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

Diff of /code/trunk/ChangeLog

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

revision 158 by ph10, Tue May 1 10:30:14 2007 UTC revision 1439 by ph10, Fri Jan 10 16:25:55 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-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    13. The configure.ac file always checked for pthread support when JIT was
64        enabled. This is not used in Windows, so I have put this test inside a
65        check for the presence of windows.h (which was already tested for).
66    
67    
68    Version 8.34 15-December-2013
69    -----------------------------
70    
71    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
72        executable memory. Patch inspired by Carsten Klein.
73    
74    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
75        this macro is never tested and has no effect, because the work to support
76        coverage involves only compiling and linking options and special targets in
77        the Makefile. The comment in config.h implied that defining the macro would
78        enable coverage support, which is totally false. There was also support for
79        setting this macro in the CMake files (my fault, I just copied it from
80        configure). SUPPORT_GCOV has now been removed.
81    
82    3.  Make a small performance improvement in strlen16() and strlen32() in
83        pcretest.
84    
85    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
86        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
87        these cases). There was also one in pcretest.c.
88    
89    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
90    
91    6.  In UTF mode, the code for checking whether a group could match an empty
92        string (which is used for indefinitely repeated groups to allow for
93        breaking an infinite loop) was broken when the group contained a repeated
94        negated single-character class with a character that occupied more than one
95        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
96        UTF-8 mode). The effect was undefined: the group might or might not be
97        deemed as matching an empty string, or the program might have crashed.
98    
99    7.  The code for checking whether a group could match an empty string was not
100        recognizing that \h, \H, \v, \V, and \R must match a character.
101    
102    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
103        an empty string. If it can, pcretest shows this in its information output.
104    
105    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
106        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
107        matched by pcre_exec() without using JIT:
108    
109        (a) If the rest of the pattern did not match after a maximal run of
110            grapheme clusters, the code for backing up to try with fewer of them
111            did not always back up over a full grapheme when characters that do not
112            have the modifier quality were involved, e.g. Hangul syllables.
113    
114        (b) If the match point in a subject started with modifier character, and
115            there was no match, the code could incorrectly back up beyond the match
116            point, and potentially beyond the first character in the subject,
117            leading to a segfault or an incorrect match result.
118    
119    10. A conditional group with an assertion condition could lead to PCRE
120        recording an incorrect first data item for a match if no other first data
121        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
122        first data item, and therefore matched "ca" after "c" instead of at the
123        start.
124    
125    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
126        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
127    
128    12. The source of pcregrep now includes z/OS-specific code so that it can be
129        compiled for z/OS as part of the special z/OS distribution.
130    
131    13. Added the -T and -TM options to pcretest.
132    
133    14. The code in pcre_compile.c for creating the table of named capturing groups
134        has been refactored. Instead of creating the table dynamically during the
135        actual compiling pass, the information is remembered during the pre-compile
136        pass (on the stack unless there are more than 20 named groups, in which
137        case malloc() is used) and the whole table is created before the actual
138        compile happens. This has simplified the code (it is now nearly 150 lines
139        shorter) and prepared the way for better handling of references to groups
140        with duplicate names.
141    
142    15. A back reference to a named subpattern when there is more than one of the
143        same name now checks them in the order in which they appear in the pattern.
144        The first one that is set is used for the reference. Previously only the
145        first one was inspected. This change makes PCRE more compatible with Perl.
146    
147    16. Unicode character properties were updated from Unicode 6.3.0.
148    
149    17. The compile-time code for auto-possessification has been refactored, based
150        on a patch by Zoltan Herczeg. It now happens after instead of during
151        compilation. The code is cleaner, and more cases are handled. The option
152        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
153        options in pcretest are provided to set it. It can also be set by
154        (*NO_AUTO_POSSESS) at the start of a pattern.
155    
156    18. The character VT has been added to the default ("C" locale) set of
157        characters that match \s and are generally treated as white space,
158        following this same change in Perl 5.18. There is now no difference between
159        "Perl space" and "POSIX space". Whether VT is treated as white space in
160        other locales depends on the locale.
161    
162    19. The code for checking named groups as conditions, either for being set or
163        for being recursed, has been refactored (this is related to 14 and 15
164        above). Processing unduplicated named groups should now be as fast at
165        numerical groups, and processing duplicated groups should be faster than
166        before.
167    
168    20. Two patches to the CMake build system, by Alexander Barkov:
169    
170          (1) Replace the "source" command by "." in CMakeLists.txt because
171              "source" is a bash-ism.
172    
173          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
174              without these the CMake build does not work on Solaris.
175    
176    21. Perl has changed its handling of \8 and \9. If there is no previously
177        encountered capturing group of those numbers, they are treated as the
178        literal characters 8 and 9 instead of a binary zero followed by the
179        literals. PCRE now does the same.
180    
181    22. Following Perl, added \o{} to specify codepoints in octal, making it
182        possible to specify values greater than 0777 and also making them
183        unambiguous.
184    
185    23. Perl now gives an error for missing closing braces after \x{... instead of
186        treating the string as literal. PCRE now does the same.
187    
188    24. RunTest used to grumble if an inappropriate test was selected explicitly,
189        but just skip it when running all tests. This make it awkward to run ranges
190        of tests when one of them was inappropriate. Now it just skips any
191        inappropriate tests, as it always did when running all tests.
192    
193    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
194        character types such as \d or \w, too many callouts were inserted, and the
195        data that they returned was rubbish.
196    
197    26. In UCP mode, \s was not matching two of the characters that Perl matches,
198        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
199        were matched by \h. The code has now been refactored so that the lists of
200        the horizontal and vertical whitespace characters used for \h and \v (which
201        are defined only in one place) are now also used for \s.
202    
203    27. Add JIT support for the 64 bit TileGX architecture.
204        Patch by Jiong Wang (Tilera Corporation).
205    
206    28. Possessive quantifiers for classes (both explicit and automatically
207        generated) now use special opcodes instead of wrapping in ONCE brackets.
208    
209    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
210        (because it's meaningless), this was not happening when PCRE_CASELESS was
211        set. Not wrong, but inefficient.
212    
213    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
214        properties for \w, \d, etc) is present in a test regex. Otherwise if the
215        test contains no characters greater than 255, Perl doesn't realise it
216        should be using Unicode semantics.
217    
218    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
219        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
220        does in Unicode mode.
221    
222    32. Added the "forbid" facility to pcretest so that putting tests into the
223        wrong test files can sometimes be quickly detected.
224    
225    33. There is now a limit (default 250) on the depth of nesting of parentheses.
226        This limit is imposed to control the amount of system stack used at compile
227        time. It can be changed at build time by --with-parens-nest-limit=xxx or
228        the equivalent in CMake.
229    
230    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
231        errors. Perl warns for these when in warning mode, but PCRE has no facility
232        for giving warnings.
233    
234    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
235        However, this was not working for (?!) because it is optimized to (*FAIL),
236        for which PCRE does not allow quantifiers. The optimization is now disabled
237        when a quantifier follows (?!). I can't see any use for this, but it makes
238        things uniform.
239    
240    36. Perl no longer allows group names to start with digits, so I have made this
241        change also in PCRE. It simplifies the code a bit.
242    
243    37. In extended mode, Perl ignores spaces before a + that indicates a
244        possessive quantifier. PCRE allowed a space before the quantifier, but not
245        before the possessive +. It now does.
246    
247    38. The use of \K (reset reported match start) within a repeated possessive
248        group such as (a\Kb)*+ was not working.
249    
250    40. Document that the same character tables must be used at compile time and
251        run time, and that the facility to pass tables to pcre_exec() and
252        pcre_dfa_exec() is for use only with saved/restored patterns.
253    
254    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
255        features for Builds with MSVC:
256    
257        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
258           builds). The .rc files can be used to set FileDescription and many other
259           attributes.
260    
261        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
262           This allows higher-level build scripts which want .pdb files to avoid
263           hard-coding the exact files needed."
264    
265    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
266        mean "start of word" and "end of word", respectively, as a transition aid.
267    
268    43. A minimizing repeat of a class containing codepoints greater than 255 in
269        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
270        compiled to use the heap for recursion.
271    
272    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
273        is configured.
274    
275    
276    Version 8.33 28-May-2013
277    ------------------------
278    
279    1.  Added 'U' to some constants that are compared to unsigned integers, to
280        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
281        variables that are added to signed variables, to ensure the result is
282        signed and can be negated.
283    
284    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
285        CMake config files.
286    
287    3.  Revise the creation of config.h.generic so that all boolean macros are
288        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
289        overriding via -D on the command line possible.
290    
291    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
292        to unsigned int is reported to make a quite noticeable speed difference in
293        a specific Windows environment. Testing on Linux did also appear to show
294        some benefit (and it is clearly not harmful). Also fixed the definition of
295        Xop which should be unsigned.
296    
297    5.  Related to (4), changing the definition of the intermediate variable cc
298        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
299        improvements.
300    
301    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
302        unnecessary spaces.
303    
304    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
305        and later.
306    
307    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
308    
309    9.  Optimizing fast_forward_start_bits in JIT.
310    
311    10. Adding support for callouts in JIT, and fixing some issues revealed
312        during this work. Namely:
313    
314        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
315    
316        (b) Minimum length was not checked before the matching is started.
317    
318    11. The value of capture_last that is passed to callouts was incorrect in some
319        cases when there was a capture on one path that was subsequently abandoned
320        after a backtrack. Also, the capture_last value is now reset after a
321        recursion, since all captures are also reset in this case.
322    
323    12. The interpreter no longer returns the "too many substrings" error in the
324        case when an overflowing capture is in a branch that is subsequently
325        abandoned after a backtrack.
326    
327    13. In the pathological case when an offset vector of size 2 is used, pcretest
328        now prints out the matched string after a yield of 0 or 1.
329    
330    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
331        Only supported by the JIT compiler at the moment.
332    
333    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
334    
335    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
336        offset of the starting point of the matching process, provided the offsets
337        vector is large enough.
338    
339    17. The \A escape now records a lookbehind value of 1, though its execution
340        does not actually inspect the previous character. This is to ensure that,
341        in partial multi-segment matching, at least one character from the old
342        segment is retained when a new segment is processed. Otherwise, if there
343        are no lookbehinds in the pattern, \A might match incorrectly at the start
344        of a new segment.
345    
346    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
347    
348    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
349        gives some modest performance improvement in 8-bit mode.
350    
351    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
352        be expressed in certain programming languages using Universal Character
353        Names.
354    
355    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
356        which points out that "non characters" are not "characters that may not
357        appear in Unicode strings" but rather "characters that are reserved for
358        internal use and have only local meaning".
359    
360    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
361        there was a conditional group that depended on an assertion, if the
362        assertion was false, the callout that immediately followed the alternation
363        in the condition was skipped when pcre_exec() was used for matching.
364    
365    23. Allow an explicit callout to be inserted before an assertion that is the
366        condition for a conditional group, for compatibility with automatic
367        callouts, which always insert a callout at this point.
368    
369    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
370        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
371    
372    25. (*PRUNE) is now supported by the JIT compiler.
373    
374    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
375    
376    27. Fix the case where there are two or more SKIPs with arguments that may be
377        ignored.
378    
379    28. (*SKIP) is now supported by the JIT compiler.
380    
381    29. (*THEN) is now supported by the JIT compiler.
382    
383    30. Update RunTest with additional test selector options.
