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

Diff of /code/trunk/ChangeLog

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

revision 65 by nigel, Sat Feb 24 21:40:08 2007 UTC revision 1342 by ph10, Mon Jul 1 12:28:12 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.34 xx-xxxx-201x
5    --------------------------
6    
7    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8        executable memory. Patch inspired by Carsten Klein.
9    
10    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11        this macro is never tested and has no effect, because the work to support
12        coverage involves only compiling and linking options and special targets in
13        the Makefile. The comment in config.h implied that defining the macro would
14        enable coverage support, which is totally false. There was also support for
15        setting this macro in the CMake files (my fault, I just copied it from
16        configure). SUPPORT_GCOV has now been removed.
17    
18    
19    Version 8.33 28-May-2013
20    --------------------------
21    
22    1.  Added 'U' to some constants that are compared to unsigned integers, to
23        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
24        variables that are added to signed variables, to ensure the result is
25        signed and can be negated.
26    
27    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
28        CMake config files.
29    
30    3.  Revise the creation of config.h.generic so that all boolean macros are
31        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
32        overriding via -D on the command line possible.
33    
34    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
35        to unsigned int is reported to make a quite noticeable speed difference in
36        a specific Windows environment. Testing on Linux did also appear to show
37        some benefit (and it is clearly not harmful). Also fixed the definition of
38        Xop which should be unsigned.
39    
40    5.  Related to (4), changing the definition of the intermediate variable cc
41        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
42        improvements.
43    
44    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
45        unnecessary spaces.
46    
47    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
48        and later.
49    
50    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
51    
52    9.  Optimizing fast_forward_start_bits in JIT.
53    
54    10. Adding support for callouts in JIT, and fixing some issues revealed
55        during this work. Namely:
56    
57        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
58    
59        (b) Minimum length was not checked before the matching is started.
60    
61    11. The value of capture_last that is passed to callouts was incorrect in some
62        cases when there was a capture on one path that was subsequently abandoned
63        after a backtrack. Also, the capture_last value is now reset after a
64        recursion, since all captures are also reset in this case.
65    
66    12. The interpreter no longer returns the "too many substrings" error in the
67        case when an overflowing capture is in a branch that is subsequently
68        abandoned after a backtrack.
69    
70    13. In the pathological case when an offset vector of size 2 is used, pcretest
71        now prints out the matched string after a yield of 0 or 1.
72    
73    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
74        Only supported by the JIT compiler at the moment.
75    
76    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
77    
78    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
79        offset of the starting point of the matching process, provided the offsets
80        vector is large enough.
81    
82    17. The \A escape now records a lookbehind value of 1, though its execution
83        does not actually inspect the previous character. This is to ensure that,
84        in partial multi-segment matching, at least one character from the old
85        segment is retained when a new segment is processed. Otherwise, if there
86        are no lookbehinds in the pattern, \A might match incorrectly at the start
87        of a new segment.
88    
89    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
90    
91    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
92        gives some modest performance improvement in 8-bit mode.
93    
94    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
95        be expressed in certain programming languages using Universal Character
96        Names.
97    
98    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
99        which points out that "non characters" are not "characters that may not
100        appear in Unicode strings" but rather "characters that are reserved for
101        internal use and have only local meaning".
102    
103    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
104        there was a conditional group that depended on an assertion, if the
105        assertion was false, the callout that immediately followed the alternation
106        in the condition was skipped when pcre_exec() was used for matching.
107    
108    23. Allow an explicit callout to be inserted before an assertion that is the
109        condition for a conditional group, for compatibility with automatic
110        callouts, which always insert a callout at this point.
111    
112    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
113        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
114    
115    25. (*PRUNE) is now supported by the JIT compiler.
116    
117    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
118    
119    27. Fix the case where there are two or more SKIPs with arguments that may be
120        ignored.
121    
122    28. (*SKIP) is now supported by the JIT compiler.
123    
124    29. (*THEN) is now supported by the JIT compiler.
125    
126    30. Update RunTest with additional test selector options.
127    
128    31. The way PCRE handles backtracking verbs has been changed in two ways.
129    
130        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
131        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
132        backtracking. In some cases this makes it more Perl-compatible, but Perl's
133        rather obscure rules do not always do the same thing.
134    
135        (2) Previously, backtracking verbs were confined within assertions. This is
136        no longer the case for positive assertions, except for (*ACCEPT). Again,
137        this sometimes improves Perl compatibility, and sometimes does not.
138    
139    32. A number of tests that were in test 2 because Perl did things differently
140        have been moved to test 1, because either Perl or PCRE has changed, and
141        these tests are now compatible.
142    
143    32. Backtracking control verbs are now handled in the same way in JIT and
144        interpreter.
145    
146    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
147        contained a forward subroutine reference caused a compile error.
148    
149    34. Auto-detect and optimize limited repetitions in JIT.
150    
151    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
152        blocking (*UTF) etc.
153    
154    36. In the interpreter, maximizing pattern repetitions for characters and
155        character types now use tail recursion, which reduces stack usage.
156    
157    37. The value of the max lookbehind was not correctly preserved if a compiled
158        and saved regex was reloaded on a host of different endianness.
159    
160    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
161        of the compiled pattern block, expand the flags field from 16 to 32 bits
162        because it was almost full.
163    
164    39. Try madvise first before posix_madvise.
165    
166    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
167        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
168        original change locked out all matches of empty strings. This has been
169        changed so that one match of an empty string per line is recognized.
170        Subsequent searches on the same line (for colouring or for --only-matching,
171        for example) do not recognize empty strings.
172    
173    41. Applied a user patch to fix a number of spelling mistakes in comments.
174    
175    42. Data lines longer than 65536 caused pcretest to crash.
176    
177    43. Clarified the data type for length and startoffset arguments for pcre_exec
178        and pcre_dfa_exec in the function-specific man pages, where they were
179        explicitly stated to be in bytes, never having been updated. I also added
180        some clarification to the pcreapi man page.
181    
182    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
183        a segmentation fault.
184    
185    
186    Version 8.32 30-November-2012
187    -----------------------------
188    
189    1.  Improved JIT compiler optimizations for first character search and single
190        character iterators.
191    
192    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
193        Patch by Daniel Richard G.
194    
195    3.  Single character iterator optimizations in the JIT compiler.
196    
197    4.  Improved JIT compiler optimizations for character ranges.
198    
199    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
200        Reported by Giuseppe D'Angelo.
201    
202    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
203        of a line, was being set incorrectly in cases where .* appeared inside
204        atomic brackets at the start of a pattern, or where there was a subsequent
205        *PRUNE or *SKIP.
206    
207    7.  Improved instruction cache flush for POWER/PowerPC.
208        Patch by Daniel Richard G.
209    
210    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
211        grep:
212    
213        (a) There is now no limit to the number of patterns to be matched.
214    
215        (b) An error is given if a pattern is too long.
216    
217        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
218            are now supported.
219    
220        (d) --exclude-from and --include-from (multiple use) have been added.
221    
222        (e) Exclusions and inclusions now apply to all files and directories, not
223            just to those obtained from scanning a directory recursively.
224    
225        (f) Multiple uses of -f and --file-list are now supported.
226    
227        (g) In a Windows environment, the default for -d has been changed from
228            "read" (the GNU grep default) to "skip", because otherwise the presence
229            of a directory in the file list provokes an error.
230    
231        (h) The documentation has been revised and clarified in places.
232    
233    9.  Improve the matching speed of capturing brackets.
234    
235    10. Changed the meaning of \X so that it now matches a Unicode extended
236        grapheme cluster.
237    
238    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
239        out POSIX threads when JIT support is configured.
240    
241    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
242    
243    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
244        structure could go wrong in environments where size_t is not the same size
245        as int.
246    
247    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
248        set.
249    
250    15. The EBCDIC support had decayed; later updates to the code had included
251        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
252        general tidy up of EBCDIC-related issues, and the documentation was also
253        not quite right. There is now a test that can be run on ASCII systems to
254        check some of the EBCDIC-related things (but is it not a full test).
255    
256    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
257        in a small tidy to the code.
258    
259    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
260    
261    18. If the --only-matching (-o) option in pcregrep is specified multiple
262        times, each one causes appropriate output. For example, -o1 -o2 outputs the
263        substrings matched by the 1st and 2nd capturing parentheses. A separating
264        string can be specified by --om-separator (default empty).
265    
266    19. Improving the first n character searches.
267    
268    20. Turn case lists for horizontal and vertical white space into macros so that
269        they are defined only once.
270    
271    21. This set of changes together give more compatible Unicode case-folding
272        behaviour for characters that have more than one other case when UCP
273        support is available.
274    
275        (a) The Unicode property table now has offsets into a new table of sets of
276            three or more characters that are case-equivalent. The MultiStage2.py
277            script that generates these tables (the pcre_ucd.c file) now scans
278            CaseFolding.txt instead of UnicodeData.txt for character case
279            information.
280    
281        (b) The code for adding characters or ranges of characters to a character
282            class has been abstracted into a generalized function that also handles
283            case-independence. In UTF-mode with UCP support, this uses the new data
284            to handle characters with more than one other case.
285    
286        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
287            whose other case is greater than 256 are now correctly matched
288            caselessly. Previously, the high codepoint matched the low one, but not
289            vice versa.
290    
291        (d) The processing of \h, \H, \v, and \ in character classes now makes use
292            of the new class addition function, using character lists defined as
293            macros alongside the case definitions of 20 above.
294    
295        (e) Caseless back references now work with characters that have more than
296            one other case.
297    
298        (f) General caseless matching of characters with more than one other case
299            is supported.
300    
301    22. Unicode character properties were updated from Unicode 6.2.0
302    
303    23. Improved CMake support under Windows. Patch by Daniel Richard G.
304    
305    24. Add support for 32-bit character strings, and UTF-32
306    
307    25. Major JIT compiler update (code refactoring and bugfixing).
308        Experimental Sparc 32 support is added.
309    
310    26. Applied a modified version of Daniel Richard G's patch to create
311        pcre.h.generic and config.h.generic by "make" instead of in the
312        PrepareRelease script.
313    
314    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
315        pcre_compile.c when checking for a zero character.
316    
317    28. Introducing a native interface for JIT. Through this interface, the compiled
318        machine code can be directly executed. The purpose of this interface is to
319        provide fast pattern matching, so several sanity checks are not performed.
320        However, feature tests are still performed. The new interface provides
321        1.4x speedup compared to the old one.
322    
323    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
324        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
325        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
326    
327    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
328        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
329        the "old" RFC 2279). Instead, it ended up passing a negative length to
330        pcre_exec().
331    
332    31. Add support for GCC's visibility feature to hide internal functions.
333    
334    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
335        "unknown -C option" after outputting 0 or 1.
336    
337    33. There is now support for generating a code coverage report for the test
338        suite in environments where gcc is the compiler and lcov is installed. This
339        is mainly for the benefit of the developers.
340    
341    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
342        unaddressable using valgrind annotations, allowing valgrind to detect
343        invalid memory accesses. This is mainly for the benefit of the developers.
344    
345    25. (*UTF) can now be used to start a pattern in any of the three libraries.
346    
347    26. Give configure error if --enable-cpp but no C++ compiler found.
348    
349    
350    Version 8.31 06-July-2012
351    -------------------------
352    
353    1.  Fixing a wrong JIT test case and some compiler warnings.
354    
355    2.  Removed a bashism from the RunTest script.
356    
357    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
358        to unsigned type, result still unsigned" that was given by an MS compiler
359        on encountering the code "-sizeof(xxx)".
360    
361    4.  Partial matching support is added to the JIT compiler.
362    
363    5.  Fixed several bugs concerned with partial matching of items that consist
364        of more than one character:
365    
366        (a) /^(..)\1/ did not partially match "aba" because checking references was
367            done on an "all or nothing" basis. This also applied to repeated
368            references.
369    
370        (b) \R did not give a hard partial match if \r was found at the end of the
371            subject.
372    
373        (c) \X did not give a hard partial match after matching one or more
374            characters at the end of the subject.
375    
376        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
377            a partial match for the string "\r".
378    
379        (e) When newline was set to CRLF, the metacharacter "." did not recognize
380            a partial match for a CR character at the end of the subject string.
381    
382    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
383        running pcretest, the text "(JIT)" added to the output whenever JIT is
384        actually used to run the match.
385    
386    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
387        or /S+[+] with a digit between 1 and 7.
388    
389    8.  OP_NOT now supports any UTF character not just single-byte ones.
390    
391    9.  (*MARK) control verb is now supported by the JIT compiler.
392    
393    10. The command "./RunTest list" lists the available tests without actually
394        running any of them. (Because I keep forgetting what they all are.)
395    
396    11. Add PCRE_INFO_MAXLOOKBEHIND.
397    
398    12. Applied a (slightly modified) user-supplied patch that improves performance
399        when the heap is used for recursion (compiled with --disable-stack-for-
400        recursion). Instead of malloc and free for each heap frame each time a
401        logical recursion happens, frames are retained on a chain and re-used where
402        possible. This sometimes gives as much as 30% improvement.
403    
404    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
405        call.
406    
407    14. As documented, (*COMMIT) is now confined to within a positive assertion.
408    
409    15. It is now possible to link pcretest with libedit as an alternative to
410        libreadline.
411    
412    16. (*COMMIT) control verb is now supported by the JIT compiler.
413    
414    17. The Unicode data tables have been updated to Unicode 6.1.0.
415    
416    18. Added --file-list option to pcregrep.
417    
418    19. Added binary file support to pcregrep, including the -a, --binary-files,
419        -I, and --text options.
420    
421    20. The madvise function is renamed for posix_madvise for QNX compatibility
422        reasons. Fixed by Giuseppe D'Angelo.
423    
424    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
425    
426    22. Changed the option for creating symbolic links for 16-bit man pages from
427        -s to -sf so that re-installing does not cause issues.
428    
429    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
430    
431    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
432        matches in certain environments (the workspace was not being correctly
433        retained). Also added to pcre_dfa_exec() a simple plausibility check on
434        some of the workspace data at the beginning of a restart.
435    
436    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
437        was not doing so when it should - probably a typo introduced by SVN 528
438        (change 8.10/14).
439    
440    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
441        \w+ when the character tables indicated that \x{c4} was a word character.
442        There were several related cases, all because the tests for doing a table
443        lookup were testing for characters less than 127 instead of 255.
444    
445    27. If a pattern contains capturing parentheses that are not used in a match,
446        their slots in the ovector are set to -1. For those that are higher than
447        any matched groups, this happens at the end of processing. In the case when
448        there were back references that the ovector was too small to contain
449        (causing temporary malloc'd memory to be used during matching), and the
450        highest capturing number was not used, memory off the end of the ovector
451        was incorrectly being set to -1. (It was using the size of the temporary
452        memory instead of the true size.)
453    
454    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
455        ovector size, it uses memory at the end of the block that it has got.
456    
457    29. Check for an overlong MARK name and give an error at compile time. The
458        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
459    
460    30. JIT compiler update.
461    
462    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
463        Rill for the patch.
464    
465    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
466    
467    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
468    
469    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
470        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
471        was enabled.
472    
473    35. Improve JIT code generation for greedy plus quantifier.
474    
475    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
476        1 to "aa" instead of to an empty string. The bug affected repeated groups
477        that could potentially match an empty string.
478    
479    37. Optimizing single character iterators in JIT.
480    
481    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
482        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
483        codepoints that are too big for the mode are faulted, and in a UTF mode,
484        disallowed codepoints are also faulted.
485    
486    39. If PCRE was compiled with UTF support, in three places in the DFA
487        matcher there was code that should only have been obeyed in UTF mode, but
488        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
489        processing when bytes with values greater than 127 were present. In 16-bit
490        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
491        both cases the values are those that cannot be the first data item in a UTF
492        character. The three items that might have provoked this were recursions,
493        possessively repeated groups, and atomic groups.
494    
495    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
496        and pcregrep, because some OS require shared objects to be explicitly
497        passed to ld, causing the link step to fail if they are not.
498    
499    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
500        mode, patterns that started with \h* or \R* might be incorrectly matched.
501    
502    
503    Version 8.30 04-February-2012
504    -----------------------------
505    
506    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
507        name is defined in ctype.h.
508    
509    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
510        only in quite long subpatterns.
511    
512    3.  Removed the function pcre_info(), which has been obsolete and deprecated
513        since it was replaced by pcre_fullinfo() in February 2000.
514    
515    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
516        match a (*MARK), and the match failed at the start of the subject, a
517        reference to memory before the start of the subject could occur. This bug
518        was introduced by fix 17 of release 8.21.
