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

Diff of /code/trunk/ChangeLog

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

revision 141 by ph10, Fri Mar 30 15:46:27 2007 UTC revision 1300 by zherczeg, Mon Mar 25 12:34:26 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.1 12-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    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75    
76    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
77        gives some modest performance improvement in 8-bit mode.
78    
79    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
80        be expressed in certain programming languages using Universal Character
81        Names.
82    
83    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
84        which points out that "non characters" are not "characters that may not
85        appear in Unicode strings" but rather "characters that are reserved for
86        internal use and have only local meaning".
87    
88    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
89        there was a conditional group that depended on an assertion, if the
90        assertion was false, the callout that immediately followed the alternation
91        in the condition was skipped when pcre_exec() was used for matching.
92    
93    23. Allow an explicit callout to be inserted before an assertion that is the
94        condition for a conditional group, for compatibility with automatic
95        callouts, which always insert a callout at this point.
96    
97    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
98        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
99    
100    25. (*PRUNE) is now supported by the JIT compiler.
101    
102    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
103    
104    27. Fix the case where there are two or more SKIPs with arguments that may be
105        ignored.
106    
107    28. (*SKIP) is now supported by the JIT compiler.
108    
109    29. (*THEN) is now supported by the JIT compiler.
110    
111    30. Update RunTest with additional test selector options.
112    
113    31. The way PCRE handles backtracking verbs has been changed in to ways.
114    
115        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
116        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
117        backtracking. In some cases this makes it more Perl-compatible, but Perl's
118        rather obscure rules do not always do the same thing.
119    
120        (2) Previously, backtracking verbs were confined within assertions. This is
121        no longer the case, except for (*ACCEPT). Again, this sometimes improves
122        Perl compatibility, and sometimes does not.
123    
124    32. A number of tests that were in test 2 because Perl did things differently
125        have been moved to test 1, because either Perl or PCRE has changed, and
126        these tests are now compatible.
127    
128    32. Control verbs are handled in the same way in JIT and interpreter.
129    
130    
131    Version 8.32 30-November-2012
132    -----------------------------
133    
134    1.  Improved JIT compiler optimizations for first character search and single
135        character iterators.
136    
137    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
138        Patch by Daniel Richard G.
139    
140    3.  Single character iterator optimizations in the JIT compiler.
141    
142    4.  Improved JIT compiler optimizations for character ranges.
143    
144    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
145        Reported by Giuseppe D'Angelo.
146    
147    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
148        of a line, was being set incorrectly in cases where .* appeared inside
149        atomic brackets at the start of a pattern, or where there was a subsequent
150        *PRUNE or *SKIP.
151    
152    7.  Improved instruction cache flush for POWER/PowerPC.
153        Patch by Daniel Richard G.
154    
155    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
156        grep:
157    
158        (a) There is now no limit to the number of patterns to be matched.
159    
160        (b) An error is given if a pattern is too long.
161    
162        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
163            are now supported.
164    
165        (d) --exclude-from and --include-from (multiple use) have been added.
166    
167        (e) Exclusions and inclusions now apply to all files and directories, not
168            just to those obtained from scanning a directory recursively.
169    
170        (f) Multiple uses of -f and --file-list are now supported.
171    
172        (g) In a Windows environment, the default for -d has been changed from
173            "read" (the GNU grep default) to "skip", because otherwise the presence
174            of a directory in the file list provokes an error.
175    
176        (h) The documentation has been revised and clarified in places.
177    
178    9.  Improve the matching speed of capturing brackets.
179    
180    10. Changed the meaning of \X so that it now matches a Unicode extended
181        grapheme cluster.
182    
183    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
184        out POSIX threads when JIT support is configured.
185    
186    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
187    
188    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
189        structure could go wrong in environments where size_t is not the same size
190        as int.
191    
192    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
193        set.
194    
195    15. The EBCDIC support had decayed; later updates to the code had included
196        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
197        general tidy up of EBCDIC-related issues, and the documentation was also
198        not quite right. There is now a test that can be run on ASCII systems to
199        check some of the EBCDIC-related things (but is it not a full test).
200    
201    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
202        in a small tidy to the code.
203    
204    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
205    
206    18. If the --only-matching (-o) option in pcregrep is specified multiple
207        times, each one causes appropriate output. For example, -o1 -o2 outputs the
208        substrings matched by the 1st and 2nd capturing parentheses. A separating
209        string can be specified by --om-separator (default empty).
210    
211    19. Improving the first n character searches.
212    
213    20. Turn case lists for horizontal and vertical white space into macros so that
214        they are defined only once.
215    
216    21. This set of changes together give more compatible Unicode case-folding
217        behaviour for characters that have more than one other case when UCP
218        support is available.
219    
220        (a) The Unicode property table now has offsets into a new table of sets of
221            three or more characters that are case-equivalent. The MultiStage2.py
222            script that generates these tables (the pcre_ucd.c file) now scans
223            CaseFolding.txt instead of UnicodeData.txt for character case
224            information.
225    
226        (b) The code for adding characters or ranges of characters to a character
227            class has been abstracted into a generalized function that also handles
228            case-independence. In UTF-mode with UCP support, this uses the new data
229            to handle characters with more than one other case.
230    
231        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
232            whose other case is greater than 256 are now correctly matched
233            caselessly. Previously, the high codepoint matched the low one, but not
234            vice versa.
235    
236        (d) The processing of \h, \H, \v, and \ in character classes now makes use
237            of the new class addition function, using character lists defined as
238            macros alongside the case definitions of 20 above.
239    
240        (e) Caseless back references now work with characters that have more than
241            one other case.
242    
243        (f) General caseless matching of characters with more than one other case
244            is supported.
245    
246    22. Unicode character properties were updated from Unicode 6.2.0
247    
248    23. Improved CMake support under Windows. Patch by Daniel Richard G.
249    
250    24. Add support for 32-bit character strings, and UTF-32
251    
252    25. Major JIT compiler update (code refactoring and bugfixing).
253        Experimental Sparc 32 support is added.
254    
255    26. Applied a modified version of Daniel Richard G's patch to create
256        pcre.h.generic and config.h.generic by "make" instead of in the
257        PrepareRelease script.
258    
259    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
260        pcre_compile.c when checking for a zero character.
261    
262    28. Introducing a native interface for JIT. Through this interface, the compiled
263        machine code can be directly executed. The purpose of this interface is to
264        provide fast pattern matching, so several sanity checks are not performed.
265        However, feature tests are still performed. The new interface provides
266        1.4x speedup compared to the old one.
267    
268    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
269        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
270        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
271    
272    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
273        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
274        the "old" RFC 2279). Instead, it ended up passing a negative length to
275        pcre_exec().
276    
277    31. Add support for GCC's visibility feature to hide internal functions.
278    
279    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
280        "unknown -C option" after outputting 0 or 1.
281    
282    33. There is now support for generating a code coverage report for the test
283        suite in environments where gcc is the compiler and lcov is installed. This
284        is mainly for the benefit of the developers.
285    
286    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
287        unaddressable using valgrind annotations, allowing valgrind to detect
288        invalid memory accesses. This is mainly for the benefit of the developers.
289    
290    25. (*UTF) can now be used to start a pattern in any of the three libraries.
291    
292    26. Give configure error if --enable-cpp but no C++ compiler found.
293    
294    
295    Version 8.31 06-July-2012
296    -------------------------
297    
298    1.  Fixing a wrong JIT test case and some compiler warnings.
299    
300    2.  Removed a bashism from the RunTest script.
301    
302    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
303        to unsigned type, result still unsigned" that was given by an MS compiler
304        on encountering the code "-sizeof(xxx)".
305    
306    4.  Partial matching support is added to the JIT compiler.
307    
308    5.  Fixed several bugs concerned with partial matching of items that consist
309        of more than one character:
310    
311        (a) /^(..)\1/ did not partially match "aba" because checking references was
312            done on an "all or nothing" basis. This also applied to repeated
313            references.
314    
315        (b) \R did not give a hard partial match if \r was found at the end of the
316            subject.
317    
318        (c) \X did not give a hard partial match after matching one or more
319            characters at the end of the subject.
320    
321        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
322            a partial match for the string "\r".
323    
324        (e) When newline was set to CRLF, the metacharacter "." did not recognize
325            a partial match for a CR character at the end of the subject string.
326    
327    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
328        running pcretest, the text "(JIT)" added to the output whenever JIT is
329        actually used to run the match.
330    
331    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
332        or /S+[+] with a digit between 1 and 7.
333    
334    8.  OP_NOT now supports any UTF character not just single-byte ones.
335    
336    9.  (*MARK) control verb is now supported by the JIT compiler.
337    
338    10. The command "./RunTest list" lists the available tests without actually
339        running any of them. (Because I keep forgetting what they all are.)
340    
341    11. Add PCRE_INFO_MAXLOOKBEHIND.
342    
343    12. Applied a (slightly modified) user-supplied patch that improves performance
344        when the heap is used for recursion (compiled with --disable-stack-for-
345        recursion). Instead of malloc and free for each heap frame each time a
346        logical recursion happens, frames are retained on a chain and re-used where
347        possible. This sometimes gives as much as 30% improvement.
348    
349    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
350        call.
351    
352    14. As documented, (*COMMIT) is now confined to within a positive assertion.
353    
354    15. It is now possible to link pcretest with libedit as an alternative to
355        libreadline.
356    
357    16. (*COMMIT) control verb is now supported by the JIT compiler.
358    
359    17. The Unicode data tables have been updated to Unicode 6.1.0.
360    
361    18. Added --file-list option to pcregrep.
362    
363    19. Added binary file support to pcregrep, including the -a, --binary-files,
364        -I, and --text options.
365    
366    20. The madvise function is renamed for posix_madvise for QNX compatibility
367        reasons. Fixed by Giuseppe D'Angelo.
368    
369    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
370    
371    22. Changed the option for creating symbolic links for 16-bit man pages from
372        -s to -sf so that re-installing does not cause issues.
373    
374    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
375    
376    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
377        matches in certain environments (the workspace was not being correctly
378        retained). Also added to pcre_dfa_exec() a simple plausibility check on
379        some of the workspace data at the beginning of a restart.