384    
385    31. The way PCRE handles backtracking verbs has been changed in two ways.
386    
387        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
388        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
389        backtracking. In some cases this makes it more Perl-compatible, but Perl's
390        rather obscure rules do not always do the same thing.
391    
392        (2) Previously, backtracking verbs were confined within assertions. This is
393        no longer the case for positive assertions, except for (*ACCEPT). Again,
394        this sometimes improves Perl compatibility, and sometimes does not.
395    
396    32. A number of tests that were in test 2 because Perl did things differently
397        have been moved to test 1, because either Perl or PCRE has changed, and
398        these tests are now compatible.
399    
400    32. Backtracking control verbs are now handled in the same way in JIT and
401        interpreter.
402    
403    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
404        contained a forward subroutine reference caused a compile error.
405    
406    34. Auto-detect and optimize limited repetitions in JIT.
407    
408    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
409        blocking (*UTF) etc.
410    
411    36. In the interpreter, maximizing pattern repetitions for characters and
412        character types now use tail recursion, which reduces stack usage.
413    
414    37. The value of the max lookbehind was not correctly preserved if a compiled
415        and saved regex was reloaded on a host of different endianness.
416    
417    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
418        of the compiled pattern block, expand the flags field from 16 to 32 bits
419        because it was almost full.
420    
421    39. Try madvise first before posix_madvise.
422    
423    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
424        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
425        original change locked out all matches of empty strings. This has been
426        changed so that one match of an empty string per line is recognized.
427        Subsequent searches on the same line (for colouring or for --only-matching,
428        for example) do not recognize empty strings.
429    
430    41. Applied a user patch to fix a number of spelling mistakes in comments.
431    
432    42. Data lines longer than 65536 caused pcretest to crash.
433    
434    43. Clarified the data type for length and startoffset arguments for pcre_exec
435        and pcre_dfa_exec in the function-specific man pages, where they were
436        explicitly stated to be in bytes, never having been updated. I also added
437        some clarification to the pcreapi man page.
438    
439    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
440        a segmentation fault.
441    
442    
443    Version 8.32 30-November-2012
444    -----------------------------
445    
446    1.  Improved JIT compiler optimizations for first character search and single
447        character iterators.
448    
449    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
450        Patch by Daniel Richard G.
451    
452    3.  Single character iterator optimizations in the JIT compiler.
453    
454    4.  Improved JIT compiler optimizations for character ranges.
455    
456    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
457        Reported by Giuseppe D'Angelo.
458    
459    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
460        of a line, was being set incorrectly in cases where .* appeared inside
461        atomic brackets at the start of a pattern, or where there was a subsequent
462        *PRUNE or *SKIP.
463    
464    7.  Improved instruction cache flush for POWER/PowerPC.
465        Patch by Daniel Richard G.
466    
467    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
468        grep:
469    
470        (a) There is now no limit to the number of patterns to be matched.
471    
472        (b) An error is given if a pattern is too long.
473    
474        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
475            are now supported.
476    
477        (d) --exclude-from and --include-from (multiple use) have been added.
478    
479        (e) Exclusions and inclusions now apply to all files and directories, not
480            just to those obtained from scanning a directory recursively.
481    
482        (f) Multiple uses of -f and --file-list are now supported.
483    
484        (g) In a Windows environment, the default for -d has been changed from
485            "read" (the GNU grep default) to "skip", because otherwise the presence
486            of a directory in the file list provokes an error.
487    
488        (h) The documentation has been revised and clarified in places.
489    
490    9.  Improve the matching speed of capturing brackets.
491    
492    10. Changed the meaning of \X so that it now matches a Unicode extended
493        grapheme cluster.
494    
495    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
496        out POSIX threads when JIT support is configured.
497    
498    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
499    
500    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
501        structure could go wrong in environments where size_t is not the same size
502        as int.
503    
504    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
505        set.
506    
507    15. The EBCDIC support had decayed; later updates to the code had included
508        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
509        general tidy up of EBCDIC-related issues, and the documentation was also
510        not quite right. There is now a test that can be run on ASCII systems to
511        check some of the EBCDIC-related things (but is it not a full test).
512    
513    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
514        in a small tidy to the code.
515    
516    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
517    
518    18. If the --only-matching (-o) option in pcregrep is specified multiple
519        times, each one causes appropriate output. For example, -o1 -o2 outputs the
520        substrings matched by the 1st and 2nd capturing parentheses. A separating
521        string can be specified by --om-separator (default empty).
522    
523    19. Improving the first n character searches.
524    
525    20. Turn case lists for horizontal and vertical white space into macros so that
526        they are defined only once.
527    
528    21. This set of changes together give more compatible Unicode case-folding
529        behaviour for characters that have more than one other case when UCP
530        support is available.
531    
532        (a) The Unicode property table now has offsets into a new table of sets of
533            three or more characters that are case-equivalent. The MultiStage2.py
534            script that generates these tables (the pcre_ucd.c file) now scans
535            CaseFolding.txt instead of UnicodeData.txt for character case
536            information.
537    
538        (b) The code for adding characters or ranges of characters to a character
539            class has been abstracted into a generalized function that also handles
540            case-independence. In UTF-mode with UCP support, this uses the new data
541            to handle characters with more than one other case.
542    
543        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
544            whose other case is greater than 256 are now correctly matched
545            caselessly. Previously, the high codepoint matched the low one, but not
546            vice versa.
547    
548        (d) The processing of \h, \H, \v, and \ in character classes now makes use
549            of the new class addition function, using character lists defined as
550            macros alongside the case definitions of 20 above.
551    
552        (e) Caseless back references now work with characters that have more than
553            one other case.
554    
555        (f) General caseless matching of characters with more than one other case
556            is supported.
557    
558    22. Unicode character properties were updated from Unicode 6.2.0
559    
560    23. Improved CMake support under Windows. Patch by Daniel Richard G.
561    
562    24. Add support for 32-bit character strings, and UTF-32
563    
564    25. Major JIT compiler update (code refactoring and bugfixing).
565        Experimental Sparc 32 support is added.
566    
567    26. Applied a modified version of Daniel Richard G's patch to create
568        pcre.h.generic and config.h.generic by "make" instead of in the
569        PrepareRelease script.
570    
571    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
572        pcre_compile.c when checking for a zero character.
573    
574    28. Introducing a native interface for JIT. Through this interface, the compiled
575        machine code can be directly executed. The purpose of this interface is to
576        provide fast pattern matching, so several sanity checks are not performed.
577        However, feature tests are still performed. The new interface provides
578        1.4x speedup compared to the old one.
579    
580    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
581        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
582        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
583    
584    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
585        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
586        the "old" RFC 2279). Instead, it ended up passing a negative length to
587        pcre_exec().
588    
589    31. Add support for GCC's visibility feature to hide internal functions.
590    
591    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
592        "unknown -C option" after outputting 0 or 1.
593    
594    33. There is now support for generating a code coverage report for the test
595        suite in environments where gcc is the compiler and lcov is installed. This
596        is mainly for the benefit of the developers.
597    
598    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
599        unaddressable using valgrind annotations, allowing valgrind to detect
600        invalid memory accesses. This is mainly for the benefit of the developers.
601    
602    25. (*UTF) can now be used to start a pattern in any of the three libraries.
603    
604    26. Give configure error if --enable-cpp but no C++ compiler found.
605    
606    
607    Version 8.31 06-July-2012
608    -------------------------
609    
610    1.  Fixing a wrong JIT test case and some compiler warnings.
611    
612    2.  Removed a bashism from the RunTest script.
613    
614    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
615        to unsigned type, result still unsigned" that was given by an MS compiler
616        on encountering the code "-sizeof(xxx)".
617    
618    4.  Partial matching support is added to the JIT compiler.
619    
620    5.  Fixed several bugs concerned with partial matching of items that consist
621        of more than one character:
622    
623        (a) /^(..)\1/ did not partially match "aba" because checking references was
624            done on an "all or nothing" basis. This also applied to repeated
625            references.
626    
627        (b) \R did not give a hard partial match if \r was found at the end of the
628            subject.
629    
630        (c) \X did not give a hard partial match after matching one or more
631            characters at the end of the subject.
632    
633        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
634            a partial match for the string "\r".
635    
636        (e) When newline was set to CRLF, the metacharacter "." did not recognize
637            a partial match for a CR character at the end of the subject string.
638    
639    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
640        running pcretest, the text "(JIT)" added to the output whenever JIT is
641        actually used to run the match.
642    
643    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
644        or /S+[+] with a digit between 1 and 7.
645    
646    8.  OP_NOT now supports any UTF character not just single-byte ones.
647    
648    9.  (*MARK) control verb is now supported by the JIT compiler.
649    
650    10. The command "./RunTest list" lists the available tests without actually
651        running any of them. (Because I keep forgetting what they all are.)
652    
653    11. Add PCRE_INFO_MAXLOOKBEHIND.
654    
655    12. Applied a (slightly modified) user-supplied patch that improves performance
656        when the heap is used for recursion (compiled with --disable-stack-for-
657        recursion). Instead of malloc and free for each heap frame each time a
658        logical recursion happens, frames are retained on a chain and re-used where
659        possible. This sometimes gives as much as 30% improvement.
660    
661    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
662        call.
663    
664    14. As documented, (*COMMIT) is now confined to within a positive assertion.
665    
666    15. It is now possible to link pcretest with libedit as an alternative to
667        libreadline.
668    
669    16. (*COMMIT) control verb is now supported by the JIT compiler.
670    
671    17. The Unicode data tables have been updated to Unicode 6.1.0.
672    
673    18. Added --file-list option to pcregrep.
674    
675    19. Added binary file support to pcregrep, including the -a, --binary-files,
676        -I, and --text options.
677    
678    20. The madvise function is renamed for posix_madvise for QNX compatibility
679        reasons. Fixed by Giuseppe D'Angelo.
680    
681    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
682    
683    22. Changed the option for creating symbolic links for 16-bit man pages from
684        -s to -sf so that re-installing does not cause issues.
685    
686    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
687    
688    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
689        matches in certain environments (the workspace was not being correctly
690        retained). Also added to pcre_dfa_exec() a simple plausibility check on
691        some of the workspace data at the beginning of a restart.
692    
693    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
694        was not doing so when it should - probably a typo introduced by SVN 528
695        (change 8.10/14).
696    
697    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
698        \w+ when the character tables indicated that \x{c4} was a word character.
699        There were several related cases, all because the tests for doing a table
700        lookup were testing for characters less than 127 instead of 255.
701    
702    27. If a pattern contains capturing parentheses that are not used in a match,
703        their slots in the ovector are set to -1. For those that are higher than
704        any matched groups, this happens at the end of processing. In the case when
705        there were back references that the ovector was too small to contain
706        (causing temporary malloc'd memory to be used during matching), and the
707        highest capturing number was not used, memory off the end of the ovector
708        was incorrectly being set to -1. (It was using the size of the temporary
709        memory instead of the true size.)
710    
711    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
712        ovector size, it uses memory at the end of the block that it has got.
713    
714    29. Check for an overlong MARK name and give an error at compile time. The
715        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
716    
717    30. JIT compiler update.
718    
719    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
720        Rill for the patch.
721    
722    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
723    
724    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
725    
726    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
727        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
728        was enabled.
729    
730    35. Improve JIT code generation for greedy plus quantifier.
731    
732    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
733        1 to "aa" instead of to an empty string. The bug affected repeated groups
734        that could potentially match an empty string.
