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

Diff of /code/trunk/ChangeLog

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

revision 113 by ph10, Fri Mar 9 10:13:26 2007 UTC revision 1253 by ph10, Fri Feb 22 11:38:35 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.1 05-Mar-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    
75    Version 8.32 30-November-2012
76    -----------------------------
77    
78    1.  Improved JIT compiler optimizations for first character search and single
79        character iterators.
80    
81    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
82        Patch by Daniel Richard G.
83    
84    3.  Single character iterator optimizations in the JIT compiler.
85    
86    4.  Improved JIT compiler optimizations for character ranges.
87    
88    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
89        Reported by Giuseppe D'Angelo.
90    
91    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
92        of a line, was being set incorrectly in cases where .* appeared inside
93        atomic brackets at the start of a pattern, or where there was a subsequent
94        *PRUNE or *SKIP.
95    
96    7.  Improved instruction cache flush for POWER/PowerPC.
97        Patch by Daniel Richard G.
98    
99    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
100        grep:
101    
102        (a) There is now no limit to the number of patterns to be matched.
103    
104        (b) An error is given if a pattern is too long.
105    
106        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
107            are now supported.
108    
109        (d) --exclude-from and --include-from (multiple use) have been added.
110    
111        (e) Exclusions and inclusions now apply to all files and directories, not
112            just to those obtained from scanning a directory recursively.
113    
114        (f) Multiple uses of -f and --file-list are now supported.
115    
116        (g) In a Windows environment, the default for -d has been changed from
117            "read" (the GNU grep default) to "skip", because otherwise the presence
118            of a directory in the file list provokes an error.
119    
120        (h) The documentation has been revised and clarified in places.
121    
122    9.  Improve the matching speed of capturing brackets.
123    
124    10. Changed the meaning of \X so that it now matches a Unicode extended
125        grapheme cluster.
126    
127    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
128        out POSIX threads when JIT support is configured.
129    
130    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
131    
132    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
133        structure could go wrong in environments where size_t is not the same size
134        as int.
135    
136    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
137        set.
138    
139    15. The EBCDIC support had decayed; later updates to the code had included
140        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
141        general tidy up of EBCDIC-related issues, and the documentation was also
142        not quite right. There is now a test that can be run on ASCII systems to
143        check some of the EBCDIC-related things (but is it not a full test).
144    
145    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
146        in a small tidy to the code.
147    
148    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
149    
150    18. If the --only-matching (-o) option in pcregrep is specified multiple
151        times, each one causes appropriate output. For example, -o1 -o2 outputs the
152        substrings matched by the 1st and 2nd capturing parentheses. A separating
153        string can be specified by --om-separator (default empty).
154    
155    19. Improving the first n character searches.
156    
157    20. Turn case lists for horizontal and vertical white space into macros so that
158        they are defined only once.
159    
160    21. This set of changes together give more compatible Unicode case-folding
161        behaviour for characters that have more than one other case when UCP
162        support is available.
163    
164        (a) The Unicode property table now has offsets into a new table of sets of
165            three or more characters that are case-equivalent. The MultiStage2.py
166            script that generates these tables (the pcre_ucd.c file) now scans
167            CaseFolding.txt instead of UnicodeData.txt for character case
168            information.
169    
170        (b) The code for adding characters or ranges of characters to a character
171            class has been abstracted into a generalized function that also handles
172            case-independence. In UTF-mode with UCP support, this uses the new data
173            to handle characters with more than one other case.
174    
175        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
176            whose other case is greater than 256 are now correctly matched
177            caselessly. Previously, the high codepoint matched the low one, but not
178            vice versa.
179    
180        (d) The processing of \h, \H, \v, and \ in character classes now makes use
181            of the new class addition function, using character lists defined as
182            macros alongside the case definitions of 20 above.
183    
184        (e) Caseless back references now work with characters that have more than
185            one other case.
186    
187        (f) General caseless matching of characters with more than one other case
188            is supported.
189    
190    22. Unicode character properties were updated from Unicode 6.2.0
191    
192    23. Improved CMake support under Windows. Patch by Daniel Richard G.
193    
194    24. Add support for 32-bit character strings, and UTF-32
195    
196    25. Major JIT compiler update (code refactoring and bugfixing).
197        Experimental Sparc 32 support is added.
198    
199    26. Applied a modified version of Daniel Richard G's patch to create
200        pcre.h.generic and config.h.generic by "make" instead of in the
201        PrepareRelease script.
202    
203    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
204        pcre_compile.c when checking for a zero character.
205    
206    28. Introducing a native interface for JIT. Through this interface, the compiled
207        machine code can be directly executed. The purpose of this interface is to
208        provide fast pattern matching, so several sanity checks are not performed.
209        However, feature tests are still performed. The new interface provides
210        1.4x speedup compared to the old one.
211    
212    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
213        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
214        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
215    
216    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
217        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
218        the "old" RFC 2279). Instead, it ended up passing a negative length to
219        pcre_exec().
220    
221    31. Add support for GCC's visibility feature to hide internal functions.
222    
223    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
224        "unknown -C option" after outputting 0 or 1.
225    
226    33. There is now support for generating a code coverage report for the test
227        suite in environments where gcc is the compiler and lcov is installed. This
228        is mainly for the benefit of the developers.
229    
230    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
231        unaddressable using valgrind annotations, allowing valgrind to detect
232        invalid memory accesses. This is mainly for the benefit of the developers.
233    
234    25. (*UTF) can now be used to start a pattern in any of the three libraries.
235    
236    26. Give configure error if --enable-cpp but no C++ compiler found.
237    
238    
239    Version 8.31 06-July-2012
240    -------------------------
241    
242    1.  Fixing a wrong JIT test case and some compiler warnings.
243    
244    2.  Removed a bashism from the RunTest script.
245    
246    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
247        to unsigned type, result still unsigned" that was given by an MS compiler
248        on encountering the code "-sizeof(xxx)".
249    
250    4.  Partial matching support is added to the JIT compiler.
251    
252    5.  Fixed several bugs concerned with partial matching of items that consist
253        of more than one character:
254    
255        (a) /^(..)\1/ did not partially match "aba" because checking references was
256            done on an "all or nothing" basis. This also applied to repeated
257            references.
258    
259        (b) \R did not give a hard partial match if \r was found at the end of the
260            subject.
261    
262        (c) \X did not give a hard partial match after matching one or more
263            characters at the end of the subject.
264    
265        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
266            a partial match for the string "\r".
267    
268        (e) When newline was set to CRLF, the metacharacter "." did not recognize
269            a partial match for a CR character at the end of the subject string.
270    
271    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
272        running pcretest, the text "(JIT)" added to the output whenever JIT is
273        actually used to run the match.
274    
275    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
276        or /S+[+] with a digit between 1 and 7.
277    
278    8.  OP_NOT now supports any UTF character not just single-byte ones.
279    
280    9.  (*MARK) control verb is now supported by the JIT compiler.
281    
282    10. The command "./RunTest list" lists the available tests without actually
283        running any of them. (Because I keep forgetting what they all are.)
284    
285    11. Add PCRE_INFO_MAXLOOKBEHIND.
286    
287    12. Applied a (slightly modified) user-supplied patch that improves performance
288        when the heap is used for recursion (compiled with --disable-stack-for-
289        recursion). Instead of malloc and free for each heap frame each time a
290        logical recursion happens, frames are retained on a chain and re-used where
291        possible. This sometimes gives as much as 30% improvement.
292    
293    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
294        call.
295    
296    14. As documented, (*COMMIT) is now confined to within a positive assertion.
297    
298    15. It is now possible to link pcretest with libedit as an alternative to
299        libreadline.
300    
301    16. (*COMMIT) control verb is now supported by the JIT compiler.
302    
303    17. The Unicode data tables have been updated to Unicode 6.1.0.
304    
305    18. Added --file-list option to pcregrep.
306    
307    19. Added binary file support to pcregrep, including the -a, --binary-files,
308        -I, and --text options.
309    
310    20. The madvise function is renamed for posix_madvise for QNX compatibility
311        reasons. Fixed by Giuseppe D'Angelo.
312    
313    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
314    
315    22. Changed the option for creating symbolic links for 16-bit man pages from
316        -s to -sf so that re-installing does not cause issues.
317    
318    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
319    
320    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
321        matches in certain environments (the workspace was not being correctly
322        retained). Also added to pcre_dfa_exec() a simple plausibility check on
323        some of the workspace data at the beginning of a restart.
324    
325    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
326        was not doing so when it should - probably a typo introduced by SVN 528
327        (change 8.10/14).
328    
329    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
330        \w+ when the character tables indicated that \x{c4} was a word character.
331        There were several related cases, all because the tests for doing a table
332        lookup were testing for characters less than 127 instead of 255.
333    
334    27. If a pattern contains capturing parentheses that are not used in a match,
335        their slots in the ovector are set to -1. For those that are higher than
336        any matched groups, this happens at the end of processing. In the case when
337        there were back references that the ovector was too small to contain
338        (causing temporary malloc'd memory to be used during matching), and the
339        highest capturing number was not used, memory off the end of the ovector
340        was incorrectly being set to -1. (It was using the size of the temporary
341        memory instead of the true size.)
342    
343    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
344        ovector size, it uses memory at the end of the block that it has got.
345    
346    29. Check for an overlong MARK name and give an error at compile time. The
347        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
348    
349    30. JIT compiler update.
350    
351    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
352        Rill for the patch.
353    
354    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
355    
356    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
357    
358    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
359        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
360        was enabled.
361    
362    35. Improve JIT code generation for greedy plus quantifier.
363    
364    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
365        1 to "aa" instead of to an empty string. The bug affected repeated groups
366        that could potentially match an empty string.
367    
368    37. Optimizing single character iterators in JIT.
369    
370    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
371        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
372        codepoints that are too big for the mode are faulted, and in a UTF mode,
373        disallowed codepoints are also faulted.
374    
375    39. If PCRE was compiled with UTF support, in three places in the DFA
376        matcher there was code that should only have been obeyed in UTF mode, but
377        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
378        processing when bytes with values greater than 127 were present. In 16-bit
379        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
380        both cases the values are those that cannot be the first data item in a UTF
381        character. The three items that might have provoked this were recursions,
382        possessively repeated groups, and atomic groups.
383    
384    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
385        and pcregrep, because some OS require shared objects to be explicitly
386        passed to ld, causing the link step to fail if they are not.
387    
388    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
389        mode, patterns that started with \h* or \R* might be incorrectly matched.