380    
381    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
382        was not doing so when it should - probably a typo introduced by SVN 528
383        (change 8.10/14).
384    
385    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
386        \w+ when the character tables indicated that \x{c4} was a word character.
387        There were several related cases, all because the tests for doing a table
388        lookup were testing for characters less than 127 instead of 255.
389    
390    27. If a pattern contains capturing parentheses that are not used in a match,
391        their slots in the ovector are set to -1. For those that are higher than
392        any matched groups, this happens at the end of processing. In the case when
393        there were back references that the ovector was too small to contain
394        (causing temporary malloc'd memory to be used during matching), and the
395        highest capturing number was not used, memory off the end of the ovector
396        was incorrectly being set to -1. (It was using the size of the temporary
397        memory instead of the true size.)
398    
399    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
400        ovector size, it uses memory at the end of the block that it has got.
401    
402    29. Check for an overlong MARK name and give an error at compile time. The
403        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
404    
405    30. JIT compiler update.
406    
407    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
408        Rill for the patch.
409    
410    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
411    
412    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
413    
414    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
415        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
416        was enabled.
417    
418    35. Improve JIT code generation for greedy plus quantifier.
419    
420    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
421        1 to "aa" instead of to an empty string. The bug affected repeated groups
422        that could potentially match an empty string.
423    
424    37. Optimizing single character iterators in JIT.
425    
426    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
427        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
428        codepoints that are too big for the mode are faulted, and in a UTF mode,
429        disallowed codepoints are also faulted.
430    
431    39. If PCRE was compiled with UTF support, in three places in the DFA
432        matcher there was code that should only have been obeyed in UTF mode, but
433        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
434        processing when bytes with values greater than 127 were present. In 16-bit
435        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
436        both cases the values are those that cannot be the first data item in a UTF
437        character. The three items that might have provoked this were recursions,
438        possessively repeated groups, and atomic groups.
439    
440    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
441        and pcregrep, because some OS require shared objects to be explicitly
442        passed to ld, causing the link step to fail if they are not.
443    
444    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
445        mode, patterns that started with \h* or \R* might be incorrectly matched.
446    
447    
448    Version 8.30 04-February-2012
449    -----------------------------
450    
451    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
452        name is defined in ctype.h.
453    
454    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
455        only in quite long subpatterns.
456    
457    3.  Removed the function pcre_info(), which has been obsolete and deprecated
458        since it was replaced by pcre_fullinfo() in February 2000.
459    
460    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
461        match a (*MARK), and the match failed at the start of the subject, a
462        reference to memory before the start of the subject could occur. This bug
463        was introduced by fix 17 of release 8.21.
464    
465    5.  A reference to an unset group with zero minimum repetition was giving
466        totally wrong answers (in non-JavaScript-compatibility mode). For example,
467        /(another)?(\1?)test/ matched against "hello world test". This bug was
468        introduced in release 8.13.
469    
470    6.  Add support for 16-bit character strings (a large amount of work involving
471        many changes and refactorings).
472    
473    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
474        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
475        from a file.
476    
477    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
478        rounding is not applied in this particular case).
479    
480    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
481        if they appear, or are escaped, in patterns.
482    
483    10. Get rid of a number of -Wunused-but-set-variable warnings.
484    
485    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
486        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
487        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
488        also returns the mark "x". This bug applied to capturing parentheses,
489        non-capturing parentheses, and atomic parentheses. It also applied to some
490        assertions.
491    
492    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
493        information out of configure.ac instead of relying on pcre.h.generic, which
494        is not stored in the repository.
495    
496    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
497        -lreadline.
498    
499    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
500    
501    15. Applied Graycode's patch to put the top-level frame on the stack rather
502        than the heap when not using the stack for recursion. This gives a
503        performance improvement in many cases when recursion is not deep.
504    
505    16. Experimental code added to "pcretest -C" to output the stack frame size.
506    
507    
508    Version 8.21 12-Dec-2011
509    ------------------------
510    
511    1.  Updating the JIT compiler.
512    
513    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
514        are added as well.
515    
516    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
517        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
518        calling _pcre_jit_exec. Some extra comments are added.
519    
520    4.  (*MARK) settings inside atomic groups that do not contain any capturing
521        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
522        was introduced by change 18 for 8.20.
523    
524    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
525        ECMA-262 standard.
526    
527    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
528        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
529        This bug was probably introduced by change 9 of 8.13.
530    
531    7.  While fixing 6 above, I noticed that a number of other items were being
532        incorrectly rejected as "not fixed length". This arose partly because newer
533        opcodes had not been added to the fixed-length checking code. I have (a)
534        corrected the bug and added tests for these items, and (b) arranged for an
535        error to occur if an unknown opcode is encountered while checking for fixed
536        length instead of just assuming "not fixed length". The items that were
537        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
538        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
539        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
540    
541    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
542        being incorrectly compiled and would have given unpredicatble results.
543    
544    9.  A possessively repeated subpattern with minimum repeat count greater than
545        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
546        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
547        the first (A) could occur when it should not.
548    
549    10. Add a cast and remove a redundant test from the code.
550    
551    11. JIT should use pcre_malloc/pcre_free for allocation.
552    
553    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
554        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
555        is anything other than /usr, -L is still shown).
556    
557    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
558    
559    14. Perl does not support \N without a following name in a [] class; PCRE now
560        also gives an error.
561    
562    15. If a forward reference was repeated with an upper limit of around 2000,
563        it caused the error "internal error: overran compiling workspace". The
564        maximum number of forward references (including repeats) was limited by the
565        internal workspace, and dependent on the LINK_SIZE. The code has been
566        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
567        the default depends on LINK_SIZE. There is a new upper limit (for safety)
568        of around 200,000 forward references. While doing this, I also speeded up
569        the filling in of repeated forward references.
570    
571    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
572        incorrectly expecting the subject to contain another "a" after the start.
573    
574    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
575        in the match, the SKIP should be ignored. This was not happening; instead
576        the SKIP was being treated as NOMATCH. For patterns such as
577        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
578        tested.
579    
580    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
581        now much more compatible with Perl, in particular in cases where the result
582        is a non-match for a non-anchored pattern. For example, if
583        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
584        "m", where previously it did not return a name. A side effect of this
585        change is that for partial matches, the last encountered mark name is
586        returned, as for non matches. A number of tests that were previously not
587        Perl-compatible have been moved into the Perl-compatible test files. The
588        refactoring has had the pleasing side effect of removing one argument from
589        the match() function, thus reducing its stack requirements.
590    
591    19. If the /S+ option was used in pcretest to study a pattern using JIT,
592        subsequent uses of /S (without +) incorrectly behaved like /S+.
593    
594    21. Retrieve executable code size support for the JIT compiler and fixing
595        some warnings.
596    
597    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
598        not work when the shorter character appeared right at the end of the
599        subject string.
600    
601    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
602        systems.
603    
604    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
605        output it when the /M option is used in pcretest.
606    
607    25. The CheckMan script was not being included in the distribution. Also, added
608        an explicit "perl" to run Perl scripts from the PrepareRelease script
609        because this is reportedly needed in Windows.
610    
611    26. If study data was being save in a file and studying had not found a set of
612        "starts with" bytes for the pattern, the data written to the file (though
613        never used) was taken from uninitialized memory and so caused valgrind to
614        complain.
615    
616    27. Updated RunTest.bat as provided by Sheri Pierce.
617    
618    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
619    
620    29. Computation of memory usage for the table of capturing group names was
621        giving an unnecessarily large value.
622    
623    
624    Version 8.20 21-Oct-2011
625    ------------------------
626    
627    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
628        a POSIX class. After further experiments with Perl, which convinced me that
629        Perl has bugs and confusions, a closing square bracket is no longer allowed
630        in a POSIX name. This bug also affected patterns with classes that started
631        with full stops.
632    
633    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
634        captured substring, but while checking the failing first alternative,
635        substring 1 is temporarily captured. If the output vector supplied to
636        pcre_exec() was not big enough for this capture, the yield of the function
637        was still zero ("insufficient space for captured substrings"). This cannot
638        be totally fixed without adding another stack variable, which seems a lot
639        of expense for a edge case. However, I have improved the situation in cases
640        such as /(a)(b)x|abc/ matched against "abc", where the return code
641        indicates that fewer than the maximum number of slots in the ovector have
642        been set.
643    
644    3.  Related to (2) above: when there are more back references in a pattern than
645        slots in the output vector, pcre_exec() uses temporary memory during
646        matching, and copies in the captures as far as possible afterwards. It was
647        using the entire output vector, but this conflicts with the specification
648        that only 2/3 is used for passing back captured substrings. Now it uses
649        only the first 2/3, for compatibility. This is, of course, another edge
650        case.
651    
652    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
653        main code base, and can be used by building with --enable-jit. When this is
654        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
655        runtime --no-jit option is given.
656    
657    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
658        ovector, the return from the function was zero, implying that there were
659        other matches that did not fit. The correct "exactly full" value is now
660        returned.
661    
662    6.  If a subpattern that was called recursively or as a subroutine contained
663        (*PRUNE) or any other control that caused it to give a non-standard return,
664        invalid errors such as "Error -26 (nested recursion at the same subject
665        position)" or even infinite loops could occur.
666    
667    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
668        computing the minimum length on reaching *ACCEPT, and so ended up with the
669        wrong value of 1 rather than 0. Further investigation indicates that
670        computing a minimum subject length in the presence of *ACCEPT is difficult
671        (think back references, subroutine calls), and so I have changed the code
672        so that no minimum is registered for a pattern that contains *ACCEPT.
673    
674    8.  If (*THEN) was present in the first (true) branch of a conditional group,
675        it was not handled as intended. [But see 16 below.]
676    
677    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
678        Sheri Pierce.
679    
680    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
681        the first byte in a match must be "a".
682    
683    11. Change 17 for 8.13 increased the recursion depth for patterns like
684        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
685        pattern contains any instances of (*THEN). If it does not, the old
686        optimizations are restored. It would be nice to do this on a per-group
687        basis, but at the moment that is not feasible.
688    
689    12. In some environments, the output of pcretest -C is CRLF terminated. This
690        broke RunTest's code that checks for the link size. A single white space
691        character after the value is now allowed for.
