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

Diff of /code/trunk/ChangeLog

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

revision 425 by ph10, Mon Aug 17 14:48:48 2009 UTC revision 1473 by ph10, Mon Apr 21 17:52:38 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 8.00 ??-???-??  Version 8.36 xx-xxx-2014
5    ------------------------
6    
7    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
8        -Wmissing-field-initializers and -Wunused-parameter.
9    
10    2.  The tests for quantifiers being too big (greater than 65535) were being
11        applied after reading the number, and stupidly assuming that integer
12        overflow would give a negative number. The tests are now applied as the
13        numbers are read.
14    
15    3.  Tidy code in pcre_exec.c where two branches that used to be different are
16        now the same.
17    
18    
19    Version 8.35 04-April-2014
20    --------------------------
21    
22    1.  A new flag is set, when property checks are present in an XCLASS.
23        When this flag is not set, PCRE can perform certain optimizations
24        such as studying these XCLASS-es.
25    
26    2.  The auto-possessification of character sets were improved: a normal
27        and an extended character set can be compared now. Furthermore
28        the JIT compiler optimizes more character set checks.
29    
30    3.  Got rid of some compiler warnings for potentially uninitialized variables
31        that show up only when compiled with -O2.
32    
33    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
34        of a match later then the end of the match. The pcretest program was not
35        handling the case sensibly - it was outputting from the start to the next
36        binary zero. It now reports this situation in a message, and outputs the
37        text from the end to the start.
38    
39    5.  Fast forward search is improved in JIT. Instead of the first three
40        characters, any three characters with fixed position can be searched.
41        Search order: first, last, middle.
42    
43    6.  Improve character range checks in JIT. Characters are read by an inprecise
44        function now, which returns with an unknown value if the character code is
45        above a certain treshold (e.g: 256). The only limitation is that the value
46        must be bigger than the treshold as well. This function is useful, when
47        the characters above the treshold are handled in the same way.
48    
49    7.  The macros whose names start with RAWUCHAR are placeholders for a future
50        mode in which only the bottom 21 bits of 32-bit data items are used. To
51        make this more memorable for those maintaining the code, the names have
52        been changed to start with UCHAR21, and an extensive comment has been added
53        to their definition.
54    
55    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
56        to the export list in Makefile.am (they were accidentally omitted from the
57        8.34 tarball).
58    
59    9.  The informational output from pcretest used the phrase "starting byte set"
60        which is inappropriate for the 16-bit and 32-bit libraries. As the output
61        for "first char" and "need char" really means "non-UTF-char", I've changed
62        "byte" to "char", and slightly reworded the output. The documentation about
63        these values has also been (I hope) clarified.
64    
65    10. Another JIT related optimization: use table jumps for selecting the correct
66        backtracking path, when more than four alternatives are present inside a
67        bracket.
68    
69    11. Empty match is not possible, when the minimum length is greater than zero,
70        and there is no \K in the pattern. JIT should avoid empty match checks in
71        such cases.
72    
73    12. In a caseless character class with UCP support, when a character with more
74        than one alternative case was not the first character of a range, not all
75        the alternative cases were added to the class. For example, s and \x{17f}
76        are both alternative cases for S: the class [RST] was handled correctly,
77        but [R-T] was not.
78    
79    13. The configure.ac file always checked for pthread support when JIT was
80        enabled. This is not used in Windows, so I have put this test inside a
81        check for the presence of windows.h (which was already tested for).
82    
83    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
84        The algorithm provides a way to skip certain starting offsets, and usually
85        faster than linear prefix searches.
86    
87    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
88        as for 'fr_FR' and 'french'. For some reason, however, it then used the
89        Windows-specific input and output files, which have 'french' screwed in.
90        So this could never have worked. One of the problems with locales is that
91        they aren't always the same. I have now updated RunTest so that it checks
92        the output of the locale test (test 3) against three different output
93        files, and it allows the test to pass if any one of them matches. With luck
94        this should make the test pass on some versions of Solaris where it was
95        failing. Because of the uncertainty, the script did not used to stop if
96        test 3 failed; it now does. If further versions of a French locale ever
97        come to light, they can now easily be added.
