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

Diff of /code/trunk/ChangeLog

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

revision 591 by ph10, Sat Jan 15 17:26:57 2011 UTC revision 1414 by zherczeg, Sun Dec 22 16:27:35 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.35-RC1 xx-xxxx-201x
5    -----------------------------
6    
7    1.  A new flag is set, when property checks are present in an XCLASS.
8        When this flag is not set, PCRE can perform certain optimizations
9        such as studying these XCLASS-es.
10    
11    
12    Version 8.34 15-December-2013
13    -----------------------------
14    
15    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
16        executable memory. Patch inspired by Carsten Klein.
17    
18    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
19        this macro is never tested and has no effect, because the work to support
20        coverage involves only compiling and linking options and special targets in
21        the Makefile. The comment in config.h implied that defining the macro would
22        enable coverage support, which is totally false. There was also support for
23        setting this macro in the CMake files (my fault, I just copied it from
24        configure). SUPPORT_GCOV has now been removed.
25    
26    3.  Make a small performance improvement in strlen16() and strlen32() in
27        pcretest.
28    
29    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
30        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
31        these cases). There was also one in pcretest.c.
32    
33    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
34    
35    6.  In UTF mode, the code for checking whether a group could match an empty
36        string (which is used for indefinitely repeated groups to allow for
37        breaking an infinite loop) was broken when the group contained a repeated
38        negated single-character class with a character that occupied more than one
39        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
40        UTF-8 mode). The effect was undefined: the group might or might not be
41        deemed as matching an empty string, or the program might have crashed.
42    
43    7.  The code for checking whether a group could match an empty string was not
44        recognizing that \h, \H, \v, \V, and \R must match a character.
45    
46    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
47        an empty string. If it can, pcretest shows this in its information output.
48    
49    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
50        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
51        matched by pcre_exec() without using JIT:
52    
53        (a) If the rest of the pattern did not match after a maximal run of
54            grapheme clusters, the code for backing up to try with fewer of them
55            did not always back up over a full grapheme when characters that do not
56            have the modifier quality were involved, e.g. Hangul syllables.
57    
58        (b) If the match point in a subject started with modifier character, and
59            there was no match, the code could incorrectly back up beyond the match
60            point, and potentially beyond the first character in the subject,
61            leading to a segfault or an incorrect match result.
62    
63    10. A conditional group with an assertion condition could lead to PCRE
64        recording an incorrect first data item for a match if no other first data
65        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
66        first data item, and therefore matched "ca" after "c" instead of at the
67        start.
68    
69    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
70        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
71    
72    12. The source of pcregrep now includes z/OS-specific code so that it can be
73        compiled for z/OS as part of the special z/OS distribution.
74    
75    13. Added the -T and -TM options to pcretest.
76    
77    14. The code in pcre_compile.c for creating the table of named capturing groups
78        has been refactored. Instead of creating the table dynamically during the
79        actual compiling pass, the information is remembered during the pre-compile
80        pass (on the stack unless there are more than 20 named groups, in which
81        case malloc() is used) and the whole table is created before the actual
82        compile happens. This has simplified the code (it is now nearly 150 lines
83        shorter) and prepared the way for better handling of references to groups
84        with duplicate names.
85    
86    15. A back reference to a named subpattern when there is more than one of the
87        same name now checks them in the order in which they appear in the pattern.
88        The first one that is set is used for the reference. Previously only the
89        first one was inspected. This change makes PCRE more compatible with Perl.
90    
91    16. Unicode character properties were updated from Unicode 6.3.0.
92    
93    17. The compile-time code for auto-possessification has been refactored, based
94        on a patch by Zoltan Herczeg. It now happens after instead of during
95        compilation. The code is cleaner, and more cases are handled. The option
96        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
97        options in pcretest are provided to set it. It can also be set by
98        (*NO_AUTO_POSSESS) at the start of a pattern.
99    
100    18. The character VT has been added to the default ("C" locale) set of
101        characters that match \s and are generally treated as white space,
102        following this same change in Perl 5.18. There is now no difference between
103        "Perl space" and "POSIX space". Whether VT is treated as white space in
104        other locales depends on the locale.
105    
106    19. The code for checking named groups as conditions, either for being set or
107        for being recursed, has been refactored (this is related to 14 and 15
108        above). Processing unduplicated named groups should now be as fast at
109        numerical groups, and processing duplicated groups should be faster than
110        before.
111    
112    20. Two patches to the CMake build system, by Alexander Barkov:
113    
114          (1) Replace the "source" command by "." in CMakeLists.txt because
115              "source" is a bash-ism.
116    
117          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
118              without these the CMake build does not work on Solaris.
119    
120    21. Perl has changed its handling of \8 and \9. If there is no previously
121        encountered capturing group of those numbers, they are treated as the
122        literal characters 8 and 9 instead of a binary zero followed by the
123        literals. PCRE now does the same.
124    
125    22. Following Perl, added \o{} to specify codepoints in octal, making it
126        possible to specify values greater than 0777 and also making them
127        unambiguous.
128    
129    23. Perl now gives an error for missing closing braces after \x{... instead of
130        treating the string as literal. PCRE now does the same.
131    
132    24. RunTest used to grumble if an inappropriate test was selected explicitly,
133        but just skip it when running all tests. This make it awkward to run ranges
134        of tests when one of them was inappropriate. Now it just skips any
135        inappropriate tests, as it always did when running all tests.
136    
137    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
138        character types such as \d or \w, too many callouts were inserted, and the
139        data that they returned was rubbish.
