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

Diff of /code/trunk/ChangeLog

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

revision 45 by nigel, Sat Feb 24 21:39:25 2007 UTC revision 1417 by ph10, Tue Dec 24 18:03:06 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.35-RC1 xx-xxxx-201x
5    -----------------------------
6    
7    1.  A new flag is set, when property checks are present in an XCLASS.
8        When this flag is not set, PCRE can perform certain optimizations
9        such as studying these XCLASS-es.
10    
11    2.  The auto-possessification of character sets were improved: a normal
12        and an extended character set can be compared now. Furthermore
13        the JIT compiler optimizes more character set checks.
14    
15    3.  Got rid of some compiler warnings for potentially uninitialized variables
16        that show up only when compiled with -O2.
17    
18    
19    Version 8.34 15-December-2013
20    -----------------------------
21    
22    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
23        executable memory. Patch inspired by Carsten Klein.
24    
25    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
26        this macro is never tested and has no effect, because the work to support
27        coverage involves only compiling and linking options and special targets in
28        the Makefile. The comment in config.h implied that defining the macro would
29        enable coverage support, which is totally false. There was also support for
30        setting this macro in the CMake files (my fault, I just copied it from
31        configure). SUPPORT_GCOV has now been removed.
32    
33    3.  Make a small performance improvement in strlen16() and strlen32() in
34        pcretest.
35    
36    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
37        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
38        these cases). There was also one in pcretest.c.
39    
40    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
41    
42    6.  In UTF mode, the code for checking whether a group could match an empty
43        string (which is used for indefinitely repeated groups to allow for
44        breaking an infinite loop) was broken when the group contained a repeated
45        negated single-character class with a character that occupied more than one
46        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
47        UTF-8 mode). The effect was undefined: the group might or might not be
48        deemed as matching an empty string, or the program might have crashed.
49    
50    7.  The code for checking whether a group could match an empty string was not
51        recognizing that \h, \H, \v, \V, and \R must match a character.
52    
53    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
54        an empty string. If it can, pcretest shows this in its information output.
55    
56    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
57        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
58        matched by pcre_exec() without using JIT:
59    
60        (a) If the rest of the pattern did not match after a maximal run of
61            grapheme clusters, the code for backing up to try with fewer of them
62            did not always back up over a full grapheme when characters that do not
63            have the modifier quality were involved, e.g. Hangul syllables.
64    
65        (b) If the match point in a subject started with modifier character, and
66            there was no match, the code could incorrectly back up beyond the match
67            point, and potentially beyond the first character in the subject,
68            leading to a segfault or an incorrect match result.
69    
70    10. A conditional group with an assertion condition could lead to PCRE
71        recording an incorrect first data item for a match if no other first data
72        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
73        first data item, and therefore matched "ca" after "c" instead of at the
74        start.
75    
76    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
77        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
78    
79    12. The source of pcregrep now includes z/OS-specific code so that it can be
80        compiled for z/OS as part of the special z/OS distribution.
81    
82    13. Added the -T and -TM options to pcretest.
83    
84    14. The code in pcre_compile.c for creating the table of named capturing groups
85        has been refactored. Instead of creating the table dynamically during the
86        actual compiling pass, the information is remembered during the pre-compile
87        pass (on the stack unless there are more than 20 named groups, in which
88        case malloc() is used) and the whole table is created before the actual
89        compile happens. This has simplified the code (it is now nearly 150 lines
90        shorter) and prepared the way for better handling of references to groups
91        with duplicate names.
92    
93    15. A back reference to a named subpattern when there is more than one of the
94        same name now checks them in the order in which they appear in the pattern.
95        The first one that is set is used for the reference. Previously only the
96        first one was inspected. This change makes PCRE more compatible with Perl.
97    
98    16. Unicode character properties were updated from Unicode 6.3.0.
99    
100    17. The compile-time code for auto-possessification has been refactored, based
101        on a patch by Zoltan Herczeg. It now happens after instead of during
102        compilation. The code is cleaner, and more cases are handled. The option
103        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
104        options in pcretest are provided to set it. It can also be set by
105        (*NO_AUTO_POSSESS) at the start of a pattern.
106    
107    18. The character VT has been added to the default ("C" locale) set of
108        characters that match \s and are generally treated as white space,
109        following this same change in Perl 5.18. There is now no difference between
110        "Perl space" and "POSIX space". Whether VT is treated as white space in
111        other locales depends on the locale.
112    
113    19. The code for checking named groups as conditions, either for being set or
114        for being recursed, has been refactored (this is related to 14 and 15
115        above). Processing unduplicated named groups should now be as fast at
116        numerical groups, and processing duplicated groups should be faster than
117        before.
118    
119    20. Two patches to the CMake build system, by Alexander Barkov:
120    
121          (1) Replace the "source" command by "." in CMakeLists.txt because
122              "source" is a bash-ism.
123    
124          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
125              without these the CMake build does not work on Solaris.
126    
127    21. Perl has changed its handling of \8 and \9. If there is no previously
128        encountered capturing group of those numbers, they are treated as the
129        literal characters 8 and 9 instead of a binary zero followed by the
130        literals. PCRE now does the same.
131    
132    22. Following Perl, added \o{} to specify codepoints in octal, making it
133        possible to specify values greater than 0777 and also making them
134        unambiguous.
135    
136    23. Perl now gives an error for missing closing braces after \x{... instead of
137        treating the string as literal. PCRE now does the same.
138    
139    24. RunTest used to grumble if an inappropriate test was selected explicitly,
140        but just skip it when running all tests. This make it awkward to run ranges
141        of tests when one of them was inappropriate. Now it just skips any
142        inappropriate tests, as it always did when running all tests.
143    
144    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
145        character types such as \d or \w, too many callouts were inserted, and the
146        data that they returned was rubbish.
147    
148    26. In UCP mode, \s was not matching two of the characters that Perl matches,
149        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
150        were matched by \h. The code has now been refactored so that the lists of
151        the horizontal and vertical whitespace characters used for \h and \v (which
152        are defined only in one place) are now also used for \s.
153    
154    27. Add JIT support for the 64 bit TileGX architecture.
155        Patch by Jiong Wang (Tilera Corporation).
156    
157    28. Possessive quantifiers for classes (both explicit and automatically
158        generated) now use special opcodes instead of wrapping in ONCE brackets.
159    
160    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
161        (because it's meaningless), this was not happening when PCRE_CASELESS was
162        set. Not wrong, but inefficient.
163    
164    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
165        properties for \w, \d, etc) is present in a test regex. Otherwise if the
166        test contains no characters greater than 255, Perl doesn't realise it
167        should be using Unicode semantics.
168    
169    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
170        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
171        does in Unicode mode.
172    
173    32. Added the "forbid" facility to pcretest so that putting tests into the
174        wrong test files can sometimes be quickly detected.
175    
176    33. There is now a limit (default 250) on the depth of nesting of parentheses.
177        This limit is imposed to control the amount of system stack used at compile
178        time. It can be changed at build time by --with-parens-nest-limit=xxx or
179        the equivalent in CMake.
180    
181    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
182        errors. Perl warns for these when in warning mode, but PCRE has no facility
183        for giving warnings.
184    
185    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
186        However, this was not working for (?!) because it is optimized to (*FAIL),
187        for which PCRE does not allow quantifiers. The optimization is now disabled
188        when a quantifier follows (?!). I can't see any use for this, but it makes
189        things uniform.
190    
191    36. Perl no longer allows group names to start with digits, so I have made this
192        change also in PCRE. It simplifies the code a bit.
193    
194    37. In extended mode, Perl ignores spaces before a + that indicates a
195        possessive quantifier. PCRE allowed a space before the quantifier, but not
196        before the possessive +. It now does.
197    
198    38. The use of \K (reset reported match start) within a repeated possessive
199        group such as (a\Kb)*+ was not working.
200    
201    40. Document that the same character tables must be used at compile time and
202        run time, and that the facility to pass tables to pcre_exec() and
203        pcre_dfa_exec() is for use only with saved/restored patterns.
204    
205    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
206        features for Builds with MSVC:
207    
208        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
209           builds). The .rc files can be used to set FileDescription and many other
210           attributes.
211    
212        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
213           This allows higher-level build scripts which want .pdb files to avoid
214           hard-coding the exact files needed."
215    
216    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
217        mean "start of word" and "end of word", respectively, as a transition aid.
218    
219    43. A minimizing repeat of a class containing codepoints greater than 255 in
220        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
221        compiled to use the heap for recursion.
222    
223    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
224        is configured.
225    
226    
227    Version 8.33 28-May-2013
228    ------------------------
229    
230    1.  Added 'U' to some constants that are compared to unsigned integers, to
231        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
232        variables that are added to signed variables, to ensure the result is
233        signed and can be negated.
234    
235    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
236        CMake config files.
237    
238    3.  Revise the creation of config.h.generic so that all boolean macros are
239        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
240        overriding via -D on the command line possible.
241    
242    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
243        to unsigned int is reported to make a quite noticeable speed difference in
244        a specific Windows environment. Testing on Linux did also appear to show
245        some benefit (and it is clearly not harmful). Also fixed the definition of
246        Xop which should be unsigned.
247    
248    5.  Related to (4), changing the definition of the intermediate variable cc
249        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
250        improvements.
251    
252    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
253        unnecessary spaces.
254    
255    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
256        and later.
257    
258    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
259    
260    9.  Optimizing fast_forward_start_bits in JIT.
261    
262    10. Adding support for callouts in JIT, and fixing some issues revealed
263        during this work. Namely:
264    
265        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
266    
267        (b) Minimum length was not checked before the matching is started.
268    
269    11. The value of capture_last that is passed to callouts was incorrect in some
270        cases when there was a capture on one path that was subsequently abandoned
271        after a backtrack. Also, the capture_last value is now reset after a
272        recursion, since all captures are also reset in this case.
273    
274    12. The interpreter no longer returns the "too many substrings" error in the
275        case when an overflowing capture is in a branch that is subsequently
276        abandoned after a backtrack.
277    
278    13. In the pathological case when an offset vector of size 2 is used, pcretest
279        now prints out the matched string after a yield of 0 or 1.
280    
281    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
282        Only supported by the JIT compiler at the moment.
283    
284    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
285    
286    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
287        offset of the starting point of the matching process, provided the offsets
288        vector is large enough.
289    
290    17. The \A escape now records a lookbehind value of 1, though its execution
291        does not actually inspect the previous character. This is to ensure that,
292        in partial multi-segment matching, at least one character from the old
293        segment is retained when a new segment is processed. Otherwise, if there
294        are no lookbehinds in the pattern, \A might match incorrectly at the start
295        of a new segment.
296    
297    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
298    
299    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
300        gives some modest performance improvement in 8-bit mode.
301    
302    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
303        be expressed in certain programming languages using Universal Character
304        Names.
305    
306    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
307        which points out that "non characters" are not "characters that may not
308        appear in Unicode strings" but rather "characters that are reserved for
309        internal use and have only local meaning".
310    
311    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
312        there was a conditional group that depended on an assertion, if the
313        assertion was false, the callout that immediately followed the alternation
314        in the condition was skipped when pcre_exec() was used for matching.
315    
316    23. Allow an explicit callout to be inserted before an assertion that is the
317        condition for a conditional group, for compatibility with automatic
318        callouts, which always insert a callout at this point.
319    
320    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
321        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
322    
323    25. (*PRUNE) is now supported by the JIT compiler.
324    
325    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
326    
327    27. Fix the case where there are two or more SKIPs with arguments that may be
328        ignored.
329    
330    28. (*SKIP) is now supported by the JIT compiler.
331    
332    29. (*THEN) is now supported by the JIT compiler.
333    
334    30. Update RunTest with additional test selector options.
335    
336    31. The way PCRE handles backtracking verbs has been changed in two ways.
337    
338        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
339        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
340        backtracking. In some cases this makes it more Perl-compatible, but Perl's
341        rather obscure rules do not always do the same thing.
342    
343        (2) Previously, backtracking verbs were confined within assertions. This is
344        no longer the case for positive assertions, except for (*ACCEPT). Again,
345        this sometimes improves Perl compatibility, and sometimes does not.
346    
347    32. A number of tests that were in test 2 because Perl did things differently
348        have been moved to test 1, because either Perl or PCRE has changed, and
349        these tests are now compatible.
350    
351    32. Backtracking control verbs are now handled in the same way in JIT and
352        interpreter.
353    
354    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
355        contained a forward subroutine reference caused a compile error.
356    
357    34. Auto-detect and optimize limited repetitions in JIT.
358    
359    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
360        blocking (*UTF) etc.
361    
362    36. In the interpreter, maximizing pattern repetitions for characters and
363        character types now use tail recursion, which reduces stack usage.
364    
365    37. The value of the max lookbehind was not correctly preserved if a compiled
366        and saved regex was reloaded on a host of different endianness.
367    
368    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
369        of the compiled pattern block, expand the flags field from 16 to 32 bits
370        because it was almost full.
371    
372    39. Try madvise first before posix_madvise.
373    
374    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
375        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
376        original change locked out all matches of empty strings. This has been
377        changed so that one match of an empty string per line is recognized.
378        Subsequent searches on the same line (for colouring or for --only-matching,
379        for example) do not recognize empty strings.
380    
381    41. Applied a user patch to fix a number of spelling mistakes in comments.
382    
383    42. Data lines longer than 65536 caused pcretest to crash.
384    
385    43. Clarified the data type for length and startoffset arguments for pcre_exec
386        and pcre_dfa_exec in the function-specific man pages, where they were
387        explicitly stated to be in bytes, never having been updated. I also added
388        some clarification to the pcreapi man page.
389    
390    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
391        a segmentation fault.
392    
393    
394    Version 8.32 30-November-2012
395    -----------------------------
396    
397    1.  Improved JIT compiler optimizations for first character search and single
398        character iterators.
399    
400    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
401        Patch by Daniel Richard G.
402    
403    3.  Single character iterator optimizations in the JIT compiler.
404    
405    4.  Improved JIT compiler optimizations for character ranges.
406    
407    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
408        Reported by Giuseppe D'Angelo.
409    
410    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
411        of a line, was being set incorrectly in cases where .* appeared inside
412        atomic brackets at the start of a pattern, or where there was a subsequent
413        *PRUNE or *SKIP.
414    
415    7.  Improved instruction cache flush for POWER/PowerPC.
416        Patch by Daniel Richard G.
417    
418    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
419        grep:
420    
421        (a) There is now no limit to the number of patterns to be matched.
422    
423        (b) An error is given if a pattern is too long.
424    
425        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
426            are now supported.
427    
428        (d) --exclude-from and --include-from (multiple use) have been added.
429    
430        (e) Exclusions and inclusions now apply to all files and directories, not
431            just to those obtained from scanning a directory recursively.
432    
433        (f) Multiple uses of -f and --file-list are now supported.
434    
435        (g) In a Windows environment, the default for -d has been changed from
436            "read" (the GNU grep default) to "skip", because otherwise the presence
437            of a directory in the file list provokes an error.
438    
439        (h) The documentation has been revised and clarified in places.
440    
441    9.  Improve the matching speed of capturing brackets.
442    
443    10. Changed the meaning of \X so that it now matches a Unicode extended
444        grapheme cluster.
445    
446    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
447        out POSIX threads when JIT support is configured.
448    
449    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
450    
451    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
452        structure could go wrong in environments where size_t is not the same size
453        as int.
454    
455    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
456        set.
457    
458    15. The EBCDIC support had decayed; later updates to the code had included
459        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
460        general tidy up of EBCDIC-related issues, and the documentation was also
461        not quite right. There is now a test that can be run on ASCII systems to
462        check some of the EBCDIC-related things (but is it not a full test).
