292    
1293    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1294        so that the demonstration program is easily available in environments where
1295        PCRE has not been installed from source.
1296    
1297    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1298        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1299        library.
1300    
1301    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1302        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1303        is not the first non-POSIX option to be added. Clearly some people find
1304        these options useful.
1305    
1306    17. If a caller to the POSIX matching function regexec() passes a non-zero
1307        value for nmatch with a NULL value for pmatch, the value of
1308        nmatch is forced to zero.
1309    
1310    18. RunGrepTest did not have a test for the availability of the -u option of
1311        the diff command, as RunTest does. It now checks in the same way as
1312        RunTest, and also checks for the -b option.
1313    
1314    19. If an odd number of negated classes containing just a single character
1315        interposed, within parentheses, between a forward reference to a named
1316        subpattern and the definition of the subpattern, compilation crashed with
1317        an internal error, complaining that it could not find the referenced
1318        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1319        [The bug was that it was starting one character too far in when skipping
1320        over the character class, thus treating the ] as data rather than
1321        terminating the class. This meant it could skip too much.]
1322    
1323    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1324        /g option in pcretest when the pattern contains \K, which makes it possible
1325        to have an empty string match not at the start, even when the pattern is
1326        anchored. Updated pcretest and pcredemo to use this option.
1327    
1328    21. If the maximum number of capturing subpatterns in a recursion was greater
1329        than the maximum at the outer level, the higher number was returned, but
1330        with unset values at the outer level. The correct (outer level) value is
1331        now given.
1332    
1333    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1334        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1335        make it do so. The string so far is captured, making this feature
1336        compatible with Perl.
1337    
1338    23. The tests have been re-organized, adding tests 11 and 12, to make it
1339        possible to check the Perl 5.10 features against Perl 5.10.
1340    
1341    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1342        pattern matches a fixed length string. PCRE did not allow this; now it
1343        does. Neither allows recursion.
1344    
1345    25. I finally figured out how to implement a request to provide the minimum
1346        length of subject string that was needed in order to match a given pattern.
1347        (It was back references and recursion that I had previously got hung up
1348        on.) This code has now been added to pcre_study(); it finds a lower bound
1349        to the length of subject needed. It is not necessarily the greatest lower
1350        bound, but using it to avoid searching strings that are too short does give
1351        some useful speed-ups. The value is available to calling programs via
1352        pcre_fullinfo().
1353    
1354    26. While implementing 25, I discovered to my embarrassment that pcretest had
1355        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1356        study optimizations had never been tested with that matching function.
1357        Oops. What is worse, even when it was passed study data, there was a bug in
1358        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1359        were also very few tests of studied patterns with pcre_dfa_exec().
1360    
1361    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1362        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1363        on the other side of the coin, they are no longer allowed to have different
1364        names, because these cannot be distinguished in PCRE, and this has caused
1365        confusion. (This is a difference from Perl.)
1366    
1367    28. When duplicate subpattern names are present (necessarily with different
1368        numbers, as required by 27 above), and a test is made by name in a
1369        conditional pattern, either for a subpattern having been matched, or for
1370        recursion in such a pattern, all the associated numbered subpatterns are
1371        tested, and the overall condition is true if the condition is true for any
1372        one of them. This is the way Perl works, and is also more like the way
1373        testing by number works.
1374    
1375    
1376    Version 7.9 11-Apr-09
1377    ---------------------
1378    
1379    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1380        (pcretest), all targets were linked against these libraries. This included
1381        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1382        libraries. This caused unwanted dependencies to be created. This problem
1383        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1384        pcretest is linked with readline.
1385    
1386    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1387        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1388        moved outside it again, because FALSE and TRUE are already defined in AIX,
1389        but BOOL is not.
1390    
1391    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1392        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1393    
1394    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1395        hyphens) following file names and line numbers when outputting matching
1396        lines. This is not true; no spaces are inserted. I have also clarified the
1397        wording for the --colour (or --color) option.
1398    
1399    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1400        was not coloured; this is different to GNU grep, so I have changed it to be
1401        the same.
1402    
1403    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1404        each matching line was coloured. Now it goes on to look for further matches
1405        of any of the test patterns, which is the same behaviour as GNU grep.