390    
391    
392    Version 8.30 04-February-2012
393    -----------------------------
394    
395    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
396        name is defined in ctype.h.
397    
398    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
399        only in quite long subpatterns.
400    
401    3.  Removed the function pcre_info(), which has been obsolete and deprecated
402        since it was replaced by pcre_fullinfo() in February 2000.
403    
404    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
405        match a (*MARK), and the match failed at the start of the subject, a
406        reference to memory before the start of the subject could occur. This bug
407        was introduced by fix 17 of release 8.21.
408    
409    5.  A reference to an unset group with zero minimum repetition was giving
410        totally wrong answers (in non-JavaScript-compatibility mode). For example,
411        /(another)?(\1?)test/ matched against "hello world test". This bug was
412        introduced in release 8.13.
413    
414    6.  Add support for 16-bit character strings (a large amount of work involving
415        many changes and refactorings).
416    
417    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
418        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
419        from a file.
420    
421    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
422        rounding is not applied in this particular case).
423    
424    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
425        if they appear, or are escaped, in patterns.
426    
427    10. Get rid of a number of -Wunused-but-set-variable warnings.
428    
429    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
430        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
431        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
432        also returns the mark "x". This bug applied to capturing parentheses,
433        non-capturing parentheses, and atomic parentheses. It also applied to some
434        assertions.
435    
436    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
437        information out of configure.ac instead of relying on pcre.h.generic, which
438        is not stored in the repository.
439    
440    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
441        -lreadline.
442    
443    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
444    
445    15. Applied Graycode's patch to put the top-level frame on the stack rather
446        than the heap when not using the stack for recursion. This gives a
447        performance improvement in many cases when recursion is not deep.
448    
449    16. Experimental code added to "pcretest -C" to output the stack frame size.
450    
451    
452    Version 8.21 12-Dec-2011
453    ------------------------
454    
455    1.  Updating the JIT compiler.
456    
457    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
458        are added as well.
459    
460    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
461        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
462        calling _pcre_jit_exec. Some extra comments are added.
463    
464    4.  (*MARK) settings inside atomic groups that do not contain any capturing
465        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
466        was introduced by change 18 for 8.20.
467    
468    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
469        ECMA-262 standard.
470    
471    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
472        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
473        This bug was probably introduced by change 9 of 8.13.
474    
475    7.  While fixing 6 above, I noticed that a number of other items were being
476        incorrectly rejected as "not fixed length". This arose partly because newer
477        opcodes had not been added to the fixed-length checking code. I have (a)
478        corrected the bug and added tests for these items, and (b) arranged for an
479        error to occur if an unknown opcode is encountered while checking for fixed
480        length instead of just assuming "not fixed length". The items that were
481        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
482        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
483        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
484    
485    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
486        being incorrectly compiled and would have given unpredicatble results.
487    
488    9.  A possessively repeated subpattern with minimum repeat count greater than
489        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
490        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
491        the first (A) could occur when it should not.
492    
493    10. Add a cast and remove a redundant test from the code.
494    
495    11. JIT should use pcre_malloc/pcre_free for allocation.
496    
497    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
498        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
499        is anything other than /usr, -L is still shown).
500    
501    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
502    
503    14. Perl does not support \N without a following name in a [] class; PCRE now
504        also gives an error.
505    
506    15. If a forward reference was repeated with an upper limit of around 2000,
507        it caused the error "internal error: overran compiling workspace". The
508        maximum number of forward references (including repeats) was limited by the
509        internal workspace, and dependent on the LINK_SIZE. The code has been
510        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
511        the default depends on LINK_SIZE. There is a new upper limit (for safety)
512        of around 200,000 forward references. While doing this, I also speeded up
513        the filling in of repeated forward references.
514    
515    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
516        incorrectly expecting the subject to contain another "a" after the start.
517    
518    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
519        in the match, the SKIP should be ignored. This was not happening; instead
520        the SKIP was being treated as NOMATCH. For patterns such as
521        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
522        tested.
523    
524    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
525        now much more compatible with Perl, in particular in cases where the result
526        is a non-match for a non-anchored pattern. For example, if
527        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
528        "m", where previously it did not return a name. A side effect of this
529        change is that for partial matches, the last encountered mark name is
530        returned, as for non matches. A number of tests that were previously not
531        Perl-compatible have been moved into the Perl-compatible test files. The
532        refactoring has had the pleasing side effect of removing one argument from
533        the match() function, thus reducing its stack requirements.
534    
535    19. If the /S+ option was used in pcretest to study a pattern using JIT,
536        subsequent uses of /S (without +) incorrectly behaved like /S+.
537    
538    21. Retrieve executable code size support for the JIT compiler and fixing
539        some warnings.
540    
541    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
542        not work when the shorter character appeared right at the end of the
543        subject string.
544    
545    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
546        systems.
547    
548    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
549        output it when the /M option is used in pcretest.
550    
551    25. The CheckMan script was not being included in the distribution. Also, added
552        an explicit "perl" to run Perl scripts from the PrepareRelease script
553        because this is reportedly needed in Windows.
554    
555    26. If study data was being save in a file and studying had not found a set of
556        "starts with" bytes for the pattern, the data written to the file (though
557        never used) was taken from uninitialized memory and so caused valgrind to
558        complain.
559    
560    27. Updated RunTest.bat as provided by Sheri Pierce.
561    
562    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
563    
564    29. Computation of memory usage for the table of capturing group names was
565        giving an unnecessarily large value.
566    
567    
568    Version 8.20 21-Oct-2011
569    ------------------------
570    
571    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
572        a POSIX class. After further experiments with Perl, which convinced me that
573        Perl has bugs and confusions, a closing square bracket is no longer allowed
574        in a POSIX name. This bug also affected patterns with classes that started
575        with full stops.
576    
577    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
578        captured substring, but while checking the failing first alternative,
579        substring 1 is temporarily captured. If the output vector supplied to
580        pcre_exec() was not big enough for this capture, the yield of the function
581        was still zero ("insufficient space for captured substrings"). This cannot
582        be totally fixed without adding another stack variable, which seems a lot
583        of expense for a edge case. However, I have improved the situation in cases
584        such as /(a)(b)x|abc/ matched against "abc", where the return code
585        indicates that fewer than the maximum number of slots in the ovector have
586        been set.
587    
588    3.  Related to (2) above: when there are more back references in a pattern than
589        slots in the output vector, pcre_exec() uses temporary memory during
590        matching, and copies in the captures as far as possible afterwards. It was
591        using the entire output vector, but this conflicts with the specification
592        that only 2/3 is used for passing back captured substrings. Now it uses
593        only the first 2/3, for compatibility. This is, of course, another edge
594        case.
595    
596    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
597        main code base, and can be used by building with --enable-jit. When this is
598        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
599        runtime --no-jit option is given.
600    
601    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
602        ovector, the return from the function was zero, implying that there were
603        other matches that did not fit. The correct "exactly full" value is now
604        returned.
605    
606    6.  If a subpattern that was called recursively or as a subroutine contained
607        (*PRUNE) or any other control that caused it to give a non-standard return,
608        invalid errors such as "Error -26 (nested recursion at the same subject
609        position)" or even infinite loops could occur.
610    
611    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
612        computing the minimum length on reaching *ACCEPT, and so ended up with the
613        wrong value of 1 rather than 0. Further investigation indicates that
614        computing a minimum subject length in the presence of *ACCEPT is difficult
615        (think back references, subroutine calls), and so I have changed the code
616        so that no minimum is registered for a pattern that contains *ACCEPT.
617    
618    8.  If (*THEN) was present in the first (true) branch of a conditional group,
619        it was not handled as intended. [But see 16 below.]
620    
621    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
622        Sheri Pierce.
623    
624    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
625        the first byte in a match must be "a".
626    
627    11. Change 17 for 8.13 increased the recursion depth for patterns like
628        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
629        pattern contains any instances of (*THEN). If it does not, the old
630        optimizations are restored. It would be nice to do this on a per-group
631        basis, but at the moment that is not feasible.
632    
633    12. In some environments, the output of pcretest -C is CRLF terminated. This
634        broke RunTest's code that checks for the link size. A single white space
635        character after the value is now allowed for.
636    
637    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
638        For "fr", it uses the Windows-specific input and output files.
639    
640    14. If (*THEN) appeared in a group that was called recursively or as a
641        subroutine, it did not work as intended. [But see next item.]
642    
643    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
644        pattern fragments (but not containing any | characters). If A and B are
645        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
646        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
647        D. In other words, Perl considers parentheses that do not contain any |
648        characters to be part of a surrounding alternative, whereas PCRE was
649        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
650        differently. PCRE now behaves in the same way as Perl, except in the case
651        of subroutine/recursion calls such as (?1) which have in any case always
652        been different (but PCRE had them first :-).
653    
654    16. Related to 15 above: Perl does not treat the | in a conditional group as
655        creating alternatives. Such a group is treated in the same way as an
656        ordinary group without any | characters when processing (*THEN). PCRE has
657        been changed to match Perl's behaviour.
658    
659    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
660        RunGrepTest script failed.
661    
662    18. Change 22 for version 13 caused atomic groups to use more stack. This is
663        inevitable for groups that contain captures, but it can lead to a lot of
664        stack use in large patterns. The old behaviour has been restored for atomic
665        groups that do not contain any capturing parentheses.
666    
667    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
668        suppress the check for a minimum subject length at run time. (If it was
669        given to pcre_exec() or pcre_dfa_exec() it did work.)
670    
671    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
672        fail to work when decoding hex characters in data strings in EBCDIC
673        environments.
674    
675    21. It appears that in at least one Mac OS environment, the isxdigit() function
676        is implemented as a macro that evaluates to its argument more than once,
677        contravening the C 90 Standard (I haven't checked a later standard). There
678        was an instance in pcretest which caused it to go wrong when processing
679        \x{...} escapes in subject strings. The has been rewritten to avoid using
680        things like p++ in the argument of isxdigit().
681    
682    
683    Version 8.13 16-Aug-2011
684    ------------------------
685    
686    1.  The Unicode data tables have been updated to Unicode 6.0.0.
687    
688    2.  Two minor typos in pcre_internal.h have been fixed.
689    
690    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
691        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
692        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
693    
694    4.  There were a number of related bugs in the code for matching backrefences
695        caselessly in UTF-8 mode when codes for the characters concerned were
696        different numbers of bytes. For example, U+023A and U+2C65 are an upper
697        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
698        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
699        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
700        2-byte code at the end of the subject (it thought there wasn't enough data
701        left).