692    
693    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
694        For "fr", it uses the Windows-specific input and output files.
695    
696    14. If (*THEN) appeared in a group that was called recursively or as a
697        subroutine, it did not work as intended. [But see next item.]
698    
699    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
700        pattern fragments (but not containing any | characters). If A and B are
701        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
702        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
703        D. In other words, Perl considers parentheses that do not contain any |
704        characters to be part of a surrounding alternative, whereas PCRE was
705        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
706        differently. PCRE now behaves in the same way as Perl, except in the case
707        of subroutine/recursion calls such as (?1) which have in any case always
708        been different (but PCRE had them first :-).
709    
710    16. Related to 15 above: Perl does not treat the | in a conditional group as
711        creating alternatives. Such a group is treated in the same way as an
712        ordinary group without any | characters when processing (*THEN). PCRE has
713        been changed to match Perl's behaviour.
714    
715    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
716        RunGrepTest script failed.
717    
718    18. Change 22 for version 13 caused atomic groups to use more stack. This is
719        inevitable for groups that contain captures, but it can lead to a lot of
720        stack use in large patterns. The old behaviour has been restored for atomic
721        groups that do not contain any capturing parentheses.
722    
723    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
724        suppress the check for a minimum subject length at run time. (If it was
725        given to pcre_exec() or pcre_dfa_exec() it did work.)
726    
727    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
728        fail to work when decoding hex characters in data strings in EBCDIC
729        environments.
730    
731    21. It appears that in at least one Mac OS environment, the isxdigit() function
732        is implemented as a macro that evaluates to its argument more than once,
733        contravening the C 90 Standard (I haven't checked a later standard). There
734        was an instance in pcretest which caused it to go wrong when processing
735        \x{...} escapes in subject strings. The has been rewritten to avoid using
736        things like p++ in the argument of isxdigit().
737    
738    
739    Version 8.13 16-Aug-2011
740    ------------------------
741    
742    1.  The Unicode data tables have been updated to Unicode 6.0.0.
743    
744    2.  Two minor typos in pcre_internal.h have been fixed.
745    
746    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
747        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
748        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
749    
750    4.  There were a number of related bugs in the code for matching backrefences
751        caselessly in UTF-8 mode when codes for the characters concerned were
752        different numbers of bytes. For example, U+023A and U+2C65 are an upper
753        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
754        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
755        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
756        2-byte code at the end of the subject (it thought there wasn't enough data
757        left).
758    
759    5.  Comprehensive information about what went wrong is now returned by
760        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
761        as the output vector has at least 2 elements. The offset of the start of
762        the failing character and a reason code are placed in the vector.
763    
764    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
765        now returned is for the first byte of the failing character, instead of the
766        last byte inspected. This is an incompatible change, but I hope it is small
767        enough not to be a problem. It makes the returned offset consistent with
768        pcre_exec() and pcre_dfa_exec().
769    
770    7.  pcretest now gives a text phrase as well as the error number when
771        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
772        failure, the offset and reason code are output.
773    
774    8.  When \R was used with a maximizing quantifier it failed to skip backwards
775        over a \r\n pair if the subsequent match failed. Instead, it just skipped
776        back over a single character (\n). This seems wrong (because it treated the
777        two characters as a single entity when going forwards), conflicts with the
778        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
779        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
780        has been changed.
781    
782    9.  Some internal refactoring has changed the processing so that the handling
783        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
784        time (the PCRE_DOTALL option was changed this way some time ago: version
785        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
786        which was always a bit of a fudge. It also means that there is one less
787        argument for the match() function, which reduces its stack requirements
788        slightly. This change also fixes an incompatibility with Perl: the pattern
789        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
790    
791    10. More internal refactoring has drastically reduced the number of recursive
792        calls to match() for possessively repeated groups such as (abc)++ when
793        using pcre_exec().
794    
795    11. While implementing 10, a number of bugs in the handling of groups were
796        discovered and fixed:
797    
798        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
799        (a|)*(?1) gave a compile-time internal error.
800        ((a|)+)+  did not notice that the outer group could match an empty string.
801        (^a|^)+   was not marked as anchored.
802        (.*a|.*)+ was not marked as matching at start or after a newline.
803    
804    12. Yet more internal refactoring has removed another argument from the match()
805        function. Special calls to this function are now indicated by setting a
806        value in a variable in the "match data" data block.
807    
808    13. Be more explicit in pcre_study() instead of relying on "default" for
809        opcodes that mean there is no starting character; this means that when new
810        ones are added and accidentally left out of pcre_study(), testing should
811        pick them up.
812    
813    14. The -s option of pcretest has been documented for ages as being an old
814        synonym of -m (show memory usage). I have changed it to mean "force study
815        for every regex", that is, assume /S for every regex. This is similar to -i
816        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
817        using it. It makes it easier to run collections of tests with and without
818        study enabled, and thereby test pcre_study() more easily. All the standard
819        tests are now run with and without -s (but some patterns can be marked as
820        "never study" - see 20 below).
821    
822    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
823        restoration of the capturing data to the outer values was not happening
824        correctly.
825    
826    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
827        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
828        pattern had matched an empty string, and so incorrectly returned a no
829        match.
830    
831    17. There was optimizing code for the last branch of non-capturing parentheses,
832        and also for the obeyed branch of a conditional subexpression, which used
833        tail recursion to cut down on stack usage. Unfortunately, now that there is
834        the possibility of (*THEN) occurring in these branches, tail recursion is
835        no longer possible because the return has to be checked for (*THEN). These
836        two optimizations have therefore been removed. [But see 8.20/11 above.]
837    
838    18. If a pattern containing \R was studied, it was assumed that \R always
839        matched two bytes, thus causing the minimum subject length to be
840        incorrectly computed because \R can also match just one byte.
841    
842    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
843        was incorrectly computed.
844    
845    20. If /S is present twice on a test pattern in pcretest input, it now
846        *disables* studying, thereby overriding the use of -s on the command line
847        (see 14 above). This is necessary for one or two tests to keep the output
848        identical in both cases.
849    
850    21. When (*ACCEPT) was used in an assertion that matched an empty string and
851        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
852    
853    22. When an atomic group that contained a capturing parenthesis was
854        successfully matched, but the branch in which it appeared failed, the
855        capturing was not being forgotten if a higher numbered group was later
856        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
857        group 1 to "a", when in fact it should be unset. This applied to multi-
858        branched capturing and non-capturing groups, repeated or not, and also to
859        positive assertions (capturing in negative assertions does not happen
860        in PCRE) and also to nested atomic groups.
861    
862    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
863        subject after a captured substring, to make it easier to tell which of a
864        number of identical substrings has been captured.
865    
866    24. The way atomic groups are processed by pcre_exec() has been changed so that
867        if they are repeated, backtracking one repetition now resets captured
868        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
869        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
870        "aaa". Previously, it would have been "a". As part of this code
871        refactoring, the way recursive calls are handled has also been changed.
872    
873    25. If an assertion condition captured any substrings, they were not passed
874        back unless some other capturing happened later. For example, if
875        (?(?=(a))a) was matched against "a", no capturing was returned.
876    
877    26. When studying a pattern that contained subroutine calls or assertions,
878        the code for finding the minimum length of a possible match was handling
879        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
880        group 1 called group 2 while simultaneously a separate group 2 called group
881        1). A stack overflow occurred in this case. I have fixed this by limiting
882        the recursion depth to 10.
883    
884    27. Updated RunTest.bat in the distribution to the version supplied by Tom
885        Fortmann. This supports explicit test numbers on the command line, and has
886        argument validation and error reporting.
887    
888    28. An instance of \X with an unlimited repeat could fail if at any point the
889        first character it looked at was a mark character.
890    
891    29. Some minor code refactoring concerning Unicode properties and scripts
892        should reduce the stack requirement of match() slightly.
893    
894    30. Added the '=' option to pcretest to check the setting of unused capturing
895        slots at the end of the pattern, which are documented as being -1, but are
896        not included in the return count.
897    
898    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
899        compiled something random. Now it gives a compile-time error (as does
900        Perl).
901    
902    32. A *MARK encountered during the processing of a positive assertion is now
903        recorded and passed back (compatible with Perl).
904    
905    33. If --only-matching or --colour was set on a pcregrep call whose pattern
906        had alternative anchored branches, the search for a second match in a line
907        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
908        matched the line "0102" twice. The same bug affected patterns that started
909        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
910        twice.
911    
912    34. Previously, PCRE did not allow quantification of assertions. However, Perl
913        does, and because of capturing effects, quantifying parenthesized
914        assertions may at times be useful. Quantifiers are now allowed for
915        parenthesized assertions.
916    
917    35. A minor code tidy in pcre_compile() when checking options for \R usage.
918    
919    36. \g was being checked for fancy things in a character class, when it should
920        just be a literal "g".
921    
922    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
923        appearance of a nested POSIX class supersedes an apparent external class.
924        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
925        unescaped square brackets may also appear as part of class names. For
926        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
927        more like Perl. (But see 8.20/1 above.)
928    
929    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
930        was because it thought it was \N{name}, which is not supported).
931    
932    39. Add minix to OS list not supporting the -S option in pcretest.
933    
934    40. PCRE tries to detect cases of infinite recursion at compile time, but it
935        cannot analyze patterns in sufficient detail to catch mutual recursions
936        such as ((?1))((?2)). There is now a runtime test that gives an error if a
937        subgroup is called recursively as a subpattern for a second time at the
938        same position in the subject string. In previous releases this might have
939        been caught by the recursion limit, or it might have run out of stack.
940    
941    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
942        happen only once. PCRE was, however incorrectly giving a compile time error
943        "recursive call could loop indefinitely" because it cannot analyze the
944        pattern in sufficient detail. The compile time test no longer happens when
945        PCRE is compiling a conditional subpattern, but actual runaway loops are
946        now caught at runtime (see 40 above).
947    
948    42. It seems that Perl allows any characters other than a closing parenthesis
949        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
950        has been changed to be the same.