735    
736    37. Optimizing single character iterators in JIT.
737    
738    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
739        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
740        codepoints that are too big for the mode are faulted, and in a UTF mode,
741        disallowed codepoints are also faulted.
742    
743    39. If PCRE was compiled with UTF support, in three places in the DFA
744        matcher there was code that should only have been obeyed in UTF mode, but
745        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
746        processing when bytes with values greater than 127 were present. In 16-bit
747        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
748        both cases the values are those that cannot be the first data item in a UTF
749        character. The three items that might have provoked this were recursions,
750        possessively repeated groups, and atomic groups.
751    
752    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
753        and pcregrep, because some OS require shared objects to be explicitly
754        passed to ld, causing the link step to fail if they are not.
755    
756    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
757        mode, patterns that started with \h* or \R* might be incorrectly matched.
758    
759    
760    Version 8.30 04-February-2012
761    -----------------------------
762    
763    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
764        name is defined in ctype.h.
765    
766    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
767        only in quite long subpatterns.
768    
769    3.  Removed the function pcre_info(), which has been obsolete and deprecated
770        since it was replaced by pcre_fullinfo() in February 2000.
771    
772    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
773        match a (*MARK), and the match failed at the start of the subject, a
774        reference to memory before the start of the subject could occur. This bug
775        was introduced by fix 17 of release 8.21.
776    
777    5.  A reference to an unset group with zero minimum repetition was giving
778        totally wrong answers (in non-JavaScript-compatibility mode). For example,
779        /(another)?(\1?)test/ matched against "hello world test". This bug was
780        introduced in release 8.13.
781    
782    6.  Add support for 16-bit character strings (a large amount of work involving
783        many changes and refactorings).
784    
785    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
786        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
787        from a file.
788    
789    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
790        rounding is not applied in this particular case).
791    
792    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
793        if they appear, or are escaped, in patterns.
794    
795    10. Get rid of a number of -Wunused-but-set-variable warnings.
796    
797    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
798        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
799        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
800        also returns the mark "x". This bug applied to capturing parentheses,
801        non-capturing parentheses, and atomic parentheses. It also applied to some
802        assertions.
803    
804    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
805        information out of configure.ac instead of relying on pcre.h.generic, which
806        is not stored in the repository.
807    
808    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
809        -lreadline.
810    
811    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
812    
813    15. Applied Graycode's patch to put the top-level frame on the stack rather
814        than the heap when not using the stack for recursion. This gives a
815        performance improvement in many cases when recursion is not deep.
816    
817    16. Experimental code added to "pcretest -C" to output the stack frame size.
818    
819    
820    Version 8.21 12-Dec-2011
821    ------------------------
822    
823    1.  Updating the JIT compiler.
824    
825    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
826        are added as well.
827    
828    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
829        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
830        calling _pcre_jit_exec. Some extra comments are added.
831    
832    4.  (*MARK) settings inside atomic groups that do not contain any capturing
833        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
834        was introduced by change 18 for 8.20.
835    
836    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
837        ECMA-262 standard.
838    
839    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
840        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
841        This bug was probably introduced by change 9 of 8.13.
842    
843    7.  While fixing 6 above, I noticed that a number of other items were being
844        incorrectly rejected as "not fixed length". This arose partly because newer
845        opcodes had not been added to the fixed-length checking code. I have (a)
846        corrected the bug and added tests for these items, and (b) arranged for an
847        error to occur if an unknown opcode is encountered while checking for fixed
848        length instead of just assuming "not fixed length". The items that were
849        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
850        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
851        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
852    
853    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
854        being incorrectly compiled and would have given unpredicatble results.
855    
856    9.  A possessively repeated subpattern with minimum repeat count greater than
857        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
858        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
859        the first (A) could occur when it should not.
860    
861    10. Add a cast and remove a redundant test from the code.
862    
863    11. JIT should use pcre_malloc/pcre_free for allocation.
864    
865    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
866        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
867        is anything other than /usr, -L is still shown).
868    
869    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
870    
871    14. Perl does not support \N without a following name in a [] class; PCRE now
872        also gives an error.
873    
874    15. If a forward reference was repeated with an upper limit of around 2000,
875        it caused the error "internal error: overran compiling workspace". The
876        maximum number of forward references (including repeats) was limited by the
877        internal workspace, and dependent on the LINK_SIZE. The code has been
878        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
879        the default depends on LINK_SIZE. There is a new upper limit (for safety)
880        of around 200,000 forward references. While doing this, I also speeded up
881        the filling in of repeated forward references.
882    
883    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
884        incorrectly expecting the subject to contain another "a" after the start.
885    
886    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
887        in the match, the SKIP should be ignored. This was not happening; instead
888        the SKIP was being treated as NOMATCH. For patterns such as
889        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
890        tested.
891    
892    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
893        now much more compatible with Perl, in particular in cases where the result
894        is a non-match for a non-anchored pattern. For example, if
895        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
896        "m", where previously it did not return a name. A side effect of this
897        change is that for partial matches, the last encountered mark name is
898        returned, as for non matches. A number of tests that were previously not
899        Perl-compatible have been moved into the Perl-compatible test files. The
900        refactoring has had the pleasing side effect of removing one argument from
901        the match() function, thus reducing its stack requirements.
902    
903    19. If the /S+ option was used in pcretest to study a pattern using JIT,
904        subsequent uses of /S (without +) incorrectly behaved like /S+.
905    
906    21. Retrieve executable code size support for the JIT compiler and fixing
907        some warnings.
908    
909    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
910        not work when the shorter character appeared right at the end of the
911        subject string.
912    
913    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
914        systems.
915    
916    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
917        output it when the /M option is used in pcretest.
918    
919    25. The CheckMan script was not being included in the distribution. Also, added
920        an explicit "perl" to run Perl scripts from the PrepareRelease script
921        because this is reportedly needed in Windows.
922    
923    26. If study data was being save in a file and studying had not found a set of
924        "starts with" bytes for the pattern, the data written to the file (though
925        never used) was taken from uninitialized memory and so caused valgrind to
926        complain.
927    
928    27. Updated RunTest.bat as provided by Sheri Pierce.
929    
930    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
931    
932    29. Computation of memory usage for the table of capturing group names was
933        giving an unnecessarily large value.
934    
935    
936    Version 8.20 21-Oct-2011
937    ------------------------
938    
939    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
940        a POSIX class. After further experiments with Perl, which convinced me that
941        Perl has bugs and confusions, a closing square bracket is no longer allowed
942        in a POSIX name. This bug also affected patterns with classes that started
943        with full stops.
944    
945    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
946        captured substring, but while checking the failing first alternative,
947        substring 1 is temporarily captured. If the output vector supplied to
948        pcre_exec() was not big enough for this capture, the yield of the function
949        was still zero ("insufficient space for captured substrings"). This cannot
950        be totally fixed without adding another stack variable, which seems a lot
951        of expense for a edge case. However, I have improved the situation in cases
952        such as /(a)(b)x|abc/ matched against "abc", where the return code
953        indicates that fewer than the maximum number of slots in the ovector have
954        been set.
955    
956    3.  Related to (2) above: when there are more back references in a pattern than
957        slots in the output vector, pcre_exec() uses temporary memory during
958        matching, and copies in the captures as far as possible afterwards. It was
959        using the entire output vector, but this conflicts with the specification
960        that only 2/3 is used for passing back captured substrings. Now it uses
961        only the first 2/3, for compatibility. This is, of course, another edge
962        case.
963    
964    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
965        main code base, and can be used by building with --enable-jit. When this is
966        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
967        runtime --no-jit option is given.
968    
969    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
970        ovector, the return from the function was zero, implying that there were
971        other matches that did not fit. The correct "exactly full" value is now
972        returned.
973    
974    6.  If a subpattern that was called recursively or as a subroutine contained
975        (*PRUNE) or any other control that caused it to give a non-standard return,
976        invalid errors such as "Error -26 (nested recursion at the same subject
977        position)" or even infinite loops could occur.
978    
979    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
980        computing the minimum length on reaching *ACCEPT, and so ended up with the
981        wrong value of 1 rather than 0. Further investigation indicates that
982        computing a minimum subject length in the presence of *ACCEPT is difficult
983        (think back references, subroutine calls), and so I have changed the code
984        so that no minimum is registered for a pattern that contains *ACCEPT.
985    
986    8.  If (*THEN) was present in the first (true) branch of a conditional group,
987        it was not handled as intended. [But see 16 below.]
988    
989    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
990        Sheri Pierce.
991    
992    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
993        the first byte in a match must be "a".
994    
995    11. Change 17 for 8.13 increased the recursion depth for patterns like
996        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
997        pattern contains any instances of (*THEN). If it does not, the old
998        optimizations are restored. It would be nice to do this on a per-group
999        basis, but at the moment that is not feasible.
1000    
1001    12. In some environments, the output of pcretest -C is CRLF terminated. This
1002        broke RunTest's code that checks for the link size. A single white space
1003        character after the value is now allowed for.
1004    
1005    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1006        For "fr", it uses the Windows-specific input and output files.
1007    
1008    14. If (*THEN) appeared in a group that was called recursively or as a
1009        subroutine, it did not work as intended. [But see next item.]
1010    
1011    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1012        pattern fragments (but not containing any | characters). If A and B are
1013        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1014        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1015        D. In other words, Perl considers parentheses that do not contain any |
1016        characters to be part of a surrounding alternative, whereas PCRE was
1017        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1018        differently. PCRE now behaves in the same way as Perl, except in the case
1019        of subroutine/recursion calls such as (?1) which have in any case always
1020        been different (but PCRE had them first :-).
1021    
1022    16. Related to 15 above: Perl does not treat the | in a conditional group as
1023        creating alternatives. Such a group is treated in the same way as an
1024        ordinary group without any | characters when processing (*THEN). PCRE has
1025        been changed to match Perl's behaviour.
1026    
1027    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1028        RunGrepTest script failed.
1029    
1030    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1031        inevitable for groups that contain captures, but it can lead to a lot of
1032        stack use in large patterns. The old behaviour has been restored for atomic
1033        groups that do not contain any capturing parentheses.
1034    
1035    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1036        suppress the check for a minimum subject length at run time. (If it was
1037        given to pcre_exec() or pcre_dfa_exec() it did work.)
1038    
1039    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1040        fail to work when decoding hex characters in data strings in EBCDIC
1041        environments.
1042    
1043    21. It appears that in at least one Mac OS environment, the isxdigit() function
1044        is implemented as a macro that evaluates to its argument more than once,
1045        contravening the C 90 Standard (I haven't checked a later standard). There
1046        was an instance in pcretest which caused it to go wrong when processing
1047        \x{...} escapes in subject strings. The has been rewritten to avoid using
1048        things like p++ in the argument of isxdigit().
1049    
1050    
1051    Version 8.13 16-Aug-2011
1052    ------------------------
1053    
1054    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1055    
1056    2.  Two minor typos in pcre_internal.h have been fixed.
1057    
1058    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1059        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1060        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1061    
1062    4.  There were a number of related bugs in the code for matching backrefences
1063        caselessly in UTF-8 mode when codes for the characters concerned were
1064        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1065        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1066        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1067        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1068        2-byte code at the end of the subject (it thought there wasn't enough data
1069        left).