463    
464    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
465        in a small tidy to the code.
466    
467    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
468    
469    18. If the --only-matching (-o) option in pcregrep is specified multiple
470        times, each one causes appropriate output. For example, -o1 -o2 outputs the
471        substrings matched by the 1st and 2nd capturing parentheses. A separating
472        string can be specified by --om-separator (default empty).
473    
474    19. Improving the first n character searches.
475    
476    20. Turn case lists for horizontal and vertical white space into macros so that
477        they are defined only once.
478    
479    21. This set of changes together give more compatible Unicode case-folding
480        behaviour for characters that have more than one other case when UCP
481        support is available.
482    
483        (a) The Unicode property table now has offsets into a new table of sets of
484            three or more characters that are case-equivalent. The MultiStage2.py
485            script that generates these tables (the pcre_ucd.c file) now scans
486            CaseFolding.txt instead of UnicodeData.txt for character case
487            information.
488    
489        (b) The code for adding characters or ranges of characters to a character
490            class has been abstracted into a generalized function that also handles
491            case-independence. In UTF-mode with UCP support, this uses the new data
492            to handle characters with more than one other case.
493    
494        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
495            whose other case is greater than 256 are now correctly matched
496            caselessly. Previously, the high codepoint matched the low one, but not
497            vice versa.
498    
499        (d) The processing of \h, \H, \v, and \ in character classes now makes use
500            of the new class addition function, using character lists defined as
501            macros alongside the case definitions of 20 above.
502    
503        (e) Caseless back references now work with characters that have more than
504            one other case.
505    
506        (f) General caseless matching of characters with more than one other case
507            is supported.
508    
509    22. Unicode character properties were updated from Unicode 6.2.0
510    
511    23. Improved CMake support under Windows. Patch by Daniel Richard G.
512    
513    24. Add support for 32-bit character strings, and UTF-32
514    
515    25. Major JIT compiler update (code refactoring and bugfixing).
516        Experimental Sparc 32 support is added.
517    
518    26. Applied a modified version of Daniel Richard G's patch to create
519        pcre.h.generic and config.h.generic by "make" instead of in the
520        PrepareRelease script.
521    
522    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
523        pcre_compile.c when checking for a zero character.
524    
525    28. Introducing a native interface for JIT. Through this interface, the compiled
526        machine code can be directly executed. The purpose of this interface is to
527        provide fast pattern matching, so several sanity checks are not performed.
528        However, feature tests are still performed. The new interface provides
529        1.4x speedup compared to the old one.
530    
531    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
532        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
533        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
534    
535    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
536        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
537        the "old" RFC 2279). Instead, it ended up passing a negative length to
538        pcre_exec().
539    
540    31. Add support for GCC's visibility feature to hide internal functions.
541    
542    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
543        "unknown -C option" after outputting 0 or 1.
544    
545    33. There is now support for generating a code coverage report for the test
546        suite in environments where gcc is the compiler and lcov is installed. This
547        is mainly for the benefit of the developers.
548    
549    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
550        unaddressable using valgrind annotations, allowing valgrind to detect
551        invalid memory accesses. This is mainly for the benefit of the developers.
552    
553    25. (*UTF) can now be used to start a pattern in any of the three libraries.
554    
555    26. Give configure error if --enable-cpp but no C++ compiler found.
556    
557    
558    Version 8.31 06-July-2012
559    -------------------------
560    
561    1.  Fixing a wrong JIT test case and some compiler warnings.
562    
563    2.  Removed a bashism from the RunTest script.
564    
565    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
566        to unsigned type, result still unsigned" that was given by an MS compiler
567        on encountering the code "-sizeof(xxx)".
568    
569    4.  Partial matching support is added to the JIT compiler.
570    
571    5.  Fixed several bugs concerned with partial matching of items that consist
572        of more than one character:
573    
574        (a) /^(..)\1/ did not partially match "aba" because checking references was
575            done on an "all or nothing" basis. This also applied to repeated
576            references.
577    
578        (b) \R did not give a hard partial match if \r was found at the end of the
579            subject.
580    
581        (c) \X did not give a hard partial match after matching one or more
582            characters at the end of the subject.
583    
584        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
585            a partial match for the string "\r".
586    
587        (e) When newline was set to CRLF, the metacharacter "." did not recognize
588            a partial match for a CR character at the end of the subject string.
589    
590    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
591        running pcretest, the text "(JIT)" added to the output whenever JIT is
592        actually used to run the match.
593    
594    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
595        or /S+[+] with a digit between 1 and 7.
596    
597    8.  OP_NOT now supports any UTF character not just single-byte ones.
598    
599    9.  (*MARK) control verb is now supported by the JIT compiler.
600    
601    10. The command "./RunTest list" lists the available tests without actually
602        running any of them. (Because I keep forgetting what they all are.)
603    
604    11. Add PCRE_INFO_MAXLOOKBEHIND.
605    
606    12. Applied a (slightly modified) user-supplied patch that improves performance
607        when the heap is used for recursion (compiled with --disable-stack-for-
608        recursion). Instead of malloc and free for each heap frame each time a
609        logical recursion happens, frames are retained on a chain and re-used where
610        possible. This sometimes gives as much as 30% improvement.
611    
612    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
613        call.
614    
615    14. As documented, (*COMMIT) is now confined to within a positive assertion.
616    
617    15. It is now possible to link pcretest with libedit as an alternative to
618        libreadline.
619    
620    16. (*COMMIT) control verb is now supported by the JIT compiler.
621    
622    17. The Unicode data tables have been updated to Unicode 6.1.0.
623    
624    18. Added --file-list option to pcregrep.
625    
626    19. Added binary file support to pcregrep, including the -a, --binary-files,
627        -I, and --text options.
628    
629    20. The madvise function is renamed for posix_madvise for QNX compatibility
630        reasons. Fixed by Giuseppe D'Angelo.
631    
632    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
633    
634    22. Changed the option for creating symbolic links for 16-bit man pages from
635        -s to -sf so that re-installing does not cause issues.
636    
637    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
638    
639    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
640        matches in certain environments (the workspace was not being correctly
641        retained). Also added to pcre_dfa_exec() a simple plausibility check on
642        some of the workspace data at the beginning of a restart.
643    
644    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
645        was not doing so when it should - probably a typo introduced by SVN 528
646        (change 8.10/14).
647    
648    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
649        \w+ when the character tables indicated that \x{c4} was a word character.
650        There were several related cases, all because the tests for doing a table
651        lookup were testing for characters less than 127 instead of 255.
652    
653    27. If a pattern contains capturing parentheses that are not used in a match,
654        their slots in the ovector are set to -1. For those that are higher than
655        any matched groups, this happens at the end of processing. In the case when
656        there were back references that the ovector was too small to contain
657        (causing temporary malloc'd memory to be used during matching), and the
658        highest capturing number was not used, memory off the end of the ovector
659        was incorrectly being set to -1. (It was using the size of the temporary
660        memory instead of the true size.)
661    
662    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
663        ovector size, it uses memory at the end of the block that it has got.
664    
665    29. Check for an overlong MARK name and give an error at compile time. The
666        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
667    
668    30. JIT compiler update.
669    
670    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
671        Rill for the patch.
672    
673    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
674    
675    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
676    
677    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
678        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
679        was enabled.
680    
681    35. Improve JIT code generation for greedy plus quantifier.
682    
683    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
684        1 to "aa" instead of to an empty string. The bug affected repeated groups
685        that could potentially match an empty string.
686    
687    37. Optimizing single character iterators in JIT.
688    
689    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
690        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
691        codepoints that are too big for the mode are faulted, and in a UTF mode,
692        disallowed codepoints are also faulted.
693    
694    39. If PCRE was compiled with UTF support, in three places in the DFA
695        matcher there was code that should only have been obeyed in UTF mode, but
696        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
697        processing when bytes with values greater than 127 were present. In 16-bit
698        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
699        both cases the values are those that cannot be the first data item in a UTF
700        character. The three items that might have provoked this were recursions,
701        possessively repeated groups, and atomic groups.
702    
703    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
704        and pcregrep, because some OS require shared objects to be explicitly
705        passed to ld, causing the link step to fail if they are not.
706    
707    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
708        mode, patterns that started with \h* or \R* might be incorrectly matched.
709    
710    
711    Version 8.30 04-February-2012
712    -----------------------------
713    
714    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
715        name is defined in ctype.h.
716    
717    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
718        only in quite long subpatterns.
719    
720    3.  Removed the function pcre_info(), which has been obsolete and deprecated
721        since it was replaced by pcre_fullinfo() in February 2000.
722    
723    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
724        match a (*MARK), and the match failed at the start of the subject, a
725        reference to memory before the start of the subject could occur. This bug
726        was introduced by fix 17 of release 8.21.
727    
728    5.  A reference to an unset group with zero minimum repetition was giving
729        totally wrong answers (in non-JavaScript-compatibility mode). For example,
730        /(another)?(\1?)test/ matched against "hello world test". This bug was
731        introduced in release 8.13.
732    
733    6.  Add support for 16-bit character strings (a large amount of work involving
734        many changes and refactorings).
735    
736    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
737        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
738        from a file.
739    
740    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
741        rounding is not applied in this particular case).
742    
743    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
744        if they appear, or are escaped, in patterns.
745    
746    10. Get rid of a number of -Wunused-but-set-variable warnings.
747    
748    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
749        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
750        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
751        also returns the mark "x". This bug applied to capturing parentheses,
752        non-capturing parentheses, and atomic parentheses. It also applied to some
753        assertions.
754    
755    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
756        information out of configure.ac instead of relying on pcre.h.generic, which
757        is not stored in the repository.
758    
759    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
760        -lreadline.
761    
762    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
763    
764    15. Applied Graycode's patch to put the top-level frame on the stack rather
765        than the heap when not using the stack for recursion. This gives a
766        performance improvement in many cases when recursion is not deep.
767    
768    16. Experimental code added to "pcretest -C" to output the stack frame size.
769    
770    
771    Version 8.21 12-Dec-2011
772    ------------------------
773    
774    1.  Updating the JIT compiler.
775    
776    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
777        are added as well.
778    
779    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
780        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
781        calling _pcre_jit_exec. Some extra comments are added.
782    
783    4.  (*MARK) settings inside atomic groups that do not contain any capturing
784        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
785        was introduced by change 18 for 8.20.
786    
787    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
788        ECMA-262 standard.
789    
790    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
791        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
792        This bug was probably introduced by change 9 of 8.13.
793    
794    7.  While fixing 6 above, I noticed that a number of other items were being
795        incorrectly rejected as "not fixed length". This arose partly because newer
796        opcodes had not been added to the fixed-length checking code. I have (a)
797        corrected the bug and added tests for these items, and (b) arranged for an
798        error to occur if an unknown opcode is encountered while checking for fixed
799        length instead of just assuming "not fixed length". The items that were
800        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
801        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
802        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
803    
804    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
805        being incorrectly compiled and would have given unpredicatble results.
806    
807    9.  A possessively repeated subpattern with minimum repeat count greater than
808        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
809        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
810        the first (A) could occur when it should not.
811    
812    10. Add a cast and remove a redundant test from the code.
813    
814    11. JIT should use pcre_malloc/pcre_free for allocation.
815    
816    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
817        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
818        is anything other than /usr, -L is still shown).
819    
820    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
821    
822    14. Perl does not support \N without a following name in a [] class; PCRE now
823        also gives an error.
824    
825    15. If a forward reference was repeated with an upper limit of around 2000,
826        it caused the error "internal error: overran compiling workspace". The
827        maximum number of forward references (including repeats) was limited by the
828        internal workspace, and dependent on the LINK_SIZE. The code has been
829        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
830        the default depends on LINK_SIZE. There is a new upper limit (for safety)
831        of around 200,000 forward references. While doing this, I also speeded up
832        the filling in of repeated forward references.
833    
834    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
835        incorrectly expecting the subject to contain another "a" after the start.
836    
837    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
838        in the match, the SKIP should be ignored. This was not happening; instead
839        the SKIP was being treated as NOMATCH. For patterns such as
840        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
841        tested.
842    
843    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
844        now much more compatible with Perl, in particular in cases where the result
845        is a non-match for a non-anchored pattern. For example, if
846        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
847        "m", where previously it did not return a name. A side effect of this
848        change is that for partial matches, the last encountered mark name is
849        returned, as for non matches. A number of tests that were previously not
850        Perl-compatible have been moved into the Perl-compatible test files. The
851        refactoring has had the pleasing side effect of removing one argument from
852        the match() function, thus reducing its stack requirements.
853    
854    19. If the /S+ option was used in pcretest to study a pattern using JIT,
855        subsequent uses of /S (without +) incorrectly behaved like /S+.
856    
857    21. Retrieve executable code size support for the JIT compiler and fixing
858        some warnings.
859    
860    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
861        not work when the shorter character appeared right at the end of the
862        subject string.
863    
864    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
865        systems.
866    
867    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
868        output it when the /M option is used in pcretest.
869    
870    25. The CheckMan script was not being included in the distribution. Also, added
871        an explicit "perl" to run Perl scripts from the PrepareRelease script
872        because this is reportedly needed in Windows.
873    
874    26. If study data was being save in a file and studying had not found a set of
875        "starts with" bytes for the pattern, the data written to the file (though
876        never used) was taken from uninitialized memory and so caused valgrind to
877        complain.
878    
879    27. Updated RunTest.bat as provided by Sheri Pierce.
880    
881    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
882    
883    29. Computation of memory usage for the table of capturing group names was
884        giving an unnecessarily large value.
885    
886    
887    Version 8.20 21-Oct-2011
888    ------------------------
889    
890    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
891        a POSIX class. After further experiments with Perl, which convinced me that
892        Perl has bugs and confusions, a closing square bracket is no longer allowed
893        in a POSIX name. This bug also affected patterns with classes that started
894        with full stops.
895    
896    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
897        captured substring, but while checking the failing first alternative,
898        substring 1 is temporarily captured. If the output vector supplied to
899        pcre_exec() was not big enough for this capture, the yield of the function
900        was still zero ("insufficient space for captured substrings"). This cannot
901        be totally fixed without adding another stack variable, which seems a lot
902        of expense for a edge case. However, I have improved the situation in cases
903        such as /(a)(b)x|abc/ matched against "abc", where the return code
904        indicates that fewer than the maximum number of slots in the ovector have
905        been set.
906    
907    3.  Related to (2) above: when there are more back references in a pattern than
908        slots in the output vector, pcre_exec() uses temporary memory during
909        matching, and copies in the captures as far as possible afterwards. It was
910        using the entire output vector, but this conflicts with the specification
911        that only 2/3 is used for passing back captured substrings. Now it uses
912        only the first 2/3, for compatibility. This is, of course, another edge
913        case.
914    
915    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
916        main code base, and can be used by building with --enable-jit. When this is
917        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
918        runtime --no-jit option is given.
919    
920    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
921        ovector, the return from the function was zero, implying that there were
922        other matches that did not fit. The correct "exactly full" value is now
923        returned.
924    
925    6.  If a subpattern that was called recursively or as a subroutine contained
926        (*PRUNE) or any other control that caused it to give a non-standard return,
927        invalid errors such as "Error -26 (nested recursion at the same subject
928        position)" or even infinite loops could occur.
929    
930    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
931        computing the minimum length on reaching *ACCEPT, and so ended up with the
932        wrong value of 1 rather than 0. Further investigation indicates that
933        computing a minimum subject length in the presence of *ACCEPT is difficult
934        (think back references, subroutine calls), and so I have changed the code
935        so that no minimum is registered for a pattern that contains *ACCEPT.
