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

Diff of /code/trunk/ChangeLog

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

revision 172 by ph10, Tue Jun 5 10:40:13 2007 UTC revision 1438 by ph10, Fri Jan 10 16:13:10 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 05-June-07  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    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
19        of a match later then the end of the match. The pcretest program was not
20        handling the case sensibly - it was outputting from the start to the next
21        binary zero. It now reports this situation in a message, and outputs the
22        text from the end to the start.
23    
24    5.  Fast forward search is improved in JIT. Instead of the first three
25        characters, any three characters with fixed position can be searched.
26        Search order: first, last, middle.
27    
28    6.  Improve character range checks in JIT. Characters are read by an inprecise
29        function now, which returns with an unknown value if the character code is
30        above a certain treshold (e.g: 256). The only limitation is that the value
31        must be bigger than the treshold as well. This function is useful, when
32        the characters above the treshold are handled in the same way.
33    
34    7.  The macros whose names start with RAWUCHAR are placeholders for a future
35        mode in which only the bottom 21 bits of 32-bit data items are used. To
36        make this more memorable for those maintaining the code, the names have
37        been changed to start with UCHAR21, and an extensive comment has been added
38        to their definition.
39    
40    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
41        to the export list in Makefile.am (they were accidentally omitted from the
42        8.34 tarball).
43    
44    9.  The informational output from pcretest used the phrase "starting byte set"
45        which is inappropriate for the 16-bit and 32-bit libraries. As the output
46        for "first char" and "need char" really means "non-UTF-char", I've changed
47        "byte" to "char", and slightly reworded the output. The documentation about
48        these values has also been (I hope) clarified.
49    
50    10. Another JIT related optimization: use table jumps for selecting the correct
51        backtracking path, when more than four alternatives are present inside a
52        bracket.
53    
54    11. Empty match is not possible, when the minimum length is greater than zero,
55        and there is no \K in the pattern. Remove these unnecessary checks form JIT.
56    
57    12. In a caseless character class with UCP support, when a character with more
58        than one alternative case was not the first character of a range, not all
59        the alternative cases were added to the class. For example, s and \x{17f}
60        are both alternative cases for S: the class [RST] was handled correctly,
61        but [R-T] was not.
62    
63    
64    Version 8.34 15-December-2013
65    -----------------------------
66    
67    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
68        executable memory. Patch inspired by Carsten Klein.
69    
70    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
71        this macro is never tested and has no effect, because the work to support
72        coverage involves only compiling and linking options and special targets in
73        the Makefile. The comment in config.h implied that defining the macro would
74        enable coverage support, which is totally false. There was also support for
75        setting this macro in the CMake files (my fault, I just copied it from
76        configure). SUPPORT_GCOV has now been removed.
77    
78    3.  Make a small performance improvement in strlen16() and strlen32() in
79        pcretest.
80    
81    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
82        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
83        these cases). There was also one in pcretest.c.
84    
85    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
86    
87    6.  In UTF mode, the code for checking whether a group could match an empty
88        string (which is used for indefinitely repeated groups to allow for
89        breaking an infinite loop) was broken when the group contained a repeated
90        negated single-character class with a character that occupied more than one
91        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
92        UTF-8 mode). The effect was undefined: the group might or might not be
93        deemed as matching an empty string, or the program might have crashed.
94    
95    7.  The code for checking whether a group could match an empty string was not
96        recognizing that \h, \H, \v, \V, and \R must match a character.
97    
98    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
99        an empty string. If it can, pcretest shows this in its information output.
100    
101    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
102        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
103        matched by pcre_exec() without using JIT:
104    
105        (a) If the rest of the pattern did not match after a maximal run of
106            grapheme clusters, the code for backing up to try with fewer of them
107            did not always back up over a full grapheme when characters that do not
108            have the modifier quality were involved, e.g. Hangul syllables.
109    
110        (b) If the match point in a subject started with modifier character, and
111            there was no match, the code could incorrectly back up beyond the match
112            point, and potentially beyond the first character in the subject,
113            leading to a segfault or an incorrect match result.
114    
115    10. A conditional group with an assertion condition could lead to PCRE
116        recording an incorrect first data item for a match if no other first data
117        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
118        first data item, and therefore matched "ca" after "c" instead of at the
119        start.
120    
121    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
122        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
123    
124    12. The source of pcregrep now includes z/OS-specific code so that it can be
125        compiled for z/OS as part of the special z/OS distribution.
126    
127    13. Added the -T and -TM options to pcretest.
128    
129    14. The code in pcre_compile.c for creating the table of named capturing groups
130        has been refactored. Instead of creating the table dynamically during the
131        actual compiling pass, the information is remembered during the pre-compile
132        pass (on the stack unless there are more than 20 named groups, in which
133        case malloc() is used) and the whole table is created before the actual
134        compile happens. This has simplified the code (it is now nearly 150 lines
135        shorter) and prepared the way for better handling of references to groups
136        with duplicate names.
137    
138    15. A back reference to a named subpattern when there is more than one of the
139        same name now checks them in the order in which they appear in the pattern.
140        The first one that is set is used for the reference. Previously only the
141        first one was inspected. This change makes PCRE more compatible with Perl.
142    
143    16. Unicode character properties were updated from Unicode 6.3.0.
144    
145    17. The compile-time code for auto-possessification has been refactored, based
146        on a patch by Zoltan Herczeg. It now happens after instead of during
147        compilation. The code is cleaner, and more cases are handled. The option
148        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
149        options in pcretest are provided to set it. It can also be set by
150        (*NO_AUTO_POSSESS) at the start of a pattern.
151    
152    18. The character VT has been added to the default ("C" locale) set of
153        characters that match \s and are generally treated as white space,
154        following this same change in Perl 5.18. There is now no difference between
155        "Perl space" and "POSIX space". Whether VT is treated as white space in
156        other locales depends on the locale.
157    
158    19. The code for checking named groups as conditions, either for being set or
159        for being recursed, has been refactored (this is related to 14 and 15
160        above). Processing unduplicated named groups should now be as fast at
161        numerical groups, and processing duplicated groups should be faster than
162        before.
163    
164    20. Two patches to the CMake build system, by Alexander Barkov:
165    
166          (1) Replace the "source" command by "." in CMakeLists.txt because
167              "source" is a bash-ism.
168    
169          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
170              without these the CMake build does not work on Solaris.
171    
172    21. Perl has changed its handling of \8 and \9. If there is no previously
173        encountered capturing group of those numbers, they are treated as the
174        literal characters 8 and 9 instead of a binary zero followed by the
175        literals. PCRE now does the same.
176    
177    22. Following Perl, added \o{} to specify codepoints in octal, making it
178        possible to specify values greater than 0777 and also making them
179        unambiguous.
180    
181    23. Perl now gives an error for missing closing braces after \x{... instead of
182        treating the string as literal. PCRE now does the same.
183    
184    24. RunTest used to grumble if an inappropriate test was selected explicitly,
185        but just skip it when running all tests. This make it awkward to run ranges
186        of tests when one of them was inappropriate. Now it just skips any
187        inappropriate tests, as it always did when running all tests.
188    
189    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
190        character types such as \d or \w, too many callouts were inserted, and the
191        data that they returned was rubbish.
192    
193    26. In UCP mode, \s was not matching two of the characters that Perl matches,
194        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
195        were matched by \h. The code has now been refactored so that the lists of
196        the horizontal and vertical whitespace characters used for \h and \v (which
197        are defined only in one place) are now also used for \s.
198    
199    27. Add JIT support for the 64 bit TileGX architecture.
200        Patch by Jiong Wang (Tilera Corporation).
201    
202    28. Possessive quantifiers for classes (both explicit and automatically
203        generated) now use special opcodes instead of wrapping in ONCE brackets.
204    
205    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
206        (because it's meaningless), this was not happening when PCRE_CASELESS was
207        set. Not wrong, but inefficient.
208    
209    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
210        properties for \w, \d, etc) is present in a test regex. Otherwise if the
211        test contains no characters greater than 255, Perl doesn't realise it
212        should be using Unicode semantics.
213    
214    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
215        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
216        does in Unicode mode.
217    
218    32. Added the "forbid" facility to pcretest so that putting tests into the
219        wrong test files can sometimes be quickly detected.
220    
221    33. There is now a limit (default 250) on the depth of nesting of parentheses.
222        This limit is imposed to control the amount of system stack used at compile
223        time. It can be changed at build time by --with-parens-nest-limit=xxx or
224        the equivalent in CMake.
225    
226    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
227        errors. Perl warns for these when in warning mode, but PCRE has no facility
228        for giving warnings.
229    
230    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
231        However, this was not working for (?!) because it is optimized to (*FAIL),
232        for which PCRE does not allow quantifiers. The optimization is now disabled
233        when a quantifier follows (?!). I can't see any use for this, but it makes
234        things uniform.
235    
236    36. Perl no longer allows group names to start with digits, so I have made this
237        change also in PCRE. It simplifies the code a bit.
238    
239    37. In extended mode, Perl ignores spaces before a + that indicates a
240        possessive quantifier. PCRE allowed a space before the quantifier, but not
241        before the possessive +. It now does.
242    
243    38. The use of \K (reset reported match start) within a repeated possessive
244        group such as (a\Kb)*+ was not working.
245    
246    40. Document that the same character tables must be used at compile time and
247        run time, and that the facility to pass tables to pcre_exec() and
248        pcre_dfa_exec() is for use only with saved/restored patterns.
249    
250    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
251        features for Builds with MSVC:
252    
253        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
254           builds). The .rc files can be used to set FileDescription and many other
255           attributes.
256    
257        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
258           This allows higher-level build scripts which want .pdb files to avoid
259           hard-coding the exact files needed."
260    
261    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
262        mean "start of word" and "end of word", respectively, as a transition aid.
263    
264    43. A minimizing repeat of a class containing codepoints greater than 255 in
265        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
266        compiled to use the heap for recursion.
267    
268    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
269        is configured.
270    
271    
272    Version 8.33 28-May-2013
273    ------------------------
274    
275    1.  Added 'U' to some constants that are compared to unsigned integers, to
276        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
277        variables that are added to signed variables, to ensure the result is
278        signed and can be negated.
279    
280    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
281        CMake config files.
282    
283    3.  Revise the creation of config.h.generic so that all boolean macros are
284        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
285        overriding via -D on the command line possible.
286    
287    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
288        to unsigned int is reported to make a quite noticeable speed difference in
289        a specific Windows environment. Testing on Linux did also appear to show
290        some benefit (and it is clearly not harmful). Also fixed the definition of
291        Xop which should be unsigned.
292    
293    5.  Related to (4), changing the definition of the intermediate variable cc
294        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
295        improvements.
296    
297    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
298        unnecessary spaces.
299    
300    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
301        and later.
302    
303    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
304    
305    9.  Optimizing fast_forward_start_bits in JIT.
306    
307    10. Adding support for callouts in JIT, and fixing some issues revealed
308        during this work. Namely:
309    
310        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
311    
312        (b) Minimum length was not checked before the matching is started.
313    
314    11. The value of capture_last that is passed to callouts was incorrect in some
315        cases when there was a capture on one path that was subsequently abandoned
316        after a backtrack. Also, the capture_last value is now reset after a
317        recursion, since all captures are also reset in this case.
318    
319    12. The interpreter no longer returns the "too many substrings" error in the
320        case when an overflowing capture is in a branch that is subsequently
321        abandoned after a backtrack.