1070    
1071    5.  Comprehensive information about what went wrong is now returned by
1072        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1073        as the output vector has at least 2 elements. The offset of the start of
1074        the failing character and a reason code are placed in the vector.
1075    
1076    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1077        now returned is for the first byte of the failing character, instead of the
1078        last byte inspected. This is an incompatible change, but I hope it is small
1079        enough not to be a problem. It makes the returned offset consistent with
1080        pcre_exec() and pcre_dfa_exec().
1081    
1082    7.  pcretest now gives a text phrase as well as the error number when
1083        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1084        failure, the offset and reason code are output.
1085    
1086    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1087        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1088        back over a single character (\n). This seems wrong (because it treated the
1089        two characters as a single entity when going forwards), conflicts with the
1090        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1091        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1092        has been changed.
1093    
1094    9.  Some internal refactoring has changed the processing so that the handling
1095        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1096        time (the PCRE_DOTALL option was changed this way some time ago: version
1097        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1098        which was always a bit of a fudge. It also means that there is one less
1099        argument for the match() function, which reduces its stack requirements
1100        slightly. This change also fixes an incompatibility with Perl: the pattern
1101        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1102    
1103    10. More internal refactoring has drastically reduced the number of recursive
1104        calls to match() for possessively repeated groups such as (abc)++ when
1105        using pcre_exec().
1106    
1107    11. While implementing 10, a number of bugs in the handling of groups were
1108        discovered and fixed:
1109    
1110        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1111        (a|)*(?1) gave a compile-time internal error.
1112        ((a|)+)+  did not notice that the outer group could match an empty string.
1113        (^a|^)+   was not marked as anchored.
1114        (.*a|.*)+ was not marked as matching at start or after a newline.
1115    
1116    12. Yet more internal refactoring has removed another argument from the match()
1117        function. Special calls to this function are now indicated by setting a
1118        value in a variable in the "match data" data block.
1119    
1120    13. Be more explicit in pcre_study() instead of relying on "default" for
1121        opcodes that mean there is no starting character; this means that when new
1122        ones are added and accidentally left out of pcre_study(), testing should
1123        pick them up.
1124    
1125    14. The -s option of pcretest has been documented for ages as being an old
1126        synonym of -m (show memory usage). I have changed it to mean "force study
1127        for every regex", that is, assume /S for every regex. This is similar to -i
1128        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1129        using it. It makes it easier to run collections of tests with and without
1130        study enabled, and thereby test pcre_study() more easily. All the standard
1131        tests are now run with and without -s (but some patterns can be marked as
1132        "never study" - see 20 below).
1133    
1134    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1135        restoration of the capturing data to the outer values was not happening
1136        correctly.
1137    
1138    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1139        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1140        pattern had matched an empty string, and so incorrectly returned a no
1141        match.
1142    
1143    17. There was optimizing code for the last branch of non-capturing parentheses,
1144        and also for the obeyed branch of a conditional subexpression, which used
1145        tail recursion to cut down on stack usage. Unfortunately, now that there is
1146        the possibility of (*THEN) occurring in these branches, tail recursion is
1147        no longer possible because the return has to be checked for (*THEN). These
1148        two optimizations have therefore been removed. [But see 8.20/11 above.]
1149    
1150    18. If a pattern containing \R was studied, it was assumed that \R always
1151        matched two bytes, thus causing the minimum subject length to be
1152        incorrectly computed because \R can also match just one byte.
1153    
1154    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1155        was incorrectly computed.
1156    
1157    20. If /S is present twice on a test pattern in pcretest input, it now
1158        *disables* studying, thereby overriding the use of -s on the command line
1159        (see 14 above). This is necessary for one or two tests to keep the output
1160        identical in both cases.
1161    
1162    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1163        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1164    
1165    22. When an atomic group that contained a capturing parenthesis was
1166        successfully matched, but the branch in which it appeared failed, the
1167        capturing was not being forgotten if a higher numbered group was later
1168        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1169        group 1 to "a", when in fact it should be unset. This applied to multi-
1170        branched capturing and non-capturing groups, repeated or not, and also to
1171        positive assertions (capturing in negative assertions does not happen
1172        in PCRE) and also to nested atomic groups.
1173    
1174    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1175        subject after a captured substring, to make it easier to tell which of a
1176        number of identical substrings has been captured.
1177    
1178    24. The way atomic groups are processed by pcre_exec() has been changed so that
1179        if they are repeated, backtracking one repetition now resets captured
1180        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1181        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1182        "aaa". Previously, it would have been "a". As part of this code
1183        refactoring, the way recursive calls are handled has also been changed.
1184    
1185    25. If an assertion condition captured any substrings, they were not passed
1186        back unless some other capturing happened later. For example, if
1187        (?(?=(a))a) was matched against "a", no capturing was returned.
1188    
1189    26. When studying a pattern that contained subroutine calls or assertions,
1190        the code for finding the minimum length of a possible match was handling
1191        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1192        group 1 called group 2 while simultaneously a separate group 2 called group
1193        1). A stack overflow occurred in this case. I have fixed this by limiting
1194        the recursion depth to 10.
1195    
1196    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1197        Fortmann. This supports explicit test numbers on the command line, and has
1198        argument validation and error reporting.
1199    
1200    28. An instance of \X with an unlimited repeat could fail if at any point the
1201        first character it looked at was a mark character.
1202    
1203    29. Some minor code refactoring concerning Unicode properties and scripts
1204        should reduce the stack requirement of match() slightly.
1205    
1206    30. Added the '=' option to pcretest to check the setting of unused capturing
1207        slots at the end of the pattern, which are documented as being -1, but are
1208        not included in the return count.
1209    
1210    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1211        compiled something random. Now it gives a compile-time error (as does
1212        Perl).
1213    
1214    32. A *MARK encountered during the processing of a positive assertion is now
1215        recorded and passed back (compatible with Perl).
1216    
1217    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1218        had alternative anchored branches, the search for a second match in a line
1219        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1220        matched the line "0102" twice. The same bug affected patterns that started
1221        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1222        twice.
1223    
1224    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1225        does, and because of capturing effects, quantifying parenthesized
1226        assertions may at times be useful. Quantifiers are now allowed for
1227        parenthesized assertions.
1228    
1229    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1230    
1231    36. \g was being checked for fancy things in a character class, when it should
1232        just be a literal "g".
1233    
1234    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1235        appearance of a nested POSIX class supersedes an apparent external class.
1236        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1237        unescaped square brackets may also appear as part of class names. For
1238        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1239        more like Perl. (But see 8.20/1 above.)
1240    
1241    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1242        was because it thought it was \N{name}, which is not supported).
1243    
1244    39. Add minix to OS list not supporting the -S option in pcretest.
1245    
1246    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1247        cannot analyze patterns in sufficient detail to catch mutual recursions
1248        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1249        subgroup is called recursively as a subpattern for a second time at the
1250        same position in the subject string. In previous releases this might have
1251        been caught by the recursion limit, or it might have run out of stack.
1252    
1253    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1254        happen only once. PCRE was, however incorrectly giving a compile time error
1255        "recursive call could loop indefinitely" because it cannot analyze the
1256        pattern in sufficient detail. The compile time test no longer happens when
1257        PCRE is compiling a conditional subpattern, but actual runaway loops are
1258        now caught at runtime (see 40 above).
1259    
1260    42. It seems that Perl allows any characters other than a closing parenthesis
1261        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1262        has been changed to be the same.
1263    
1264    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1265        as not to get warnings when autogen.sh is called. Also changed
1266        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1267    
1268    44. To help people who use pcregrep to scan files containing exceedingly long
1269        lines, the following changes have been made:
1270    
1271        (a) The default value of the buffer size parameter has been increased from
1272            8K to 20K. (The actual buffer used is three times this size.)
1273    
1274        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1275            PCRE is built.
1276    
1277        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1278            to be set at run time.
1279    
1280        (d) Numerical values in pcregrep options can be followed by K or M, for
1281            example --buffer-size=50K.
1282    
1283        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1284            given and the return code is set to 2.
1285    
1286    45. Add a pointer to the latest mark to the callout data block.
1287    
1288    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1289        partial match of an empty string instead of no match. This was specific to
1290        the use of ".".
1291    
1292    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1293        complete match instead of a partial match. This bug was dependent on both
1294        the PCRE_UTF8 and PCRE_DOTALL options being set.
1295    
1296    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1297        starting byte set, because \b was not being ignored.
1298    
1299    
1300    Version 8.12 15-Jan-2011
1301    ------------------------
1302    
1303    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1304        checks for such things as part of the documentation building process.
1305    
1306    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1307        --match-limit and --recursion-limit options (added for 8.11). In
1308        particular, this made one of the standard tests fail. (The integer value
1309        went into the wrong half of a long int.)
1310    
1311    3.  If the --colour option was given to pcregrep with -v (invert match), it
1312        did strange things, either producing crazy output, or crashing. It should,
1313        of course, ignore a request for colour when reporting lines that do not
1314        match.
1315    
1316    4.  Another pcregrep bug caused similar problems if --colour was specified with
1317        -M (multiline) and the pattern match finished with a line ending.
1318    
1319    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1320        matched in multiline mode, the following line was shown as part of the
1321        match. This seems wrong, so I have changed it.
1322    
1323    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1324        the check for further matches in the same line (so they could be coloured)
1325        to overrun the end of the current line. If another match was found, it was
1326        incorrectly shown (and then shown again when found in the next line).
1327    
1328    7.  If pcregrep was compiled under Windows, there was a reference to the
1329        function pcregrep_exit() before it was defined. I am assuming this was
1330        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1331        reported by a user. I've moved the definition above the reference.
1332    
1333    
1334    Version 8.11 10-Dec-2010
1335    ------------------------
1336    
1337    1.  (*THEN) was not working properly if there were untried alternatives prior
1338        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1339        backtracked to try for "b" instead of moving to the next alternative branch
1340        at the same level (in this case, to look for "c"). The Perl documentation
1341        is clear that when (*THEN) is backtracked onto, it goes to the "next
1342        alternative in the innermost enclosing group".
1343    
1344    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1345        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1346        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1347        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1348        (*THEN).
1349    
1350    3.  If \s appeared in a character class, it removed the VT character from
1351        the class, even if it had been included by some previous item, for example
1352        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1353        of \s, but is part of the POSIX "space" class.)
1354    
1355    4.  A partial match never returns an empty string (because you can always
1356        match an empty string at the end of the subject); however the checking for
1357        an empty string was starting at the "start of match" point. This has been
1358        changed to the "earliest inspected character" point, because the returned
1359        data for a partial match starts at this character. This means that, for
1360        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1361        (previously it gave "no match").
1362    
1363    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1364        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1365        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1366        has an implication that the given string is incomplete (because a partial
1367        match is preferred over a full match). For this reason, these items now
1368        give a partial match in this situation. [Aside: previously, the one case
1369        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1370        match rather than a full match, which was wrong by the old rules, but is
1371        now correct.]
1372    
1373    6.  There was a bug in the handling of #-introduced comments, recognized when
1374        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1375        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1376        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1377        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1378        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1379        places in pcre_compile().
1380    
1381    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1382        comments when looking ahead for named forward references to subpatterns,
1383        the only newline sequence it recognized was NL. It now handles newlines
1384        according to the set newline convention.
1385    
1386    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1387        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1388        cater for a lack of strerror(). These oversights have been fixed.
1389    
1390    9.  Added --match-limit and --recursion-limit to pcregrep.