1406    
1407    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1408        doesn't make sense to accept an empty string match in pcregrep, so I have
1409        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1410        seems to be how GNU grep behaves.
1411    
1412    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1413        start or after a newline", because the conditional assertion was not being
1414        correctly handled. The rule now is that both the assertion and what follows
1415        in the first alternative must satisfy the test.
1416    
1417    9.  If auto-callout was enabled in a pattern with a conditional group whose
1418        condition was an assertion, PCRE could crash during matching, both with
1419        pcre_exec() and pcre_dfa_exec().
1420    
1421    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1422        used for matching.
1423    
1424    11. Unicode property support in character classes was not working for
1425        characters (bytes) greater than 127 when not in UTF-8 mode.
1426    
1427    12. Added the -M command line option to pcretest.
1428    
1429    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1430    
1431    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1432    
1433    16. Added comments and documentation about mis-use of no_arg in the C++
1434        wrapper.
1435    
1436    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1437        from Martin Jerabek that uses macro names for all relevant character and
1438        string constants.
1439    
1440    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1441        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1442        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1443        these, but not everybody uses configure.
1444    
1445    19. A conditional group that had only one branch was not being correctly
1446        recognized as an item that could match an empty string. This meant that an
1447        enclosing group might also not be so recognized, causing infinite looping
1448        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1449        with the subject "ab", where knowledge that the repeated group can match
1450        nothing is needed in order to break the loop.
1451    
1452    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1453        exec(), but without supplying a callout function, matching went wrong.
1454    
1455    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1456        leak if the size of the offset vector was greater than 30. When the vector
1457        is smaller, the saved offsets during recursion go onto a local stack
1458        vector, but for larger vectors malloc() is used. It was failing to free
1459        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1460        error, in fact).
1461    
1462    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1463        heapframe that is used only when UTF-8 support is enabled. This caused no
1464        problem, but was untidy.
1465    
1466    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1467        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1468        included within another project.
1469    
1470    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1471        slightly modified by me:
1472    
1473          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1474              not building pcregrep.
1475    
1476          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1477              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1478    
1479    25. Forward references, both numeric and by name, in patterns that made use of
1480        duplicate group numbers, could behave incorrectly or give incorrect errors,
1481        because when scanning forward to find the reference group, PCRE was not
1482        taking into account the duplicate group numbers. A pattern such as
1483        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1484    
1485    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1486        the feature of a custom pointer more persuasive (as requested by a user).
1487    
1488    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1489        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1490        pre-defined.
1491    
1492    28. Added support for (*UTF8) at the start of a pattern.
1493    
1494    29. Arrange for flags added by the "release type" setting in CMake to be shown
1495        in the configuration summary.
1496    
1497    
1498    Version 7.8 05-Sep-08
1499    ---------------------
1500    
1501    1.  Replaced UCP searching code with optimized version as implemented for Ad
1502        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1503        stage table and inline lookup instead of a function, giving speed ups of 2
1504        to 5 times on some simple patterns that I tested. Permission was given to
1505        distribute the MultiStage2.py script that generates the tables (it's not in
1506        the tarball, but is in the Subversion repository).
1507    
1508    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1509        scripts.
1510    
1511    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1512        a group with a zero qualifier. The result of the study could be incorrect,
1513        or the function might crash, depending on the pattern.
1514    
1515    4.  Caseless matching was not working for non-ASCII characters in back
1516        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1517        It now works when Unicode Property Support is available.
1518    
1519    5.  In pcretest, an escape such as \x{de} in the data was always generating
1520        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1521        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1522        truncation.
1523    
1524    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1525    
1526    7.  Added two (int) casts to pcregrep when printing the difference of two
1527        pointers, in case they are 64-bit values.
1528    
1529    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1530        test 2 if it fails.
1531    
1532    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1533        and a #define of that name to empty if it is not externally set. This is to
1534        allow users of MSVC to set it if necessary.
1535    
1536    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1537        the convenience functions in the pcre_get.c source file.
1538    
1539    11. An option change at the start of a pattern that had top-level alternatives
1540        could cause overwriting and/or a crash. This command provoked a crash in
1541        some environments:
1542    
1543          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1544    
1545        This potential security problem was recorded as CVE-2008-2371.
1546    
1547    12. For a pattern where the match had to start at the beginning or immediately
1548        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1549        pcre_dfa_exec() could read past the end of the passed subject if there was
1550        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1551        pcretest so that it places the subject at the end of its malloc-ed buffer.