322    
323    13. In the pathological case when an offset vector of size 2 is used, pcretest
324        now prints out the matched string after a yield of 0 or 1.
325    
326    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
327        Only supported by the JIT compiler at the moment.
328    
329    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
330    
331    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
332        offset of the starting point of the matching process, provided the offsets
333        vector is large enough.
334    
335    17. The \A escape now records a lookbehind value of 1, though its execution
336        does not actually inspect the previous character. This is to ensure that,
337        in partial multi-segment matching, at least one character from the old
338        segment is retained when a new segment is processed. Otherwise, if there
339        are no lookbehinds in the pattern, \A might match incorrectly at the start
340        of a new segment.
341    
342    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
343    
344    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
345        gives some modest performance improvement in 8-bit mode.
346    
347    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
348        be expressed in certain programming languages using Universal Character
349        Names.
350    
351    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
352        which points out that "non characters" are not "characters that may not
353        appear in Unicode strings" but rather "characters that are reserved for
354        internal use and have only local meaning".
355    
356    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
357        there was a conditional group that depended on an assertion, if the
358        assertion was false, the callout that immediately followed the alternation
359        in the condition was skipped when pcre_exec() was used for matching.
360    
361    23. Allow an explicit callout to be inserted before an assertion that is the
362        condition for a conditional group, for compatibility with automatic
363        callouts, which always insert a callout at this point.
364    
365    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
366        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
367    
368    25. (*PRUNE) is now supported by the JIT compiler.
369    
370    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
371    
372    27. Fix the case where there are two or more SKIPs with arguments that may be
373        ignored.
374    
375    28. (*SKIP) is now supported by the JIT compiler.
376    
377    29. (*THEN) is now supported by the JIT compiler.
378    
379    30. Update RunTest with additional test selector options.
380    
381    31. The way PCRE handles backtracking verbs has been changed in two ways.
382    
383        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
384        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
385        backtracking. In some cases this makes it more Perl-compatible, but Perl's
386        rather obscure rules do not always do the same thing.
387    
388        (2) Previously, backtracking verbs were confined within assertions. This is
389        no longer the case for positive assertions, except for (*ACCEPT). Again,
390        this sometimes improves Perl compatibility, and sometimes does not.
391    
392    32. A number of tests that were in test 2 because Perl did things differently
393        have been moved to test 1, because either Perl or PCRE has changed, and
394        these tests are now compatible.
395    
396    32. Backtracking control verbs are now handled in the same way in JIT and
397        interpreter.
398    
399    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
400        contained a forward subroutine reference caused a compile error.
401    
402    34. Auto-detect and optimize limited repetitions in JIT.
403    
404    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
405        blocking (*UTF) etc.
406    
407    36. In the interpreter, maximizing pattern repetitions for characters and
408        character types now use tail recursion, which reduces stack usage.
409    
410    37. The value of the max lookbehind was not correctly preserved if a compiled
411        and saved regex was reloaded on a host of different endianness.
412    
413    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
414        of the compiled pattern block, expand the flags field from 16 to 32 bits
415        because it was almost full.
416    
417    39. Try madvise first before posix_madvise.
418    
419    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
420        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
421        original change locked out all matches of empty strings. This has been
422        changed so that one match of an empty string per line is recognized.
423        Subsequent searches on the same line (for colouring or for --only-matching,
424        for example) do not recognize empty strings.
425    
426    41. Applied a user patch to fix a number of spelling mistakes in comments.
427    
428    42. Data lines longer than 65536 caused pcretest to crash.
429    
430    43. Clarified the data type for length and startoffset arguments for pcre_exec
431        and pcre_dfa_exec in the function-specific man pages, where they were
432        explicitly stated to be in bytes, never having been updated. I also added
433        some clarification to the pcreapi man page.
434    
435    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
436        a segmentation fault.
437    
438    
439    Version 8.32 30-November-2012
440    -----------------------------
441    
442    1.  Improved JIT compiler optimizations for first character search and single
443        character iterators.
444    
445    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
446        Patch by Daniel Richard G.
447    
448    3.  Single character iterator optimizations in the JIT compiler.
449    
450    4.  Improved JIT compiler optimizations for character ranges.
451    
452    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
453        Reported by Giuseppe D'Angelo.
454    
455    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
456        of a line, was being set incorrectly in cases where .* appeared inside
457        atomic brackets at the start of a pattern, or where there was a subsequent
458        *PRUNE or *SKIP.
459    
460    7.  Improved instruction cache flush for POWER/PowerPC.
461        Patch by Daniel Richard G.
462    
463    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
464        grep:
465    
466        (a) There is now no limit to the number of patterns to be matched.
467    
468        (b) An error is given if a pattern is too long.
469    
470        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
471            are now supported.
472    
473        (d) --exclude-from and --include-from (multiple use) have been added.
474    
475        (e) Exclusions and inclusions now apply to all files and directories, not
476            just to those obtained from scanning a directory recursively.
477    
478        (f) Multiple uses of -f and --file-list are now supported.
479    
480        (g) In a Windows environment, the default for -d has been changed from
481            "read" (the GNU grep default) to "skip", because otherwise the presence
482            of a directory in the file list provokes an error.
483    
484        (h) The documentation has been revised and clarified in places.
485    
486    9.  Improve the matching speed of capturing brackets.
487    
488    10. Changed the meaning of \X so that it now matches a Unicode extended
489        grapheme cluster.
490    
491    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
492        out POSIX threads when JIT support is configured.
493    
494    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
495    
496    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
497        structure could go wrong in environments where size_t is not the same size
498        as int.
499    
500    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
501        set.
502    
503    15. The EBCDIC support had decayed; later updates to the code had included
504        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
505        general tidy up of EBCDIC-related issues, and the documentation was also
506        not quite right. There is now a test that can be run on ASCII systems to
507        check some of the EBCDIC-related things (but is it not a full test).
508    
509    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
510        in a small tidy to the code.
511    
512    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
513    
514    18. If the --only-matching (-o) option in pcregrep is specified multiple
515        times, each one causes appropriate output. For example, -o1 -o2 outputs the
516        substrings matched by the 1st and 2nd capturing parentheses. A separating
517        string can be specified by --om-separator (default empty).
518    
519    19. Improving the first n character searches.
520    
521    20. Turn case lists for horizontal and vertical white space into macros so that
522        they are defined only once.
523    
524    21. This set of changes together give more compatible Unicode case-folding
525        behaviour for characters that have more than one other case when UCP
526        support is available.
527    
528        (a) The Unicode property table now has offsets into a new table of sets of
529            three or more characters that are case-equivalent. The MultiStage2.py
530            script that generates these tables (the pcre_ucd.c file) now scans
531            CaseFolding.txt instead of UnicodeData.txt for character case
532            information.
533    
534        (b) The code for adding characters or ranges of characters to a character
535            class has been abstracted into a generalized function that also handles
536            case-independence. In UTF-mode with UCP support, this uses the new data
537            to handle characters with more than one other case.
538    
539        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
540            whose other case is greater than 256 are now correctly matched
541            caselessly. Previously, the high codepoint matched the low one, but not
542            vice versa.
543    
544        (d) The processing of \h, \H, \v, and \ in character classes now makes use
545            of the new class addition function, using character lists defined as
546            macros alongside the case definitions of 20 above.
547    
548        (e) Caseless back references now work with characters that have more than
549            one other case.
550    
551        (f) General caseless matching of characters with more than one other case
552            is supported.
553    
554    22. Unicode character properties were updated from Unicode 6.2.0
555    
556    23. Improved CMake support under Windows. Patch by Daniel Richard G.
557    
558    24. Add support for 32-bit character strings, and UTF-32
559    
560    25. Major JIT compiler update (code refactoring and bugfixing).
561        Experimental Sparc 32 support is added.
562    
563    26. Applied a modified version of Daniel Richard G's patch to create
564        pcre.h.generic and config.h.generic by "make" instead of in the
565        PrepareRelease script.
566    
567    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
568        pcre_compile.c when checking for a zero character.
569    
570    28. Introducing a native interface for JIT. Through this interface, the compiled
571        machine code can be directly executed. The purpose of this interface is to
572        provide fast pattern matching, so several sanity checks are not performed.
573        However, feature tests are still performed. The new interface provides
574        1.4x speedup compared to the old one.
575    
576    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
577        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
578        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
579    
580    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
581        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
582        the "old" RFC 2279). Instead, it ended up passing a negative length to
583        pcre_exec().
584    
585    31. Add support for GCC's visibility feature to hide internal functions.
586    
587    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
588        "unknown -C option" after outputting 0 or 1.
589    
590    33. There is now support for generating a code coverage report for the test
591        suite in environments where gcc is the compiler and lcov is installed. This
592        is mainly for the benefit of the developers.
593    
594    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
595        unaddressable using valgrind annotations, allowing valgrind to detect
596        invalid memory accesses. This is mainly for the benefit of the developers.
597    
598    25. (*UTF) can now be used to start a pattern in any of the three libraries.
599    
600    26. Give configure error if --enable-cpp but no C++ compiler found.
601    
602    
603    Version 8.31 06-July-2012
604    -------------------------
605    
606    1.  Fixing a wrong JIT test case and some compiler warnings.
607    
608    2.  Removed a bashism from the RunTest script.
609    
610    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
611        to unsigned type, result still unsigned" that was given by an MS compiler
612        on encountering the code "-sizeof(xxx)".
613    
614    4.  Partial matching support is added to the JIT compiler.
615    
616    5.  Fixed several bugs concerned with partial matching of items that consist
617        of more than one character:
618    
619        (a) /^(..)\1/ did not partially match "aba" because checking references was
620            done on an "all or nothing" basis. This also applied to repeated
621            references.
622    
623        (b) \R did not give a hard partial match if \r was found at the end of the
624            subject.
625    
626        (c) \X did not give a hard partial match after matching one or more
627            characters at the end of the subject.
628    
629        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
630            a partial match for the string "\r".
631    
632        (e) When newline was set to CRLF, the metacharacter "." did not recognize
633            a partial match for a CR character at the end of the subject string.
634    
635    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
636        running pcretest, the text "(JIT)" added to the output whenever JIT is
637        actually used to run the match.
638    
639    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
640        or /S+[+] with a digit between 1 and 7.
641    
642    8.  OP_NOT now supports any UTF character not just single-byte ones.
643    
644    9.  (*MARK) control verb is now supported by the JIT compiler.
645    
646    10. The command "./RunTest list" lists the available tests without actually
647        running any of them. (Because I keep forgetting what they all are.)
648    
649    11. Add PCRE_INFO_MAXLOOKBEHIND.
650    
651    12. Applied a (slightly modified) user-supplied patch that improves performance
652        when the heap is used for recursion (compiled with --disable-stack-for-
653        recursion). Instead of malloc and free for each heap frame each time a
654        logical recursion happens, frames are retained on a chain and re-used where
655        possible. This sometimes gives as much as 30% improvement.
656    
657    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
658        call.
659    
660    14. As documented, (*COMMIT) is now confined to within a positive assertion.
661    
662    15. It is now possible to link pcretest with libedit as an alternative to
663        libreadline.
664    
665    16. (*COMMIT) control verb is now supported by the JIT compiler.
666    
667    17. The Unicode data tables have been updated to Unicode 6.1.0.
668    
669    18. Added --file-list option to pcregrep.
670    
671    19. Added binary file support to pcregrep, including the -a, --binary-files,
672        -I, and --text options.
673    
674    20. The madvise function is renamed for posix_madvise for QNX compatibility
675        reasons. Fixed by Giuseppe D'Angelo.
