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

Diff of /code/trunk/ChangeLog

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

revision 193 by ph10, Mon Jul 30 11:34:18 2007 UTC revision 1254 by ph10, Fri Feb 22 12:46:26 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 05-Jul-07  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    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
13        CMake config files.
14    
15    3.  Revise the creation of config.h.generic so that all boolean macros are
16        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17        overriding via -D on the command line possible.
18    
19    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20        to unsigned int is reported to make a quite noticeable speed difference in
21        a specific Windows environment. Testing on Linux did also appear to show
22        some benefit (and it is clearly not harmful). Also fixed the definition of
23        Xop which should be unsigned.
24    
25    5.  Related to (4), changing the definition of the intermediate variable cc
26        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27        improvements.
28    
29    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
30        unnecessary spaces.
31    
32    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33        and later.
34    
35    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36    
37    9.  Optimizing fast_forward_start_bits in JIT.
38    
39    10. Adding experimental support for callouts in JIT, and fixing some
40        issues revealed during this work. Namely:
41    
42        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43    
44        (b) Minimum length was not checked before the matching is started.
45    
46    11. The value of capture_last that is passed to callouts was incorrect in some
47        cases when there was a capture on one path that was subsequently abandoned
48        after a backtrack. Also, the capture_last value is now reset after a
49        recursion, since all captures are also reset in this case.
50    
51    12. The interpreter no longer returns the "too many substrings" error in the
52        case when an overflowing capture is in a branch that is subsequently
53        abandoned after a backtrack.
54    
55    13. In the pathological case when an offset vector of size 2 is used, pcretest
56        now prints out the matched string after a yield of 0 or 1.
57    
58    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59        Only supported by the JIT compiler at the moment.
60    
61    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62    
63    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
64        offset of the starting point of the matching process, provided the offsets
65        vector is large enough.
66    
67    17. The \A escape now records a lookbehind value of 1, though its execution
68        does not actually inspect the previous character. This is to ensure that,
69        in partial multi-segment matching, at least one character from the old
70        segment is retained when a new segment is processed. Otherwise, if there
71        are no lookbehinds in the pattern, \A might match incorrectly at the start
72        of a new segment.
73    
74    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75    
76    
77    Version 8.32 30-November-2012
78    -----------------------------
79    
80    1.  Improved JIT compiler optimizations for first character search and single
81        character iterators.
82    
83    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
84        Patch by Daniel Richard G.
85    
86    3.  Single character iterator optimizations in the JIT compiler.
87    
88    4.  Improved JIT compiler optimizations for character ranges.
89    
90    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
91        Reported by Giuseppe D'Angelo.
92    
93    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
94        of a line, was being set incorrectly in cases where .* appeared inside
95        atomic brackets at the start of a pattern, or where there was a subsequent
96        *PRUNE or *SKIP.
97    
98    7.  Improved instruction cache flush for POWER/PowerPC.
99        Patch by Daniel Richard G.
100    
101    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
102        grep:
103    
104        (a) There is now no limit to the number of patterns to be matched.
105    
106        (b) An error is given if a pattern is too long.
107    
108        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
109            are now supported.
110    
111        (d) --exclude-from and --include-from (multiple use) have been added.
112    
113        (e) Exclusions and inclusions now apply to all files and directories, not
114            just to those obtained from scanning a directory recursively.
115    
116        (f) Multiple uses of -f and --file-list are now supported.
117    
118        (g) In a Windows environment, the default for -d has been changed from
119            "read" (the GNU grep default) to "skip", because otherwise the presence
120            of a directory in the file list provokes an error.
121    
122        (h) The documentation has been revised and clarified in places.
123    
124    9.  Improve the matching speed of capturing brackets.
125    
126    10. Changed the meaning of \X so that it now matches a Unicode extended
127        grapheme cluster.
128    
129    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
130        out POSIX threads when JIT support is configured.
131    
132    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
133    
134    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
135        structure could go wrong in environments where size_t is not the same size
136        as int.
137    
138    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
139        set.
140    
141    15. The EBCDIC support had decayed; later updates to the code had included
142        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
143        general tidy up of EBCDIC-related issues, and the documentation was also
144        not quite right. There is now a test that can be run on ASCII systems to
145        check some of the EBCDIC-related things (but is it not a full test).
146    
147    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
148        in a small tidy to the code.
149    
150    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
151    
152    18. If the --only-matching (-o) option in pcregrep is specified multiple
153        times, each one causes appropriate output. For example, -o1 -o2 outputs the
154        substrings matched by the 1st and 2nd capturing parentheses. A separating
155        string can be specified by --om-separator (default empty).
156    
157    19. Improving the first n character searches.
158    
159    20. Turn case lists for horizontal and vertical white space into macros so that
160        they are defined only once.
161    
162    21. This set of changes together give more compatible Unicode case-folding
163        behaviour for characters that have more than one other case when UCP
164        support is available.
165    
166        (a) The Unicode property table now has offsets into a new table of sets of
167            three or more characters that are case-equivalent. The MultiStage2.py
168            script that generates these tables (the pcre_ucd.c file) now scans
169            CaseFolding.txt instead of UnicodeData.txt for character case
170            information.
171    
172        (b) The code for adding characters or ranges of characters to a character
173            class has been abstracted into a generalized function that also handles
174            case-independence. In UTF-mode with UCP support, this uses the new data
175            to handle characters with more than one other case.
176    
177        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
178            whose other case is greater than 256 are now correctly matched
179            caselessly. Previously, the high codepoint matched the low one, but not
180            vice versa.
181    
182        (d) The processing of \h, \H, \v, and \ in character classes now makes use
183            of the new class addition function, using character lists defined as
184            macros alongside the case definitions of 20 above.
185    
186        (e) Caseless back references now work with characters that have more than
187            one other case.
188    
189        (f) General caseless matching of characters with more than one other case
190            is supported.
191    
192    22. Unicode character properties were updated from Unicode 6.2.0
193    
194    23. Improved CMake support under Windows. Patch by Daniel Richard G.
195    
196    24. Add support for 32-bit character strings, and UTF-32
197    
198    25. Major JIT compiler update (code refactoring and bugfixing).
199        Experimental Sparc 32 support is added.
200    
201    26. Applied a modified version of Daniel Richard G's patch to create
202        pcre.h.generic and config.h.generic by "make" instead of in the
203        PrepareRelease script.
204    
205    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
206        pcre_compile.c when checking for a zero character.
207    
208    28. Introducing a native interface for JIT. Through this interface, the compiled
209        machine code can be directly executed. The purpose of this interface is to
210        provide fast pattern matching, so several sanity checks are not performed.
211        However, feature tests are still performed. The new interface provides
212        1.4x speedup compared to the old one.
213    
214    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
215        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
216        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
217    
218    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
219        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
220        the "old" RFC 2279). Instead, it ended up passing a negative length to
221        pcre_exec().
222    
223    31. Add support for GCC's visibility feature to hide internal functions.
224    
225    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
226        "unknown -C option" after outputting 0 or 1.
227    
228    33. There is now support for generating a code coverage report for the test
229        suite in environments where gcc is the compiler and lcov is installed. This
230        is mainly for the benefit of the developers.
231    
232    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
233        unaddressable using valgrind annotations, allowing valgrind to detect
234        invalid memory accesses. This is mainly for the benefit of the developers.
235    
236    25. (*UTF) can now be used to start a pattern in any of the three libraries.
237    
238    26. Give configure error if --enable-cpp but no C++ compiler found.
239    
240    
241    Version 8.31 06-July-2012
242    -------------------------
243    
244    1.  Fixing a wrong JIT test case and some compiler warnings.
245    
246    2.  Removed a bashism from the RunTest script.
247    
248    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
249        to unsigned type, result still unsigned" that was given by an MS compiler
250        on encountering the code "-sizeof(xxx)".
251    
252    4.  Partial matching support is added to the JIT compiler.
253    
254    5.  Fixed several bugs concerned with partial matching of items that consist
255        of more than one character:
256    
257        (a) /^(..)\1/ did not partially match "aba" because checking references was
258            done on an "all or nothing" basis. This also applied to repeated
259            references.
260    
261        (b) \R did not give a hard partial match if \r was found at the end of the
262            subject.
263    
264        (c) \X did not give a hard partial match after matching one or more
265            characters at the end of the subject.
266    
267        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
268            a partial match for the string "\r".
269    
270        (e) When newline was set to CRLF, the metacharacter "." did not recognize
271            a partial match for a CR character at the end of the subject string.
272    
273    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
274        running pcretest, the text "(JIT)" added to the output whenever JIT is
275        actually used to run the match.
276    
277    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
278        or /S+[+] with a digit between 1 and 7.
279    
280    8.  OP_NOT now supports any UTF character not just single-byte ones.
281    
282    9.  (*MARK) control verb is now supported by the JIT compiler.
283    
284    10. The command "./RunTest list" lists the available tests without actually
285        running any of them. (Because I keep forgetting what they all are.)
286    
287    11. Add PCRE_INFO_MAXLOOKBEHIND.
288    
289    12. Applied a (slightly modified) user-supplied patch that improves performance
290        when the heap is used for recursion (compiled with --disable-stack-for-
291        recursion). Instead of malloc and free for each heap frame each time a
292        logical recursion happens, frames are retained on a chain and re-used where
293        possible. This sometimes gives as much as 30% improvement.
294    
295    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
296        call.
297    
298    14. As documented, (*COMMIT) is now confined to within a positive assertion.
299    
300    15. It is now possible to link pcretest with libedit as an alternative to
301        libreadline.
302    
303    16. (*COMMIT) control verb is now supported by the JIT compiler.
304    
305    17. The Unicode data tables have been updated to Unicode 6.1.0.
306    
307    18. Added --file-list option to pcregrep.
308    
309    19. Added binary file support to pcregrep, including the -a, --binary-files,
310        -I, and --text options.