951    
952    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
953        as not to get warnings when autogen.sh is called. Also changed
954        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
955    
956    44. To help people who use pcregrep to scan files containing exceedingly long
957        lines, the following changes have been made:
958    
959        (a) The default value of the buffer size parameter has been increased from
960            8K to 20K. (The actual buffer used is three times this size.)
961    
962        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
963            PCRE is built.
964    
965        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
966            to be set at run time.
967    
968        (d) Numerical values in pcregrep options can be followed by K or M, for
969            example --buffer-size=50K.
970    
971        (e) If a line being scanned overflows pcregrep's buffer, an error is now
972            given and the return code is set to 2.
973    
974    45. Add a pointer to the latest mark to the callout data block.
975    
976    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
977        partial match of an empty string instead of no match. This was specific to
978        the use of ".".
979    
980    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
981        complete match instead of a partial match. This bug was dependent on both
982        the PCRE_UTF8 and PCRE_DOTALL options being set.
983    
984    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
985        starting byte set, because \b was not being ignored.
986    
987    
988    Version 8.12 15-Jan-2011
989    ------------------------
990    
991    1.  Fixed some typos in the markup of the man pages, and wrote a script that
992        checks for such things as part of the documentation building process.
993    
994    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
995        --match-limit and --recursion-limit options (added for 8.11). In
996        particular, this made one of the standard tests fail. (The integer value
997        went into the wrong half of a long int.)
998    
999    3.  If the --colour option was given to pcregrep with -v (invert match), it
1000        did strange things, either producing crazy output, or crashing. It should,
1001        of course, ignore a request for colour when reporting lines that do not
1002        match.
1003    
1004    4.  Another pcregrep bug caused similar problems if --colour was specified with
1005        -M (multiline) and the pattern match finished with a line ending.
1006    
1007    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1008        matched in multiline mode, the following line was shown as part of the
1009        match. This seems wrong, so I have changed it.
1010    
1011    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1012        the check for further matches in the same line (so they could be coloured)
1013        to overrun the end of the current line. If another match was found, it was
1014        incorrectly shown (and then shown again when found in the next line).
1015    
1016    7.  If pcregrep was compiled under Windows, there was a reference to the
1017        function pcregrep_exit() before it was defined. I am assuming this was
1018        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1019        reported by a user. I've moved the definition above the reference.
1020    
1021    
1022    Version 8.11 10-Dec-2010
1023    ------------------------
1024    
1025    1.  (*THEN) was not working properly if there were untried alternatives prior
1026        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1027        backtracked to try for "b" instead of moving to the next alternative branch
1028        at the same level (in this case, to look for "c"). The Perl documentation
1029        is clear that when (*THEN) is backtracked onto, it goes to the "next
1030        alternative in the innermost enclosing group".
1031    
1032    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1033        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1034        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1035        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1036        (*THEN).
1037    
1038    3.  If \s appeared in a character class, it removed the VT character from
1039        the class, even if it had been included by some previous item, for example
1040        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1041        of \s, but is part of the POSIX "space" class.)
1042    
1043    4.  A partial match never returns an empty string (because you can always
1044        match an empty string at the end of the subject); however the checking for
1045        an empty string was starting at the "start of match" point. This has been
1046        changed to the "earliest inspected character" point, because the returned
1047        data for a partial match starts at this character. This means that, for
1048        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1049        (previously it gave "no match").
1050    
1051    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1052        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1053        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1054        has an implication that the given string is incomplete (because a partial
1055        match is preferred over a full match). For this reason, these items now
1056        give a partial match in this situation. [Aside: previously, the one case
1057        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1058        match rather than a full match, which was wrong by the old rules, but is
1059        now correct.]
1060    
1061    6.  There was a bug in the handling of #-introduced comments, recognized when
1062        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1063        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1064        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1065        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1066        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1067        places in pcre_compile().
1068    
1069    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1070        comments when looking ahead for named forward references to subpatterns,
1071        the only newline sequence it recognized was NL. It now handles newlines
1072        according to the set newline convention.
1073    
1074    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1075        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1076        cater for a lack of strerror(). These oversights have been fixed.
1077    
1078    9.  Added --match-limit and --recursion-limit to pcregrep.
1079    
1080    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1081    
1082    11. When the -o option was used, pcregrep was setting a return code of 1, even
1083        when matches were found, and --line-buffered was not being honoured.
1084    
1085    12. Added an optional parentheses number to the -o and --only-matching options
1086        of pcregrep.
1087    
1088    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1089        can match an empty string. The code to do it in pcretest and pcredemo
1090        needed fixing:
1091    
1092        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1093            only one byte after an empty string match just before CRLF (this case
1094            just got forgotten; "any" and "anycrlf" were OK).
1095    
1096        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1097            mode when an empty string match preceded an ASCII character followed by
1098            a non-ASCII character. (The code for advancing by one character rather
1099            than one byte was nonsense.)
1100    
1101        (c) The pcredemo.c sample program did not have any code at all to handle
1102            the cases when CRLF is a valid newline sequence.
1103    
1104    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1105        as a starting offset was within the subject string. There is now a new
1106        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1107        negative or greater than the length of the string. In order to test this,
1108        pcretest is extended to allow the setting of negative starting offsets.
1109    
1110    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1111        starting offset points to the beginning of a UTF-8 character was
1112        unnecessarily clumsy. I tidied it up.
1113    
1114    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1115        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1116    
1117    17. Nobody had reported that the --include_dir option, which was added in
1118        release 7.7 should have been called --include-dir (hyphen, not underscore)
1119        for compatibility with GNU grep. I have changed it to --include-dir, but
1120        left --include_dir as an undocumented synonym, and the same for
1121        --exclude-dir, though that is not available in GNU grep, at least as of
1122        release 2.5.4.
1123    
1124    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1125        characters from a string of bytes) have been redefined so as not to use
1126        loops, in order to improve performance in some environments. At the same
1127        time, I abstracted some of the common code into auxiliary macros to save
1128        repetition (this should not affect the compiled code).
1129    
1130    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1131        compile-time error is now given if \c is not followed by an ASCII
1132        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1133        different, and any byte value is allowed.)
1134    
1135    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1136        START_OPTIMIZE option, which is now allowed at compile time - but just
1137        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1138        to pcregrep and other applications that have no direct access to PCRE
1139        options. The new /Y option in pcretest sets this option when calling
1140        pcre_compile().
1141    
1142    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1143        back references. Groups containing recursive back references were forced to
1144        be atomic by that change, but in the case of named groups, the amount of
1145        memory required was incorrectly computed, leading to "Failed: internal
1146        error: code overflow". This has been fixed.
1147    
1148    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1149        pcretest.c, to avoid build problems in some Borland environments.
1150    
1151    
1152    Version 8.10 25-Jun-2010
1153    ------------------------
1154    
1155    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1156        THEN.
1157    
1158    2.  (*ACCEPT) was not working when inside an atomic group.
1159    
1160    3.  Inside a character class, \B is treated as a literal by default, but
1161        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1162        causes the error). The code is unchanged, but I tidied the documentation.
1163    
1164    4.  Inside a character class, PCRE always treated \R and \X as literals,
1165        whereas Perl faults them if its -w option is set. I have changed PCRE so
1166        that it faults them when PCRE_EXTRA is set.
1167    
1168    5.  Added support for \N, which always matches any character other than
1169        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1170    
1171    6.  When compiling pcregrep with newer versions of gcc which may have
1172        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1173        declared with attribute warn_unused_result" were given. Just casting the
1174        result to (void) does not stop the warnings; a more elaborate fudge is
1175        needed. I've used a macro to implement this.
1176    
1177    7.  Minor change to pcretest.c to avoid a compiler warning.
1178    
1179    8.  Added four artifical Unicode properties to help with an option to make
1180        \s etc use properties (see next item). The new properties are: Xan
1181        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1182    
1183    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1184        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1185        this option. Modified pcretest to add /W to test this facility. Added
1186        REG_UCP to make it available via the POSIX interface.
1187    
1188    10. Added --line-buffered to pcregrep.
1189    
1190    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1191        studied, and the match started with a letter with a code point greater than
1192        127 whose first byte was different to the first byte of the other case of
1193        the letter, the other case of this starting letter was not recognized
1194        (#976).
1195    
1196    12. If a pattern that was studied started with a repeated Unicode property
1197        test, for example, \p{Nd}+, there was the theoretical possibility of
1198        setting up an incorrect bitmap of starting bytes, but fortunately it could
1199        not have actually happened in practice until change 8 above was made (it
1200        added property types that matched character-matching opcodes).
1201    
1202    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1203        possible starting bytes for non-anchored patterns.
1204    
1205    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1206        \R, and also a number of cases that involve Unicode properties, both
1207        explicit and implicit when PCRE_UCP is set.
1208    
1209    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1210        input, it could crash or give wrong results if characters with values
1211        greater than 0xc0 were present in the subject string. (Detail: it assumed
1212        UTF-8 input when processing these items.)
1213    
1214    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1215        size_t is 64-bit (#991).
1216    
1217    17. Added a check for running out of memory when PCRE is compiled with
1218        --disable-stack-for-recursion (#990).
1219    
1220    18. If the last data line in a file for pcretest does not have a newline on
1221        the end, a newline was missing in the output.
1222    
1223    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1224        less than 128) in its various bitmaps. However, there is a facility for
1225        generating tables according to the current locale when PCRE is compiled. It
1226        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1227        characters, are recognized by isspace() and therefore were getting set in
1228        these tables, and indeed these tables seem to approximate to ISO 8859. This
1229        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1230        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1231        which of course cannot start UTF-8 characters. I have changed the code so
1232        that only real ASCII characters (less than 128) and the correct starting
1233        bytes for UTF-8 encodings are set for characters greater than 127 when in
1234        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1235        altogether.)
1236    
1237    20. Added the /T option to pcretest so as to be able to run tests with non-
1238        standard character tables, thus making it possible to include the tests
1239        used for 19 above in the standard set of tests.
1240    
1241    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1242        reference to a subpattern the other side of a comment that contains an
1243        opening parenthesis caused either an internal compiling error, or a
1244        reference to the wrong subpattern.
1245    
1246    
1247    Version 8.02 19-Mar-2010
1248    ------------------------
1249    
1250    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1251    
1252    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1253        configured.