98    
99    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
100        there was a message during ./configure, but it did not stop. This now
101        provokes an error. The invalid example in README has been corrected.
102        If a value less than the minimum is given, the minimum value has always
103        been used, but now a warning is given.
104    
105    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
106        was a bug in the test system, which is now fixed. Also, the list of various
107        configurations that are tested for each release did not have one with both
108        16/32 bits and --enable-bar-anycrlf. It now does.
109    
110    18. pcretest was missing "-C bsr" for displaying the \R default setting.
111    
112    19. Little endian PowerPC systems are supported now by the JIT compiler.
113    
114    20. The fast forward newline mechanism could enter to an infinite loop on
115        certain invalid UTF-8 input. Although we don't support these cases
116        this issue can be fixed by a performance optimization.
117    
118    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
119        not take account if existing stack usage. There is now a new global
120        variable called pcre_stack_guard that can be set to point to an external
121        function to check stack availability. It is called at the start of
122        processing every parenthesized group.
123    
124    22. A typo in the code meant that in ungreedy mode the max/min qualifier
125        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
126        not match "ab".
127    
128    23. When UTF was disabled, the JIT program reported some incorrect compile
129        errors. These messages are silenced now.
130    
131    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
132        compiler.
133    
134    25. Change all the temporary files used in RunGrepTest to be different to those
135        used by RunTest so that the tests can be run simultaneously, for example by
136        "make -j check".
137    
138    
139    Version 8.34 15-December-2013
140    -----------------------------
141    
142    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
143        executable memory. Patch inspired by Carsten Klein.
144    
145    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
146        this macro is never tested and has no effect, because the work to support
147        coverage involves only compiling and linking options and special targets in
148        the Makefile. The comment in config.h implied that defining the macro would
149        enable coverage support, which is totally false. There was also support for
150        setting this macro in the CMake files (my fault, I just copied it from
151        configure). SUPPORT_GCOV has now been removed.
152    
153    3.  Make a small performance improvement in strlen16() and strlen32() in
154        pcretest.
155    
156    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
157        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
158        these cases). There was also one in pcretest.c.
159    
160    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
161    
162    6.  In UTF mode, the code for checking whether a group could match an empty
163        string (which is used for indefinitely repeated groups to allow for
164        breaking an infinite loop) was broken when the group contained a repeated
165        negated single-character class with a character that occupied more than one
166        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
167        UTF-8 mode). The effect was undefined: the group might or might not be
168        deemed as matching an empty string, or the program might have crashed.
169    
170    7.  The code for checking whether a group could match an empty string was not
171        recognizing that \h, \H, \v, \V, and \R must match a character.
172    
173    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
174        an empty string. If it can, pcretest shows this in its information output.
175    
176    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
177        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
178        matched by pcre_exec() without using JIT:
179    
180        (a) If the rest of the pattern did not match after a maximal run of
181            grapheme clusters, the code for backing up to try with fewer of them
182            did not always back up over a full grapheme when characters that do not
183            have the modifier quality were involved, e.g. Hangul syllables.
184    
185        (b) If the match point in a subject started with modifier character, and
186            there was no match, the code could incorrectly back up beyond the match
187            point, and potentially beyond the first character in the subject,
188            leading to a segfault or an incorrect match result.
189    
190    10. A conditional group with an assertion condition could lead to PCRE
191        recording an incorrect first data item for a match if no other first data
192        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
193        first data item, and therefore matched "ca" after "c" instead of at the
194        start.
195    
196    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
197        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
198    
199    12. The source of pcregrep now includes z/OS-specific code so that it can be
200        compiled for z/OS as part of the special z/OS distribution.
201    
202    13. Added the -T and -TM options to pcretest.
203    
204    14. The code in pcre_compile.c for creating the table of named capturing groups
205        has been refactored. Instead of creating the table dynamically during the
206        actual compiling pass, the information is remembered during the pre-compile
207        pass (on the stack unless there are more than 20 named groups, in which
208        case malloc() is used) and the whole table is created before the actual
209        compile happens. This has simplified the code (it is now nearly 150 lines
210        shorter) and prepared the way for better handling of references to groups
211        with duplicate names.