936    
937    8.  If (*THEN) was present in the first (true) branch of a conditional group,
938        it was not handled as intended. [But see 16 below.]
939    
940    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
941        Sheri Pierce.
942    
943    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
944        the first byte in a match must be "a".
945    
946    11. Change 17 for 8.13 increased the recursion depth for patterns like
947        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
948        pattern contains any instances of (*THEN). If it does not, the old
949        optimizations are restored. It would be nice to do this on a per-group
950        basis, but at the moment that is not feasible.
951    
952    12. In some environments, the output of pcretest -C is CRLF terminated. This
953        broke RunTest's code that checks for the link size. A single white space
954        character after the value is now allowed for.
955    
956    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
957        For "fr", it uses the Windows-specific input and output files.
958    
959    14. If (*THEN) appeared in a group that was called recursively or as a
960        subroutine, it did not work as intended. [But see next item.]
961    
962    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
963        pattern fragments (but not containing any | characters). If A and B are
964        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
965        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
966        D. In other words, Perl considers parentheses that do not contain any |
967        characters to be part of a surrounding alternative, whereas PCRE was
968        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
969        differently. PCRE now behaves in the same way as Perl, except in the case
970        of subroutine/recursion calls such as (?1) which have in any case always
971        been different (but PCRE had them first :-).
972    
973    16. Related to 15 above: Perl does not treat the | in a conditional group as
974        creating alternatives. Such a group is treated in the same way as an
975        ordinary group without any | characters when processing (*THEN). PCRE has
976        been changed to match Perl's behaviour.
977    
978    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
979        RunGrepTest script failed.
980    
981    18. Change 22 for version 13 caused atomic groups to use more stack. This is
982        inevitable for groups that contain captures, but it can lead to a lot of
983        stack use in large patterns. The old behaviour has been restored for atomic
984        groups that do not contain any capturing parentheses.
985    
986    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
987        suppress the check for a minimum subject length at run time. (If it was
988        given to pcre_exec() or pcre_dfa_exec() it did work.)
989    
990    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
991        fail to work when decoding hex characters in data strings in EBCDIC
992        environments.
993    
994    21. It appears that in at least one Mac OS environment, the isxdigit() function
995        is implemented as a macro that evaluates to its argument more than once,
996        contravening the C 90 Standard (I haven't checked a later standard). There
997        was an instance in pcretest which caused it to go wrong when processing
998        \x{...} escapes in subject strings. The has been rewritten to avoid using
999        things like p++ in the argument of isxdigit().
1000    
1001    
1002    Version 8.13 16-Aug-2011
1003    ------------------------
1004    
1005    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1006    
1007    2.  Two minor typos in pcre_internal.h have been fixed.
1008    
1009    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1010        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1011        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1012    
1013    4.  There were a number of related bugs in the code for matching backrefences
1014        caselessly in UTF-8 mode when codes for the characters concerned were
1015        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1016        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1017        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1018        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1019        2-byte code at the end of the subject (it thought there wasn't enough data
1020        left).
1021    
1022    5.  Comprehensive information about what went wrong is now returned by
1023        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1024        as the output vector has at least 2 elements. The offset of the start of
1025        the failing character and a reason code are placed in the vector.
1026    
1027    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1028        now returned is for the first byte of the failing character, instead of the
1029        last byte inspected. This is an incompatible change, but I hope it is small
1030        enough not to be a problem. It makes the returned offset consistent with
1031        pcre_exec() and pcre_dfa_exec().
1032    
1033    7.  pcretest now gives a text phrase as well as the error number when
1034        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1035        failure, the offset and reason code are output.
1036    
1037    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1038        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1039        back over a single character (\n). This seems wrong (because it treated the
1040        two characters as a single entity when going forwards), conflicts with the
1041        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1042        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1043        has been changed.
1044    
1045    9.  Some internal refactoring has changed the processing so that the handling
1046        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1047        time (the PCRE_DOTALL option was changed this way some time ago: version
1048        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1049        which was always a bit of a fudge. It also means that there is one less
1050        argument for the match() function, which reduces its stack requirements
1051        slightly. This change also fixes an incompatibility with Perl: the pattern
1052        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1053    
1054    10. More internal refactoring has drastically reduced the number of recursive
1055        calls to match() for possessively repeated groups such as (abc)++ when
1056        using pcre_exec().
1057    
1058    11. While implementing 10, a number of bugs in the handling of groups were
1059        discovered and fixed:
1060    
1061        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1062        (a|)*(?1) gave a compile-time internal error.
1063        ((a|)+)+  did not notice that the outer group could match an empty string.
1064        (^a|^)+   was not marked as anchored.
1065        (.*a|.*)+ was not marked as matching at start or after a newline.
1066    
1067    12. Yet more internal refactoring has removed another argument from the match()
1068        function. Special calls to this function are now indicated by setting a
1069        value in a variable in the "match data" data block.
1070    
1071    13. Be more explicit in pcre_study() instead of relying on "default" for
1072        opcodes that mean there is no starting character; this means that when new
1073        ones are added and accidentally left out of pcre_study(), testing should
1074        pick them up.
1075    
1076    14. The -s option of pcretest has been documented for ages as being an old
1077        synonym of -m (show memory usage). I have changed it to mean "force study
1078        for every regex", that is, assume /S for every regex. This is similar to -i
1079        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1080        using it. It makes it easier to run collections of tests with and without
1081        study enabled, and thereby test pcre_study() more easily. All the standard
1082        tests are now run with and without -s (but some patterns can be marked as
1083        "never study" - see 20 below).
1084    
1085    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1086        restoration of the capturing data to the outer values was not happening
1087        correctly.
1088    
1089    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1090        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1091        pattern had matched an empty string, and so incorrectly returned a no
1092        match.
1093    
1094    17. There was optimizing code for the last branch of non-capturing parentheses,
1095        and also for the obeyed branch of a conditional subexpression, which used
1096        tail recursion to cut down on stack usage. Unfortunately, now that there is
1097        the possibility of (*THEN) occurring in these branches, tail recursion is
1098        no longer possible because the return has to be checked for (*THEN). These
1099        two optimizations have therefore been removed. [But see 8.20/11 above.]
1100    
1101    18. If a pattern containing \R was studied, it was assumed that \R always
1102        matched two bytes, thus causing the minimum subject length to be
1103        incorrectly computed because \R can also match just one byte.
1104    
1105    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1106        was incorrectly computed.
1107    
1108    20. If /S is present twice on a test pattern in pcretest input, it now
1109        *disables* studying, thereby overriding the use of -s on the command line
1110        (see 14 above). This is necessary for one or two tests to keep the output
1111        identical in both cases.
1112    
1113    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1114        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1115    
1116    22. When an atomic group that contained a capturing parenthesis was
1117        successfully matched, but the branch in which it appeared failed, the
1118        capturing was not being forgotten if a higher numbered group was later
1119        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1120        group 1 to "a", when in fact it should be unset. This applied to multi-
1121        branched capturing and non-capturing groups, repeated or not, and also to
1122        positive assertions (capturing in negative assertions does not happen
1123        in PCRE) and also to nested atomic groups.
1124    
1125    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1126        subject after a captured substring, to make it easier to tell which of a
1127        number of identical substrings has been captured.
1128    
1129    24. The way atomic groups are processed by pcre_exec() has been changed so that
1130        if they are repeated, backtracking one repetition now resets captured
1131        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1132        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1133        "aaa". Previously, it would have been "a". As part of this code
1134        refactoring, the way recursive calls are handled has also been changed.
1135    
1136    25. If an assertion condition captured any substrings, they were not passed
1137        back unless some other capturing happened later. For example, if
1138        (?(?=(a))a) was matched against "a", no capturing was returned.
1139    
1140    26. When studying a pattern that contained subroutine calls or assertions,
1141        the code for finding the minimum length of a possible match was handling
1142        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1143        group 1 called group 2 while simultaneously a separate group 2 called group
1144        1). A stack overflow occurred in this case. I have fixed this by limiting
1145        the recursion depth to 10.
1146    
1147    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1148        Fortmann. This supports explicit test numbers on the command line, and has
1149        argument validation and error reporting.
1150    
1151    28. An instance of \X with an unlimited repeat could fail if at any point the
1152        first character it looked at was a mark character.
1153    
1154    29. Some minor code refactoring concerning Unicode properties and scripts
1155        should reduce the stack requirement of match() slightly.
1156    
1157    30. Added the '=' option to pcretest to check the setting of unused capturing
1158        slots at the end of the pattern, which are documented as being -1, but are
1159        not included in the return count.
1160    
1161    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1162        compiled something random. Now it gives a compile-time error (as does
1163        Perl).
1164    
1165    32. A *MARK encountered during the processing of a positive assertion is now
1166        recorded and passed back (compatible with Perl).
1167    
1168    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1169        had alternative anchored branches, the search for a second match in a line
1170        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1171        matched the line "0102" twice. The same bug affected patterns that started
1172        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1173        twice.
1174    
1175    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1176        does, and because of capturing effects, quantifying parenthesized
1177        assertions may at times be useful. Quantifiers are now allowed for
1178        parenthesized assertions.
1179    
1180    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1181    
1182    36. \g was being checked for fancy things in a character class, when it should
1183        just be a literal "g".
1184    
1185    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1186        appearance of a nested POSIX class supersedes an apparent external class.
1187        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1188        unescaped square brackets may also appear as part of class names. For
1189        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1190        more like Perl. (But see 8.20/1 above.)
1191    
1192    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1193        was because it thought it was \N{name}, which is not supported).
1194    
1195    39. Add minix to OS list not supporting the -S option in pcretest.
1196    
1197    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1198        cannot analyze patterns in sufficient detail to catch mutual recursions
1199        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1200        subgroup is called recursively as a subpattern for a second time at the
1201        same position in the subject string. In previous releases this might have
1202        been caught by the recursion limit, or it might have run out of stack.
1203    
1204    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1205        happen only once. PCRE was, however incorrectly giving a compile time error
1206        "recursive call could loop indefinitely" because it cannot analyze the
1207        pattern in sufficient detail. The compile time test no longer happens when
1208        PCRE is compiling a conditional subpattern, but actual runaway loops are
1209        now caught at runtime (see 40 above).
1210    
1211    42. It seems that Perl allows any characters other than a closing parenthesis
1212        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1213        has been changed to be the same.
1214    
1215    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1216        as not to get warnings when autogen.sh is called. Also changed
1217        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1218    
1219    44. To help people who use pcregrep to scan files containing exceedingly long
1220        lines, the following changes have been made:
1221    
1222        (a) The default value of the buffer size parameter has been increased from
1223            8K to 20K. (The actual buffer used is three times this size.)
1224    
1225        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1226            PCRE is built.
1227    
1228        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1229            to be set at run time.
1230    
1231        (d) Numerical values in pcregrep options can be followed by K or M, for
1232            example --buffer-size=50K.
1233    
1234        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1235            given and the return code is set to 2.
1236    
1237    45. Add a pointer to the latest mark to the callout data block.
1238    
1239    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1240        partial match of an empty string instead of no match. This was specific to
1241        the use of ".".
1242    
1243    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1244        complete match instead of a partial match. This bug was dependent on both
1245        the PCRE_UTF8 and PCRE_DOTALL options being set.
1246    
1247    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1248        starting byte set, because \b was not being ignored.
1249    
1250    
1251    Version 8.12 15-Jan-2011
1252    ------------------------
1253    
1254    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1255        checks for such things as part of the documentation building process.
1256    
1257    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1258        --match-limit and --recursion-limit options (added for 8.11). In
1259        particular, this made one of the standard tests fail. (The integer value
1260        went into the wrong half of a long int.)
1261    
1262    3.  If the --colour option was given to pcregrep with -v (invert match), it
1263        did strange things, either producing crazy output, or crashing. It should,
1264        of course, ignore a request for colour when reporting lines that do not
1265        match.
1266    
1267    4.  Another pcregrep bug caused similar problems if --colour was specified with
1268        -M (multiline) and the pattern match finished with a line ending.
1269    
1270    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1271        matched in multiline mode, the following line was shown as part of the
1272        match. This seems wrong, so I have changed it.
1273    
1274    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1275        the check for further matches in the same line (so they could be coloured)
1276        to overrun the end of the current line. If another match was found, it was
1277        incorrectly shown (and then shown again when found in the next line).
1278    
1279    7.  If pcregrep was compiled under Windows, there was a reference to the
1280        function pcregrep_exit() before it was defined. I am assuming this was
1281        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1282        reported by a user. I've moved the definition above the reference.
1283    
1284    
1285    Version 8.11 10-Dec-2010
1286    ------------------------
1287    
1288    1.  (*THEN) was not working properly if there were untried alternatives prior
1289        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1290        backtracked to try for "b" instead of moving to the next alternative branch
1291        at the same level (in this case, to look for "c"). The Perl documentation
1292        is clear that when (*THEN) is backtracked onto, it goes to the "next
1293        alternative in the innermost enclosing group".
1294    
1295    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1296        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1297        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1298        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1299        (*THEN).
1300    
1301    3.  If \s appeared in a character class, it removed the VT character from
1302        the class, even if it had been included by some previous item, for example
1303        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1304        of \s, but is part of the POSIX "space" class.)
1305    
1306    4.  A partial match never returns an empty string (because you can always
1307        match an empty string at the end of the subject); however the checking for
1308        an empty string was starting at the "start of match" point. This has been
1309        changed to the "earliest inspected character" point, because the returned
1310        data for a partial match starts at this character. This means that, for
1311        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1312        (previously it gave "no match").
1313    
1314    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1315        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1316        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1317        has an implication that the given string is incomplete (because a partial
1318        match is preferred over a full match). For this reason, these items now
1319        give a partial match in this situation. [Aside: previously, the one case
1320        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1321        match rather than a full match, which was wrong by the old rules, but is
1322        now correct.]
1323    
1324    6.  There was a bug in the handling of #-introduced comments, recognized when
1325        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1326        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1327        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1328        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1329        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1330        places in pcre_compile().
1331    
1332    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1333        comments when looking ahead for named forward references to subpatterns,
1334        the only newline sequence it recognized was NL. It now handles newlines
1335        according to the set newline convention.
1336    
1337    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1338        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1339        cater for a lack of strerror(). These oversights have been fixed.
1340    
1341    9.  Added --match-limit and --recursion-limit to pcregrep.
1342    
1343    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1344    
1345    11. When the -o option was used, pcregrep was setting a return code of 1, even
1346        when matches were found, and --line-buffered was not being honoured.
1347    
1348    12. Added an optional parentheses number to the -o and --only-matching options
1349        of pcregrep.
1350    
1351    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1352        can match an empty string. The code to do it in pcretest and pcredemo
1353        needed fixing:
1354    
1355        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1356            only one byte after an empty string match just before CRLF (this case
1357            just got forgotten; "any" and "anycrlf" were OK).
1358    
1359        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1360            mode when an empty string match preceded an ASCII character followed by
1361            a non-ASCII character. (The code for advancing by one character rather
1362            than one byte was nonsense.)
1363    
1364        (c) The pcredemo.c sample program did not have any code at all to handle
1365            the cases when CRLF is a valid newline sequence.
1366    
1367    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1368        as a starting offset was within the subject string. There is now a new
1369        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1370        negative or greater than the length of the string. In order to test this,
1371        pcretest is extended to allow the setting of negative starting offsets.
1372    
1373    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1374        starting offset points to the beginning of a UTF-8 character was
1375        unnecessarily clumsy. I tidied it up.