702    
703    5.  Comprehensive information about what went wrong is now returned by
704        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
705        as the output vector has at least 2 elements. The offset of the start of
706        the failing character and a reason code are placed in the vector.
707    
708    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
709        now returned is for the first byte of the failing character, instead of the
710        last byte inspected. This is an incompatible change, but I hope it is small
711        enough not to be a problem. It makes the returned offset consistent with
712        pcre_exec() and pcre_dfa_exec().
713    
714    7.  pcretest now gives a text phrase as well as the error number when
715        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
716        failure, the offset and reason code are output.
717    
718    8.  When \R was used with a maximizing quantifier it failed to skip backwards
719        over a \r\n pair if the subsequent match failed. Instead, it just skipped
720        back over a single character (\n). This seems wrong (because it treated the
721        two characters as a single entity when going forwards), conflicts with the
722        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
723        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
724        has been changed.
725    
726    9.  Some internal refactoring has changed the processing so that the handling
727        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
728        time (the PCRE_DOTALL option was changed this way some time ago: version
729        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
730        which was always a bit of a fudge. It also means that there is one less
731        argument for the match() function, which reduces its stack requirements
732        slightly. This change also fixes an incompatibility with Perl: the pattern
733        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
734    
735    10. More internal refactoring has drastically reduced the number of recursive
736        calls to match() for possessively repeated groups such as (abc)++ when
737        using pcre_exec().
738    
739    11. While implementing 10, a number of bugs in the handling of groups were
740        discovered and fixed:
741    
742        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
743        (a|)*(?1) gave a compile-time internal error.
744        ((a|)+)+  did not notice that the outer group could match an empty string.
745        (^a|^)+   was not marked as anchored.
746        (.*a|.*)+ was not marked as matching at start or after a newline.
747    
748    12. Yet more internal refactoring has removed another argument from the match()
749        function. Special calls to this function are now indicated by setting a
750        value in a variable in the "match data" data block.
751    
752    13. Be more explicit in pcre_study() instead of relying on "default" for
753        opcodes that mean there is no starting character; this means that when new
754        ones are added and accidentally left out of pcre_study(), testing should
755        pick them up.
756    
757    14. The -s option of pcretest has been documented for ages as being an old
758        synonym of -m (show memory usage). I have changed it to mean "force study
759        for every regex", that is, assume /S for every regex. This is similar to -i
760        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
761        using it. It makes it easier to run collections of tests with and without
762        study enabled, and thereby test pcre_study() more easily. All the standard
763        tests are now run with and without -s (but some patterns can be marked as
764        "never study" - see 20 below).
765    
766    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
767        restoration of the capturing data to the outer values was not happening
768        correctly.
769    
770    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
771        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
772        pattern had matched an empty string, and so incorrectly returned a no
773        match.
774    
775    17. There was optimizing code for the last branch of non-capturing parentheses,
776        and also for the obeyed branch of a conditional subexpression, which used
777        tail recursion to cut down on stack usage. Unfortunately, now that there is
778        the possibility of (*THEN) occurring in these branches, tail recursion is
779        no longer possible because the return has to be checked for (*THEN). These
780        two optimizations have therefore been removed. [But see 8.20/11 above.]
781    
782    18. If a pattern containing \R was studied, it was assumed that \R always
783        matched two bytes, thus causing the minimum subject length to be
784        incorrectly computed because \R can also match just one byte.
785    
786    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
787        was incorrectly computed.
788    
789    20. If /S is present twice on a test pattern in pcretest input, it now
790        *disables* studying, thereby overriding the use of -s on the command line
791        (see 14 above). This is necessary for one or two tests to keep the output
792        identical in both cases.
793    
794    21. When (*ACCEPT) was used in an assertion that matched an empty string and
795        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
796    
797    22. When an atomic group that contained a capturing parenthesis was
798        successfully matched, but the branch in which it appeared failed, the
799        capturing was not being forgotten if a higher numbered group was later
800        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
801        group 1 to "a", when in fact it should be unset. This applied to multi-
802        branched capturing and non-capturing groups, repeated or not, and also to
803        positive assertions (capturing in negative assertions does not happen
804        in PCRE) and also to nested atomic groups.
805    
806    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
807        subject after a captured substring, to make it easier to tell which of a
808        number of identical substrings has been captured.
809    
810    24. The way atomic groups are processed by pcre_exec() has been changed so that
811        if they are repeated, backtracking one repetition now resets captured
812        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
813        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
814        "aaa". Previously, it would have been "a". As part of this code
815        refactoring, the way recursive calls are handled has also been changed.
816    
817    25. If an assertion condition captured any substrings, they were not passed
818        back unless some other capturing happened later. For example, if
819        (?(?=(a))a) was matched against "a", no capturing was returned.
820    
821    26. When studying a pattern that contained subroutine calls or assertions,
822        the code for finding the minimum length of a possible match was handling
823        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
824        group 1 called group 2 while simultaneously a separate group 2 called group
825        1). A stack overflow occurred in this case. I have fixed this by limiting
826        the recursion depth to 10.
827    
828    27. Updated RunTest.bat in the distribution to the version supplied by Tom
829        Fortmann. This supports explicit test numbers on the command line, and has
830        argument validation and error reporting.
831    
832    28. An instance of \X with an unlimited repeat could fail if at any point the
833        first character it looked at was a mark character.
834    
835    29. Some minor code refactoring concerning Unicode properties and scripts
836        should reduce the stack requirement of match() slightly.
837    
838    30. Added the '=' option to pcretest to check the setting of unused capturing
839        slots at the end of the pattern, which are documented as being -1, but are
840        not included in the return count.
841    
842    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
843        compiled something random. Now it gives a compile-time error (as does
844        Perl).
845    
846    32. A *MARK encountered during the processing of a positive assertion is now
847        recorded and passed back (compatible with Perl).
848    
849    33. If --only-matching or --colour was set on a pcregrep call whose pattern
850        had alternative anchored branches, the search for a second match in a line
851        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
852        matched the line "0102" twice. The same bug affected patterns that started
853        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
854        twice.
855    
856    34. Previously, PCRE did not allow quantification of assertions. However, Perl
857        does, and because of capturing effects, quantifying parenthesized
858        assertions may at times be useful. Quantifiers are now allowed for
859        parenthesized assertions.
860    
861    35. A minor code tidy in pcre_compile() when checking options for \R usage.
862    
863    36. \g was being checked for fancy things in a character class, when it should
864        just be a literal "g".
865    
866    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
867        appearance of a nested POSIX class supersedes an apparent external class.
868        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
869        unescaped square brackets may also appear as part of class names. For
870        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
871        more like Perl. (But see 8.20/1 above.)
872    
873    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
874        was because it thought it was \N{name}, which is not supported).
875    
876    39. Add minix to OS list not supporting the -S option in pcretest.
877    
878    40. PCRE tries to detect cases of infinite recursion at compile time, but it
879        cannot analyze patterns in sufficient detail to catch mutual recursions
880        such as ((?1))((?2)). There is now a runtime test that gives an error if a
881        subgroup is called recursively as a subpattern for a second time at the
882        same position in the subject string. In previous releases this might have
883        been caught by the recursion limit, or it might have run out of stack.
884    
885    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
886        happen only once. PCRE was, however incorrectly giving a compile time error
887        "recursive call could loop indefinitely" because it cannot analyze the
888        pattern in sufficient detail. The compile time test no longer happens when
889        PCRE is compiling a conditional subpattern, but actual runaway loops are
890        now caught at runtime (see 40 above).
891    
892    42. It seems that Perl allows any characters other than a closing parenthesis
893        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
894        has been changed to be the same.
895    
896    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
897        as not to get warnings when autogen.sh is called. Also changed
898        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
899    
900    44. To help people who use pcregrep to scan files containing exceedingly long
901        lines, the following changes have been made:
902    
903        (a) The default value of the buffer size parameter has been increased from
904            8K to 20K. (The actual buffer used is three times this size.)
905    
906        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
907            PCRE is built.
908    
909        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
910            to be set at run time.
911    
912        (d) Numerical values in pcregrep options can be followed by K or M, for
913            example --buffer-size=50K.
914    
915        (e) If a line being scanned overflows pcregrep's buffer, an error is now
916            given and the return code is set to 2.
917    
918    45. Add a pointer to the latest mark to the callout data block.
919    
920    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
921        partial match of an empty string instead of no match. This was specific to
922        the use of ".".
923    
924    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
925        complete match instead of a partial match. This bug was dependent on both
926        the PCRE_UTF8 and PCRE_DOTALL options being set.
927    
928    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
929        starting byte set, because \b was not being ignored.
930    
931    
932    Version 8.12 15-Jan-2011
933    ------------------------
934    
935    1.  Fixed some typos in the markup of the man pages, and wrote a script that
936        checks for such things as part of the documentation building process.
937    
938    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
939        --match-limit and --recursion-limit options (added for 8.11). In
940        particular, this made one of the standard tests fail. (The integer value
941        went into the wrong half of a long int.)
942    
943    3.  If the --colour option was given to pcregrep with -v (invert match), it
944        did strange things, either producing crazy output, or crashing. It should,
945        of course, ignore a request for colour when reporting lines that do not
946        match.
947    
948    4.  Another pcregrep bug caused similar problems if --colour was specified with
949        -M (multiline) and the pattern match finished with a line ending.
950    
951    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
952        matched in multiline mode, the following line was shown as part of the
953        match. This seems wrong, so I have changed it.
954    
955    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
956        the check for further matches in the same line (so they could be coloured)
957        to overrun the end of the current line. If another match was found, it was
958        incorrectly shown (and then shown again when found in the next line).
959    
960    7.  If pcregrep was compiled under Windows, there was a reference to the
961        function pcregrep_exit() before it was defined. I am assuming this was
962        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
963        reported by a user. I've moved the definition above the reference.
964    
965    
966    Version 8.11 10-Dec-2010
967    ------------------------
968    
969    1.  (*THEN) was not working properly if there were untried alternatives prior
970        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
971        backtracked to try for "b" instead of moving to the next alternative branch
972        at the same level (in this case, to look for "c"). The Perl documentation
973        is clear that when (*THEN) is backtracked onto, it goes to the "next
974        alternative in the innermost enclosing group".
975    
976    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
977        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
978        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
979        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
980        (*THEN).