676    
677    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
678    
679    22. Changed the option for creating symbolic links for 16-bit man pages from
680        -s to -sf so that re-installing does not cause issues.
681    
682    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
683    
684    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
685        matches in certain environments (the workspace was not being correctly
686        retained). Also added to pcre_dfa_exec() a simple plausibility check on
687        some of the workspace data at the beginning of a restart.
688    
689    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
690        was not doing so when it should - probably a typo introduced by SVN 528
691        (change 8.10/14).
692    
693    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
694        \w+ when the character tables indicated that \x{c4} was a word character.
695        There were several related cases, all because the tests for doing a table
696        lookup were testing for characters less than 127 instead of 255.
697    
698    27. If a pattern contains capturing parentheses that are not used in a match,
699        their slots in the ovector are set to -1. For those that are higher than
700        any matched groups, this happens at the end of processing. In the case when
701        there were back references that the ovector was too small to contain
702        (causing temporary malloc'd memory to be used during matching), and the
703        highest capturing number was not used, memory off the end of the ovector
704        was incorrectly being set to -1. (It was using the size of the temporary
705        memory instead of the true size.)
706    
707    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
708        ovector size, it uses memory at the end of the block that it has got.
709    
710    29. Check for an overlong MARK name and give an error at compile time. The
711        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
712    
713    30. JIT compiler update.
714    
715    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
716        Rill for the patch.
717    
718    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
719    
720    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
721    
722    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
723        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
724        was enabled.
725    
726    35. Improve JIT code generation for greedy plus quantifier.
727    
728    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
729        1 to "aa" instead of to an empty string. The bug affected repeated groups
730        that could potentially match an empty string.
731    
732    37. Optimizing single character iterators in JIT.
733    
734    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
735        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
736        codepoints that are too big for the mode are faulted, and in a UTF mode,
737        disallowed codepoints are also faulted.
738    
739    39. If PCRE was compiled with UTF support, in three places in the DFA
740        matcher there was code that should only have been obeyed in UTF mode, but
741        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
742        processing when bytes with values greater than 127 were present. In 16-bit
743        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
744        both cases the values are those that cannot be the first data item in a UTF
745        character. The three items that might have provoked this were recursions,
746        possessively repeated groups, and atomic groups.
747    
748    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
749        and pcregrep, because some OS require shared objects to be explicitly
750        passed to ld, causing the link step to fail if they are not.
751    
752    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
753        mode, patterns that started with \h* or \R* might be incorrectly matched.
754    
755    
756    Version 8.30 04-February-2012
757    -----------------------------
758    
759    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
760        name is defined in ctype.h.
761    
762    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
763        only in quite long subpatterns.
764    
765    3.  Removed the function pcre_info(), which has been obsolete and deprecated
766        since it was replaced by pcre_fullinfo() in February 2000.
767    
768    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
769        match a (*MARK), and the match failed at the start of the subject, a
770        reference to memory before the start of the subject could occur. This bug
771        was introduced by fix 17 of release 8.21.
772    
773    5.  A reference to an unset group with zero minimum repetition was giving
774        totally wrong answers (in non-JavaScript-compatibility mode). For example,
775        /(another)?(\1?)test/ matched against "hello world test". This bug was
776        introduced in release 8.13.
777    
778    6.  Add support for 16-bit character strings (a large amount of work involving
779        many changes and refactorings).
780    
781    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
782        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
783        from a file.
784    
785    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
786        rounding is not applied in this particular case).
787    
788    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
789        if they appear, or are escaped, in patterns.
790    
791    10. Get rid of a number of -Wunused-but-set-variable warnings.
792    
793    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
794        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
795        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
796        also returns the mark "x". This bug applied to capturing parentheses,
797        non-capturing parentheses, and atomic parentheses. It also applied to some
798        assertions.
799    
800    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
801        information out of configure.ac instead of relying on pcre.h.generic, which
802        is not stored in the repository.
803    
804    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
805        -lreadline.
806    
807    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
808    
809    15. Applied Graycode's patch to put the top-level frame on the stack rather
810        than the heap when not using the stack for recursion. This gives a
811        performance improvement in many cases when recursion is not deep.
812    
813    16. Experimental code added to "pcretest -C" to output the stack frame size.
814    
815    
816    Version 8.21 12-Dec-2011
817    ------------------------
818    
819    1.  Updating the JIT compiler.
820    
821    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
822        are added as well.
823    
824    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
825        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
826        calling _pcre_jit_exec. Some extra comments are added.
827    
828    4.  (*MARK) settings inside atomic groups that do not contain any capturing
829        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
830        was introduced by change 18 for 8.20.
831    
832    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
833        ECMA-262 standard.
834    
835    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
836        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
837        This bug was probably introduced by change 9 of 8.13.
838    
839    7.  While fixing 6 above, I noticed that a number of other items were being
840        incorrectly rejected as "not fixed length". This arose partly because newer
841        opcodes had not been added to the fixed-length checking code. I have (a)
842        corrected the bug and added tests for these items, and (b) arranged for an
843        error to occur if an unknown opcode is encountered while checking for fixed
844        length instead of just assuming "not fixed length". The items that were
845        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
846        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
847        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
848    
849    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
850        being incorrectly compiled and would have given unpredicatble results.
851    
852    9.  A possessively repeated subpattern with minimum repeat count greater than
853        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
854        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
855        the first (A) could occur when it should not.
856    
857    10. Add a cast and remove a redundant test from the code.
858    
859    11. JIT should use pcre_malloc/pcre_free for allocation.
860    
861    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
862        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
863        is anything other than /usr, -L is still shown).
864    
865    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
866    
867    14. Perl does not support \N without a following name in a [] class; PCRE now
868        also gives an error.
869    
870    15. If a forward reference was repeated with an upper limit of around 2000,
871        it caused the error "internal error: overran compiling workspace". The
872        maximum number of forward references (including repeats) was limited by the
873        internal workspace, and dependent on the LINK_SIZE. The code has been
874        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
875        the default depends on LINK_SIZE. There is a new upper limit (for safety)
876        of around 200,000 forward references. While doing this, I also speeded up
877        the filling in of repeated forward references.
878    
879    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
880        incorrectly expecting the subject to contain another "a" after the start.
881    
882    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
883        in the match, the SKIP should be ignored. This was not happening; instead
884        the SKIP was being treated as NOMATCH. For patterns such as
885        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
886        tested.
887    
888    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
889        now much more compatible with Perl, in particular in cases where the result
890        is a non-match for a non-anchored pattern. For example, if
891        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
892        "m", where previously it did not return a name. A side effect of this
893        change is that for partial matches, the last encountered mark name is
894        returned, as for non matches. A number of tests that were previously not
895        Perl-compatible have been moved into the Perl-compatible test files. The
896        refactoring has had the pleasing side effect of removing one argument from
897        the match() function, thus reducing its stack requirements.
898    
899    19. If the /S+ option was used in pcretest to study a pattern using JIT,
900        subsequent uses of /S (without +) incorrectly behaved like /S+.
901    
902    21. Retrieve executable code size support for the JIT compiler and fixing
903        some warnings.
904    
905    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
906        not work when the shorter character appeared right at the end of the
907        subject string.
908    
909    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
910        systems.
911    
912    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
913        output it when the /M option is used in pcretest.
914    
915    25. The CheckMan script was not being included in the distribution. Also, added
916        an explicit "perl" to run Perl scripts from the PrepareRelease script
917        because this is reportedly needed in Windows.
918    
919    26. If study data was being save in a file and studying had not found a set of
920        "starts with" bytes for the pattern, the data written to the file (though
921        never used) was taken from uninitialized memory and so caused valgrind to
922        complain.
923    
924    27. Updated RunTest.bat as provided by Sheri Pierce.
925    
926    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
927    
928    29. Computation of memory usage for the table of capturing group names was
929        giving an unnecessarily large value.
930    
931    
932    Version 8.20 21-Oct-2011
933    ------------------------
934    
935    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
936        a POSIX class. After further experiments with Perl, which convinced me that
937        Perl has bugs and confusions, a closing square bracket is no longer allowed
938        in a POSIX name. This bug also affected patterns with classes that started
939        with full stops.
940    
941    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
942        captured substring, but while checking the failing first alternative,
943        substring 1 is temporarily captured. If the output vector supplied to
944        pcre_exec() was not big enough for this capture, the yield of the function
945        was still zero ("insufficient space for captured substrings"). This cannot
946        be totally fixed without adding another stack variable, which seems a lot
947        of expense for a edge case. However, I have improved the situation in cases
948        such as /(a)(b)x|abc/ matched against "abc", where the return code
949        indicates that fewer than the maximum number of slots in the ovector have
950        been set.
951    
952    3.  Related to (2) above: when there are more back references in a pattern than
953        slots in the output vector, pcre_exec() uses temporary memory during
954        matching, and copies in the captures as far as possible afterwards. It was
955        using the entire output vector, but this conflicts with the specification
956        that only 2/3 is used for passing back captured substrings. Now it uses
957        only the first 2/3, for compatibility. This is, of course, another edge
958        case.
959    
960    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
961        main code base, and can be used by building with --enable-jit. When this is
962        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
963        runtime --no-jit option is given.
964    
965    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
966        ovector, the return from the function was zero, implying that there were
967        other matches that did not fit. The correct "exactly full" value is now
968        returned.
969    
970    6.  If a subpattern that was called recursively or as a subroutine contained
971        (*PRUNE) or any other control that caused it to give a non-standard return,
972        invalid errors such as "Error -26 (nested recursion at the same subject
973        position)" or even infinite loops could occur.
974    
975    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
976        computing the minimum length on reaching *ACCEPT, and so ended up with the
977        wrong value of 1 rather than 0. Further investigation indicates that
978        computing a minimum subject length in the presence of *ACCEPT is difficult
979        (think back references, subroutine calls), and so I have changed the code
980        so that no minimum is registered for a pattern that contains *ACCEPT.
981    
982    8.  If (*THEN) was present in the first (true) branch of a conditional group,
983        it was not handled as intended. [But see 16 below.]
984    
985    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
986        Sheri Pierce.
987    
988    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
989        the first byte in a match must be "a".
990    
991    11. Change 17 for 8.13 increased the recursion depth for patterns like
992        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
993        pattern contains any instances of (*THEN). If it does not, the old
994        optimizations are restored. It would be nice to do this on a per-group
995        basis, but at the moment that is not feasible.
996    
997    12. In some environments, the output of pcretest -C is CRLF terminated. This
998        broke RunTest's code that checks for the link size. A single white space
999        character after the value is now allowed for.
1000    
1001    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1002        For "fr", it uses the Windows-specific input and output files.
1003    
1004    14. If (*THEN) appeared in a group that was called recursively or as a
1005        subroutine, it did not work as intended. [But see next item.]
1006    
1007    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1008        pattern fragments (but not containing any | characters). If A and B are
1009        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1010        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1011        D. In other words, Perl considers parentheses that do not contain any |
1012        characters to be part of a surrounding alternative, whereas PCRE was
1013        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1014        differently. PCRE now behaves in the same way as Perl, except in the case
1015        of subroutine/recursion calls such as (?1) which have in any case always
1016        been different (but PCRE had them first :-).
1017    
1018    16. Related to 15 above: Perl does not treat the | in a conditional group as
1019        creating alternatives. Such a group is treated in the same way as an
1020        ordinary group without any | characters when processing (*THEN). PCRE has
1021        been changed to match Perl's behaviour.
1022    
1023    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1024        RunGrepTest script failed.