1376    
1377    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1378        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1379    
1380    17. Nobody had reported that the --include_dir option, which was added in
1381        release 7.7 should have been called --include-dir (hyphen, not underscore)
1382        for compatibility with GNU grep. I have changed it to --include-dir, but
1383        left --include_dir as an undocumented synonym, and the same for
1384        --exclude-dir, though that is not available in GNU grep, at least as of
1385        release 2.5.4.
1386    
1387    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1388        characters from a string of bytes) have been redefined so as not to use
1389        loops, in order to improve performance in some environments. At the same
1390        time, I abstracted some of the common code into auxiliary macros to save
1391        repetition (this should not affect the compiled code).
1392    
1393    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1394        compile-time error is now given if \c is not followed by an ASCII
1395        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1396        different, and any byte value is allowed.)
1397    
1398    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1399        START_OPTIMIZE option, which is now allowed at compile time - but just
1400        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1401        to pcregrep and other applications that have no direct access to PCRE
1402        options. The new /Y option in pcretest sets this option when calling
1403        pcre_compile().
1404    
1405    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1406        back references. Groups containing recursive back references were forced to
1407        be atomic by that change, but in the case of named groups, the amount of
1408        memory required was incorrectly computed, leading to "Failed: internal
1409        error: code overflow". This has been fixed.
1410    
1411    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1412        pcretest.c, to avoid build problems in some Borland environments.
1413    
1414    
1415    Version 8.10 25-Jun-2010
1416    ------------------------
1417    
1418    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1419        THEN.
1420    
1421    2.  (*ACCEPT) was not working when inside an atomic group.
1422    
1423    3.  Inside a character class, \B is treated as a literal by default, but
1424        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1425        causes the error). The code is unchanged, but I tidied the documentation.
1426    
1427    4.  Inside a character class, PCRE always treated \R and \X as literals,
1428        whereas Perl faults them if its -w option is set. I have changed PCRE so
1429        that it faults them when PCRE_EXTRA is set.
1430    
1431    5.  Added support for \N, which always matches any character other than
1432        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1433    
1434    6.  When compiling pcregrep with newer versions of gcc which may have
1435        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1436        declared with attribute warn_unused_result" were given. Just casting the
1437        result to (void) does not stop the warnings; a more elaborate fudge is
1438        needed. I've used a macro to implement this.
1439    
1440    7.  Minor change to pcretest.c to avoid a compiler warning.
1441    
1442    8.  Added four artifical Unicode properties to help with an option to make
1443        \s etc use properties (see next item). The new properties are: Xan
1444        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1445    
1446    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1447        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1448        this option. Modified pcretest to add /W to test this facility. Added
1449        REG_UCP to make it available via the POSIX interface.
1450    
1451    10. Added --line-buffered to pcregrep.
1452    
1453    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1454        studied, and the match started with a letter with a code point greater than
1455        127 whose first byte was different to the first byte of the other case of
1456        the letter, the other case of this starting letter was not recognized
1457        (#976).
1458    
1459    12. If a pattern that was studied started with a repeated Unicode property
1460        test, for example, \p{Nd}+, there was the theoretical possibility of
1461        setting up an incorrect bitmap of starting bytes, but fortunately it could
1462        not have actually happened in practice until change 8 above was made (it
1463        added property types that matched character-matching opcodes).
1464    
1465    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1466        possible starting bytes for non-anchored patterns.
1467    
1468    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1469        \R, and also a number of cases that involve Unicode properties, both
1470        explicit and implicit when PCRE_UCP is set.
1471    
1472    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1473        input, it could crash or give wrong results if characters with values
1474        greater than 0xc0 were present in the subject string. (Detail: it assumed
1475        UTF-8 input when processing these items.)
1476    
1477    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1478        size_t is 64-bit (#991).
1479    
1480    17. Added a check for running out of memory when PCRE is compiled with
1481        --disable-stack-for-recursion (#990).
1482    
1483    18. If the last data line in a file for pcretest does not have a newline on
1484        the end, a newline was missing in the output.
1485    
1486    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1487        less than 128) in its various bitmaps. However, there is a facility for
1488        generating tables according to the current locale when PCRE is compiled. It
1489        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1490        characters, are recognized by isspace() and therefore were getting set in
1491        these tables, and indeed these tables seem to approximate to ISO 8859. This
1492        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1493        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1494        which of course cannot start UTF-8 characters. I have changed the code so
1495        that only real ASCII characters (less than 128) and the correct starting
1496        bytes for UTF-8 encodings are set for characters greater than 127 when in
1497        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1498        altogether.)
1499    
1500    20. Added the /T option to pcretest so as to be able to run tests with non-
1501        standard character tables, thus making it possible to include the tests
1502        used for 19 above in the standard set of tests.
1503    
1504    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1505        reference to a subpattern the other side of a comment that contains an
1506        opening parenthesis caused either an internal compiling error, or a
1507        reference to the wrong subpattern.
1508    
1509    
1510    Version 8.02 19-Mar-2010
1511    ------------------------
1512    
1513    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1514    
1515    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1516        configured.
1517    
1518    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1519        original author of that file, following a query about its status.
1520    
1521    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1522        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1523    
1524    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1525        quantifier applied to a forward-referencing subroutine call, could compile
1526        incorrect code or give the error "internal error: previously-checked
1527        referenced subpattern not found".
1528    
1529    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1530        variables to point to external functions. For these systems, therefore,
1531        pcre_malloc etc. are now initialized to local functions that call the
1532        relevant global functions.
1533    
1534    7.  There were two entries missing in the vectors called coptable and poptable
1535        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1536        I've fixed the data, and added a kludgy way of testing at compile time that
1537        the lengths are correct (equal to the number of opcodes).
1538    
1539    8.  Following on from 7, I added a similar kludge to check the length of the
1540        eint vector in pcreposix.c.
1541    
1542    9.  Error texts for pcre_compile() are held as one long string to avoid too
1543        much relocation at load time. To find a text, the string is searched,
1544        counting zeros. There was no check for running off the end of the string,
1545        which could happen if a new error number was added without updating the
1546        string.
1547    
1548    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1549    
1550    11. \K was not working if it appeared in an atomic group or in a group that
1551        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1552        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1553        the assertion is positive, but not if it is negative.
1554    
1555    12. Change 11 fortuitously reduced the size of the stack frame used in the
1556        "match()" function of pcre_exec.c by one pointer. Forthcoming
1557        implementation of support for (*MARK) will need an extra pointer on the
1558        stack; I have reserved it now, so that the stack frame size does not
1559        decrease.
1560    
1561    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1562        item in branch that calls a recursion is a subroutine call - as in the
1563        second branch in the above example - was incorrectly given the compile-
1564        time error "recursive call could loop indefinitely" because pcre_compile()
1565        was not correctly checking the subroutine for matching a non-empty string.
1566    
1567    14. The checks for overrunning compiling workspace could trigger after an
1568        overrun had occurred. This is a "should never occur" error, but it can be
1569        triggered by pathological patterns such as hundreds of nested parentheses.
1570        The checks now trigger 100 bytes before the end of the workspace.
1571    
1572    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1573    
1574    
1575    Version 8.01 19-Jan-2010
1576    ------------------------
1577    
1578    1.  If a pattern contained a conditional subpattern with only one branch (in
1579        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1580        computed the wrong minimum data length (which is of course zero for such
1581        subpatterns). This could cause incorrect "no match" results.
1582    
1583    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1584        the pattern is reset in the first branch, pcre_compile() failed with
1585        "internal error: code overflow at offset...". This happened only when
1586        the reset was to the original external option setting. (An optimization
1587        abstracts leading options settings into an external setting, which was the
1588        cause of this.)
1589    
1590    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1591        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1592        assertion pattern did not match (meaning that the assertion was true), it
1593        was incorrectly treated as false if the SKIP had been reached during the
1594        matching. This also applied to assertions used as conditions.
1595    
1596    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1597        assertion subpattern, including such a pattern used as a condition,
1598        unpredictable results occurred, instead of the error return
1599        PCRE_ERROR_DFA_UITEM.
1600    
1601    5.  The C++ GlobalReplace function was not working like Perl for the special
1602        situation when an empty string is matched. It now does the fancy magic
1603        stuff that is necessary.
1604    
1605    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1606        removed. (These were left over from very, very early versions of PCRE.)
1607    
1608    7.  Some cosmetic changes to the code to make life easier when compiling it
1609        as part of something else:
1610    
1611        (a) Change DEBUG to PCRE_DEBUG.
1612    
1613        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1614            called "current" as "current_branch", to prevent a collision with the
1615            Linux macro when compiled as a kernel module.
1616    
1617        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1618            prevent a collision with the Linux macro when compiled as a kernel
1619            module.
1620    
1621    8.  In pcre_compile() there are some checks for integer overflows that used to
1622        cast potentially large values to (double). This has been changed to that
1623        when building, a check for int64_t is made, and if it is found, it is used
1624        instead, thus avoiding the use of floating point arithmetic. (There is no
1625        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1626        double.
1627    
1628    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1629        2005 (difference between two addresses compared to an unsigned value).
1630    
1631    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1632        custom one, because of the following reported problem in Windows:
1633    
1634          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1635              under Win32.
1636          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1637              therefore missing the function definition.
1638          - The compiler thus generates a "C" signature for the test function.
1639          - The linker fails to find the "C" function.
1640          - PCRE fails to configure if asked to do so against libbz2.
1641    
1642    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1643        messages were output:
1644    
1645          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1646          rerunning libtoolize, to keep the correct libtool macros in-tree.
1647          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1648    
1649        I have done both of these things.
1650    
1651    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1652        most of the time, it *can* run out if it is given a pattern that contains a
1653        runaway infinite recursion. I updated the discussion in the pcrestack man
1654        page.
1655    
1656    13. Now that we have gone to the x.xx style of version numbers, the minor
1657        version may start with zero. Using 08 or 09 is a bad idea because users
1658        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1659        interpreted as invalid octal numbers. I've updated the previous comment in
1660        configure.ac, and also added a check that gives an error if 08 or 09 are
1661        used.
1662    
1663    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1664        causing partial matching to fail when the end of the subject matched \W
1665        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1666    
1667    15. There were some discrepancies between the declarations in pcre_internal.h
1668        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1669        their definitions. The declarations used "const uschar *" and the
1670        definitions used USPTR. Even though USPTR is normally defined as "const
1671        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1672        reported that: "This difference in casting confuses some C++ compilers, for
1673        example, SunCC recognizes above declarations as different functions and
1674        generates broken code for hbpcre." I have changed the declarations to use
1675        USPTR.
1676    
1677    16. GNU libtool is named differently on some systems. The autogen.sh script now
1678        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1679        (FreeBSD).
1680    
1681    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1682        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1683        comment: "Figure out how to create a longlong from a string: strtoll and
1684        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1685        instance, but it only takes 2 args instead of 3!"
1686    
1687    18. A subtle bug concerned with back references has been fixed by a change of
1688        specification, with a corresponding code fix. A pattern such as
1689        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1690        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1691        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1692        same bug. Such groups have to be quantified to be useful, or contained
1693        inside another quantified group. (If there's no repetition, the reference
1694        can never match.) The problem arises because, having left the group and
1695        moved on to the rest of the pattern, a later failure that backtracks into
1696        the group uses the captured value from the final iteration of the group
1697        rather than the correct earlier one. I have fixed this in PCRE by forcing
1698        any group that contains a reference to itself to be an atomic group; that
1699        is, there cannot be any backtracking into it once it has completed. This is
1700        similar to recursive and subroutine calls.
1701    
1702    
1703    Version 8.00 19-Oct-09
1704    ----------------------
1705    
1706    1.  The table for translating pcre_compile() error codes into POSIX error codes
1707        was out-of-date, and there was no check on the pcre_compile() error code
1708        being within the table. This could lead to an OK return being given in
1709        error.
1710    
1711    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1712        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1713        in a Windows environment.
1714    
1715    3.  The pcregrep --count option prints the count for each file even when it is
1716        zero, as does GNU grep. However, pcregrep was also printing all files when
1717        --files-with-matches was added. Now, when both options are given, it prints
1718        counts only for those files that have at least one match. (GNU grep just
1719        prints the file name in this circumstance, but including the count seems
1720        more useful - otherwise, why use --count?) Also ensured that the
1721        combination -clh just lists non-zero counts, with no names.
1722    
1723    4.  The long form of the pcregrep -F option was incorrectly implemented as
1724        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1725        but it seems right to fix it, and I didn't think it was worth preserving
1726        the old behaviour.
1727    
1728    5.  The command line items --regex=pattern and --regexp=pattern were not
1729        recognized by pcregrep, which required --regex pattern or --regexp pattern
1730        (with a space rather than an '='). The man page documented the '=' forms,
1731        which are compatible with GNU grep; these now work.
1732    
1733    6.  No libpcreposix.pc file was created for pkg-config; there was just
1734        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1735    
1736    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1737        when UCP support is not needed, by modifying the Python script that
1738        generates it from Unicode data files. This should not matter if the module
1739        is correctly used as a library, but I received one complaint about 50K of
1740        unwanted data. My guess is that the person linked everything into his
1741        program rather than using a library. Anyway, it does no harm.
1742    
1743    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1744        was a minimum greater than 1 for a wide character in a possessive
1745        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1746        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1747        character. Chaos in the form of incorrect output or a compiling loop could
1748        result.
1749    
1750    9.  The restrictions on what a pattern can contain when partial matching is
1751        requested for pcre_exec() have been removed. All patterns can now be
1752        partially matched by this function. In addition, if there are at least two
1753        slots in the offset vector, the offset of the earliest inspected character
1754        for the match and the offset of the end of the subject are set in them when
1755        PCRE_ERROR_PARTIAL is returned.
1756    
1757    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1758        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1759        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1760        and may be more useful for multi-segment matching.
1761    
1762    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1763        used to be given if ever the end of the subject was reached; now it is
1764        given only if matching could not proceed because another character was
1765        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1766        string "Z", which now yields "no match" instead of "partial match". In the
1767        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1768        final character ended with (*FAIL).
1769    
1770    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1771        if the pattern had a "must contain" character that was already found in the
1772        earlier partial match, unless partial matching was again requested. For
1773        example, with the pattern /dog.(body)?/, the "must contain" character is
1774        "g". If the first part-match was for the string "dog", restarting with
1775        "sbody" failed. This bug has been fixed.
1776    
1777    13. The string returned by pcre_dfa_exec() after a partial match has been
1778        changed so that it starts at the first inspected character rather than the
1779        first character of the match. This makes a difference only if the pattern
1780        starts with a lookbehind assertion or \b or \B (\K is not supported by
1781        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1782        matching functions compatible, and I think it's the right thing to do.
1783    
1784    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1785        so that the demonstration program is easily available in environments where
1786        PCRE has not been installed from source.
1787    
1788    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1789        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1790        library.
1791    
1792    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1793        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1794        is not the first non-POSIX option to be added. Clearly some people find
1795        these options useful.
1796    
1797    17. If a caller to the POSIX matching function regexec() passes a non-zero
1798        value for nmatch with a NULL value for pmatch, the value of
1799        nmatch is forced to zero.
1800    
1801    18. RunGrepTest did not have a test for the availability of the -u option of
1802        the diff command, as RunTest does. It now checks in the same way as
1803        RunTest, and also checks for the -b option.
1804    
1805    19. If an odd number of negated classes containing just a single character
1806        interposed, within parentheses, between a forward reference to a named
1807        subpattern and the definition of the subpattern, compilation crashed with
1808        an internal error, complaining that it could not find the referenced
1809        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1810        [The bug was that it was starting one character too far in when skipping
1811        over the character class, thus treating the ] as data rather than
1812        terminating the class. This meant it could skip too much.]