212    
213    15. A back reference to a named subpattern when there is more than one of the
214        same name now checks them in the order in which they appear in the pattern.
215        The first one that is set is used for the reference. Previously only the
216        first one was inspected. This change makes PCRE more compatible with Perl.
217    
218    16. Unicode character properties were updated from Unicode 6.3.0.
219    
220    17. The compile-time code for auto-possessification has been refactored, based
221        on a patch by Zoltan Herczeg. It now happens after instead of during
222        compilation. The code is cleaner, and more cases are handled. The option
223        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
224        options in pcretest are provided to set it. It can also be set by
225        (*NO_AUTO_POSSESS) at the start of a pattern.
226    
227    18. The character VT has been added to the default ("C" locale) set of
228        characters that match \s and are generally treated as white space,
229        following this same change in Perl 5.18. There is now no difference between
230        "Perl space" and "POSIX space". Whether VT is treated as white space in
231        other locales depends on the locale.
232    
233    19. The code for checking named groups as conditions, either for being set or
234        for being recursed, has been refactored (this is related to 14 and 15
235        above). Processing unduplicated named groups should now be as fast at
236        numerical groups, and processing duplicated groups should be faster than
237        before.
238    
239    20. Two patches to the CMake build system, by Alexander Barkov:
240    
241          (1) Replace the "source" command by "." in CMakeLists.txt because
242              "source" is a bash-ism.
243    
244          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
245              without these the CMake build does not work on Solaris.
246    
247    21. Perl has changed its handling of \8 and \9. If there is no previously
248        encountered capturing group of those numbers, they are treated as the
249        literal characters 8 and 9 instead of a binary zero followed by the
250        literals. PCRE now does the same.
251    
252    22. Following Perl, added \o{} to specify codepoints in octal, making it
253        possible to specify values greater than 0777 and also making them
254        unambiguous.
255    
256    23. Perl now gives an error for missing closing braces after \x{... instead of
257        treating the string as literal. PCRE now does the same.
258    
259    24. RunTest used to grumble if an inappropriate test was selected explicitly,
260        but just skip it when running all tests. This make it awkward to run ranges
261        of tests when one of them was inappropriate. Now it just skips any
262        inappropriate tests, as it always did when running all tests.
263    
264    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
265        character types such as \d or \w, too many callouts were inserted, and the
266        data that they returned was rubbish.
267    
268    26. In UCP mode, \s was not matching two of the characters that Perl matches,
269        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
270        were matched by \h. The code has now been refactored so that the lists of
271        the horizontal and vertical whitespace characters used for \h and \v (which
272        are defined only in one place) are now also used for \s.
273    
274    27. Add JIT support for the 64 bit TileGX architecture.
275        Patch by Jiong Wang (Tilera Corporation).
276    
277    28. Possessive quantifiers for classes (both explicit and automatically
278        generated) now use special opcodes instead of wrapping in ONCE brackets.
279    
280    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
281        (because it's meaningless), this was not happening when PCRE_CASELESS was
282        set. Not wrong, but inefficient.
283    
284    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
285        properties for \w, \d, etc) is present in a test regex. Otherwise if the
286        test contains no characters greater than 255, Perl doesn't realise it
287        should be using Unicode semantics.
288    
289    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
290        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
291        does in Unicode mode.
292    
293    32. Added the "forbid" facility to pcretest so that putting tests into the
294        wrong test files can sometimes be quickly detected.
295    
296    33. There is now a limit (default 250) on the depth of nesting of parentheses.
297        This limit is imposed to control the amount of system stack used at compile
298        time. It can be changed at build time by --with-parens-nest-limit=xxx or
299        the equivalent in CMake.
300    
301    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
302        errors. Perl warns for these when in warning mode, but PCRE has no facility
303        for giving warnings.
304    
305    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
306        However, this was not working for (?!) because it is optimized to (*FAIL),
307        for which PCRE does not allow quantifiers. The optimization is now disabled
308        when a quantifier follows (?!). I can't see any use for this, but it makes
309        things uniform.