311    
312    20. The madvise function is renamed for posix_madvise for QNX compatibility
313        reasons. Fixed by Giuseppe D'Angelo.
314    
315    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
316    
317    22. Changed the option for creating symbolic links for 16-bit man pages from
318        -s to -sf so that re-installing does not cause issues.
319    
320    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
321    
322    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
323        matches in certain environments (the workspace was not being correctly
324        retained). Also added to pcre_dfa_exec() a simple plausibility check on
325        some of the workspace data at the beginning of a restart.
326    
327    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
328        was not doing so when it should - probably a typo introduced by SVN 528
329        (change 8.10/14).
330    
331    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
332        \w+ when the character tables indicated that \x{c4} was a word character.
333        There were several related cases, all because the tests for doing a table
334        lookup were testing for characters less than 127 instead of 255.
335    
336    27. If a pattern contains capturing parentheses that are not used in a match,
337        their slots in the ovector are set to -1. For those that are higher than
338        any matched groups, this happens at the end of processing. In the case when
339        there were back references that the ovector was too small to contain
340        (causing temporary malloc'd memory to be used during matching), and the
341        highest capturing number was not used, memory off the end of the ovector
342        was incorrectly being set to -1. (It was using the size of the temporary
343        memory instead of the true size.)
344    
345    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
346        ovector size, it uses memory at the end of the block that it has got.
347    
348    29. Check for an overlong MARK name and give an error at compile time. The
349        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
350    
351    30. JIT compiler update.
352    
353    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
354        Rill for the patch.
355    
356    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
357    
358    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
359    
360    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
361        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
362        was enabled.
363    
364    35. Improve JIT code generation for greedy plus quantifier.
365    
366    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
367        1 to "aa" instead of to an empty string. The bug affected repeated groups
368        that could potentially match an empty string.
369    
370    37. Optimizing single character iterators in JIT.
371    
372    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
373        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
374        codepoints that are too big for the mode are faulted, and in a UTF mode,
375        disallowed codepoints are also faulted.
376    
377    39. If PCRE was compiled with UTF support, in three places in the DFA
378        matcher there was code that should only have been obeyed in UTF mode, but
379        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
380        processing when bytes with values greater than 127 were present. In 16-bit
381        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
382        both cases the values are those that cannot be the first data item in a UTF
383        character. The three items that might have provoked this were recursions,
384        possessively repeated groups, and atomic groups.
385    
386    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
387        and pcregrep, because some OS require shared objects to be explicitly
388        passed to ld, causing the link step to fail if they are not.
389    
390    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
391        mode, patterns that started with \h* or \R* might be incorrectly matched.
392    
393    
394    Version 8.30 04-February-2012
395    -----------------------------
396    
397    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
398        name is defined in ctype.h.
399    
400    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
401        only in quite long subpatterns.
402    
403    3.  Removed the function pcre_info(), which has been obsolete and deprecated
404        since it was replaced by pcre_fullinfo() in February 2000.
405    
406    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
407        match a (*MARK), and the match failed at the start of the subject, a
408        reference to memory before the start of the subject could occur. This bug
409        was introduced by fix 17 of release 8.21.
410    
411    5.  A reference to an unset group with zero minimum repetition was giving
412        totally wrong answers (in non-JavaScript-compatibility mode). For example,
413        /(another)?(\1?)test/ matched against "hello world test". This bug was
414        introduced in release 8.13.
415    
416    6.  Add support for 16-bit character strings (a large amount of work involving
417        many changes and refactorings).
418    
419    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
420        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
421        from a file.
422    
423    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
424        rounding is not applied in this particular case).
425    
426    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
427        if they appear, or are escaped, in patterns.
428    
429    10. Get rid of a number of -Wunused-but-set-variable warnings.
430    
431    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
432        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
433        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
434        also returns the mark "x". This bug applied to capturing parentheses,
435        non-capturing parentheses, and atomic parentheses. It also applied to some
436        assertions.
437    
438    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
439        information out of configure.ac instead of relying on pcre.h.generic, which
440        is not stored in the repository.
441    
442    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
443        -lreadline.
444    
445    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
446    
447    15. Applied Graycode's patch to put the top-level frame on the stack rather
448        than the heap when not using the stack for recursion. This gives a
449        performance improvement in many cases when recursion is not deep.
450    
451    16. Experimental code added to "pcretest -C" to output the stack frame size.
452    
453    
454    Version 8.21 12-Dec-2011
455    ------------------------
456    
457    1.  Updating the JIT compiler.
458    
459    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
460        are added as well.
461    
462    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
463        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
464        calling _pcre_jit_exec. Some extra comments are added.
465    
466    4.  (*MARK) settings inside atomic groups that do not contain any capturing
467        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
468        was introduced by change 18 for 8.20.
469    
470    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
471        ECMA-262 standard.
472    
473    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
474        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
475        This bug was probably introduced by change 9 of 8.13.
476    
477    7.  While fixing 6 above, I noticed that a number of other items were being
478        incorrectly rejected as "not fixed length". This arose partly because newer
479        opcodes had not been added to the fixed-length checking code. I have (a)
480        corrected the bug and added tests for these items, and (b) arranged for an
481        error to occur if an unknown opcode is encountered while checking for fixed
482        length instead of just assuming "not fixed length". The items that were
483        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
484        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
485        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
486    
487    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
488        being incorrectly compiled and would have given unpredicatble results.
489    
490    9.  A possessively repeated subpattern with minimum repeat count greater than
491        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
492        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
493        the first (A) could occur when it should not.
494    
495    10. Add a cast and remove a redundant test from the code.
496    
497    11. JIT should use pcre_malloc/pcre_free for allocation.
498    
499    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
500        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
501        is anything other than /usr, -L is still shown).
502    
503    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
504    
505    14. Perl does not support \N without a following name in a [] class; PCRE now
506        also gives an error.
507    
508    15. If a forward reference was repeated with an upper limit of around 2000,
509        it caused the error "internal error: overran compiling workspace". The
510        maximum number of forward references (including repeats) was limited by the
511        internal workspace, and dependent on the LINK_SIZE. The code has been
512        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
513        the default depends on LINK_SIZE. There is a new upper limit (for safety)
514        of around 200,000 forward references. While doing this, I also speeded up
515        the filling in of repeated forward references.
516    
517    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
518        incorrectly expecting the subject to contain another "a" after the start.
519    
520    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
521        in the match, the SKIP should be ignored. This was not happening; instead
522        the SKIP was being treated as NOMATCH. For patterns such as
523        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
524        tested.
525    
526    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
527        now much more compatible with Perl, in particular in cases where the result
528        is a non-match for a non-anchored pattern. For example, if
529        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
530        "m", where previously it did not return a name. A side effect of this
531        change is that for partial matches, the last encountered mark name is
532        returned, as for non matches. A number of tests that were previously not
533        Perl-compatible have been moved into the Perl-compatible test files. The
534        refactoring has had the pleasing side effect of removing one argument from
535        the match() function, thus reducing its stack requirements.
536    
537    19. If the /S+ option was used in pcretest to study a pattern using JIT,
538        subsequent uses of /S (without +) incorrectly behaved like /S+.
539    
540    21. Retrieve executable code size support for the JIT compiler and fixing
541        some warnings.
542    
543    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
544        not work when the shorter character appeared right at the end of the
545        subject string.
546    
547    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
548        systems.
549    
550    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
551        output it when the /M option is used in pcretest.
552    
553    25. The CheckMan script was not being included in the distribution. Also, added
554        an explicit "perl" to run Perl scripts from the PrepareRelease script
555        because this is reportedly needed in Windows.
556    
557    26. If study data was being save in a file and studying had not found a set of
558        "starts with" bytes for the pattern, the data written to the file (though
559        never used) was taken from uninitialized memory and so caused valgrind to
560        complain.
561    
562    27. Updated RunTest.bat as provided by Sheri Pierce.
563    
564    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
565    
566    29. Computation of memory usage for the table of capturing group names was
567        giving an unnecessarily large value.
568    
569    
570    Version 8.20 21-Oct-2011
571    ------------------------
572    
573    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
574        a POSIX class. After further experiments with Perl, which convinced me that
575        Perl has bugs and confusions, a closing square bracket is no longer allowed
576        in a POSIX name. This bug also affected patterns with classes that started
577        with full stops.
578    
579    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
580        captured substring, but while checking the failing first alternative,
581        substring 1 is temporarily captured. If the output vector supplied to
582        pcre_exec() was not big enough for this capture, the yield of the function
583        was still zero ("insufficient space for captured substrings"). This cannot
584        be totally fixed without adding another stack variable, which seems a lot
585        of expense for a edge case. However, I have improved the situation in cases
586        such as /(a)(b)x|abc/ matched against "abc", where the return code
587        indicates that fewer than the maximum number of slots in the ovector have
588        been set.
589    
590    3.  Related to (2) above: when there are more back references in a pattern than
591        slots in the output vector, pcre_exec() uses temporary memory during
592        matching, and copies in the captures as far as possible afterwards. It was
593        using the entire output vector, but this conflicts with the specification
594        that only 2/3 is used for passing back captured substrings. Now it uses
595        only the first 2/3, for compatibility. This is, of course, another edge
596        case.
597    
598    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
599        main code base, and can be used by building with --enable-jit. When this is
600        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
601        runtime --no-jit option is given.
602    
603    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
604        ovector, the return from the function was zero, implying that there were
605        other matches that did not fit. The correct "exactly full" value is now
606        returned.
607    
608    6.  If a subpattern that was called recursively or as a subroutine contained
609        (*PRUNE) or any other control that caused it to give a non-standard return,
610        invalid errors such as "Error -26 (nested recursion at the same subject
611        position)" or even infinite loops could occur.
612    
613    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
614        computing the minimum length on reaching *ACCEPT, and so ended up with the
615        wrong value of 1 rather than 0. Further investigation indicates that
616        computing a minimum subject length in the presence of *ACCEPT is difficult
617        (think back references, subroutine calls), and so I have changed the code
618        so that no minimum is registered for a pattern that contains *ACCEPT.