1813    
1814    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1815        /g option in pcretest when the pattern contains \K, which makes it possible
1816        to have an empty string match not at the start, even when the pattern is
1817        anchored. Updated pcretest and pcredemo to use this option.
1818    
1819    21. If the maximum number of capturing subpatterns in a recursion was greater
1820        than the maximum at the outer level, the higher number was returned, but
1821        with unset values at the outer level. The correct (outer level) value is
1822        now given.
1823    
1824    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1825        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1826        make it do so. The string so far is captured, making this feature
1827        compatible with Perl.
1828    
1829    23. The tests have been re-organized, adding tests 11 and 12, to make it
1830        possible to check the Perl 5.10 features against Perl 5.10.
1831    
1832    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1833        pattern matches a fixed length string. PCRE did not allow this; now it
1834        does. Neither allows recursion.
1835    
1836    25. I finally figured out how to implement a request to provide the minimum
1837        length of subject string that was needed in order to match a given pattern.
1838        (It was back references and recursion that I had previously got hung up
1839        on.) This code has now been added to pcre_study(); it finds a lower bound
1840        to the length of subject needed. It is not necessarily the greatest lower
1841        bound, but using it to avoid searching strings that are too short does give
1842        some useful speed-ups. The value is available to calling programs via
1843        pcre_fullinfo().
1844    
1845    26. While implementing 25, I discovered to my embarrassment that pcretest had
1846        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1847        study optimizations had never been tested with that matching function.
1848        Oops. What is worse, even when it was passed study data, there was a bug in
1849        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1850        were also very few tests of studied patterns with pcre_dfa_exec().
1851    
1852    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1853        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1854        on the other side of the coin, they are no longer allowed to have different
1855        names, because these cannot be distinguished in PCRE, and this has caused
1856        confusion. (This is a difference from Perl.)
1857    
1858    28. When duplicate subpattern names are present (necessarily with different
1859        numbers, as required by 27 above), and a test is made by name in a
1860        conditional pattern, either for a subpattern having been matched, or for
1861        recursion in such a pattern, all the associated numbered subpatterns are
1862        tested, and the overall condition is true if the condition is true for any
1863        one of them. This is the way Perl works, and is also more like the way
1864        testing by number works.
1865    
1866    
1867    Version 7.9 11-Apr-09
1868    ---------------------
1869    
1870    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1871        (pcretest), all targets were linked against these libraries. This included
1872        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1873        libraries. This caused unwanted dependencies to be created. This problem
1874        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1875        pcretest is linked with readline.
1876    
1877    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1878        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1879        moved outside it again, because FALSE and TRUE are already defined in AIX,
1880        but BOOL is not.
1881    
1882    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1883        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1884    
1885    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1886        hyphens) following file names and line numbers when outputting matching
1887        lines. This is not true; no spaces are inserted. I have also clarified the
1888        wording for the --colour (or --color) option.
1889    
1890    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1891        was not coloured; this is different to GNU grep, so I have changed it to be
1892        the same.
1893    
1894    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1895        each matching line was coloured. Now it goes on to look for further matches
1896        of any of the test patterns, which is the same behaviour as GNU grep.
1897    
1898    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1899        doesn't make sense to accept an empty string match in pcregrep, so I have
1900        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1901        seems to be how GNU grep behaves. [But see later change 40 for release
1902        8.33.]
1903    
1904    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1905        start or after a newline", because the conditional assertion was not being
1906        correctly handled. The rule now is that both the assertion and what follows
1907        in the first alternative must satisfy the test.
1908    
1909    9.  If auto-callout was enabled in a pattern with a conditional group whose
1910        condition was an assertion, PCRE could crash during matching, both with
1911        pcre_exec() and pcre_dfa_exec().
1912    
1913    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1914        used for matching.
1915    
1916    11. Unicode property support in character classes was not working for
1917        characters (bytes) greater than 127 when not in UTF-8 mode.
1918    
1919    12. Added the -M command line option to pcretest.
1920    
1921    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1922    
1923    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1924    
1925    16. Added comments and documentation about mis-use of no_arg in the C++
1926        wrapper.
1927    
1928    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1929        from Martin Jerabek that uses macro names for all relevant character and
1930        string constants.
1931    
1932    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1933        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1934        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1935        these, but not everybody uses configure.
1936    
1937    19. A conditional group that had only one branch was not being correctly
1938        recognized as an item that could match an empty string. This meant that an
1939        enclosing group might also not be so recognized, causing infinite looping
1940        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1941        with the subject "ab", where knowledge that the repeated group can match
1942        nothing is needed in order to break the loop.
1943    
1944    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1945        exec(), but without supplying a callout function, matching went wrong.
1946    
1947    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1948        leak if the size of the offset vector was greater than 30. When the vector
1949        is smaller, the saved offsets during recursion go onto a local stack
1950        vector, but for larger vectors malloc() is used. It was failing to free
1951        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1952        error, in fact).
1953    
1954    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1955        heapframe that is used only when UTF-8 support is enabled. This caused no
1956        problem, but was untidy.
1957    
1958    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1959        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1960        included within another project.
1961    
1962    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1963        slightly modified by me:
1964    
1965          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1966              not building pcregrep.
1967    
1968          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1969              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1970    
1971    25. Forward references, both numeric and by name, in patterns that made use of
1972        duplicate group numbers, could behave incorrectly or give incorrect errors,
1973        because when scanning forward to find the reference group, PCRE was not
1974        taking into account the duplicate group numbers. A pattern such as
1975        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1976    
1977    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1978        the feature of a custom pointer more persuasive (as requested by a user).
1979    
1980    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1981        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1982        pre-defined.
1983    
1984    28. Added support for (*UTF8) at the start of a pattern.
1985    
1986    29. Arrange for flags added by the "release type" setting in CMake to be shown
1987        in the configuration summary.
1988    
1989    
1990    Version 7.8 05-Sep-08
1991    ---------------------
1992    
1993    1.  Replaced UCP searching code with optimized version as implemented for Ad
1994        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1995        stage table and inline lookup instead of a function, giving speed ups of 2
1996        to 5 times on some simple patterns that I tested. Permission was given to
1997        distribute the MultiStage2.py script that generates the tables (it's not in
1998        the tarball, but is in the Subversion repository).
1999    
2000    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2001        scripts.
2002    
2003    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2004        a group with a zero qualifier. The result of the study could be incorrect,
2005        or the function might crash, depending on the pattern.
2006    
2007    4.  Caseless matching was not working for non-ASCII characters in back
2008        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2009        It now works when Unicode Property Support is available.
2010    
2011    5.  In pcretest, an escape such as \x{de} in the data was always generating
2012        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2013        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2014        truncation.
2015    
2016    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2017    
2018    7.  Added two (int) casts to pcregrep when printing the difference of two
2019        pointers, in case they are 64-bit values.
2020    
2021    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2022        test 2 if it fails.
2023    
2024    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2025        and a #define of that name to empty if it is not externally set. This is to
2026        allow users of MSVC to set it if necessary.
2027    
2028    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2029        the convenience functions in the pcre_get.c source file.
2030    
2031    11. An option change at the start of a pattern that had top-level alternatives
2032        could cause overwriting and/or a crash. This command provoked a crash in
2033        some environments:
2034    
2035          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2036    
2037        This potential security problem was recorded as CVE-2008-2371.
2038    
2039    12. For a pattern where the match had to start at the beginning or immediately
2040        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2041        pcre_dfa_exec() could read past the end of the passed subject if there was
2042        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2043        pcretest so that it places the subject at the end of its malloc-ed buffer.
2044    
2045    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2046        exec() might read past the end of the data buffer in UTF-8 mode.
2047    
2048    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2049        the data contained the byte 0x85 as part of a UTF-8 character within its
2050        first line. This applied both to normal and DFA matching.
2051    
2052    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2053        /^[^d]*?$/8 failed to match "abc".
2054    
2055    16. Added a missing copyright notice to pcrecpp_internal.h.
2056    
2057    17. Make it more clear in the documentation that values returned from
2058        pcre_exec() in ovector are byte offsets, not character counts.
2059    
2060    18. Tidied a few places to stop certain compilers from issuing warnings.
2061    
2062    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2063        supplied by Stefan Weber. I made a further small update for 7.8 because
2064        there is a change of source arrangements: the pcre_searchfuncs.c module is
2065        replaced by pcre_ucd.c.
2066    
2067    
2068    Version 7.7 07-May-08
2069    ---------------------
2070    
2071    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2072        a string to a long long, pretend we don't even have a long long." This is
2073        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2074    
2075    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2076        pre-7.6 versions, which defined a global no_arg variable instead of putting
2077        it in the RE class. (See also #8 below.)
2078    
2079    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2080        Lopes.
2081    
2082    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2083    
2084        (1) The include/exclude patterns were being applied to the whole pathnames
2085            of files, instead of just to the final components.
2086    
2087        (2) If there was more than one level of directory, the subdirectories were
2088            skipped unless they satisfied the include/exclude conditions. This is
2089            inconsistent with GNU grep (and could even be seen as contrary to the
2090            pcregrep specification - which I improved to make it absolutely clear).
2091            The action now is always to scan all levels of directory, and just
2092            apply the include/exclude patterns to regular files.
2093    
2094    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2095        --exclude_dir in the tests to avoid scanning .svn directories.
2096    
2097    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2098        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2099        doesn't support NULs in patterns.
2100    
2101    7.  Added some missing "const"s to declarations of static tables in
2102        pcre_compile.c and pcre_dfa_exec.c.
2103    
2104    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2105        caused by fix #2  above. (Subsequently also a second patch to fix the
2106        first patch. And a third patch - this was a messy problem.)
2107    
2108    9.  Applied Craig's patch to remove the use of push_back().
2109    
2110    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2111        matching function regexec().
2112    
2113    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2114        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2115        references. PCRE supports relative numbers with this syntax (I don't think
2116        Oniguruma does).
2117    
2118    12. Previously, a group with a zero repeat such as (...){0} was completely
2119        omitted from the compiled regex. However, this means that if the group
2120        was called as a subroutine from elsewhere in the pattern, things went wrong
2121        (an internal error was given). Such groups are now left in the compiled
2122        pattern, with a new opcode that causes them to be skipped at execution
2123        time.
2124    
2125    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2126        to the way PCRE behaves:
2127    
2128        (a) A lone ] character is dis-allowed (Perl treats it as data).
2129    
2130        (b) A back reference to an unmatched subpattern matches an empty string
2131            (Perl fails the current match path).
2132    
2133        (c) A data ] in a character class must be notated as \] because if the
2134            first data character in a class is ], it defines an empty class. (In
2135            Perl it is not possible to have an empty class.) The empty class []
2136            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2137            The negative empty class [^] matches any one character, independently
2138            of the DOTALL setting.
2139    
2140    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2141        non-existent subpattern following a character class starting with ']' and
2142        containing () gave an internal compiling error instead of "reference to
2143        non-existent subpattern". Fortunately, when the pattern did exist, the
2144        compiled code was correct. (When scanning forwards to check for the
2145        existence of the subpattern, it was treating the data ']' as terminating
2146        the class, so got the count wrong. When actually compiling, the reference
2147        was subsequently set up correctly.)
2148    
2149    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2150        it was being rejected as not supported by pcre_dfa_exec(), even though
2151        other assertions are supported. I have made pcre_dfa_exec() support
2152        (*FAIL).
2153    
2154    16. The implementation of 13c above involved the invention of a new opcode,
2155        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2156        cannot be changed at match time, I realized I could make a small
2157        improvement to matching performance by compiling OP_ALLANY instead of
2158        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2159        on the OP_ANY path.
2160    
2161    17. Compiling pcretest on Windows with readline support failed without the
2162        following two fixes: (1) Make the unistd.h include conditional on
2163        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2164    
2165    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2166        ncurses library to be included for pcretest when ReadLine support is
2167        requested, but also to allow for it to be overridden. This patch came from
2168        Daniel Bergström.
2169    
2170    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2171        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2172        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2173        spotting this.
2174    
2175    
2176    Version 7.6 28-Jan-08
2177    ---------------------
2178    
2179    1.  A character class containing a very large number of characters with
2180        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2181        overflow.
2182    
2183    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2184        HAVE_LONG_LONG is not defined.
2185    
2186    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2187        bring them up to date and include new features. This patch includes:
2188    
2189        - Fixed PH's badly added libz and libbz2 support.
2190        - Fixed a problem with static linking.
2191        - Added pcredemo. [But later removed - see 7 below.]
2192        - Fixed dftables problem and added an option.
2193        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2194            HAVE_LONG_LONG.
2195        - Added readline support for pcretest.
2196        - Added an listing of the option settings after cmake has run.
2197    
2198    4.  A user submitted a patch to Makefile that makes it easy to create
2199        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2200        Makefile.am that cause it to include this special target, without
2201        affecting anything else. Note that the same mingw target plus all
2202        the other distribution libraries and programs are now supported
2203        when configuring with CMake (see 6 below) instead of with
2204        Configure/Make.
2205    
2206    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2207        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2208        exported in the Windows port". It has not yet been confirmed that the patch
2209        solves the problem, but it does no harm.
2210    
2211    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2212        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2213        with CMake, and also correct the comment about stack recursion.
2214    
2215    7.  Remove the automatic building of pcredemo from the ./configure system and
2216        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2217        of a program that users should build themselves after PCRE is installed, so
2218        building it automatically is not really right. What is more, it gave
2219        trouble in some build environments.
2220    
2221    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2222    
2223    
2224    Version 7.5 10-Jan-08
2225    ---------------------
2226    
2227    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2228        values in parens when parsing an RE using the C++ wrapper."
2229    
2230    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2231        Characters greater than 255 were excluded from the class instead of being
2232        included.
2233    
2234    3.  The same bug as (2) above applied to negated POSIX classes such as
2235        [:^space:].
2236    
2237    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2238        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2239        I have changed it.
2240    
2241    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2242        first named subpattern) and a construct such as (?&a) would reference the
2243        first named subpattern whose name started with "a" (in other words, the
2244        length check was missing). Both these problems are fixed. "Subpattern name
2245        expected" is now given for (?&) (a zero-length name), and this patch also
2246        makes it give the same error for \k'' (previously it complained that that
2247        was a reference to a non-existent subpattern).
2248    
2249    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2250        this is right because (?- can be followed by option settings as well as by
2251        digits. I have, however, made the messages clearer.
2252    
2253    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2254        than the number used in the conditional) now cause a compile-time error.
2255        This is actually not compatible with Perl, which accepts such patterns, but
2256        treats the conditional as always being FALSE (as PCRE used to), but it
2257        seems to me that giving a diagnostic is better.
2258    
2259    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2260        and messages.
2261    
2262    9.  Fix two occurrences of "backslash" in comments that should have been
2263        "backspace".
2264    
2265    10. Remove two redundant lines of code that can never be obeyed (their function
2266        was moved elsewhere).
2267    
2268    11. The program that makes PCRE's Unicode character property table had a bug
2269        which caused it to generate incorrect table entries for sequences of
2270        characters that have the same character type, but are in different scripts.