981    
982    3.  If \s appeared in a character class, it removed the VT character from
983        the class, even if it had been included by some previous item, for example
984        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
985        of \s, but is part of the POSIX "space" class.)
986    
987    4.  A partial match never returns an empty string (because you can always
988        match an empty string at the end of the subject); however the checking for
989        an empty string was starting at the "start of match" point. This has been
990        changed to the "earliest inspected character" point, because the returned
991        data for a partial match starts at this character. This means that, for
992        example, /(?<=abc)def/ gives a partial match for the subject "abc"
993        (previously it gave "no match").
994    
995    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
996        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
997        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
998        has an implication that the given string is incomplete (because a partial
999        match is preferred over a full match). For this reason, these items now
1000        give a partial match in this situation. [Aside: previously, the one case
1001        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1002        match rather than a full match, which was wrong by the old rules, but is
1003        now correct.]
1004    
1005    6.  There was a bug in the handling of #-introduced comments, recognized when
1006        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1007        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1008        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1009        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1010        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1011        places in pcre_compile().
1012    
1013    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1014        comments when looking ahead for named forward references to subpatterns,
1015        the only newline sequence it recognized was NL. It now handles newlines
1016        according to the set newline convention.
1017    
1018    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1019        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1020        cater for a lack of strerror(). These oversights have been fixed.
1021    
1022    9.  Added --match-limit and --recursion-limit to pcregrep.
1023    
1024    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1025    
1026    11. When the -o option was used, pcregrep was setting a return code of 1, even
1027        when matches were found, and --line-buffered was not being honoured.
1028    
1029    12. Added an optional parentheses number to the -o and --only-matching options
1030        of pcregrep.
1031    
1032    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1033        can match an empty string. The code to do it in pcretest and pcredemo
1034        needed fixing:
1035    
1036        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1037            only one byte after an empty string match just before CRLF (this case
1038            just got forgotten; "any" and "anycrlf" were OK).
1039    
1040        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1041            mode when an empty string match preceded an ASCII character followed by
1042            a non-ASCII character. (The code for advancing by one character rather
1043            than one byte was nonsense.)
1044    
1045        (c) The pcredemo.c sample program did not have any code at all to handle
1046            the cases when CRLF is a valid newline sequence.
1047    
1048    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1049        as a starting offset was within the subject string. There is now a new
1050        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1051        negative or greater than the length of the string. In order to test this,
1052        pcretest is extended to allow the setting of negative starting offsets.
1053    
1054    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1055        starting offset points to the beginning of a UTF-8 character was
1056        unnecessarily clumsy. I tidied it up.
1057    
1058    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1059        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1060    
1061    17. Nobody had reported that the --include_dir option, which was added in
1062        release 7.7 should have been called --include-dir (hyphen, not underscore)
1063        for compatibility with GNU grep. I have changed it to --include-dir, but
1064        left --include_dir as an undocumented synonym, and the same for
1065        --exclude-dir, though that is not available in GNU grep, at least as of
1066        release 2.5.4.
1067    
1068    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1069        characters from a string of bytes) have been redefined so as not to use
1070        loops, in order to improve performance in some environments. At the same
1071        time, I abstracted some of the common code into auxiliary macros to save
1072        repetition (this should not affect the compiled code).
1073    
1074    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1075        compile-time error is now given if \c is not followed by an ASCII
1076        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1077        different, and any byte value is allowed.)
1078    
1079    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1080        START_OPTIMIZE option, which is now allowed at compile time - but just
1081        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1082        to pcregrep and other applications that have no direct access to PCRE
1083        options. The new /Y option in pcretest sets this option when calling
1084        pcre_compile().
1085    
1086    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1087        back references. Groups containing recursive back references were forced to
1088        be atomic by that change, but in the case of named groups, the amount of
1089        memory required was incorrectly computed, leading to "Failed: internal
1090        error: code overflow". This has been fixed.
1091    
1092    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1093        pcretest.c, to avoid build problems in some Borland environments.
1094    
1095    
1096    Version 8.10 25-Jun-2010
1097    ------------------------
1098    
1099    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1100        THEN.
1101    
1102    2.  (*ACCEPT) was not working when inside an atomic group.
1103    
1104    3.  Inside a character class, \B is treated as a literal by default, but
1105        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1106        causes the error). The code is unchanged, but I tidied the documentation.
1107    
1108    4.  Inside a character class, PCRE always treated \R and \X as literals,
1109        whereas Perl faults them if its -w option is set. I have changed PCRE so
1110        that it faults them when PCRE_EXTRA is set.
1111    
1112    5.  Added support for \N, which always matches any character other than
1113        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1114    
1115    6.  When compiling pcregrep with newer versions of gcc which may have
1116        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1117        declared with attribute warn_unused_result" were given. Just casting the
1118        result to (void) does not stop the warnings; a more elaborate fudge is
1119        needed. I've used a macro to implement this.
1120    
1121    7.  Minor change to pcretest.c to avoid a compiler warning.
1122    
1123    8.  Added four artifical Unicode properties to help with an option to make
1124        \s etc use properties (see next item). The new properties are: Xan
1125        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1126    
1127    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1128        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1129        this option. Modified pcretest to add /W to test this facility. Added
1130        REG_UCP to make it available via the POSIX interface.
1131    
1132    10. Added --line-buffered to pcregrep.
1133    
1134    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1135        studied, and the match started with a letter with a code point greater than
1136        127 whose first byte was different to the first byte of the other case of
1137        the letter, the other case of this starting letter was not recognized
1138        (#976).
1139    
1140    12. If a pattern that was studied started with a repeated Unicode property
1141        test, for example, \p{Nd}+, there was the theoretical possibility of
1142        setting up an incorrect bitmap of starting bytes, but fortunately it could
1143        not have actually happened in practice until change 8 above was made (it
1144        added property types that matched character-matching opcodes).
1145    
1146    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1147        possible starting bytes for non-anchored patterns.
1148    
1149    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1150        \R, and also a number of cases that involve Unicode properties, both
1151        explicit and implicit when PCRE_UCP is set.
1152    
1153    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1154        input, it could crash or give wrong results if characters with values
1155        greater than 0xc0 were present in the subject string. (Detail: it assumed
1156        UTF-8 input when processing these items.)
1157    
1158    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1159        size_t is 64-bit (#991).
1160    
1161    17. Added a check for running out of memory when PCRE is compiled with
1162        --disable-stack-for-recursion (#990).
1163    
1164    18. If the last data line in a file for pcretest does not have a newline on
1165        the end, a newline was missing in the output.
1166    
1167    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1168        less than 128) in its various bitmaps. However, there is a facility for
1169        generating tables according to the current locale when PCRE is compiled. It
1170        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1171        characters, are recognized by isspace() and therefore were getting set in
1172        these tables, and indeed these tables seem to approximate to ISO 8859. This
1173        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1174        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1175        which of course cannot start UTF-8 characters. I have changed the code so
1176        that only real ASCII characters (less than 128) and the correct starting
1177        bytes for UTF-8 encodings are set for characters greater than 127 when in
1178        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1179        altogether.)
1180    
1181    20. Added the /T option to pcretest so as to be able to run tests with non-
1182        standard character tables, thus making it possible to include the tests
1183        used for 19 above in the standard set of tests.
1184    
1185    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1186        reference to a subpattern the other side of a comment that contains an
1187        opening parenthesis caused either an internal compiling error, or a
1188        reference to the wrong subpattern.
1189    
1190    
1191    Version 8.02 19-Mar-2010
1192    ------------------------
1193    
1194    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1195    
1196    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1197        configured.
1198    
1199    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1200        original author of that file, following a query about its status.
1201    
1202    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1203        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1204    
1205    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1206        quantifier applied to a forward-referencing subroutine call, could compile
1207        incorrect code or give the error "internal error: previously-checked
1208        referenced subpattern not found".
1209    
1210    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1211        variables to point to external functions. For these systems, therefore,
1212        pcre_malloc etc. are now initialized to local functions that call the
1213        relevant global functions.
1214    
1215    7.  There were two entries missing in the vectors called coptable and poptable
1216        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1217        I've fixed the data, and added a kludgy way of testing at compile time that
1218        the lengths are correct (equal to the number of opcodes).
1219    
1220    8.  Following on from 7, I added a similar kludge to check the length of the
1221        eint vector in pcreposix.c.
1222    
1223    9.  Error texts for pcre_compile() are held as one long string to avoid too
1224        much relocation at load time. To find a text, the string is searched,
1225        counting zeros. There was no check for running off the end of the string,
1226        which could happen if a new error number was added without updating the
1227        string.
1228    
1229    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1230    
1231    11. \K was not working if it appeared in an atomic group or in a group that
1232        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1233        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1234        the assertion is positive, but not if it is negative.
1235    
1236    12. Change 11 fortuitously reduced the size of the stack frame used in the
1237        "match()" function of pcre_exec.c by one pointer. Forthcoming
1238        implementation of support for (*MARK) will need an extra pointer on the
1239        stack; I have reserved it now, so that the stack frame size does not
1240        decrease.
1241    
1242    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1243        item in branch that calls a recursion is a subroutine call - as in the
1244        second branch in the above example - was incorrectly given the compile-
1245        time error "recursive call could loop indefinitely" because pcre_compile()
1246        was not correctly checking the subroutine for matching a non-empty string.
1247    
1248    14. The checks for overrunning compiling workspace could trigger after an
1249        overrun had occurred. This is a "should never occur" error, but it can be
1250        triggered by pathological patterns such as hundreds of nested parentheses.
1251        The checks now trigger 100 bytes before the end of the workspace.
1252    
1253    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1254    
1255    
1256    Version 8.01 19-Jan-2010
1257    ------------------------
1258    
1259    1.  If a pattern contained a conditional subpattern with only one branch (in
1260        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1261        computed the wrong minimum data length (which is of course zero for such
1262        subpatterns). This could cause incorrect "no match" results.
1263    
1264    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1265        the pattern is reset in the first branch, pcre_compile() failed with
1266        "internal error: code overflow at offset...". This happened only when
1267        the reset was to the original external option setting. (An optimization
1268        abstracts leading options settings into an external setting, which was the
1269        cause of this.)
1270    
1271    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1272        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1273        assertion pattern did not match (meaning that the assertion was true), it
1274        was incorrectly treated as false if the SKIP had been reached during the
1275        matching. This also applied to assertions used as conditions.