1025    
1026    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1027        inevitable for groups that contain captures, but it can lead to a lot of
1028        stack use in large patterns. The old behaviour has been restored for atomic
1029        groups that do not contain any capturing parentheses.
1030    
1031    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1032        suppress the check for a minimum subject length at run time. (If it was
1033        given to pcre_exec() or pcre_dfa_exec() it did work.)
1034    
1035    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1036        fail to work when decoding hex characters in data strings in EBCDIC
1037        environments.
1038    
1039    21. It appears that in at least one Mac OS environment, the isxdigit() function
1040        is implemented as a macro that evaluates to its argument more than once,
1041        contravening the C 90 Standard (I haven't checked a later standard). There
1042        was an instance in pcretest which caused it to go wrong when processing
1043        \x{...} escapes in subject strings. The has been rewritten to avoid using
1044        things like p++ in the argument of isxdigit().
1045    
1046    
1047    Version 8.13 16-Aug-2011
1048    ------------------------
1049    
1050    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1051    
1052    2.  Two minor typos in pcre_internal.h have been fixed.
1053    
1054    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1055        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1056        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1057    
1058    4.  There were a number of related bugs in the code for matching backrefences
1059        caselessly in UTF-8 mode when codes for the characters concerned were
1060        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1061        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1062        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1063        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1064        2-byte code at the end of the subject (it thought there wasn't enough data
1065        left).
1066    
1067    5.  Comprehensive information about what went wrong is now returned by
1068        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1069        as the output vector has at least 2 elements. The offset of the start of
1070        the failing character and a reason code are placed in the vector.
1071    
1072    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1073        now returned is for the first byte of the failing character, instead of the
1074        last byte inspected. This is an incompatible change, but I hope it is small
1075        enough not to be a problem. It makes the returned offset consistent with
1076        pcre_exec() and pcre_dfa_exec().
1077    
1078    7.  pcretest now gives a text phrase as well as the error number when
1079        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1080        failure, the offset and reason code are output.
1081    
1082    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1083        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1084        back over a single character (\n). This seems wrong (because it treated the
1085        two characters as a single entity when going forwards), conflicts with the
1086        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1087        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1088        has been changed.
1089    
1090    9.  Some internal refactoring has changed the processing so that the handling
1091        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1092        time (the PCRE_DOTALL option was changed this way some time ago: version
1093        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1094        which was always a bit of a fudge. It also means that there is one less
1095        argument for the match() function, which reduces its stack requirements
1096        slightly. This change also fixes an incompatibility with Perl: the pattern
1097        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1098    
1099    10. More internal refactoring has drastically reduced the number of recursive
1100        calls to match() for possessively repeated groups such as (abc)++ when
1101        using pcre_exec().
1102    
1103    11. While implementing 10, a number of bugs in the handling of groups were
1104        discovered and fixed:
1105    
1106        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1107        (a|)*(?1) gave a compile-time internal error.
1108        ((a|)+)+  did not notice that the outer group could match an empty string.
1109        (^a|^)+   was not marked as anchored.
1110        (.*a|.*)+ was not marked as matching at start or after a newline.
1111    
1112    12. Yet more internal refactoring has removed another argument from the match()
1113        function. Special calls to this function are now indicated by setting a
1114        value in a variable in the "match data" data block.
1115    
1116    13. Be more explicit in pcre_study() instead of relying on "default" for
1117        opcodes that mean there is no starting character; this means that when new
1118        ones are added and accidentally left out of pcre_study(), testing should
1119        pick them up.
1120    
1121    14. The -s option of pcretest has been documented for ages as being an old
1122        synonym of -m (show memory usage). I have changed it to mean "force study
1123        for every regex", that is, assume /S for every regex. This is similar to -i
1124        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1125        using it. It makes it easier to run collections of tests with and without
1126        study enabled, and thereby test pcre_study() more easily. All the standard
1127        tests are now run with and without -s (but some patterns can be marked as
1128        "never study" - see 20 below).
1129    
1130    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1131        restoration of the capturing data to the outer values was not happening
1132        correctly.
1133    
1134    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1135        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1136        pattern had matched an empty string, and so incorrectly returned a no
1137        match.
1138    
1139    17. There was optimizing code for the last branch of non-capturing parentheses,
1140        and also for the obeyed branch of a conditional subexpression, which used
1141        tail recursion to cut down on stack usage. Unfortunately, now that there is
1142        the possibility of (*THEN) occurring in these branches, tail recursion is
1143        no longer possible because the return has to be checked for (*THEN). These
1144        two optimizations have therefore been removed. [But see 8.20/11 above.]
1145    
1146    18. If a pattern containing \R was studied, it was assumed that \R always
1147        matched two bytes, thus causing the minimum subject length to be
1148        incorrectly computed because \R can also match just one byte.
1149    
1150    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1151        was incorrectly computed.
1152    
1153    20. If /S is present twice on a test pattern in pcretest input, it now
1154        *disables* studying, thereby overriding the use of -s on the command line
1155        (see 14 above). This is necessary for one or two tests to keep the output
1156        identical in both cases.
1157    
1158    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1159        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1160    
1161    22. When an atomic group that contained a capturing parenthesis was
1162        successfully matched, but the branch in which it appeared failed, the
1163        capturing was not being forgotten if a higher numbered group was later
1164        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1165        group 1 to "a", when in fact it should be unset. This applied to multi-
1166        branched capturing and non-capturing groups, repeated or not, and also to
1167        positive assertions (capturing in negative assertions does not happen
1168        in PCRE) and also to nested atomic groups.
1169    
1170    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1171        subject after a captured substring, to make it easier to tell which of a
1172        number of identical substrings has been captured.
1173    
1174    24. The way atomic groups are processed by pcre_exec() has been changed so that
1175        if they are repeated, backtracking one repetition now resets captured
1176        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1177        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1178        "aaa". Previously, it would have been "a". As part of this code
1179        refactoring, the way recursive calls are handled has also been changed.
1180    
1181    25. If an assertion condition captured any substrings, they were not passed
1182        back unless some other capturing happened later. For example, if
1183        (?(?=(a))a) was matched against "a", no capturing was returned.
1184    
1185    26. When studying a pattern that contained subroutine calls or assertions,
1186        the code for finding the minimum length of a possible match was handling
1187        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1188        group 1 called group 2 while simultaneously a separate group 2 called group
1189        1). A stack overflow occurred in this case. I have fixed this by limiting
1190        the recursion depth to 10.
1191    
1192    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1193        Fortmann. This supports explicit test numbers on the command line, and has
1194        argument validation and error reporting.
1195    
1196    28. An instance of \X with an unlimited repeat could fail if at any point the
1197        first character it looked at was a mark character.
1198    
1199    29. Some minor code refactoring concerning Unicode properties and scripts
1200        should reduce the stack requirement of match() slightly.
1201    
1202    30. Added the '=' option to pcretest to check the setting of unused capturing
1203        slots at the end of the pattern, which are documented as being -1, but are
1204        not included in the return count.
1205    
1206    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1207        compiled something random. Now it gives a compile-time error (as does
1208        Perl).
1209    
1210    32. A *MARK encountered during the processing of a positive assertion is now
1211        recorded and passed back (compatible with Perl).
1212    
1213    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1214        had alternative anchored branches, the search for a second match in a line
1215        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1216        matched the line "0102" twice. The same bug affected patterns that started
1217        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1218        twice.
1219    
1220    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1221        does, and because of capturing effects, quantifying parenthesized
1222        assertions may at times be useful. Quantifiers are now allowed for
1223        parenthesized assertions.
1224    
1225    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1226    
1227    36. \g was being checked for fancy things in a character class, when it should
1228        just be a literal "g".
1229    
1230    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1231        appearance of a nested POSIX class supersedes an apparent external class.
1232        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1233        unescaped square brackets may also appear as part of class names. For
1234        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1235        more like Perl. (But see 8.20/1 above.)
1236    
1237    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1238        was because it thought it was \N{name}, which is not supported).
1239    
1240    39. Add minix to OS list not supporting the -S option in pcretest.
1241    
1242    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1243        cannot analyze patterns in sufficient detail to catch mutual recursions
1244        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1245        subgroup is called recursively as a subpattern for a second time at the
1246        same position in the subject string. In previous releases this might have
1247        been caught by the recursion limit, or it might have run out of stack.
1248    
1249    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1250        happen only once. PCRE was, however incorrectly giving a compile time error
1251        "recursive call could loop indefinitely" because it cannot analyze the
1252        pattern in sufficient detail. The compile time test no longer happens when
1253        PCRE is compiling a conditional subpattern, but actual runaway loops are
1254        now caught at runtime (see 40 above).
1255    
1256    42. It seems that Perl allows any characters other than a closing parenthesis
1257        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1258        has been changed to be the same.
1259    
1260    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1261        as not to get warnings when autogen.sh is called. Also changed
1262        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1263    
1264    44. To help people who use pcregrep to scan files containing exceedingly long
1265        lines, the following changes have been made:
1266    
1267        (a) The default value of the buffer size parameter has been increased from
1268            8K to 20K. (The actual buffer used is three times this size.)
1269    
1270        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1271            PCRE is built.
1272    
1273        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1274            to be set at run time.
1275    
1276        (d) Numerical values in pcregrep options can be followed by K or M, for
1277            example --buffer-size=50K.
1278    
1279        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1280            given and the return code is set to 2.
1281    
1282    45. Add a pointer to the latest mark to the callout data block.
1283    
1284    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1285        partial match of an empty string instead of no match. This was specific to
1286        the use of ".".
1287    
1288    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1289        complete match instead of a partial match. This bug was dependent on both
1290        the PCRE_UTF8 and PCRE_DOTALL options being set.
1291    
1292    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1293        starting byte set, because \b was not being ignored.
1294    
1295    
1296    Version 8.12 15-Jan-2011
1297    ------------------------
1298    
1299    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1300        checks for such things as part of the documentation building process.
1301    
1302    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1303        --match-limit and --recursion-limit options (added for 8.11). In
1304        particular, this made one of the standard tests fail. (The integer value
1305        went into the wrong half of a long int.)
1306    
1307    3.  If the --colour option was given to pcregrep with -v (invert match), it
1308        did strange things, either producing crazy output, or crashing. It should,
1309        of course, ignore a request for colour when reporting lines that do not
1310        match.
1311    
1312    4.  Another pcregrep bug caused similar problems if --colour was specified with
1313        -M (multiline) and the pattern match finished with a line ending.
1314    
1315    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1316        matched in multiline mode, the following line was shown as part of the
1317        match. This seems wrong, so I have changed it.
1318    
1319    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1320        the check for further matches in the same line (so they could be coloured)
1321        to overrun the end of the current line. If another match was found, it was
1322        incorrectly shown (and then shown again when found in the next line).
1323    
1324    7.  If pcregrep was compiled under Windows, there was a reference to the
1325        function pcregrep_exit() before it was defined. I am assuming this was
1326        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1327        reported by a user. I've moved the definition above the reference.
1328    
1329    
1330    Version 8.11 10-Dec-2010
1331    ------------------------
1332    
1333    1.  (*THEN) was not working properly if there were untried alternatives prior
1334        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1335        backtracked to try for "b" instead of moving to the next alternative branch
1336        at the same level (in this case, to look for "c"). The Perl documentation
1337        is clear that when (*THEN) is backtracked onto, it goes to the "next
1338        alternative in the innermost enclosing group".
1339    
1340    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1341        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1342        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1343        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1344        (*THEN).