2271        It amalgamated them into a single range, with the script of the first of
2272        them. In other words, some characters were in the wrong script. There were
2273        thirteen such cases, affecting characters in the following ranges:
2274    
2275          U+002b0 - U+002c1
2276          U+0060c - U+0060d
2277          U+0061e - U+00612
2278          U+0064b - U+0065e
2279          U+0074d - U+0076d
2280          U+01800 - U+01805
2281          U+01d00 - U+01d77
2282          U+01d9b - U+01dbf
2283          U+0200b - U+0200f
2284          U+030fc - U+030fe
2285          U+03260 - U+0327f
2286          U+0fb46 - U+0fbb1
2287          U+10450 - U+1049d
2288    
2289    12. The -o option (show only the matching part of a line) for pcregrep was not
2290        compatible with GNU grep in that, if there was more than one match in a
2291        line, it showed only the first of them. It now behaves in the same way as
2292        GNU grep.
2293    
2294    13. If the -o and -v options were combined for pcregrep, it printed a blank
2295        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2296        does the same. The return code can be used to tell if there were any
2297        non-matching lines.
2298    
2299    14. Added --file-offsets and --line-offsets to pcregrep.
2300    
2301    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2302        infinitely looping recursion. The bug was that positive lookaheads were not
2303        being skipped when checking for a possible empty match (negative lookaheads
2304        and both kinds of lookbehind were skipped).
2305    
2306    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2307        inclusion of <windows.h> to before rather than after the definition of
2308        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2309    
2310    17. Specifying a possessive quantifier with a specific limit for a Unicode
2311        character property caused pcre_compile() to compile bad code, which led at
2312        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2313        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2314        caused the error; without that there was no problem.
2315    
2316    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2317    
2318    19. Added --enable-pcretest-libreadline.
2319    
2320    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2321        RE::GlobalReplace(). As a result, the number of replacements returned was
2322        double what it should be. I removed one of the increments, but Craig sent a
2323        later patch that removed the other one (the right fix) and added unit tests
2324        that check the return values (which was not done before).
2325    
2326    21. Several CMake things:
2327    
2328        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2329            the names libpcre and libpcreposix, not just pcre and pcreposix.
2330    
2331        (2) The above change means that pcretest and pcregrep are now correctly
2332            linked with the newly-built libraries, not previously installed ones.
2333    
2334        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2335    
2336    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2337        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2338        UTF-8 newline character). The key issue is that the pattern starts .*;
2339        this means that the match must be either at the beginning, or after a
2340        newline. The bug was in the code for advancing after a failed match and
2341        checking that the new position followed a newline. It was not taking
2342        account of UTF-8 characters correctly.
2343    
2344    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2345        character classes. PCRE was not treating the sequence [:...:] as a
2346        character class unless the ... were all letters. Perl, however, seems to
2347        allow any characters between [: and :], though of course it rejects as
2348        unknown any "names" that contain non-letters, because all the known class
2349        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2350        for example, whereas PCRE did not - it did not recognize a POSIX character
2351        class. This seemed a bit dangerous, so the code has been changed to be
2352        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2353        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2354        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2355        Perl does, and where it didn't before.
2356    
2357    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2358        Windows environments %n is disabled by default.
2359    
2360    
2361    Version 7.4 21-Sep-07
2362    ---------------------
2363    
2364    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2365        means that a class such as [\s] counted as "explicit reference to CR or
2366        LF". That isn't really right - the whole point of the change was to try to
2367        help when there was an actual mention of one of the two characters. So now
2368        the change happens only if \r or \n (or a literal CR or LF) character is
2369        encountered.
2370    
2371    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2372        of both had grown to the point where there were only 3 bits left.
2373        Fortunately, there was spare space in the data structure, and so I have
2374        moved the internal flags into a new 16-bit field to free up more option
2375        bits.
2376    
2377    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2378        but did not set the internal JCHANGED flag - either of these is enough to
2379        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2380        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2381        start sets both bits.
2382    
2383    4.  Added options (at build time, compile time, exec time) to change \R from
2384        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2385    
2386    5.  doc/pcresyntax.html was missing from the distribution.
2387    
2388    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2389        compatibility, even though it is no longer used.
2390    
2391    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2392        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2393        windows.h file is present (where different names are used). [This was
2394        reversed later after testing - see 16 below.]
2395    
2396    8.  Changed all #include <config.h> to #include "config.h". There were also
2397        some further <pcre.h> cases that I changed to "pcre.h".
2398    
2399    9.  When pcregrep was used with the --colour option, it missed the line ending
2400        sequence off the lines that it output.
2401    
2402    10. It was pointed out to me that arrays of string pointers cause lots of
2403        relocations when a shared library is dynamically loaded. A technique of
2404        using a single long string with a table of offsets can drastically reduce
2405        these. I have refactored PCRE in four places to do this. The result is
2406        dramatic:
2407    
2408          Originally:                          290
2409          After changing UCP table:            187
2410          After changing error message table:   43
2411          After changing table of "verbs"       36
2412          After changing table of Posix names   22
2413    
2414        Thanks to the folks working on Gregex for glib for this insight.
2415    
2416    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2417        unicode-properties was also set.
2418    
2419    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2420    
2421    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2422        checked only for CRLF.
2423    
2424    14. Added casts to pcretest.c to avoid compiler warnings.
2425    
2426    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2427    
2428    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2429        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2430        entirely. This removes changes made in 7 above.
2431    
2432    17. The CMake files have been updated, and there is now more information about
2433        building with CMake in the NON-UNIX-USE document.
2434    
2435    
2436    Version 7.3 28-Aug-07
2437    ---------------------
2438    
2439     1. In the rejigging of the build system that eventually resulted in 7.1, the
2440        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2441        brackets there is not right, since it causes compilers to look for an
2442        installed pcre.h, not the version that is in the source that is being
2443        compiled (which of course may be different). I have changed it back to:
2444    
2445          #include "pcre.h"
2446    
2447        I have a vague recollection that the change was concerned with compiling in
2448        different directories, but in the new build system, that is taken care of
2449        by the VPATH setting the Makefile.
2450    
2451     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2452        when the subject happened to end in the byte 0x85 (e.g. if the last
2453        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2454        characters but of course it shouldn't be taken as a newline when it is part
2455        of another character. The bug was that, for an unlimited repeat of . in
2456        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2457        characters when looking for a newline.
2458    
2459     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2460    
2461     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2462        in debug output.
2463    
2464     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2465        long printing in the pcrecpp unittest when running under MinGW.
2466    
2467     6. ESC_K was left out of the EBCDIC table.
2468    
2469     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2470        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2471        limit also applies to "virtual nesting" when a pattern is recursive, and in
2472        this case 1000 isn't so big. I have been able to remove this limit at the
2473        expense of backing off one optimization in certain circumstances. Normally,
2474        when pcre_exec() would call its internal match() function recursively and
2475        immediately return the result unconditionally, it uses a "tail recursion"
2476        feature to save stack. However, when a subpattern that can match an empty
2477        string has an unlimited repetition quantifier, it no longer makes this
2478        optimization. That gives it a stack frame in which to save the data for
2479        checking that an empty string has been matched. Previously this was taken
2480        from the 1000-entry workspace that had been reserved. So now there is no
2481        explicit limit, but more stack is used.
2482    
2483     8. Applied Daniel's patches to solve problems with the import/export magic
2484        syntax that is required for Windows, and which was going wrong for the
2485        pcreposix and pcrecpp parts of the library. These were overlooked when this
2486        problem was solved for the main library.
2487    
2488     9. There were some crude static tests to avoid integer overflow when computing
2489        the size of patterns that contain repeated groups with explicit upper
2490        limits. As the maximum quantifier is 65535, the maximum group length was
2491        set at 30,000 so that the product of these two numbers did not overflow a
2492        32-bit integer. However, it turns out that people want to use groups that
2493        are longer than 30,000 bytes (though not repeat them that many times).
2494        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2495        made it possible to implement the integer overflow checks in a much more
2496        dynamic way, which I have now done. The artificial limitation on group
2497        length has been removed - we now have only the limit on the total length of
2498        the compiled pattern, which depends on the LINK_SIZE setting.
2499    
2500    10. Fixed a bug in the documentation for get/copy named substring when
2501        duplicate names are permitted. If none of the named substrings are set, the
2502        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2503        empty string.
2504    
2505    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2506        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2507        because the ] is interpreted as the first data character and the
2508        terminating ] is not found. PCRE has been made compatible with Perl in this
2509        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2510        cause memory overwriting.
2511    
2512    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2513        string has been matched (to stop an infinite loop). It was not recognizing
2514        a conditional subpattern that could match an empty string if that
2515        subpattern was within another subpattern. For example, it looped when
2516        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2517        condition was not nested. This bug has been fixed.
2518    
2519    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2520        past the start of the subject in the presence of bytes with the top bit
2521        set, for example "\x8aBCD".
2522    
2523    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2524        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2525    
2526    14. Optimized (?!) to (*FAIL).
2527    
2528    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2529        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2530        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2531        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2532        does: it's just the validity check that is more restrictive.
2533    
2534    16. Inserted checks for integer overflows during escape sequence (backslash)
2535        processing, and also fixed erroneous offset values for syntax errors during
2536        backslash processing.
2537    
2538    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2539        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2540    
2541    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2542        caused an overrun.
2543    
2544    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2545        something other than just ASCII characters) inside a group that had an
2546        unlimited repeat caused a loop at compile time (while checking to see
2547        whether the group could match an empty string).
2548    
2549    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2550        [\P{Any}] did so. (Error in the code for printing property names.)
2551    
2552    21. An orphan \E inside a character class could cause a crash.
2553    
2554    22. A repeated capturing bracket such as (A)? could cause a wild memory
2555        reference during compilation.
2556    
2557    23. There are several functions in pcre_compile() that scan along a compiled
2558        expression for various reasons (e.g. to see if it's fixed length for look
2559        behind). There were bugs in these functions when a repeated \p or \P was
2560        present in the pattern. These operators have additional parameters compared
2561        with \d, etc, and these were not being taken into account when moving along
2562        the compiled data. Specifically:
2563    
2564        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2565            length.
2566    
2567        (b) An item such as \pL+ within a repeated group could cause crashes or
2568            loops.
2569    
2570        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2571            "reference to non-existent subpattern" error.
2572    
2573        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2574    
2575    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2576        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2577    
2578    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2579    
2580    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2581        character were causing crashes (broken optimization).
2582    
2583    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2584        \p or \P) caused a compile-time loop.
2585    
2586    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2587        break. For example, the unstudied pattern [\r\n]A does not match the string
2588        "\r\nA" because change 7.0/46 below moves the current point on by two
2589        characters after failing to match at the start. However, the pattern \nA
2590        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2591        the same is true. There doesn't seem any very clean way out of this, but
2592        what I have chosen to do makes the common cases work: PCRE now takes note
2593        of whether there can be an explicit match for \r or \n anywhere in the
2594        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2595        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2596        pattern has explicit CR or LF references.
2597    
2598    29. Added (*CR) etc for changing newline setting at start of pattern.
2599    
2600    
2601    Version 7.2 19-Jun-07
2602    ---------------------
2603    
2604     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2605        which is apparently normally available under Windows.
2606    
2607     2. Re-jig the pcregrep tests with different newline settings in an attempt
2608        to make them independent of the local environment's newline setting.
2609    
2610     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2611    
2612     4. Some of the "internals" tests were previously cut out when the link size
2613        was not 2, because the output contained actual offsets. The recent new
2614        "Z" feature of pcretest means that these can be cut out, making the tests
2615        usable with all link sizes.
2616    
2617     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2618        stack recursion. This gives a massive performance boost under BSD, but just
2619        a small improvement under Linux. However, it saves one field in the frame
2620        in all cases.
2621    
2622     6. Added more features from the forthcoming Perl 5.10:
2623    
2624        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2625            recursion call. It refers to the nth most recently opened parentheses.
2626    
2627        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2628            to be opened parentheses.
2629    
2630        (c) Conditions that refer to capturing parentheses can be specified
2631            relatively, for example, (?(-2)... or (?(+3)...
2632    
2633        (d) \K resets the start of the current match so that everything before
2634            is not part of it.
2635    
2636        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2637    
2638        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2639            reference syntax.
2640    
2641        (g) (?| introduces a group in which the numbering of parentheses in each
2642            alternative starts with the same number.
2643    
2644        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2645    
2646     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2647        PCRE_INFO_JCHANGED.
2648    
2649     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2650        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2651        for detecting groups that can match an empty string.
2652    
2653     9. A pattern with a very large number of alternatives (more than several
2654        hundred) was running out of internal workspace during the pre-compile
2655        phase, where pcre_compile() figures out how much memory will be needed. A
2656        bit of new cunning has reduced the workspace needed for groups with
2657        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2658        workspace instead of running out of the 4096 that are available.
2659    
2660    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2661    
2662    11. Applied patch from Google to remove an optimization that didn't quite work.
2663        The report of the bug said:
2664    
2665          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2666          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2667          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2668    
2669    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2670        it matched the wrong number of bytes.
2671    
2672    
2673    Version 7.1 24-Apr-07
2674    ---------------------
2675    
2676     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2677        that is more "standard", making use of automake and other Autotools. There
2678        is some re-arrangement of the files and adjustment of comments consequent
2679        on this.
2680    
2681     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2682        for recursive directory scanning broke on some systems because the files
2683        are not scanned in any specific order and on different systems the order
2684        was different. A call to "sort" has been inserted into RunGrepTest for the
2685        approprate test as a short-term fix. In the longer term there may be an
2686        alternative.
2687    
2688     3. I had an email from Eric Raymond about problems translating some of PCRE's
2689        man pages to HTML (despite the fact that I distribute HTML pages, some
2690        people do their own conversions for various reasons). The problems
2691        concerned the use of low-level troff macros .br and .in. I have therefore
2692        removed all such uses from the man pages (some were redundant, some could
2693        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2694        HTML has been updated to handle .nf/.fi and to complain if it encounters
2695        .br or .in.
2696    
2697     4. Updated comments in configure.ac that get placed in config.h.in and also
2698        arranged for config.h to be included in the distribution, with the name
2699        config.h.generic, for the benefit of those who have to compile without
2700        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2701    
2702     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2703        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2704        makevp.bat for the current PCRE, using the additional files
2705        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2706    
2707     6. A Windows user reported a minor discrepancy with test 2, which turned out
2708        to be caused by a trailing space on an input line that had got lost in his
2709        copy. The trailing space was an accident, so I've just removed it.
2710    
2711     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2712        that is needed.
2713    
2714     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2715        as "const" (a) because they are and (b) because it helps the PHP
2716        maintainers who have recently made a script to detect big data structures
2717        in the php code that should be moved to the .rodata section. I remembered
2718        to update Builducptable as well, so it won't revert if ucptable.h is ever
2719        re-created.
2720    
2721     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2722        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2723        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2724        support is not required. This saves 1.5-2K of code, which is important in
2725        some applications.
2726    
2727        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2728        so as not to refer to the tables, even though these functions will never be
2729        called when UTF-8 support is disabled. Otherwise there are problems with a
2730        shared library.
2731    
2732    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2733    
2734        (a) It was defining its arguments as char * instead of void *.
2735    
2736        (b) It was assuming that all moves were upwards in memory; this was true
2737            a long time ago when I wrote it, but is no longer the case.
2738    
2739        The emulated memove() is provided for those environments that have neither
2740        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2741        is clearly not the case, as these two bugs were recently reported.
2742    
2743    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2744        and Detrail to create the HTML documentation, the .txt form of the man
2745        pages, and it removes trailing spaces from listed files. It also creates
2746        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2747        case, it wraps all the #defines with #ifndefs. This script should be run
2748        before "make dist".
2749    
2750    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2751        with Unicode property support.
2752    
2753        (a) For a maximizing quantifier, if the two different cases of the
2754            character were of different lengths in their UTF-8 codings (there are
2755            some cases like this - I found 11), and the matching function had to
2756            back up over a mixture of the two cases, it incorrectly assumed they
2757            were both the same length.