1391    
1392    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1393    
1394    11. When the -o option was used, pcregrep was setting a return code of 1, even
1395        when matches were found, and --line-buffered was not being honoured.
1396    
1397    12. Added an optional parentheses number to the -o and --only-matching options
1398        of pcregrep.
1399    
1400    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1401        can match an empty string. The code to do it in pcretest and pcredemo
1402        needed fixing:
1403    
1404        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1405            only one byte after an empty string match just before CRLF (this case
1406            just got forgotten; "any" and "anycrlf" were OK).
1407    
1408        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1409            mode when an empty string match preceded an ASCII character followed by
1410            a non-ASCII character. (The code for advancing by one character rather
1411            than one byte was nonsense.)
1412    
1413        (c) The pcredemo.c sample program did not have any code at all to handle
1414            the cases when CRLF is a valid newline sequence.
1415    
1416    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1417        as a starting offset was within the subject string. There is now a new
1418        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1419        negative or greater than the length of the string. In order to test this,
1420        pcretest is extended to allow the setting of negative starting offsets.
1421    
1422    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1423        starting offset points to the beginning of a UTF-8 character was
1424        unnecessarily clumsy. I tidied it up.
1425    
1426    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1427        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1428    
1429    17. Nobody had reported that the --include_dir option, which was added in
1430        release 7.7 should have been called --include-dir (hyphen, not underscore)
1431        for compatibility with GNU grep. I have changed it to --include-dir, but
1432        left --include_dir as an undocumented synonym, and the same for
1433        --exclude-dir, though that is not available in GNU grep, at least as of
1434        release 2.5.4.
1435    
1436    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1437        characters from a string of bytes) have been redefined so as not to use
1438        loops, in order to improve performance in some environments. At the same
1439        time, I abstracted some of the common code into auxiliary macros to save
1440        repetition (this should not affect the compiled code).
1441    
1442    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1443        compile-time error is now given if \c is not followed by an ASCII
1444        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1445        different, and any byte value is allowed.)
1446    
1447    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1448        START_OPTIMIZE option, which is now allowed at compile time - but just
1449        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1450        to pcregrep and other applications that have no direct access to PCRE
1451        options. The new /Y option in pcretest sets this option when calling
1452        pcre_compile().
1453    
1454    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1455        back references. Groups containing recursive back references were forced to
1456        be atomic by that change, but in the case of named groups, the amount of
1457        memory required was incorrectly computed, leading to "Failed: internal
1458        error: code overflow". This has been fixed.
1459    
1460    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1461        pcretest.c, to avoid build problems in some Borland environments.
1462    
1463    
1464    Version 8.10 25-Jun-2010
1465    ------------------------
1466    
1467    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1468        THEN.
1469    
1470    2.  (*ACCEPT) was not working when inside an atomic group.
1471    
1472    3.  Inside a character class, \B is treated as a literal by default, but
1473        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1474        causes the error). The code is unchanged, but I tidied the documentation.
1475    
1476    4.  Inside a character class, PCRE always treated \R and \X as literals,
1477        whereas Perl faults them if its -w option is set. I have changed PCRE so
1478        that it faults them when PCRE_EXTRA is set.
1479    
1480    5.  Added support for \N, which always matches any character other than
1481        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1482    
1483    6.  When compiling pcregrep with newer versions of gcc which may have
1484        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1485        declared with attribute warn_unused_result" were given. Just casting the
1486        result to (void) does not stop the warnings; a more elaborate fudge is
1487        needed. I've used a macro to implement this.
1488    
1489    7.  Minor change to pcretest.c to avoid a compiler warning.
1490    
1491    8.  Added four artifical Unicode properties to help with an option to make
1492        \s etc use properties (see next item). The new properties are: Xan
1493        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1494    
1495    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1496        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1497        this option. Modified pcretest to add /W to test this facility. Added
1498        REG_UCP to make it available via the POSIX interface.
1499    
1500    10. Added --line-buffered to pcregrep.
1501    
1502    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1503        studied, and the match started with a letter with a code point greater than
1504        127 whose first byte was different to the first byte of the other case of
1505        the letter, the other case of this starting letter was not recognized
1506        (#976).
1507    
1508    12. If a pattern that was studied started with a repeated Unicode property
1509        test, for example, \p{Nd}+, there was the theoretical possibility of
1510        setting up an incorrect bitmap of starting bytes, but fortunately it could
1511        not have actually happened in practice until change 8 above was made (it
1512        added property types that matched character-matching opcodes).
1513    
1514    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1515        possible starting bytes for non-anchored patterns.
1516    
1517    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1518        \R, and also a number of cases that involve Unicode properties, both
1519        explicit and implicit when PCRE_UCP is set.
1520    
1521    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1522        input, it could crash or give wrong results if characters with values
1523        greater than 0xc0 were present in the subject string. (Detail: it assumed
1524        UTF-8 input when processing these items.)
1525    
1526    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1527        size_t is 64-bit (#991).
1528    
1529    17. Added a check for running out of memory when PCRE is compiled with
1530        --disable-stack-for-recursion (#990).
1531    
1532    18. If the last data line in a file for pcretest does not have a newline on
1533        the end, a newline was missing in the output.
1534    
1535    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1536        less than 128) in its various bitmaps. However, there is a facility for
1537        generating tables according to the current locale when PCRE is compiled. It
1538        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1539        characters, are recognized by isspace() and therefore were getting set in
1540        these tables, and indeed these tables seem to approximate to ISO 8859. This
1541        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1542        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1543        which of course cannot start UTF-8 characters. I have changed the code so
1544        that only real ASCII characters (less than 128) and the correct starting
1545        bytes for UTF-8 encodings are set for characters greater than 127 when in
1546        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1547        altogether.)
1548    
1549    20. Added the /T option to pcretest so as to be able to run tests with non-
1550        standard character tables, thus making it possible to include the tests
1551        used for 19 above in the standard set of tests.
1552    
1553    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1554        reference to a subpattern the other side of a comment that contains an
1555        opening parenthesis caused either an internal compiling error, or a
1556        reference to the wrong subpattern.
1557    
1558    
1559    Version 8.02 19-Mar-2010
1560    ------------------------
1561    
1562    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1563    
1564    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1565        configured.
1566    
1567    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1568        original author of that file, following a query about its status.
1569    
1570    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1571        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1572    
1573    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1574        quantifier applied to a forward-referencing subroutine call, could compile
1575        incorrect code or give the error "internal error: previously-checked
1576        referenced subpattern not found".
1577    
1578    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1579        variables to point to external functions. For these systems, therefore,
1580        pcre_malloc etc. are now initialized to local functions that call the
1581        relevant global functions.
1582    
1583    7.  There were two entries missing in the vectors called coptable and poptable
1584        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1585        I've fixed the data, and added a kludgy way of testing at compile time that
1586        the lengths are correct (equal to the number of opcodes).
1587    
1588    8.  Following on from 7, I added a similar kludge to check the length of the
1589        eint vector in pcreposix.c.
1590    
1591    9.  Error texts for pcre_compile() are held as one long string to avoid too
1592        much relocation at load time. To find a text, the string is searched,
1593        counting zeros. There was no check for running off the end of the string,
1594        which could happen if a new error number was added without updating the
1595        string.
1596    
1597    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1598    
1599    11. \K was not working if it appeared in an atomic group or in a group that
1600        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1601        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1602        the assertion is positive, but not if it is negative.
1603    
1604    12. Change 11 fortuitously reduced the size of the stack frame used in the
1605        "match()" function of pcre_exec.c by one pointer. Forthcoming
1606        implementation of support for (*MARK) will need an extra pointer on the
1607        stack; I have reserved it now, so that the stack frame size does not
1608        decrease.
1609    
1610    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1611        item in branch that calls a recursion is a subroutine call - as in the
1612        second branch in the above example - was incorrectly given the compile-
1613        time error "recursive call could loop indefinitely" because pcre_compile()
1614        was not correctly checking the subroutine for matching a non-empty string.
1615    
1616    14. The checks for overrunning compiling workspace could trigger after an
1617        overrun had occurred. This is a "should never occur" error, but it can be
1618        triggered by pathological patterns such as hundreds of nested parentheses.
1619        The checks now trigger 100 bytes before the end of the workspace.
1620    
1621    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1622    
1623    
1624    Version 8.01 19-Jan-2010
1625    ------------------------
1626    
1627    1.  If a pattern contained a conditional subpattern with only one branch (in
1628        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1629        computed the wrong minimum data length (which is of course zero for such
1630        subpatterns). This could cause incorrect "no match" results.
1631    
1632    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1633        the pattern is reset in the first branch, pcre_compile() failed with
1634        "internal error: code overflow at offset...". This happened only when
1635        the reset was to the original external option setting. (An optimization
1636        abstracts leading options settings into an external setting, which was the
1637        cause of this.)
1638    
1639    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1640        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1641        assertion pattern did not match (meaning that the assertion was true), it
1642        was incorrectly treated as false if the SKIP had been reached during the
1643        matching. This also applied to assertions used as conditions.
1644    
1645    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1646        assertion subpattern, including such a pattern used as a condition,
1647        unpredictable results occurred, instead of the error return
1648        PCRE_ERROR_DFA_UITEM.
1649    
1650    5.  The C++ GlobalReplace function was not working like Perl for the special
1651        situation when an empty string is matched. It now does the fancy magic
1652        stuff that is necessary.
1653    
1654    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1655        removed. (These were left over from very, very early versions of PCRE.)
1656    
1657    7.  Some cosmetic changes to the code to make life easier when compiling it
1658        as part of something else:
1659    
1660        (a) Change DEBUG to PCRE_DEBUG.
1661    
1662        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1663            called "current" as "current_branch", to prevent a collision with the
1664            Linux macro when compiled as a kernel module.
1665    
1666        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1667            prevent a collision with the Linux macro when compiled as a kernel
1668            module.
1669    
1670    8.  In pcre_compile() there are some checks for integer overflows that used to
1671        cast potentially large values to (double). This has been changed to that
1672        when building, a check for int64_t is made, and if it is found, it is used
1673        instead, thus avoiding the use of floating point arithmetic. (There is no
1674        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1675        double.
1676    
1677    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1678        2005 (difference between two addresses compared to an unsigned value).
1679    
1680    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1681        custom one, because of the following reported problem in Windows:
1682    
1683          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1684              under Win32.
1685          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1686              therefore missing the function definition.
1687          - The compiler thus generates a "C" signature for the test function.
1688          - The linker fails to find the "C" function.
1689          - PCRE fails to configure if asked to do so against libbz2.
1690    
1691    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1692        messages were output:
1693    
1694          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1695          rerunning libtoolize, to keep the correct libtool macros in-tree.
1696          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1697    
1698        I have done both of these things.
1699    
1700    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1701        most of the time, it *can* run out if it is given a pattern that contains a
1702        runaway infinite recursion. I updated the discussion in the pcrestack man
1703        page.
1704    
1705    13. Now that we have gone to the x.xx style of version numbers, the minor
1706        version may start with zero. Using 08 or 09 is a bad idea because users
1707        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1708        interpreted as invalid octal numbers. I've updated the previous comment in
1709        configure.ac, and also added a check that gives an error if 08 or 09 are
1710        used.