1276    
1277    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1278        assertion subpattern, including such a pattern used as a condition,
1279        unpredictable results occurred, instead of the error return
1280        PCRE_ERROR_DFA_UITEM.
1281    
1282    5.  The C++ GlobalReplace function was not working like Perl for the special
1283        situation when an empty string is matched. It now does the fancy magic
1284        stuff that is necessary.
1285    
1286    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1287        removed. (These were left over from very, very early versions of PCRE.)
1288    
1289    7.  Some cosmetic changes to the code to make life easier when compiling it
1290        as part of something else:
1291    
1292        (a) Change DEBUG to PCRE_DEBUG.
1293    
1294        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1295            called "current" as "current_branch", to prevent a collision with the
1296            Linux macro when compiled as a kernel module.
1297    
1298        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1299            prevent a collision with the Linux macro when compiled as a kernel
1300            module.
1301    
1302    8.  In pcre_compile() there are some checks for integer overflows that used to
1303        cast potentially large values to (double). This has been changed to that
1304        when building, a check for int64_t is made, and if it is found, it is used
1305        instead, thus avoiding the use of floating point arithmetic. (There is no
1306        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1307        double.
1308    
1309    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1310        2005 (difference between two addresses compared to an unsigned value).
1311    
1312    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1313        custom one, because of the following reported problem in Windows:
1314    
1315          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1316              under Win32.
1317          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1318              therefore missing the function definition.
1319          - The compiler thus generates a "C" signature for the test function.
1320          - The linker fails to find the "C" function.
1321          - PCRE fails to configure if asked to do so against libbz2.
1322    
1323    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1324        messages were output:
1325    
1326          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1327          rerunning libtoolize, to keep the correct libtool macros in-tree.
1328          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1329    
1330        I have done both of these things.
1331    
1332    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1333        most of the time, it *can* run out if it is given a pattern that contains a
1334        runaway infinite recursion. I updated the discussion in the pcrestack man
1335        page.
1336    
1337    13. Now that we have gone to the x.xx style of version numbers, the minor
1338        version may start with zero. Using 08 or 09 is a bad idea because users
1339        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1340        interpreted as invalid octal numbers. I've updated the previous comment in
1341        configure.ac, and also added a check that gives an error if 08 or 09 are
1342        used.
1343    
1344    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1345        causing partial matching to fail when the end of the subject matched \W
1346        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1347    
1348    15. There were some discrepancies between the declarations in pcre_internal.h
1349        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1350        their definitions. The declarations used "const uschar *" and the
1351        definitions used USPTR. Even though USPTR is normally defined as "const
1352        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1353        reported that: "This difference in casting confuses some C++ compilers, for
1354        example, SunCC recognizes above declarations as different functions and
1355        generates broken code for hbpcre." I have changed the declarations to use
1356        USPTR.
1357    
1358    16. GNU libtool is named differently on some systems. The autogen.sh script now
1359        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1360        (FreeBSD).
1361    
1362    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1363        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1364        comment: "Figure out how to create a longlong from a string: strtoll and
1365        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1366        instance, but it only takes 2 args instead of 3!"
1367    
1368    18. A subtle bug concerned with back references has been fixed by a change of
1369        specification, with a corresponding code fix. A pattern such as
1370        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1371        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1372        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1373        same bug. Such groups have to be quantified to be useful, or contained
1374        inside another quantified group. (If there's no repetition, the reference
1375        can never match.) The problem arises because, having left the group and
1376        moved on to the rest of the pattern, a later failure that backtracks into
1377        the group uses the captured value from the final iteration of the group
1378        rather than the correct earlier one. I have fixed this in PCRE by forcing
1379        any group that contains a reference to itself to be an atomic group; that
1380        is, there cannot be any backtracking into it once it has completed. This is
1381        similar to recursive and subroutine calls.
1382    
1383    
1384    Version 8.00 19-Oct-09
1385    ----------------------
1386    
1387    1.  The table for translating pcre_compile() error codes into POSIX error codes
1388        was out-of-date, and there was no check on the pcre_compile() error code
1389        being within the table. This could lead to an OK return being given in
1390        error.
1391    
1392    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1393        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1394        in a Windows environment.
1395    
1396    3.  The pcregrep --count option prints the count for each file even when it is
1397        zero, as does GNU grep. However, pcregrep was also printing all files when
1398        --files-with-matches was added. Now, when both options are given, it prints
1399        counts only for those files that have at least one match. (GNU grep just
1400        prints the file name in this circumstance, but including the count seems
1401        more useful - otherwise, why use --count?) Also ensured that the
1402        combination -clh just lists non-zero counts, with no names.
1403    
1404    4.  The long form of the pcregrep -F option was incorrectly implemented as
1405        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1406        but it seems right to fix it, and I didn't think it was worth preserving
1407        the old behaviour.
1408    
1409    5.  The command line items --regex=pattern and --regexp=pattern were not
1410        recognized by pcregrep, which required --regex pattern or --regexp pattern
1411        (with a space rather than an '='). The man page documented the '=' forms,
1412        which are compatible with GNU grep; these now work.
1413    
1414    6.  No libpcreposix.pc file was created for pkg-config; there was just
1415        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1416    
1417    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1418        when UCP support is not needed, by modifying the Python script that
1419        generates it from Unicode data files. This should not matter if the module
1420        is correctly used as a library, but I received one complaint about 50K of
1421        unwanted data. My guess is that the person linked everything into his
1422        program rather than using a library. Anyway, it does no harm.
1423    
1424    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1425        was a minimum greater than 1 for a wide character in a possessive
1426        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1427        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1428        character. Chaos in the form of incorrect output or a compiling loop could
1429        result.
1430    
1431    9.  The restrictions on what a pattern can contain when partial matching is
1432        requested for pcre_exec() have been removed. All patterns can now be
1433        partially matched by this function. In addition, if there are at least two
1434        slots in the offset vector, the offset of the earliest inspected character
1435        for the match and the offset of the end of the subject are set in them when
1436        PCRE_ERROR_PARTIAL is returned.
1437    
1438    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1439        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1440        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1441        and may be more useful for multi-segment matching.
1442    
1443    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1444        used to be given if ever the end of the subject was reached; now it is
1445        given only if matching could not proceed because another character was
1446        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1447        string "Z", which now yields "no match" instead of "partial match". In the
1448        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1449        final character ended with (*FAIL).
1450    
1451    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1452        if the pattern had a "must contain" character that was already found in the
1453        earlier partial match, unless partial matching was again requested. For
1454        example, with the pattern /dog.(body)?/, the "must contain" character is
1455        "g". If the first part-match was for the string "dog", restarting with
1456        "sbody" failed. This bug has been fixed.
1457    
1458    13. The string returned by pcre_dfa_exec() after a partial match has been
1459        changed so that it starts at the first inspected character rather than the
1460        first character of the match. This makes a difference only if the pattern
1461        starts with a lookbehind assertion or \b or \B (\K is not supported by
1462        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1463        matching functions compatible, and I think it's the right thing to do.
1464    
1465    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1466        so that the demonstration program is easily available in environments where
1467        PCRE has not been installed from source.
1468    
1469    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1470        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1471        library.
1472    
1473    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1474        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1475        is not the first non-POSIX option to be added. Clearly some people find
1476        these options useful.
1477    
1478    17. If a caller to the POSIX matching function regexec() passes a non-zero
1479        value for nmatch with a NULL value for pmatch, the value of
1480        nmatch is forced to zero.
1481    
1482    18. RunGrepTest did not have a test for the availability of the -u option of
1483        the diff command, as RunTest does. It now checks in the same way as
1484        RunTest, and also checks for the -b option.
1485    
1486    19. If an odd number of negated classes containing just a single character
1487        interposed, within parentheses, between a forward reference to a named
1488        subpattern and the definition of the subpattern, compilation crashed with
1489        an internal error, complaining that it could not find the referenced
1490        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1491        [The bug was that it was starting one character too far in when skipping
1492        over the character class, thus treating the ] as data rather than
1493        terminating the class. This meant it could skip too much.]
1494    
1495    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1496        /g option in pcretest when the pattern contains \K, which makes it possible
1497        to have an empty string match not at the start, even when the pattern is
1498        anchored. Updated pcretest and pcredemo to use this option.
1499    
1500    21. If the maximum number of capturing subpatterns in a recursion was greater
1501        than the maximum at the outer level, the higher number was returned, but
1502        with unset values at the outer level. The correct (outer level) value is
1503        now given.
1504    
1505    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1506        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1507        make it do so. The string so far is captured, making this feature
1508        compatible with Perl.
1509    
1510    23. The tests have been re-organized, adding tests 11 and 12, to make it
1511        possible to check the Perl 5.10 features against Perl 5.10.
1512    
1513    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1514        pattern matches a fixed length string. PCRE did not allow this; now it
1515        does. Neither allows recursion.
1516    
1517    25. I finally figured out how to implement a request to provide the minimum
1518        length of subject string that was needed in order to match a given pattern.
1519        (It was back references and recursion that I had previously got hung up
1520        on.) This code has now been added to pcre_study(); it finds a lower bound
1521        to the length of subject needed. It is not necessarily the greatest lower
1522        bound, but using it to avoid searching strings that are too short does give
1523        some useful speed-ups. The value is available to calling programs via
1524        pcre_fullinfo().
1525    
1526    26. While implementing 25, I discovered to my embarrassment that pcretest had
1527        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1528        study optimizations had never been tested with that matching function.
1529        Oops. What is worse, even when it was passed study data, there was a bug in
1530        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1531        were also very few tests of studied patterns with pcre_dfa_exec().
1532    
1533    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1534        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1535        on the other side of the coin, they are no longer allowed to have different
1536        names, because these cannot be distinguished in PCRE, and this has caused
1537        confusion. (This is a difference from Perl.)
1538    
1539    28. When duplicate subpattern names are present (necessarily with different
1540        numbers, as required by 27 above), and a test is made by name in a
1541        conditional pattern, either for a subpattern having been matched, or for
1542        recursion in such a pattern, all the associated numbered subpatterns are
1543        tested, and the overall condition is true if the condition is true for any
1544        one of them. This is the way Perl works, and is also more like the way
1545        testing by number works.
1546    
1547    
1548    Version 7.9 11-Apr-09
1549    ---------------------
1550    
1551    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1552        (pcretest), all targets were linked against these libraries. This included
1553        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1554        libraries. This caused unwanted dependencies to be created. This problem
1555        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1556        pcretest is linked with readline.
1557    
1558    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1559        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1560        moved outside it again, because FALSE and TRUE are already defined in AIX,
1561        but BOOL is not.