619    
620    8.  If (*THEN) was present in the first (true) branch of a conditional group,
621        it was not handled as intended. [But see 16 below.]
622    
623    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
624        Sheri Pierce.
625    
626    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
627        the first byte in a match must be "a".
628    
629    11. Change 17 for 8.13 increased the recursion depth for patterns like
630        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
631        pattern contains any instances of (*THEN). If it does not, the old
632        optimizations are restored. It would be nice to do this on a per-group
633        basis, but at the moment that is not feasible.
634    
635    12. In some environments, the output of pcretest -C is CRLF terminated. This
636        broke RunTest's code that checks for the link size. A single white space
637        character after the value is now allowed for.
638    
639    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
640        For "fr", it uses the Windows-specific input and output files.
641    
642    14. If (*THEN) appeared in a group that was called recursively or as a
643        subroutine, it did not work as intended. [But see next item.]
644    
645    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
646        pattern fragments (but not containing any | characters). If A and B are
647        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
648        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
649        D. In other words, Perl considers parentheses that do not contain any |
650        characters to be part of a surrounding alternative, whereas PCRE was
651        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
652        differently. PCRE now behaves in the same way as Perl, except in the case
653        of subroutine/recursion calls such as (?1) which have in any case always
654        been different (but PCRE had them first :-).
655    
656    16. Related to 15 above: Perl does not treat the | in a conditional group as
657        creating alternatives. Such a group is treated in the same way as an
658        ordinary group without any | characters when processing (*THEN). PCRE has
659        been changed to match Perl's behaviour.
660    
661    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
662        RunGrepTest script failed.
663    
664    18. Change 22 for version 13 caused atomic groups to use more stack. This is
665        inevitable for groups that contain captures, but it can lead to a lot of
666        stack use in large patterns. The old behaviour has been restored for atomic
667        groups that do not contain any capturing parentheses.
668    
669    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
670        suppress the check for a minimum subject length at run time. (If it was
671        given to pcre_exec() or pcre_dfa_exec() it did work.)
672    
673    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
674        fail to work when decoding hex characters in data strings in EBCDIC
675        environments.
676    
677    21. It appears that in at least one Mac OS environment, the isxdigit() function
678        is implemented as a macro that evaluates to its argument more than once,
679        contravening the C 90 Standard (I haven't checked a later standard). There
680        was an instance in pcretest which caused it to go wrong when processing
681        \x{...} escapes in subject strings. The has been rewritten to avoid using
682        things like p++ in the argument of isxdigit().
683    
684    
685    Version 8.13 16-Aug-2011
686    ------------------------
687    
688    1.  The Unicode data tables have been updated to Unicode 6.0.0.
689    
690    2.  Two minor typos in pcre_internal.h have been fixed.
691    
692    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
693        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
694        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
695    
696    4.  There were a number of related bugs in the code for matching backrefences
697        caselessly in UTF-8 mode when codes for the characters concerned were
698        different numbers of bytes. For example, U+023A and U+2C65 are an upper
699        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
700        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
701        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
702        2-byte code at the end of the subject (it thought there wasn't enough data
703        left).
704    
705    5.  Comprehensive information about what went wrong is now returned by
706        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
707        as the output vector has at least 2 elements. The offset of the start of
708        the failing character and a reason code are placed in the vector.
709    
710    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
711        now returned is for the first byte of the failing character, instead of the
712        last byte inspected. This is an incompatible change, but I hope it is small
713        enough not to be a problem. It makes the returned offset consistent with
714        pcre_exec() and pcre_dfa_exec().
715    
716    7.  pcretest now gives a text phrase as well as the error number when
717        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
718        failure, the offset and reason code are output.
719    
720    8.  When \R was used with a maximizing quantifier it failed to skip backwards
721        over a \r\n pair if the subsequent match failed. Instead, it just skipped
722        back over a single character (\n). This seems wrong (because it treated the
723        two characters as a single entity when going forwards), conflicts with the
724        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
725        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
726        has been changed.
727    
728    9.  Some internal refactoring has changed the processing so that the handling
729        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
730        time (the PCRE_DOTALL option was changed this way some time ago: version
731        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
732        which was always a bit of a fudge. It also means that there is one less
733        argument for the match() function, which reduces its stack requirements
734        slightly. This change also fixes an incompatibility with Perl: the pattern
735        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
736    
737    10. More internal refactoring has drastically reduced the number of recursive
738        calls to match() for possessively repeated groups such as (abc)++ when
739        using pcre_exec().
740    
741    11. While implementing 10, a number of bugs in the handling of groups were
742        discovered and fixed:
743    
744        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
745        (a|)*(?1) gave a compile-time internal error.
746        ((a|)+)+  did not notice that the outer group could match an empty string.
747        (^a|^)+   was not marked as anchored.
748        (.*a|.*)+ was not marked as matching at start or after a newline.
749    
750    12. Yet more internal refactoring has removed another argument from the match()
751        function. Special calls to this function are now indicated by setting a
752        value in a variable in the "match data" data block.
753    
754    13. Be more explicit in pcre_study() instead of relying on "default" for
755        opcodes that mean there is no starting character; this means that when new
756        ones are added and accidentally left out of pcre_study(), testing should
757        pick them up.
758    
759    14. The -s option of pcretest has been documented for ages as being an old
760        synonym of -m (show memory usage). I have changed it to mean "force study
761        for every regex", that is, assume /S for every regex. This is similar to -i
762        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
763        using it. It makes it easier to run collections of tests with and without
764        study enabled, and thereby test pcre_study() more easily. All the standard
765        tests are now run with and without -s (but some patterns can be marked as
766        "never study" - see 20 below).
767    
768    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
769        restoration of the capturing data to the outer values was not happening
770        correctly.
771    
772    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
773        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
774        pattern had matched an empty string, and so incorrectly returned a no
775        match.
776    
777    17. There was optimizing code for the last branch of non-capturing parentheses,
778        and also for the obeyed branch of a conditional subexpression, which used
779        tail recursion to cut down on stack usage. Unfortunately, now that there is
780        the possibility of (*THEN) occurring in these branches, tail recursion is
781        no longer possible because the return has to be checked for (*THEN). These
782        two optimizations have therefore been removed. [But see 8.20/11 above.]
783    
784    18. If a pattern containing \R was studied, it was assumed that \R always
785        matched two bytes, thus causing the minimum subject length to be
786        incorrectly computed because \R can also match just one byte.
787    
788    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
789        was incorrectly computed.
790    
791    20. If /S is present twice on a test pattern in pcretest input, it now
792        *disables* studying, thereby overriding the use of -s on the command line
793        (see 14 above). This is necessary for one or two tests to keep the output
794        identical in both cases.
795    
796    21. When (*ACCEPT) was used in an assertion that matched an empty string and
797        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
798    
799    22. When an atomic group that contained a capturing parenthesis was
800        successfully matched, but the branch in which it appeared failed, the
801        capturing was not being forgotten if a higher numbered group was later
802        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
803        group 1 to "a", when in fact it should be unset. This applied to multi-
804        branched capturing and non-capturing groups, repeated or not, and also to
805        positive assertions (capturing in negative assertions does not happen
806        in PCRE) and also to nested atomic groups.
807    
808    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
809        subject after a captured substring, to make it easier to tell which of a
810        number of identical substrings has been captured.
811    
812    24. The way atomic groups are processed by pcre_exec() has been changed so that
813        if they are repeated, backtracking one repetition now resets captured
814        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
815        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
816        "aaa". Previously, it would have been "a". As part of this code
817        refactoring, the way recursive calls are handled has also been changed.
818    
819    25. If an assertion condition captured any substrings, they were not passed
820        back unless some other capturing happened later. For example, if
821        (?(?=(a))a) was matched against "a", no capturing was returned.
822    
823    26. When studying a pattern that contained subroutine calls or assertions,
824        the code for finding the minimum length of a possible match was handling
825        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
826        group 1 called group 2 while simultaneously a separate group 2 called group
827        1). A stack overflow occurred in this case. I have fixed this by limiting
828        the recursion depth to 10.
829    
830    27. Updated RunTest.bat in the distribution to the version supplied by Tom
831        Fortmann. This supports explicit test numbers on the command line, and has
832        argument validation and error reporting.
833    
834    28. An instance of \X with an unlimited repeat could fail if at any point the
835        first character it looked at was a mark character.
836    
837    29. Some minor code refactoring concerning Unicode properties and scripts
838        should reduce the stack requirement of match() slightly.
839    
840    30. Added the '=' option to pcretest to check the setting of unused capturing
841        slots at the end of the pattern, which are documented as being -1, but are
842        not included in the return count.
843    
844    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
845        compiled something random. Now it gives a compile-time error (as does
846        Perl).
847    
848    32. A *MARK encountered during the processing of a positive assertion is now
849        recorded and passed back (compatible with Perl).
850    
851    33. If --only-matching or --colour was set on a pcregrep call whose pattern
852        had alternative anchored branches, the search for a second match in a line
853        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
854        matched the line "0102" twice. The same bug affected patterns that started
855        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
856        twice.
857    
858    34. Previously, PCRE did not allow quantification of assertions. However, Perl
859        does, and because of capturing effects, quantifying parenthesized
860        assertions may at times be useful. Quantifiers are now allowed for
861        parenthesized assertions.
862    
863    35. A minor code tidy in pcre_compile() when checking options for \R usage.
864    
865    36. \g was being checked for fancy things in a character class, when it should
866        just be a literal "g".
867    
868    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
869        appearance of a nested POSIX class supersedes an apparent external class.
870        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
871        unescaped square brackets may also appear as part of class names. For
872        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
873        more like Perl. (But see 8.20/1 above.)
874    
875    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
876        was because it thought it was \N{name}, which is not supported).