310    
311    36. Perl no longer allows group names to start with digits, so I have made this
312        change also in PCRE. It simplifies the code a bit.
313    
314    37. In extended mode, Perl ignores spaces before a + that indicates a
315        possessive quantifier. PCRE allowed a space before the quantifier, but not
316        before the possessive +. It now does.
317    
318    38. The use of \K (reset reported match start) within a repeated possessive
319        group such as (a\Kb)*+ was not working.
320    
321    40. Document that the same character tables must be used at compile time and
322        run time, and that the facility to pass tables to pcre_exec() and
323        pcre_dfa_exec() is for use only with saved/restored patterns.
324    
325    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
326        features for Builds with MSVC:
327    
328        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
329           builds). The .rc files can be used to set FileDescription and many other
330           attributes.
331    
332        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
333           This allows higher-level build scripts which want .pdb files to avoid
334           hard-coding the exact files needed."
335    
336    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
337        mean "start of word" and "end of word", respectively, as a transition aid.
338    
339    43. A minimizing repeat of a class containing codepoints greater than 255 in
340        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
341        compiled to use the heap for recursion.
342    
343    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
344        is configured.
345    
346    
347    Version 8.33 28-May-2013
348    ------------------------
349    
350    1.  Added 'U' to some constants that are compared to unsigned integers, to
351        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
352        variables that are added to signed variables, to ensure the result is
353        signed and can be negated.
354    
355    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
356        CMake config files.
357    
358    3.  Revise the creation of config.h.generic so that all boolean macros are
359        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
360        overriding via -D on the command line possible.
361    
362    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
363        to unsigned int is reported to make a quite noticeable speed difference in
364        a specific Windows environment. Testing on Linux did also appear to show
365        some benefit (and it is clearly not harmful). Also fixed the definition of
366        Xop which should be unsigned.
367    
368    5.  Related to (4), changing the definition of the intermediate variable cc
369        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
370        improvements.
371    
372    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
373        unnecessary spaces.
374    
375    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
376        and later.
377    
378    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
379    
380    9.  Optimizing fast_forward_start_bits in JIT.
381    
382    10. Adding support for callouts in JIT, and fixing some issues revealed
383        during this work. Namely:
384    
385        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
386    
387        (b) Minimum length was not checked before the matching is started.
388    
389    11. The value of capture_last that is passed to callouts was incorrect in some
390        cases when there was a capture on one path that was subsequently abandoned
391        after a backtrack. Also, the capture_last value is now reset after a
392        recursion, since all captures are also reset in this case.
393    
394    12. The interpreter no longer returns the "too many substrings" error in the
395        case when an overflowing capture is in a branch that is subsequently
396        abandoned after a backtrack.
397    
398    13. In the pathological case when an offset vector of size 2 is used, pcretest
399        now prints out the matched string after a yield of 0 or 1.
400    
401    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
402        Only supported by the JIT compiler at the moment.
403    
404    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
405    
406    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
407        offset of the starting point of the matching process, provided the offsets
408        vector is large enough.
409    
410    17. The \A escape now records a lookbehind value of 1, though its execution
411        does not actually inspect the previous character. This is to ensure that,
412        in partial multi-segment matching, at least one character from the old
413        segment is retained when a new segment is processed. Otherwise, if there
414        are no lookbehinds in the pattern, \A might match incorrectly at the start
415        of a new segment.
416    
417    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
418    
419    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
420        gives some modest performance improvement in 8-bit mode.
421    
422    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
423        be expressed in certain programming languages using Universal Character
424        Names.
425    
426    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
427        which points out that "non characters" are not "characters that may not
428        appear in Unicode strings" but rather "characters that are reserved for
429        internal use and have only local meaning".
430    
431    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
432        there was a conditional group that depended on an assertion, if the
433        assertion was false, the callout that immediately followed the alternation
434        in the condition was skipped when pcre_exec() was used for matching.
435    
436    23. Allow an explicit callout to be inserted before an assertion that is the
437        condition for a conditional group, for compatibility with automatic
438        callouts, which always insert a callout at this point.