1562    
1563    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1564        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1565    
1566    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1567        hyphens) following file names and line numbers when outputting matching
1568        lines. This is not true; no spaces are inserted. I have also clarified the
1569        wording for the --colour (or --color) option.
1570    
1571    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1572        was not coloured; this is different to GNU grep, so I have changed it to be
1573        the same.
1574    
1575    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1576        each matching line was coloured. Now it goes on to look for further matches
1577        of any of the test patterns, which is the same behaviour as GNU grep.
1578    
1579    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1580        doesn't make sense to accept an empty string match in pcregrep, so I have
1581        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1582        seems to be how GNU grep behaves.
1583    
1584    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1585        start or after a newline", because the conditional assertion was not being
1586        correctly handled. The rule now is that both the assertion and what follows
1587        in the first alternative must satisfy the test.
1588    
1589    9.  If auto-callout was enabled in a pattern with a conditional group whose
1590        condition was an assertion, PCRE could crash during matching, both with
1591        pcre_exec() and pcre_dfa_exec().
1592    
1593    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1594        used for matching.
1595    
1596    11. Unicode property support in character classes was not working for
1597        characters (bytes) greater than 127 when not in UTF-8 mode.
1598    
1599    12. Added the -M command line option to pcretest.
1600    
1601    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1602    
1603    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1604    
1605    16. Added comments and documentation about mis-use of no_arg in the C++
1606        wrapper.
1607    
1608    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1609        from Martin Jerabek that uses macro names for all relevant character and
1610        string constants.
1611    
1612    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1613        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1614        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1615        these, but not everybody uses configure.
1616    
1617    19. A conditional group that had only one branch was not being correctly
1618        recognized as an item that could match an empty string. This meant that an
1619        enclosing group might also not be so recognized, causing infinite looping
1620        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1621        with the subject "ab", where knowledge that the repeated group can match
1622        nothing is needed in order to break the loop.
1623    
1624    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1625        exec(), but without supplying a callout function, matching went wrong.
1626    
1627    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1628        leak if the size of the offset vector was greater than 30. When the vector
1629        is smaller, the saved offsets during recursion go onto a local stack
1630        vector, but for larger vectors malloc() is used. It was failing to free
1631        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1632        error, in fact).
1633    
1634    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1635        heapframe that is used only when UTF-8 support is enabled. This caused no
1636        problem, but was untidy.
1637    
1638    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1639        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1640        included within another project.
1641    
1642    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1643        slightly modified by me:
1644    
1645          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1646              not building pcregrep.
1647    
1648          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1649              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1650    
1651    25. Forward references, both numeric and by name, in patterns that made use of
1652        duplicate group numbers, could behave incorrectly or give incorrect errors,
1653        because when scanning forward to find the reference group, PCRE was not
1654        taking into account the duplicate group numbers. A pattern such as
1655        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1656    
1657    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1658        the feature of a custom pointer more persuasive (as requested by a user).
1659    
1660    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1661        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1662        pre-defined.
1663    
1664    28. Added support for (*UTF8) at the start of a pattern.
1665    
1666    29. Arrange for flags added by the "release type" setting in CMake to be shown
1667        in the configuration summary.
1668    
1669    
1670    Version 7.8 05-Sep-08
1671    ---------------------
1672    
1673    1.  Replaced UCP searching code with optimized version as implemented for Ad
1674        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1675        stage table and inline lookup instead of a function, giving speed ups of 2
1676        to 5 times on some simple patterns that I tested. Permission was given to
1677        distribute the MultiStage2.py script that generates the tables (it's not in
1678        the tarball, but is in the Subversion repository).
1679    
1680    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1681        scripts.
1682    
1683    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1684        a group with a zero qualifier. The result of the study could be incorrect,
1685        or the function might crash, depending on the pattern.
1686    
1687    4.  Caseless matching was not working for non-ASCII characters in back
1688        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1689        It now works when Unicode Property Support is available.
1690    
1691    5.  In pcretest, an escape such as \x{de} in the data was always generating
1692        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1693        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1694        truncation.
1695    
1696    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1697    
1698    7.  Added two (int) casts to pcregrep when printing the difference of two
1699        pointers, in case they are 64-bit values.
1700    
1701    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1702        test 2 if it fails.
1703    
1704    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1705        and a #define of that name to empty if it is not externally set. This is to
1706        allow users of MSVC to set it if necessary.
1707    
1708    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1709        the convenience functions in the pcre_get.c source file.
1710    
1711    11. An option change at the start of a pattern that had top-level alternatives
1712        could cause overwriting and/or a crash. This command provoked a crash in
1713        some environments:
1714    
1715          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1716    
1717        This potential security problem was recorded as CVE-2008-2371.
1718    
1719    12. For a pattern where the match had to start at the beginning or immediately
1720        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1721        pcre_dfa_exec() could read past the end of the passed subject if there was
1722        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1723        pcretest so that it places the subject at the end of its malloc-ed buffer.
1724    
1725    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1726        exec() might read past the end of the data buffer in UTF-8 mode.
1727    
1728    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1729        the data contained the byte 0x85 as part of a UTF-8 character within its
1730        first line. This applied both to normal and DFA matching.
1731    
1732    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1733        /^[^d]*?$/8 failed to match "abc".
1734    
1735    16. Added a missing copyright notice to pcrecpp_internal.h.
1736    
1737    17. Make it more clear in the documentation that values returned from
1738        pcre_exec() in ovector are byte offsets, not character counts.
1739    
1740    18. Tidied a few places to stop certain compilers from issuing warnings.
1741    
1742    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1743        supplied by Stefan Weber. I made a further small update for 7.8 because
1744        there is a change of source arrangements: the pcre_searchfuncs.c module is
1745        replaced by pcre_ucd.c.
1746    
1747    
1748    Version 7.7 07-May-08
1749    ---------------------
1750    
1751    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1752        a string to a long long, pretend we don't even have a long long." This is
1753        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1754    
1755    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1756        pre-7.6 versions, which defined a global no_arg variable instead of putting
1757        it in the RE class. (See also #8 below.)
1758    
1759    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1760        Lopes.
1761    
1762    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1763    
1764        (1) The include/exclude patterns were being applied to the whole pathnames
1765            of files, instead of just to the final components.
1766    
1767        (2) If there was more than one level of directory, the subdirectories were
1768            skipped unless they satisfied the include/exclude conditions. This is
1769            inconsistent with GNU grep (and could even be seen as contrary to the
1770            pcregrep specification - which I improved to make it absolutely clear).
1771            The action now is always to scan all levels of directory, and just
1772            apply the include/exclude patterns to regular files.
1773    
1774    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1775        --exclude_dir in the tests to avoid scanning .svn directories.
1776    
1777    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1778        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1779        doesn't support NULs in patterns.
1780    
1781    7.  Added some missing "const"s to declarations of static tables in
1782        pcre_compile.c and pcre_dfa_exec.c.
1783    
1784    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1785        caused by fix #2  above. (Subsequently also a second patch to fix the
1786        first patch. And a third patch - this was a messy problem.)
1787    
1788    9.  Applied Craig's patch to remove the use of push_back().
1789    
1790    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1791        matching function regexec().
1792    
1793    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1794        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1795        references. PCRE supports relative numbers with this syntax (I don't think
1796        Oniguruma does).
1797    
1798    12. Previously, a group with a zero repeat such as (...){0} was completely
1799        omitted from the compiled regex. However, this means that if the group
1800        was called as a subroutine from elsewhere in the pattern, things went wrong
1801        (an internal error was given). Such groups are now left in the compiled
1802        pattern, with a new opcode that causes them to be skipped at execution
1803        time.
1804    
1805    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1806        to the way PCRE behaves:
1807    
1808        (a) A lone ] character is dis-allowed (Perl treats it as data).
1809    
1810        (b) A back reference to an unmatched subpattern matches an empty string
1811            (Perl fails the current match path).
1812    
1813        (c) A data ] in a character class must be notated as \] because if the
1814            first data character in a class is ], it defines an empty class. (In
1815            Perl it is not possible to have an empty class.) The empty class []
1816            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1817            The negative empty class [^] matches any one character, independently
1818            of the DOTALL setting.
1819    
1820    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1821        non-existent subpattern following a character class starting with ']' and
1822        containing () gave an internal compiling error instead of "reference to
1823        non-existent subpattern". Fortunately, when the pattern did exist, the
1824        compiled code was correct. (When scanning forwards to check for the
1825        existencd of the subpattern, it was treating the data ']' as terminating
1826        the class, so got the count wrong. When actually compiling, the reference
1827        was subsequently set up correctly.)
1828    
1829    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1830        it was being rejected as not supported by pcre_dfa_exec(), even though
1831        other assertions are supported. I have made pcre_dfa_exec() support
1832        (*FAIL).
1833    
1834    16. The implementation of 13c above involved the invention of a new opcode,
1835        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1836        cannot be changed at match time, I realized I could make a small
1837        improvement to matching performance by compiling OP_ALLANY instead of
1838        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1839        on the OP_ANY path.
1840    
1841    17. Compiling pcretest on Windows with readline support failed without the
1842        following two fixes: (1) Make the unistd.h include conditional on
1843        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1844    
1845    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1846        ncurses library to be included for pcretest when ReadLine support is
1847        requested, but also to allow for it to be overridden. This patch came from
1848        Daniel Bergström.
1849    
1850    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1851        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1852        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1853        spotting this.
1854    
1855    
1856    Version 7.6 28-Jan-08
1857    ---------------------
1858    
1859    1.  A character class containing a very large number of characters with
1860        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1861        overflow.
1862    
1863    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1864        HAVE_LONG_LONG is not defined.
1865    
1866    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1867        bring them up to date and include new features. This patch includes:
1868    
1869        - Fixed PH's badly added libz and libbz2 support.
1870        - Fixed a problem with static linking.
1871        - Added pcredemo. [But later removed - see 7 below.]
1872        - Fixed dftables problem and added an option.
1873        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1874            HAVE_LONG_LONG.
1875        - Added readline support for pcretest.
1876        - Added an listing of the option settings after cmake has run.
1877    
1878    4.  A user submitted a patch to Makefile that makes it easy to create
1879        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1880        Makefile.am that cause it to include this special target, without
1881        affecting anything else. Note that the same mingw target plus all
1882        the other distribution libraries and programs are now supported
1883        when configuring with CMake (see 6 below) instead of with
1884        Configure/Make.
1885    
1886    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1887        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1888        exported in the Windows port". It has not yet been confirmed that the patch
1889        solves the problem, but it does no harm.