519    
520    5.  A reference to an unset group with zero minimum repetition was giving
521        totally wrong answers (in non-JavaScript-compatibility mode). For example,
522        /(another)?(\1?)test/ matched against "hello world test". This bug was
523        introduced in release 8.13.
524    
525    6.  Add support for 16-bit character strings (a large amount of work involving
526        many changes and refactorings).
527    
528    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
529        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
530        from a file.
531    
532    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
533        rounding is not applied in this particular case).
534    
535    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
536        if they appear, or are escaped, in patterns.
537    
538    10. Get rid of a number of -Wunused-but-set-variable warnings.
539    
540    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
541        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
542        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
543        also returns the mark "x". This bug applied to capturing parentheses,
544        non-capturing parentheses, and atomic parentheses. It also applied to some
545        assertions.
546    
547    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
548        information out of configure.ac instead of relying on pcre.h.generic, which
549        is not stored in the repository.
550    
551    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
552        -lreadline.
553    
554    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
555    
556    15. Applied Graycode's patch to put the top-level frame on the stack rather
557        than the heap when not using the stack for recursion. This gives a
558        performance improvement in many cases when recursion is not deep.
559    
560    16. Experimental code added to "pcretest -C" to output the stack frame size.
561    
562    
563    Version 8.21 12-Dec-2011
564    ------------------------
565    
566    1.  Updating the JIT compiler.
567    
568    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
569        are added as well.
570    
571    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
572        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
573        calling _pcre_jit_exec. Some extra comments are added.
574    
575    4.  (*MARK) settings inside atomic groups that do not contain any capturing
576        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
577        was introduced by change 18 for 8.20.
578    
579    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
580        ECMA-262 standard.
581    
582    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
583        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
584        This bug was probably introduced by change 9 of 8.13.
585    
586    7.  While fixing 6 above, I noticed that a number of other items were being
587        incorrectly rejected as "not fixed length". This arose partly because newer
588        opcodes had not been added to the fixed-length checking code. I have (a)
589        corrected the bug and added tests for these items, and (b) arranged for an
590        error to occur if an unknown opcode is encountered while checking for fixed
591        length instead of just assuming "not fixed length". The items that were
592        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
593        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
594        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
595    
596    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
597        being incorrectly compiled and would have given unpredicatble results.
598    
599    9.  A possessively repeated subpattern with minimum repeat count greater than
600        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
601        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
602        the first (A) could occur when it should not.
603    
604    10. Add a cast and remove a redundant test from the code.
605    
606    11. JIT should use pcre_malloc/pcre_free for allocation.
607    
608    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
609        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
610        is anything other than /usr, -L is still shown).
611    
612    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
613    
614    14. Perl does not support \N without a following name in a [] class; PCRE now
615        also gives an error.
616    
617    15. If a forward reference was repeated with an upper limit of around 2000,
618        it caused the error "internal error: overran compiling workspace". The
619        maximum number of forward references (including repeats) was limited by the
620        internal workspace, and dependent on the LINK_SIZE. The code has been
621        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
622        the default depends on LINK_SIZE. There is a new upper limit (for safety)
623        of around 200,000 forward references. While doing this, I also speeded up
624        the filling in of repeated forward references.
625    
626    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
627        incorrectly expecting the subject to contain another "a" after the start.
628    
629    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
630        in the match, the SKIP should be ignored. This was not happening; instead
631        the SKIP was being treated as NOMATCH. For patterns such as
632        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
633        tested.
634    
635    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
636        now much more compatible with Perl, in particular in cases where the result
637        is a non-match for a non-anchored pattern. For example, if
638        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
639        "m", where previously it did not return a name. A side effect of this
640        change is that for partial matches, the last encountered mark name is
641        returned, as for non matches. A number of tests that were previously not
642        Perl-compatible have been moved into the Perl-compatible test files. The
643        refactoring has had the pleasing side effect of removing one argument from
644        the match() function, thus reducing its stack requirements.
645    
646    19. If the /S+ option was used in pcretest to study a pattern using JIT,
647        subsequent uses of /S (without +) incorrectly behaved like /S+.
648    
649    21. Retrieve executable code size support for the JIT compiler and fixing
650        some warnings.
651    
652    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
653        not work when the shorter character appeared right at the end of the
654        subject string.
655    
656    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
657        systems.
658    
659    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
660        output it when the /M option is used in pcretest.
661    
662    25. The CheckMan script was not being included in the distribution. Also, added
663        an explicit "perl" to run Perl scripts from the PrepareRelease script
664        because this is reportedly needed in Windows.
665    
666    26. If study data was being save in a file and studying had not found a set of
667        "starts with" bytes for the pattern, the data written to the file (though
668        never used) was taken from uninitialized memory and so caused valgrind to
669        complain.
670    
671    27. Updated RunTest.bat as provided by Sheri Pierce.
672    
673    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
674    
675    29. Computation of memory usage for the table of capturing group names was
676        giving an unnecessarily large value.
677    
678    
679    Version 8.20 21-Oct-2011
680    ------------------------
681    
682    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
683        a POSIX class. After further experiments with Perl, which convinced me that
684        Perl has bugs and confusions, a closing square bracket is no longer allowed
685        in a POSIX name. This bug also affected patterns with classes that started
686        with full stops.
687    
688    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
689        captured substring, but while checking the failing first alternative,
690        substring 1 is temporarily captured. If the output vector supplied to
691        pcre_exec() was not big enough for this capture, the yield of the function
692        was still zero ("insufficient space for captured substrings"). This cannot
693        be totally fixed without adding another stack variable, which seems a lot
694        of expense for a edge case. However, I have improved the situation in cases
695        such as /(a)(b)x|abc/ matched against "abc", where the return code
696        indicates that fewer than the maximum number of slots in the ovector have
697        been set.
698    
699    3.  Related to (2) above: when there are more back references in a pattern than
700        slots in the output vector, pcre_exec() uses temporary memory during
701        matching, and copies in the captures as far as possible afterwards. It was
702        using the entire output vector, but this conflicts with the specification
703        that only 2/3 is used for passing back captured substrings. Now it uses
704        only the first 2/3, for compatibility. This is, of course, another edge
705        case.
706    
707    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
708        main code base, and can be used by building with --enable-jit. When this is
709        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
710        runtime --no-jit option is given.
711    
712    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
713        ovector, the return from the function was zero, implying that there were
714        other matches that did not fit. The correct "exactly full" value is now
715        returned.
716    
717    6.  If a subpattern that was called recursively or as a subroutine contained
718        (*PRUNE) or any other control that caused it to give a non-standard return,
719        invalid errors such as "Error -26 (nested recursion at the same subject
720        position)" or even infinite loops could occur.
721    
722    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
723        computing the minimum length on reaching *ACCEPT, and so ended up with the
724        wrong value of 1 rather than 0. Further investigation indicates that
725        computing a minimum subject length in the presence of *ACCEPT is difficult
726        (think back references, subroutine calls), and so I have changed the code
727        so that no minimum is registered for a pattern that contains *ACCEPT.
728    
729    8.  If (*THEN) was present in the first (true) branch of a conditional group,
730        it was not handled as intended. [But see 16 below.]
731    
732    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
733        Sheri Pierce.
734    
735    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
736        the first byte in a match must be "a".
737    
738    11. Change 17 for 8.13 increased the recursion depth for patterns like
739        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
740        pattern contains any instances of (*THEN). If it does not, the old
741        optimizations are restored. It would be nice to do this on a per-group
742        basis, but at the moment that is not feasible.
743    
744    12. In some environments, the output of pcretest -C is CRLF terminated. This
745        broke RunTest's code that checks for the link size. A single white space
746        character after the value is now allowed for.
747    
748    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
749        For "fr", it uses the Windows-specific input and output files.
750    
751    14. If (*THEN) appeared in a group that was called recursively or as a
752        subroutine, it did not work as intended. [But see next item.]
753    
754    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
755        pattern fragments (but not containing any | characters). If A and B are
756        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
757        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
758        D. In other words, Perl considers parentheses that do not contain any |
759        characters to be part of a surrounding alternative, whereas PCRE was
760        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
761        differently. PCRE now behaves in the same way as Perl, except in the case
762        of subroutine/recursion calls such as (?1) which have in any case always
763        been different (but PCRE had them first :-).
764    
765    16. Related to 15 above: Perl does not treat the | in a conditional group as
766        creating alternatives. Such a group is treated in the same way as an
767        ordinary group without any | characters when processing (*THEN). PCRE has
768        been changed to match Perl's behaviour.
769    
770    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
771        RunGrepTest script failed.
772    
773    18. Change 22 for version 13 caused atomic groups to use more stack. This is
774        inevitable for groups that contain captures, but it can lead to a lot of
775        stack use in large patterns. The old behaviour has been restored for atomic
776        groups that do not contain any capturing parentheses.
777    
778    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
779        suppress the check for a minimum subject length at run time. (If it was
780        given to pcre_exec() or pcre_dfa_exec() it did work.)
781    
782    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
783        fail to work when decoding hex characters in data strings in EBCDIC
784        environments.
785    
786    21. It appears that in at least one Mac OS environment, the isxdigit() function
787        is implemented as a macro that evaluates to its argument more than once,
788        contravening the C 90 Standard (I haven't checked a later standard). There
789        was an instance in pcretest which caused it to go wrong when processing
790        \x{...} escapes in subject strings. The has been rewritten to avoid using
791        things like p++ in the argument of isxdigit().
792    
793    
794    Version 8.13 16-Aug-2011
795    ------------------------
796    
797    1.  The Unicode data tables have been updated to Unicode 6.0.0.
798    
799    2.  Two minor typos in pcre_internal.h have been fixed.
800    
801    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
802        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
803        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
804    
805    4.  There were a number of related bugs in the code for matching backrefences
806        caselessly in UTF-8 mode when codes for the characters concerned were
807        different numbers of bytes. For example, U+023A and U+2C65 are an upper
808        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
809        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
810        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
811        2-byte code at the end of the subject (it thought there wasn't enough data
812        left).
813    
814    5.  Comprehensive information about what went wrong is now returned by
815        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
816        as the output vector has at least 2 elements. The offset of the start of
817        the failing character and a reason code are placed in the vector.
818    
819    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
820        now returned is for the first byte of the failing character, instead of the
821        last byte inspected. This is an incompatible change, but I hope it is small
822        enough not to be a problem. It makes the returned offset consistent with
823        pcre_exec() and pcre_dfa_exec().
824    
825    7.  pcretest now gives a text phrase as well as the error number when
826        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
827        failure, the offset and reason code are output.
828    
829    8.  When \R was used with a maximizing quantifier it failed to skip backwards
830        over a \r\n pair if the subsequent match failed. Instead, it just skipped
831        back over a single character (\n). This seems wrong (because it treated the
832        two characters as a single entity when going forwards), conflicts with the
833        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
834        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
835        has been changed.
836    
837    9.  Some internal refactoring has changed the processing so that the handling
838        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
839        time (the PCRE_DOTALL option was changed this way some time ago: version
840        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
841        which was always a bit of a fudge. It also means that there is one less
842        argument for the match() function, which reduces its stack requirements
843        slightly. This change also fixes an incompatibility with Perl: the pattern
844        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
845    
846    10. More internal refactoring has drastically reduced the number of recursive
847        calls to match() for possessively repeated groups such as (abc)++ when
848        using pcre_exec().
849    
850    11. While implementing 10, a number of bugs in the handling of groups were
851        discovered and fixed:
852    
853        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
854        (a|)*(?1) gave a compile-time internal error.
855        ((a|)+)+  did not notice that the outer group could match an empty string.
856        (^a|^)+   was not marked as anchored.
857        (.*a|.*)+ was not marked as matching at start or after a newline.
858    
859    12. Yet more internal refactoring has removed another argument from the match()
860        function. Special calls to this function are now indicated by setting a
861        value in a variable in the "match data" data block.
862    
863    13. Be more explicit in pcre_study() instead of relying on "default" for
864        opcodes that mean there is no starting character; this means that when new
865        ones are added and accidentally left out of pcre_study(), testing should
866        pick them up.
867    
868    14. The -s option of pcretest has been documented for ages as being an old
869        synonym of -m (show memory usage). I have changed it to mean "force study
870        for every regex", that is, assume /S for every regex. This is similar to -i
871        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
872        using it. It makes it easier to run collections of tests with and without
873        study enabled, and thereby test pcre_study() more easily. All the standard
874        tests are now run with and without -s (but some patterns can be marked as
875        "never study" - see 20 below).
876    
877    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
878        restoration of the capturing data to the outer values was not happening
879        correctly.
880    
881    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
882        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
883        pattern had matched an empty string, and so incorrectly returned a no
884        match.
885    
886    17. There was optimizing code for the last branch of non-capturing parentheses,
887        and also for the obeyed branch of a conditional subexpression, which used
888        tail recursion to cut down on stack usage. Unfortunately, now that there is
889        the possibility of (*THEN) occurring in these branches, tail recursion is
890        no longer possible because the return has to be checked for (*THEN). These
891        two optimizations have therefore been removed. [But see 8.20/11 above.]
892    
893    18. If a pattern containing \R was studied, it was assumed that \R always
894        matched two bytes, thus causing the minimum subject length to be
895        incorrectly computed because \R can also match just one byte.
896    
897    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
898        was incorrectly computed.
899    
900    20. If /S is present twice on a test pattern in pcretest input, it now
901        *disables* studying, thereby overriding the use of -s on the command line
902        (see 14 above). This is necessary for one or two tests to keep the output
903        identical in both cases.
904    
905    21. When (*ACCEPT) was used in an assertion that matched an empty string and
906        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
907    
908    22. When an atomic group that contained a capturing parenthesis was
909        successfully matched, but the branch in which it appeared failed, the
910        capturing was not being forgotten if a higher numbered group was later
911        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
912        group 1 to "a", when in fact it should be unset. This applied to multi-
913        branched capturing and non-capturing groups, repeated or not, and also to
914        positive assertions (capturing in negative assertions does not happen
915        in PCRE) and also to nested atomic groups.
916    
917    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
918        subject after a captured substring, to make it easier to tell which of a
919        number of identical substrings has been captured.
920    
921    24. The way atomic groups are processed by pcre_exec() has been changed so that
922        if they are repeated, backtracking one repetition now resets captured
923        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
924        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
925        "aaa". Previously, it would have been "a". As part of this code
926        refactoring, the way recursive calls are handled has also been changed.
927    
928    25. If an assertion condition captured any substrings, they were not passed
929        back unless some other capturing happened later. For example, if
930        (?(?=(a))a) was matched against "a", no capturing was returned.
931    
932    26. When studying a pattern that contained subroutine calls or assertions,
933        the code for finding the minimum length of a possible match was handling
934        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
935        group 1 called group 2 while simultaneously a separate group 2 called group
936        1). A stack overflow occurred in this case. I have fixed this by limiting
937        the recursion depth to 10.
938    
939    27. Updated RunTest.bat in the distribution to the version supplied by Tom
940        Fortmann. This supports explicit test numbers on the command line, and has
941        argument validation and error reporting.
942    
943    28. An instance of \X with an unlimited repeat could fail if at any point the
944        first character it looked at was a mark character.
945    
946    29. Some minor code refactoring concerning Unicode properties and scripts
947        should reduce the stack requirement of match() slightly.
948    
949    30. Added the '=' option to pcretest to check the setting of unused capturing
950        slots at the end of the pattern, which are documented as being -1, but are
951        not included in the return count.
952    
953    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
954        compiled something random. Now it gives a compile-time error (as does
955        Perl).
956    
957    32. A *MARK encountered during the processing of a positive assertion is now
958        recorded and passed back (compatible with Perl).
959    
960    33. If --only-matching or --colour was set on a pcregrep call whose pattern
961        had alternative anchored branches, the search for a second match in a line
962        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
963        matched the line "0102" twice. The same bug affected patterns that started
964        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
965        twice.
966    
967    34. Previously, PCRE did not allow quantification of assertions. However, Perl
968        does, and because of capturing effects, quantifying parenthesized
969        assertions may at times be useful. Quantifiers are now allowed for
970        parenthesized assertions.
971    
972    35. A minor code tidy in pcre_compile() when checking options for \R usage.
973    
974    36. \g was being checked for fancy things in a character class, when it should
975        just be a literal "g".
976    
977    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
978        appearance of a nested POSIX class supersedes an apparent external class.
979        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
980        unescaped square brackets may also appear as part of class names. For
981        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
982        more like Perl. (But see 8.20/1 above.)
983    
984    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
985        was because it thought it was \N{name}, which is not supported).