140    
141    26. In UCP mode, \s was not matching two of the characters that Perl matches,
142        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
143        were matched by \h. The code has now been refactored so that the lists of
144        the horizontal and vertical whitespace characters used for \h and \v (which
145        are defined only in one place) are now also used for \s.
146    
147    27. Add JIT support for the 64 bit TileGX architecture.
148        Patch by Jiong Wang (Tilera Corporation).
149    
150    28. Possessive quantifiers for classes (both explicit and automatically
151        generated) now use special opcodes instead of wrapping in ONCE brackets.
152    
153    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
154        (because it's meaningless), this was not happening when PCRE_CASELESS was
155        set. Not wrong, but inefficient.
156    
157    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
158        properties for \w, \d, etc) is present in a test regex. Otherwise if the
159        test contains no characters greater than 255, Perl doesn't realise it
160        should be using Unicode semantics.
161    
162    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
163        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
164        does in Unicode mode.
165    
166    32. Added the "forbid" facility to pcretest so that putting tests into the
167        wrong test files can sometimes be quickly detected.
168    
169    33. There is now a limit (default 250) on the depth of nesting of parentheses.
170        This limit is imposed to control the amount of system stack used at compile
171        time. It can be changed at build time by --with-parens-nest-limit=xxx or
172        the equivalent in CMake.
173    
174    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
175        errors. Perl warns for these when in warning mode, but PCRE has no facility
176        for giving warnings.
177    
178    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
179        However, this was not working for (?!) because it is optimized to (*FAIL),
180        for which PCRE does not allow quantifiers. The optimization is now disabled
181        when a quantifier follows (?!). I can't see any use for this, but it makes
182        things uniform.
183    
184    36. Perl no longer allows group names to start with digits, so I have made this
185        change also in PCRE. It simplifies the code a bit.
186    
187    37. In extended mode, Perl ignores spaces before a + that indicates a
188        possessive quantifier. PCRE allowed a space before the quantifier, but not
189        before the possessive +. It now does.
190    
191    38. The use of \K (reset reported match start) within a repeated possessive
192        group such as (a\Kb)*+ was not working.
193    
194    40. Document that the same character tables must be used at compile time and
195        run time, and that the facility to pass tables to pcre_exec() and
196        pcre_dfa_exec() is for use only with saved/restored patterns.
197    
198    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
199        features for Builds with MSVC:
200    
201        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
202           builds). The .rc files can be used to set FileDescription and many other
203           attributes.
204    
205        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
206           This allows higher-level build scripts which want .pdb files to avoid
207           hard-coding the exact files needed."
208    
209    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
210        mean "start of word" and "end of word", respectively, as a transition aid.
211    
212    43. A minimizing repeat of a class containing codepoints greater than 255 in
213        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
214        compiled to use the heap for recursion.
215    
216    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
217        is configured.
218    
219    
220    Version 8.33 28-May-2013
221    ------------------------
222    
223    1.  Added 'U' to some constants that are compared to unsigned integers, to
224        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
225        variables that are added to signed variables, to ensure the result is
226        signed and can be negated.
227    
228    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
229        CMake config files.
230    
231    3.  Revise the creation of config.h.generic so that all boolean macros are
232        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
233        overriding via -D on the command line possible.
234    
235    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
236        to unsigned int is reported to make a quite noticeable speed difference in
237        a specific Windows environment. Testing on Linux did also appear to show
238        some benefit (and it is clearly not harmful). Also fixed the definition of
239        Xop which should be unsigned.
240    
241    5.  Related to (4), changing the definition of the intermediate variable cc
242        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
243        improvements.
244    
245    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
246        unnecessary spaces.
247    
248    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
249        and later.
250    
251    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
252    
253    9.  Optimizing fast_forward_start_bits in JIT.
254    
255    10. Adding support for callouts in JIT, and fixing some issues revealed
256        during this work. Namely:
257    
258        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
259    
260        (b) Minimum length was not checked before the matching is started.
261    
262    11. The value of capture_last that is passed to callouts was incorrect in some
263        cases when there was a capture on one path that was subsequently abandoned
264        after a backtrack. Also, the capture_last value is now reset after a
265        recursion, since all captures are also reset in this case.
266    
267    12. The interpreter no longer returns the "too many substrings" error in the
268        case when an overflowing capture is in a branch that is subsequently
269        abandoned after a backtrack.
270    
271    13. In the pathological case when an offset vector of size 2 is used, pcretest
272        now prints out the matched string after a yield of 0 or 1.
273    
274    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
275        Only supported by the JIT compiler at the moment.
276    
277    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
278    
279    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
280        offset of the starting point of the matching process, provided the offsets
281        vector is large enough.
282    
283    17. The \A escape now records a lookbehind value of 1, though its execution
284        does not actually inspect the previous character. This is to ensure that,
285        in partial multi-segment matching, at least one character from the old
286        segment is retained when a new segment is processed. Otherwise, if there
287        are no lookbehinds in the pattern, \A might match incorrectly at the start
288        of a new segment.
289    
290    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
291    
292    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
293        gives some modest performance improvement in 8-bit mode.
294    
295    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
296        be expressed in certain programming languages using Universal Character
297        Names.
298    
299    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
300        which points out that "non characters" are not "characters that may not
301        appear in Unicode strings" but rather "characters that are reserved for
302        internal use and have only local meaning".
303    
304    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
305        there was a conditional group that depended on an assertion, if the
306        assertion was false, the callout that immediately followed the alternation
307        in the condition was skipped when pcre_exec() was used for matching.
308    
309    23. Allow an explicit callout to be inserted before an assertion that is the
310        condition for a conditional group, for compatibility with automatic
311        callouts, which always insert a callout at this point.