1345    
1346    3.  If \s appeared in a character class, it removed the VT character from
1347        the class, even if it had been included by some previous item, for example
1348        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1349        of \s, but is part of the POSIX "space" class.)
1350    
1351    4.  A partial match never returns an empty string (because you can always
1352        match an empty string at the end of the subject); however the checking for
1353        an empty string was starting at the "start of match" point. This has been
1354        changed to the "earliest inspected character" point, because the returned
1355        data for a partial match starts at this character. This means that, for
1356        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1357        (previously it gave "no match").
1358    
1359    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1360        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1361        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1362        has an implication that the given string is incomplete (because a partial
1363        match is preferred over a full match). For this reason, these items now
1364        give a partial match in this situation. [Aside: previously, the one case
1365        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1366        match rather than a full match, which was wrong by the old rules, but is
1367        now correct.]
1368    
1369    6.  There was a bug in the handling of #-introduced comments, recognized when
1370        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1371        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1372        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1373        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1374        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1375        places in pcre_compile().
1376    
1377    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1378        comments when looking ahead for named forward references to subpatterns,
1379        the only newline sequence it recognized was NL. It now handles newlines
1380        according to the set newline convention.
1381    
1382    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1383        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1384        cater for a lack of strerror(). These oversights have been fixed.
1385    
1386    9.  Added --match-limit and --recursion-limit to pcregrep.
1387    
1388    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1389    
1390    11. When the -o option was used, pcregrep was setting a return code of 1, even
1391        when matches were found, and --line-buffered was not being honoured.
1392    
1393    12. Added an optional parentheses number to the -o and --only-matching options
1394        of pcregrep.
1395    
1396    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1397        can match an empty string. The code to do it in pcretest and pcredemo
1398        needed fixing:
1399    
1400        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1401            only one byte after an empty string match just before CRLF (this case
1402            just got forgotten; "any" and "anycrlf" were OK).
1403    
1404        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1405            mode when an empty string match preceded an ASCII character followed by
1406            a non-ASCII character. (The code for advancing by one character rather
1407            than one byte was nonsense.)
1408    
1409        (c) The pcredemo.c sample program did not have any code at all to handle
1410            the cases when CRLF is a valid newline sequence.
1411    
1412    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1413        as a starting offset was within the subject string. There is now a new
1414        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1415        negative or greater than the length of the string. In order to test this,
1416        pcretest is extended to allow the setting of negative starting offsets.
1417    
1418    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1419        starting offset points to the beginning of a UTF-8 character was
1420        unnecessarily clumsy. I tidied it up.
1421    
1422    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1423        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1424    
1425    17. Nobody had reported that the --include_dir option, which was added in
1426        release 7.7 should have been called --include-dir (hyphen, not underscore)
1427        for compatibility with GNU grep. I have changed it to --include-dir, but
1428        left --include_dir as an undocumented synonym, and the same for
1429        --exclude-dir, though that is not available in GNU grep, at least as of
1430        release 2.5.4.
1431    
1432    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1433        characters from a string of bytes) have been redefined so as not to use
1434        loops, in order to improve performance in some environments. At the same
1435        time, I abstracted some of the common code into auxiliary macros to save
1436        repetition (this should not affect the compiled code).
1437    
1438    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1439        compile-time error is now given if \c is not followed by an ASCII
1440        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1441        different, and any byte value is allowed.)
1442    
1443    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1444        START_OPTIMIZE option, which is now allowed at compile time - but just
1445        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1446        to pcregrep and other applications that have no direct access to PCRE
1447        options. The new /Y option in pcretest sets this option when calling
1448        pcre_compile().
1449    
1450    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1451        back references. Groups containing recursive back references were forced to
1452        be atomic by that change, but in the case of named groups, the amount of
1453        memory required was incorrectly computed, leading to "Failed: internal
1454        error: code overflow". This has been fixed.
1455    
1456    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1457        pcretest.c, to avoid build problems in some Borland environments.
1458    
1459    
1460    Version 8.10 25-Jun-2010
1461    ------------------------
1462    
1463    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1464        THEN.
1465    
1466    2.  (*ACCEPT) was not working when inside an atomic group.
1467    
1468    3.  Inside a character class, \B is treated as a literal by default, but
1469        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1470        causes the error). The code is unchanged, but I tidied the documentation.
1471    
1472    4.  Inside a character class, PCRE always treated \R and \X as literals,
1473        whereas Perl faults them if its -w option is set. I have changed PCRE so
1474        that it faults them when PCRE_EXTRA is set.
1475    
1476    5.  Added support for \N, which always matches any character other than
1477        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1478    
1479    6.  When compiling pcregrep with newer versions of gcc which may have
1480        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1481        declared with attribute warn_unused_result" were given. Just casting the
1482        result to (void) does not stop the warnings; a more elaborate fudge is
1483        needed. I've used a macro to implement this.
1484    
1485    7.  Minor change to pcretest.c to avoid a compiler warning.
1486    
1487    8.  Added four artifical Unicode properties to help with an option to make
1488        \s etc use properties (see next item). The new properties are: Xan
1489        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1490    
1491    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1492        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1493        this option. Modified pcretest to add /W to test this facility. Added
1494        REG_UCP to make it available via the POSIX interface.
1495    
1496    10. Added --line-buffered to pcregrep.
1497    
1498    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1499        studied, and the match started with a letter with a code point greater than
1500        127 whose first byte was different to the first byte of the other case of
1501        the letter, the other case of this starting letter was not recognized
1502        (#976).
1503    
1504    12. If a pattern that was studied started with a repeated Unicode property
1505        test, for example, \p{Nd}+, there was the theoretical possibility of
1506        setting up an incorrect bitmap of starting bytes, but fortunately it could
1507        not have actually happened in practice until change 8 above was made (it
1508        added property types that matched character-matching opcodes).
1509    
1510    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1511        possible starting bytes for non-anchored patterns.
1512    
1513    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1514        \R, and also a number of cases that involve Unicode properties, both
1515        explicit and implicit when PCRE_UCP is set.
1516    
1517    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1518        input, it could crash or give wrong results if characters with values
1519        greater than 0xc0 were present in the subject string. (Detail: it assumed
1520        UTF-8 input when processing these items.)
1521    
1522    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1523        size_t is 64-bit (#991).
1524    
1525    17. Added a check for running out of memory when PCRE is compiled with
1526        --disable-stack-for-recursion (#990).
1527    
1528    18. If the last data line in a file for pcretest does not have a newline on
1529        the end, a newline was missing in the output.
1530    
1531    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1532        less than 128) in its various bitmaps. However, there is a facility for
1533        generating tables according to the current locale when PCRE is compiled. It
1534        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1535        characters, are recognized by isspace() and therefore were getting set in
1536        these tables, and indeed these tables seem to approximate to ISO 8859. This
1537        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1538        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1539        which of course cannot start UTF-8 characters. I have changed the code so
1540        that only real ASCII characters (less than 128) and the correct starting
1541        bytes for UTF-8 encodings are set for characters greater than 127 when in
1542        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1543        altogether.)
1544    
1545    20. Added the /T option to pcretest so as to be able to run tests with non-
1546        standard character tables, thus making it possible to include the tests
1547        used for 19 above in the standard set of tests.
1548    
1549    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1550        reference to a subpattern the other side of a comment that contains an
1551        opening parenthesis caused either an internal compiling error, or a
1552        reference to the wrong subpattern.
1553    
1554    
1555    Version 8.02 19-Mar-2010
1556    ------------------------
1557    
1558    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1559    
1560    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1561        configured.
1562    
1563    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1564        original author of that file, following a query about its status.
1565    
1566    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1567        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1568    
1569    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1570        quantifier applied to a forward-referencing subroutine call, could compile
1571        incorrect code or give the error "internal error: previously-checked
1572        referenced subpattern not found".
1573    
1574    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1575        variables to point to external functions. For these systems, therefore,
1576        pcre_malloc etc. are now initialized to local functions that call the
1577        relevant global functions.
1578    
1579    7.  There were two entries missing in the vectors called coptable and poptable
1580        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1581        I've fixed the data, and added a kludgy way of testing at compile time that
1582        the lengths are correct (equal to the number of opcodes).
1583    
1584    8.  Following on from 7, I added a similar kludge to check the length of the
1585        eint vector in pcreposix.c.
1586    
1587    9.  Error texts for pcre_compile() are held as one long string to avoid too
1588        much relocation at load time. To find a text, the string is searched,
1589        counting zeros. There was no check for running off the end of the string,
1590        which could happen if a new error number was added without updating the
1591        string.
1592    
1593    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1594    
1595    11. \K was not working if it appeared in an atomic group or in a group that
1596        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1597        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1598        the assertion is positive, but not if it is negative.
1599    
1600    12. Change 11 fortuitously reduced the size of the stack frame used in the
1601        "match()" function of pcre_exec.c by one pointer. Forthcoming
1602        implementation of support for (*MARK) will need an extra pointer on the
1603        stack; I have reserved it now, so that the stack frame size does not
1604        decrease.
1605    
1606    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1607        item in branch that calls a recursion is a subroutine call - as in the
1608        second branch in the above example - was incorrectly given the compile-
1609        time error "recursive call could loop indefinitely" because pcre_compile()
1610        was not correctly checking the subroutine for matching a non-empty string.
1611    
1612    14. The checks for overrunning compiling workspace could trigger after an
1613        overrun had occurred. This is a "should never occur" error, but it can be
1614        triggered by pathological patterns such as hundreds of nested parentheses.
1615        The checks now trigger 100 bytes before the end of the workspace.
1616    
1617    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1618    
1619    
1620    Version 8.01 19-Jan-2010
1621    ------------------------
1622    
1623    1.  If a pattern contained a conditional subpattern with only one branch (in
1624        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1625        computed the wrong minimum data length (which is of course zero for such
1626        subpatterns). This could cause incorrect "no match" results.
1627    
1628    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1629        the pattern is reset in the first branch, pcre_compile() failed with
1630        "internal error: code overflow at offset...". This happened only when
1631        the reset was to the original external option setting. (An optimization
1632        abstracts leading options settings into an external setting, which was the
1633        cause of this.)
1634    
1635    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1636        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1637        assertion pattern did not match (meaning that the assertion was true), it
1638        was incorrectly treated as false if the SKIP had been reached during the
1639        matching. This also applied to assertions used as conditions.
1640    
1641    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1642        assertion subpattern, including such a pattern used as a condition,
1643        unpredictable results occurred, instead of the error return
1644        PCRE_ERROR_DFA_UITEM.
1645    
1646    5.  The C++ GlobalReplace function was not working like Perl for the special
1647        situation when an empty string is matched. It now does the fancy magic
1648        stuff that is necessary.
1649    
1650    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1651        removed. (These were left over from very, very early versions of PCRE.)
1652    
1653    7.  Some cosmetic changes to the code to make life easier when compiling it
1654        as part of something else:
1655    
1656        (a) Change DEBUG to PCRE_DEBUG.
1657    
1658        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1659            called "current" as "current_branch", to prevent a collision with the
1660            Linux macro when compiled as a kernel module.
1661    
1662        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1663            prevent a collision with the Linux macro when compiled as a kernel
1664            module.
1665    
1666    8.  In pcre_compile() there are some checks for integer overflows that used to
1667        cast potentially large values to (double). This has been changed to that
1668        when building, a check for int64_t is made, and if it is found, it is used
1669        instead, thus avoiding the use of floating point arithmetic. (There is no
1670        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1671        double.