1711    
1712    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1713        causing partial matching to fail when the end of the subject matched \W
1714        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1715    
1716    15. There were some discrepancies between the declarations in pcre_internal.h
1717        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1718        their definitions. The declarations used "const uschar *" and the
1719        definitions used USPTR. Even though USPTR is normally defined as "const
1720        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1721        reported that: "This difference in casting confuses some C++ compilers, for
1722        example, SunCC recognizes above declarations as different functions and
1723        generates broken code for hbpcre." I have changed the declarations to use
1724        USPTR.
1725    
1726    16. GNU libtool is named differently on some systems. The autogen.sh script now
1727        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1728        (FreeBSD).
1729    
1730    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1731        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1732        comment: "Figure out how to create a longlong from a string: strtoll and
1733        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1734        instance, but it only takes 2 args instead of 3!"
1735    
1736    18. A subtle bug concerned with back references has been fixed by a change of
1737        specification, with a corresponding code fix. A pattern such as
1738        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1739        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1740        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1741        same bug. Such groups have to be quantified to be useful, or contained
1742        inside another quantified group. (If there's no repetition, the reference
1743        can never match.) The problem arises because, having left the group and
1744        moved on to the rest of the pattern, a later failure that backtracks into
1745        the group uses the captured value from the final iteration of the group
1746        rather than the correct earlier one. I have fixed this in PCRE by forcing
1747        any group that contains a reference to itself to be an atomic group; that
1748        is, there cannot be any backtracking into it once it has completed. This is
1749        similar to recursive and subroutine calls.
1750    
1751    
1752    Version 8.00 19-Oct-09
1753    ----------------------
1754    
1755    1.  The table for translating pcre_compile() error codes into POSIX error codes
1756        was out-of-date, and there was no check on the pcre_compile() error code
1757        being within the table. This could lead to an OK return being given in
1758        error.
1759    
1760    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1761        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1762        in a Windows environment.
1763    
1764    3.  The pcregrep --count option prints the count for each file even when it is
1765        zero, as does GNU grep. However, pcregrep was also printing all files when
1766        --files-with-matches was added. Now, when both options are given, it prints
1767        counts only for those files that have at least one match. (GNU grep just
1768        prints the file name in this circumstance, but including the count seems
1769        more useful - otherwise, why use --count?) Also ensured that the
1770        combination -clh just lists non-zero counts, with no names.
1771    
1772    4.  The long form of the pcregrep -F option was incorrectly implemented as
1773        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1774        but it seems right to fix it, and I didn't think it was worth preserving
1775        the old behaviour.
1776    
1777    5.  The command line items --regex=pattern and --regexp=pattern were not
1778        recognized by pcregrep, which required --regex pattern or --regexp pattern
1779        (with a space rather than an '='). The man page documented the '=' forms,
1780        which are compatible with GNU grep; these now work.
1781    
1782    6.  No libpcreposix.pc file was created for pkg-config; there was just
1783        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1784    
1785    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1786        when UCP support is not needed, by modifying the Python script that
1787        generates it from Unicode data files. This should not matter if the module
1788        is correctly used as a library, but I received one complaint about 50K of
1789        unwanted data. My guess is that the person linked everything into his
1790        program rather than using a library. Anyway, it does no harm.
1791    
1792    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1793        was a minimum greater than 1 for a wide character in a possessive
1794        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1795        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1796        character. Chaos in the form of incorrect output or a compiling loop could
1797        result.
1798    
1799    9.  The restrictions on what a pattern can contain when partial matching is
1800        requested for pcre_exec() have been removed. All patterns can now be
1801        partially matched by this function. In addition, if there are at least two
1802        slots in the offset vector, the offset of the earliest inspected character
1803        for the match and the offset of the end of the subject are set in them when
1804        PCRE_ERROR_PARTIAL is returned.
1805    
1806    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1807        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1808        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1809        and may be more useful for multi-segment matching.
1810    
1811    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1812        used to be given if ever the end of the subject was reached; now it is
1813        given only if matching could not proceed because another character was
1814        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1815        string "Z", which now yields "no match" instead of "partial match". In the
1816        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1817        final character ended with (*FAIL).
1818    
1819    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1820        if the pattern had a "must contain" character that was already found in the
1821        earlier partial match, unless partial matching was again requested. For
1822        example, with the pattern /dog.(body)?/, the "must contain" character is
1823        "g". If the first part-match was for the string "dog", restarting with
1824        "sbody" failed. This bug has been fixed.
1825    
1826    13. The string returned by pcre_dfa_exec() after a partial match has been
1827        changed so that it starts at the first inspected character rather than the
1828        first character of the match. This makes a difference only if the pattern
1829        starts with a lookbehind assertion or \b or \B (\K is not supported by
1830        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1831        matching functions compatible, and I think it's the right thing to do.
1832    
1833    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1834        so that the demonstration program is easily available in environments where
1835        PCRE has not been installed from source.
1836    
1837    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1838        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1839        library.
1840    
1841    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1842        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1843        is not the first non-POSIX option to be added. Clearly some people find
1844        these options useful.
1845    
1846    17. If a caller to the POSIX matching function regexec() passes a non-zero
1847        value for nmatch with a NULL value for pmatch, the value of
1848        nmatch is forced to zero.
1849    
1850    18. RunGrepTest did not have a test for the availability of the -u option of
1851        the diff command, as RunTest does. It now checks in the same way as
1852        RunTest, and also checks for the -b option.
1853    
1854    19. If an odd number of negated classes containing just a single character
1855        interposed, within parentheses, between a forward reference to a named
1856        subpattern and the definition of the subpattern, compilation crashed with
1857        an internal error, complaining that it could not find the referenced
1858        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1859        [The bug was that it was starting one character too far in when skipping
1860        over the character class, thus treating the ] as data rather than
1861        terminating the class. This meant it could skip too much.]
1862    
1863    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1864        /g option in pcretest when the pattern contains \K, which makes it possible
1865        to have an empty string match not at the start, even when the pattern is
1866        anchored. Updated pcretest and pcredemo to use this option.
1867    
1868    21. If the maximum number of capturing subpatterns in a recursion was greater
1869        than the maximum at the outer level, the higher number was returned, but
1870        with unset values at the outer level. The correct (outer level) value is
1871        now given.
1872    
1873    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1874        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1875        make it do so. The string so far is captured, making this feature
1876        compatible with Perl.
1877    
1878    23. The tests have been re-organized, adding tests 11 and 12, to make it
1879        possible to check the Perl 5.10 features against Perl 5.10.
1880    
1881    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1882        pattern matches a fixed length string. PCRE did not allow this; now it
1883        does. Neither allows recursion.
1884    
1885    25. I finally figured out how to implement a request to provide the minimum
1886        length of subject string that was needed in order to match a given pattern.
1887        (It was back references and recursion that I had previously got hung up
1888        on.) This code has now been added to pcre_study(); it finds a lower bound
1889        to the length of subject needed. It is not necessarily the greatest lower
1890        bound, but using it to avoid searching strings that are too short does give
1891        some useful speed-ups. The value is available to calling programs via
1892        pcre_fullinfo().
1893    
1894    26. While implementing 25, I discovered to my embarrassment that pcretest had
1895        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1896        study optimizations had never been tested with that matching function.
1897        Oops. What is worse, even when it was passed study data, there was a bug in
1898        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1899        were also very few tests of studied patterns with pcre_dfa_exec().
1900    
1901    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1902        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1903        on the other side of the coin, they are no longer allowed to have different
1904        names, because these cannot be distinguished in PCRE, and this has caused
1905        confusion. (This is a difference from Perl.)
1906    
1907    28. When duplicate subpattern names are present (necessarily with different
1908        numbers, as required by 27 above), and a test is made by name in a
1909        conditional pattern, either for a subpattern having been matched, or for
1910        recursion in such a pattern, all the associated numbered subpatterns are
1911        tested, and the overall condition is true if the condition is true for any
1912        one of them. This is the way Perl works, and is also more like the way
1913        testing by number works.
1914    
1915    
1916    Version 7.9 11-Apr-09
1917    ---------------------
1918    
1919    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1920        (pcretest), all targets were linked against these libraries. This included
1921        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1922        libraries. This caused unwanted dependencies to be created. This problem
1923        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1924        pcretest is linked with readline.
1925    
1926    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1927        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1928        moved outside it again, because FALSE and TRUE are already defined in AIX,
1929        but BOOL is not.
1930    
1931    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1932        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1933    
1934    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1935        hyphens) following file names and line numbers when outputting matching
1936        lines. This is not true; no spaces are inserted. I have also clarified the
1937        wording for the --colour (or --color) option.
1938    
1939    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1940        was not coloured; this is different to GNU grep, so I have changed it to be
1941        the same.
1942    
1943    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1944        each matching line was coloured. Now it goes on to look for further matches
1945        of any of the test patterns, which is the same behaviour as GNU grep.
1946    
1947    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1948        doesn't make sense to accept an empty string match in pcregrep, so I have
1949        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1950        seems to be how GNU grep behaves. [But see later change 40 for release
1951        8.33.]
1952    
1953    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1954        start or after a newline", because the conditional assertion was not being
1955        correctly handled. The rule now is that both the assertion and what follows
1956        in the first alternative must satisfy the test.
1957    
1958    9.  If auto-callout was enabled in a pattern with a conditional group whose
1959        condition was an assertion, PCRE could crash during matching, both with
1960        pcre_exec() and pcre_dfa_exec().
1961    
1962    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1963        used for matching.
1964    
1965    11. Unicode property support in character classes was not working for
1966        characters (bytes) greater than 127 when not in UTF-8 mode.
1967    
1968    12. Added the -M command line option to pcretest.
1969    
1970    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1971    
1972    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1973    
1974    16. Added comments and documentation about mis-use of no_arg in the C++
1975        wrapper.
1976    
1977    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1978        from Martin Jerabek that uses macro names for all relevant character and
1979        string constants.
1980    
1981    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1982        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1983        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1984        these, but not everybody uses configure.
1985    
1986    19. A conditional group that had only one branch was not being correctly
1987        recognized as an item that could match an empty string. This meant that an
1988        enclosing group might also not be so recognized, causing infinite looping
1989        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1990        with the subject "ab", where knowledge that the repeated group can match
1991        nothing is needed in order to break the loop.
1992    
1993    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1994        exec(), but without supplying a callout function, matching went wrong.
1995    
1996    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1997        leak if the size of the offset vector was greater than 30. When the vector
1998        is smaller, the saved offsets during recursion go onto a local stack
1999        vector, but for larger vectors malloc() is used. It was failing to free
2000        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2001        error, in fact).
2002    
2003    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2004        heapframe that is used only when UTF-8 support is enabled. This caused no
2005        problem, but was untidy.
2006    
2007    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2008        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2009        included within another project.
2010    
2011    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2012        slightly modified by me:
2013    
2014          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2015              not building pcregrep.
2016    
2017          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2018              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2019    
2020    25. Forward references, both numeric and by name, in patterns that made use of
2021        duplicate group numbers, could behave incorrectly or give incorrect errors,
2022        because when scanning forward to find the reference group, PCRE was not
2023        taking into account the duplicate group numbers. A pattern such as
2024        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2025    
2026    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2027        the feature of a custom pointer more persuasive (as requested by a user).
2028    
2029    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2030        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2031        pre-defined.
2032    
2033    28. Added support for (*UTF8) at the start of a pattern.
2034    
2035    29. Arrange for flags added by the "release type" setting in CMake to be shown
2036        in the configuration summary.