877    
878    39. Add minix to OS list not supporting the -S option in pcretest.
879    
880    40. PCRE tries to detect cases of infinite recursion at compile time, but it
881        cannot analyze patterns in sufficient detail to catch mutual recursions
882        such as ((?1))((?2)). There is now a runtime test that gives an error if a
883        subgroup is called recursively as a subpattern for a second time at the
884        same position in the subject string. In previous releases this might have
885        been caught by the recursion limit, or it might have run out of stack.
886    
887    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
888        happen only once. PCRE was, however incorrectly giving a compile time error
889        "recursive call could loop indefinitely" because it cannot analyze the
890        pattern in sufficient detail. The compile time test no longer happens when
891        PCRE is compiling a conditional subpattern, but actual runaway loops are
892        now caught at runtime (see 40 above).
893    
894    42. It seems that Perl allows any characters other than a closing parenthesis
895        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
896        has been changed to be the same.
897    
898    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
899        as not to get warnings when autogen.sh is called. Also changed
900        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
901    
902    44. To help people who use pcregrep to scan files containing exceedingly long
903        lines, the following changes have been made:
904    
905        (a) The default value of the buffer size parameter has been increased from
906            8K to 20K. (The actual buffer used is three times this size.)
907    
908        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
909            PCRE is built.
910    
911        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
912            to be set at run time.
913    
914        (d) Numerical values in pcregrep options can be followed by K or M, for
915            example --buffer-size=50K.
916    
917        (e) If a line being scanned overflows pcregrep's buffer, an error is now
918            given and the return code is set to 2.
919    
920    45. Add a pointer to the latest mark to the callout data block.
921    
922    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
923        partial match of an empty string instead of no match. This was specific to
924        the use of ".".
925    
926    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
927        complete match instead of a partial match. This bug was dependent on both
928        the PCRE_UTF8 and PCRE_DOTALL options being set.
929    
930    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
931        starting byte set, because \b was not being ignored.
932    
933    
934    Version 8.12 15-Jan-2011
935    ------------------------
936    
937    1.  Fixed some typos in the markup of the man pages, and wrote a script that
938        checks for such things as part of the documentation building process.
939    
940    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
941        --match-limit and --recursion-limit options (added for 8.11). In
942        particular, this made one of the standard tests fail. (The integer value
943        went into the wrong half of a long int.)
944    
945    3.  If the --colour option was given to pcregrep with -v (invert match), it
946        did strange things, either producing crazy output, or crashing. It should,
947        of course, ignore a request for colour when reporting lines that do not
948        match.
949    
950    4.  Another pcregrep bug caused similar problems if --colour was specified with
951        -M (multiline) and the pattern match finished with a line ending.
952    
953    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
954        matched in multiline mode, the following line was shown as part of the
955        match. This seems wrong, so I have changed it.
956    
957    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
958        the check for further matches in the same line (so they could be coloured)
959        to overrun the end of the current line. If another match was found, it was
960        incorrectly shown (and then shown again when found in the next line).
961    
962    7.  If pcregrep was compiled under Windows, there was a reference to the
963        function pcregrep_exit() before it was defined. I am assuming this was
964        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
965        reported by a user. I've moved the definition above the reference.
966    
967    
968    Version 8.11 10-Dec-2010
969    ------------------------
970    
971    1.  (*THEN) was not working properly if there were untried alternatives prior
972        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
973        backtracked to try for "b" instead of moving to the next alternative branch
974        at the same level (in this case, to look for "c"). The Perl documentation
975        is clear that when (*THEN) is backtracked onto, it goes to the "next
976        alternative in the innermost enclosing group".
977    
978    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
979        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
980        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
981        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
982        (*THEN).
983    
984    3.  If \s appeared in a character class, it removed the VT character from
985        the class, even if it had been included by some previous item, for example
986        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
987        of \s, but is part of the POSIX "space" class.)
988    
989    4.  A partial match never returns an empty string (because you can always
990        match an empty string at the end of the subject); however the checking for
991        an empty string was starting at the "start of match" point. This has been
992        changed to the "earliest inspected character" point, because the returned
993        data for a partial match starts at this character. This means that, for
994        example, /(?<=abc)def/ gives a partial match for the subject "abc"
995        (previously it gave "no match").
996    
997    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
998        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
999        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1000        has an implication that the given string is incomplete (because a partial
1001        match is preferred over a full match). For this reason, these items now
1002        give a partial match in this situation. [Aside: previously, the one case
1003        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1004        match rather than a full match, which was wrong by the old rules, but is
1005        now correct.]
1006    
1007    6.  There was a bug in the handling of #-introduced comments, recognized when
1008        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1009        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1010        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1011        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1012        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1013        places in pcre_compile().
1014    
1015    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1016        comments when looking ahead for named forward references to subpatterns,
1017        the only newline sequence it recognized was NL. It now handles newlines
1018        according to the set newline convention.
1019    
1020    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1021        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1022        cater for a lack of strerror(). These oversights have been fixed.
1023    
1024    9.  Added --match-limit and --recursion-limit to pcregrep.
1025    
1026    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1027    
1028    11. When the -o option was used, pcregrep was setting a return code of 1, even
1029        when matches were found, and --line-buffered was not being honoured.
1030    
1031    12. Added an optional parentheses number to the -o and --only-matching options
1032        of pcregrep.
1033    
1034    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1035        can match an empty string. The code to do it in pcretest and pcredemo
1036        needed fixing:
1037    
1038        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1039            only one byte after an empty string match just before CRLF (this case
1040            just got forgotten; "any" and "anycrlf" were OK).
1041    
1042        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1043            mode when an empty string match preceded an ASCII character followed by
1044            a non-ASCII character. (The code for advancing by one character rather
1045            than one byte was nonsense.)
1046    
1047        (c) The pcredemo.c sample program did not have any code at all to handle
1048            the cases when CRLF is a valid newline sequence.
1049    
1050    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1051        as a starting offset was within the subject string. There is now a new
1052        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1053        negative or greater than the length of the string. In order to test this,
1054        pcretest is extended to allow the setting of negative starting offsets.
1055    
1056    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1057        starting offset points to the beginning of a UTF-8 character was
1058        unnecessarily clumsy. I tidied it up.
1059    
1060    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1061        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1062    
1063    17. Nobody had reported that the --include_dir option, which was added in
1064        release 7.7 should have been called --include-dir (hyphen, not underscore)
1065        for compatibility with GNU grep. I have changed it to --include-dir, but
1066        left --include_dir as an undocumented synonym, and the same for
1067        --exclude-dir, though that is not available in GNU grep, at least as of
1068        release 2.5.4.
1069    
1070    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1071        characters from a string of bytes) have been redefined so as not to use
1072        loops, in order to improve performance in some environments. At the same
1073        time, I abstracted some of the common code into auxiliary macros to save
1074        repetition (this should not affect the compiled code).
1075    
1076    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1077        compile-time error is now given if \c is not followed by an ASCII
1078        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1079        different, and any byte value is allowed.)
1080    
1081    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1082        START_OPTIMIZE option, which is now allowed at compile time - but just
1083        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1084        to pcregrep and other applications that have no direct access to PCRE
1085        options. The new /Y option in pcretest sets this option when calling
1086        pcre_compile().
1087    
1088    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1089        back references. Groups containing recursive back references were forced to
1090        be atomic by that change, but in the case of named groups, the amount of
1091        memory required was incorrectly computed, leading to "Failed: internal
1092        error: code overflow". This has been fixed.
1093    
1094    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1095        pcretest.c, to avoid build problems in some Borland environments.
1096    
1097    
1098    Version 8.10 25-Jun-2010
1099    ------------------------
1100    
1101    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1102        THEN.
1103    
1104    2.  (*ACCEPT) was not working when inside an atomic group.
1105    
1106    3.  Inside a character class, \B is treated as a literal by default, but
1107        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1108        causes the error). The code is unchanged, but I tidied the documentation.
1109    
1110    4.  Inside a character class, PCRE always treated \R and \X as literals,
1111        whereas Perl faults them if its -w option is set. I have changed PCRE so
1112        that it faults them when PCRE_EXTRA is set.
1113    
1114    5.  Added support for \N, which always matches any character other than
1115        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1116    
1117    6.  When compiling pcregrep with newer versions of gcc which may have
1118        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1119        declared with attribute warn_unused_result" were given. Just casting the
1120        result to (void) does not stop the warnings; a more elaborate fudge is
1121        needed. I've used a macro to implement this.
1122    
1123    7.  Minor change to pcretest.c to avoid a compiler warning.
1124    
1125    8.  Added four artifical Unicode properties to help with an option to make
1126        \s etc use properties (see next item). The new properties are: Xan
1127        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1128    
1129    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1130        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1131        this option. Modified pcretest to add /W to test this facility. Added
1132        REG_UCP to make it available via the POSIX interface.
1133    
1134    10. Added --line-buffered to pcregrep.
1135    
1136    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1137        studied, and the match started with a letter with a code point greater than
1138        127 whose first byte was different to the first byte of the other case of
1139        the letter, the other case of this starting letter was not recognized
1140        (#976).
1141    
1142    12. If a pattern that was studied started with a repeated Unicode property
1143        test, for example, \p{Nd}+, there was the theoretical possibility of
1144        setting up an incorrect bitmap of starting bytes, but fortunately it could
1145        not have actually happened in practice until change 8 above was made (it
1146        added property types that matched character-matching opcodes).
1147    
1148    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1149        possible starting bytes for non-anchored patterns.
1150    
1151    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1152        \R, and also a number of cases that involve Unicode properties, both
1153        explicit and implicit when PCRE_UCP is set.
1154    
1155    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1156        input, it could crash or give wrong results if characters with values
1157        greater than 0xc0 were present in the subject string. (Detail: it assumed
1158        UTF-8 input when processing these items.)
1159    
1160    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1161        size_t is 64-bit (#991).