1552    
1553    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1554        exec() might read past the end of the data buffer in UTF-8 mode.
1555    
1556    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1557        the data contained the byte 0x85 as part of a UTF-8 character within its
1558        first line. This applied both to normal and DFA matching.
1559    
1560    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1561        /^[^d]*?$/8 failed to match "abc".
1562    
1563    16. Added a missing copyright notice to pcrecpp_internal.h.
1564    
1565    17. Make it more clear in the documentation that values returned from
1566        pcre_exec() in ovector are byte offsets, not character counts.
1567    
1568    18. Tidied a few places to stop certain compilers from issuing warnings.
1569    
1570    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1571        supplied by Stefan Weber. I made a further small update for 7.8 because
1572        there is a change of source arrangements: the pcre_searchfuncs.c module is
1573        replaced by pcre_ucd.c.
1574    
1575    
1576    Version 7.7 07-May-08
1577    ---------------------
1578    
1579    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1580        a string to a long long, pretend we don't even have a long long." This is
1581        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1582    
1583    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1584        pre-7.6 versions, which defined a global no_arg variable instead of putting
1585        it in the RE class. (See also #8 below.)
1586    
1587    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1588        Lopes.
1589    
1590    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1591    
1592        (1) The include/exclude patterns were being applied to the whole pathnames
1593            of files, instead of just to the final components.
1594    
1595        (2) If there was more than one level of directory, the subdirectories were
1596            skipped unless they satisfied the include/exclude conditions. This is
1597            inconsistent with GNU grep (and could even be seen as contrary to the
1598            pcregrep specification - which I improved to make it absolutely clear).
1599            The action now is always to scan all levels of directory, and just
1600            apply the include/exclude patterns to regular files.
1601    
1602    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1603        --exclude_dir in the tests to avoid scanning .svn directories.
1604    
1605    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1606        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1607        doesn't support NULs in patterns.
1608    
1609    7.  Added some missing "const"s to declarations of static tables in
1610        pcre_compile.c and pcre_dfa_exec.c.
1611    
1612    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1613        caused by fix #2  above. (Subsequently also a second patch to fix the
1614        first patch. And a third patch - this was a messy problem.)
1615    
1616    9.  Applied Craig's patch to remove the use of push_back().
1617    
1618    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1619        matching function regexec().
1620    
1621    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1622        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1623        references. PCRE supports relative numbers with this syntax (I don't think
1624        Oniguruma does).
1625    
1626    12. Previously, a group with a zero repeat such as (...){0} was completely
1627        omitted from the compiled regex. However, this means that if the group
1628        was called as a subroutine from elsewhere in the pattern, things went wrong
1629        (an internal error was given). Such groups are now left in the compiled
1630        pattern, with a new opcode that causes them to be skipped at execution
1631        time.
1632    
1633    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1634        to the way PCRE behaves:
1635    
1636        (a) A lone ] character is dis-allowed (Perl treats it as data).
1637    
1638        (b) A back reference to an unmatched subpattern matches an empty string
1639            (Perl fails the current match path).
1640    
1641        (c) A data ] in a character class must be notated as \] because if the
1642            first data character in a class is ], it defines an empty class. (In
1643            Perl it is not possible to have an empty class.) The empty class []
1644            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1645            The negative empty class [^] matches any one character, independently
1646            of the DOTALL setting.
1647    
1648    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1649        non-existent subpattern following a character class starting with ']' and
1650        containing () gave an internal compiling error instead of "reference to
1651        non-existent subpattern". Fortunately, when the pattern did exist, the
1652        compiled code was correct. (When scanning forwards to check for the
1653        existencd of the subpattern, it was treating the data ']' as terminating
1654        the class, so got the count wrong. When actually compiling, the reference
1655        was subsequently set up correctly.)
1656    
1657    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1658        it was being rejected as not supported by pcre_dfa_exec(), even though
1659        other assertions are supported. I have made pcre_dfa_exec() support
1660        (*FAIL).
1661    
1662    16. The implementation of 13c above involved the invention of a new opcode,
1663        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1664        cannot be changed at match time, I realized I could make a small
1665        improvement to matching performance by compiling OP_ALLANY instead of
1666        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1667        on the OP_ANY path.