2037    
2038    
2039    Version 7.8 05-Sep-08
2040    ---------------------
2041    
2042    1.  Replaced UCP searching code with optimized version as implemented for Ad
2043        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2044        stage table and inline lookup instead of a function, giving speed ups of 2
2045        to 5 times on some simple patterns that I tested. Permission was given to
2046        distribute the MultiStage2.py script that generates the tables (it's not in
2047        the tarball, but is in the Subversion repository).
2048    
2049    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2050        scripts.
2051    
2052    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2053        a group with a zero qualifier. The result of the study could be incorrect,
2054        or the function might crash, depending on the pattern.
2055    
2056    4.  Caseless matching was not working for non-ASCII characters in back
2057        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2058        It now works when Unicode Property Support is available.
2059    
2060    5.  In pcretest, an escape such as \x{de} in the data was always generating
2061        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2062        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2063        truncation.
2064    
2065    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2066    
2067    7.  Added two (int) casts to pcregrep when printing the difference of two
2068        pointers, in case they are 64-bit values.
2069    
2070    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2071        test 2 if it fails.
2072    
2073    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2074        and a #define of that name to empty if it is not externally set. This is to
2075        allow users of MSVC to set it if necessary.
2076    
2077    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2078        the convenience functions in the pcre_get.c source file.
2079    
2080    11. An option change at the start of a pattern that had top-level alternatives
2081        could cause overwriting and/or a crash. This command provoked a crash in
2082        some environments:
2083    
2084          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2085    
2086        This potential security problem was recorded as CVE-2008-2371.
2087    
2088    12. For a pattern where the match had to start at the beginning or immediately
2089        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2090        pcre_dfa_exec() could read past the end of the passed subject if there was
2091        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2092        pcretest so that it places the subject at the end of its malloc-ed buffer.
2093    
2094    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2095        exec() might read past the end of the data buffer in UTF-8 mode.
2096    
2097    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2098        the data contained the byte 0x85 as part of a UTF-8 character within its
2099        first line. This applied both to normal and DFA matching.
2100    
2101    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2102        /^[^d]*?$/8 failed to match "abc".
2103    
2104    16. Added a missing copyright notice to pcrecpp_internal.h.
2105    
2106    17. Make it more clear in the documentation that values returned from
2107        pcre_exec() in ovector are byte offsets, not character counts.
2108    
2109    18. Tidied a few places to stop certain compilers from issuing warnings.
2110    
2111    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2112        supplied by Stefan Weber. I made a further small update for 7.8 because
2113        there is a change of source arrangements: the pcre_searchfuncs.c module is
2114        replaced by pcre_ucd.c.
2115    
2116    
2117    Version 7.7 07-May-08
2118    ---------------------
2119    
2120    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2121        a string to a long long, pretend we don't even have a long long." This is
2122        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2123    
2124    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2125        pre-7.6 versions, which defined a global no_arg variable instead of putting
2126        it in the RE class. (See also #8 below.)
2127    
2128    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2129        Lopes.
2130    
2131    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2132    
2133        (1) The include/exclude patterns were being applied to the whole pathnames
2134            of files, instead of just to the final components.
2135    
2136        (2) If there was more than one level of directory, the subdirectories were
2137            skipped unless they satisfied the include/exclude conditions. This is
2138            inconsistent with GNU grep (and could even be seen as contrary to the
2139            pcregrep specification - which I improved to make it absolutely clear).
2140            The action now is always to scan all levels of directory, and just
2141            apply the include/exclude patterns to regular files.
2142    
2143    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2144        --exclude_dir in the tests to avoid scanning .svn directories.
2145    
2146    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2147        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2148        doesn't support NULs in patterns.
2149    
2150    7.  Added some missing "const"s to declarations of static tables in
2151        pcre_compile.c and pcre_dfa_exec.c.
2152    
2153    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2154        caused by fix #2  above. (Subsequently also a second patch to fix the
2155        first patch. And a third patch - this was a messy problem.)
2156    
2157    9.  Applied Craig's patch to remove the use of push_back().
2158    
2159    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2160        matching function regexec().
2161    
2162    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2163        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2164        references. PCRE supports relative numbers with this syntax (I don't think
2165        Oniguruma does).
2166    
2167    12. Previously, a group with a zero repeat such as (...){0} was completely
2168        omitted from the compiled regex. However, this means that if the group
2169        was called as a subroutine from elsewhere in the pattern, things went wrong
2170        (an internal error was given). Such groups are now left in the compiled
2171        pattern, with a new opcode that causes them to be skipped at execution
2172        time.
2173    
2174    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2175        to the way PCRE behaves:
2176    
2177        (a) A lone ] character is dis-allowed (Perl treats it as data).
2178    
2179        (b) A back reference to an unmatched subpattern matches an empty string
2180            (Perl fails the current match path).
2181    
2182        (c) A data ] in a character class must be notated as \] because if the
2183            first data character in a class is ], it defines an empty class. (In
2184            Perl it is not possible to have an empty class.) The empty class []
2185            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2186            The negative empty class [^] matches any one character, independently
2187            of the DOTALL setting.
2188    
2189    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2190        non-existent subpattern following a character class starting with ']' and
2191        containing () gave an internal compiling error instead of "reference to
2192        non-existent subpattern". Fortunately, when the pattern did exist, the
2193        compiled code was correct. (When scanning forwards to check for the
2194        existence of the subpattern, it was treating the data ']' as terminating
2195        the class, so got the count wrong. When actually compiling, the reference
2196        was subsequently set up correctly.)
2197    
2198    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2199        it was being rejected as not supported by pcre_dfa_exec(), even though
2200        other assertions are supported. I have made pcre_dfa_exec() support
2201        (*FAIL).
2202    
2203    16. The implementation of 13c above involved the invention of a new opcode,
2204        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2205        cannot be changed at match time, I realized I could make a small
2206        improvement to matching performance by compiling OP_ALLANY instead of
2207        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2208        on the OP_ANY path.
2209    
2210    17. Compiling pcretest on Windows with readline support failed without the
2211        following two fixes: (1) Make the unistd.h include conditional on
2212        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2213    
2214    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2215        ncurses library to be included for pcretest when ReadLine support is
2216        requested, but also to allow for it to be overridden. This patch came from
2217        Daniel Bergström.
2218    
2219    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2220        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2221        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2222        spotting this.
2223    
2224    
2225    Version 7.6 28-Jan-08
2226    ---------------------
2227    
2228    1.  A character class containing a very large number of characters with
2229        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2230        overflow.
2231    
2232    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2233        HAVE_LONG_LONG is not defined.
2234    
2235    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2236        bring them up to date and include new features. This patch includes:
2237    
2238        - Fixed PH's badly added libz and libbz2 support.
2239        - Fixed a problem with static linking.
2240        - Added pcredemo. [But later removed - see 7 below.]
2241        - Fixed dftables problem and added an option.
2242        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2243            HAVE_LONG_LONG.
2244        - Added readline support for pcretest.
2245        - Added an listing of the option settings after cmake has run.
2246    
2247    4.  A user submitted a patch to Makefile that makes it easy to create
2248        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2249        Makefile.am that cause it to include this special target, without
2250        affecting anything else. Note that the same mingw target plus all
2251        the other distribution libraries and programs are now supported
2252        when configuring with CMake (see 6 below) instead of with
2253        Configure/Make.
2254    
2255    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2256        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2257        exported in the Windows port". It has not yet been confirmed that the patch
2258        solves the problem, but it does no harm.
2259    
2260    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2261        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2262        with CMake, and also correct the comment about stack recursion.
2263    
2264    7.  Remove the automatic building of pcredemo from the ./configure system and
2265        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2266        of a program that users should build themselves after PCRE is installed, so
2267        building it automatically is not really right. What is more, it gave
2268        trouble in some build environments.
2269    
2270    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2271    
2272    
2273    Version 7.5 10-Jan-08
2274    ---------------------
2275    
2276    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2277        values in parens when parsing an RE using the C++ wrapper."
2278    
2279    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2280        Characters greater than 255 were excluded from the class instead of being
2281        included.
2282    
2283    3.  The same bug as (2) above applied to negated POSIX classes such as
2284        [:^space:].
2285    
2286    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2287        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2288        I have changed it.
2289    
2290    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2291        first named subpattern) and a construct such as (?&a) would reference the
2292        first named subpattern whose name started with "a" (in other words, the
2293        length check was missing). Both these problems are fixed. "Subpattern name
2294        expected" is now given for (?&) (a zero-length name), and this patch also
2295        makes it give the same error for \k'' (previously it complained that that
2296        was a reference to a non-existent subpattern).
2297    
2298    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2299        this is right because (?- can be followed by option settings as well as by
2300        digits. I have, however, made the messages clearer.
2301    
2302    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2303        than the number used in the conditional) now cause a compile-time error.
2304        This is actually not compatible with Perl, which accepts such patterns, but
2305        treats the conditional as always being FALSE (as PCRE used to), but it
2306        seems to me that giving a diagnostic is better.
2307    
2308    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2309        and messages.
2310    
2311    9.  Fix two occurrences of "backslash" in comments that should have been
2312        "backspace".
2313    
2314    10. Remove two redundant lines of code that can never be obeyed (their function
2315        was moved elsewhere).
2316    
2317    11. The program that makes PCRE's Unicode character property table had a bug
2318        which caused it to generate incorrect table entries for sequences of
2319        characters that have the same character type, but are in different scripts.
2320        It amalgamated them into a single range, with the script of the first of
2321        them. In other words, some characters were in the wrong script. There were
2322        thirteen such cases, affecting characters in the following ranges:
2323    
2324          U+002b0 - U+002c1
2325          U+0060c - U+0060d
2326          U+0061e - U+00612
2327          U+0064b - U+0065e
2328          U+0074d - U+0076d
2329          U+01800 - U+01805
2330          U+01d00 - U+01d77
2331          U+01d9b - U+01dbf
2332          U+0200b - U+0200f
2333          U+030fc - U+030fe
2334          U+03260 - U+0327f
2335          U+0fb46 - U+0fbb1
2336          U+10450 - U+1049d
2337    
2338    12. The -o option (show only the matching part of a line) for pcregrep was not
2339        compatible with GNU grep in that, if there was more than one match in a
2340        line, it showed only the first of them. It now behaves in the same way as
2341        GNU grep.
2342    
2343    13. If the -o and -v options were combined for pcregrep, it printed a blank
2344        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2345        does the same. The return code can be used to tell if there were any
2346        non-matching lines.
2347    
2348    14. Added --file-offsets and --line-offsets to pcregrep.
2349    
2350    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2351        infinitely looping recursion. The bug was that positive lookaheads were not
2352        being skipped when checking for a possible empty match (negative lookaheads
2353        and both kinds of lookbehind were skipped).
2354    
2355    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2356        inclusion of <windows.h> to before rather than after the definition of
2357        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2358    
2359    17. Specifying a possessive quantifier with a specific limit for a Unicode
2360        character property caused pcre_compile() to compile bad code, which led at
2361        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2362        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2363        caused the error; without that there was no problem.
2364    
2365    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2366    
2367    19. Added --enable-pcretest-libreadline.
2368    
2369    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2370        RE::GlobalReplace(). As a result, the number of replacements returned was
2371        double what it should be. I removed one of the increments, but Craig sent a
2372        later patch that removed the other one (the right fix) and added unit tests
2373        that check the return values (which was not done before).
2374    
2375    21. Several CMake things:
2376    
2377        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2378            the names libpcre and libpcreposix, not just pcre and pcreposix.