986    
987    39. Add minix to OS list not supporting the -S option in pcretest.
988    
989    40. PCRE tries to detect cases of infinite recursion at compile time, but it
990        cannot analyze patterns in sufficient detail to catch mutual recursions
991        such as ((?1))((?2)). There is now a runtime test that gives an error if a
992        subgroup is called recursively as a subpattern for a second time at the
993        same position in the subject string. In previous releases this might have
994        been caught by the recursion limit, or it might have run out of stack.
995    
996    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
997        happen only once. PCRE was, however incorrectly giving a compile time error
998        "recursive call could loop indefinitely" because it cannot analyze the
999        pattern in sufficient detail. The compile time test no longer happens when
1000        PCRE is compiling a conditional subpattern, but actual runaway loops are
1001        now caught at runtime (see 40 above).
1002    
1003    42. It seems that Perl allows any characters other than a closing parenthesis
1004        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1005        has been changed to be the same.
1006    
1007    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1008        as not to get warnings when autogen.sh is called. Also changed
1009        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1010    
1011    44. To help people who use pcregrep to scan files containing exceedingly long
1012        lines, the following changes have been made:
1013    
1014        (a) The default value of the buffer size parameter has been increased from
1015            8K to 20K. (The actual buffer used is three times this size.)
1016    
1017        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1018            PCRE is built.
1019    
1020        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1021            to be set at run time.
1022    
1023        (d) Numerical values in pcregrep options can be followed by K or M, for
1024            example --buffer-size=50K.
1025    
1026        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1027            given and the return code is set to 2.
1028    
1029    45. Add a pointer to the latest mark to the callout data block.
1030    
1031    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1032        partial match of an empty string instead of no match. This was specific to
1033        the use of ".".
1034    
1035    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1036        complete match instead of a partial match. This bug was dependent on both
1037        the PCRE_UTF8 and PCRE_DOTALL options being set.
1038    
1039    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1040        starting byte set, because \b was not being ignored.
1041    
1042    
1043    Version 8.12 15-Jan-2011
1044    ------------------------
1045    
1046    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1047        checks for such things as part of the documentation building process.
1048    
1049    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1050        --match-limit and --recursion-limit options (added for 8.11). In
1051        particular, this made one of the standard tests fail. (The integer value
1052        went into the wrong half of a long int.)
1053    
1054    3.  If the --colour option was given to pcregrep with -v (invert match), it
1055        did strange things, either producing crazy output, or crashing. It should,
1056        of course, ignore a request for colour when reporting lines that do not
1057        match.
1058    
1059    4.  Another pcregrep bug caused similar problems if --colour was specified with
1060        -M (multiline) and the pattern match finished with a line ending.
1061    
1062    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1063        matched in multiline mode, the following line was shown as part of the
1064        match. This seems wrong, so I have changed it.
1065    
1066    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1067        the check for further matches in the same line (so they could be coloured)
1068        to overrun the end of the current line. If another match was found, it was
1069        incorrectly shown (and then shown again when found in the next line).
1070    
1071    7.  If pcregrep was compiled under Windows, there was a reference to the
1072        function pcregrep_exit() before it was defined. I am assuming this was
1073        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1074        reported by a user. I've moved the definition above the reference.
1075    
1076    
1077    Version 8.11 10-Dec-2010
1078    ------------------------
1079    
1080    1.  (*THEN) was not working properly if there were untried alternatives prior
1081        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1082        backtracked to try for "b" instead of moving to the next alternative branch
1083        at the same level (in this case, to look for "c"). The Perl documentation
1084        is clear that when (*THEN) is backtracked onto, it goes to the "next
1085        alternative in the innermost enclosing group".
1086    
1087    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1088        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1089        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1090        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1091        (*THEN).
1092    
1093    3.  If \s appeared in a character class, it removed the VT character from
1094        the class, even if it had been included by some previous item, for example
1095        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1096        of \s, but is part of the POSIX "space" class.)
1097    
1098    4.  A partial match never returns an empty string (because you can always
1099        match an empty string at the end of the subject); however the checking for
1100        an empty string was starting at the "start of match" point. This has been
1101        changed to the "earliest inspected character" point, because the returned
1102        data for a partial match starts at this character. This means that, for
1103        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1104        (previously it gave "no match").
1105    
1106    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1107        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1108        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1109        has an implication that the given string is incomplete (because a partial
1110        match is preferred over a full match). For this reason, these items now
1111        give a partial match in this situation. [Aside: previously, the one case
1112        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1113        match rather than a full match, which was wrong by the old rules, but is
1114        now correct.]
1115    
1116    6.  There was a bug in the handling of #-introduced comments, recognized when
1117        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1118        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1119        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1120        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1121        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1122        places in pcre_compile().
1123    
1124    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1125        comments when looking ahead for named forward references to subpatterns,
1126        the only newline sequence it recognized was NL. It now handles newlines
1127        according to the set newline convention.
1128    
1129    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1130        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1131        cater for a lack of strerror(). These oversights have been fixed.
1132    
1133    9.  Added --match-limit and --recursion-limit to pcregrep.
1134    
1135    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1136    
1137    11. When the -o option was used, pcregrep was setting a return code of 1, even
1138        when matches were found, and --line-buffered was not being honoured.
1139    
1140    12. Added an optional parentheses number to the -o and --only-matching options
1141        of pcregrep.
1142    
1143    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1144        can match an empty string. The code to do it in pcretest and pcredemo
1145        needed fixing:
1146    
1147        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1148            only one byte after an empty string match just before CRLF (this case
1149            just got forgotten; "any" and "anycrlf" were OK).
1150    
1151        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1152            mode when an empty string match preceded an ASCII character followed by
1153            a non-ASCII character. (The code for advancing by one character rather
1154            than one byte was nonsense.)
1155    
1156        (c) The pcredemo.c sample program did not have any code at all to handle
1157            the cases when CRLF is a valid newline sequence.
1158    
1159    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1160        as a starting offset was within the subject string. There is now a new
1161        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1162        negative or greater than the length of the string. In order to test this,
1163        pcretest is extended to allow the setting of negative starting offsets.
1164    
1165    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1166        starting offset points to the beginning of a UTF-8 character was
1167        unnecessarily clumsy. I tidied it up.
1168    
1169    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1170        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1171    
1172    17. Nobody had reported that the --include_dir option, which was added in
1173        release 7.7 should have been called --include-dir (hyphen, not underscore)
1174        for compatibility with GNU grep. I have changed it to --include-dir, but
1175        left --include_dir as an undocumented synonym, and the same for
1176        --exclude-dir, though that is not available in GNU grep, at least as of
1177        release 2.5.4.
1178    
1179    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1180        characters from a string of bytes) have been redefined so as not to use
1181        loops, in order to improve performance in some environments. At the same
1182        time, I abstracted some of the common code into auxiliary macros to save
1183        repetition (this should not affect the compiled code).
1184    
1185    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1186        compile-time error is now given if \c is not followed by an ASCII
1187        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1188        different, and any byte value is allowed.)
1189    
1190    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1191        START_OPTIMIZE option, which is now allowed at compile time - but just
1192        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1193        to pcregrep and other applications that have no direct access to PCRE
1194        options. The new /Y option in pcretest sets this option when calling
1195        pcre_compile().
1196    
1197    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1198        back references. Groups containing recursive back references were forced to
1199        be atomic by that change, but in the case of named groups, the amount of
1200        memory required was incorrectly computed, leading to "Failed: internal
1201        error: code overflow". This has been fixed.
1202    
1203    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1204        pcretest.c, to avoid build problems in some Borland environments.
1205    
1206    
1207    Version 8.10 25-Jun-2010
1208    ------------------------
1209    
1210    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1211        THEN.
1212    
1213    2.  (*ACCEPT) was not working when inside an atomic group.
1214    
1215    3.  Inside a character class, \B is treated as a literal by default, but
1216        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1217        causes the error). The code is unchanged, but I tidied the documentation.
1218    
1219    4.  Inside a character class, PCRE always treated \R and \X as literals,
1220        whereas Perl faults them if its -w option is set. I have changed PCRE so
1221        that it faults them when PCRE_EXTRA is set.
1222    
1223    5.  Added support for \N, which always matches any character other than
1224        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1225    
1226    6.  When compiling pcregrep with newer versions of gcc which may have
1227        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1228        declared with attribute warn_unused_result" were given. Just casting the
1229        result to (void) does not stop the warnings; a more elaborate fudge is
1230        needed. I've used a macro to implement this.
1231    
1232    7.  Minor change to pcretest.c to avoid a compiler warning.
1233    
1234    8.  Added four artifical Unicode properties to help with an option to make
1235        \s etc use properties (see next item). The new properties are: Xan
1236        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1237    
1238    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1239        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1240        this option. Modified pcretest to add /W to test this facility. Added
1241        REG_UCP to make it available via the POSIX interface.
1242    
1243    10. Added --line-buffered to pcregrep.
1244    
1245    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1246        studied, and the match started with a letter with a code point greater than
1247        127 whose first byte was different to the first byte of the other case of
1248        the letter, the other case of this starting letter was not recognized
1249        (#976).
1250    
1251    12. If a pattern that was studied started with a repeated Unicode property
1252        test, for example, \p{Nd}+, there was the theoretical possibility of
1253        setting up an incorrect bitmap of starting bytes, but fortunately it could
1254        not have actually happened in practice until change 8 above was made (it
1255        added property types that matched character-matching opcodes).
1256    
1257    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1258        possible starting bytes for non-anchored patterns.
1259    
1260    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1261        \R, and also a number of cases that involve Unicode properties, both
1262        explicit and implicit when PCRE_UCP is set.
1263    
1264    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1265        input, it could crash or give wrong results if characters with values
1266        greater than 0xc0 were present in the subject string. (Detail: it assumed
1267        UTF-8 input when processing these items.)
1268    
1269    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1270        size_t is 64-bit (#991).
1271    
1272    17. Added a check for running out of memory when PCRE is compiled with
1273        --disable-stack-for-recursion (#990).
1274    
1275    18. If the last data line in a file for pcretest does not have a newline on
1276        the end, a newline was missing in the output.
1277    
1278    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1279        less than 128) in its various bitmaps. However, there is a facility for
1280        generating tables according to the current locale when PCRE is compiled. It
1281        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1282        characters, are recognized by isspace() and therefore were getting set in
1283        these tables, and indeed these tables seem to approximate to ISO 8859. This
1284        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1285        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1286        which of course cannot start UTF-8 characters. I have changed the code so
1287        that only real ASCII characters (less than 128) and the correct starting
1288        bytes for UTF-8 encodings are set for characters greater than 127 when in
1289        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1290        altogether.)
1291    
1292    20. Added the /T option to pcretest so as to be able to run tests with non-
1293        standard character tables, thus making it possible to include the tests
1294        used for 19 above in the standard set of tests.
1295    
1296    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1297        reference to a subpattern the other side of a comment that contains an
1298        opening parenthesis caused either an internal compiling error, or a
1299        reference to the wrong subpattern.
1300    
1301    
1302    Version 8.02 19-Mar-2010
1303    ------------------------
1304    
1305    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1306    
1307    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1308        configured.
1309    
1310    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1311        original author of that file, following a query about its status.
1312    
1313    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1314        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1315    
1316    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1317        quantifier applied to a forward-referencing subroutine call, could compile
1318        incorrect code or give the error "internal error: previously-checked
1319        referenced subpattern not found".
1320    
1321    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1322        variables to point to external functions. For these systems, therefore,
1323        pcre_malloc etc. are now initialized to local functions that call the
1324        relevant global functions.
1325    
1326    7.  There were two entries missing in the vectors called coptable and poptable
1327        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1328        I've fixed the data, and added a kludgy way of testing at compile time that
1329        the lengths are correct (equal to the number of opcodes).
1330    
1331    8.  Following on from 7, I added a similar kludge to check the length of the
1332        eint vector in pcreposix.c.
1333    
1334    9.  Error texts for pcre_compile() are held as one long string to avoid too
1335        much relocation at load time. To find a text, the string is searched,
1336        counting zeros. There was no check for running off the end of the string,
1337        which could happen if a new error number was added without updating the
1338        string.
1339    
1340    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1341    
1342    11. \K was not working if it appeared in an atomic group or in a group that
1343        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1344        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1345        the assertion is positive, but not if it is negative.
1346    
1347    12. Change 11 fortuitously reduced the size of the stack frame used in the
1348        "match()" function of pcre_exec.c by one pointer. Forthcoming
1349        implementation of support for (*MARK) will need an extra pointer on the
1350        stack; I have reserved it now, so that the stack frame size does not
1351        decrease.
1352    
1353    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1354        item in branch that calls a recursion is a subroutine call - as in the
1355        second branch in the above example - was incorrectly given the compile-
1356        time error "recursive call could loop indefinitely" because pcre_compile()
1357        was not correctly checking the subroutine for matching a non-empty string.
1358    
1359    14. The checks for overrunning compiling workspace could trigger after an
1360        overrun had occurred. This is a "should never occur" error, but it can be
1361        triggered by pathological patterns such as hundreds of nested parentheses.
1362        The checks now trigger 100 bytes before the end of the workspace.
1363    
1364    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1365    
1366    
1367    Version 8.01 19-Jan-2010
1368    ------------------------
1369    
1370    1.  If a pattern contained a conditional subpattern with only one branch (in
1371        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1372        computed the wrong minimum data length (which is of course zero for such
1373        subpatterns). This could cause incorrect "no match" results.
1374    
1375    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1376        the pattern is reset in the first branch, pcre_compile() failed with
1377        "internal error: code overflow at offset...". This happened only when
1378        the reset was to the original external option setting. (An optimization
1379        abstracts leading options settings into an external setting, which was the
1380        cause of this.)
1381    
1382    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1383        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1384        assertion pattern did not match (meaning that the assertion was true), it
1385        was incorrectly treated as false if the SKIP had been reached during the
1386        matching. This also applied to assertions used as conditions.
1387    
1388    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1389        assertion subpattern, including such a pattern used as a condition,
1390        unpredictable results occurred, instead of the error return
1391        PCRE_ERROR_DFA_UITEM.
1392    
1393    5.  The C++ GlobalReplace function was not working like Perl for the special
1394        situation when an empty string is matched. It now does the fancy magic
1395        stuff that is necessary.
1396    
1397    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1398        removed. (These were left over from very, very early versions of PCRE.)
1399    
1400    7.  Some cosmetic changes to the code to make life easier when compiling it
1401        as part of something else:
1402    
1403        (a) Change DEBUG to PCRE_DEBUG.
1404    
1405        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1406            called "current" as "current_branch", to prevent a collision with the
1407            Linux macro when compiled as a kernel module.
1408    
1409        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1410            prevent a collision with the Linux macro when compiled as a kernel
1411            module.
1412    
1413    8.  In pcre_compile() there are some checks for integer overflows that used to
1414        cast potentially large values to (double). This has been changed to that
1415        when building, a check for int64_t is made, and if it is found, it is used
1416        instead, thus avoiding the use of floating point arithmetic. (There is no
1417        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1418        double.
1419    
1420    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1421        2005 (difference between two addresses compared to an unsigned value).
1422    
1423    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1424        custom one, because of the following reported problem in Windows:
1425    
1426          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1427              under Win32.
1428          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1429              therefore missing the function definition.
1430          - The compiler thus generates a "C" signature for the test function.
1431          - The linker fails to find the "C" function.
1432          - PCRE fails to configure if asked to do so against libbz2.
1433    
1434    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1435        messages were output:
1436    
1437          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1438          rerunning libtoolize, to keep the correct libtool macros in-tree.
1439          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1440    
1441        I have done both of these things.
1442    
1443    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1444        most of the time, it *can* run out if it is given a pattern that contains a
1445        runaway infinite recursion. I updated the discussion in the pcrestack man
1446        page.
1447    
1448    13. Now that we have gone to the x.xx style of version numbers, the minor
1449        version may start with zero. Using 08 or 09 is a bad idea because users
1450        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1451        interpreted as invalid octal numbers. I've updated the previous comment in
1452        configure.ac, and also added a check that gives an error if 08 or 09 are
1453        used.
1454    
1455    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1456        causing partial matching to fail when the end of the subject matched \W
1457        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1458    
1459    15. There were some discrepancies between the declarations in pcre_internal.h
1460        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1461        their definitions. The declarations used "const uschar *" and the
1462        definitions used USPTR. Even though USPTR is normally defined as "const
1463        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1464        reported that: "This difference in casting confuses some C++ compilers, for
1465        example, SunCC recognizes above declarations as different functions and
1466        generates broken code for hbpcre." I have changed the declarations to use
1467        USPTR.
1468    
1469    16. GNU libtool is named differently on some systems. The autogen.sh script now
1470        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1471        (FreeBSD).