1162    
1163    17. Added a check for running out of memory when PCRE is compiled with
1164        --disable-stack-for-recursion (#990).
1165    
1166    18. If the last data line in a file for pcretest does not have a newline on
1167        the end, a newline was missing in the output.
1168    
1169    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1170        less than 128) in its various bitmaps. However, there is a facility for
1171        generating tables according to the current locale when PCRE is compiled. It
1172        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1173        characters, are recognized by isspace() and therefore were getting set in
1174        these tables, and indeed these tables seem to approximate to ISO 8859. This
1175        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1176        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1177        which of course cannot start UTF-8 characters. I have changed the code so
1178        that only real ASCII characters (less than 128) and the correct starting
1179        bytes for UTF-8 encodings are set for characters greater than 127 when in
1180        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1181        altogether.)
1182    
1183    20. Added the /T option to pcretest so as to be able to run tests with non-
1184        standard character tables, thus making it possible to include the tests
1185        used for 19 above in the standard set of tests.
1186    
1187    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1188        reference to a subpattern the other side of a comment that contains an
1189        opening parenthesis caused either an internal compiling error, or a
1190        reference to the wrong subpattern.
1191    
1192    
1193    Version 8.02 19-Mar-2010
1194    ------------------------
1195    
1196    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1197    
1198    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1199        configured.
1200    
1201    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1202        original author of that file, following a query about its status.
1203    
1204    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1205        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1206    
1207    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1208        quantifier applied to a forward-referencing subroutine call, could compile
1209        incorrect code or give the error "internal error: previously-checked
1210        referenced subpattern not found".
1211    
1212    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1213        variables to point to external functions. For these systems, therefore,
1214        pcre_malloc etc. are now initialized to local functions that call the
1215        relevant global functions.
1216    
1217    7.  There were two entries missing in the vectors called coptable and poptable
1218        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1219        I've fixed the data, and added a kludgy way of testing at compile time that
1220        the lengths are correct (equal to the number of opcodes).
1221    
1222    8.  Following on from 7, I added a similar kludge to check the length of the
1223        eint vector in pcreposix.c.
1224    
1225    9.  Error texts for pcre_compile() are held as one long string to avoid too
1226        much relocation at load time. To find a text, the string is searched,
1227        counting zeros. There was no check for running off the end of the string,
1228        which could happen if a new error number was added without updating the
1229        string.
1230    
1231    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1232    
1233    11. \K was not working if it appeared in an atomic group or in a group that
1234        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1235        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1236        the assertion is positive, but not if it is negative.
1237    
1238    12. Change 11 fortuitously reduced the size of the stack frame used in the
1239        "match()" function of pcre_exec.c by one pointer. Forthcoming
1240        implementation of support for (*MARK) will need an extra pointer on the
1241        stack; I have reserved it now, so that the stack frame size does not
1242        decrease.
1243    
1244    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1245        item in branch that calls a recursion is a subroutine call - as in the
1246        second branch in the above example - was incorrectly given the compile-
1247        time error "recursive call could loop indefinitely" because pcre_compile()
1248        was not correctly checking the subroutine for matching a non-empty string.
1249    
1250    14. The checks for overrunning compiling workspace could trigger after an
1251        overrun had occurred. This is a "should never occur" error, but it can be
1252        triggered by pathological patterns such as hundreds of nested parentheses.
1253        The checks now trigger 100 bytes before the end of the workspace.
1254    
1255    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1256    
1257    
1258    Version 8.01 19-Jan-2010
1259    ------------------------
1260    
1261    1.  If a pattern contained a conditional subpattern with only one branch (in
1262        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1263        computed the wrong minimum data length (which is of course zero for such
1264        subpatterns). This could cause incorrect "no match" results.
1265    
1266    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1267        the pattern is reset in the first branch, pcre_compile() failed with
1268        "internal error: code overflow at offset...". This happened only when
1269        the reset was to the original external option setting. (An optimization
1270        abstracts leading options settings into an external setting, which was the
1271        cause of this.)
1272    
1273    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1274        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1275        assertion pattern did not match (meaning that the assertion was true), it
1276        was incorrectly treated as false if the SKIP had been reached during the
1277        matching. This also applied to assertions used as conditions.
1278    
1279    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1280        assertion subpattern, including such a pattern used as a condition,
1281        unpredictable results occurred, instead of the error return
1282        PCRE_ERROR_DFA_UITEM.
1283    
1284    5.  The C++ GlobalReplace function was not working like Perl for the special
1285        situation when an empty string is matched. It now does the fancy magic
1286        stuff that is necessary.
1287    
1288    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1289        removed. (These were left over from very, very early versions of PCRE.)
1290    
1291    7.  Some cosmetic changes to the code to make life easier when compiling it
1292        as part of something else:
1293    
1294        (a) Change DEBUG to PCRE_DEBUG.
1295    
1296        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1297            called "current" as "current_branch", to prevent a collision with the
1298            Linux macro when compiled as a kernel module.
1299    
1300        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1301            prevent a collision with the Linux macro when compiled as a kernel
1302            module.
1303    
1304    8.  In pcre_compile() there are some checks for integer overflows that used to
1305        cast potentially large values to (double). This has been changed to that
1306        when building, a check for int64_t is made, and if it is found, it is used
1307        instead, thus avoiding the use of floating point arithmetic. (There is no
1308        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1309        double.
1310    
1311    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1312        2005 (difference between two addresses compared to an unsigned value).
1313    
1314    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1315        custom one, because of the following reported problem in Windows:
1316    
1317          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1318              under Win32.
1319          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1320              therefore missing the function definition.
1321          - The compiler thus generates a "C" signature for the test function.
1322          - The linker fails to find the "C" function.
1323          - PCRE fails to configure if asked to do so against libbz2.
1324    
1325    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1326        messages were output:
1327    
1328          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1329          rerunning libtoolize, to keep the correct libtool macros in-tree.
1330          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1331    
1332        I have done both of these things.
1333    
1334    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1335        most of the time, it *can* run out if it is given a pattern that contains a
1336        runaway infinite recursion. I updated the discussion in the pcrestack man
1337        page.
1338    
1339    13. Now that we have gone to the x.xx style of version numbers, the minor
1340        version may start with zero. Using 08 or 09 is a bad idea because users
1341        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1342        interpreted as invalid octal numbers. I've updated the previous comment in
1343        configure.ac, and also added a check that gives an error if 08 or 09 are
1344        used.
1345    
1346    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1347        causing partial matching to fail when the end of the subject matched \W
1348        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1349    
1350    15. There were some discrepancies between the declarations in pcre_internal.h
1351        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1352        their definitions. The declarations used "const uschar *" and the
1353        definitions used USPTR. Even though USPTR is normally defined as "const
1354        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1355        reported that: "This difference in casting confuses some C++ compilers, for
1356        example, SunCC recognizes above declarations as different functions and
1357        generates broken code for hbpcre." I have changed the declarations to use
1358        USPTR.
1359    
1360    16. GNU libtool is named differently on some systems. The autogen.sh script now
1361        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1362        (FreeBSD).
1363    
1364    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1365        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1366        comment: "Figure out how to create a longlong from a string: strtoll and
1367        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1368        instance, but it only takes 2 args instead of 3!"
1369    
1370    18. A subtle bug concerned with back references has been fixed by a change of
1371        specification, with a corresponding code fix. A pattern such as
1372        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1373        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1374        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1375        same bug. Such groups have to be quantified to be useful, or contained
1376        inside another quantified group. (If there's no repetition, the reference
1377        can never match.) The problem arises because, having left the group and
1378        moved on to the rest of the pattern, a later failure that backtracks into
1379        the group uses the captured value from the final iteration of the group
1380        rather than the correct earlier one. I have fixed this in PCRE by forcing
1381        any group that contains a reference to itself to be an atomic group; that
1382        is, there cannot be any backtracking into it once it has completed. This is
1383        similar to recursive and subroutine calls.
1384    
1385    
1386    Version 8.00 19-Oct-09
1387    ----------------------
1388    
1389    1.  The table for translating pcre_compile() error codes into POSIX error codes
1390        was out-of-date, and there was no check on the pcre_compile() error code
1391        being within the table. This could lead to an OK return being given in
1392        error.
1393    
1394    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1395        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1396        in a Windows environment.
1397    
1398    3.  The pcregrep --count option prints the count for each file even when it is
1399        zero, as does GNU grep. However, pcregrep was also printing all files when
1400        --files-with-matches was added. Now, when both options are given, it prints
1401        counts only for those files that have at least one match. (GNU grep just
1402        prints the file name in this circumstance, but including the count seems
1403        more useful - otherwise, why use --count?) Also ensured that the
1404        combination -clh just lists non-zero counts, with no names.
1405    
1406    4.  The long form of the pcregrep -F option was incorrectly implemented as
1407        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1408        but it seems right to fix it, and I didn't think it was worth preserving
1409        the old behaviour.
1410    
1411    5.  The command line items --regex=pattern and --regexp=pattern were not
1412        recognized by pcregrep, which required --regex pattern or --regexp pattern
1413        (with a space rather than an '='). The man page documented the '=' forms,
1414        which are compatible with GNU grep; these now work.
1415    
1416    6.  No libpcreposix.pc file was created for pkg-config; there was just
1417        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1418    
1419    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1420        when UCP support is not needed, by modifying the Python script that
1421        generates it from Unicode data files. This should not matter if the module
1422        is correctly used as a library, but I received one complaint about 50K of
1423        unwanted data. My guess is that the person linked everything into his
1424        program rather than using a library. Anyway, it does no harm.
1425    
1426    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1427        was a minimum greater than 1 for a wide character in a possessive
1428        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1429        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1430        character. Chaos in the form of incorrect output or a compiling loop could
1431        result.