1668    
1669    17. Compiling pcretest on Windows with readline support failed without the
1670        following two fixes: (1) Make the unistd.h include conditional on
1671        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1672    
1673    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1674        ncurses library to be included for pcretest when ReadLine support is
1675        requested, but also to allow for it to be overridden. This patch came from
1676        Daniel Bergström.
1677    
1678    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1679        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1680        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1681        spotting this.
1682    
1683    
1684    Version 7.6 28-Jan-08
1685    ---------------------
1686    
1687    1.  A character class containing a very large number of characters with
1688        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1689        overflow.
1690    
1691    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1692        HAVE_LONG_LONG is not defined.
1693    
1694    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1695        bring them up to date and include new features. This patch includes:
1696    
1697        - Fixed PH's badly added libz and libbz2 support.
1698        - Fixed a problem with static linking.
1699        - Added pcredemo. [But later removed - see 7 below.]
1700        - Fixed dftables problem and added an option.
1701        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1702            HAVE_LONG_LONG.
1703        - Added readline support for pcretest.
1704        - Added an listing of the option settings after cmake has run.
1705    
1706    4.  A user submitted a patch to Makefile that makes it easy to create
1707        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1708        Makefile.am that cause it to include this special target, without
1709        affecting anything else. Note that the same mingw target plus all
1710        the other distribution libraries and programs are now supported
1711        when configuring with CMake (see 6 below) instead of with
1712        Configure/Make.
1713    
1714    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1715        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1716        exported in the Windows port". It has not yet been confirmed that the patch
1717        solves the problem, but it does no harm.
1718    
1719    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1720        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1721        with CMake, and also correct the comment about stack recursion.
1722    
1723    7.  Remove the automatic building of pcredemo from the ./configure system and
1724        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1725        of a program that users should build themselves after PCRE is installed, so
1726        building it automatically is not really right. What is more, it gave
1727        trouble in some build environments.
1728    
1729    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1730    
1731    
1732    Version 7.5 10-Jan-08
1733    ---------------------
1734    
1735    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1736        values in parens when parsing an RE using the C++ wrapper."
1737    
1738    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1739        Characters greater than 255 were excluded from the class instead of being
1740        included.
1741    
1742    3.  The same bug as (2) above applied to negated POSIX classes such as
1743        [:^space:].
1744    
1745    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1746        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1747        I have changed it.
1748    
1749    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1750        first named subpattern) and a construct such as (?&a) would reference the
1751        first named subpattern whose name started with "a" (in other words, the
1752        length check was missing). Both these problems are fixed. "Subpattern name
1753        expected" is now given for (?&) (a zero-length name), and this patch also
1754        makes it give the same error for \k'' (previously it complained that that
1755        was a reference to a non-existent subpattern).
1756    
1757    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1758        this is right because (?- can be followed by option settings as well as by
1759        digits. I have, however, made the messages clearer.
1760    
1761    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1762        than the number used in the conditional) now cause a compile-time error.
1763        This is actually not compatible with Perl, which accepts such patterns, but
1764        treats the conditional as always being FALSE (as PCRE used to), but it
1765        seems to me that giving a diagnostic is better.
1766    
1767    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1768        and messages.
1769    
1770    9.  Fix two occurrences of "backslash" in comments that should have been
1771        "backspace".
1772    
1773    10. Remove two redundant lines of code that can never be obeyed (their function
1774        was moved elsewhere).
1775    
1776    11. The program that makes PCRE's Unicode character property table had a bug
1777        which caused it to generate incorrect table entries for sequences of
1778        characters that have the same character type, but are in different scripts.
1779        It amalgamated them into a single range, with the script of the first of
1780        them. In other words, some characters were in the wrong script. There were
1781        thirteen such cases, affecting characters in the following ranges:
1782    
1783          U+002b0 - U+002c1
1784          U+0060c - U+0060d
1785          U+0061e - U+00612
1786          U+0064b - U+0065e
1787          U+0074d - U+0076d
1788          U+01800 - U+01805
1789          U+01d00 - U+01d77
1790          U+01d9b - U+01dbf
1791          U+0200b - U+0200f
1792          U+030fc - U+030fe
1793          U+03260 - U+0327f
1794          U+0fb46 - U+0fbb1
1795          U+10450 - U+1049d
1796    
1797    12. The -o option (show only the matching part of a line) for pcregrep was not
1798        compatible with GNU grep in that, if there was more than one match in a
1799        line, it showed only the first of them. It now behaves in the same way as
1800        GNU grep.
1801    
1802    13. If the -o and -v options were combined for pcregrep, it printed a blank
1803        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1804        does the same. The return code can be used to tell if there were any
1805        non-matching lines.