1472    
1473    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1474        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1475        comment: "Figure out how to create a longlong from a string: strtoll and
1476        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1477        instance, but it only takes 2 args instead of 3!"
1478    
1479    18. A subtle bug concerned with back references has been fixed by a change of
1480        specification, with a corresponding code fix. A pattern such as
1481        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1482        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1483        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1484        same bug. Such groups have to be quantified to be useful, or contained
1485        inside another quantified group. (If there's no repetition, the reference
1486        can never match.) The problem arises because, having left the group and
1487        moved on to the rest of the pattern, a later failure that backtracks into
1488        the group uses the captured value from the final iteration of the group
1489        rather than the correct earlier one. I have fixed this in PCRE by forcing
1490        any group that contains a reference to itself to be an atomic group; that
1491        is, there cannot be any backtracking into it once it has completed. This is
1492        similar to recursive and subroutine calls.
1493    
1494    
1495    Version 8.00 19-Oct-09
1496    ----------------------
1497    
1498    1.  The table for translating pcre_compile() error codes into POSIX error codes
1499        was out-of-date, and there was no check on the pcre_compile() error code
1500        being within the table. This could lead to an OK return being given in
1501        error.
1502    
1503    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1504        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1505        in a Windows environment.
1506    
1507    3.  The pcregrep --count option prints the count for each file even when it is
1508        zero, as does GNU grep. However, pcregrep was also printing all files when
1509        --files-with-matches was added. Now, when both options are given, it prints
1510        counts only for those files that have at least one match. (GNU grep just
1511        prints the file name in this circumstance, but including the count seems
1512        more useful - otherwise, why use --count?) Also ensured that the
1513        combination -clh just lists non-zero counts, with no names.
1514    
1515    4.  The long form of the pcregrep -F option was incorrectly implemented as
1516        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1517        but it seems right to fix it, and I didn't think it was worth preserving
1518        the old behaviour.
1519    
1520    5.  The command line items --regex=pattern and --regexp=pattern were not
1521        recognized by pcregrep, which required --regex pattern or --regexp pattern
1522        (with a space rather than an '='). The man page documented the '=' forms,
1523        which are compatible with GNU grep; these now work.
1524    
1525    6.  No libpcreposix.pc file was created for pkg-config; there was just
1526        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1527    
1528    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1529        when UCP support is not needed, by modifying the Python script that
1530        generates it from Unicode data files. This should not matter if the module
1531        is correctly used as a library, but I received one complaint about 50K of
1532        unwanted data. My guess is that the person linked everything into his
1533        program rather than using a library. Anyway, it does no harm.
1534    
1535    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1536        was a minimum greater than 1 for a wide character in a possessive
1537        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1538        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1539        character. Chaos in the form of incorrect output or a compiling loop could
1540        result.
1541    
1542    9.  The restrictions on what a pattern can contain when partial matching is
1543        requested for pcre_exec() have been removed. All patterns can now be
1544        partially matched by this function. In addition, if there are at least two
1545        slots in the offset vector, the offset of the earliest inspected character
1546        for the match and the offset of the end of the subject are set in them when
1547        PCRE_ERROR_PARTIAL is returned.
1548    
1549    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1550        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1551        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1552        and may be more useful for multi-segment matching.
1553    
1554    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1555        used to be given if ever the end of the subject was reached; now it is
1556        given only if matching could not proceed because another character was
1557        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1558        string "Z", which now yields "no match" instead of "partial match". In the
1559        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1560        final character ended with (*FAIL).
1561    
1562    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1563        if the pattern had a "must contain" character that was already found in the
1564        earlier partial match, unless partial matching was again requested. For
1565        example, with the pattern /dog.(body)?/, the "must contain" character is
1566        "g". If the first part-match was for the string "dog", restarting with
1567        "sbody" failed. This bug has been fixed.
1568    
1569    13. The string returned by pcre_dfa_exec() after a partial match has been
1570        changed so that it starts at the first inspected character rather than the
1571        first character of the match. This makes a difference only if the pattern
1572        starts with a lookbehind assertion or \b or \B (\K is not supported by
1573        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1574        matching functions compatible, and I think it's the right thing to do.
1575    
1576    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1577        so that the demonstration program is easily available in environments where
1578        PCRE has not been installed from source.
1579    
1580    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1581        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1582        library.
1583    
1584    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1585        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1586        is not the first non-POSIX option to be added. Clearly some people find
1587        these options useful.
1588    
1589    17. If a caller to the POSIX matching function regexec() passes a non-zero
1590        value for nmatch with a NULL value for pmatch, the value of
1591        nmatch is forced to zero.
1592    
1593    18. RunGrepTest did not have a test for the availability of the -u option of
1594        the diff command, as RunTest does. It now checks in the same way as
1595        RunTest, and also checks for the -b option.
1596    
1597    19. If an odd number of negated classes containing just a single character
1598        interposed, within parentheses, between a forward reference to a named
1599        subpattern and the definition of the subpattern, compilation crashed with
1600        an internal error, complaining that it could not find the referenced
1601        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1602        [The bug was that it was starting one character too far in when skipping
1603        over the character class, thus treating the ] as data rather than
1604        terminating the class. This meant it could skip too much.]
1605    
1606    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1607        /g option in pcretest when the pattern contains \K, which makes it possible
1608        to have an empty string match not at the start, even when the pattern is
1609        anchored. Updated pcretest and pcredemo to use this option.
1610    
1611    21. If the maximum number of capturing subpatterns in a recursion was greater
1612        than the maximum at the outer level, the higher number was returned, but
1613        with unset values at the outer level. The correct (outer level) value is
1614        now given.
1615    
1616    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1617        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1618        make it do so. The string so far is captured, making this feature
1619        compatible with Perl.
1620    
1621    23. The tests have been re-organized, adding tests 11 and 12, to make it
1622        possible to check the Perl 5.10 features against Perl 5.10.
1623    
1624    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1625        pattern matches a fixed length string. PCRE did not allow this; now it
1626        does. Neither allows recursion.
1627    
1628    25. I finally figured out how to implement a request to provide the minimum
1629        length of subject string that was needed in order to match a given pattern.
1630        (It was back references and recursion that I had previously got hung up
1631        on.) This code has now been added to pcre_study(); it finds a lower bound
1632        to the length of subject needed. It is not necessarily the greatest lower
1633        bound, but using it to avoid searching strings that are too short does give
1634        some useful speed-ups. The value is available to calling programs via
1635        pcre_fullinfo().
1636    
1637    26. While implementing 25, I discovered to my embarrassment that pcretest had
1638        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1639        study optimizations had never been tested with that matching function.
1640        Oops. What is worse, even when it was passed study data, there was a bug in
1641        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1642        were also very few tests of studied patterns with pcre_dfa_exec().
1643    
1644    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1645        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1646        on the other side of the coin, they are no longer allowed to have different
1647        names, because these cannot be distinguished in PCRE, and this has caused
1648        confusion. (This is a difference from Perl.)
1649    
1650    28. When duplicate subpattern names are present (necessarily with different
1651        numbers, as required by 27 above), and a test is made by name in a
1652        conditional pattern, either for a subpattern having been matched, or for
1653        recursion in such a pattern, all the associated numbered subpatterns are
1654        tested, and the overall condition is true if the condition is true for any
1655        one of them. This is the way Perl works, and is also more like the way
1656        testing by number works.
1657    
1658    
1659    Version 7.9 11-Apr-09
1660    ---------------------
1661    
1662    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1663        (pcretest), all targets were linked against these libraries. This included
1664        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1665        libraries. This caused unwanted dependencies to be created. This problem
1666        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1667        pcretest is linked with readline.
1668    
1669    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1670        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1671        moved outside it again, because FALSE and TRUE are already defined in AIX,
1672        but BOOL is not.
1673    
1674    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1675        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1676    
1677    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1678        hyphens) following file names and line numbers when outputting matching
1679        lines. This is not true; no spaces are inserted. I have also clarified the
1680        wording for the --colour (or --color) option.
1681    
1682    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1683        was not coloured; this is different to GNU grep, so I have changed it to be
1684        the same.
1685    
1686    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1687        each matching line was coloured. Now it goes on to look for further matches
1688        of any of the test patterns, which is the same behaviour as GNU grep.
1689    
1690    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1691        doesn't make sense to accept an empty string match in pcregrep, so I have
1692        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1693        seems to be how GNU grep behaves. [But see later change 40 for release
1694        8.33.]
1695    
1696    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1697        start or after a newline", because the conditional assertion was not being
1698        correctly handled. The rule now is that both the assertion and what follows
1699        in the first alternative must satisfy the test.
1700    
1701    9.  If auto-callout was enabled in a pattern with a conditional group whose
1702        condition was an assertion, PCRE could crash during matching, both with
1703        pcre_exec() and pcre_dfa_exec().
1704    
1705    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1706        used for matching.
1707    
1708    11. Unicode property support in character classes was not working for
1709        characters (bytes) greater than 127 when not in UTF-8 mode.
1710    
1711    12. Added the -M command line option to pcretest.
1712    
1713    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1714    
1715    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1716    
1717    16. Added comments and documentation about mis-use of no_arg in the C++
1718        wrapper.
1719    
1720    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1721        from Martin Jerabek that uses macro names for all relevant character and
1722        string constants.
1723    
1724    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1725        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1726        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1727        these, but not everybody uses configure.
1728    
1729    19. A conditional group that had only one branch was not being correctly
1730        recognized as an item that could match an empty string. This meant that an
1731        enclosing group might also not be so recognized, causing infinite looping
1732        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1733        with the subject "ab", where knowledge that the repeated group can match
1734        nothing is needed in order to break the loop.
1735    
1736    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1737        exec(), but without supplying a callout function, matching went wrong.
1738    
1739    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1740        leak if the size of the offset vector was greater than 30. When the vector
1741        is smaller, the saved offsets during recursion go onto a local stack
1742        vector, but for larger vectors malloc() is used. It was failing to free
1743        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1744        error, in fact).
1745    
1746    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1747        heapframe that is used only when UTF-8 support is enabled. This caused no
1748        problem, but was untidy.
1749    
1750    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1751        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1752        included within another project.
1753    
1754    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1755        slightly modified by me:
1756    
1757          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1758              not building pcregrep.
1759    
1760          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1761              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1762    
1763    25. Forward references, both numeric and by name, in patterns that made use of
1764        duplicate group numbers, could behave incorrectly or give incorrect errors,
1765        because when scanning forward to find the reference group, PCRE was not
1766        taking into account the duplicate group numbers. A pattern such as
1767        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1768    
1769    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1770        the feature of a custom pointer more persuasive (as requested by a user).
1771    
1772    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1773        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1774        pre-defined.
1775    
1776    28. Added support for (*UTF8) at the start of a pattern.
1777    
1778    29. Arrange for flags added by the "release type" setting in CMake to be shown
1779        in the configuration summary.
1780    
1781    
1782    Version 7.8 05-Sep-08
1783    ---------------------
1784    
1785    1.  Replaced UCP searching code with optimized version as implemented for Ad
1786        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1787        stage table and inline lookup instead of a function, giving speed ups of 2
1788        to 5 times on some simple patterns that I tested. Permission was given to
1789        distribute the MultiStage2.py script that generates the tables (it's not in
1790        the tarball, but is in the Subversion repository).
1791    
1792    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1793        scripts.
1794    
1795    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1796        a group with a zero qualifier. The result of the study could be incorrect,
1797        or the function might crash, depending on the pattern.
1798    
1799    4.  Caseless matching was not working for non-ASCII characters in back
1800        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1801        It now works when Unicode Property Support is available.
1802    
1803    5.  In pcretest, an escape such as \x{de} in the data was always generating
1804        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1805        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1806        truncation.
1807    
1808    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1809    
1810    7.  Added two (int) casts to pcregrep when printing the difference of two
1811        pointers, in case they are 64-bit values.
1812    
1813    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1814        test 2 if it fails.
1815    
1816    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1817        and a #define of that name to empty if it is not externally set. This is to
1818        allow users of MSVC to set it if necessary.
1819    
1820    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1821        the convenience functions in the pcre_get.c source file.
1822    
1823    11. An option change at the start of a pattern that had top-level alternatives
1824        could cause overwriting and/or a crash. This command provoked a crash in
1825        some environments:
1826    
1827          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1828    
1829        This potential security problem was recorded as CVE-2008-2371.
1830    
1831    12. For a pattern where the match had to start at the beginning or immediately
1832        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1833        pcre_dfa_exec() could read past the end of the passed subject if there was
1834        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1835        pcretest so that it places the subject at the end of its malloc-ed buffer.
1836    
1837    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1838        exec() might read past the end of the data buffer in UTF-8 mode.
1839    
1840    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1841        the data contained the byte 0x85 as part of a UTF-8 character within its
1842        first line. This applied both to normal and DFA matching.
1843    
1844    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1845        /^[^d]*?$/8 failed to match "abc".
1846    
1847    16. Added a missing copyright notice to pcrecpp_internal.h.
1848    
1849    17. Make it more clear in the documentation that values returned from
1850        pcre_exec() in ovector are byte offsets, not character counts.
1851    
1852    18. Tidied a few places to stop certain compilers from issuing warnings.
1853    
1854    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1855        supplied by Stefan Weber. I made a further small update for 7.8 because
1856        there is a change of source arrangements: the pcre_searchfuncs.c module is
1857        replaced by pcre_ucd.c.
1858    
1859    
1860    Version 7.7 07-May-08
1861    ---------------------
1862    
1863    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1864        a string to a long long, pretend we don't even have a long long." This is
1865        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1866    
1867    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1868        pre-7.6 versions, which defined a global no_arg variable instead of putting
1869        it in the RE class. (See also #8 below.)
1870    
1871    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1872        Lopes.
1873    
1874    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1875    
1876        (1) The include/exclude patterns were being applied to the whole pathnames
1877            of files, instead of just to the final components.
1878    
1879        (2) If there was more than one level of directory, the subdirectories were
1880            skipped unless they satisfied the include/exclude conditions. This is
1881            inconsistent with GNU grep (and could even be seen as contrary to the
1882            pcregrep specification - which I improved to make it absolutely clear).
1883            The action now is always to scan all levels of directory, and just
1884            apply the include/exclude patterns to regular files.
1885    
1886    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1887        --exclude_dir in the tests to avoid scanning .svn directories.
1888    
1889    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1890        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1891        doesn't support NULs in patterns.
1892    
1893    7.  Added some missing "const"s to declarations of static tables in
1894        pcre_compile.c and pcre_dfa_exec.c.
1895    
1896    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1897        caused by fix #2  above. (Subsequently also a second patch to fix the
1898        first patch. And a third patch - this was a messy problem.)
1899    
1900    9.  Applied Craig's patch to remove the use of push_back().
1901    
1902    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1903        matching function regexec().
1904    
1905    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1906        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1907        references. PCRE supports relative numbers with this syntax (I don't think
1908        Oniguruma does).
1909    
1910    12. Previously, a group with a zero repeat such as (...){0} was completely
1911        omitted from the compiled regex. However, this means that if the group
1912        was called as a subroutine from elsewhere in the pattern, things went wrong
1913        (an internal error was given). Such groups are now left in the compiled
1914        pattern, with a new opcode that causes them to be skipped at execution
1915        time.
1916    
1917    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1918        to the way PCRE behaves:
1919    
1920        (a) A lone ] character is dis-allowed (Perl treats it as data).
1921    
1922        (b) A back reference to an unmatched subpattern matches an empty string
1923            (Perl fails the current match path).
1924    
1925        (c) A data ] in a character class must be notated as \] because if the
1926            first data character in a class is ], it defines an empty class. (In
1927            Perl it is not possible to have an empty class.) The empty class []
1928            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1929            The negative empty class [^] matches any one character, independently
1930            of the DOTALL setting.
1931    
1932    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1933        non-existent subpattern following a character class starting with ']' and
1934        containing () gave an internal compiling error instead of "reference to
1935        non-existent subpattern". Fortunately, when the pattern did exist, the
1936        compiled code was correct. (When scanning forwards to check for the
1937        existence of the subpattern, it was treating the data ']' as terminating
1938        the class, so got the count wrong. When actually compiling, the reference
1939        was subsequently set up correctly.)
1940    
1941    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1942        it was being rejected as not supported by pcre_dfa_exec(), even though
1943        other assertions are supported. I have made pcre_dfa_exec() support
1944        (*FAIL).