2379    
2380        (2) The above change means that pcretest and pcregrep are now correctly
2381            linked with the newly-built libraries, not previously installed ones.
2382    
2383        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2384    
2385    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2386        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2387        UTF-8 newline character). The key issue is that the pattern starts .*;
2388        this means that the match must be either at the beginning, or after a
2389        newline. The bug was in the code for advancing after a failed match and
2390        checking that the new position followed a newline. It was not taking
2391        account of UTF-8 characters correctly.
2392    
2393    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2394        character classes. PCRE was not treating the sequence [:...:] as a
2395        character class unless the ... were all letters. Perl, however, seems to
2396        allow any characters between [: and :], though of course it rejects as
2397        unknown any "names" that contain non-letters, because all the known class
2398        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2399        for example, whereas PCRE did not - it did not recognize a POSIX character
2400        class. This seemed a bit dangerous, so the code has been changed to be
2401        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2402        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2403        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2404        Perl does, and where it didn't before.
2405    
2406    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2407        Windows environments %n is disabled by default.
2408    
2409    
2410    Version 7.4 21-Sep-07
2411    ---------------------
2412    
2413    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2414        means that a class such as [\s] counted as "explicit reference to CR or
2415        LF". That isn't really right - the whole point of the change was to try to
2416        help when there was an actual mention of one of the two characters. So now
2417        the change happens only if \r or \n (or a literal CR or LF) character is
2418        encountered.
2419    
2420    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2421        of both had grown to the point where there were only 3 bits left.
2422        Fortunately, there was spare space in the data structure, and so I have
2423        moved the internal flags into a new 16-bit field to free up more option
2424        bits.
2425    
2426    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2427        but did not set the internal JCHANGED flag - either of these is enough to
2428        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2429        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2430        start sets both bits.
2431    
2432    4.  Added options (at build time, compile time, exec time) to change \R from
2433        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2434    
2435    5.  doc/pcresyntax.html was missing from the distribution.
2436    
2437    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2438        compatibility, even though it is no longer used.
2439    
2440    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2441        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2442        windows.h file is present (where different names are used). [This was
2443        reversed later after testing - see 16 below.]
2444    
2445    8.  Changed all #include <config.h> to #include "config.h". There were also
2446        some further <pcre.h> cases that I changed to "pcre.h".
2447    
2448    9.  When pcregrep was used with the --colour option, it missed the line ending
2449        sequence off the lines that it output.
2450    
2451    10. It was pointed out to me that arrays of string pointers cause lots of
2452        relocations when a shared library is dynamically loaded. A technique of
2453        using a single long string with a table of offsets can drastically reduce
2454        these. I have refactored PCRE in four places to do this. The result is
2455        dramatic:
2456    
2457          Originally:                          290
2458          After changing UCP table:            187
2459          After changing error message table:   43
2460          After changing table of "verbs"       36
2461          After changing table of Posix names   22
2462    
2463        Thanks to the folks working on Gregex for glib for this insight.
2464    
2465    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2466        unicode-properties was also set.
2467    
2468    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2469    
2470    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2471        checked only for CRLF.
2472    
2473    14. Added casts to pcretest.c to avoid compiler warnings.
2474    
2475    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2476    
2477    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2478        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2479        entirely. This removes changes made in 7 above.
2480    
2481    17. The CMake files have been updated, and there is now more information about
2482        building with CMake in the NON-UNIX-USE document.
2483    
2484    
2485    Version 7.3 28-Aug-07
2486    ---------------------
2487    
2488     1. In the rejigging of the build system that eventually resulted in 7.1, the
2489        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2490        brackets there is not right, since it causes compilers to look for an
2491        installed pcre.h, not the version that is in the source that is being
2492        compiled (which of course may be different). I have changed it back to:
2493    
2494          #include "pcre.h"
2495    
2496        I have a vague recollection that the change was concerned with compiling in
2497        different directories, but in the new build system, that is taken care of
2498        by the VPATH setting the Makefile.
2499    
2500     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2501        when the subject happened to end in the byte 0x85 (e.g. if the last
2502        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2503        characters but of course it shouldn't be taken as a newline when it is part
2504        of another character. The bug was that, for an unlimited repeat of . in
2505        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2506        characters when looking for a newline.
2507    
2508     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2509    
2510     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2511        in debug output.
2512    
2513     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2514        long printing in the pcrecpp unittest when running under MinGW.
2515    
2516     6. ESC_K was left out of the EBCDIC table.
2517    
2518     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2519        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2520        limit also applies to "virtual nesting" when a pattern is recursive, and in
2521        this case 1000 isn't so big. I have been able to remove this limit at the
2522        expense of backing off one optimization in certain circumstances. Normally,
2523        when pcre_exec() would call its internal match() function recursively and
2524        immediately return the result unconditionally, it uses a "tail recursion"
2525        feature to save stack. However, when a subpattern that can match an empty
2526        string has an unlimited repetition quantifier, it no longer makes this
2527        optimization. That gives it a stack frame in which to save the data for
2528        checking that an empty string has been matched. Previously this was taken
2529        from the 1000-entry workspace that had been reserved. So now there is no
2530        explicit limit, but more stack is used.
2531    
2532     8. Applied Daniel's patches to solve problems with the import/export magic
2533        syntax that is required for Windows, and which was going wrong for the
2534        pcreposix and pcrecpp parts of the library. These were overlooked when this
2535        problem was solved for the main library.
2536    
2537     9. There were some crude static tests to avoid integer overflow when computing
2538        the size of patterns that contain repeated groups with explicit upper
2539        limits. As the maximum quantifier is 65535, the maximum group length was
2540        set at 30,000 so that the product of these two numbers did not overflow a
2541        32-bit integer. However, it turns out that people want to use groups that
2542        are longer than 30,000 bytes (though not repeat them that many times).
2543        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2544        made it possible to implement the integer overflow checks in a much more
2545        dynamic way, which I have now done. The artificial limitation on group
2546        length has been removed - we now have only the limit on the total length of
2547        the compiled pattern, which depends on the LINK_SIZE setting.
2548    
2549    10. Fixed a bug in the documentation for get/copy named substring when
2550        duplicate names are permitted. If none of the named substrings are set, the
2551        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2552        empty string.
2553    
2554    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2555        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2556        because the ] is interpreted as the first data character and the
2557        terminating ] is not found. PCRE has been made compatible with Perl in this
2558        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2559        cause memory overwriting.
2560    
2561    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2562        string has been matched (to stop an infinite loop). It was not recognizing
2563        a conditional subpattern that could match an empty string if that
2564        subpattern was within another subpattern. For example, it looped when
2565        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2566        condition was not nested. This bug has been fixed.
2567    
2568    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2569        past the start of the subject in the presence of bytes with the top bit
2570        set, for example "\x8aBCD".
2571    
2572    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2573        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2574    
2575    14. Optimized (?!) to (*FAIL).
2576    
2577    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2578        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2579        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2580        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2581        does: it's just the validity check that is more restrictive.
2582    
2583    16. Inserted checks for integer overflows during escape sequence (backslash)
2584        processing, and also fixed erroneous offset values for syntax errors during
2585        backslash processing.
2586    
2587    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2588        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2589    
2590    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2591        caused an overrun.
2592    
2593    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2594        something other than just ASCII characters) inside a group that had an
2595        unlimited repeat caused a loop at compile time (while checking to see
2596        whether the group could match an empty string).
2597    
2598    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2599        [\P{Any}] did so. (Error in the code for printing property names.)
2600    
2601    21. An orphan \E inside a character class could cause a crash.
2602    
2603    22. A repeated capturing bracket such as (A)? could cause a wild memory
2604        reference during compilation.
2605    
2606    23. There are several functions in pcre_compile() that scan along a compiled
2607        expression for various reasons (e.g. to see if it's fixed length for look
2608        behind). There were bugs in these functions when a repeated \p or \P was
2609        present in the pattern. These operators have additional parameters compared
2610        with \d, etc, and these were not being taken into account when moving along
2611        the compiled data. Specifically:
2612    
2613        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2614            length.
2615    
2616        (b) An item such as \pL+ within a repeated group could cause crashes or
2617            loops.
2618    
2619        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2620            "reference to non-existent subpattern" error.
2621    
2622        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2623    
2624    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2625        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2626    
2627    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2628    
2629    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2630        character were causing crashes (broken optimization).
2631    
2632    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2633        \p or \P) caused a compile-time loop.
2634    
2635    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2636        break. For example, the unstudied pattern [\r\n]A does not match the string
2637        "\r\nA" because change 7.0/46 below moves the current point on by two
2638        characters after failing to match at the start. However, the pattern \nA
2639        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2640        the same is true. There doesn't seem any very clean way out of this, but
2641        what I have chosen to do makes the common cases work: PCRE now takes note
2642        of whether there can be an explicit match for \r or \n anywhere in the
2643        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2644        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2645        pattern has explicit CR or LF references.
2646    
2647    29. Added (*CR) etc for changing newline setting at start of pattern.
2648    
2649    
2650    Version 7.2 19-Jun-07
2651  ---------------------  ---------------------
2652    
2653   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,
2654      which is apparently normally available under Windows.      which is apparently normally available under Windows.
2655    
2656   2. Re-jig the pcregrep tests with different newline settings in an attempt   2. Re-jig the pcregrep tests with different newline settings in an attempt
2657      to make them independent of the local environment's newline setting.      to make them independent of the local environment's newline setting.
2658    
2659     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2660    
2661     4. Some of the "internals" tests were previously cut out when the link size
2662        was not 2, because the output contained actual offsets. The recent new
2663        "Z" feature of pcretest means that these can be cut out, making the tests
2664        usable with all link sizes.
2665    
2666     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2667        stack recursion. This gives a massive performance boost under BSD, but just
2668        a small improvement under Linux. However, it saves one field in the frame
2669        in all cases.
2670    
2671     6. Added more features from the forthcoming Perl 5.10:
2672    
2673        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2674            recursion call. It refers to the nth most recently opened parentheses.
2675    
2676        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2677            to be opened parentheses.
2678    
2679        (c) Conditions that refer to capturing parentheses can be specified
2680            relatively, for example, (?(-2)... or (?(+3)...
2681    
2682        (d) \K resets the start of the current match so that everything before
2683            is not part of it.
2684    
2685        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2686    
2687        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2688            reference syntax.
2689    
2690        (g) (?| introduces a group in which the numbering of parentheses in each
2691            alternative starts with the same number.
2692    
2693        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2694    
2695     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2696        PCRE_INFO_JCHANGED.
2697    
2698     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2699        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2700        for detecting groups that can match an empty string.
2701    
2702     9. A pattern with a very large number of alternatives (more than several
2703        hundred) was running out of internal workspace during the pre-compile
2704        phase, where pcre_compile() figures out how much memory will be needed. A
2705        bit of new cunning has reduced the workspace needed for groups with
2706        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2707        workspace instead of running out of the 4096 that are available.
2708    
2709    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2710    
2711    11. Applied patch from Google to remove an optimization that didn't quite work.
2712        The report of the bug said:
2713    
2714          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2715          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2716          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2717    
2718    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2719        it matched the wrong number of bytes.
2720    
2721    
2722  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
2723  ---------------------  ---------------------

Legend:
Removed from v.158  
changed lines
  Added in v.1439

  ViewVC Help
Powered by ViewVC 1.1.5