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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.127  
changed lines
  Added in v.1432

  ViewVC Help
Powered by ViewVC 1.1.5