1672    
1673    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1674        2005 (difference between two addresses compared to an unsigned value).
1675    
1676    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1677        custom one, because of the following reported problem in Windows:
1678    
1679          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1680              under Win32.
1681          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1682              therefore missing the function definition.
1683          - The compiler thus generates a "C" signature for the test function.
1684          - The linker fails to find the "C" function.
1685          - PCRE fails to configure if asked to do so against libbz2.
1686    
1687    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1688        messages were output:
1689    
1690          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1691          rerunning libtoolize, to keep the correct libtool macros in-tree.
1692          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1693    
1694        I have done both of these things.
1695    
1696    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1697        most of the time, it *can* run out if it is given a pattern that contains a
1698        runaway infinite recursion. I updated the discussion in the pcrestack man
1699        page.
1700    
1701    13. Now that we have gone to the x.xx style of version numbers, the minor
1702        version may start with zero. Using 08 or 09 is a bad idea because users
1703        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1704        interpreted as invalid octal numbers. I've updated the previous comment in
1705        configure.ac, and also added a check that gives an error if 08 or 09 are
1706        used.
1707    
1708    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1709        causing partial matching to fail when the end of the subject matched \W
1710        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1711    
1712    15. There were some discrepancies between the declarations in pcre_internal.h
1713        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1714        their definitions. The declarations used "const uschar *" and the
1715        definitions used USPTR. Even though USPTR is normally defined as "const
1716        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1717        reported that: "This difference in casting confuses some C++ compilers, for
1718        example, SunCC recognizes above declarations as different functions and
1719        generates broken code for hbpcre." I have changed the declarations to use
1720        USPTR.
1721    
1722    16. GNU libtool is named differently on some systems. The autogen.sh script now
1723        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1724        (FreeBSD).
1725    
1726    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1727        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1728        comment: "Figure out how to create a longlong from a string: strtoll and
1729        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1730        instance, but it only takes 2 args instead of 3!"
1731    
1732    18. A subtle bug concerned with back references has been fixed by a change of
1733        specification, with a corresponding code fix. A pattern such as
1734        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1735        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1736        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1737        same bug. Such groups have to be quantified to be useful, or contained
1738        inside another quantified group. (If there's no repetition, the reference
1739        can never match.) The problem arises because, having left the group and
1740        moved on to the rest of the pattern, a later failure that backtracks into
1741        the group uses the captured value from the final iteration of the group
1742        rather than the correct earlier one. I have fixed this in PCRE by forcing
1743        any group that contains a reference to itself to be an atomic group; that
1744        is, there cannot be any backtracking into it once it has completed. This is
1745        similar to recursive and subroutine calls.
1746    
1747    
1748    Version 8.00 19-Oct-09
1749    ----------------------
1750    
1751    1.  The table for translating pcre_compile() error codes into POSIX error codes
1752        was out-of-date, and there was no check on the pcre_compile() error code
1753        being within the table. This could lead to an OK return being given in
1754        error.
1755    
1756    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1757        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1758        in a Windows environment.
1759    
1760    3.  The pcregrep --count option prints the count for each file even when it is
1761        zero, as does GNU grep. However, pcregrep was also printing all files when
1762        --files-with-matches was added. Now, when both options are given, it prints
1763        counts only for those files that have at least one match. (GNU grep just
1764        prints the file name in this circumstance, but including the count seems
1765        more useful - otherwise, why use --count?) Also ensured that the
1766        combination -clh just lists non-zero counts, with no names.
1767    
1768    4.  The long form of the pcregrep -F option was incorrectly implemented as
1769        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1770        but it seems right to fix it, and I didn't think it was worth preserving
1771        the old behaviour.
1772    
1773    5.  The command line items --regex=pattern and --regexp=pattern were not
1774        recognized by pcregrep, which required --regex pattern or --regexp pattern
1775        (with a space rather than an '='). The man page documented the '=' forms,
1776        which are compatible with GNU grep; these now work.
1777    
1778    6.  No libpcreposix.pc file was created for pkg-config; there was just
1779        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1780    
1781    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1782        when UCP support is not needed, by modifying the Python script that
1783        generates it from Unicode data files. This should not matter if the module
1784        is correctly used as a library, but I received one complaint about 50K of
1785        unwanted data. My guess is that the person linked everything into his
1786        program rather than using a library. Anyway, it does no harm.
1787    
1788    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1789        was a minimum greater than 1 for a wide character in a possessive
1790        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1791        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1792        character. Chaos in the form of incorrect output or a compiling loop could
1793        result.
1794    
1795    9.  The restrictions on what a pattern can contain when partial matching is
1796        requested for pcre_exec() have been removed. All patterns can now be
1797        partially matched by this function. In addition, if there are at least two
1798        slots in the offset vector, the offset of the earliest inspected character
1799        for the match and the offset of the end of the subject are set in them when
1800        PCRE_ERROR_PARTIAL is returned.
1801    
1802    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1803        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1804        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1805        and may be more useful for multi-segment matching.
1806    
1807    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1808        used to be given if ever the end of the subject was reached; now it is
1809        given only if matching could not proceed because another character was
1810        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1811        string "Z", which now yields "no match" instead of "partial match". In the
1812        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1813        final character ended with (*FAIL).
1814    
1815    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1816        if the pattern had a "must contain" character that was already found in the
1817        earlier partial match, unless partial matching was again requested. For
1818        example, with the pattern /dog.(body)?/, the "must contain" character is
1819        "g". If the first part-match was for the string "dog", restarting with
1820        "sbody" failed. This bug has been fixed.
1821    
1822    13. The string returned by pcre_dfa_exec() after a partial match has been
1823        changed so that it starts at the first inspected character rather than the
1824        first character of the match. This makes a difference only if the pattern
1825        starts with a lookbehind assertion or \b or \B (\K is not supported by
1826        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1827        matching functions compatible, and I think it's the right thing to do.
1828    
1829    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1830        so that the demonstration program is easily available in environments where
1831        PCRE has not been installed from source.
1832    
1833    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1834        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1835        library.
1836    
1837    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1838        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1839        is not the first non-POSIX option to be added. Clearly some people find
1840        these options useful.
1841    
1842    17. If a caller to the POSIX matching function regexec() passes a non-zero
1843        value for nmatch with a NULL value for pmatch, the value of
1844        nmatch is forced to zero.
1845    
1846    18. RunGrepTest did not have a test for the availability of the -u option of
1847        the diff command, as RunTest does. It now checks in the same way as
1848        RunTest, and also checks for the -b option.
1849    
1850    19. If an odd number of negated classes containing just a single character
1851        interposed, within parentheses, between a forward reference to a named
1852        subpattern and the definition of the subpattern, compilation crashed with
1853        an internal error, complaining that it could not find the referenced
1854        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1855        [The bug was that it was starting one character too far in when skipping
1856        over the character class, thus treating the ] as data rather than
1857        terminating the class. This meant it could skip too much.]
1858    
1859    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1860        /g option in pcretest when the pattern contains \K, which makes it possible
1861        to have an empty string match not at the start, even when the pattern is
1862        anchored. Updated pcretest and pcredemo to use this option.
1863    
1864    21. If the maximum number of capturing subpatterns in a recursion was greater
1865        than the maximum at the outer level, the higher number was returned, but
1866        with unset values at the outer level. The correct (outer level) value is
1867        now given.
1868    
1869    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1870        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1871        make it do so. The string so far is captured, making this feature
1872        compatible with Perl.
1873    
1874    23. The tests have been re-organized, adding tests 11 and 12, to make it
1875        possible to check the Perl 5.10 features against Perl 5.10.
1876    
1877    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1878        pattern matches a fixed length string. PCRE did not allow this; now it
1879        does. Neither allows recursion.
1880    
1881    25. I finally figured out how to implement a request to provide the minimum
1882        length of subject string that was needed in order to match a given pattern.
1883        (It was back references and recursion that I had previously got hung up
1884        on.) This code has now been added to pcre_study(); it finds a lower bound
1885        to the length of subject needed. It is not necessarily the greatest lower
1886        bound, but using it to avoid searching strings that are too short does give
1887        some useful speed-ups. The value is available to calling programs via
1888        pcre_fullinfo().
1889    
1890    26. While implementing 25, I discovered to my embarrassment that pcretest had
1891        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1892        study optimizations had never been tested with that matching function.
1893        Oops. What is worse, even when it was passed study data, there was a bug in
1894        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1895        were also very few tests of studied patterns with pcre_dfa_exec().
1896    
1897    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1898        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1899        on the other side of the coin, they are no longer allowed to have different
1900        names, because these cannot be distinguished in PCRE, and this has caused
1901        confusion. (This is a difference from Perl.)
1902    
1903    28. When duplicate subpattern names are present (necessarily with different
1904        numbers, as required by 27 above), and a test is made by name in a
1905        conditional pattern, either for a subpattern having been matched, or for
1906        recursion in such a pattern, all the associated numbered subpatterns are
1907        tested, and the overall condition is true if the condition is true for any
1908        one of them. This is the way Perl works, and is also more like the way
1909        testing by number works.
1910    
1911    
1912    Version 7.9 11-Apr-09
1913    ---------------------
1914    
1915    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1916        (pcretest), all targets were linked against these libraries. This included
1917        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1918        libraries. This caused unwanted dependencies to be created. This problem
1919        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1920        pcretest is linked with readline.
1921    
1922    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1923        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1924        moved outside it again, because FALSE and TRUE are already defined in AIX,
1925        but BOOL is not.
1926    
1927    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1928        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1929    
1930    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1931        hyphens) following file names and line numbers when outputting matching
1932        lines. This is not true; no spaces are inserted. I have also clarified the
1933        wording for the --colour (or --color) option.
1934    
1935    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1936        was not coloured; this is different to GNU grep, so I have changed it to be
1937        the same.
1938    
1939    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1940        each matching line was coloured. Now it goes on to look for further matches
1941        of any of the test patterns, which is the same behaviour as GNU grep.
1942    
1943    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1944        doesn't make sense to accept an empty string match in pcregrep, so I have
1945        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1946        seems to be how GNU grep behaves. [But see later change 40 for release
1947        8.33.]
1948    
1949    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1950        start or after a newline", because the conditional assertion was not being
1951        correctly handled. The rule now is that both the assertion and what follows
1952        in the first alternative must satisfy the test.
1953    
1954    9.  If auto-callout was enabled in a pattern with a conditional group whose
1955        condition was an assertion, PCRE could crash during matching, both with
1956        pcre_exec() and pcre_dfa_exec().
1957    
1958    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1959        used for matching.
1960    
1961    11. Unicode property support in character classes was not working for
1962        characters (bytes) greater than 127 when not in UTF-8 mode.
1963    
1964    12. Added the -M command line option to pcretest.
1965    
1966    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1967    
1968    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1969    
1970    16. Added comments and documentation about mis-use of no_arg in the C++
1971        wrapper.
1972    
1973    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1974        from Martin Jerabek that uses macro names for all relevant character and
1975        string constants.
1976    
1977    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1978        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1979        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1980        these, but not everybody uses configure.
1981    
1982    19. A conditional group that had only one branch was not being correctly
1983        recognized as an item that could match an empty string. This meant that an
1984        enclosing group might also not be so recognized, causing infinite looping
1985        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1986        with the subject "ab", where knowledge that the repeated group can match
1987        nothing is needed in order to break the loop.
1988    
1989    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1990        exec(), but without supplying a callout function, matching went wrong.