312    
313    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
314        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
315    
316    25. (*PRUNE) is now supported by the JIT compiler.
317    
318    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
319    
320    27. Fix the case where there are two or more SKIPs with arguments that may be
321        ignored.
322    
323    28. (*SKIP) is now supported by the JIT compiler.
324    
325    29. (*THEN) is now supported by the JIT compiler.
326    
327    30. Update RunTest with additional test selector options.
328    
329    31. The way PCRE handles backtracking verbs has been changed in two ways.
330    
331        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
332        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
333        backtracking. In some cases this makes it more Perl-compatible, but Perl's
334        rather obscure rules do not always do the same thing.
335    
336        (2) Previously, backtracking verbs were confined within assertions. This is
337        no longer the case for positive assertions, except for (*ACCEPT). Again,
338        this sometimes improves Perl compatibility, and sometimes does not.
339    
340    32. A number of tests that were in test 2 because Perl did things differently
341        have been moved to test 1, because either Perl or PCRE has changed, and
342        these tests are now compatible.
343    
344    32. Backtracking control verbs are now handled in the same way in JIT and
345        interpreter.
346    
347    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
348        contained a forward subroutine reference caused a compile error.
349    
350    34. Auto-detect and optimize limited repetitions in JIT.
351    
352    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
353        blocking (*UTF) etc.
354    
355    36. In the interpreter, maximizing pattern repetitions for characters and
356        character types now use tail recursion, which reduces stack usage.
357    
358    37. The value of the max lookbehind was not correctly preserved if a compiled
359        and saved regex was reloaded on a host of different endianness.
360    
361    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
362        of the compiled pattern block, expand the flags field from 16 to 32 bits
363        because it was almost full.
364    
365    39. Try madvise first before posix_madvise.
366    
367    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
368        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
369        original change locked out all matches of empty strings. This has been
370        changed so that one match of an empty string per line is recognized.
371        Subsequent searches on the same line (for colouring or for --only-matching,
372        for example) do not recognize empty strings.
373    
374    41. Applied a user patch to fix a number of spelling mistakes in comments.
375    
376    42. Data lines longer than 65536 caused pcretest to crash.
377    
378    43. Clarified the data type for length and startoffset arguments for pcre_exec
379        and pcre_dfa_exec in the function-specific man pages, where they were
380        explicitly stated to be in bytes, never having been updated. I also added
381        some clarification to the pcreapi man page.
382    
383    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
384        a segmentation fault.
385    
386    
387    Version 8.32 30-November-2012
388    -----------------------------
389    
390    1.  Improved JIT compiler optimizations for first character search and single
391        character iterators.
392    
393    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
394        Patch by Daniel Richard G.
395    
396    3.  Single character iterator optimizations in the JIT compiler.
397    
398    4.  Improved JIT compiler optimizations for character ranges.
399    
400    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
401        Reported by Giuseppe D'Angelo.
402    
403    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
404        of a line, was being set incorrectly in cases where .* appeared inside
405        atomic brackets at the start of a pattern, or where there was a subsequent
406        *PRUNE or *SKIP.
407    
408    7.  Improved instruction cache flush for POWER/PowerPC.
409        Patch by Daniel Richard G.
410    
411    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
412        grep:
413    
414        (a) There is now no limit to the number of patterns to be matched.
415    
416        (b) An error is given if a pattern is too long.
417    
418        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
419            are now supported.
420    
421        (d) --exclude-from and --include-from (multiple use) have been added.
422    
423        (e) Exclusions and inclusions now apply to all files and directories, not
424            just to those obtained from scanning a directory recursively.
425    
426        (f) Multiple uses of -f and --file-list are now supported.
427    
428        (g) In a Windows environment, the default for -d has been changed from
429            "read" (the GNU grep default) to "skip", because otherwise the presence
430            of a directory in the file list provokes an error.
431    
432        (h) The documentation has been revised and clarified in places.
433    
434    9.  Improve the matching speed of capturing brackets.
435    
436    10. Changed the meaning of \X so that it now matches a Unicode extended
437        grapheme cluster.
438    
439    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
440        out POSIX threads when JIT support is configured.
441    
442    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
443    
444    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
445        structure could go wrong in environments where size_t is not the same size
446        as int.
447    
448    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
449        set.
450    
451    15. The EBCDIC support had decayed; later updates to the code had included
452        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
453        general tidy up of EBCDIC-related issues, and the documentation was also
454        not quite right. There is now a test that can be run on ASCII systems to
455        check some of the EBCDIC-related things (but is it not a full test).
456    
457    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
458        in a small tidy to the code.
459    
460    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
461    
462    18. If the --only-matching (-o) option in pcregrep is specified multiple
463        times, each one causes appropriate output. For example, -o1 -o2 outputs the
464        substrings matched by the 1st and 2nd capturing parentheses. A separating
465        string can be specified by --om-separator (default empty).
466    
467    19. Improving the first n character searches.
468    
469    20. Turn case lists for horizontal and vertical white space into macros so that
470        they are defined only once.
471    
472    21. This set of changes together give more compatible Unicode case-folding
473        behaviour for characters that have more than one other case when UCP
474        support is available.
475    
476        (a) The Unicode property table now has offsets into a new table of sets of
477            three or more characters that are case-equivalent. The MultiStage2.py
478            script that generates these tables (the pcre_ucd.c file) now scans
479            CaseFolding.txt instead of UnicodeData.txt for character case
480            information.
481    
482        (b) The code for adding characters or ranges of characters to a character
483            class has been abstracted into a generalized function that also handles
484            case-independence. In UTF-mode with UCP support, this uses the new data
485            to handle characters with more than one other case.