1254    
1255    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1256        original author of that file, following a query about its status.
1257    
1258    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1259        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1260    
1261    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1262        quantifier applied to a forward-referencing subroutine call, could compile
1263        incorrect code or give the error "internal error: previously-checked
1264        referenced subpattern not found".
1265    
1266    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1267        variables to point to external functions. For these systems, therefore,
1268        pcre_malloc etc. are now initialized to local functions that call the
1269        relevant global functions.
1270    
1271    7.  There were two entries missing in the vectors called coptable and poptable
1272        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1273        I've fixed the data, and added a kludgy way of testing at compile time that
1274        the lengths are correct (equal to the number of opcodes).
1275    
1276    8.  Following on from 7, I added a similar kludge to check the length of the
1277        eint vector in pcreposix.c.
1278    
1279    9.  Error texts for pcre_compile() are held as one long string to avoid too
1280        much relocation at load time. To find a text, the string is searched,
1281        counting zeros. There was no check for running off the end of the string,
1282        which could happen if a new error number was added without updating the
1283        string.
1284    
1285    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1286    
1287    11. \K was not working if it appeared in an atomic group or in a group that
1288        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1289        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1290        the assertion is positive, but not if it is negative.
1291    
1292    12. Change 11 fortuitously reduced the size of the stack frame used in the
1293        "match()" function of pcre_exec.c by one pointer. Forthcoming
1294        implementation of support for (*MARK) will need an extra pointer on the
1295        stack; I have reserved it now, so that the stack frame size does not
1296        decrease.
1297    
1298    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1299        item in branch that calls a recursion is a subroutine call - as in the
1300        second branch in the above example - was incorrectly given the compile-
1301        time error "recursive call could loop indefinitely" because pcre_compile()
1302        was not correctly checking the subroutine for matching a non-empty string.
1303    
1304    14. The checks for overrunning compiling workspace could trigger after an
1305        overrun had occurred. This is a "should never occur" error, but it can be
1306        triggered by pathological patterns such as hundreds of nested parentheses.
1307        The checks now trigger 100 bytes before the end of the workspace.
1308    
1309    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1310    
1311    
1312    Version 8.01 19-Jan-2010
1313    ------------------------
1314    
1315    1.  If a pattern contained a conditional subpattern with only one branch (in
1316        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1317        computed the wrong minimum data length (which is of course zero for such
1318        subpatterns). This could cause incorrect "no match" results.
1319    
1320    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1321        the pattern is reset in the first branch, pcre_compile() failed with
1322        "internal error: code overflow at offset...". This happened only when
1323        the reset was to the original external option setting. (An optimization
1324        abstracts leading options settings into an external setting, which was the
1325        cause of this.)
1326    
1327    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1328        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1329        assertion pattern did not match (meaning that the assertion was true), it
1330        was incorrectly treated as false if the SKIP had been reached during the
1331        matching. This also applied to assertions used as conditions.
1332    
1333    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1334        assertion subpattern, including such a pattern used as a condition,
1335        unpredictable results occurred, instead of the error return
1336        PCRE_ERROR_DFA_UITEM.
1337    
1338    5.  The C++ GlobalReplace function was not working like Perl for the special
1339        situation when an empty string is matched. It now does the fancy magic
1340        stuff that is necessary.
1341    
1342    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1343        removed. (These were left over from very, very early versions of PCRE.)
1344    
1345    7.  Some cosmetic changes to the code to make life easier when compiling it
1346        as part of something else:
1347    
1348        (a) Change DEBUG to PCRE_DEBUG.
1349    
1350        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1351            called "current" as "current_branch", to prevent a collision with the
1352            Linux macro when compiled as a kernel module.
1353    
1354        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1355            prevent a collision with the Linux macro when compiled as a kernel
1356            module.
1357    
1358    8.  In pcre_compile() there are some checks for integer overflows that used to
1359        cast potentially large values to (double). This has been changed to that
1360        when building, a check for int64_t is made, and if it is found, it is used
1361        instead, thus avoiding the use of floating point arithmetic. (There is no
1362        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1363        double.
1364    
1365    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1366        2005 (difference between two addresses compared to an unsigned value).
1367    
1368    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1369        custom one, because of the following reported problem in Windows:
1370    
1371          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1372              under Win32.
1373          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1374              therefore missing the function definition.
1375          - The compiler thus generates a "C" signature for the test function.
1376          - The linker fails to find the "C" function.
1377          - PCRE fails to configure if asked to do so against libbz2.
1378    
1379    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1380        messages were output:
1381    
1382          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1383          rerunning libtoolize, to keep the correct libtool macros in-tree.
1384          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1385    
1386        I have done both of these things.
1387    
1388    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1389        most of the time, it *can* run out if it is given a pattern that contains a
1390        runaway infinite recursion. I updated the discussion in the pcrestack man
1391        page.
1392    
1393    13. Now that we have gone to the x.xx style of version numbers, the minor
1394        version may start with zero. Using 08 or 09 is a bad idea because users
1395        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1396        interpreted as invalid octal numbers. I've updated the previous comment in
1397        configure.ac, and also added a check that gives an error if 08 or 09 are
1398        used.
1399    
1400    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1401        causing partial matching to fail when the end of the subject matched \W
1402        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1403    
1404    15. There were some discrepancies between the declarations in pcre_internal.h
1405        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1406        their definitions. The declarations used "const uschar *" and the
1407        definitions used USPTR. Even though USPTR is normally defined as "const
1408        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1409        reported that: "This difference in casting confuses some C++ compilers, for
1410        example, SunCC recognizes above declarations as different functions and
1411        generates broken code for hbpcre." I have changed the declarations to use
1412        USPTR.
1413    
1414    16. GNU libtool is named differently on some systems. The autogen.sh script now
1415        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1416        (FreeBSD).
1417    
1418    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1419        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1420        comment: "Figure out how to create a longlong from a string: strtoll and
1421        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1422        instance, but it only takes 2 args instead of 3!"
1423    
1424    18. A subtle bug concerned with back references has been fixed by a change of
1425        specification, with a corresponding code fix. A pattern such as
1426        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1427        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1428        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1429        same bug. Such groups have to be quantified to be useful, or contained
1430        inside another quantified group. (If there's no repetition, the reference
1431        can never match.) The problem arises because, having left the group and
1432        moved on to the rest of the pattern, a later failure that backtracks into
1433        the group uses the captured value from the final iteration of the group
1434        rather than the correct earlier one. I have fixed this in PCRE by forcing
1435        any group that contains a reference to itself to be an atomic group; that
1436        is, there cannot be any backtracking into it once it has completed. This is
1437        similar to recursive and subroutine calls.
1438    
1439    
1440    Version 8.00 19-Oct-09
1441    ----------------------
1442    
1443    1.  The table for translating pcre_compile() error codes into POSIX error codes
1444        was out-of-date, and there was no check on the pcre_compile() error code
1445        being within the table. This could lead to an OK return being given in
1446        error.
1447    
1448    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1449        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1450        in a Windows environment.
1451    
1452    3.  The pcregrep --count option prints the count for each file even when it is
1453        zero, as does GNU grep. However, pcregrep was also printing all files when
1454        --files-with-matches was added. Now, when both options are given, it prints
1455        counts only for those files that have at least one match. (GNU grep just
1456        prints the file name in this circumstance, but including the count seems
1457        more useful - otherwise, why use --count?) Also ensured that the
1458        combination -clh just lists non-zero counts, with no names.
1459    
1460    4.  The long form of the pcregrep -F option was incorrectly implemented as
1461        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1462        but it seems right to fix it, and I didn't think it was worth preserving
1463        the old behaviour.
1464    
1465    5.  The command line items --regex=pattern and --regexp=pattern were not
1466        recognized by pcregrep, which required --regex pattern or --regexp pattern
1467        (with a space rather than an '='). The man page documented the '=' forms,
1468        which are compatible with GNU grep; these now work.
1469    
1470    6.  No libpcreposix.pc file was created for pkg-config; there was just
1471        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1472    
1473    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1474        when UCP support is not needed, by modifying the Python script that
1475        generates it from Unicode data files. This should not matter if the module
1476        is correctly used as a library, but I received one complaint about 50K of
1477        unwanted data. My guess is that the person linked everything into his
1478        program rather than using a library. Anyway, it does no harm.
1479    
1480    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1481        was a minimum greater than 1 for a wide character in a possessive
1482        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1483        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1484        character. Chaos in the form of incorrect output or a compiling loop could
1485        result.
1486    
1487    9.  The restrictions on what a pattern can contain when partial matching is
1488        requested for pcre_exec() have been removed. All patterns can now be
1489        partially matched by this function. In addition, if there are at least two
1490        slots in the offset vector, the offset of the earliest inspected character
1491        for the match and the offset of the end of the subject are set in them when
1492        PCRE_ERROR_PARTIAL is returned.
1493    
1494    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1495        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1496        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1497        and may be more useful for multi-segment matching.
1498    
1499    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1500        used to be given if ever the end of the subject was reached; now it is
1501        given only if matching could not proceed because another character was
1502        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1503        string "Z", which now yields "no match" instead of "partial match". In the
1504        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1505        final character ended with (*FAIL).
1506    
1507    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1508        if the pattern had a "must contain" character that was already found in the
1509        earlier partial match, unless partial matching was again requested. For
1510        example, with the pattern /dog.(body)?/, the "must contain" character is
1511        "g". If the first part-match was for the string "dog", restarting with
1512        "sbody" failed. This bug has been fixed.
1513    
1514    13. The string returned by pcre_dfa_exec() after a partial match has been
1515        changed so that it starts at the first inspected character rather than the
1516        first character of the match. This makes a difference only if the pattern
1517        starts with a lookbehind assertion or \b or \B (\K is not supported by
1518        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1519        matching functions compatible, and I think it's the right thing to do.
1520    
1521    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1522        so that the demonstration program is easily available in environments where
1523        PCRE has not been installed from source.
1524    
1525    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1526        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1527        library.
1528    
1529    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1530        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1531        is not the first non-POSIX option to be added. Clearly some people find
1532        these options useful.
1533    
1534    17. If a caller to the POSIX matching function regexec() passes a non-zero
1535        value for nmatch with a NULL value for pmatch, the value of
1536        nmatch is forced to zero.