1890    
1891    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1892        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1893        with CMake, and also correct the comment about stack recursion.
1894    
1895    7.  Remove the automatic building of pcredemo from the ./configure system and
1896        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1897        of a program that users should build themselves after PCRE is installed, so
1898        building it automatically is not really right. What is more, it gave
1899        trouble in some build environments.
1900    
1901    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1902    
1903    
1904    Version 7.5 10-Jan-08
1905    ---------------------
1906    
1907    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1908        values in parens when parsing an RE using the C++ wrapper."
1909    
1910    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1911        Characters greater than 255 were excluded from the class instead of being
1912        included.
1913    
1914    3.  The same bug as (2) above applied to negated POSIX classes such as
1915        [:^space:].
1916    
1917    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1918        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1919        I have changed it.
1920    
1921    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1922        first named subpattern) and a construct such as (?&a) would reference the
1923        first named subpattern whose name started with "a" (in other words, the
1924        length check was missing). Both these problems are fixed. "Subpattern name
1925        expected" is now given for (?&) (a zero-length name), and this patch also
1926        makes it give the same error for \k'' (previously it complained that that
1927        was a reference to a non-existent subpattern).
1928    
1929    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1930        this is right because (?- can be followed by option settings as well as by
1931        digits. I have, however, made the messages clearer.
1932    
1933    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1934        than the number used in the conditional) now cause a compile-time error.
1935        This is actually not compatible with Perl, which accepts such patterns, but
1936        treats the conditional as always being FALSE (as PCRE used to), but it
1937        seems to me that giving a diagnostic is better.
1938    
1939    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1940        and messages.
1941    
1942    9.  Fix two occurrences of "backslash" in comments that should have been
1943        "backspace".
1944    
1945    10. Remove two redundant lines of code that can never be obeyed (their function
1946        was moved elsewhere).
1947    
1948    11. The program that makes PCRE's Unicode character property table had a bug
1949        which caused it to generate incorrect table entries for sequences of
1950        characters that have the same character type, but are in different scripts.
1951        It amalgamated them into a single range, with the script of the first of
1952        them. In other words, some characters were in the wrong script. There were
1953        thirteen such cases, affecting characters in the following ranges:
1954    
1955          U+002b0 - U+002c1
1956          U+0060c - U+0060d
1957          U+0061e - U+00612
1958          U+0064b - U+0065e
1959          U+0074d - U+0076d
1960          U+01800 - U+01805
1961          U+01d00 - U+01d77
1962          U+01d9b - U+01dbf
1963          U+0200b - U+0200f
1964          U+030fc - U+030fe
1965          U+03260 - U+0327f
1966          U+0fb46 - U+0fbb1
1967          U+10450 - U+1049d
1968    
1969    12. The -o option (show only the matching part of a line) for pcregrep was not
1970        compatible with GNU grep in that, if there was more than one match in a
1971        line, it showed only the first of them. It now behaves in the same way as
1972        GNU grep.
1973    
1974    13. If the -o and -v options were combined for pcregrep, it printed a blank
1975        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1976        does the same. The return code can be used to tell if there were any
1977        non-matching lines.
1978    
1979    14. Added --file-offsets and --line-offsets to pcregrep.
1980    
1981    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1982        infinitely looping recursion. The bug was that positive lookaheads were not
1983        being skipped when checking for a possible empty match (negative lookaheads
1984        and both kinds of lookbehind were skipped).
1985    
1986    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1987        inclusion of <windows.h> to before rather than after the definition of
1988        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1989    
1990    17. Specifying a possessive quantifier with a specific limit for a Unicode
1991        character property caused pcre_compile() to compile bad code, which led at
1992        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1993        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1994        caused the error; without that there was no problem.
1995    
1996    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1997    
1998    19. Added --enable-pcretest-libreadline.
1999    
2000    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2001        RE::GlobalReplace(). As a result, the number of replacements returned was
2002        double what it should be. I removed one of the increments, but Craig sent a
2003        later patch that removed the other one (the right fix) and added unit tests
2004        that check the return values (which was not done before).
2005    
2006    21. Several CMake things:
2007    
2008        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2009            the names libpcre and libpcreposix, not just pcre and pcreposix.
2010    
2011        (2) The above change means that pcretest and pcregrep are now correctly
2012            linked with the newly-built libraries, not previously installed ones.
2013    
2014        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2015    
2016    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2017        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2018        UTF-8 newline character). The key issue is that the pattern starts .*;
2019        this means that the match must be either at the beginning, or after a
2020        newline. The bug was in the code for advancing after a failed match and
2021        checking that the new position followed a newline. It was not taking
2022        account of UTF-8 characters correctly.
2023    
2024    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2025        character classes. PCRE was not treating the sequence [:...:] as a
2026        character class unless the ... were all letters. Perl, however, seems to
2027        allow any characters between [: and :], though of course it rejects as
2028        unknown any "names" that contain non-letters, because all the known class
2029        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2030        for example, whereas PCRE did not - it did not recognize a POSIX character
2031        class. This seemed a bit dangerous, so the code has been changed to be
2032        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2033        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2034        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2035        Perl does, and where it didn't before.
2036    
2037    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2038        Windows environments %n is disabled by default.
2039    
2040    
2041    Version 7.4 21-Sep-07
2042    ---------------------
2043    
2044    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2045        means that a class such as [\s] counted as "explicit reference to CR or
2046        LF". That isn't really right - the whole point of the change was to try to
2047        help when there was an actual mention of one of the two characters. So now
2048        the change happens only if \r or \n (or a literal CR or LF) character is
2049        encountered.
2050    
2051    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2052        of both had grown to the point where there were only 3 bits left.
2053        Fortunately, there was spare space in the data structure, and so I have
2054        moved the internal flags into a new 16-bit field to free up more option
2055        bits.
2056    
2057    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2058        but did not set the internal JCHANGED flag - either of these is enough to
2059        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2060        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2061        start sets both bits.
2062    
2063    4.  Added options (at build time, compile time, exec time) to change \R from
2064        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2065    
2066    5.  doc/pcresyntax.html was missing from the distribution.
2067    
2068    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2069        compatibility, even though it is no longer used.
2070    
2071    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2072        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2073        windows.h file is present (where different names are used). [This was
2074        reversed later after testing - see 16 below.]
2075    
2076    8.  Changed all #include <config.h> to #include "config.h". There were also
2077        some further <pcre.h> cases that I changed to "pcre.h".
2078    
2079    9.  When pcregrep was used with the --colour option, it missed the line ending
2080        sequence off the lines that it output.
2081    
2082    10. It was pointed out to me that arrays of string pointers cause lots of
2083        relocations when a shared library is dynamically loaded. A technique of
2084        using a single long string with a table of offsets can drastically reduce
2085        these. I have refactored PCRE in four places to do this. The result is
2086        dramatic:
2087    
2088          Originally:                          290
2089          After changing UCP table:            187
2090          After changing error message table:   43
2091          After changing table of "verbs"       36
2092          After changing table of Posix names   22
2093    
2094        Thanks to the folks working on Gregex for glib for this insight.
2095    
2096    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2097        unicode-properties was also set.
2098    
2099    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2100    
2101    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2102        checked only for CRLF.
2103    
2104    14. Added casts to pcretest.c to avoid compiler warnings.
2105    
2106    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2107    
2108    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2109        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2110        entirely. This removes changes made in 7 above.
2111    
2112    17. The CMake files have been updated, and there is now more information about
2113        building with CMake in the NON-UNIX-USE document.
2114    
2115    
2116    Version 7.3 28-Aug-07
2117    ---------------------
2118    
2119     1. In the rejigging of the build system that eventually resulted in 7.1, the
2120        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2121        brackets there is not right, since it causes compilers to look for an
2122        installed pcre.h, not the version that is in the source that is being
2123        compiled (which of course may be different). I have changed it back to:
2124    
2125          #include "pcre.h"
2126    
2127        I have a vague recollection that the change was concerned with compiling in
2128        different directories, but in the new build system, that is taken care of
2129        by the VPATH setting the Makefile.
2130    
2131     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2132        when the subject happened to end in the byte 0x85 (e.g. if the last
2133        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2134        characters but of course it shouldn't be taken as a newline when it is part
2135        of another character. The bug was that, for an unlimited repeat of . in
2136        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2137        characters when looking for a newline.
2138    
2139     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2140    
2141     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2142        in debug output.
2143    
2144     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2145        long printing in the pcrecpp unittest when running under MinGW.
2146    
2147     6. ESC_K was left out of the EBCDIC table.
2148    
2149     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2150        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2151        limit also applies to "virtual nesting" when a pattern is recursive, and in
2152        this case 1000 isn't so big. I have been able to remove this limit at the
2153        expense of backing off one optimization in certain circumstances. Normally,
2154        when pcre_exec() would call its internal match() function recursively and
2155        immediately return the result unconditionally, it uses a "tail recursion"
2156        feature to save stack. However, when a subpattern that can match an empty
2157        string has an unlimited repetition quantifier, it no longer makes this
2158        optimization. That gives it a stack frame in which to save the data for
2159        checking that an empty string has been matched. Previously this was taken
2160        from the 1000-entry workspace that had been reserved. So now there is no
2161        explicit limit, but more stack is used.
2162    
2163     8. Applied Daniel's patches to solve problems with the import/export magic
2164        syntax that is required for Windows, and which was going wrong for the
2165        pcreposix and pcrecpp parts of the library. These were overlooked when this
2166        problem was solved for the main library.
2167    
2168     9. There were some crude static tests to avoid integer overflow when computing
2169        the size of patterns that contain repeated groups with explicit upper
2170        limits. As the maximum quantifier is 65535, the maximum group length was
2171        set at 30,000 so that the product of these two numbers did not overflow a
2172        32-bit integer. However, it turns out that people want to use groups that
2173        are longer than 30,000 bytes (though not repeat them that many times).
2174        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2175        made it possible to implement the integer overflow checks in a much more
2176        dynamic way, which I have now done. The artificial limitation on group
2177        length has been removed - we now have only the limit on the total length of
2178        the compiled pattern, which depends on the LINK_SIZE setting.
2179    
2180    10. Fixed a bug in the documentation for get/copy named substring when
2181        duplicate names are permitted. If none of the named substrings are set, the
2182        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2183        empty string.
2184    
2185    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2186        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2187        because the ] is interpreted as the first data character and the
2188        terminating ] is not found. PCRE has been made compatible with Perl in this
2189        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2190        cause memory overwriting.