1991    
1992    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1993        leak if the size of the offset vector was greater than 30. When the vector
1994        is smaller, the saved offsets during recursion go onto a local stack
1995        vector, but for larger vectors malloc() is used. It was failing to free
1996        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1997        error, in fact).
1998    
1999    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2000        heapframe that is used only when UTF-8 support is enabled. This caused no
2001        problem, but was untidy.
2002    
2003    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2004        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2005        included within another project.
2006    
2007    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2008        slightly modified by me:
2009    
2010          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2011              not building pcregrep.
2012    
2013          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2014              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2015    
2016    25. Forward references, both numeric and by name, in patterns that made use of
2017        duplicate group numbers, could behave incorrectly or give incorrect errors,
2018        because when scanning forward to find the reference group, PCRE was not
2019        taking into account the duplicate group numbers. A pattern such as
2020        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2021    
2022    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2023        the feature of a custom pointer more persuasive (as requested by a user).
2024    
2025    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2026        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2027        pre-defined.
2028    
2029    28. Added support for (*UTF8) at the start of a pattern.
2030    
2031    29. Arrange for flags added by the "release type" setting in CMake to be shown
2032        in the configuration summary.
2033    
2034    
2035    Version 7.8 05-Sep-08
2036    ---------------------
2037    
2038    1.  Replaced UCP searching code with optimized version as implemented for Ad
2039        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2040        stage table and inline lookup instead of a function, giving speed ups of 2
2041        to 5 times on some simple patterns that I tested. Permission was given to
2042        distribute the MultiStage2.py script that generates the tables (it's not in
2043        the tarball, but is in the Subversion repository).
2044    
2045    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2046        scripts.
2047    
2048    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2049        a group with a zero qualifier. The result of the study could be incorrect,
2050        or the function might crash, depending on the pattern.
2051    
2052    4.  Caseless matching was not working for non-ASCII characters in back
2053        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2054        It now works when Unicode Property Support is available.
2055    
2056    5.  In pcretest, an escape such as \x{de} in the data was always generating
2057        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2058        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2059        truncation.
2060    
2061    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2062    
2063    7.  Added two (int) casts to pcregrep when printing the difference of two
2064        pointers, in case they are 64-bit values.
2065    
2066    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2067        test 2 if it fails.
2068    
2069    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2070        and a #define of that name to empty if it is not externally set. This is to
2071        allow users of MSVC to set it if necessary.
2072    
2073    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2074        the convenience functions in the pcre_get.c source file.
2075    
2076    11. An option change at the start of a pattern that had top-level alternatives
2077        could cause overwriting and/or a crash. This command provoked a crash in
2078        some environments:
2079    
2080          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2081    
2082        This potential security problem was recorded as CVE-2008-2371.
2083    
2084    12. For a pattern where the match had to start at the beginning or immediately
2085        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2086        pcre_dfa_exec() could read past the end of the passed subject if there was
2087        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2088        pcretest so that it places the subject at the end of its malloc-ed buffer.
2089    
2090    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2091        exec() might read past the end of the data buffer in UTF-8 mode.
2092    
2093    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2094        the data contained the byte 0x85 as part of a UTF-8 character within its
2095        first line. This applied both to normal and DFA matching.
2096    
2097    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2098        /^[^d]*?$/8 failed to match "abc".
2099    
2100    16. Added a missing copyright notice to pcrecpp_internal.h.
2101    
2102    17. Make it more clear in the documentation that values returned from
2103        pcre_exec() in ovector are byte offsets, not character counts.
2104    
2105    18. Tidied a few places to stop certain compilers from issuing warnings.
2106    
2107    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2108        supplied by Stefan Weber. I made a further small update for 7.8 because
2109        there is a change of source arrangements: the pcre_searchfuncs.c module is
2110        replaced by pcre_ucd.c.
2111    
2112    
2113    Version 7.7 07-May-08
2114    ---------------------
2115    
2116    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2117        a string to a long long, pretend we don't even have a long long." This is
2118        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2119    
2120    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2121        pre-7.6 versions, which defined a global no_arg variable instead of putting
2122        it in the RE class. (See also #8 below.)
2123    
2124    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2125        Lopes.
2126    
2127    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2128    
2129        (1) The include/exclude patterns were being applied to the whole pathnames
2130            of files, instead of just to the final components.
2131    
2132        (2) If there was more than one level of directory, the subdirectories were
2133            skipped unless they satisfied the include/exclude conditions. This is
2134            inconsistent with GNU grep (and could even be seen as contrary to the
2135            pcregrep specification - which I improved to make it absolutely clear).
2136            The action now is always to scan all levels of directory, and just
2137            apply the include/exclude patterns to regular files.
2138    
2139    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2140        --exclude_dir in the tests to avoid scanning .svn directories.
2141    
2142    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2143        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2144        doesn't support NULs in patterns.
2145    
2146    7.  Added some missing "const"s to declarations of static tables in
2147        pcre_compile.c and pcre_dfa_exec.c.
2148    
2149    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2150        caused by fix #2  above. (Subsequently also a second patch to fix the
2151        first patch. And a third patch - this was a messy problem.)
2152    
2153    9.  Applied Craig's patch to remove the use of push_back().
2154    
2155    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2156        matching function regexec().
2157    
2158    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2159        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2160        references. PCRE supports relative numbers with this syntax (I don't think
2161        Oniguruma does).
2162    
2163    12. Previously, a group with a zero repeat such as (...){0} was completely
2164        omitted from the compiled regex. However, this means that if the group
2165        was called as a subroutine from elsewhere in the pattern, things went wrong
2166        (an internal error was given). Such groups are now left in the compiled
2167        pattern, with a new opcode that causes them to be skipped at execution
2168        time.
2169    
2170    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2171        to the way PCRE behaves:
2172    
2173        (a) A lone ] character is dis-allowed (Perl treats it as data).
2174    
2175        (b) A back reference to an unmatched subpattern matches an empty string
2176            (Perl fails the current match path).
2177    
2178        (c) A data ] in a character class must be notated as \] because if the
2179            first data character in a class is ], it defines an empty class. (In
2180            Perl it is not possible to have an empty class.) The empty class []
2181            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2182            The negative empty class [^] matches any one character, independently
2183            of the DOTALL setting.
2184    
2185    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2186        non-existent subpattern following a character class starting with ']' and
2187        containing () gave an internal compiling error instead of "reference to
2188        non-existent subpattern". Fortunately, when the pattern did exist, the
2189        compiled code was correct. (When scanning forwards to check for the
2190        existence of the subpattern, it was treating the data ']' as terminating
2191        the class, so got the count wrong. When actually compiling, the reference
2192        was subsequently set up correctly.)
2193    
2194    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2195        it was being rejected as not supported by pcre_dfa_exec(), even though
2196        other assertions are supported. I have made pcre_dfa_exec() support
2197        (*FAIL).
2198    
2199    16. The implementation of 13c above involved the invention of a new opcode,
2200        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2201        cannot be changed at match time, I realized I could make a small
2202        improvement to matching performance by compiling OP_ALLANY instead of
2203        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2204        on the OP_ANY path.
2205    
2206    17. Compiling pcretest on Windows with readline support failed without the
2207        following two fixes: (1) Make the unistd.h include conditional on
2208        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2209    
2210    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2211        ncurses library to be included for pcretest when ReadLine support is
2212        requested, but also to allow for it to be overridden. This patch came from
2213        Daniel Bergström.
2214    
2215    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2216        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2217        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2218        spotting this.
2219    
2220    
2221    Version 7.6 28-Jan-08
2222    ---------------------
2223    
2224    1.  A character class containing a very large number of characters with
2225        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2226        overflow.
2227    
2228    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2229        HAVE_LONG_LONG is not defined.
2230    
2231    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2232        bring them up to date and include new features. This patch includes:
2233    
2234        - Fixed PH's badly added libz and libbz2 support.
2235        - Fixed a problem with static linking.
2236        - Added pcredemo. [But later removed - see 7 below.]
2237        - Fixed dftables problem and added an option.
2238        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2239            HAVE_LONG_LONG.
2240        - Added readline support for pcretest.
2241        - Added an listing of the option settings after cmake has run.
2242    
2243    4.  A user submitted a patch to Makefile that makes it easy to create
2244        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2245        Makefile.am that cause it to include this special target, without
2246        affecting anything else. Note that the same mingw target plus all
2247        the other distribution libraries and programs are now supported
2248        when configuring with CMake (see 6 below) instead of with
2249        Configure/Make.
2250    
2251    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2252        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2253        exported in the Windows port". It has not yet been confirmed that the patch
2254        solves the problem, but it does no harm.
2255    
2256    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2257        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2258        with CMake, and also correct the comment about stack recursion.
2259    
2260    7.  Remove the automatic building of pcredemo from the ./configure system and
2261        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2262        of a program that users should build themselves after PCRE is installed, so
2263        building it automatically is not really right. What is more, it gave
2264        trouble in some build environments.
2265    
2266    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2267    
2268    
2269    Version 7.5 10-Jan-08
2270    ---------------------
2271    
2272    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2273        values in parens when parsing an RE using the C++ wrapper."
2274    
2275    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2276        Characters greater than 255 were excluded from the class instead of being
2277        included.
2278    
2279    3.  The same bug as (2) above applied to negated POSIX classes such as
2280        [:^space:].
2281    
2282    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2283        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2284        I have changed it.
2285    
2286    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2287        first named subpattern) and a construct such as (?&a) would reference the
2288        first named subpattern whose name started with "a" (in other words, the
2289        length check was missing). Both these problems are fixed. "Subpattern name
2290        expected" is now given for (?&) (a zero-length name), and this patch also
2291        makes it give the same error for \k'' (previously it complained that that
2292        was a reference to a non-existent subpattern).
2293    
2294    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2295        this is right because (?- can be followed by option settings as well as by
2296        digits. I have, however, made the messages clearer.
2297    
2298    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2299        than the number used in the conditional) now cause a compile-time error.
2300        This is actually not compatible with Perl, which accepts such patterns, but
2301        treats the conditional as always being FALSE (as PCRE used to), but it
2302        seems to me that giving a diagnostic is better.
2303    
2304    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2305        and messages.
2306    
2307    9.  Fix two occurrences of "backslash" in comments that should have been
2308        "backspace".
2309    
2310    10. Remove two redundant lines of code that can never be obeyed (their function
2311        was moved elsewhere).
2312    
2313    11. The program that makes PCRE's Unicode character property table had a bug
2314        which caused it to generate incorrect table entries for sequences of
2315        characters that have the same character type, but are in different scripts.
2316        It amalgamated them into a single range, with the script of the first of
2317        them. In other words, some characters were in the wrong script. There were
2318        thirteen such cases, affecting characters in the following ranges:
2319    
2320          U+002b0 - U+002c1
2321          U+0060c - U+0060d
2322          U+0061e - U+00612
2323          U+0064b - U+0065e
2324          U+0074d - U+0076d
2325          U+01800 - U+01805
2326          U+01d00 - U+01d77
2327          U+01d9b - U+01dbf
2328          U+0200b - U+0200f
2329          U+030fc - U+030fe
2330          U+03260 - U+0327f
2331          U+0fb46 - U+0fbb1
2332          U+10450 - U+1049d
2333    
2334    12. The -o option (show only the matching part of a line) for pcregrep was not
2335        compatible with GNU grep in that, if there was more than one match in a
2336        line, it showed only the first of them. It now behaves in the same way as
2337        GNU grep.
2338    
2339    13. If the -o and -v options were combined for pcregrep, it printed a blank
2340        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2341        does the same. The return code can be used to tell if there were any
2342        non-matching lines.