1945    
1946    16. The implementation of 13c above involved the invention of a new opcode,
1947        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1948        cannot be changed at match time, I realized I could make a small
1949        improvement to matching performance by compiling OP_ALLANY instead of
1950        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1951        on the OP_ANY path.
1952    
1953    17. Compiling pcretest on Windows with readline support failed without the
1954        following two fixes: (1) Make the unistd.h include conditional on
1955        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1956    
1957    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1958        ncurses library to be included for pcretest when ReadLine support is
1959        requested, but also to allow for it to be overridden. This patch came from
1960        Daniel Bergström.
1961    
1962    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1963        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1964        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1965        spotting this.
1966    
1967    
1968    Version 7.6 28-Jan-08
1969    ---------------------
1970    
1971    1.  A character class containing a very large number of characters with
1972        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1973        overflow.
1974    
1975    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1976        HAVE_LONG_LONG is not defined.
1977    
1978    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1979        bring them up to date and include new features. This patch includes:
1980    
1981        - Fixed PH's badly added libz and libbz2 support.
1982        - Fixed a problem with static linking.
1983        - Added pcredemo. [But later removed - see 7 below.]
1984        - Fixed dftables problem and added an option.
1985        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1986            HAVE_LONG_LONG.
1987        - Added readline support for pcretest.
1988        - Added an listing of the option settings after cmake has run.
1989    
1990    4.  A user submitted a patch to Makefile that makes it easy to create
1991        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1992        Makefile.am that cause it to include this special target, without
1993        affecting anything else. Note that the same mingw target plus all
1994        the other distribution libraries and programs are now supported
1995        when configuring with CMake (see 6 below) instead of with
1996        Configure/Make.
1997    
1998    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1999        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2000        exported in the Windows port". It has not yet been confirmed that the patch
2001        solves the problem, but it does no harm.
2002    
2003    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2004        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2005        with CMake, and also correct the comment about stack recursion.
2006    
2007    7.  Remove the automatic building of pcredemo from the ./configure system and
2008        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2009        of a program that users should build themselves after PCRE is installed, so
2010        building it automatically is not really right. What is more, it gave
2011        trouble in some build environments.
2012    
2013    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2014    
2015    
2016    Version 7.5 10-Jan-08
2017    ---------------------
2018    
2019    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2020        values in parens when parsing an RE using the C++ wrapper."
2021    
2022    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2023        Characters greater than 255 were excluded from the class instead of being
2024        included.
2025    
2026    3.  The same bug as (2) above applied to negated POSIX classes such as
2027        [:^space:].
2028    
2029    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2030        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2031        I have changed it.
2032    
2033    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2034        first named subpattern) and a construct such as (?&a) would reference the
2035        first named subpattern whose name started with "a" (in other words, the
2036        length check was missing). Both these problems are fixed. "Subpattern name
2037        expected" is now given for (?&) (a zero-length name), and this patch also
2038        makes it give the same error for \k'' (previously it complained that that
2039        was a reference to a non-existent subpattern).
2040    
2041    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2042        this is right because (?- can be followed by option settings as well as by
2043        digits. I have, however, made the messages clearer.
2044    
2045    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2046        than the number used in the conditional) now cause a compile-time error.
2047        This is actually not compatible with Perl, which accepts such patterns, but
2048        treats the conditional as always being FALSE (as PCRE used to), but it
2049        seems to me that giving a diagnostic is better.
2050    
2051    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2052        and messages.
2053    
2054    9.  Fix two occurrences of "backslash" in comments that should have been
2055        "backspace".
2056    
2057    10. Remove two redundant lines of code that can never be obeyed (their function
2058        was moved elsewhere).
2059    
2060    11. The program that makes PCRE's Unicode character property table had a bug
2061        which caused it to generate incorrect table entries for sequences of
2062        characters that have the same character type, but are in different scripts.
2063        It amalgamated them into a single range, with the script of the first of
2064        them. In other words, some characters were in the wrong script. There were
2065        thirteen such cases, affecting characters in the following ranges:
2066    
2067          U+002b0 - U+002c1
2068          U+0060c - U+0060d
2069          U+0061e - U+00612
2070          U+0064b - U+0065e
2071          U+0074d - U+0076d
2072          U+01800 - U+01805
2073          U+01d00 - U+01d77
2074          U+01d9b - U+01dbf
2075          U+0200b - U+0200f
2076          U+030fc - U+030fe
2077          U+03260 - U+0327f
2078          U+0fb46 - U+0fbb1
2079          U+10450 - U+1049d
2080    
2081    12. The -o option (show only the matching part of a line) for pcregrep was not
2082        compatible with GNU grep in that, if there was more than one match in a
2083        line, it showed only the first of them. It now behaves in the same way as
2084        GNU grep.
2085    
2086    13. If the -o and -v options were combined for pcregrep, it printed a blank
2087        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2088        does the same. The return code can be used to tell if there were any
2089        non-matching lines.
2090    
2091    14. Added --file-offsets and --line-offsets to pcregrep.
2092    
2093    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2094        infinitely looping recursion. The bug was that positive lookaheads were not
2095        being skipped when checking for a possible empty match (negative lookaheads
2096        and both kinds of lookbehind were skipped).
2097    
2098    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2099        inclusion of <windows.h> to before rather than after the definition of
2100        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2101    
2102    17. Specifying a possessive quantifier with a specific limit for a Unicode
2103        character property caused pcre_compile() to compile bad code, which led at
2104        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2105        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2106        caused the error; without that there was no problem.
2107    
2108    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2109    
2110    19. Added --enable-pcretest-libreadline.
2111    
2112    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2113        RE::GlobalReplace(). As a result, the number of replacements returned was
2114        double what it should be. I removed one of the increments, but Craig sent a
2115        later patch that removed the other one (the right fix) and added unit tests
2116        that check the return values (which was not done before).
2117    
2118    21. Several CMake things:
2119    
2120        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2121            the names libpcre and libpcreposix, not just pcre and pcreposix.
2122    
2123        (2) The above change means that pcretest and pcregrep are now correctly
2124            linked with the newly-built libraries, not previously installed ones.
2125    
2126        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2127    
2128    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2129        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2130        UTF-8 newline character). The key issue is that the pattern starts .*;
2131        this means that the match must be either at the beginning, or after a
2132        newline. The bug was in the code for advancing after a failed match and
2133        checking that the new position followed a newline. It was not taking
2134        account of UTF-8 characters correctly.
2135    
2136    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2137        character classes. PCRE was not treating the sequence [:...:] as a
2138        character class unless the ... were all letters. Perl, however, seems to
2139        allow any characters between [: and :], though of course it rejects as
2140        unknown any "names" that contain non-letters, because all the known class
2141        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2142        for example, whereas PCRE did not - it did not recognize a POSIX character
2143        class. This seemed a bit dangerous, so the code has been changed to be
2144        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2145        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2146        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2147        Perl does, and where it didn't before.
2148    
2149    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2150        Windows environments %n is disabled by default.
2151    
2152    
2153    Version 7.4 21-Sep-07
2154    ---------------------
2155    
2156    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2157        means that a class such as [\s] counted as "explicit reference to CR or
2158        LF". That isn't really right - the whole point of the change was to try to
2159        help when there was an actual mention of one of the two characters. So now
2160        the change happens only if \r or \n (or a literal CR or LF) character is
2161        encountered.
2162    
2163    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2164        of both had grown to the point where there were only 3 bits left.
2165        Fortunately, there was spare space in the data structure, and so I have
2166        moved the internal flags into a new 16-bit field to free up more option
2167        bits.
2168    
2169    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2170        but did not set the internal JCHANGED flag - either of these is enough to
2171        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2172        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2173        start sets both bits.
2174    
2175    4.  Added options (at build time, compile time, exec time) to change \R from
2176        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2177    
2178    5.  doc/pcresyntax.html was missing from the distribution.
2179    
2180    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2181        compatibility, even though it is no longer used.
2182    
2183    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2184        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2185        windows.h file is present (where different names are used). [This was
2186        reversed later after testing - see 16 below.]
2187    
2188    8.  Changed all #include <config.h> to #include "config.h". There were also
2189        some further <pcre.h> cases that I changed to "pcre.h".
2190    
2191    9.  When pcregrep was used with the --colour option, it missed the line ending
2192        sequence off the lines that it output.
2193    
2194    10. It was pointed out to me that arrays of string pointers cause lots of
2195        relocations when a shared library is dynamically loaded. A technique of
2196        using a single long string with a table of offsets can drastically reduce
2197        these. I have refactored PCRE in four places to do this. The result is
2198        dramatic:
2199    
2200          Originally:                          290
2201          After changing UCP table:            187
2202          After changing error message table:   43
2203          After changing table of "verbs"       36
2204          After changing table of Posix names   22
2205    
2206        Thanks to the folks working on Gregex for glib for this insight.
2207    
2208    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2209        unicode-properties was also set.
2210    
2211    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2212    
2213    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2214        checked only for CRLF.
2215    
2216    14. Added casts to pcretest.c to avoid compiler warnings.
2217    
2218    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2219    
2220    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2221        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2222        entirely. This removes changes made in 7 above.
2223    
2224    17. The CMake files have been updated, and there is now more information about
2225        building with CMake in the NON-UNIX-USE document.
2226    
2227    
2228    Version 7.3 28-Aug-07
2229    ---------------------
2230    
2231     1. In the rejigging of the build system that eventually resulted in 7.1, the
2232        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2233        brackets there is not right, since it causes compilers to look for an
2234        installed pcre.h, not the version that is in the source that is being
2235        compiled (which of course may be different). I have changed it back to:
2236    
2237          #include "pcre.h"
2238    
2239        I have a vague recollection that the change was concerned with compiling in
2240        different directories, but in the new build system, that is taken care of
2241        by the VPATH setting the Makefile.
2242    
2243     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2244        when the subject happened to end in the byte 0x85 (e.g. if the last
2245        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2246        characters but of course it shouldn't be taken as a newline when it is part
2247        of another character. The bug was that, for an unlimited repeat of . in
2248        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2249        characters when looking for a newline.
2250    
2251     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2252    
2253     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2254        in debug output.
2255    
2256     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2257        long printing in the pcrecpp unittest when running under MinGW.
2258    
2259     6. ESC_K was left out of the EBCDIC table.
2260    
2261     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2262        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2263        limit also applies to "virtual nesting" when a pattern is recursive, and in
2264        this case 1000 isn't so big. I have been able to remove this limit at the
2265        expense of backing off one optimization in certain circumstances. Normally,
2266        when pcre_exec() would call its internal match() function recursively and
2267        immediately return the result unconditionally, it uses a "tail recursion"
2268        feature to save stack. However, when a subpattern that can match an empty
2269        string has an unlimited repetition quantifier, it no longer makes this
2270        optimization. That gives it a stack frame in which to save the data for
2271        checking that an empty string has been matched. Previously this was taken
2272        from the 1000-entry workspace that had been reserved. So now there is no
2273        explicit limit, but more stack is used.
2274    
2275     8. Applied Daniel's patches to solve problems with the import/export magic
2276        syntax that is required for Windows, and which was going wrong for the
2277        pcreposix and pcrecpp parts of the library. These were overlooked when this
2278        problem was solved for the main library.
2279    
2280     9. There were some crude static tests to avoid integer overflow when computing
2281        the size of patterns that contain repeated groups with explicit upper
2282        limits. As the maximum quantifier is 65535, the maximum group length was
2283        set at 30,000 so that the product of these two numbers did not overflow a
2284        32-bit integer. However, it turns out that people want to use groups that
2285        are longer than 30,000 bytes (though not repeat them that many times).
2286        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2287        made it possible to implement the integer overflow checks in a much more
2288        dynamic way, which I have now done. The artificial limitation on group
2289        length has been removed - we now have only the limit on the total length of
2290        the compiled pattern, which depends on the LINK_SIZE setting.
2291    
2292    10. Fixed a bug in the documentation for get/copy named substring when
2293        duplicate names are permitted. If none of the named substrings are set, the
2294        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2295        empty string.
2296    
2297    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2298        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2299        because the ] is interpreted as the first data character and the
2300        terminating ] is not found. PCRE has been made compatible with Perl in this
2301        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2302        cause memory overwriting.
2303    
2304    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2305        string has been matched (to stop an infinite loop). It was not recognizing
2306        a conditional subpattern that could match an empty string if that
2307        subpattern was within another subpattern. For example, it looped when
2308        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2309        condition was not nested. This bug has been fixed.
2310    
2311    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2312        past the start of the subject in the presence of bytes with the top bit
2313        set, for example "\x8aBCD".
2314    
2315    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2316        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2317    
2318    14. Optimized (?!) to (*FAIL).
2319    
2320    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2321        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2322        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2323        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2324        does: it's just the validity check that is more restrictive.
2325    
2326    16. Inserted checks for integer overflows during escape sequence (backslash)
2327        processing, and also fixed erroneous offset values for syntax errors during
2328        backslash processing.
2329    
2330    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2331        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2332    
2333    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2334        caused an overrun.
2335    
2336    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2337        something other than just ASCII characters) inside a group that had an
2338        unlimited repeat caused a loop at compile time (while checking to see
2339        whether the group could match an empty string).
2340    
2341    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2342        [\P{Any}] did so. (Error in the code for printing property names.)
2343    
2344    21. An orphan \E inside a character class could cause a crash.
2345    
2346    22. A repeated capturing bracket such as (A)? could cause a wild memory
2347        reference during compilation.
2348    
2349    23. There are several functions in pcre_compile() that scan along a compiled
2350        expression for various reasons (e.g. to see if it's fixed length for look
2351        behind). There were bugs in these functions when a repeated \p or \P was
2352        present in the pattern. These operators have additional parameters compared
2353        with \d, etc, and these were not being taken into account when moving along
2354        the compiled data. Specifically:
2355    
2356        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2357            length.
2358    
2359        (b) An item such as \pL+ within a repeated group could cause crashes or
2360            loops.
2361    
2362        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2363            "reference to non-existent subpattern" error.
2364    
2365        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2366    
2367    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2368        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2369    
2370    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2371    
2372    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2373        character were causing crashes (broken optimization).
2374    
2375    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2376        \p or \P) caused a compile-time loop.
2377    
2378    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2379        break. For example, the unstudied pattern [\r\n]A does not match the string
2380        "\r\nA" because change 7.0/46 below moves the current point on by two
2381        characters after failing to match at the start. However, the pattern \nA
2382        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2383        the same is true. There doesn't seem any very clean way out of this, but
2384        what I have chosen to do makes the common cases work: PCRE now takes note
2385        of whether there can be an explicit match for \r or \n anywhere in the
2386        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2387        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2388        pattern has explicit CR or LF references.
2389    
2390    29. Added (*CR) etc for changing newline setting at start of pattern.
2391    
2392    
2393    Version 7.2 19-Jun-07
2394    ---------------------
2395    
2396     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2397        which is apparently normally available under Windows.
2398    
2399     2. Re-jig the pcregrep tests with different newline settings in an attempt
2400        to make them independent of the local environment's newline setting.
2401    
2402     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2403    
2404     4. Some of the "internals" tests were previously cut out when the link size
2405        was not 2, because the output contained actual offsets. The recent new
2406        "Z" feature of pcretest means that these can be cut out, making the tests
2407        usable with all link sizes.
2408    
2409     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2410        stack recursion. This gives a massive performance boost under BSD, but just
2411        a small improvement under Linux. However, it saves one field in the frame
2412        in all cases.
2413    
2414     6. Added more features from the forthcoming Perl 5.10:
2415    
2416        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2417            recursion call. It refers to the nth most recently opened parentheses.
2418    
2419        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2420            to be opened parentheses.
2421    
2422        (c) Conditions that refer to capturing parentheses can be specified
2423            relatively, for example, (?(-2)... or (?(+3)...
2424    
2425        (d) \K resets the start of the current match so that everything before
2426            is not part of it.
2427    
2428        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2429    
2430        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2431            reference syntax.
2432    
2433        (g) (?| introduces a group in which the numbering of parentheses in each
2434            alternative starts with the same number.
2435    
2436        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2437    
2438     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2439        PCRE_INFO_JCHANGED.
2440    
2441     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2442        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2443        for detecting groups that can match an empty string.
2444    
2445     9. A pattern with a very large number of alternatives (more than several
2446        hundred) was running out of internal workspace during the pre-compile
2447        phase, where pcre_compile() figures out how much memory will be needed. A
2448        bit of new cunning has reduced the workspace needed for groups with
2449        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2450        workspace instead of running out of the 4096 that are available.
