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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.324  
changed lines
  Added in v.1233

  ViewVC Help
Powered by ViewVC 1.1.5