486    
487        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
488            whose other case is greater than 256 are now correctly matched
489            caselessly. Previously, the high codepoint matched the low one, but not
490            vice versa.
491    
492        (d) The processing of \h, \H, \v, and \ in character classes now makes use
493            of the new class addition function, using character lists defined as
494            macros alongside the case definitions of 20 above.
495    
496        (e) Caseless back references now work with characters that have more than
497            one other case.
498    
499        (f) General caseless matching of characters with more than one other case
500            is supported.
501    
502    22. Unicode character properties were updated from Unicode 6.2.0
503    
504    23. Improved CMake support under Windows. Patch by Daniel Richard G.
505    
506    24. Add support for 32-bit character strings, and UTF-32
507    
508    25. Major JIT compiler update (code refactoring and bugfixing).
509        Experimental Sparc 32 support is added.
510    
511    26. Applied a modified version of Daniel Richard G's patch to create
512        pcre.h.generic and config.h.generic by "make" instead of in the
513        PrepareRelease script.
514    
515    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
516        pcre_compile.c when checking for a zero character.
517    
518    28. Introducing a native interface for JIT. Through this interface, the compiled
519        machine code can be directly executed. The purpose of this interface is to
520        provide fast pattern matching, so several sanity checks are not performed.
521        However, feature tests are still performed. The new interface provides
522        1.4x speedup compared to the old one.
523    
524    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
525        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
526        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
527    
528    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
529        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
530        the "old" RFC 2279). Instead, it ended up passing a negative length to
531        pcre_exec().
532    
533    31. Add support for GCC's visibility feature to hide internal functions.
534    
535    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
536        "unknown -C option" after outputting 0 or 1.
537    
538    33. There is now support for generating a code coverage report for the test
539        suite in environments where gcc is the compiler and lcov is installed. This
540        is mainly for the benefit of the developers.
541    
542    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
543        unaddressable using valgrind annotations, allowing valgrind to detect
544        invalid memory accesses. This is mainly for the benefit of the developers.
545    
546    25. (*UTF) can now be used to start a pattern in any of the three libraries.
547    
548    26. Give configure error if --enable-cpp but no C++ compiler found.
549    
550    
551    Version 8.31 06-July-2012
552    -------------------------
553    
554    1.  Fixing a wrong JIT test case and some compiler warnings.
555    
556    2.  Removed a bashism from the RunTest script.
557    
558    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
559        to unsigned type, result still unsigned" that was given by an MS compiler
560        on encountering the code "-sizeof(xxx)".
561    
562    4.  Partial matching support is added to the JIT compiler.
563    
564    5.  Fixed several bugs concerned with partial matching of items that consist
565        of more than one character:
566    
567        (a) /^(..)\1/ did not partially match "aba" because checking references was
568            done on an "all or nothing" basis. This also applied to repeated
569            references.
570    
571        (b) \R did not give a hard partial match if \r was found at the end of the
572            subject.
573    
574        (c) \X did not give a hard partial match after matching one or more
575            characters at the end of the subject.
576    
577        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
578            a partial match for the string "\r".
579    
580        (e) When newline was set to CRLF, the metacharacter "." did not recognize
581            a partial match for a CR character at the end of the subject string.
582    
583    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
584        running pcretest, the text "(JIT)" added to the output whenever JIT is
585        actually used to run the match.
586    
587    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
588        or /S+[+] with a digit between 1 and 7.
589    
590    8.  OP_NOT now supports any UTF character not just single-byte ones.
591    
592    9.  (*MARK) control verb is now supported by the JIT compiler.
593    
594    10. The command "./RunTest list" lists the available tests without actually
595        running any of them. (Because I keep forgetting what they all are.)
596    
597    11. Add PCRE_INFO_MAXLOOKBEHIND.
598    
599    12. Applied a (slightly modified) user-supplied patch that improves performance
600        when the heap is used for recursion (compiled with --disable-stack-for-
601        recursion). Instead of malloc and free for each heap frame each time a
602        logical recursion happens, frames are retained on a chain and re-used where
603        possible. This sometimes gives as much as 30% improvement.
604    
605    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
606        call.
607    
608    14. As documented, (*COMMIT) is now confined to within a positive assertion.
609    
610    15. It is now possible to link pcretest with libedit as an alternative to
611        libreadline.
612    
613    16. (*COMMIT) control verb is now supported by the JIT compiler.
614    
615    17. The Unicode data tables have been updated to Unicode 6.1.0.
616    
617    18. Added --file-list option to pcregrep.
618    
619    19. Added binary file support to pcregrep, including the -a, --binary-files,
620        -I, and --text options.
621    
622    20. The madvise function is renamed for posix_madvise for QNX compatibility
623        reasons. Fixed by Giuseppe D'Angelo.
624    
625    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
626    
627    22. Changed the option for creating symbolic links for 16-bit man pages from
628        -s to -sf so that re-installing does not cause issues.
629    
630    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
631    
632    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
633        matches in certain environments (the workspace was not being correctly
634        retained). Also added to pcre_dfa_exec() a simple plausibility check on
635        some of the workspace data at the beginning of a restart.
636    
637    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
638        was not doing so when it should - probably a typo introduced by SVN 528
639        (change 8.10/14).
640    
641    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
642        \w+ when the character tables indicated that \x{c4} was a word character.
643        There were several related cases, all because the tests for doing a table
644        lookup were testing for characters less than 127 instead of 255.