2758    
2759        (b) When PCRE was configured to use the heap rather than the stack for
2760            recursion during matching, it was not correctly preserving the data for
2761            the other case of a UTF-8 character when checking ahead for a match
2762            while processing a minimizing repeat. If the check also involved
2763            matching a wide character, but failed, corruption could cause an
2764            erroneous result when trying to check for a repeat of the original
2765            character.
2766    
2767    13. Some tidying changes to the testing mechanism:
2768    
2769        (a) The RunTest script now detects the internal link size and whether there
2770            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2771            values substituted by "configure". (The RunGrepTest script already did
2772            this for UTF-8.) The configure.ac script no longer substitutes the
2773            relevant variables.
2774    
2775        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2776            with length and offset values. This means that the output is different
2777            for different internal link sizes. Test 2 is skipped for link sizes
2778            other than 2 because of this, bypassing the problem. Unfortunately,
2779            there was also a test in test 3 (the locale tests) that used /B and
2780            failed for link sizes other than 2. Rather than cut the whole test out,
2781            I have added a new /Z option to pcretest that replaces the length and
2782            offset values with spaces. This is now used to make test 3 independent
2783            of link size. (Test 2 will be tidied up later.)
2784    
2785    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2786        segmentation fault instead of returning the appropriate error message.
2787    
2788    15. In multiline mode when the newline sequence was set to "any", the pattern
2789        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2790        This doesn't seem right; it now treats the CRLF combination as the line
2791        ending, and so does not match in that case. It's only a pattern such as ^$
2792        that would hit this one: something like ^ABC$ would have failed after \r
2793        and then tried again after \r\n.
2794    
2795    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2796        in an attempt to make files that differ only in their line terminators
2797        compare equal. This works on Linux.
2798    
2799    17. Under certain error circumstances pcregrep might try to free random memory
2800        as it exited. This is now fixed, thanks to valgrind.
2801    
2802    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2803        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2804        was because its rules for how to advance for /g after matching an empty
2805        string at the end of a line did not allow for this case. They now check for
2806        it specially.
2807    
2808    20. pcretest is supposed to handle patterns and data of any length, by
2809        extending its buffers when necessary. It was getting this wrong when the
2810        buffer for a data line had to be extended.
2811    
2812    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2813        CRLF as a newline sequence.
2814    
2815    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2816        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2817        I have nevertheless tidied it up.
2818    
2819    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2820    
2821    24. Added a man page for pcre-config.
2822    
2823    
2824    Version 7.0 19-Dec-06
2825    ---------------------
2826    
2827     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2828        moving to gcc 4.1.1.
2829    
2830     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2831        sys/time.h, which is documented as needed for this function. It doesn't
2832        seem to matter on Linux, but it showed up on some releases of OS X.
2833    
2834     3. It seems that there are systems where bytes whose values are greater than
2835        127 match isprint() in the "C" locale. The "C" locale should be the
2836        default when a C program starts up. In most systems, only ASCII printing
2837        characters match isprint(). This difference caused the output from pcretest
2838        to vary, making some of the tests fail. I have changed pcretest so that:
2839    
2840        (a) When it is outputting text in the compiled version of a pattern, bytes
2841            other than 32-126 are always shown as hex escapes.
2842    
2843        (b) When it is outputting text that is a matched part of a subject string,
2844            it does the same, unless a different locale has been set for the match
2845            (using the /L modifier). In this case, it uses isprint() to decide.
2846    
2847     4. Fixed a major bug that caused incorrect computation of the amount of memory
2848        required for a compiled pattern when options that changed within the
2849        pattern affected the logic of the preliminary scan that determines the
2850        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2851        that the computed length was too small. The symptoms of this bug were
2852        either the PCRE error "internal error: code overflow" from pcre_compile(),
2853        or a glibc crash with a message such as "pcretest: free(): invalid next
2854        size (fast)". Examples of patterns that provoked this bug (shown in
2855        pcretest format) are:
2856    
2857          /(?-x: )/x
2858          /(?x)(?-x: \s*#\s*)/
2859          /((?i)[\x{c0}])/8
2860          /(?i:[\x{c0}])/8
2861    
2862        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2863        is now done differently.
2864    
2865     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2866        wrapper classes; (b) implement a new function in the C++ scanner that is
2867        more efficient than the old way of doing things because it avoids levels of
2868        recursion in the regex matching; (c) add a paragraph to the documentation
2869        for the FullMatch() function.
2870    
2871     6. The escape sequence \n was being treated as whatever was defined as
2872        "newline". Not only was this contrary to the documentation, which states
2873        that \n is character 10 (hex 0A), but it also went horribly wrong when
2874        "newline" was defined as CRLF. This has been fixed.
2875    
2876     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2877        was being set to -1 for the "end of line" case (supposedly a value that no
2878        character can have). Though this value is never used (the check for end of
2879        line is "zero bytes in current character"), it caused compiler complaints.
2880        I've changed it to 0xffffffff.
2881    
2882     8. In pcre_version.c, the version string was being built by a sequence of
2883        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2884        string (as it is for production releases) called a macro with an empty
2885        argument. The C standard says the result of this is undefined. The gcc
2886        compiler treats it as an empty string (which was what was wanted) but it is
2887        reported that Visual C gives an error. The source has been hacked around to
2888        avoid this problem.
2889    
2890     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2891        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2892        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2893        of them did).
2894    
2895    10. Originally, pcretest opened its input and output without "b"; then I was
2896        told that "b" was needed in some environments, so it was added for release
2897        5.0 to both the input and output. (It makes no difference on Unix-like
2898        systems.) Later I was told that it is wrong for the input on Windows. I've
2899        now abstracted the modes into two macros, to make it easier to fiddle with
2900        them, and removed "b" from the input mode under Windows.
2901    
2902    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2903    
2904    12. Added -help and --help to pcretest as an official way of being reminded
2905        of the options.
2906    
2907    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2908        and pcrecpp.cc because they annoy compilers at high warning levels.
2909    
2910    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2911    
2912    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2913        scripts are not C programs :-) and which was not noticed because it works
2914        on Linux.
2915    
2916    16. pcretest is supposed to handle any length of pattern and data line (as one
2917        line or as a continued sequence of lines) by extending its input buffer if
2918        necessary. This feature was broken for very long pattern lines, leading to
2919        a string of junk being passed to pcre_compile() if the pattern was longer
2920        than about 50K.
2921    
2922    17. I have done a major re-factoring of the way pcre_compile() computes the
2923        amount of memory needed for a compiled pattern. Previously, there was code
2924        that made a preliminary scan of the pattern in order to do this. That was
2925        OK when PCRE was new, but as the facilities have expanded, it has become
2926        harder and harder to keep it in step with the real compile phase, and there
2927        have been a number of bugs (see for example, 4 above). I have now found a
2928        cunning way of running the real compile function in a "fake" mode that
2929        enables it to compute how much memory it would need, while actually only
2930        ever using a few hundred bytes of working memory and without too many
2931        tests of the mode. This should make future maintenance and development
2932        easier. A side effect of this work is that the limit of 200 on the nesting
2933        depth of parentheses has been removed (though this was never a serious
2934        limitation, I suspect). However, there is a downside: pcre_compile() now
2935        runs more slowly than before (30% or more, depending on the pattern). I
2936        hope this isn't a big issue. There is no effect on runtime performance.
2937    
2938    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2939        newline (only possible for the last line of a file) and it was a
2940        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2941    
2942    19. Added additional timing features to pcretest. (1) The -tm option now times
2943        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2944        separate command line item, by a number that specifies the number of
2945        repeats to use when timing. The default is 50000; this gives better
2946        precision, but takes uncomfortably long for very large patterns.
2947    
2948    20. Extended pcre_study() to be more clever in cases where a branch of a
2949        subpattern has no definite first character. For example, (a*|b*)[cd] would
2950        previously give no result from pcre_study(). Now it recognizes that the
2951        first character must be a, b, c, or d.
2952    
2953    21. There was an incorrect error "recursive call could loop indefinitely" if
2954        a subpattern (or the entire pattern) that was being tested for matching an
2955        empty string contained only one non-empty item after a nested subpattern.
2956        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2957        incorrectly, because the \d was being skipped in the check.
2958    
2959    22. The pcretest program now has a new pattern option /B and a command line
2960        option -b, which is equivalent to adding /B to every pattern. This causes
2961        it to show the compiled bytecode, without the additional information that
2962        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2963        is the same as /B/I).
2964    
2965    23. A new optimization is now able automatically to treat some sequences such
2966        as a*b as a*+b. More specifically, if something simple (such as a character
2967        or a simple class like \d) has an unlimited quantifier, and is followed by
2968        something that cannot possibly match the quantified thing, the quantifier
2969        is automatically "possessified".
2970    
2971    24. A recursive reference to a subpattern whose number was greater than 39
2972        went wrong under certain circumstances in UTF-8 mode. This bug could also
2973        have affected the operation of pcre_study().
2974    
2975    25. Realized that a little bit of performance could be had by replacing
2976        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2977    
2978    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2979    
2980    27. Possessive quantifiers such as a++ were previously implemented by turning
2981        them into atomic groups such as ($>a+). Now they have their own opcodes,
2982        which improves performance. This includes the automatically created ones
2983        from 23 above.
2984    
2985    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2986        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2987        the first matched character to be a colon. This applied both to named and
2988        numbered groups.
2989    
2990    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2991    
2992    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2993        building PCRE on Windows easier, so I have included it in the distribution.
2994    
2995    31. There is now a check in pcretest against a ridiculously large number being
2996        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2997        loop, the loop is abandoned.
2998    
2999    32. Forward references to subpatterns in conditions such as (?(2)...) where
3000        subpattern 2 is defined later cause pcre_compile() to search forwards in
3001        the pattern for the relevant set of parentheses. This search went wrong
3002        when there were unescaped parentheses in a character class, parentheses
3003        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3004    
3005    33. "Subroutine" calls and backreferences were previously restricted to
3006        referencing subpatterns earlier in the regex. This restriction has now
3007        been removed.
3008    
3009    34. Added a number of extra features that are going to be in Perl 5.10. On the
3010        whole, these are just syntactic alternatives for features that PCRE had
3011        previously implemented using the Python syntax or my own invention. The
3012        other formats are all retained for compatibility.
3013    
3014        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3015            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3016            also .NET compatible.
3017    
3018        (b) A recursion or subroutine call to a named group can now be defined as
3019            (?&name) as well as (?P>name).
3020    
3021        (c) A backreference to a named group can now be defined as \k<name> or
3022            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3023            5.10, are also .NET compatible.
3024    
3025        (d) A conditional reference to a named group can now use the syntax
3026            (?(<name>) or (?('name') as well as (?(name).
3027    
3028        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3029            groups (named and numbered) that are never evaluated inline, but can be
3030            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3031            is always false. There may be only one alternative in such a group.
3032    
3033        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3034            as the simple (?(R). The condition is true only if the most recent
3035            recursion is that of the given number or name. It does not search out
3036            through the entire recursion stack.
3037    
3038        (g) The escape \gN or \g{N} has been added, where N is a positive or
3039            negative number, specifying an absolute or relative reference.
3040    
3041    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3042        some "unreachable code" warnings.
3043    
3044    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3045        things, this adds five new scripts.
3046    
3047    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3048        There were also incompatibilities regarding the handling of \Q..\E inside
3049        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3050        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3051    
3052    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3053        matches an empty string, and forcibly breaks the loop. There were bugs in
3054        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3055        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3056        separate and independent bugs (that affected different cases) have been
3057        fixed.
3058    
3059    39. Refactored the code to abolish the use of different opcodes for small
3060        capturing bracket numbers. This is a tidy that I avoided doing when I
3061        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3062        The new approach is not only tidier, it makes it possible to reduce the
3063        memory needed to fix the previous bug (38).
3064    
3065    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3066        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3067        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3068        mode.
3069    
3070    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3071        report.
3072    
3073    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3074        copy construction and assignment in the C++ wrapper.
3075    
3076    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3077        couple of bugs that could have given wrong results in the "--newline=crlf"
3078        case.
3079    
3080    44. Added a number of casts and did some reorganization of signed/unsigned int
3081        variables following suggestions from Dair Grant. Also renamed the variable
3082        "this" as "item" because it is a C++ keyword.
3083    
3084    45. Arranged for dftables to add
3085    
3086          #include "pcre_internal.h"
3087    
3088        to pcre_chartables.c because without it, gcc 4.x may remove the array
3089        definition from the final binary if PCRE is built into a static library and
3090        dead code stripping is activated.
3091    
3092    46. For an unanchored pattern, if a match attempt fails at the start of a
3093        newline sequence, and the newline setting is CRLF or ANY, and the next two
3094        characters are CRLF, advance by two characters instead of one.
3095    
3096    
3097    Version 6.7 04-Jul-06
3098    ---------------------
3099    
3100     1. In order to handle tests when input lines are enormously long, pcretest has
3101        been re-factored so that it automatically extends its buffers when
3102        necessary. The code is crude, but this _is_ just a test program. The
3103        default size has been increased from 32K to 50K.
3104    
3105     2. The code in pcre_study() was using the value of the re argument before
3106        testing it for NULL. (Of course, in any sensible call of the function, it
3107        won't be NULL.)
3108    
3109     3. The memmove() emulation function in pcre_internal.h, which is used on
3110        systems that lack both memmove() and bcopy() - that is, hardly ever -
3111        was missing a "static" storage class specifier.
3112    
3113     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3114        containing an extended class (one that cannot be represented by a bitmap
3115        because it contains high-valued characters or Unicode property items, e.g.
3116        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3117        pattern, but PCRE should not loop if you do not (it no longer does).
3118        [Detail: two cases were found: (a) a repeated subpattern containing an
3119        extended class; (b) a recursive reference to a subpattern that followed a
3120        previous extended class. It wasn't skipping over the extended class
3121        correctly when UTF-8 mode was not set.]
3122    
3123     5. A negated single-character class was not being recognized as fixed-length
3124        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3125        compile error "lookbehind assertion is not fixed length".
3126    
3127     6. The RunPerlTest auxiliary script was showing an unexpected difference
3128        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3129        write a Perl script that can interpret lines of an input file either as
3130        byte characters or as UTF-8, which is what "perltest" was being required to
3131        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3132        can't do is switch easily at run time between having the "use utf8;" pragma
3133        or not. In the end, I fudged it by using the RunPerlTest script to insert
3134        "use utf8;" explicitly for the UTF-8 tests.
3135    
3136     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3137        the end of the subject string, contrary to the documentation and to what
3138        Perl does. This was true of both matching functions. Now it matches only at
3139        the start of the subject and immediately after *internal* newlines.
3140    
3141     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3142        a pointer to an int instead of a pointer to an unsigned long int. This
3143        caused problems on 64-bit systems.
3144    
3145     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3146        instance of the 'standard' template library not being so standard".
3147    
3148    10. There was no check on the number of named subpatterns nor the maximum
3149        length of a subpattern name. The product of these values is used to compute
3150        the size of the memory block for a compiled pattern. By supplying a very
3151        long subpattern name and a large number of named subpatterns, the size
3152        computation could be caused to overflow. This is now prevented by limiting
3153        the length of names to 32 characters, and the number of named subpatterns
3154        to 10,000.
3155    
3156    11. Subpatterns that are repeated with specific counts have to be replicated in
3157        the compiled pattern. The size of memory for this was computed from the
3158        length of the subpattern and the repeat count. The latter is limited to
3159        65535, but there was no limit on the former, meaning that integer overflow
3160        could in principle occur. The compiled length of a repeated subpattern is
3161        now limited to 30,000 bytes in order to prevent this.