2451    
2452    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2453    
2454    11. Applied patch from Google to remove an optimization that didn't quite work.
2455        The report of the bug said:
2456    
2457          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2458          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2459          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2460    
2461    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2462        it matched the wrong number of bytes.
2463    
2464    
2465    Version 7.1 24-Apr-07
2466    ---------------------
2467    
2468     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2469        that is more "standard", making use of automake and other Autotools. There
2470        is some re-arrangement of the files and adjustment of comments consequent
2471        on this.
2472    
2473     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2474        for recursive directory scanning broke on some systems because the files
2475        are not scanned in any specific order and on different systems the order
2476        was different. A call to "sort" has been inserted into RunGrepTest for the
2477        approprate test as a short-term fix. In the longer term there may be an
2478        alternative.
2479    
2480     3. I had an email from Eric Raymond about problems translating some of PCRE's
2481        man pages to HTML (despite the fact that I distribute HTML pages, some
2482        people do their own conversions for various reasons). The problems
2483        concerned the use of low-level troff macros .br and .in. I have therefore
2484        removed all such uses from the man pages (some were redundant, some could
2485        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2486        HTML has been updated to handle .nf/.fi and to complain if it encounters
2487        .br or .in.
2488    
2489     4. Updated comments in configure.ac that get placed in config.h.in and also
2490        arranged for config.h to be included in the distribution, with the name
2491        config.h.generic, for the benefit of those who have to compile without
2492        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2493    
2494     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2495        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2496        makevp.bat for the current PCRE, using the additional files
2497        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2498    
2499     6. A Windows user reported a minor discrepancy with test 2, which turned out
2500        to be caused by a trailing space on an input line that had got lost in his
2501        copy. The trailing space was an accident, so I've just removed it.
2502    
2503     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2504        that is needed.
2505    
2506     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2507        as "const" (a) because they are and (b) because it helps the PHP
2508        maintainers who have recently made a script to detect big data structures
2509        in the php code that should be moved to the .rodata section. I remembered
2510        to update Builducptable as well, so it won't revert if ucptable.h is ever
2511        re-created.
2512    
2513     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2514        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2515        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2516        support is not required. This saves 1.5-2K of code, which is important in
2517        some applications.
2518    
2519        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2520        so as not to refer to the tables, even though these functions will never be
2521        called when UTF-8 support is disabled. Otherwise there are problems with a
2522        shared library.
2523    
2524    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2525    
2526        (a) It was defining its arguments as char * instead of void *.
2527    
2528        (b) It was assuming that all moves were upwards in memory; this was true
2529            a long time ago when I wrote it, but is no longer the case.
2530    
2531        The emulated memove() is provided for those environments that have neither
2532        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2533        is clearly not the case, as these two bugs were recently reported.
2534    
2535    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2536        and Detrail to create the HTML documentation, the .txt form of the man
2537        pages, and it removes trailing spaces from listed files. It also creates
2538        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2539        case, it wraps all the #defines with #ifndefs. This script should be run
2540        before "make dist".
2541    
2542    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2543        with Unicode property support.
2544    
2545        (a) For a maximizing quantifier, if the two different cases of the
2546            character were of different lengths in their UTF-8 codings (there are
2547            some cases like this - I found 11), and the matching function had to
2548            back up over a mixture of the two cases, it incorrectly assumed they
2549            were both the same length.
2550    
2551        (b) When PCRE was configured to use the heap rather than the stack for
2552            recursion during matching, it was not correctly preserving the data for
2553            the other case of a UTF-8 character when checking ahead for a match
2554            while processing a minimizing repeat. If the check also involved
2555            matching a wide character, but failed, corruption could cause an
2556            erroneous result when trying to check for a repeat of the original
2557            character.
2558    
2559    13. Some tidying changes to the testing mechanism:
2560    
2561        (a) The RunTest script now detects the internal link size and whether there
2562            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2563            values substituted by "configure". (The RunGrepTest script already did
2564            this for UTF-8.) The configure.ac script no longer substitutes the
2565            relevant variables.
2566    
2567        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2568            with length and offset values. This means that the output is different
2569            for different internal link sizes. Test 2 is skipped for link sizes
2570            other than 2 because of this, bypassing the problem. Unfortunately,
2571            there was also a test in test 3 (the locale tests) that used /B and
2572            failed for link sizes other than 2. Rather than cut the whole test out,
2573            I have added a new /Z option to pcretest that replaces the length and
2574            offset values with spaces. This is now used to make test 3 independent
2575            of link size. (Test 2 will be tidied up later.)
2576    
2577    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2578        segmentation fault instead of returning the appropriate error message.
2579    
2580    15. In multiline mode when the newline sequence was set to "any", the pattern
2581        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2582        This doesn't seem right; it now treats the CRLF combination as the line
2583        ending, and so does not match in that case. It's only a pattern such as ^$
2584        that would hit this one: something like ^ABC$ would have failed after \r
2585        and then tried again after \r\n.
2586    
2587    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2588        in an attempt to make files that differ only in their line terminators
2589        compare equal. This works on Linux.
2590    
2591    17. Under certain error circumstances pcregrep might try to free random memory
2592        as it exited. This is now fixed, thanks to valgrind.
2593    
2594    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2595        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2596        was because its rules for how to advance for /g after matching an empty
2597        string at the end of a line did not allow for this case. They now check for
2598        it specially.
2599    
2600    20. pcretest is supposed to handle patterns and data of any length, by
2601        extending its buffers when necessary. It was getting this wrong when the
2602        buffer for a data line had to be extended.
2603    
2604    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2605        CRLF as a newline sequence.
2606    
2607    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2608        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2609        I have nevertheless tidied it up.
2610    
2611    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2612    
2613    24. Added a man page for pcre-config.
2614    
2615    
2616    Version 7.0 19-Dec-06
2617    ---------------------
2618    
2619     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2620        moving to gcc 4.1.1.
2621    
2622     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2623        sys/time.h, which is documented as needed for this function. It doesn't
2624        seem to matter on Linux, but it showed up on some releases of OS X.
2625    
2626     3. It seems that there are systems where bytes whose values are greater than
2627        127 match isprint() in the "C" locale. The "C" locale should be the
2628        default when a C program starts up. In most systems, only ASCII printing
2629        characters match isprint(). This difference caused the output from pcretest
2630        to vary, making some of the tests fail. I have changed pcretest so that:
2631    
2632        (a) When it is outputting text in the compiled version of a pattern, bytes
2633            other than 32-126 are always shown as hex escapes.
2634    
2635        (b) When it is outputting text that is a matched part of a subject string,
2636            it does the same, unless a different locale has been set for the match
2637            (using the /L modifier). In this case, it uses isprint() to decide.
2638    
2639     4. Fixed a major bug that caused incorrect computation of the amount of memory
2640        required for a compiled pattern when options that changed within the
2641        pattern affected the logic of the preliminary scan that determines the
2642        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2643        that the computed length was too small. The symptoms of this bug were
2644        either the PCRE error "internal error: code overflow" from pcre_compile(),
2645        or a glibc crash with a message such as "pcretest: free(): invalid next
2646        size (fast)". Examples of patterns that provoked this bug (shown in
2647        pcretest format) are:
2648    
2649          /(?-x: )/x
2650          /(?x)(?-x: \s*#\s*)/
2651          /((?i)[\x{c0}])/8
2652          /(?i:[\x{c0}])/8
2653    
2654        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2655        is now done differently.
2656    
2657     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2658        wrapper classes; (b) implement a new function in the C++ scanner that is
2659        more efficient than the old way of doing things because it avoids levels of
2660        recursion in the regex matching; (c) add a paragraph to the documentation
2661        for the FullMatch() function.
2662    
2663     6. The escape sequence \n was being treated as whatever was defined as
2664        "newline". Not only was this contrary to the documentation, which states
2665        that \n is character 10 (hex 0A), but it also went horribly wrong when
2666        "newline" was defined as CRLF. This has been fixed.
2667    
2668     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2669        was being set to -1 for the "end of line" case (supposedly a value that no
2670        character can have). Though this value is never used (the check for end of
2671        line is "zero bytes in current character"), it caused compiler complaints.
2672        I've changed it to 0xffffffff.
2673    
2674     8. In pcre_version.c, the version string was being built by a sequence of
2675        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2676        string (as it is for production releases) called a macro with an empty
2677        argument. The C standard says the result of this is undefined. The gcc
2678        compiler treats it as an empty string (which was what was wanted) but it is
2679        reported that Visual C gives an error. The source has been hacked around to
2680        avoid this problem.
2681    
2682     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2683        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2684        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2685        of them did).
2686    
2687    10. Originally, pcretest opened its input and output without "b"; then I was
2688        told that "b" was needed in some environments, so it was added for release
2689        5.0 to both the input and output. (It makes no difference on Unix-like
2690        systems.) Later I was told that it is wrong for the input on Windows. I've
2691        now abstracted the modes into two macros, to make it easier to fiddle with
2692        them, and removed "b" from the input mode under Windows.
2693    
2694    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2695    
2696    12. Added -help and --help to pcretest as an official way of being reminded
2697        of the options.
2698    
2699    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2700        and pcrecpp.cc because they annoy compilers at high warning levels.
2701    
2702    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2703    
2704    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2705        scripts are not C programs :-) and which was not noticed because it works
2706        on Linux.
2707    
2708    16. pcretest is supposed to handle any length of pattern and data line (as one
2709        line or as a continued sequence of lines) by extending its input buffer if
2710        necessary. This feature was broken for very long pattern lines, leading to
2711        a string of junk being passed to pcre_compile() if the pattern was longer
2712        than about 50K.
2713    
2714    17. I have done a major re-factoring of the way pcre_compile() computes the
2715        amount of memory needed for a compiled pattern. Previously, there was code
2716        that made a preliminary scan of the pattern in order to do this. That was
2717        OK when PCRE was new, but as the facilities have expanded, it has become
2718        harder and harder to keep it in step with the real compile phase, and there
2719        have been a number of bugs (see for example, 4 above). I have now found a
2720        cunning way of running the real compile function in a "fake" mode that
2721        enables it to compute how much memory it would need, while actually only
2722        ever using a few hundred bytes of working memory and without too many
2723        tests of the mode. This should make future maintenance and development
2724        easier. A side effect of this work is that the limit of 200 on the nesting
2725        depth of parentheses has been removed (though this was never a serious
2726        limitation, I suspect). However, there is a downside: pcre_compile() now
2727        runs more slowly than before (30% or more, depending on the pattern). I
2728        hope this isn't a big issue. There is no effect on runtime performance.
2729    
2730    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2731        newline (only possible for the last line of a file) and it was a
2732        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2733    
2734    19. Added additional timing features to pcretest. (1) The -tm option now times
2735        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2736        separate command line item, by a number that specifies the number of
2737        repeats to use when timing. The default is 50000; this gives better
2738        precision, but takes uncomfortably long for very large patterns.
2739    
2740    20. Extended pcre_study() to be more clever in cases where a branch of a
2741        subpattern has no definite first character. For example, (a*|b*)[cd] would
2742        previously give no result from pcre_study(). Now it recognizes that the
2743        first character must be a, b, c, or d.
2744    
2745    21. There was an incorrect error "recursive call could loop indefinitely" if
2746        a subpattern (or the entire pattern) that was being tested for matching an
2747        empty string contained only one non-empty item after a nested subpattern.
2748        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2749        incorrectly, because the \d was being skipped in the check.
2750    
2751    22. The pcretest program now has a new pattern option /B and a command line
2752        option -b, which is equivalent to adding /B to every pattern. This causes
2753        it to show the compiled bytecode, without the additional information that
2754        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2755        is the same as /B/I).
2756    
2757    23. A new optimization is now able automatically to treat some sequences such
2758        as a*b as a*+b. More specifically, if something simple (such as a character
2759        or a simple class like \d) has an unlimited quantifier, and is followed by
2760        something that cannot possibly match the quantified thing, the quantifier
2761        is automatically "possessified".
2762    
2763    24. A recursive reference to a subpattern whose number was greater than 39
2764        went wrong under certain circumstances in UTF-8 mode. This bug could also
2765        have affected the operation of pcre_study().
2766    
2767    25. Realized that a little bit of performance could be had by replacing
2768        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2769    
2770    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2771    
2772    27. Possessive quantifiers such as a++ were previously implemented by turning
2773        them into atomic groups such as ($>a+). Now they have their own opcodes,
2774        which improves performance. This includes the automatically created ones
2775        from 23 above.
2776    
2777    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2778        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2779        the first matched character to be a colon. This applied both to named and
2780        numbered groups.
2781    
2782    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2783    
2784    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2785        building PCRE on Windows easier, so I have included it in the distribution.
2786    
2787    31. There is now a check in pcretest against a ridiculously large number being
2788        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2789        loop, the loop is abandoned.
2790    
2791    32. Forward references to subpatterns in conditions such as (?(2)...) where
2792        subpattern 2 is defined later cause pcre_compile() to search forwards in
2793        the pattern for the relevant set of parentheses. This search went wrong
2794        when there were unescaped parentheses in a character class, parentheses
2795        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2796    
2797    33. "Subroutine" calls and backreferences were previously restricted to
2798        referencing subpatterns earlier in the regex. This restriction has now
2799        been removed.
2800    
2801    34. Added a number of extra features that are going to be in Perl 5.10. On the
2802        whole, these are just syntactic alternatives for features that PCRE had
2803        previously implemented using the Python syntax or my own invention. The
2804        other formats are all retained for compatibility.
2805    
2806        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2807            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2808            also .NET compatible.
2809    
2810        (b) A recursion or subroutine call to a named group can now be defined as
2811            (?&name) as well as (?P>name).
2812    
2813        (c) A backreference to a named group can now be defined as \k<name> or
2814            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2815            5.10, are also .NET compatible.
2816    
2817        (d) A conditional reference to a named group can now use the syntax
2818            (?(<name>) or (?('name') as well as (?(name).
2819    
2820        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2821            groups (named and numbered) that are never evaluated inline, but can be
2822            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2823            is always false. There may be only one alternative in such a group.
2824    
2825        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2826            as the simple (?(R). The condition is true only if the most recent
2827            recursion is that of the given number or name. It does not search out
2828            through the entire recursion stack.
2829    
2830        (g) The escape \gN or \g{N} has been added, where N is a positive or
2831            negative number, specifying an absolute or relative reference.
2832    
2833    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2834        some "unreachable code" warnings.
2835    
2836    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2837        things, this adds five new scripts.
2838    
2839    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2840        There were also incompatibilities regarding the handling of \Q..\E inside
2841        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2842        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2843    
2844    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2845        matches an empty string, and forcibly breaks the loop. There were bugs in
2846        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2847        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2848        separate and independent bugs (that affected different cases) have been
2849        fixed.
2850    
2851    39. Refactored the code to abolish the use of different opcodes for small
2852        capturing bracket numbers. This is a tidy that I avoided doing when I
2853        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2854        The new approach is not only tidier, it makes it possible to reduce the
2855        memory needed to fix the previous bug (38).
2856    
2857    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2858        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2859        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2860        mode.
2861    
2862    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2863        report.
2864    
2865    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2866        copy construction and assignment in the C++ wrapper.
2867    
2868    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2869        couple of bugs that could have given wrong results in the "--newline=crlf"
2870        case.
2871    
2872    44. Added a number of casts and did some reorganization of signed/unsigned int
2873        variables following suggestions from Dair Grant. Also renamed the variable
2874        "this" as "item" because it is a C++ keyword.
2875    
2876    45. Arranged for dftables to add
2877    
2878          #include "pcre_internal.h"
2879    
2880        to pcre_chartables.c because without it, gcc 4.x may remove the array
2881        definition from the final binary if PCRE is built into a static library and
2882        dead code stripping is activated.
2883    
2884    46. For an unanchored pattern, if a match attempt fails at the start of a
2885        newline sequence, and the newline setting is CRLF or ANY, and the next two
2886        characters are CRLF, advance by two characters instead of one.
2887    
2888    
2889    Version 6.7 04-Jul-06
2890    ---------------------
2891    
2892     1. In order to handle tests when input lines are enormously long, pcretest has
2893        been re-factored so that it automatically extends its buffers when
2894        necessary. The code is crude, but this _is_ just a test program. The
2895        default size has been increased from 32K to 50K.
2896    
2897     2. The code in pcre_study() was using the value of the re argument before
2898        testing it for NULL. (Of course, in any sensible call of the function, it
2899        won't be NULL.)