645    
646    27. If a pattern contains capturing parentheses that are not used in a match,
647        their slots in the ovector are set to -1. For those that are higher than
648        any matched groups, this happens at the end of processing. In the case when
649        there were back references that the ovector was too small to contain
650        (causing temporary malloc'd memory to be used during matching), and the
651        highest capturing number was not used, memory off the end of the ovector
652        was incorrectly being set to -1. (It was using the size of the temporary
653        memory instead of the true size.)
654    
655    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
656        ovector size, it uses memory at the end of the block that it has got.
657    
658    29. Check for an overlong MARK name and give an error at compile time. The
659        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
660    
661    30. JIT compiler update.
662    
663    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
664        Rill for the patch.
665    
666    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
667    
668    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
669    
670    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
671        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
672        was enabled.
673    
674    35. Improve JIT code generation for greedy plus quantifier.
675    
676    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
677        1 to "aa" instead of to an empty string. The bug affected repeated groups
678        that could potentially match an empty string.
679    
680    37. Optimizing single character iterators in JIT.
681    
682    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
683        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
684        codepoints that are too big for the mode are faulted, and in a UTF mode,
685        disallowed codepoints are also faulted.
686    
687    39. If PCRE was compiled with UTF support, in three places in the DFA
688        matcher there was code that should only have been obeyed in UTF mode, but
689        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
690        processing when bytes with values greater than 127 were present. In 16-bit
691        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
692        both cases the values are those that cannot be the first data item in a UTF
693        character. The three items that might have provoked this were recursions,
694        possessively repeated groups, and atomic groups.
695    
696    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
697        and pcregrep, because some OS require shared objects to be explicitly
698        passed to ld, causing the link step to fail if they are not.
699    
700    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
701        mode, patterns that started with \h* or \R* might be incorrectly matched.
702    
703    
704    Version 8.30 04-February-2012
705    -----------------------------
706    
707    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
708        name is defined in ctype.h.
709    
710    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
711        only in quite long subpatterns.
712    
713    3.  Removed the function pcre_info(), which has been obsolete and deprecated
714        since it was replaced by pcre_fullinfo() in February 2000.
715    
716    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
717        match a (*MARK), and the match failed at the start of the subject, a
718        reference to memory before the start of the subject could occur. This bug
719        was introduced by fix 17 of release 8.21.
720    
721    5.  A reference to an unset group with zero minimum repetition was giving
722        totally wrong answers (in non-JavaScript-compatibility mode). For example,
723        /(another)?(\1?)test/ matched against "hello world test". This bug was
724        introduced in release 8.13.
725    
726    6.  Add support for 16-bit character strings (a large amount of work involving
727        many changes and refactorings).
728    
729    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
730        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
731        from a file.
732    
733    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
734        rounding is not applied in this particular case).
735    
736    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
737        if they appear, or are escaped, in patterns.
738    
739    10. Get rid of a number of -Wunused-but-set-variable warnings.
740    
741    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
742        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
743        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
744        also returns the mark "x". This bug applied to capturing parentheses,
745        non-capturing parentheses, and atomic parentheses. It also applied to some
746        assertions.
747    
748    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
749        information out of configure.ac instead of relying on pcre.h.generic, which
750        is not stored in the repository.
751    
752    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
753        -lreadline.
754    
755    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
756    
757    15. Applied Graycode's patch to put the top-level frame on the stack rather
758        than the heap when not using the stack for recursion. This gives a
759        performance improvement in many cases when recursion is not deep.
760    
761    16. Experimental code added to "pcretest -C" to output the stack frame size.
762    
763    
764    Version 8.21 12-Dec-2011
765    ------------------------
766    
767    1.  Updating the JIT compiler.
768    
769    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
770        are added as well.
771    
772    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
773        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
774        calling _pcre_jit_exec. Some extra comments are added.
775    
776    4.  (*MARK) settings inside atomic groups that do not contain any capturing
777        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
778        was introduced by change 18 for 8.20.
779    
780    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
781        ECMA-262 standard.
782    
783    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
784        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
785        This bug was probably introduced by change 9 of 8.13.
786    
787    7.  While fixing 6 above, I noticed that a number of other items were being
788        incorrectly rejected as "not fixed length". This arose partly because newer
789        opcodes had not been added to the fixed-length checking code. I have (a)
790        corrected the bug and added tests for these items, and (b) arranged for an
791        error to occur if an unknown opcode is encountered while checking for fixed
792        length instead of just assuming "not fixed length". The items that were
793        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
794        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
795        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
796    
797    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
798        being incorrectly compiled and would have given unpredicatble results.
799    
800    9.  A possessively repeated subpattern with minimum repeat count greater than
801        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
802        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
803        the first (A) could occur when it should not.
804    
805    10. Add a cast and remove a redundant test from the code.
806    
807    11. JIT should use pcre_malloc/pcre_free for allocation.
808    
809    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
810        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
811        is anything other than /usr, -L is still shown).
812    
813    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
814    
815    14. Perl does not support \N without a following name in a [] class; PCRE now
816        also gives an error.
817    
818    15. If a forward reference was repeated with an upper limit of around 2000,
819        it caused the error "internal error: overran compiling workspace". The
820        maximum number of forward references (including repeats) was limited by the
821        internal workspace, and dependent on the LINK_SIZE. The code has been
822        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
823        the default depends on LINK_SIZE. There is a new upper limit (for safety)
824        of around 200,000 forward references. While doing this, I also speeded up
825        the filling in of repeated forward references.
826    
827    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
828        incorrectly expecting the subject to contain another "a" after the start.
829    
830    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
831        in the match, the SKIP should be ignored. This was not happening; instead
832        the SKIP was being treated as NOMATCH. For patterns such as
833        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
834        tested.