1806    
1807    14. Added --file-offsets and --line-offsets to pcregrep.
1808    
1809    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1810        infinitely looping recursion. The bug was that positive lookaheads were not
1811        being skipped when checking for a possible empty match (negative lookaheads
1812        and both kinds of lookbehind were skipped).
1813    
1814    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1815        inclusion of <windows.h> to before rather than after the definition of
1816        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1817    
1818    17. Specifying a possessive quantifier with a specific limit for a Unicode
1819        character property caused pcre_compile() to compile bad code, which led at
1820        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1821        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1822        caused the error; without that there was no problem.
1823    
1824    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1825    
1826    19. Added --enable-pcretest-libreadline.
1827    
1828    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1829        RE::GlobalReplace(). As a result, the number of replacements returned was
1830        double what it should be. I removed one of the increments, but Craig sent a
1831        later patch that removed the other one (the right fix) and added unit tests
1832        that check the return values (which was not done before).
1833    
1834    21. Several CMake things:
1835    
1836        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1837            the names libpcre and libpcreposix, not just pcre and pcreposix.
1838    
1839        (2) The above change means that pcretest and pcregrep are now correctly
1840            linked with the newly-built libraries, not previously installed ones.
1841    
1842        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1843    
1844    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1845        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1846        UTF-8 newline character). The key issue is that the pattern starts .*;
1847        this means that the match must be either at the beginning, or after a
1848        newline. The bug was in the code for advancing after a failed match and
1849        checking that the new position followed a newline. It was not taking
1850        account of UTF-8 characters correctly.
1851    
1852    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1853        character classes. PCRE was not treating the sequence [:...:] as a
1854        character class unless the ... were all letters. Perl, however, seems to
1855        allow any characters between [: and :], though of course it rejects as
1856        unknown any "names" that contain non-letters, because all the known class
1857        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1858        for example, whereas PCRE did not - it did not recognize a POSIX character
1859        class. This seemed a bit dangerous, so the code has been changed to be
1860        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1861        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1862        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1863        Perl does, and where it didn't before.
1864    
1865    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1866        Windows environments %n is disabled by default.
1867    
1868    
1869    Version 7.4 21-Sep-07
1870  ---------------------  ---------------------
1871    
1872  1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This  1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
# Line 33  Version 7.4 10-Sep-07 Line 1898  Version 7.4 10-Sep-07
1898    
1899  7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and  7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1900      strtoull to pcrecpp.cc to select the available functions in WIN32 when the      strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1901      windows.h file is present (where different names are used). [This was      windows.h file is present (where different names are used). [This was
1902      reversed later after testing - see 16 below.]      reversed later after testing - see 16 below.]
1903    
1904  8.  Changed all #include <config.h> to #include "config.h". There were also  8.  Changed all #include <config.h> to #include "config.h". There were also
# Line 63  Version 7.4 10-Sep-07 Line 1928  Version 7.4 10-Sep-07
1928    
1929  13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously  13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1930      checked only for CRLF.      checked only for CRLF.
1931    
1932  14. Added casts to pcretest.c to avoid compiler warnings.  14. Added casts to pcretest.c to avoid compiler warnings.
1933    
1934  15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.  15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1935    
1936  16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,  16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1937      and instead check for _strtoi64 explicitly, and avoid the use of snprintf()      and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1938      entirely. This removes changes made in 7 above.      entirely. This removes changes made in 7 above.
1939    
1940    17. The CMake files have been updated, and there is now more information about
1941        building with CMake in the NON-UNIX-USE document.
1942    
1943    
1944  Version 7.3 28-Aug-07  Version 7.3 28-Aug-07

Legend:
Removed from v.257  
changed lines
  Added in v.1024

  ViewVC Help
Powered by ViewVC 1.1.5