2900    
2901     3. The memmove() emulation function in pcre_internal.h, which is used on
2902        systems that lack both memmove() and bcopy() - that is, hardly ever -
2903        was missing a "static" storage class specifier.
2904    
2905     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2906        containing an extended class (one that cannot be represented by a bitmap
2907        because it contains high-valued characters or Unicode property items, e.g.
2908        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2909        pattern, but PCRE should not loop if you do not (it no longer does).
2910        [Detail: two cases were found: (a) a repeated subpattern containing an
2911        extended class; (b) a recursive reference to a subpattern that followed a
2912        previous extended class. It wasn't skipping over the extended class
2913        correctly when UTF-8 mode was not set.]
2914    
2915     5. A negated single-character class was not being recognized as fixed-length
2916        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2917        compile error "lookbehind assertion is not fixed length".
2918    
2919     6. The RunPerlTest auxiliary script was showing an unexpected difference
2920        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2921        write a Perl script that can interpret lines of an input file either as
2922        byte characters or as UTF-8, which is what "perltest" was being required to
2923        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2924        can't do is switch easily at run time between having the "use utf8;" pragma
2925        or not. In the end, I fudged it by using the RunPerlTest script to insert
2926        "use utf8;" explicitly for the UTF-8 tests.
2927    
2928     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2929        the end of the subject string, contrary to the documentation and to what
2930        Perl does. This was true of both matching functions. Now it matches only at
2931        the start of the subject and immediately after *internal* newlines.
2932    
2933     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2934        a pointer to an int instead of a pointer to an unsigned long int. This
2935        caused problems on 64-bit systems.
2936    
2937     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2938        instance of the 'standard' template library not being so standard".
2939    
2940    10. There was no check on the number of named subpatterns nor the maximum
2941        length of a subpattern name. The product of these values is used to compute
2942        the size of the memory block for a compiled pattern. By supplying a very
2943        long subpattern name and a large number of named subpatterns, the size
2944        computation could be caused to overflow. This is now prevented by limiting
2945        the length of names to 32 characters, and the number of named subpatterns
2946        to 10,000.
2947    
2948    11. Subpatterns that are repeated with specific counts have to be replicated in
2949        the compiled pattern. The size of memory for this was computed from the
2950        length of the subpattern and the repeat count. The latter is limited to
2951        65535, but there was no limit on the former, meaning that integer overflow
2952        could in principle occur. The compiled length of a repeated subpattern is
2953        now limited to 30,000 bytes in order to prevent this.
2954    
2955    12. Added the optional facility to have named substrings with the same name.
2956    
2957    13. Added the ability to use a named substring as a condition, using the
2958        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2959        are numbers (not recommended). Forward references are permitted.
2960    
2961    14. Added forward references in named backreferences (if you see what I mean).
2962    
2963    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2964        pattern could run off the end of the subject. For example, the pattern
2965        "(?s)(.{1,5})"8 did this with the subject "ab".
2966    
2967    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2968        PCRE_CASELESS was set when matching characters that were quantified with ?
2969        or *.
2970    
2971    17. A character class other than a single negated character that had a minimum
2972        but no maximum quantifier - for example [ab]{6,} - was not handled
2973        correctly by pce_dfa_exec(). It would match only one character.
2974    
2975    18. A valid (though odd) pattern that looked like a POSIX character
2976        class but used an invalid character after [ (for example [[,abc,]]) caused
2977        pcre_compile() to give the error "Failed: internal error: code overflow" or
2978        in some cases to crash with a glibc free() error. This could even happen if
2979        the pattern terminated after [[ but there just happened to be a sequence of
2980        letters, a binary zero, and a closing ] in the memory that followed.
2981    
2982    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2983        over the years. Originally (before any Unicode support), just the bottom 8
2984        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2985        output from "man perlunicode" includes this:
2986    
2987          The regular expression compiler produces polymorphic opcodes.  That
2988          is, the pattern adapts to the data and automatically switches to
2989          the Unicode character scheme when presented with Unicode data--or
2990          instead uses a traditional byte scheme when presented with byte
2991          data.
2992    
2993        Sadly, a wide octal escape does not cause a switch, and in a string with
2994        no other multibyte characters, these octal escapes are treated as before.
2995        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2996        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2997        Unicode string.
2998    
2999        I have not perpetrated such confusion in PCRE. Up till now, it took just
3000        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3001        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3002        translate to the appropriate multibyte character.
3003    
3004    29. Applied some refactoring to reduce the number of warnings from Microsoft
3005        and Borland compilers. This has included removing the fudge introduced
3006        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3007        a warning about an unused variable.
3008    
3009    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3010        characters since release 4.0, because Perl (from release 5.004) does not.
3011        [Or at least, is documented not to: some releases seem to be in conflict
3012        with the documentation.] However, when a pattern was studied with
3013        pcre_study() and all its branches started with \s, PCRE still included VT
3014        as a possible starting character. Of course, this did no harm; it just
3015        caused an unnecessary match attempt.
3016    
3017    22. Removed a now-redundant internal flag bit that recorded the fact that case
3018        dependency changed within the pattern. This was once needed for "required
3019        byte" processing, but is no longer used. This recovers a now-scarce options
3020        bit. Also moved the least significant internal flag bit to the most-
3021        significant bit of the word, which was not previously used (hangover from
3022        the days when it was an int rather than a uint) to free up another bit for
3023        the future.
3024    
3025    23. Added support for CRLF line endings as well as CR and LF. As well as the
3026        default being selectable at build time, it can now be changed at runtime
3027        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3028        specify that it is scanning data with non-default line endings.
3029    
3030    24. Changed the definition of CXXLINK to make it agree with the definition of
3031        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3032    
3033    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3034        recursions. This makes a big different to stack usage for some patterns.
3035    
3036    26. If a subpattern containing a named recursion or subroutine reference such
3037        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3038        the space required for the compiled pattern went wrong and gave too small a
3039        value. Depending on the environment, this could lead to "Failed: internal
3040        error: code overflow at offset 49" or "glibc detected double free or
3041        corruption" errors.
3042    
3043    27. Applied patches from Google (a) to support the new newline modes and (b) to
3044        advance over multibyte UTF-8 characters in GlobalReplace.
3045    
3046    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3047        difference for some implementation of PCRE in some Windows version.
3048    
3049    29. Added some extra testing facilities to pcretest:
3050    
3051        \q<number>   in a data line sets the "match limit" value
3052        \Q<number>   in a data line sets the "match recursion limt" value
3053        -S <number>  sets the stack size, where <number> is in megabytes
3054    
3055        The -S option isn't available for Windows.
3056    
3057    
3058    Version 6.6 06-Feb-06
3059    ---------------------
3060    
3061     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3062        in pcreposix.h. I have copied the definition from pcre.h.
3063    
3064     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3065        because pcre.h is no longer a built file.
3066    
3067     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3068        not normally included in the compiled code.
3069    
3070    
3071    Version 6.5 01-Feb-06
3072    ---------------------
3073    
3074     1. When using the partial match feature with pcre_dfa_exec(), it was not
3075        anchoring the second and subsequent partial matches at the new starting
3076        point. This could lead to incorrect results. For example, with the pattern
3077        /1234/, partially matching against "123" and then "a4" gave a match.
3078    
3079     2. Changes to pcregrep:
3080    
3081        (a) All non-match returns from pcre_exec() were being treated as failures
3082            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3083            error message is output. Some extra information is given for the
3084            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3085            probably the only errors that are likely to be caused by users (by
3086            specifying a regex that has nested indefinite repeats, for instance).
3087            If there are more than 20 of these errors, pcregrep is abandoned.
3088    
3089        (b) A binary zero was treated as data while matching, but terminated the
3090            output line if it was written out. This has been fixed: binary zeroes
3091            are now no different to any other data bytes.
3092    
3093        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3094            used to set a locale for matching. The --locale=xxxx long option has
3095            been added (no short equivalent) to specify a locale explicitly on the
3096            pcregrep command, overriding the environment variables.
3097    
3098        (d) When -B was used with -n, some line numbers in the output were one less
3099            than they should have been.
3100    
3101        (e) Added the -o (--only-matching) option.
3102    
3103        (f) If -A or -C was used with -c (count only), some lines of context were
3104            accidentally printed for the final match.
3105    
3106        (g) Added the -H (--with-filename) option.
3107    
3108        (h) The combination of options -rh failed to suppress file names for files
3109            that were found from directory arguments.
3110    
3111        (i) Added the -D (--devices) and -d (--directories) options.
3112    
3113        (j) Added the -F (--fixed-strings) option.
3114    
3115        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3116    
3117        (l) Added the --colo(u)r option.
3118    
3119        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3120            is not present by default.
3121    
3122     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3123        items such as (?R) or (?1), when the recursion could match a number of
3124        alternatives. If it matched one of the alternatives, but subsequently,
3125        outside the recursion, there was a failure, the code tried to back up into
3126        the recursion. However, because of the way PCRE is implemented, this is not
3127        possible, and the result was an incorrect result from the match.
3128    
3129        In order to prevent this happening, the specification of recursion has
3130        been changed so that all such subpatterns are automatically treated as
3131        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3132    
3133     4. I had overlooked the fact that, in some locales, there are characters for
3134        which isalpha() is true but neither isupper() nor islower() are true. In
3135        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3136        and ordfeminine) are like this. This affected the treatment of \w and \W
3137        when they appeared in character classes, but not when they appeared outside
3138        a character class. The bit map for "word" characters is now created
3139        separately from the results of isalnum() instead of just taking it from the
3140        upper, lower, and digit maps. (Plus the underscore character, of course.)
3141    
3142     5. The above bug also affected the handling of POSIX character classes such as
3143        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3144        permanent tables. Instead, the bit maps for such a class were previously
3145        created as the appropriate unions of the upper, lower, and digit bitmaps.
3146        Now they are created by subtraction from the [[:word:]] class, which has
3147        its own bitmap.
3148    
3149     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3150        It is created by subtracting the vertical space characters (\x09, \x0a,
3151        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3152        subtraction was done in the overall bitmap for a character class, meaning
3153        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3154        be recognized. This bug has been fixed.
3155    
3156     7. Patches from the folks at Google:
3157    
3158          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3159          real life, but is still worth protecting against".
3160    
3161          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3162          regular expressions".
3163    
3164          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3165          have it.
3166    
3167          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3168          "configure" and the latter not, in order to fix a problem somebody had
3169          with compiling the Arg class on HP-UX.
3170    
3171          (e) Improve the error-handling of the C++ wrapper a little bit.
3172    
3173          (f) New tests for checking recursion limiting.
3174    
3175     8. The pcre_memmove() function, which is used only if the environment does not
3176        have a standard memmove() function (and is therefore rarely compiled),
3177        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3178        returning a result (though PCRE never actually uses the result).
3179    
3180     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3181        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3182        returned instead of calling malloc() with an overflowing number that would
3183        most likely cause subsequent chaos.
3184    
3185    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3186    
3187    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3188        with this option is matched, the nmatch and pmatch options of regexec() are
3189        ignored.
3190    
3191    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3192        provided in case anyone wants to the the POSIX interface with UTF-8
3193        strings.
3194    
3195    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3196        C++ linking (needed for some HP-UX environments).
3197    
3198    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3199        (unused parameter) and in the pcre_printint() function (omitted "default"
3200        switch label when the default is to do nothing).
3201    
3202    15. Added some code to make it possible, when PCRE is compiled as a C++
3203        library, to replace subject pointers for pcre_exec() with a smart pointer
3204        class, thus making it possible to process discontinuous strings.
3205    
3206    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3207        much the same function. They were added by different people who were trying
3208        to make PCRE easy to compile on non-Unix systems. It has been suggested
3209        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3210        for compiling on Windows systems. I have therefore replaced it with
3211        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3212        defaults to "extern" for C or "extern C" for C++, which works fine on
3213        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3214        SCOPE with something explicit in config.h. In addition:
3215    
3216        (a) pcreposix.h still had just "extern" instead of either of these macros;
3217            I have replaced it with PCRE_DATA_SCOPE.
3218    
3219        (b) Functions such as _pcre_xclass(), which are internal to the library,
3220            but external in the C sense, all had PCRE_EXPORT in their definitions.
3221            This is apparently wrong for the Windows case, so I have removed it.
3222            (It makes no difference on Unix-like systems.)
3223    
3224    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3225        of recursive calls to match(). This is different to MATCH_LIMIT because
3226        that limits the total number of calls to match(), not all of which increase
3227        the depth of recursion. Limiting the recursion depth limits the amount of
3228        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3229        when PCRE is compiled, and changed at run time. A patch from Google adds
3230        this functionality to the C++ interface.
3231    
3232    18. Changes to the handling of Unicode character properties:
3233    
3234        (a) Updated the table to Unicode 4.1.0.
3235    
3236        (b) Recognize characters that are not in the table as "Cn" (undefined).
3237    
3238        (c) I revised the way the table is implemented to a much improved format
3239            which includes recognition of ranges. It now supports the ranges that
3240            are defined in UnicodeData.txt, and it also amalgamates other
3241            characters into ranges. This has reduced the number of entries in the
3242            table from around 16,000 to around 3,000, thus reducing its size
3243            considerably. I realized I did not need to use a tree structure after
3244            all - a binary chop search is just as efficient. Having reduced the
3245            number of entries, I extended their size from 6 bytes to 8 bytes to
3246            allow for more data.
3247    
3248        (d) Added support for Unicode script names via properties such as \p{Han}.
3249    
3250    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3251        matching that character.
3252    
3253    20. When matching a repeated Unicode property with a minimum greater than zero,
3254        (for example \pL{2,}), PCRE could look past the end of the subject if it
3255        reached it while seeking the minimum number of characters. This could
3256        happen only if some of the characters were more than one byte long, because
3257        there is a check for at least the minimum number of bytes.
3258    
3259    21. Refactored the implementation of \p and \P so as to be more general, to
3260        allow for more different types of property in future. This has changed the
3261        compiled form incompatibly. Anybody with saved compiled patterns that use
3262        \p or \P will have to recompile them.
3263    
3264    22. Added "Any" and "L&" to the supported property types.
3265    
3266    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3267        but give a compile time error if the value is greater than 0xff.
3268    
3269    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3270        accidentally not being installed or uninstalled.
3271    
3272    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3273        made were to insert the current release number. This seemed silly, because
3274        it made things harder for people building PCRE on systems that don't run
3275        "configure". I have turned pcre.h into a distributed file, no longer built
3276        by "configure", with the version identification directly included. There is
3277        no longer a pcre.h.in file.
3278    
3279        However, this change necessitated a change to the pcre-config script as
3280        well. It is built from pcre-config.in, and one of the substitutions was the
3281        release number. I have updated configure.ac so that ./configure now finds
3282        the release number by grepping pcre.h.
3283    
3284    26. Added the ability to run the tests under valgrind.
3285    
3286    
3287    Version 6.4 05-Sep-05
3288    ---------------------
3289    
3290     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3291        "--" to be printed when multiple files were scanned, even when none of the
3292        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3293        consider it to be a bug, and have restored the previous behaviour.
3294    
3295     2. A couple of code tidies to get rid of compiler warnings.
3296    
3297     3. The pcretest program used to cheat by referring to symbols in the library
3298        whose names begin with _pcre_. These are internal symbols that are not
3299        really supposed to be visible externally, and in some environments it is
3300        possible to suppress them. The cheating is now confined to including
3301        certain files from the library's source, which is a bit cleaner.
3302    
3303     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3304        file's purpose clearer.
3305    
3306     5. Reorganized pcre_ucp_findchar().
3307    
3308    
3309    Version 6.3 15-Aug-05
3310    ---------------------
3311    
3312     1. The file libpcre.pc.in did not have general read permission in the tarball.
3313    
3314     2. There were some problems when building without C++ support:
3315    
3316        (a) If C++ support was not built, "make install" and "make test" still
3317            tried to test it.
3318    
3319        (b) There were problems when the value of CXX was explicitly set. Some
3320            changes have been made to try to fix these, and ...
3321    
3322        (c) --disable-cpp can now be used to explicitly disable C++ support.
3323    
3324        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3325            backslash in a target when C++ was disabled. This confuses some
3326            versions of "make", apparently. Using an intermediate variable solves
3327            this. (Same for CPP_LOBJ.)
3328    
3329     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3330        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3331        necessary on certain architectures.
3332    
3333     4. Added a setting of -export-symbols-regex to the link command to remove
3334        those symbols that are exported in the C sense, but actually are local
3335        within the library, and not documented. Their names all begin with
3336        "_pcre_". This is not a perfect job, because (a) we have to except some
3337        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3338        available (and never for static libraries). I have made a note to try to
3339        find a way round (a) in the future.