835    
836    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
837        now much more compatible with Perl, in particular in cases where the result
838        is a non-match for a non-anchored pattern. For example, if
839        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
840        "m", where previously it did not return a name. A side effect of this
841        change is that for partial matches, the last encountered mark name is
842        returned, as for non matches. A number of tests that were previously not
843        Perl-compatible have been moved into the Perl-compatible test files. The
844        refactoring has had the pleasing side effect of removing one argument from
845        the match() function, thus reducing its stack requirements.
846    
847    19. If the /S+ option was used in pcretest to study a pattern using JIT,
848        subsequent uses of /S (without +) incorrectly behaved like /S+.
849    
850    21. Retrieve executable code size support for the JIT compiler and fixing
851        some warnings.
852    
853    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
854        not work when the shorter character appeared right at the end of the
855        subject string.
856    
857    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
858        systems.
859    
860    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
861        output it when the /M option is used in pcretest.
862    
863    25. The CheckMan script was not being included in the distribution. Also, added
864        an explicit "perl" to run Perl scripts from the PrepareRelease script
865        because this is reportedly needed in Windows.
866    
867    26. If study data was being save in a file and studying had not found a set of
868        "starts with" bytes for the pattern, the data written to the file (though
869        never used) was taken from uninitialized memory and so caused valgrind to
870        complain.
871    
872    27. Updated RunTest.bat as provided by Sheri Pierce.
873    
874    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
875    
876    29. Computation of memory usage for the table of capturing group names was
877        giving an unnecessarily large value.
878    
879    
880    Version 8.20 21-Oct-2011
881    ------------------------
882    
883    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
884        a POSIX class. After further experiments with Perl, which convinced me that
885        Perl has bugs and confusions, a closing square bracket is no longer allowed
886        in a POSIX name. This bug also affected patterns with classes that started
887        with full stops.
888    
889    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
890        captured substring, but while checking the failing first alternative,
891        substring 1 is temporarily captured. If the output vector supplied to
892        pcre_exec() was not big enough for this capture, the yield of the function
893        was still zero ("insufficient space for captured substrings"). This cannot
894        be totally fixed without adding another stack variable, which seems a lot
895        of expense for a edge case. However, I have improved the situation in cases
896        such as /(a)(b)x|abc/ matched against "abc", where the return code
897        indicates that fewer than the maximum number of slots in the ovector have
898        been set.
899    
900    3.  Related to (2) above: when there are more back references in a pattern than
901        slots in the output vector, pcre_exec() uses temporary memory during
902        matching, and copies in the captures as far as possible afterwards. It was
903        using the entire output vector, but this conflicts with the specification
904        that only 2/3 is used for passing back captured substrings. Now it uses
905        only the first 2/3, for compatibility. This is, of course, another edge
906        case.
907    
908    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
909        main code base, and can be used by building with --enable-jit. When this is
910        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
911        runtime --no-jit option is given.
912    
913    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
914        ovector, the return from the function was zero, implying that there were
915        other matches that did not fit. The correct "exactly full" value is now
916        returned.
917    
918    6.  If a subpattern that was called recursively or as a subroutine contained
919        (*PRUNE) or any other control that caused it to give a non-standard return,
920        invalid errors such as "Error -26 (nested recursion at the same subject
921        position)" or even infinite loops could occur.
922    
923    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
924        computing the minimum length on reaching *ACCEPT, and so ended up with the
925        wrong value of 1 rather than 0. Further investigation indicates that
926        computing a minimum subject length in the presence of *ACCEPT is difficult
927        (think back references, subroutine calls), and so I have changed the code
928        so that no minimum is registered for a pattern that contains *ACCEPT.
929    
930    8.  If (*THEN) was present in the first (true) branch of a conditional group,
931        it was not handled as intended. [But see 16 below.]
932    
933    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
934        Sheri Pierce.
935    
936    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
937        the first byte in a match must be "a".
938    
939    11. Change 17 for 8.13 increased the recursion depth for patterns like
940        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
941        pattern contains any instances of (*THEN). If it does not, the old
942        optimizations are restored. It would be nice to do this on a per-group
943        basis, but at the moment that is not feasible.
944    
945    12. In some environments, the output of pcretest -C is CRLF terminated. This
946        broke RunTest's code that checks for the link size. A single white space
947        character after the value is now allowed for.
948    
949    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
950        For "fr", it uses the Windows-specific input and output files.
951    
952    14. If (*THEN) appeared in a group that was called recursively or as a
953        subroutine, it did not work as intended. [But see next item.]
954    
955    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
956        pattern fragments (but not containing any | characters). If A and B are
957        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
958        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
959        D. In other words, Perl considers parentheses that do not contain any |
960        characters to be part of a surrounding alternative, whereas PCRE was
961        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
962        differently. PCRE now behaves in the same way as Perl, except in the case
963        of subroutine/recursion calls such as (?1) which have in any case always
964        been different (but PCRE had them first :-).
965    
966    16. Related to 15 above: Perl does not treat the | in a conditional group as
967        creating alternatives. Such a group is treated in the same way as an
968        ordinary group without any | characters when processing (*THEN). PCRE has
969        been changed to match Perl's behaviour.
970    
971    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
972        RunGrepTest script failed.
973    
974    18. Change 22 for version 13 caused atomic groups to use more stack. This is
975        inevitable for groups that contain captures, but it can lead to a lot of
976        stack use in large patterns. The old behaviour has been restored for atomic
977        groups that do not contain any capturing parentheses.
978    
979    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
980        suppress the check for a minimum subject length at run time. (If it was
981        given to pcre_exec() or pcre_dfa_exec() it did work.)