2343    
2344    14. Added --file-offsets and --line-offsets to pcregrep.
2345    
2346    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2347        infinitely looping recursion. The bug was that positive lookaheads were not
2348        being skipped when checking for a possible empty match (negative lookaheads
2349        and both kinds of lookbehind were skipped).
2350    
2351    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2352        inclusion of <windows.h> to before rather than after the definition of
2353        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2354    
2355    17. Specifying a possessive quantifier with a specific limit for a Unicode
2356        character property caused pcre_compile() to compile bad code, which led at
2357        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2358        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2359        caused the error; without that there was no problem.
2360    
2361    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2362    
2363    19. Added --enable-pcretest-libreadline.
2364    
2365    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2366        RE::GlobalReplace(). As a result, the number of replacements returned was
2367        double what it should be. I removed one of the increments, but Craig sent a
2368        later patch that removed the other one (the right fix) and added unit tests
2369        that check the return values (which was not done before).
2370    
2371    21. Several CMake things:
2372    
2373        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2374            the names libpcre and libpcreposix, not just pcre and pcreposix.
2375    
2376        (2) The above change means that pcretest and pcregrep are now correctly
2377            linked with the newly-built libraries, not previously installed ones.
2378    
2379        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2380    
2381    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2382        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2383        UTF-8 newline character). The key issue is that the pattern starts .*;
2384        this means that the match must be either at the beginning, or after a
2385        newline. The bug was in the code for advancing after a failed match and
2386        checking that the new position followed a newline. It was not taking
2387        account of UTF-8 characters correctly.
2388    
2389    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2390        character classes. PCRE was not treating the sequence [:...:] as a
2391        character class unless the ... were all letters. Perl, however, seems to
2392        allow any characters between [: and :], though of course it rejects as
2393        unknown any "names" that contain non-letters, because all the known class
2394        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2395        for example, whereas PCRE did not - it did not recognize a POSIX character
2396        class. This seemed a bit dangerous, so the code has been changed to be
2397        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2398        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2399        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2400        Perl does, and where it didn't before.
2401    
2402    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2403        Windows environments %n is disabled by default.
2404    
2405    
2406    Version 7.4 21-Sep-07
2407    ---------------------
2408    
2409    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2410        means that a class such as [\s] counted as "explicit reference to CR or
2411        LF". That isn't really right - the whole point of the change was to try to
2412        help when there was an actual mention of one of the two characters. So now
2413        the change happens only if \r or \n (or a literal CR or LF) character is
2414        encountered.
2415    
2416    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2417        of both had grown to the point where there were only 3 bits left.
2418        Fortunately, there was spare space in the data structure, and so I have
2419        moved the internal flags into a new 16-bit field to free up more option
2420        bits.
2421    
2422    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2423        but did not set the internal JCHANGED flag - either of these is enough to
2424        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2425        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2426        start sets both bits.
2427    
2428    4.  Added options (at build time, compile time, exec time) to change \R from
2429        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2430    
2431    5.  doc/pcresyntax.html was missing from the distribution.
2432    
2433    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2434        compatibility, even though it is no longer used.
2435    
2436    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2437        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2438        windows.h file is present (where different names are used). [This was
2439        reversed later after testing - see 16 below.]
2440    
2441    8.  Changed all #include <config.h> to #include "config.h". There were also
2442        some further <pcre.h> cases that I changed to "pcre.h".
2443    
2444    9.  When pcregrep was used with the --colour option, it missed the line ending
2445        sequence off the lines that it output.
2446    
2447    10. It was pointed out to me that arrays of string pointers cause lots of
2448        relocations when a shared library is dynamically loaded. A technique of
2449        using a single long string with a table of offsets can drastically reduce
2450        these. I have refactored PCRE in four places to do this. The result is
2451        dramatic:
2452    
2453          Originally:                          290
2454          After changing UCP table:            187
2455          After changing error message table:   43
2456          After changing table of "verbs"       36
2457          After changing table of Posix names   22
2458    
2459        Thanks to the folks working on Gregex for glib for this insight.
2460    
2461    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2462        unicode-properties was also set.
2463    
2464    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2465    
2466    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2467        checked only for CRLF.
2468    
2469    14. Added casts to pcretest.c to avoid compiler warnings.
2470    
2471    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2472    
2473    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2474        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2475        entirely. This removes changes made in 7 above.
2476    
2477    17. The CMake files have been updated, and there is now more information about
2478        building with CMake in the NON-UNIX-USE document.
2479    
2480    
2481    Version 7.3 28-Aug-07
2482    ---------------------
2483    
2484     1. In the rejigging of the build system that eventually resulted in 7.1, the
2485        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2486        brackets there is not right, since it causes compilers to look for an
2487        installed pcre.h, not the version that is in the source that is being
2488        compiled (which of course may be different). I have changed it back to:
2489    
2490          #include "pcre.h"
2491    
2492        I have a vague recollection that the change was concerned with compiling in
2493        different directories, but in the new build system, that is taken care of
2494        by the VPATH setting the Makefile.
2495    
2496     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2497        when the subject happened to end in the byte 0x85 (e.g. if the last
2498        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2499        characters but of course it shouldn't be taken as a newline when it is part
2500        of another character. The bug was that, for an unlimited repeat of . in
2501        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2502        characters when looking for a newline.
2503    
2504     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2505    
2506     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2507        in debug output.
2508    
2509     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2510        long printing in the pcrecpp unittest when running under MinGW.
2511    
2512     6. ESC_K was left out of the EBCDIC table.
2513    
2514     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2515        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2516        limit also applies to "virtual nesting" when a pattern is recursive, and in
2517        this case 1000 isn't so big. I have been able to remove this limit at the
2518        expense of backing off one optimization in certain circumstances. Normally,
2519        when pcre_exec() would call its internal match() function recursively and
2520        immediately return the result unconditionally, it uses a "tail recursion"
2521        feature to save stack. However, when a subpattern that can match an empty
2522        string has an unlimited repetition quantifier, it no longer makes this
2523        optimization. That gives it a stack frame in which to save the data for
2524        checking that an empty string has been matched. Previously this was taken
2525        from the 1000-entry workspace that had been reserved. So now there is no
2526        explicit limit, but more stack is used.
2527    
2528     8. Applied Daniel's patches to solve problems with the import/export magic
2529        syntax that is required for Windows, and which was going wrong for the
2530        pcreposix and pcrecpp parts of the library. These were overlooked when this
2531        problem was solved for the main library.
2532    
2533     9. There were some crude static tests to avoid integer overflow when computing
2534        the size of patterns that contain repeated groups with explicit upper
2535        limits. As the maximum quantifier is 65535, the maximum group length was
2536        set at 30,000 so that the product of these two numbers did not overflow a
2537        32-bit integer. However, it turns out that people want to use groups that
2538        are longer than 30,000 bytes (though not repeat them that many times).
2539        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2540        made it possible to implement the integer overflow checks in a much more
2541        dynamic way, which I have now done. The artificial limitation on group
2542        length has been removed - we now have only the limit on the total length of
2543        the compiled pattern, which depends on the LINK_SIZE setting.
2544    
2545    10. Fixed a bug in the documentation for get/copy named substring when
2546        duplicate names are permitted. If none of the named substrings are set, the
2547        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2548        empty string.
2549    
2550    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2551        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2552        because the ] is interpreted as the first data character and the
2553        terminating ] is not found. PCRE has been made compatible with Perl in this
2554        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2555        cause memory overwriting.
2556    
2557    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2558        string has been matched (to stop an infinite loop). It was not recognizing
2559        a conditional subpattern that could match an empty string if that
2560        subpattern was within another subpattern. For example, it looped when
2561        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2562        condition was not nested. This bug has been fixed.
2563    
2564    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2565        past the start of the subject in the presence of bytes with the top bit
2566        set, for example "\x8aBCD".
2567    
2568    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2569        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2570    
2571    14. Optimized (?!) to (*FAIL).
2572    
2573    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2574        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2575        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2576        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2577        does: it's just the validity check that is more restrictive.
2578    
2579    16. Inserted checks for integer overflows during escape sequence (backslash)
2580        processing, and also fixed erroneous offset values for syntax errors during
2581        backslash processing.
2582    
2583    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2584        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2585    
2586    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2587        caused an overrun.
2588    
2589    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2590        something other than just ASCII characters) inside a group that had an
2591        unlimited repeat caused a loop at compile time (while checking to see
2592        whether the group could match an empty string).
2593    
2594    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2595        [\P{Any}] did so. (Error in the code for printing property names.)
2596    
2597    21. An orphan \E inside a character class could cause a crash.
2598    
2599    22. A repeated capturing bracket such as (A)? could cause a wild memory
2600        reference during compilation.
2601    
2602    23. There are several functions in pcre_compile() that scan along a compiled
2603        expression for various reasons (e.g. to see if it's fixed length for look
2604        behind). There were bugs in these functions when a repeated \p or \P was
2605        present in the pattern. These operators have additional parameters compared
2606        with \d, etc, and these were not being taken into account when moving along
2607        the compiled data. Specifically:
2608    
2609        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2610            length.
2611    
2612        (b) An item such as \pL+ within a repeated group could cause crashes or
2613            loops.
2614    
2615        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2616            "reference to non-existent subpattern" error.
2617    
2618        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2619    
2620    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2621        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2622    
2623    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2624    
2625    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2626        character were causing crashes (broken optimization).
2627    
2628    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2629        \p or \P) caused a compile-time loop.
2630    
2631    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2632        break. For example, the unstudied pattern [\r\n]A does not match the string
2633        "\r\nA" because change 7.0/46 below moves the current point on by two
2634        characters after failing to match at the start. However, the pattern \nA
2635        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2636        the same is true. There doesn't seem any very clean way out of this, but
2637        what I have chosen to do makes the common cases work: PCRE now takes note
2638        of whether there can be an explicit match for \r or \n anywhere in the
2639        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2640        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2641        pattern has explicit CR or LF references.
2642    
2643    29. Added (*CR) etc for changing newline setting at start of pattern.
2644    
2645    
2646    Version 7.2 19-Jun-07
2647  ---------------------  ---------------------
2648    
2649   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
# Line 41  Version 7.2 05-June-07 Line 2683  Version 7.2 05-June-07
2683      (f) \g{name} is another synonym - part of Perl 5.10's unification of      (f) \g{name} is another synonym - part of Perl 5.10's unification of
2684          reference syntax.          reference syntax.
2685    
2686        (g) (?| introduces a group in which the numbering of parentheses in each
2687            alternative starts with the same number.
2688    
2689        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2690    
2691   7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and   7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2692      PCRE_INFO_JCHANGED.      PCRE_INFO_JCHANGED.
2693    
# Line 55  Version 7.2 05-June-07 Line 2702  Version 7.2 05-June-07
2702      alternatives. The 1000-alternative test pattern now uses 12 bytes of      alternatives. The 1000-alternative test pattern now uses 12 bytes of
2703      workspace instead of running out of the 4096 that are available.      workspace instead of running out of the 4096 that are available.
2704    
2705    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2706    
2707    11. Applied patch from Google to remove an optimization that didn't quite work.
2708        The report of the bug said:
2709    
2710          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2711          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2712          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2713    
2714    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2715        it matched the wrong number of bytes.
2716    
2717    
2718  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
2719  ---------------------  ---------------------

Legend:
Removed from v.172  
changed lines
  Added in v.1438

  ViewVC Help
Powered by ViewVC 1.1.5