1432    
1433    9.  The restrictions on what a pattern can contain when partial matching is
1434        requested for pcre_exec() have been removed. All patterns can now be
1435        partially matched by this function. In addition, if there are at least two
1436        slots in the offset vector, the offset of the earliest inspected character
1437        for the match and the offset of the end of the subject are set in them when
1438        PCRE_ERROR_PARTIAL is returned.
1439    
1440    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1441        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1442        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1443        and may be more useful for multi-segment matching.
1444    
1445    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1446        used to be given if ever the end of the subject was reached; now it is
1447        given only if matching could not proceed because another character was
1448        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1449        string "Z", which now yields "no match" instead of "partial match". In the
1450        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1451        final character ended with (*FAIL).
1452    
1453    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1454        if the pattern had a "must contain" character that was already found in the
1455        earlier partial match, unless partial matching was again requested. For
1456        example, with the pattern /dog.(body)?/, the "must contain" character is
1457        "g". If the first part-match was for the string "dog", restarting with
1458        "sbody" failed. This bug has been fixed.
1459    
1460    13. The string returned by pcre_dfa_exec() after a partial match has been
1461        changed so that it starts at the first inspected character rather than the
1462        first character of the match. This makes a difference only if the pattern
1463        starts with a lookbehind assertion or \b or \B (\K is not supported by
1464        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1465        matching functions compatible, and I think it's the right thing to do.
1466    
1467    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1468        so that the demonstration program is easily available in environments where
1469        PCRE has not been installed from source.
1470    
1471    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1472        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1473        library.
1474    
1475    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1476        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1477        is not the first non-POSIX option to be added. Clearly some people find
1478        these options useful.
1479    
1480    17. If a caller to the POSIX matching function regexec() passes a non-zero
1481        value for nmatch with a NULL value for pmatch, the value of
1482        nmatch is forced to zero.
1483    
1484    18. RunGrepTest did not have a test for the availability of the -u option of
1485        the diff command, as RunTest does. It now checks in the same way as
1486        RunTest, and also checks for the -b option.
1487    
1488    19. If an odd number of negated classes containing just a single character
1489        interposed, within parentheses, between a forward reference to a named
1490        subpattern and the definition of the subpattern, compilation crashed with
1491        an internal error, complaining that it could not find the referenced
1492        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1493        [The bug was that it was starting one character too far in when skipping
1494        over the character class, thus treating the ] as data rather than
1495        terminating the class. This meant it could skip too much.]
1496    
1497    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1498        /g option in pcretest when the pattern contains \K, which makes it possible
1499        to have an empty string match not at the start, even when the pattern is
1500        anchored. Updated pcretest and pcredemo to use this option.
1501    
1502    21. If the maximum number of capturing subpatterns in a recursion was greater
1503        than the maximum at the outer level, the higher number was returned, but
1504        with unset values at the outer level. The correct (outer level) value is
1505        now given.
1506    
1507    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1508        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1509        make it do so. The string so far is captured, making this feature
1510        compatible with Perl.
1511    
1512    23. The tests have been re-organized, adding tests 11 and 12, to make it
1513        possible to check the Perl 5.10 features against Perl 5.10.
1514    
1515    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1516        pattern matches a fixed length string. PCRE did not allow this; now it
1517        does. Neither allows recursion.
1518    
1519    25. I finally figured out how to implement a request to provide the minimum
1520        length of subject string that was needed in order to match a given pattern.
1521        (It was back references and recursion that I had previously got hung up
1522        on.) This code has now been added to pcre_study(); it finds a lower bound
1523        to the length of subject needed. It is not necessarily the greatest lower
1524        bound, but using it to avoid searching strings that are too short does give
1525        some useful speed-ups. The value is available to calling programs via
1526        pcre_fullinfo().
1527    
1528    26. While implementing 25, I discovered to my embarrassment that pcretest had
1529        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1530        study optimizations had never been tested with that matching function.
1531        Oops. What is worse, even when it was passed study data, there was a bug in
1532        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1533        were also very few tests of studied patterns with pcre_dfa_exec().
1534    
1535    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1536        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1537        on the other side of the coin, they are no longer allowed to have different
1538        names, because these cannot be distinguished in PCRE, and this has caused
1539        confusion. (This is a difference from Perl.)
1540    
1541    28. When duplicate subpattern names are present (necessarily with different
1542        numbers, as required by 27 above), and a test is made by name in a
1543        conditional pattern, either for a subpattern having been matched, or for
1544        recursion in such a pattern, all the associated numbered subpatterns are
1545        tested, and the overall condition is true if the condition is true for any
1546        one of them. This is the way Perl works, and is also more like the way
1547        testing by number works.
1548    
1549    
1550    Version 7.9 11-Apr-09
1551    ---------------------
1552    
1553    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1554        (pcretest), all targets were linked against these libraries. This included
1555        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1556        libraries. This caused unwanted dependencies to be created. This problem
1557        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1558        pcretest is linked with readline.
1559    
1560    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1561        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1562        moved outside it again, because FALSE and TRUE are already defined in AIX,
1563        but BOOL is not.
1564    
1565    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1566        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1567    
1568    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1569        hyphens) following file names and line numbers when outputting matching
1570        lines. This is not true; no spaces are inserted. I have also clarified the
1571        wording for the --colour (or --color) option.
1572    
1573    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1574        was not coloured; this is different to GNU grep, so I have changed it to be
1575        the same.
1576    
1577    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1578        each matching line was coloured. Now it goes on to look for further matches
1579        of any of the test patterns, which is the same behaviour as GNU grep.
1580    
1581    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1582        doesn't make sense to accept an empty string match in pcregrep, so I have
1583        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1584        seems to be how GNU grep behaves.
1585    
1586    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1587        start or after a newline", because the conditional assertion was not being
1588        correctly handled. The rule now is that both the assertion and what follows
1589        in the first alternative must satisfy the test.
1590    
1591    9.  If auto-callout was enabled in a pattern with a conditional group whose
1592        condition was an assertion, PCRE could crash during matching, both with
1593        pcre_exec() and pcre_dfa_exec().
1594    
1595    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1596        used for matching.
1597    
1598    11. Unicode property support in character classes was not working for
1599        characters (bytes) greater than 127 when not in UTF-8 mode.
1600    
1601    12. Added the -M command line option to pcretest.
1602    
1603    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1604    
1605    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1606    
1607    16. Added comments and documentation about mis-use of no_arg in the C++
1608        wrapper.
1609    
1610    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1611        from Martin Jerabek that uses macro names for all relevant character and
1612        string constants.
1613    
1614    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1615        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1616        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1617        these, but not everybody uses configure.
1618    
1619    19. A conditional group that had only one branch was not being correctly
1620        recognized as an item that could match an empty string. This meant that an
1621        enclosing group might also not be so recognized, causing infinite looping
1622        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1623        with the subject "ab", where knowledge that the repeated group can match
1624        nothing is needed in order to break the loop.
1625    
1626    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1627        exec(), but without supplying a callout function, matching went wrong.
1628    
1629    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1630        leak if the size of the offset vector was greater than 30. When the vector
1631        is smaller, the saved offsets during recursion go onto a local stack
1632        vector, but for larger vectors malloc() is used. It was failing to free
1633        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1634        error, in fact).
1635    
1636    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1637        heapframe that is used only when UTF-8 support is enabled. This caused no
1638        problem, but was untidy.
1639    
1640    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1641        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1642        included within another project.
1643    
1644    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1645        slightly modified by me:
1646    
1647          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1648              not building pcregrep.
1649    
1650          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1651              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1652    
1653    25. Forward references, both numeric and by name, in patterns that made use of
1654        duplicate group numbers, could behave incorrectly or give incorrect errors,
1655        because when scanning forward to find the reference group, PCRE was not
1656        taking into account the duplicate group numbers. A pattern such as
1657        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1658    
1659    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1660        the feature of a custom pointer more persuasive (as requested by a user).
1661    
1662    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1663        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1664        pre-defined.
1665    
1666    28. Added support for (*UTF8) at the start of a pattern.
1667    
1668    29. Arrange for flags added by the "release type" setting in CMake to be shown
1669        in the configuration summary.
1670    
1671    
1672    Version 7.8 05-Sep-08
1673    ---------------------
1674    
1675    1.  Replaced UCP searching code with optimized version as implemented for Ad
1676        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1677        stage table and inline lookup instead of a function, giving speed ups of 2
1678        to 5 times on some simple patterns that I tested. Permission was given to
1679        distribute the MultiStage2.py script that generates the tables (it's not in
1680        the tarball, but is in the Subversion repository).
1681    
1682    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1683        scripts.
1684    
1685    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1686        a group with a zero qualifier. The result of the study could be incorrect,
1687        or the function might crash, depending on the pattern.
1688    
1689    4.  Caseless matching was not working for non-ASCII characters in back
1690        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1691        It now works when Unicode Property Support is available.
1692    
1693    5.  In pcretest, an escape such as \x{de} in the data was always generating
1694        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1695        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1696        truncation.
1697    
1698    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1699    
1700    7.  Added two (int) casts to pcregrep when printing the difference of two
1701        pointers, in case they are 64-bit values.
1702    
1703    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1704        test 2 if it fails.
1705    
1706    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1707        and a #define of that name to empty if it is not externally set. This is to
1708        allow users of MSVC to set it if necessary.
1709    
1710    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1711        the convenience functions in the pcre_get.c source file.
1712    
1713    11. An option change at the start of a pattern that had top-level alternatives
1714        could cause overwriting and/or a crash. This command provoked a crash in
1715        some environments:
1716    
1717          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1718    
1719        This potential security problem was recorded as CVE-2008-2371.
1720    
1721    12. For a pattern where the match had to start at the beginning or immediately
1722        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1723        pcre_dfa_exec() could read past the end of the passed subject if there was
1724        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1725        pcretest so that it places the subject at the end of its malloc-ed buffer.
1726    
1727    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1728        exec() might read past the end of the data buffer in UTF-8 mode.
1729    
1730    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1731        the data contained the byte 0x85 as part of a UTF-8 character within its
1732        first line. This applied both to normal and DFA matching.