2191    
2192    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2193        string has been matched (to stop an infinite loop). It was not recognizing
2194        a conditional subpattern that could match an empty string if that
2195        subpattern was within another subpattern. For example, it looped when
2196        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2197        condition was not nested. This bug has been fixed.
2198    
2199    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2200        past the start of the subject in the presence of bytes with the top bit
2201        set, for example "\x8aBCD".
2202    
2203    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2204        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2205    
2206    14. Optimized (?!) to (*FAIL).
2207    
2208    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2209        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2210        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2211        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2212        does: it's just the validity check that is more restrictive.
2213    
2214    16. Inserted checks for integer overflows during escape sequence (backslash)
2215        processing, and also fixed erroneous offset values for syntax errors during
2216        backslash processing.
2217    
2218    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2219        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2220    
2221    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2222        caused an overrun.
2223    
2224    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2225        something other than just ASCII characters) inside a group that had an
2226        unlimited repeat caused a loop at compile time (while checking to see
2227        whether the group could match an empty string).
2228    
2229    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2230        [\P{Any}] did so. (Error in the code for printing property names.)
2231    
2232    21. An orphan \E inside a character class could cause a crash.
2233    
2234    22. A repeated capturing bracket such as (A)? could cause a wild memory
2235        reference during compilation.
2236    
2237    23. There are several functions in pcre_compile() that scan along a compiled
2238        expression for various reasons (e.g. to see if it's fixed length for look
2239        behind). There were bugs in these functions when a repeated \p or \P was
2240        present in the pattern. These operators have additional parameters compared
2241        with \d, etc, and these were not being taken into account when moving along
2242        the compiled data. Specifically:
2243    
2244        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2245            length.
2246    
2247        (b) An item such as \pL+ within a repeated group could cause crashes or
2248            loops.
2249    
2250        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2251            "reference to non-existent subpattern" error.
2252    
2253        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2254    
2255    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2256        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2257    
2258    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2259    
2260    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2261        character were causing crashes (broken optimization).
2262    
2263    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2264        \p or \P) caused a compile-time loop.
2265    
2266    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2267        break. For example, the unstudied pattern [\r\n]A does not match the string
2268        "\r\nA" because change 7.0/46 below moves the current point on by two
2269        characters after failing to match at the start. However, the pattern \nA
2270        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2271        the same is true. There doesn't seem any very clean way out of this, but
2272        what I have chosen to do makes the common cases work: PCRE now takes note
2273        of whether there can be an explicit match for \r or \n anywhere in the
2274        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2275        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2276        pattern has explicit CR or LF references.
2277    
2278    29. Added (*CR) etc for changing newline setting at start of pattern.
2279    
2280    
2281    Version 7.2 19-Jun-07
2282    ---------------------
2283    
2284     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2285        which is apparently normally available under Windows.
2286    
2287     2. Re-jig the pcregrep tests with different newline settings in an attempt
2288        to make them independent of the local environment's newline setting.
2289    
2290     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2291    
2292     4. Some of the "internals" tests were previously cut out when the link size
2293        was not 2, because the output contained actual offsets. The recent new
2294        "Z" feature of pcretest means that these can be cut out, making the tests
2295        usable with all link sizes.
2296    
2297     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2298        stack recursion. This gives a massive performance boost under BSD, but just
2299        a small improvement under Linux. However, it saves one field in the frame
2300        in all cases.
2301    
2302     6. Added more features from the forthcoming Perl 5.10:
2303    
2304        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2305            recursion call. It refers to the nth most recently opened parentheses.
2306    
2307        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2308            to be opened parentheses.
2309    
2310        (c) Conditions that refer to capturing parentheses can be specified
2311            relatively, for example, (?(-2)... or (?(+3)...
2312    
2313        (d) \K resets the start of the current match so that everything before
2314            is not part of it.
2315    
2316        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2317    
2318        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2319            reference syntax.
2320    
2321        (g) (?| introduces a group in which the numbering of parentheses in each
2322            alternative starts with the same number.
2323    
2324        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2325    
2326     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2327        PCRE_INFO_JCHANGED.
2328    
2329     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2330        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2331        for detecting groups that can match an empty string.
2332    
2333     9. A pattern with a very large number of alternatives (more than several
2334        hundred) was running out of internal workspace during the pre-compile
2335        phase, where pcre_compile() figures out how much memory will be needed. A
2336        bit of new cunning has reduced the workspace needed for groups with
2337        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2338        workspace instead of running out of the 4096 that are available.
2339    
2340    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2341    
2342    11. Applied patch from Google to remove an optimization that didn't quite work.
2343        The report of the bug said:
2344    
2345          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2346          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2347          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2348    
2349    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2350        it matched the wrong number of bytes.
2351    
2352    
2353    Version 7.1 24-Apr-07
2354  ---------------------  ---------------------
2355    
2356   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2357      that is more "standard", making use of automake and other autotools. There      that is more "standard", making use of automake and other Autotools. There
2358      is some re-arrangement of the files and adjustment of comments consequent      is some re-arrangement of the files and adjustment of comments consequent
2359      on this.      on this.
2360    
# Line 26  Version 7.1 05-Mar-07 Line 2375  Version 7.1 05-Mar-07
2375      .br or .in.      .br or .in.
2376    
2377   4. Updated comments in configure.ac that get placed in config.h.in and also   4. Updated comments in configure.ac that get placed in config.h.in and also
2378      arranged for config.h to be included in the distribution, with the name      arranged for config.h to be included in the distribution, with the name
2379      config.h.generic, for the benefit of those who have to compile without      config.h.generic, for the benefit of those who have to compile without
2380      Autotools (compare pcre.h, which is now distributed as pcre.h.generic).      Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2381    
2382   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2383      Weber: (1) pcre_internal.h was missing some function renames; (2) updated      Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2384      makevp.bat for the current PCRE, using the additional files !compile.txt,      makevp.bat for the current PCRE, using the additional files
2385      !linklib.txt, and pcregexp.pas.      makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2386    
2387   6. A Windows user reported a minor discrepancy with test 2, which turned out   6. A Windows user reported a minor discrepancy with test 2, which turned out
2388      to be caused by a trailing space on an input line that had got lost in his      to be caused by a trailing space on an input line that had got lost in his
# Line 60  Version 7.1 05-Mar-07 Line 2409  Version 7.1 05-Mar-07
2409      called when UTF-8 support is disabled. Otherwise there are problems with a      called when UTF-8 support is disabled. Otherwise there are problems with a
2410      shared library.      shared library.
2411    
2412  10. The emulated memmove() function in pcre_internal.h (provided for those  10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2413      environments that have neither memmove() nor bcopy()) was defining its  
2414      arguments as char * instead of void *.      (a) It was defining its arguments as char * instead of void *.
2415    
2416        (b) It was assuming that all moves were upwards in memory; this was true
2417            a long time ago when I wrote it, but is no longer the case.
2418    
2419        The emulated memove() is provided for those environments that have neither
2420        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2421        is clearly not the case, as these two bugs were recently reported.
2422    
2423  11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,  11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2424      and Detrail to create the HTML documentation, the .txt form of the man      and Detrail to create the HTML documentation, the .txt form of the man
2425      pages, and it removes trailing spaces from listed files. It also creates      pages, and it removes trailing spaces from listed files. It also creates
2426      pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter      pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2427      case, it wraps all the #defines with #ifndefs. This script should be run      case, it wraps all the #defines with #ifndefs. This script should be run
2428      before "make dist".      before "make dist".
2429    
2430    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2431        with Unicode property support.
2432    
2433        (a) For a maximizing quantifier, if the two different cases of the
2434            character were of different lengths in their UTF-8 codings (there are
2435            some cases like this - I found 11), and the matching function had to
2436            back up over a mixture of the two cases, it incorrectly assumed they
2437            were both the same length.
2438    
2439        (b) When PCRE was configured to use the heap rather than the stack for
2440            recursion during matching, it was not correctly preserving the data for
2441            the other case of a UTF-8 character when checking ahead for a match
2442            while processing a minimizing repeat. If the check also involved
2443            matching a wide character, but failed, corruption could cause an
2444            erroneous result when trying to check for a repeat of the original
2445            character.
2446    
2447    13. Some tidying changes to the testing mechanism:
2448    
2449        (a) The RunTest script now detects the internal link size and whether there
2450            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2451            values substituted by "configure". (The RunGrepTest script already did
2452            this for UTF-8.) The configure.ac script no longer substitutes the
2453            relevant variables.
2454    
2455        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2456            with length and offset values. This means that the output is different
2457            for different internal link sizes. Test 2 is skipped for link sizes
2458            other than 2 because of this, bypassing the problem. Unfortunately,
2459            there was also a test in test 3 (the locale tests) that used /B and
2460            failed for link sizes other than 2. Rather than cut the whole test out,
2461            I have added a new /Z option to pcretest that replaces the length and
2462            offset values with spaces. This is now used to make test 3 independent
2463            of link size. (Test 2 will be tidied up later.)
2464    
2465    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2466        segmentation fault instead of returning the appropriate error message.
2467    
2468    15. In multiline mode when the newline sequence was set to "any", the pattern
2469        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2470        This doesn't seem right; it now treats the CRLF combination as the line
2471        ending, and so does not match in that case. It's only a pattern such as ^$
2472        that would hit this one: something like ^ABC$ would have failed after \r
2473        and then tried again after \r\n.
2474    
2475    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2476        in an attempt to make files that differ only in their line terminators
2477        compare equal. This works on Linux.
2478    
2479    17. Under certain error circumstances pcregrep might try to free random memory
2480        as it exited. This is now fixed, thanks to valgrind.
2481    
2482    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2483        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2484        was because its rules for how to advance for /g after matching an empty
2485        string at the end of a line did not allow for this case. They now check for
2486        it specially.
2487    
2488    20. pcretest is supposed to handle patterns and data of any length, by
2489        extending its buffers when necessary. It was getting this wrong when the
2490        buffer for a data line had to be extended.
2491    
2492    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2493        CRLF as a newline sequence.
2494    
2495    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2496        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2497        I have nevertheless tidied it up.
2498    
2499    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2500    
2501    24. Added a man page for pcre-config.
2502    
2503    
2504  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06
2505  ---------------------  ---------------------

Legend:
Removed from v.113  
changed lines
  Added in v.1253

  ViewVC Help
Powered by ViewVC 1.1.5