3162    
3163    12. Added the optional facility to have named substrings with the same name.
3164    
3165    13. Added the ability to use a named substring as a condition, using the
3166        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3167        are numbers (not recommended). Forward references are permitted.
3168    
3169    14. Added forward references in named backreferences (if you see what I mean).
3170    
3171    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3172        pattern could run off the end of the subject. For example, the pattern
3173        "(?s)(.{1,5})"8 did this with the subject "ab".
3174    
3175    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3176        PCRE_CASELESS was set when matching characters that were quantified with ?
3177        or *.
3178    
3179    17. A character class other than a single negated character that had a minimum
3180        but no maximum quantifier - for example [ab]{6,} - was not handled
3181        correctly by pce_dfa_exec(). It would match only one character.
3182    
3183    18. A valid (though odd) pattern that looked like a POSIX character
3184        class but used an invalid character after [ (for example [[,abc,]]) caused
3185        pcre_compile() to give the error "Failed: internal error: code overflow" or
3186        in some cases to crash with a glibc free() error. This could even happen if
3187        the pattern terminated after [[ but there just happened to be a sequence of
3188        letters, a binary zero, and a closing ] in the memory that followed.
3189    
3190    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3191        over the years. Originally (before any Unicode support), just the bottom 8
3192        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3193        output from "man perlunicode" includes this:
3194    
3195          The regular expression compiler produces polymorphic opcodes.  That
3196          is, the pattern adapts to the data and automatically switches to
3197          the Unicode character scheme when presented with Unicode data--or
3198          instead uses a traditional byte scheme when presented with byte
3199          data.
3200    
3201        Sadly, a wide octal escape does not cause a switch, and in a string with
3202        no other multibyte characters, these octal escapes are treated as before.
3203        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3204        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3205        Unicode string.
3206    
3207        I have not perpetrated such confusion in PCRE. Up till now, it took just
3208        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3209        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3210        translate to the appropriate multibyte character.
3211    
3212    29. Applied some refactoring to reduce the number of warnings from Microsoft
3213        and Borland compilers. This has included removing the fudge introduced
3214        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3215        a warning about an unused variable.
3216    
3217    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3218        characters since release 4.0, because Perl (from release 5.004) does not.
3219        [Or at least, is documented not to: some releases seem to be in conflict
3220        with the documentation.] However, when a pattern was studied with
3221        pcre_study() and all its branches started with \s, PCRE still included VT
3222        as a possible starting character. Of course, this did no harm; it just
3223        caused an unnecessary match attempt.
3224    
3225    22. Removed a now-redundant internal flag bit that recorded the fact that case
3226        dependency changed within the pattern. This was once needed for "required
3227        byte" processing, but is no longer used. This recovers a now-scarce options
3228        bit. Also moved the least significant internal flag bit to the most-
3229        significant bit of the word, which was not previously used (hangover from
3230        the days when it was an int rather than a uint) to free up another bit for
3231        the future.
3232    
3233    23. Added support for CRLF line endings as well as CR and LF. As well as the
3234        default being selectable at build time, it can now be changed at runtime
3235        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3236        specify that it is scanning data with non-default line endings.
3237    
3238    24. Changed the definition of CXXLINK to make it agree with the definition of
3239        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3240    
3241    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3242        recursions. This makes a big different to stack usage for some patterns.
3243    
3244    26. If a subpattern containing a named recursion or subroutine reference such
3245        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3246        the space required for the compiled pattern went wrong and gave too small a
3247        value. Depending on the environment, this could lead to "Failed: internal
3248        error: code overflow at offset 49" or "glibc detected double free or
3249        corruption" errors.
3250    
3251    27. Applied patches from Google (a) to support the new newline modes and (b) to
3252        advance over multibyte UTF-8 characters in GlobalReplace.
3253    
3254    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3255        difference for some implementation of PCRE in some Windows version.
3256    
3257    29. Added some extra testing facilities to pcretest:
3258    
3259        \q<number>   in a data line sets the "match limit" value
3260        \Q<number>   in a data line sets the "match recursion limt" value
3261        -S <number>  sets the stack size, where <number> is in megabytes
3262    
3263        The -S option isn't available for Windows.
3264    
3265    
3266    Version 6.6 06-Feb-06
3267    ---------------------
3268    
3269     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3270        in pcreposix.h. I have copied the definition from pcre.h.
3271    
3272     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3273        because pcre.h is no longer a built file.
3274    
3275     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3276        not normally included in the compiled code.
3277    
3278    
3279    Version 6.5 01-Feb-06
3280    ---------------------
3281    
3282     1. When using the partial match feature with pcre_dfa_exec(), it was not
3283        anchoring the second and subsequent partial matches at the new starting
3284        point. This could lead to incorrect results. For example, with the pattern
3285        /1234/, partially matching against "123" and then "a4" gave a match.
3286    
3287     2. Changes to pcregrep:
3288    
3289        (a) All non-match returns from pcre_exec() were being treated as failures
3290            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3291            error message is output. Some extra information is given for the
3292            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3293            probably the only errors that are likely to be caused by users (by
3294            specifying a regex that has nested indefinite repeats, for instance).
3295            If there are more than 20 of these errors, pcregrep is abandoned.
3296    
3297        (b) A binary zero was treated as data while matching, but terminated the
3298            output line if it was written out. This has been fixed: binary zeroes
3299            are now no different to any other data bytes.
3300    
3301        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3302            used to set a locale for matching. The --locale=xxxx long option has
3303            been added (no short equivalent) to specify a locale explicitly on the
3304            pcregrep command, overriding the environment variables.
3305    
3306        (d) When -B was used with -n, some line numbers in the output were one less
3307            than they should have been.
3308    
3309        (e) Added the -o (--only-matching) option.
3310    
3311        (f) If -A or -C was used with -c (count only), some lines of context were
3312            accidentally printed for the final match.
3313    
3314        (g) Added the -H (--with-filename) option.
3315    
3316        (h) The combination of options -rh failed to suppress file names for files
3317            that were found from directory arguments.
3318    
3319        (i) Added the -D (--devices) and -d (--directories) options.
3320    
3321        (j) Added the -F (--fixed-strings) option.
3322    
3323        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3324    
3325        (l) Added the --colo(u)r option.
3326    
3327        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3328            is not present by default.
3329    
3330     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3331        items such as (?R) or (?1), when the recursion could match a number of
3332        alternatives. If it matched one of the alternatives, but subsequently,
3333        outside the recursion, there was a failure, the code tried to back up into
3334        the recursion. However, because of the way PCRE is implemented, this is not
3335        possible, and the result was an incorrect result from the match.
3336    
3337        In order to prevent this happening, the specification of recursion has
3338        been changed so that all such subpatterns are automatically treated as
3339        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3340    
3341     4. I had overlooked the fact that, in some locales, there are characters for
3342        which isalpha() is true but neither isupper() nor islower() are true. In
3343        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3344        and ordfeminine) are like this. This affected the treatment of \w and \W
3345        when they appeared in character classes, but not when they appeared outside
3346        a character class. The bit map for "word" characters is now created
3347        separately from the results of isalnum() instead of just taking it from the
3348        upper, lower, and digit maps. (Plus the underscore character, of course.)
3349    
3350     5. The above bug also affected the handling of POSIX character classes such as
3351        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3352        permanent tables. Instead, the bit maps for such a class were previously
3353        created as the appropriate unions of the upper, lower, and digit bitmaps.
3354        Now they are created by subtraction from the [[:word:]] class, which has
3355        its own bitmap.
3356    
3357     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3358        It is created by subtracting the vertical space characters (\x09, \x0a,
3359        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3360        subtraction was done in the overall bitmap for a character class, meaning
3361        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3362        be recognized. This bug has been fixed.
3363    
3364     7. Patches from the folks at Google:
3365    
3366          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3367          real life, but is still worth protecting against".
3368    
3369          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3370          regular expressions".
3371    
3372          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3373          have it.
3374    
3375          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3376          "configure" and the latter not, in order to fix a problem somebody had
3377          with compiling the Arg class on HP-UX.
3378    
3379          (e) Improve the error-handling of the C++ wrapper a little bit.
3380    
3381          (f) New tests for checking recursion limiting.
3382    
3383     8. The pcre_memmove() function, which is used only if the environment does not
3384        have a standard memmove() function (and is therefore rarely compiled),
3385        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3386        returning a result (though PCRE never actually uses the result).
3387    
3388     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3389        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3390        returned instead of calling malloc() with an overflowing number that would
3391        most likely cause subsequent chaos.
3392    
3393    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3394    
3395    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3396        with this option is matched, the nmatch and pmatch options of regexec() are
3397        ignored.
3398    
3399    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3400        provided in case anyone wants to the the POSIX interface with UTF-8
3401        strings.
3402    
3403    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3404        C++ linking (needed for some HP-UX environments).
3405    
3406    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3407        (unused parameter) and in the pcre_printint() function (omitted "default"
3408        switch label when the default is to do nothing).
3409    
3410    15. Added some code to make it possible, when PCRE is compiled as a C++
3411        library, to replace subject pointers for pcre_exec() with a smart pointer
3412        class, thus making it possible to process discontinuous strings.
3413    
3414    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3415        much the same function. They were added by different people who were trying
3416        to make PCRE easy to compile on non-Unix systems. It has been suggested
3417        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3418        for compiling on Windows systems. I have therefore replaced it with
3419        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3420        defaults to "extern" for C or "extern C" for C++, which works fine on
3421        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3422        SCOPE with something explicit in config.h. In addition:
3423    
3424        (a) pcreposix.h still had just "extern" instead of either of these macros;
3425            I have replaced it with PCRE_DATA_SCOPE.
3426    
3427        (b) Functions such as _pcre_xclass(), which are internal to the library,
3428            but external in the C sense, all had PCRE_EXPORT in their definitions.
3429            This is apparently wrong for the Windows case, so I have removed it.
3430            (It makes no difference on Unix-like systems.)
3431    
3432    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3433        of recursive calls to match(). This is different to MATCH_LIMIT because
3434        that limits the total number of calls to match(), not all of which increase
3435        the depth of recursion. Limiting the recursion depth limits the amount of
3436        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3437        when PCRE is compiled, and changed at run time. A patch from Google adds
3438        this functionality to the C++ interface.
3439    
3440    18. Changes to the handling of Unicode character properties:
3441    
3442        (a) Updated the table to Unicode 4.1.0.
3443    
3444        (b) Recognize characters that are not in the table as "Cn" (undefined).
3445    
3446        (c) I revised the way the table is implemented to a much improved format
3447            which includes recognition of ranges. It now supports the ranges that
3448            are defined in UnicodeData.txt, and it also amalgamates other
3449            characters into ranges. This has reduced the number of entries in the
3450            table from around 16,000 to around 3,000, thus reducing its size
3451            considerably. I realized I did not need to use a tree structure after
3452            all - a binary chop search is just as efficient. Having reduced the
3453            number of entries, I extended their size from 6 bytes to 8 bytes to
3454            allow for more data.
3455    
3456        (d) Added support for Unicode script names via properties such as \p{Han}.
3457    
3458    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3459        matching that character.
3460    
3461    20. When matching a repeated Unicode property with a minimum greater than zero,
3462        (for example \pL{2,}), PCRE could look past the end of the subject if it
3463        reached it while seeking the minimum number of characters. This could
3464        happen only if some of the characters were more than one byte long, because
3465        there is a check for at least the minimum number of bytes.
3466    
3467    21. Refactored the implementation of \p and \P so as to be more general, to
3468        allow for more different types of property in future. This has changed the
3469        compiled form incompatibly. Anybody with saved compiled patterns that use
3470        \p or \P will have to recompile them.
3471    
3472    22. Added "Any" and "L&" to the supported property types.
3473    
3474    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3475        but give a compile time error if the value is greater than 0xff.
3476    
3477    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3478        accidentally not being installed or uninstalled.
3479    
3480    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3481        made were to insert the current release number. This seemed silly, because
3482        it made things harder for people building PCRE on systems that don't run
3483        "configure". I have turned pcre.h into a distributed file, no longer built
3484        by "configure", with the version identification directly included. There is
3485        no longer a pcre.h.in file.
3486    
3487        However, this change necessitated a change to the pcre-config script as
3488        well. It is built from pcre-config.in, and one of the substitutions was the
3489        release number. I have updated configure.ac so that ./configure now finds
3490        the release number by grepping pcre.h.
3491    
3492    26. Added the ability to run the tests under valgrind.
3493    
3494    
3495    Version 6.4 05-Sep-05
3496    ---------------------
3497    
3498     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3499        "--" to be printed when multiple files were scanned, even when none of the
3500        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3501        consider it to be a bug, and have restored the previous behaviour.
3502    
3503     2. A couple of code tidies to get rid of compiler warnings.
3504    
3505     3. The pcretest program used to cheat by referring to symbols in the library
3506        whose names begin with _pcre_. These are internal symbols that are not
3507        really supposed to be visible externally, and in some environments it is
3508        possible to suppress them. The cheating is now confined to including
3509        certain files from the library's source, which is a bit cleaner.
3510    
3511     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3512        file's purpose clearer.
3513    
3514     5. Reorganized pcre_ucp_findchar().
3515    
3516    
3517    Version 6.3 15-Aug-05
3518    ---------------------
3519    
3520     1. The file libpcre.pc.in did not have general read permission in the tarball.
3521    
3522     2. There were some problems when building without C++ support:
3523    
3524        (a) If C++ support was not built, "make install" and "make test" still
3525            tried to test it.
3526    
3527        (b) There were problems when the value of CXX was explicitly set. Some
3528            changes have been made to try to fix these, and ...
3529    
3530        (c) --disable-cpp can now be used to explicitly disable C++ support.
3531    
3532        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3533            backslash in a target when C++ was disabled. This confuses some
3534            versions of "make", apparently. Using an intermediate variable solves
3535            this. (Same for CPP_LOBJ.)
3536    
3537     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3538        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3539        necessary on certain architectures.
3540    
3541     4. Added a setting of -export-symbols-regex to the link command to remove
3542        those symbols that are exported in the C sense, but actually are local
3543        within the library, and not documented. Their names all begin with
3544        "_pcre_". This is not a perfect job, because (a) we have to except some
3545        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3546        available (and never for static libraries). I have made a note to try to
3547        find a way round (a) in the future.
3548    
3549    
3550    Version 6.2 01-Aug-05
3551    ---------------------
3552    
3553     1. There was no test for integer overflow of quantifier values. A construction
3554        such as {1111111111111111} would give undefined results. What is worse, if
3555        a minimum quantifier for a parenthesized subpattern overflowed and became
3556        negative, the calculation of the memory size went wrong. This could have
3557        led to memory overwriting.
3558    
3559     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3560    
3561     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3562        operating environments where this matters.
3563    
3564     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3565        PCRE options from within the C++ wrapper.
3566    
3567     5. Named capturing subpatterns were not being correctly counted when a pattern
3568        was compiled. This caused two problems: (a) If there were more than 100
3569        such subpatterns, the calculation of the memory needed for the whole
3570        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3571        back references of the form \12, where the number was greater than 9, were
3572        not recognized as back references, even though there were sufficient
3573        previous subpatterns.
3574    
3575     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3576        versions of gcc, e.g. 2.95.4.
3577    
3578    
3579    Version 6.1 21-Jun-05
3580    ---------------------
3581    
3582     1. There was one reference to the variable "posix" in pcretest.c that was not