982    
983    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
984        fail to work when decoding hex characters in data strings in EBCDIC
985        environments.
986    
987    21. It appears that in at least one Mac OS environment, the isxdigit() function
988        is implemented as a macro that evaluates to its argument more than once,
989        contravening the C 90 Standard (I haven't checked a later standard). There
990        was an instance in pcretest which caused it to go wrong when processing
991        \x{...} escapes in subject strings. The has been rewritten to avoid using
992        things like p++ in the argument of isxdigit().
993    
994    
995    Version 8.13 16-Aug-2011
996    ------------------------
997    
998    1.  The Unicode data tables have been updated to Unicode 6.0.0.
999    
1000    2.  Two minor typos in pcre_internal.h have been fixed.
1001    
1002    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1003        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1004        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1005    
1006    4.  There were a number of related bugs in the code for matching backrefences
1007        caselessly in UTF-8 mode when codes for the characters concerned were
1008        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1009        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1010        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1011        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1012        2-byte code at the end of the subject (it thought there wasn't enough data
1013        left).
1014    
1015    5.  Comprehensive information about what went wrong is now returned by
1016        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1017        as the output vector has at least 2 elements. The offset of the start of
1018        the failing character and a reason code are placed in the vector.
1019    
1020    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1021        now returned is for the first byte of the failing character, instead of the
1022        last byte inspected. This is an incompatible change, but I hope it is small
1023        enough not to be a problem. It makes the returned offset consistent with
1024        pcre_exec() and pcre_dfa_exec().
1025    
1026    7.  pcretest now gives a text phrase as well as the error number when
1027        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1028        failure, the offset and reason code are output.
1029    
1030    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1031        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1032        back over a single character (\n). This seems wrong (because it treated the
1033        two characters as a single entity when going forwards), conflicts with the
1034        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1035        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1036        has been changed.
1037    
1038    9.  Some internal refactoring has changed the processing so that the handling
1039        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1040        time (the PCRE_DOTALL option was changed this way some time ago: version
1041        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1042        which was always a bit of a fudge. It also means that there is one less
1043        argument for the match() function, which reduces its stack requirements
1044        slightly. This change also fixes an incompatibility with Perl: the pattern
1045        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1046    
1047    10. More internal refactoring has drastically reduced the number of recursive
1048        calls to match() for possessively repeated groups such as (abc)++ when
1049        using pcre_exec().
1050    
1051    11. While implementing 10, a number of bugs in the handling of groups were
1052        discovered and fixed:
1053    
1054        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1055        (a|)*(?1) gave a compile-time internal error.
1056        ((a|)+)+  did not notice that the outer group could match an empty string.
1057        (^a|^)+   was not marked as anchored.
1058        (.*a|.*)+ was not marked as matching at start or after a newline.
1059    
1060    12. Yet more internal refactoring has removed another argument from the match()
1061        function. Special calls to this function are now indicated by setting a
1062        value in a variable in the "match data" data block.
1063    
1064    13. Be more explicit in pcre_study() instead of relying on "default" for
1065        opcodes that mean there is no starting character; this means that when new
1066        ones are added and accidentally left out of pcre_study(), testing should
1067        pick them up.
1068    
1069    14. The -s option of pcretest has been documented for ages as being an old
1070        synonym of -m (show memory usage). I have changed it to mean "force study
1071        for every regex", that is, assume /S for every regex. This is similar to -i
1072        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1073        using it. It makes it easier to run collections of tests with and without
1074        study enabled, and thereby test pcre_study() more easily. All the standard
1075        tests are now run with and without -s (but some patterns can be marked as
1076        "never study" - see 20 below).
1077    
1078    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1079        restoration of the capturing data to the outer values was not happening
1080        correctly.
1081    
1082    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1083        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1084        pattern had matched an empty string, and so incorrectly returned a no
1085        match.
1086    
1087    17. There was optimizing code for the last branch of non-capturing parentheses,
1088        and also for the obeyed branch of a conditional subexpression, which used
1089        tail recursion to cut down on stack usage. Unfortunately, now that there is
1090        the possibility of (*THEN) occurring in these branches, tail recursion is
1091        no longer possible because the return has to be checked for (*THEN). These
1092        two optimizations have therefore been removed. [But see 8.20/11 above.]
1093    
1094    18. If a pattern containing \R was studied, it was assumed that \R always
1095        matched two bytes, thus causing the minimum subject length to be
1096        incorrectly computed because \R can also match just one byte.
1097    
1098    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1099        was incorrectly computed.
1100    
1101    20. If /S is present twice on a test pattern in pcretest input, it now
1102        *disables* studying, thereby overriding the use of -s on the command line
1103        (see 14 above). This is necessary for one or two tests to keep the output
1104        identical in both cases.
1105    
1106    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1107        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1108    
1109    22. When an atomic group that contained a capturing parenthesis was
1110        successfully matched, but the branch in which it appeared failed, the
1111        capturing was not being forgotten if a higher numbered group was later
1112        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1113        group 1 to "a", when in fact it should be unset. This applied to multi-
1114        branched capturing and non-capturing groups, repeated or not, and also to
1115        positive assertions (capturing in negative assertions does not happen
1116        in PCRE) and also to nested atomic groups.
1117    
1118    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1119        subject after a captured substring, to make it easier to tell which of a
1120        number of identical substrings has been captured.
1121    
1122    24. The way atomic groups are processed by pcre_exec() has been changed so that
1123        if they are repeated, backtracking one repetition now resets captured
1124        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1125        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1126        "aaa". Previously, it would have been "a". As part of this code
1127        refactoring, the way recursive calls are handled has also been changed.