1733    
1734    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1735        /^[^d]*?$/8 failed to match "abc".
1736    
1737    16. Added a missing copyright notice to pcrecpp_internal.h.
1738    
1739    17. Make it more clear in the documentation that values returned from
1740        pcre_exec() in ovector are byte offsets, not character counts.
1741    
1742    18. Tidied a few places to stop certain compilers from issuing warnings.
1743    
1744    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1745        supplied by Stefan Weber. I made a further small update for 7.8 because
1746        there is a change of source arrangements: the pcre_searchfuncs.c module is
1747        replaced by pcre_ucd.c.
1748    
1749    
1750    Version 7.7 07-May-08
1751    ---------------------
1752    
1753    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1754        a string to a long long, pretend we don't even have a long long." This is
1755        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1756    
1757    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1758        pre-7.6 versions, which defined a global no_arg variable instead of putting
1759        it in the RE class. (See also #8 below.)
1760    
1761    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1762        Lopes.
1763    
1764    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1765    
1766        (1) The include/exclude patterns were being applied to the whole pathnames
1767            of files, instead of just to the final components.
1768    
1769        (2) If there was more than one level of directory, the subdirectories were
1770            skipped unless they satisfied the include/exclude conditions. This is
1771            inconsistent with GNU grep (and could even be seen as contrary to the
1772            pcregrep specification - which I improved to make it absolutely clear).
1773            The action now is always to scan all levels of directory, and just
1774            apply the include/exclude patterns to regular files.
1775    
1776    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1777        --exclude_dir in the tests to avoid scanning .svn directories.
1778    
1779    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1780        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1781        doesn't support NULs in patterns.
1782    
1783    7.  Added some missing "const"s to declarations of static tables in
1784        pcre_compile.c and pcre_dfa_exec.c.
1785    
1786    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1787        caused by fix #2  above. (Subsequently also a second patch to fix the
1788        first patch. And a third patch - this was a messy problem.)
1789    
1790    9.  Applied Craig's patch to remove the use of push_back().
1791    
1792    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1793        matching function regexec().
1794    
1795    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1796        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1797        references. PCRE supports relative numbers with this syntax (I don't think
1798        Oniguruma does).
1799    
1800    12. Previously, a group with a zero repeat such as (...){0} was completely
1801        omitted from the compiled regex. However, this means that if the group
1802        was called as a subroutine from elsewhere in the pattern, things went wrong
1803        (an internal error was given). Such groups are now left in the compiled
1804        pattern, with a new opcode that causes them to be skipped at execution
1805        time.
1806    
1807    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1808        to the way PCRE behaves:
1809    
1810        (a) A lone ] character is dis-allowed (Perl treats it as data).
1811    
1812        (b) A back reference to an unmatched subpattern matches an empty string
1813            (Perl fails the current match path).
1814    
1815        (c) A data ] in a character class must be notated as \] because if the
1816            first data character in a class is ], it defines an empty class. (In
1817            Perl it is not possible to have an empty class.) The empty class []
1818            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1819            The negative empty class [^] matches any one character, independently
1820            of the DOTALL setting.
1821    
1822    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1823        non-existent subpattern following a character class starting with ']' and
1824        containing () gave an internal compiling error instead of "reference to
1825        non-existent subpattern". Fortunately, when the pattern did exist, the
1826        compiled code was correct. (When scanning forwards to check for the
1827        existencd of the subpattern, it was treating the data ']' as terminating
1828        the class, so got the count wrong. When actually compiling, the reference
1829        was subsequently set up correctly.)
1830    
1831    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1832        it was being rejected as not supported by pcre_dfa_exec(), even though
1833        other assertions are supported. I have made pcre_dfa_exec() support
1834        (*FAIL).
1835    
1836    16. The implementation of 13c above involved the invention of a new opcode,
1837        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1838        cannot be changed at match time, I realized I could make a small
1839        improvement to matching performance by compiling OP_ALLANY instead of
1840        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1841        on the OP_ANY path.
1842    
1843    17. Compiling pcretest on Windows with readline support failed without the
1844        following two fixes: (1) Make the unistd.h include conditional on
1845        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1846    
1847    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1848        ncurses library to be included for pcretest when ReadLine support is
1849        requested, but also to allow for it to be overridden. This patch came from
1850        Daniel Bergström.
1851    
1852    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1853        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1854        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1855        spotting this.
1856    
1857    
1858    Version 7.6 28-Jan-08
1859    ---------------------
1860    
1861    1.  A character class containing a very large number of characters with
1862        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1863        overflow.
1864    
1865    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1866        HAVE_LONG_LONG is not defined.
1867    
1868    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1869        bring them up to date and include new features. This patch includes:
1870    
1871        - Fixed PH's badly added libz and libbz2 support.
1872        - Fixed a problem with static linking.
1873        - Added pcredemo. [But later removed - see 7 below.]
1874        - Fixed dftables problem and added an option.
1875        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1876            HAVE_LONG_LONG.
1877        - Added readline support for pcretest.
1878        - Added an listing of the option settings after cmake has run.
1879    
1880    4.  A user submitted a patch to Makefile that makes it easy to create
1881        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1882        Makefile.am that cause it to include this special target, without
1883        affecting anything else. Note that the same mingw target plus all
1884        the other distribution libraries and programs are now supported
1885        when configuring with CMake (see 6 below) instead of with
1886        Configure/Make.
1887    
1888    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1889        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1890        exported in the Windows port". It has not yet been confirmed that the patch
1891        solves the problem, but it does no harm.
1892    
1893    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1894        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1895        with CMake, and also correct the comment about stack recursion.
1896    
1897    7.  Remove the automatic building of pcredemo from the ./configure system and
1898        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1899        of a program that users should build themselves after PCRE is installed, so
1900        building it automatically is not really right. What is more, it gave
1901        trouble in some build environments.
1902    
1903    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1904    
1905    
1906    Version 7.5 10-Jan-08
1907    ---------------------
1908    
1909    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1910        values in parens when parsing an RE using the C++ wrapper."
1911    
1912    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1913        Characters greater than 255 were excluded from the class instead of being
1914        included.
1915    
1916    3.  The same bug as (2) above applied to negated POSIX classes such as
1917        [:^space:].
1918    
1919    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1920        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1921        I have changed it.
1922    
1923    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1924        first named subpattern) and a construct such as (?&a) would reference the
1925        first named subpattern whose name started with "a" (in other words, the
1926        length check was missing). Both these problems are fixed. "Subpattern name
1927        expected" is now given for (?&) (a zero-length name), and this patch also
1928        makes it give the same error for \k'' (previously it complained that that
1929        was a reference to a non-existent subpattern).
1930    
1931    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1932        this is right because (?- can be followed by option settings as well as by
1933        digits. I have, however, made the messages clearer.
1934    
1935    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1936        than the number used in the conditional) now cause a compile-time error.
1937        This is actually not compatible with Perl, which accepts such patterns, but
1938        treats the conditional as always being FALSE (as PCRE used to), but it
1939        seems to me that giving a diagnostic is better.
1940    
1941    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1942        and messages.
1943    
1944    9.  Fix two occurrences of "backslash" in comments that should have been
1945        "backspace".
1946    
1947    10. Remove two redundant lines of code that can never be obeyed (their function
1948        was moved elsewhere).
1949    
1950    11. The program that makes PCRE's Unicode character property table had a bug
1951        which caused it to generate incorrect table entries for sequences of
1952        characters that have the same character type, but are in different scripts.
1953        It amalgamated them into a single range, with the script of the first of
1954        them. In other words, some characters were in the wrong script. There were
1955        thirteen such cases, affecting characters in the following ranges:
1956    
1957          U+002b0 - U+002c1
1958          U+0060c - U+0060d
1959          U+0061e - U+00612
1960          U+0064b - U+0065e
1961          U+0074d - U+0076d
1962          U+01800 - U+01805
1963          U+01d00 - U+01d77
1964          U+01d9b - U+01dbf
1965          U+0200b - U+0200f
1966          U+030fc - U+030fe
1967          U+03260 - U+0327f
1968          U+0fb46 - U+0fbb1
1969          U+10450 - U+1049d
1970    
1971    12. The -o option (show only the matching part of a line) for pcregrep was not
1972        compatible with GNU grep in that, if there was more than one match in a
1973        line, it showed only the first of them. It now behaves in the same way as
1974        GNU grep.
1975    
1976    13. If the -o and -v options were combined for pcregrep, it printed a blank
1977        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1978        does the same. The return code can be used to tell if there were any
1979        non-matching lines.
1980    
1981    14. Added --file-offsets and --line-offsets to pcregrep.
1982    
1983    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1984        infinitely looping recursion. The bug was that positive lookaheads were not
1985        being skipped when checking for a possible empty match (negative lookaheads
1986        and both kinds of lookbehind were skipped).
1987    
1988    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1989        inclusion of <windows.h> to before rather than after the definition of
1990        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1991    
1992    17. Specifying a possessive quantifier with a specific limit for a Unicode
1993        character property caused pcre_compile() to compile bad code, which led at
1994        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1995        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1996        caused the error; without that there was no problem.
1997    
1998    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1999    
2000    19. Added --enable-pcretest-libreadline.
2001    
2002    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2003        RE::GlobalReplace(). As a result, the number of replacements returned was
2004        double what it should be. I removed one of the increments, but Craig sent a
2005        later patch that removed the other one (the right fix) and added unit tests
2006        that check the return values (which was not done before).
2007    
2008    21. Several CMake things:
2009    
2010        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2011            the names libpcre and libpcreposix, not just pcre and pcreposix.
2012    
2013        (2) The above change means that pcretest and pcregrep are now correctly
2014            linked with the newly-built libraries, not previously installed ones.
2015    
2016        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2017    
2018    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2019        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2020        UTF-8 newline character). The key issue is that the pattern starts .*;
2021        this means that the match must be either at the beginning, or after a
2022        newline. The bug was in the code for advancing after a failed match and
2023        checking that the new position followed a newline. It was not taking
2024        account of UTF-8 characters correctly.