1537    
1538    18. RunGrepTest did not have a test for the availability of the -u option of
1539        the diff command, as RunTest does. It now checks in the same way as
1540        RunTest, and also checks for the -b option.
1541    
1542    19. If an odd number of negated classes containing just a single character
1543        interposed, within parentheses, between a forward reference to a named
1544        subpattern and the definition of the subpattern, compilation crashed with
1545        an internal error, complaining that it could not find the referenced
1546        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1547        [The bug was that it was starting one character too far in when skipping
1548        over the character class, thus treating the ] as data rather than
1549        terminating the class. This meant it could skip too much.]
1550    
1551    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1552        /g option in pcretest when the pattern contains \K, which makes it possible
1553        to have an empty string match not at the start, even when the pattern is
1554        anchored. Updated pcretest and pcredemo to use this option.
1555    
1556    21. If the maximum number of capturing subpatterns in a recursion was greater
1557        than the maximum at the outer level, the higher number was returned, but
1558        with unset values at the outer level. The correct (outer level) value is
1559        now given.
1560    
1561    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1562        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1563        make it do so. The string so far is captured, making this feature
1564        compatible with Perl.
1565    
1566    23. The tests have been re-organized, adding tests 11 and 12, to make it
1567        possible to check the Perl 5.10 features against Perl 5.10.
1568    
1569    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1570        pattern matches a fixed length string. PCRE did not allow this; now it
1571        does. Neither allows recursion.
1572    
1573    25. I finally figured out how to implement a request to provide the minimum
1574        length of subject string that was needed in order to match a given pattern.
1575        (It was back references and recursion that I had previously got hung up
1576        on.) This code has now been added to pcre_study(); it finds a lower bound
1577        to the length of subject needed. It is not necessarily the greatest lower
1578        bound, but using it to avoid searching strings that are too short does give
1579        some useful speed-ups. The value is available to calling programs via
1580        pcre_fullinfo().
1581    
1582    26. While implementing 25, I discovered to my embarrassment that pcretest had
1583        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1584        study optimizations had never been tested with that matching function.
1585        Oops. What is worse, even when it was passed study data, there was a bug in
1586        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1587        were also very few tests of studied patterns with pcre_dfa_exec().
1588    
1589    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1590        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1591        on the other side of the coin, they are no longer allowed to have different
1592        names, because these cannot be distinguished in PCRE, and this has caused
1593        confusion. (This is a difference from Perl.)
1594    
1595    28. When duplicate subpattern names are present (necessarily with different
1596        numbers, as required by 27 above), and a test is made by name in a
1597        conditional pattern, either for a subpattern having been matched, or for
1598        recursion in such a pattern, all the associated numbered subpatterns are
1599        tested, and the overall condition is true if the condition is true for any
1600        one of them. This is the way Perl works, and is also more like the way
1601        testing by number works.
1602    
1603    
1604    Version 7.9 11-Apr-09
1605    ---------------------
1606    
1607    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1608        (pcretest), all targets were linked against these libraries. This included
1609        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1610        libraries. This caused unwanted dependencies to be created. This problem
1611        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1612        pcretest is linked with readline.
1613    
1614    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1615        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1616        moved outside it again, because FALSE and TRUE are already defined in AIX,
1617        but BOOL is not.
1618    
1619    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1620        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1621    
1622    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1623        hyphens) following file names and line numbers when outputting matching
1624        lines. This is not true; no spaces are inserted. I have also clarified the
1625        wording for the --colour (or --color) option.
1626    
1627    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1628        was not coloured; this is different to GNU grep, so I have changed it to be
1629        the same.
1630    
1631    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1632        each matching line was coloured. Now it goes on to look for further matches
1633        of any of the test patterns, which is the same behaviour as GNU grep.
1634    
1635    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1636        doesn't make sense to accept an empty string match in pcregrep, so I have
1637        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1638        seems to be how GNU grep behaves.
1639    
1640    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1641        start or after a newline", because the conditional assertion was not being
1642        correctly handled. The rule now is that both the assertion and what follows
1643        in the first alternative must satisfy the test.
1644    
1645    9.  If auto-callout was enabled in a pattern with a conditional group whose
1646        condition was an assertion, PCRE could crash during matching, both with
1647        pcre_exec() and pcre_dfa_exec().
1648    
1649    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1650        used for matching.
1651    
1652    11. Unicode property support in character classes was not working for
1653        characters (bytes) greater than 127 when not in UTF-8 mode.
1654    
1655    12. Added the -M command line option to pcretest.
1656    
1657    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1658    
1659    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1660    
1661    16. Added comments and documentation about mis-use of no_arg in the C++
1662        wrapper.
1663    
1664    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1665        from Martin Jerabek that uses macro names for all relevant character and
1666        string constants.
1667    
1668    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1669        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1670        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1671        these, but not everybody uses configure.
1672    
1673    19. A conditional group that had only one branch was not being correctly
1674        recognized as an item that could match an empty string. This meant that an
1675        enclosing group might also not be so recognized, causing infinite looping
1676        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1677        with the subject "ab", where knowledge that the repeated group can match
1678        nothing is needed in order to break the loop.
1679    
1680    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1681        exec(), but without supplying a callout function, matching went wrong.
1682    
1683    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1684        leak if the size of the offset vector was greater than 30. When the vector
1685        is smaller, the saved offsets during recursion go onto a local stack
1686        vector, but for larger vectors malloc() is used. It was failing to free
1687        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1688        error, in fact).
1689    
1690    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1691        heapframe that is used only when UTF-8 support is enabled. This caused no
1692        problem, but was untidy.
1693    
1694    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1695        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1696        included within another project.
1697    
1698    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1699        slightly modified by me:
1700    
1701          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1702              not building pcregrep.
1703    
1704          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1705              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1706    
1707    25. Forward references, both numeric and by name, in patterns that made use of
1708        duplicate group numbers, could behave incorrectly or give incorrect errors,
1709        because when scanning forward to find the reference group, PCRE was not
1710        taking into account the duplicate group numbers. A pattern such as
1711        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1712    
1713    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1714        the feature of a custom pointer more persuasive (as requested by a user).
1715    
1716    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1717        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1718        pre-defined.
1719    
1720    28. Added support for (*UTF8) at the start of a pattern.
1721    
1722    29. Arrange for flags added by the "release type" setting in CMake to be shown
1723        in the configuration summary.
1724    
1725    
1726    Version 7.8 05-Sep-08
1727    ---------------------
1728    
1729    1.  Replaced UCP searching code with optimized version as implemented for Ad
1730        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1731        stage table and inline lookup instead of a function, giving speed ups of 2
1732        to 5 times on some simple patterns that I tested. Permission was given to
1733        distribute the MultiStage2.py script that generates the tables (it's not in
1734        the tarball, but is in the Subversion repository).
1735    
1736    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1737        scripts.
1738    
1739    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1740        a group with a zero qualifier. The result of the study could be incorrect,
1741        or the function might crash, depending on the pattern.
1742    
1743    4.  Caseless matching was not working for non-ASCII characters in back
1744        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1745        It now works when Unicode Property Support is available.
1746    
1747    5.  In pcretest, an escape such as \x{de} in the data was always generating
1748        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1749        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1750        truncation.
1751    
1752    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1753    
1754    7.  Added two (int) casts to pcregrep when printing the difference of two
1755        pointers, in case they are 64-bit values.
1756    
1757    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1758        test 2 if it fails.
1759    
1760    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1761        and a #define of that name to empty if it is not externally set. This is to
1762        allow users of MSVC to set it if necessary.
1763    
1764    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1765        the convenience functions in the pcre_get.c source file.
1766    
1767    11. An option change at the start of a pattern that had top-level alternatives
1768        could cause overwriting and/or a crash. This command provoked a crash in
1769        some environments:
1770    
1771          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1772    
1773        This potential security problem was recorded as CVE-2008-2371.
1774    
1775    12. For a pattern where the match had to start at the beginning or immediately
1776        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1777        pcre_dfa_exec() could read past the end of the passed subject if there was
1778        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1779        pcretest so that it places the subject at the end of its malloc-ed buffer.
1780    
1781    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1782        exec() might read past the end of the data buffer in UTF-8 mode.
1783    
1784    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1785        the data contained the byte 0x85 as part of a UTF-8 character within its
1786        first line. This applied both to normal and DFA matching.
1787    
1788    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1789        /^[^d]*?$/8 failed to match "abc".
1790    
1791    16. Added a missing copyright notice to pcrecpp_internal.h.
1792    
1793    17. Make it more clear in the documentation that values returned from
1794        pcre_exec() in ovector are byte offsets, not character counts.
1795    
1796    18. Tidied a few places to stop certain compilers from issuing warnings.
1797    
1798    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1799        supplied by Stefan Weber. I made a further small update for 7.8 because
1800        there is a change of source arrangements: the pcre_searchfuncs.c module is
1801        replaced by pcre_ucd.c.
1802    
1803    
1804    Version 7.7 07-May-08
1805    ---------------------
1806    
1807    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1808        a string to a long long, pretend we don't even have a long long." This is
1809        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1810    
1811    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1812        pre-7.6 versions, which defined a global no_arg variable instead of putting
1813        it in the RE class. (See also #8 below.)
1814    
1815    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1816        Lopes.
1817    
1818    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1819    
1820        (1) The include/exclude patterns were being applied to the whole pathnames
1821            of files, instead of just to the final components.
1822    
1823        (2) If there was more than one level of directory, the subdirectories were
1824            skipped unless they satisfied the include/exclude conditions. This is
1825            inconsistent with GNU grep (and could even be seen as contrary to the
1826            pcregrep specification - which I improved to make it absolutely clear).
1827            The action now is always to scan all levels of directory, and just
1828            apply the include/exclude patterns to regular files.
1829    
1830    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1831        --exclude_dir in the tests to avoid scanning .svn directories.
1832    
1833    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1834        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1835        doesn't support NULs in patterns.
1836    
1837    7.  Added some missing "const"s to declarations of static tables in
1838        pcre_compile.c and pcre_dfa_exec.c.