1128    
1129    25. If an assertion condition captured any substrings, they were not passed
1130        back unless some other capturing happened later. For example, if
1131        (?(?=(a))a) was matched against "a", no capturing was returned.
1132    
1133    26. When studying a pattern that contained subroutine calls or assertions,
1134        the code for finding the minimum length of a possible match was handling
1135        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1136        group 1 called group 2 while simultaneously a separate group 2 called group
1137        1). A stack overflow occurred in this case. I have fixed this by limiting
1138        the recursion depth to 10.
1139    
1140    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1141        Fortmann. This supports explicit test numbers on the command line, and has
1142        argument validation and error reporting.
1143    
1144    28. An instance of \X with an unlimited repeat could fail if at any point the
1145        first character it looked at was a mark character.
1146    
1147    29. Some minor code refactoring concerning Unicode properties and scripts
1148        should reduce the stack requirement of match() slightly.
1149    
1150    30. Added the '=' option to pcretest to check the setting of unused capturing
1151        slots at the end of the pattern, which are documented as being -1, but are
1152        not included in the return count.
1153    
1154    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1155        compiled something random. Now it gives a compile-time error (as does
1156        Perl).
1157    
1158    32. A *MARK encountered during the processing of a positive assertion is now
1159        recorded and passed back (compatible with Perl).
1160    
1161    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1162        had alternative anchored branches, the search for a second match in a line
1163        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1164        matched the line "0102" twice. The same bug affected patterns that started
1165        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1166        twice.
1167    
1168    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1169        does, and because of capturing effects, quantifying parenthesized
1170        assertions may at times be useful. Quantifiers are now allowed for
1171        parenthesized assertions.
1172    
1173    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1174    
1175    36. \g was being checked for fancy things in a character class, when it should
1176        just be a literal "g".
1177    
1178    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1179        appearance of a nested POSIX class supersedes an apparent external class.
1180        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1181        unescaped square brackets may also appear as part of class names. For
1182        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1183        more like Perl. (But see 8.20/1 above.)
1184    
1185    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1186        was because it thought it was \N{name}, which is not supported).
1187    
1188    39. Add minix to OS list not supporting the -S option in pcretest.
1189    
1190    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1191        cannot analyze patterns in sufficient detail to catch mutual recursions
1192        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1193        subgroup is called recursively as a subpattern for a second time at the
1194        same position in the subject string. In previous releases this might have
1195        been caught by the recursion limit, or it might have run out of stack.
1196    
1197    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1198        happen only once. PCRE was, however incorrectly giving a compile time error
1199        "recursive call could loop indefinitely" because it cannot analyze the
1200        pattern in sufficient detail. The compile time test no longer happens when
1201        PCRE is compiling a conditional subpattern, but actual runaway loops are
1202        now caught at runtime (see 40 above).
1203    
1204    42. It seems that Perl allows any characters other than a closing parenthesis
1205        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1206        has been changed to be the same.
1207    
1208    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1209        as not to get warnings when autogen.sh is called. Also changed
1210        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1211    
1212    44. To help people who use pcregrep to scan files containing exceedingly long
1213        lines, the following changes have been made:
1214    
1215        (a) The default value of the buffer size parameter has been increased from
1216            8K to 20K. (The actual buffer used is three times this size.)
1217    
1218        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1219            PCRE is built.
1220    
1221        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1222            to be set at run time.
1223    
1224        (d) Numerical values in pcregrep options can be followed by K or M, for
1225            example --buffer-size=50K.
1226    
1227        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1228            given and the return code is set to 2.
1229    
1230    45. Add a pointer to the latest mark to the callout data block.
1231    
1232    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1233        partial match of an empty string instead of no match. This was specific to
1234        the use of ".".
1235    
1236    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1237        complete match instead of a partial match. This bug was dependent on both
1238        the PCRE_UTF8 and PCRE_DOTALL options being set.
1239    
1240    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1241        starting byte set, because \b was not being ignored.
1242    
1243    
1244  Version 8.12 15-Jan-2011  Version 8.12 15-Jan-2011
1245  ------------------------  ------------------------
1246    
# Line 651  Version 7.9 11-Apr-09 Line 1891  Version 7.9 11-Apr-09
1891  7.  A pattern that could match an empty string could cause pcregrep to loop; it  7.  A pattern that could match an empty string could cause pcregrep to loop; it
1892      doesn't make sense to accept an empty string match in pcregrep, so I have      doesn't make sense to accept an empty string match in pcregrep, so I have
1893      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1894      seems to be how GNU grep behaves.      seems to be how GNU grep behaves. [But see later change 40 for release
1895        8.33.]
1896    
1897  8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at  8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1898      start or after a newline", because the conditional assertion was not being      start or after a newline", because the conditional assertion was not being
# Line 894  Version 7.7 07-May-08 Line 2135  Version 7.7 07-May-08
2135      containing () gave an internal compiling error instead of "reference to      containing () gave an internal compiling error instead of "reference to
2136      non-existent subpattern". Fortunately, when the pattern did exist, the      non-existent subpattern". Fortunately, when the pattern did exist, the
2137      compiled code was correct. (When scanning forwards to check for the      compiled code was correct. (When scanning forwards to check for the
2138      existencd of the subpattern, it was treating the data ']' as terminating      existence of the subpattern, it was treating the data ']' as terminating
2139      the class, so got the count wrong. When actually compiling, the reference      the class, so got the count wrong. When actually compiling, the reference
2140      was subsequently set up correctly.)      was subsequently set up correctly.)
2141    

Legend:
Removed from v.591  
changed lines
  Added in v.1414

  ViewVC Help
Powered by ViewVC 1.1.5