2025    
2026    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2027        character classes. PCRE was not treating the sequence [:...:] as a
2028        character class unless the ... were all letters. Perl, however, seems to
2029        allow any characters between [: and :], though of course it rejects as
2030        unknown any "names" that contain non-letters, because all the known class
2031        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2032        for example, whereas PCRE did not - it did not recognize a POSIX character
2033        class. This seemed a bit dangerous, so the code has been changed to be
2034        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2035        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2036        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2037        Perl does, and where it didn't before.
2038    
2039    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2040        Windows environments %n is disabled by default.
2041    
2042    
2043    Version 7.4 21-Sep-07
2044    ---------------------
2045    
2046    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2047        means that a class such as [\s] counted as "explicit reference to CR or
2048        LF". That isn't really right - the whole point of the change was to try to
2049        help when there was an actual mention of one of the two characters. So now
2050        the change happens only if \r or \n (or a literal CR or LF) character is
2051        encountered.
2052    
2053    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2054        of both had grown to the point where there were only 3 bits left.
2055        Fortunately, there was spare space in the data structure, and so I have
2056        moved the internal flags into a new 16-bit field to free up more option
2057        bits.
2058    
2059    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2060        but did not set the internal JCHANGED flag - either of these is enough to
2061        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2062        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2063        start sets both bits.
2064    
2065    4.  Added options (at build time, compile time, exec time) to change \R from
2066        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2067    
2068    5.  doc/pcresyntax.html was missing from the distribution.
2069    
2070    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2071        compatibility, even though it is no longer used.
2072    
2073    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2074        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2075        windows.h file is present (where different names are used). [This was
2076        reversed later after testing - see 16 below.]
2077    
2078    8.  Changed all #include <config.h> to #include "config.h". There were also
2079        some further <pcre.h> cases that I changed to "pcre.h".
2080    
2081    9.  When pcregrep was used with the --colour option, it missed the line ending
2082        sequence off the lines that it output.
2083    
2084    10. It was pointed out to me that arrays of string pointers cause lots of
2085        relocations when a shared library is dynamically loaded. A technique of
2086        using a single long string with a table of offsets can drastically reduce
2087        these. I have refactored PCRE in four places to do this. The result is
2088        dramatic:
2089    
2090          Originally:                          290
2091          After changing UCP table:            187
2092          After changing error message table:   43
2093          After changing table of "verbs"       36
2094          After changing table of Posix names   22
2095    
2096        Thanks to the folks working on Gregex for glib for this insight.
2097    
2098    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2099        unicode-properties was also set.
2100    
2101    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2102    
2103    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2104        checked only for CRLF.
2105    
2106    14. Added casts to pcretest.c to avoid compiler warnings.
2107    
2108    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2109    
2110    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2111        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2112        entirely. This removes changes made in 7 above.
2113    
2114    17. The CMake files have been updated, and there is now more information about
2115        building with CMake in the NON-UNIX-USE document.
2116    
2117    
2118    Version 7.3 28-Aug-07
2119  ---------------------  ---------------------
2120    
2121   1. In the rejigging of the build system that eventually resulted in 7.1, the   1. In the rejigging of the build system that eventually resulted in 7.1, the
# Line 9  Version 7.3 05-Jul-07 Line 2123  Version 7.3 05-Jul-07
2123      brackets there is not right, since it causes compilers to look for an      brackets there is not right, since it causes compilers to look for an
2124      installed pcre.h, not the version that is in the source that is being      installed pcre.h, not the version that is in the source that is being
2125      compiled (which of course may be different). I have changed it back to:      compiled (which of course may be different). I have changed it back to:
2126    
2127        #include "pcre.h"        #include "pcre.h"
2128    
2129      I have a vague recollection that the change was concerned with compiling in      I have a vague recollection that the change was concerned with compiling in
2130      different directories, but in the new build system, that is taken care of      different directories, but in the new build system, that is taken care of
2131      by the VPATH setting the Makefile.      by the VPATH setting the Makefile.
2132    
2133   2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed   2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2134      when the subject happened to end in the byte 0x85 (e.g. if the last      when the subject happened to end in the byte 0x85 (e.g. if the last
2135      character was \x{1ec5}). *Character* 0x85 is one of the "any" newline      character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
# Line 23  Version 7.3 05-Jul-07 Line 2137  Version 7.3 05-Jul-07
2137      of another character. The bug was that, for an unlimited repeat of . in      of another character. The bug was that, for an unlimited repeat of . in
2138      not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by      not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2139      characters when looking for a newline.      characters when looking for a newline.
   
  3. A small performance improvement in the DOTALL UTF-8 mode .* case.  
   
  4. Debugging: adjusted the names of opcodes for different kinds of parentheses  
     in debug output.  
   
  5. Arrange to use "%I64d" instead of "%lld" for long printing in the pcrecpp  
     unittest when running under MinGW.  
2140    
2141     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2142    
2143     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2144        in debug output.
2145    
2146     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2147        long printing in the pcrecpp unittest when running under MinGW.
2148    
2149     6. ESC_K was left out of the EBCDIC table.
2150    
2151     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2152        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2153        limit also applies to "virtual nesting" when a pattern is recursive, and in
2154        this case 1000 isn't so big. I have been able to remove this limit at the
2155        expense of backing off one optimization in certain circumstances. Normally,
2156        when pcre_exec() would call its internal match() function recursively and
2157        immediately return the result unconditionally, it uses a "tail recursion"
2158        feature to save stack. However, when a subpattern that can match an empty
2159        string has an unlimited repetition quantifier, it no longer makes this
2160        optimization. That gives it a stack frame in which to save the data for
2161        checking that an empty string has been matched. Previously this was taken
2162        from the 1000-entry workspace that had been reserved. So now there is no
2163        explicit limit, but more stack is used.
2164    
2165     8. Applied Daniel's patches to solve problems with the import/export magic
2166        syntax that is required for Windows, and which was going wrong for the
2167        pcreposix and pcrecpp parts of the library. These were overlooked when this
2168        problem was solved for the main library.
2169    
2170     9. There were some crude static tests to avoid integer overflow when computing
2171        the size of patterns that contain repeated groups with explicit upper
2172        limits. As the maximum quantifier is 65535, the maximum group length was
2173        set at 30,000 so that the product of these two numbers did not overflow a
2174        32-bit integer. However, it turns out that people want to use groups that
2175        are longer than 30,000 bytes (though not repeat them that many times).
2176        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2177        made it possible to implement the integer overflow checks in a much more
2178        dynamic way, which I have now done. The artificial limitation on group
2179        length has been removed - we now have only the limit on the total length of
2180        the compiled pattern, which depends on the LINK_SIZE setting.
2181    
2182    10. Fixed a bug in the documentation for get/copy named substring when
2183        duplicate names are permitted. If none of the named substrings are set, the
2184        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2185        empty string.
2186    
2187    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2188        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2189        because the ] is interpreted as the first data character and the
2190        terminating ] is not found. PCRE has been made compatible with Perl in this
2191        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2192        cause memory overwriting.
2193    
2194    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2195        string has been matched (to stop an infinite loop). It was not recognizing
2196        a conditional subpattern that could match an empty string if that
2197        subpattern was within another subpattern. For example, it looped when
2198        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2199        condition was not nested. This bug has been fixed.
2200    
2201    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2202        past the start of the subject in the presence of bytes with the top bit
2203        set, for example "\x8aBCD".
2204    
2205    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2206        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2207    
2208    14. Optimized (?!) to (*FAIL).
2209    
2210    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2211        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2212        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2213        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2214        does: it's just the validity check that is more restrictive.
2215    
2216    16. Inserted checks for integer overflows during escape sequence (backslash)
2217        processing, and also fixed erroneous offset values for syntax errors during
2218        backslash processing.
2219    
2220    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2221        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2222    
2223    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2224        caused an overrun.
2225    
2226    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2227        something other than just ASCII characters) inside a group that had an
2228        unlimited repeat caused a loop at compile time (while checking to see
2229        whether the group could match an empty string).
2230    
2231    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2232        [\P{Any}] did so. (Error in the code for printing property names.)
2233    
2234    21. An orphan \E inside a character class could cause a crash.
2235    
2236    22. A repeated capturing bracket such as (A)? could cause a wild memory
2237        reference during compilation.
2238    
2239    23. There are several functions in pcre_compile() that scan along a compiled
2240        expression for various reasons (e.g. to see if it's fixed length for look
2241        behind). There were bugs in these functions when a repeated \p or \P was
2242        present in the pattern. These operators have additional parameters compared
2243        with \d, etc, and these were not being taken into account when moving along
2244        the compiled data. Specifically:
2245    
2246        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2247            length.
2248    
2249        (b) An item such as \pL+ within a repeated group could cause crashes or
2250            loops.
2251    
2252        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2253            "reference to non-existent subpattern" error.
2254    
2255        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2256    
2257    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2258        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2259    
2260    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2261    
2262    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2263        character were causing crashes (broken optimization).
2264    
2265    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2266        \p or \P) caused a compile-time loop.
2267    
2268    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2269        break. For example, the unstudied pattern [\r\n]A does not match the string
2270        "\r\nA" because change 7.0/46 below moves the current point on by two
2271        characters after failing to match at the start. However, the pattern \nA
2272        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2273        the same is true. There doesn't seem any very clean way out of this, but
2274        what I have chosen to do makes the common cases work: PCRE now takes note
2275        of whether there can be an explicit match for \r or \n anywhere in the
2276        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2277        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2278        pattern has explicit CR or LF references.
2279    
2280    29. Added (*CR) etc for changing newline setting at start of pattern.
2281    
2282    
2283  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07

Legend:
Removed from v.193  
changed lines
  Added in v.1254

  ViewVC Help
Powered by ViewVC 1.1.5