1839    
1840    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1841        caused by fix #2  above. (Subsequently also a second patch to fix the
1842        first patch. And a third patch - this was a messy problem.)
1843    
1844    9.  Applied Craig's patch to remove the use of push_back().
1845    
1846    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1847        matching function regexec().
1848    
1849    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1850        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1851        references. PCRE supports relative numbers with this syntax (I don't think
1852        Oniguruma does).
1853    
1854    12. Previously, a group with a zero repeat such as (...){0} was completely
1855        omitted from the compiled regex. However, this means that if the group
1856        was called as a subroutine from elsewhere in the pattern, things went wrong
1857        (an internal error was given). Such groups are now left in the compiled
1858        pattern, with a new opcode that causes them to be skipped at execution
1859        time.
1860    
1861    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1862        to the way PCRE behaves:
1863    
1864        (a) A lone ] character is dis-allowed (Perl treats it as data).
1865    
1866        (b) A back reference to an unmatched subpattern matches an empty string
1867            (Perl fails the current match path).
1868    
1869        (c) A data ] in a character class must be notated as \] because if the
1870            first data character in a class is ], it defines an empty class. (In
1871            Perl it is not possible to have an empty class.) The empty class []
1872            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1873            The negative empty class [^] matches any one character, independently
1874            of the DOTALL setting.
1875    
1876    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1877        non-existent subpattern following a character class starting with ']' and
1878        containing () gave an internal compiling error instead of "reference to
1879        non-existent subpattern". Fortunately, when the pattern did exist, the
1880        compiled code was correct. (When scanning forwards to check for the
1881        existencd of the subpattern, it was treating the data ']' as terminating
1882        the class, so got the count wrong. When actually compiling, the reference
1883        was subsequently set up correctly.)
1884    
1885    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1886        it was being rejected as not supported by pcre_dfa_exec(), even though
1887        other assertions are supported. I have made pcre_dfa_exec() support
1888        (*FAIL).
1889    
1890    16. The implementation of 13c above involved the invention of a new opcode,
1891        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1892        cannot be changed at match time, I realized I could make a small
1893        improvement to matching performance by compiling OP_ALLANY instead of
1894        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1895        on the OP_ANY path.
1896    
1897    17. Compiling pcretest on Windows with readline support failed without the
1898        following two fixes: (1) Make the unistd.h include conditional on
1899        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1900    
1901    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1902        ncurses library to be included for pcretest when ReadLine support is
1903        requested, but also to allow for it to be overridden. This patch came from
1904        Daniel Bergström.
1905    
1906    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1907        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1908        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1909        spotting this.
1910    
1911    
1912    Version 7.6 28-Jan-08
1913    ---------------------
1914    
1915    1.  A character class containing a very large number of characters with
1916        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1917        overflow.
1918    
1919    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1920        HAVE_LONG_LONG is not defined.
1921    
1922    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1923        bring them up to date and include new features. This patch includes:
1924    
1925        - Fixed PH's badly added libz and libbz2 support.
1926        - Fixed a problem with static linking.
1927        - Added pcredemo. [But later removed - see 7 below.]
1928        - Fixed dftables problem and added an option.
1929        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1930            HAVE_LONG_LONG.
1931        - Added readline support for pcretest.
1932        - Added an listing of the option settings after cmake has run.
1933    
1934    4.  A user submitted a patch to Makefile that makes it easy to create
1935        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1936        Makefile.am that cause it to include this special target, without
1937        affecting anything else. Note that the same mingw target plus all
1938        the other distribution libraries and programs are now supported
1939        when configuring with CMake (see 6 below) instead of with
1940        Configure/Make.
1941    
1942    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1943        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1944        exported in the Windows port". It has not yet been confirmed that the patch
1945        solves the problem, but it does no harm.
1946    
1947    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1948        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1949        with CMake, and also correct the comment about stack recursion.
1950    
1951    7.  Remove the automatic building of pcredemo from the ./configure system and
1952        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1953        of a program that users should build themselves after PCRE is installed, so
1954        building it automatically is not really right. What is more, it gave
1955        trouble in some build environments.
1956    
1957    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1958    
1959    
1960    Version 7.5 10-Jan-08
1961    ---------------------
1962    
1963    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1964        values in parens when parsing an RE using the C++ wrapper."
1965    
1966    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1967        Characters greater than 255 were excluded from the class instead of being
1968        included.
1969    
1970    3.  The same bug as (2) above applied to negated POSIX classes such as
1971        [:^space:].
1972    
1973    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1974        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1975        I have changed it.
1976    
1977    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1978        first named subpattern) and a construct such as (?&a) would reference the
1979        first named subpattern whose name started with "a" (in other words, the
1980        length check was missing). Both these problems are fixed. "Subpattern name
1981        expected" is now given for (?&) (a zero-length name), and this patch also
1982        makes it give the same error for \k'' (previously it complained that that
1983        was a reference to a non-existent subpattern).
1984    
1985    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1986        this is right because (?- can be followed by option settings as well as by
1987        digits. I have, however, made the messages clearer.
1988    
1989    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1990        than the number used in the conditional) now cause a compile-time error.
1991        This is actually not compatible with Perl, which accepts such patterns, but
1992        treats the conditional as always being FALSE (as PCRE used to), but it
1993        seems to me that giving a diagnostic is better.
1994    
1995    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1996        and messages.
1997    
1998    9.  Fix two occurrences of "backslash" in comments that should have been
1999        "backspace".
2000    
2001    10. Remove two redundant lines of code that can never be obeyed (their function
2002        was moved elsewhere).
2003    
2004    11. The program that makes PCRE's Unicode character property table had a bug
2005        which caused it to generate incorrect table entries for sequences of
2006        characters that have the same character type, but are in different scripts.
2007        It amalgamated them into a single range, with the script of the first of
2008        them. In other words, some characters were in the wrong script. There were
2009        thirteen such cases, affecting characters in the following ranges:
2010    
2011          U+002b0 - U+002c1
2012          U+0060c - U+0060d
2013          U+0061e - U+00612
2014          U+0064b - U+0065e
2015          U+0074d - U+0076d
2016          U+01800 - U+01805
2017          U+01d00 - U+01d77
2018          U+01d9b - U+01dbf
2019          U+0200b - U+0200f
2020          U+030fc - U+030fe
2021          U+03260 - U+0327f
2022          U+0fb46 - U+0fbb1
2023          U+10450 - U+1049d
2024    
2025    12. The -o option (show only the matching part of a line) for pcregrep was not
2026        compatible with GNU grep in that, if there was more than one match in a
2027        line, it showed only the first of them. It now behaves in the same way as
2028        GNU grep.
2029    
2030    13. If the -o and -v options were combined for pcregrep, it printed a blank
2031        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2032        does the same. The return code can be used to tell if there were any
2033        non-matching lines.
2034    
2035    14. Added --file-offsets and --line-offsets to pcregrep.
2036    
2037    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2038        infinitely looping recursion. The bug was that positive lookaheads were not
2039        being skipped when checking for a possible empty match (negative lookaheads
2040        and both kinds of lookbehind were skipped).
2041    
2042    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2043        inclusion of <windows.h> to before rather than after the definition of
2044        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2045    
2046    17. Specifying a possessive quantifier with a specific limit for a Unicode
2047        character property caused pcre_compile() to compile bad code, which led at
2048        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2049        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2050        caused the error; without that there was no problem.
2051    
2052    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2053    
2054    19. Added --enable-pcretest-libreadline.
2055    
2056    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2057        RE::GlobalReplace(). As a result, the number of replacements returned was
2058        double what it should be. I removed one of the increments, but Craig sent a
2059        later patch that removed the other one (the right fix) and added unit tests
2060        that check the return values (which was not done before).
2061    
2062    21. Several CMake things:
2063    
2064        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2065            the names libpcre and libpcreposix, not just pcre and pcreposix.
2066    
2067        (2) The above change means that pcretest and pcregrep are now correctly
2068            linked with the newly-built libraries, not previously installed ones.
2069    
2070        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2071    
2072    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2073        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2074        UTF-8 newline character). The key issue is that the pattern starts .*;
2075        this means that the match must be either at the beginning, or after a
2076        newline. The bug was in the code for advancing after a failed match and
2077        checking that the new position followed a newline. It was not taking
2078        account of UTF-8 characters correctly.
2079    
2080    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2081        character classes. PCRE was not treating the sequence [:...:] as a
2082        character class unless the ... were all letters. Perl, however, seems to
2083        allow any characters between [: and :], though of course it rejects as
2084        unknown any "names" that contain non-letters, because all the known class
2085        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2086        for example, whereas PCRE did not - it did not recognize a POSIX character
2087        class. This seemed a bit dangerous, so the code has been changed to be
2088        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2089        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2090        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2091        Perl does, and where it didn't before.
2092    
2093    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2094        Windows environments %n is disabled by default.
2095    
2096    
2097    Version 7.4 21-Sep-07
2098    ---------------------
2099    
2100    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2101        means that a class such as [\s] counted as "explicit reference to CR or
2102        LF". That isn't really right - the whole point of the change was to try to
2103        help when there was an actual mention of one of the two characters. So now
2104        the change happens only if \r or \n (or a literal CR or LF) character is
2105        encountered.
2106    
2107    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2108        of both had grown to the point where there were only 3 bits left.
2109        Fortunately, there was spare space in the data structure, and so I have
2110        moved the internal flags into a new 16-bit field to free up more option
2111        bits.
2112    
2113    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2114        but did not set the internal JCHANGED flag - either of these is enough to
2115        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2116        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2117        start sets both bits.
2118    
2119    4.  Added options (at build time, compile time, exec time) to change \R from
2120        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2121    
2122    5.  doc/pcresyntax.html was missing from the distribution.
2123    
2124    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2125        compatibility, even though it is no longer used.
2126    
2127    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2128        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2129        windows.h file is present (where different names are used). [This was
2130        reversed later after testing - see 16 below.]
2131    
2132    8.  Changed all #include <config.h> to #include "config.h". There were also
2133        some further <pcre.h> cases that I changed to "pcre.h".
2134    
2135    9.  When pcregrep was used with the --colour option, it missed the line ending
2136        sequence off the lines that it output.