3340    
3341    
3342    Version 6.2 01-Aug-05
3343    ---------------------
3344    
3345     1. There was no test for integer overflow of quantifier values. A construction
3346        such as {1111111111111111} would give undefined results. What is worse, if
3347        a minimum quantifier for a parenthesized subpattern overflowed and became
3348        negative, the calculation of the memory size went wrong. This could have
3349        led to memory overwriting.
3350    
3351     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3352    
3353     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3354        operating environments where this matters.
3355    
3356     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3357        PCRE options from within the C++ wrapper.
3358    
3359     5. Named capturing subpatterns were not being correctly counted when a pattern
3360        was compiled. This caused two problems: (a) If there were more than 100
3361        such subpatterns, the calculation of the memory needed for the whole
3362        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3363        back references of the form \12, where the number was greater than 9, were
3364        not recognized as back references, even though there were sufficient
3365        previous subpatterns.
3366    
3367     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3368        versions of gcc, e.g. 2.95.4.
3369    
3370    
3371    Version 6.1 21-Jun-05
3372    ---------------------
3373    
3374     1. There was one reference to the variable "posix" in pcretest.c that was not
3375        surrounded by "#if !defined NOPOSIX".
3376    
3377     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3378        the cross-check on the old pcre_info() function, for the benefit of the
3379        cut-down version of PCRE that is currently imported into Exim.
3380    
3381     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3382        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3383        patterns that start (?i) but I don't think that matters. The use of (?i) is
3384        just an example; this all applies to the other options as well.
3385    
3386     4. Since libtool seems to echo the compile commands it is issuing, the output
3387        from "make" can be reduced a bit by putting "@" in front of each libtool
3388        compile command.
3389    
3390     5. Patch from the folks at Google for configure.in to be a bit more thorough
3391        in checking for a suitable C++ installation before trying to compile the
3392        C++ stuff. This should fix a reported problem when a compiler was present,
3393        but no suitable headers.
3394    
3395     6. The man pages all had just "PCRE" as their title. I have changed them to
3396        be the relevant file name. I have also arranged that these names are
3397        retained in the file doc/pcre.txt, which is a concatenation in text format
3398        of all the man pages except the little individual ones for each function.
3399    
3400     7. The NON-UNIX-USE file had not been updated for the different set of source
3401        files that come with release 6. I also added a few comments about the C++
3402        wrapper.
3403    
3404    
3405    Version 6.0 07-Jun-05
3406    ---------------------
3407    
3408     1. Some minor internal re-organization to help with my DFA experiments.
3409    
3410     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3411        didn't matter for the library itself when fully configured, but did matter
3412        when compiling without UCP support, or within Exim, where the ucp files are
3413        not imported.
3414    
3415     3. Refactoring of the library code to split up the various functions into
3416        different source modules. The addition of the new DFA matching code (see
3417        below) to a single monolithic source would have made it really too
3418        unwieldy, quite apart from causing all the code to be include in a
3419        statically linked application, when only some functions are used. This is
3420        relevant even without the DFA addition now that patterns can be compiled in
3421        one application and matched in another.
3422    
3423        The downside of splitting up is that there have to be some external
3424        functions and data tables that are used internally in different modules of
3425        the library but which are not part of the API. These have all had their
3426        names changed to start with "_pcre_" so that they are unlikely to clash
3427        with other external names.
3428    
3429     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3430        a different (DFA) algorithm. Although it is slower than the original
3431        function, it does have some advantages for certain types of matching
3432        problem.
3433    
3434     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3435        including restarting after a partial match.
3436    
3437     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3438        defined when compiling for Windows was sent to me. I have put it into the
3439        code, though I have no means of testing or verifying it.
3440    
3441     7. Added the pcre_refcount() auxiliary function.
3442    
3443     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3444        match before or at the first newline in the subject string. In pcretest,
3445        the /f option on a pattern can be used to set this.
3446    
3447     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3448        would behave wrongly. This has been present in PCRE since release 4.0.
3449    
3450    10. A number of changes to the pcregrep command:
3451    
3452        (a) Refactored how -x works; insert ^(...)$ instead of setting
3453            PCRE_ANCHORED and checking the length, in preparation for adding
3454            something similar for -w.
3455    
3456        (b) Added the -w (match as a word) option.
3457    
3458        (c) Refactored the way lines are read and buffered so as to have more
3459            than one at a time available.
3460    
3461        (d) Implemented a pcregrep test script.
3462    
3463        (e) Added the -M (multiline match) option. This allows patterns to match
3464            over several lines of the subject. The buffering ensures that at least
3465            8K, or the rest of the document (whichever is the shorter) is available
3466            for matching (and similarly the previous 8K for lookbehind assertions).
3467    
3468        (f) Changed the --help output so that it now says
3469    
3470              -w, --word-regex(p)
3471    
3472            instead of two lines, one with "regex" and the other with "regexp"
3473            because that confused at least one person since the short forms are the
3474            same. (This required a bit of code, as the output is generated
3475            automatically from a table. It wasn't just a text change.)
3476    
3477        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3478            option but starts with a hyphen. Could be a pattern or a path name
3479            starting with a hyphen, for instance.
3480    
3481        (h) "-" can be given as a file name to represent stdin.
3482    
3483        (i) When file names are being printed, "(standard input)" is used for
3484            the standard input, for compatibility with GNU grep. Previously
3485            "<stdin>" was used.
3486    
3487        (j) The option --label=xxx can be used to supply a name to be used for
3488            stdin when file names are being printed. There is no short form.
3489    
3490        (k) Re-factored the options decoding logic because we are going to add
3491            two more options that take data. Such options can now be given in four
3492            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3493    
3494        (l) Added the -A, -B, and -C options for requesting that lines of context
3495            around matches be printed.
3496    
3497        (m) Added the -L option to print the names of files that do not contain
3498            any matching lines, that is, the complement of -l.
3499    
3500        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3501            continue to scan other files.
3502    
3503        (o) The -s option was incorrectly implemented. For compatibility with other
3504            greps, it now suppresses the error message for a non-existent or non-
3505            accessible file (but not the return code). There is a new option called
3506            -q that suppresses the output of matching lines, which was what -s was
3507            previously doing.
3508    
3509        (p) Added --include and --exclude options to specify files for inclusion
3510            and exclusion when recursing.
3511    
3512    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3513        Hopefully, it now does.
3514    
3515    12. Missing cast in pcre_study().
3516    
3517    13. Added an "uninstall" target to the makefile.
3518    
3519    14. Replaced "extern" in the function prototypes in Makefile.in with
3520        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3521        world, but is set differently for Windows.
3522    
3523    15. Added a second compiling function called pcre_compile2(). The only
3524        difference is that it has an extra argument, which is a pointer to an
3525        integer error code. When there is a compile-time failure, this is set
3526        non-zero, in addition to the error test pointer being set to point to an
3527        error message. The new argument may be NULL if no error number is required
3528        (but then you may as well call pcre_compile(), which is now just a
3529        wrapper). This facility is provided because some applications need a
3530        numeric error indication, but it has also enabled me to tidy up the way
3531        compile-time errors are handled in the POSIX wrapper.
3532    
3533    16. Added VPATH=.libs to the makefile; this should help when building with one
3534        prefix path and installing with another. (Or so I'm told by someone who
3535        knows more about this stuff than I do.)
3536    
3537    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3538        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3539        match everything, including newlines. This is not POSIX-compatible, but
3540        somebody wanted the feature. From pcretest it can be activated by using
3541        both the P and the s flags.
3542    
3543    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3544    
3545    19. libpcre.pc was being incorrectly installed as executable.
3546    
3547    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3548        it now also looks for '\r' so that it will work unmodified on Windows.
3549    
3550    21. Added Google's contributed C++ wrapper to the distribution.
3551    
3552    22. Added some untidy missing memory free() calls in pcretest, to keep
3553        Electric Fence happy when testing.
3554    
3555    
3556    
3557    Version 5.0 13-Sep-04
3558    ---------------------
3559    
3560     1. Internal change: literal characters are no longer packed up into items
3561        containing multiple characters in a single byte-string. Each character
3562        is now matched using a separate opcode. However, there may be more than one
3563        byte in the character in UTF-8 mode.
3564    
3565     2. The pcre_callout_block structure has two new fields: pattern_position and
3566        next_item_length. These contain the offset in the pattern to the next match
3567        item, and its length, respectively.
3568    
3569     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3570        insertion of callouts before each pattern item. Added the /C option to
3571        pcretest to make use of this.
3572    
3573     4. On the advice of a Windows user, the lines
3574    
3575          #if defined(_WIN32) || defined(WIN32)
3576          _setmode( _fileno( stdout ), 0x8000 );
3577          #endif  /* defined(_WIN32) || defined(WIN32) */
3578    
3579        have been added to the source of pcretest. This apparently does useful
3580        magic in relation to line terminators.
3581    
3582     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3583        for the benefit of those environments where the "b" makes a difference.
3584    
3585     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3586        to know about it. I have put a hack into configure.in that adds in code
3587        to set GCC=yes if CC=icc. This seems to end up at a point in the
3588        generated configure script that is early enough to affect the setting of
3589        compiler options, which is what is needed, but I have no means of testing
3590        whether it really works. (The user who reported this had patched the
3591        generated configure script, which of course I cannot do.)
3592    
3593        LATER: After change 22 below (new libtool files), the configure script
3594        seems to know about icc (and also ecc). Therefore, I have commented out
3595        this hack in configure.in.
3596    
3597     7. Added support for pkg-config (2 patches were sent in).
3598    
3599     8. Negated POSIX character classes that used a combination of internal tables
3600        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3601        [[:^ascii]]. Typically, they would match almost any characters. The other
3602        POSIX classes were not broken in this way.
3603    
3604     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3605        to find the match, as PCRE was deluded into thinking that the match had to
3606        start at the start point or following a newline. The same bug applied to
3607        patterns with negative forward assertions or any backward assertions
3608        preceding ".*" at the start, unless the pattern required a fixed first
3609        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3610    
3611    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3612        starting at the last subject character, bytes beyond the end of the subject
3613        string were read.
3614    
3615    11. Renamed the variable "class" as "classbits" to make life easier for C++
3616        users. (Previously there was a macro definition, but it apparently wasn't
3617        enough.)
3618    
3619    12. Added the new field "tables" to the extra data so that tables can be passed
3620        in at exec time, or the internal tables can be re-selected. This allows
3621        a compiled regex to be saved and re-used at a later time by a different
3622        program that might have everything at different addresses.
3623    
3624    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3625        -R library as well as a -L library.
3626    
3627    14. The debugging options of pcretest (-d on the command line or D on a
3628        pattern) showed incorrect output for anything following an extended class
3629        that contained multibyte characters and which was followed by a quantifier.
3630    
3631    15. Added optional support for general category Unicode character properties
3632        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3633        support. It adds about 90K to the size of the library. The meanings of the
3634        inbuilt class escapes such as \d and \s have NOT been changed.
3635    
3636    16. Updated pcredemo.c to include calls to free() to release the memory for the
3637        compiled pattern.
3638    
3639    17. The generated file chartables.c was being created in the source directory
3640        instead of in the building directory. This caused the build to fail if the
3641        source directory was different from the building directory, and was
3642        read-only.
3643    
3644    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3645        file. No doubt somebody will tell me if they don't make sense... Also added
3646        Dan Mooney's comments about building on OpenVMS.
3647    
3648    19. Added support for partial matching via the PCRE_PARTIAL option for
3649        pcre_exec() and the \P data escape in pcretest.
3650    
3651    20. Extended pcretest with 3 new pattern features:
3652    
3653        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3654              write the compiled pattern to the file whose name is "rest-of-line".
3655              This is a straight binary dump of the data, with the saved pointer to
3656              the character tables forced to be NULL. The study data, if any, is
3657              written too. After writing, pcretest reads a new pattern.
3658    
3659        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3660              compiled pattern from the given file. There must not be any
3661              occurrences of "<" in the file name (pretty unlikely); if there are,
3662              pcretest will instead treat the initial "<" as a pattern delimiter.
3663              After reading in the pattern, pcretest goes on to read data lines as
3664              usual.
3665    
3666        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3667              and 16-bit fields in a compiled pattern, to simulate a pattern that
3668              was compiled on a host of opposite endianness.
3669    
3670    21. The pcre-exec() function can now cope with patterns that were compiled on
3671        hosts of opposite endianness, with this restriction:
3672    
3673          As for any compiled expression that is saved and used later, the tables
3674          pointer field cannot be preserved; the extra_data field in the arguments
3675          to pcre_exec() should be used to pass in a tables address if a value
3676          other than the default internal tables were used at compile time.
3677    
3678    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3679        now diagnosed as an error. Previously, most of the time, a negative number
3680        would have been treated as zero, but if in addition "ovector" was passed as
3681        NULL, a crash could occur.
3682    
3683    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3684        new versions from the libtool 1.5 distribution (the last one is a copy of
3685        a file called libtool.m4). This seems to have fixed the need to patch
3686        "configure" to support Darwin 1.3 (which I used to do). However, I still
3687        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3688        workstation).
3689    
3690    24. Changed the PCRE licence to be the more standard "BSD" licence.
3691    
3692    
3693    Version 4.5 01-Dec-03
3694    ---------------------
3695    
3696     1. There has been some re-arrangement of the code for the match() function so
3697        that it can be compiled in a version that does not call itself recursively.
3698        Instead, it keeps those local variables that need separate instances for
3699        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3700        needs to "recurse". Keeping track of where control must go is done by means
3701        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3702        hide most of the details from the main code, and operates only if
3703        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3704        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3705        operating.
3706    
3707        To make it easier for callers to provide specially tailored get/free
3708        functions for this usage, two new functions, pcre_stack_malloc, and
3709        pcre_stack_free, are used. They are always called in strict stacking order,
3710        and the size of block requested is always the same.
3711    
3712        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3713        PCRE has been compiled to use the stack or the heap for recursion. The
3714        -C option of pcretest uses this to show which version is compiled.
3715    
3716        A new data escape \S, is added to pcretest; it causes the amounts of store
3717        obtained and freed by both kinds of malloc/free at match time to be added
3718        to the output.
3719    
3720     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3721        what's available on my current Linux desktop machine.
3722    
3723     3. When matching a UTF-8 string, the test for a valid string at the start has
3724        been extended. If start_offset is not zero, PCRE now checks that it points
3725        to a byte that is the start of a UTF-8 character. If not, it returns
3726        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3727        this is necessary because there may be backward assertions in the pattern.
3728        When matching the same subject several times, it may save resources to use
3729        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3730    
3731     4. The code for checking the validity of UTF-8 strings has been tightened so
3732        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3733        containing "overlong sequences".
3734    
3735     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3736        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3737        should have been "&", but it just so happened that all the cases this let
3738        through by mistake were picked up later in the function.
3739    
3740     6. I had used a variable called "isblank" - this is a C99 function, causing
3741        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3742    
3743     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3744        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3745        so the version is shown; (c) stop "make test" echoing "./RunTest".
3746    
3747     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3748    
3749     9. The prototype for memmove() for systems that don't have it was using
3750        size_t, but the inclusion of the header that defines size_t was later. I've
3751        moved the #includes for the C headers earlier to avoid this.
3752    
3753    10. Added some adjustments to the code to make it easier to compiler on certain
3754        special systems:
3755    
3756          (a) Some "const" qualifiers were missing.
3757          (b) Added the macro EXPORT before all exported functions; by default this
3758              is defined to be empty.
3759          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3760              that it reads its output file name as an argument instead of writing
3761              to the standard output and assuming this can be redirected.
3762    
3763    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3764        class containing characters with values greater than 255, PCRE compilation
3765        went into a loop.
3766    
3767    12. A recursive reference to a subpattern that was within another subpattern
3768        that had a minimum quantifier of zero caused PCRE to crash. For example,
3769        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3770        recursion. If the recursively-called subpattern itself had a zero repeat,
3771        that was OK.
3772    
3773    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3774        buffer into which it was copied (for escape processing) was still set at
3775        1024, so long lines caused crashes.
3776    
3777    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3778        "internal error: code overflow...". This applied to any character class
3779        that was followed by a possessive quantifier.
3780    
3781    15. Modified the Makefile to add libpcre.la as a prerequisite for
3782        libpcreposix.la because I was told this is needed for a parallel build to
3783        work.
3784    
3785    16. If a pattern that contained .* following optional items at the start was
3786        studied, the wrong optimizing data was generated, leading to matching
3787        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3788        matching string must start with a or b or c. The correct conclusion for