2137    
2138    10. It was pointed out to me that arrays of string pointers cause lots of
2139        relocations when a shared library is dynamically loaded. A technique of
2140        using a single long string with a table of offsets can drastically reduce
2141        these. I have refactored PCRE in four places to do this. The result is
2142        dramatic:
2143    
2144          Originally:                          290
2145          After changing UCP table:            187
2146          After changing error message table:   43
2147          After changing table of "verbs"       36
2148          After changing table of Posix names   22
2149    
2150        Thanks to the folks working on Gregex for glib for this insight.
2151    
2152    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2153        unicode-properties was also set.
2154    
2155    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2156    
2157    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2158        checked only for CRLF.
2159    
2160    14. Added casts to pcretest.c to avoid compiler warnings.
2161    
2162    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2163    
2164    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2165        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2166        entirely. This removes changes made in 7 above.
2167    
2168    17. The CMake files have been updated, and there is now more information about
2169        building with CMake in the NON-UNIX-USE document.
2170    
2171    
2172    Version 7.3 28-Aug-07
2173    ---------------------
2174    
2175     1. In the rejigging of the build system that eventually resulted in 7.1, the
2176        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2177        brackets there is not right, since it causes compilers to look for an
2178        installed pcre.h, not the version that is in the source that is being
2179        compiled (which of course may be different). I have changed it back to:
2180    
2181          #include "pcre.h"
2182    
2183        I have a vague recollection that the change was concerned with compiling in
2184        different directories, but in the new build system, that is taken care of
2185        by the VPATH setting the Makefile.
2186    
2187     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2188        when the subject happened to end in the byte 0x85 (e.g. if the last
2189        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2190        characters but of course it shouldn't be taken as a newline when it is part
2191        of another character. The bug was that, for an unlimited repeat of . in
2192        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2193        characters when looking for a newline.
2194    
2195     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2196    
2197     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2198        in debug output.
2199    
2200     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2201        long printing in the pcrecpp unittest when running under MinGW.
2202    
2203     6. ESC_K was left out of the EBCDIC table.
2204    
2205     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2206        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2207        limit also applies to "virtual nesting" when a pattern is recursive, and in
2208        this case 1000 isn't so big. I have been able to remove this limit at the
2209        expense of backing off one optimization in certain circumstances. Normally,
2210        when pcre_exec() would call its internal match() function recursively and
2211        immediately return the result unconditionally, it uses a "tail recursion"
2212        feature to save stack. However, when a subpattern that can match an empty
2213        string has an unlimited repetition quantifier, it no longer makes this
2214        optimization. That gives it a stack frame in which to save the data for
2215        checking that an empty string has been matched. Previously this was taken
2216        from the 1000-entry workspace that had been reserved. So now there is no
2217        explicit limit, but more stack is used.
2218    
2219     8. Applied Daniel's patches to solve problems with the import/export magic
2220        syntax that is required for Windows, and which was going wrong for the
2221        pcreposix and pcrecpp parts of the library. These were overlooked when this
2222        problem was solved for the main library.
2223    
2224     9. There were some crude static tests to avoid integer overflow when computing
2225        the size of patterns that contain repeated groups with explicit upper
2226        limits. As the maximum quantifier is 65535, the maximum group length was
2227        set at 30,000 so that the product of these two numbers did not overflow a
2228        32-bit integer. However, it turns out that people want to use groups that
2229        are longer than 30,000 bytes (though not repeat them that many times).
2230        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2231        made it possible to implement the integer overflow checks in a much more
2232        dynamic way, which I have now done. The artificial limitation on group
2233        length has been removed - we now have only the limit on the total length of
2234        the compiled pattern, which depends on the LINK_SIZE setting.
2235    
2236    10. Fixed a bug in the documentation for get/copy named substring when
2237        duplicate names are permitted. If none of the named substrings are set, the
2238        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2239        empty string.
2240    
2241    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2242        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2243        because the ] is interpreted as the first data character and the
2244        terminating ] is not found. PCRE has been made compatible with Perl in this
2245        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2246        cause memory overwriting.
2247    
2248    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2249        string has been matched (to stop an infinite loop). It was not recognizing
2250        a conditional subpattern that could match an empty string if that
2251        subpattern was within another subpattern. For example, it looped when
2252        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2253        condition was not nested. This bug has been fixed.
2254    
2255    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2256        past the start of the subject in the presence of bytes with the top bit
2257        set, for example "\x8aBCD".
2258    
2259    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2260        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2261    
2262    14. Optimized (?!) to (*FAIL).
2263    
2264    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2265        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2266        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2267        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2268        does: it's just the validity check that is more restrictive.
2269    
2270    16. Inserted checks for integer overflows during escape sequence (backslash)
2271        processing, and also fixed erroneous offset values for syntax errors during
2272        backslash processing.
2273    
2274    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2275        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2276    
2277    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2278        caused an overrun.
2279    
2280    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2281        something other than just ASCII characters) inside a group that had an
2282        unlimited repeat caused a loop at compile time (while checking to see
2283        whether the group could match an empty string).
2284    
2285    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2286        [\P{Any}] did so. (Error in the code for printing property names.)
2287    
2288    21. An orphan \E inside a character class could cause a crash.
2289    
2290    22. A repeated capturing bracket such as (A)? could cause a wild memory
2291        reference during compilation.
2292    
2293    23. There are several functions in pcre_compile() that scan along a compiled
2294        expression for various reasons (e.g. to see if it's fixed length for look
2295        behind). There were bugs in these functions when a repeated \p or \P was
2296        present in the pattern. These operators have additional parameters compared
2297        with \d, etc, and these were not being taken into account when moving along
2298        the compiled data. Specifically:
2299    
2300        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2301            length.
2302    
2303        (b) An item such as \pL+ within a repeated group could cause crashes or
2304            loops.
2305    
2306        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2307            "reference to non-existent subpattern" error.
2308    
2309        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2310    
2311    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2312        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2313    
2314    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2315    
2316    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2317        character were causing crashes (broken optimization).
2318    
2319    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2320        \p or \P) caused a compile-time loop.
2321    
2322    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2323        break. For example, the unstudied pattern [\r\n]A does not match the string
2324        "\r\nA" because change 7.0/46 below moves the current point on by two
2325        characters after failing to match at the start. However, the pattern \nA
2326        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2327        the same is true. There doesn't seem any very clean way out of this, but
2328        what I have chosen to do makes the common cases work: PCRE now takes note
2329        of whether there can be an explicit match for \r or \n anywhere in the
2330        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2331        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2332        pattern has explicit CR or LF references.
2333    
2334    29. Added (*CR) etc for changing newline setting at start of pattern.
2335    
2336    
2337    Version 7.2 19-Jun-07
2338    ---------------------
2339    
2340     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2341        which is apparently normally available under Windows.
2342    
2343     2. Re-jig the pcregrep tests with different newline settings in an attempt
2344        to make them independent of the local environment's newline setting.
2345    
2346     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2347    
2348     4. Some of the "internals" tests were previously cut out when the link size
2349        was not 2, because the output contained actual offsets. The recent new
2350        "Z" feature of pcretest means that these can be cut out, making the tests
2351        usable with all link sizes.
2352    
2353     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2354        stack recursion. This gives a massive performance boost under BSD, but just
2355        a small improvement under Linux. However, it saves one field in the frame
2356        in all cases.
2357    
2358     6. Added more features from the forthcoming Perl 5.10:
2359    
2360        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2361            recursion call. It refers to the nth most recently opened parentheses.
2362    
2363        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2364            to be opened parentheses.
2365    
2366        (c) Conditions that refer to capturing parentheses can be specified
2367            relatively, for example, (?(-2)... or (?(+3)...
2368    
2369        (d) \K resets the start of the current match so that everything before
2370            is not part of it.
2371    
2372        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2373    
2374        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2375            reference syntax.
2376    
2377        (g) (?| introduces a group in which the numbering of parentheses in each
2378            alternative starts with the same number.
2379    
2380        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2381    
2382     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2383        PCRE_INFO_JCHANGED.
2384    
2385     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2386        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2387        for detecting groups that can match an empty string.
2388    
2389     9. A pattern with a very large number of alternatives (more than several
2390        hundred) was running out of internal workspace during the pre-compile
2391        phase, where pcre_compile() figures out how much memory will be needed. A
2392        bit of new cunning has reduced the workspace needed for groups with
2393        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2394        workspace instead of running out of the 4096 that are available.
2395    
2396    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2397    
2398    11. Applied patch from Google to remove an optimization that didn't quite work.
2399        The report of the bug said:
2400    
2401          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2402          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2403          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2404    
2405    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2406        it matched the wrong number of bytes.
2407    
2408    
2409    Version 7.1 24-Apr-07
2410  ---------------------  ---------------------
2411    
2412   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
# Line 126  Version 7.1 12-Mar-07 Line 2531  Version 7.1 12-Mar-07
2531  16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"  16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2532      in an attempt to make files that differ only in their line terminators      in an attempt to make files that differ only in their line terminators
2533      compare equal. This works on Linux.      compare equal. This works on Linux.
2534    
2535  17. Under certain error circumstances pcregrep might try to free random memory  17. Under certain error circumstances pcregrep might try to free random memory
2536      as it exited. This is now fixed, thanks to valgrind.      as it exited. This is now fixed, thanks to valgrind.
2537    
2538  19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string  19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2539      "abc\r\n\r\n", it found an unwanted second match after the second \r. This      "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2540      was because its rules for how to advance for /g after matching an empty      was because its rules for how to advance for /g after matching an empty
2541      string did not allow for this case. They now check for it specially.      string at the end of a line did not allow for this case. They now check for
2542        it specially.
2543    
2544    20. pcretest is supposed to handle patterns and data of any length, by
2545        extending its buffers when necessary. It was getting this wrong when the
2546        buffer for a data line had to be extended.
2547    
2548    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2549        CRLF as a newline sequence.
2550    
2551    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2552        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2553        I have nevertheless tidied it up.
2554    
2555    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2556    
2557    24. Added a man page for pcre-config.
2558    
2559    
2560  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06

Legend:
Removed from v.141  
changed lines
  Added in v.1300

  ViewVC Help
Powered by ViewVC 1.1.5