439    
440    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
441        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
442    
443    25. (*PRUNE) is now supported by the JIT compiler.
444    
445    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
446    
447    27. Fix the case where there are two or more SKIPs with arguments that may be
448        ignored.
449    
450    28. (*SKIP) is now supported by the JIT compiler.
451    
452    29. (*THEN) is now supported by the JIT compiler.
453    
454    30. Update RunTest with additional test selector options.
455    
456    31. The way PCRE handles backtracking verbs has been changed in two ways.
457    
458        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
459        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
460        backtracking. In some cases this makes it more Perl-compatible, but Perl's
461        rather obscure rules do not always do the same thing.
462    
463        (2) Previously, backtracking verbs were confined within assertions. This is
464        no longer the case for positive assertions, except for (*ACCEPT). Again,
465        this sometimes improves Perl compatibility, and sometimes does not.
466    
467    32. A number of tests that were in test 2 because Perl did things differently
468        have been moved to test 1, because either Perl or PCRE has changed, and
469        these tests are now compatible.
470    
471    32. Backtracking control verbs are now handled in the same way in JIT and
472        interpreter.
473    
474    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
475        contained a forward subroutine reference caused a compile error.
476    
477    34. Auto-detect and optimize limited repetitions in JIT.
478    
479    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
480        blocking (*UTF) etc.
481    
482    36. In the interpreter, maximizing pattern repetitions for characters and
483        character types now use tail recursion, which reduces stack usage.
484    
485    37. The value of the max lookbehind was not correctly preserved if a compiled
486        and saved regex was reloaded on a host of different endianness.
487    
488    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
489        of the compiled pattern block, expand the flags field from 16 to 32 bits
490        because it was almost full.
491    
492    39. Try madvise first before posix_madvise.
493    
494    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
495        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
496        original change locked out all matches of empty strings. This has been
497        changed so that one match of an empty string per line is recognized.
498        Subsequent searches on the same line (for colouring or for --only-matching,
499        for example) do not recognize empty strings.
500    
501    41. Applied a user patch to fix a number of spelling mistakes in comments.
502    
503    42. Data lines longer than 65536 caused pcretest to crash.
504    
505    43. Clarified the data type for length and startoffset arguments for pcre_exec
506        and pcre_dfa_exec in the function-specific man pages, where they were
507        explicitly stated to be in bytes, never having been updated. I also added
508        some clarification to the pcreapi man page.
509    
510    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
511        a segmentation fault.
512    
513    
514    Version 8.32 30-November-2012
515    -----------------------------
516    
517    1.  Improved JIT compiler optimizations for first character search and single
518        character iterators.
519    
520    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
521        Patch by Daniel Richard G.
522    
523    3.  Single character iterator optimizations in the JIT compiler.
524    
525    4.  Improved JIT compiler optimizations for character ranges.
526    
527    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
528        Reported by Giuseppe D'Angelo.
529    
530    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
531        of a line, was being set incorrectly in cases where .* appeared inside
532        atomic brackets at the start of a pattern, or where there was a subsequent
533        *PRUNE or *SKIP.
534    
535    7.  Improved instruction cache flush for POWER/PowerPC.
536        Patch by Daniel Richard G.
537    
538    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
539        grep:
540    
541        (a) There is now no limit to the number of patterns to be matched.
542    
543        (b) An error is given if a pattern is too long.
544    
545        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
546            are now supported.
547    
548        (d) --exclude-from and --include-from (multiple use) have been added.
549    
550        (e) Exclusions and inclusions now apply to all files and directories, not
551            just to those obtained from scanning a directory recursively.
552    
553        (f) Multiple uses of -f and --file-list are now supported.
554    
555        (g) In a Windows environment, the default for -d has been changed from
556            "read" (the GNU grep default) to "skip", because otherwise the presence
557            of a directory in the file list provokes an error.
558    
559        (h) The documentation has been revised and clarified in places.
560    
561    9.  Improve the matching speed of capturing brackets.
562    
563    10. Changed the meaning of \X so that it now matches a Unicode extended
564        grapheme cluster.
565    
566    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
567        out POSIX threads when JIT support is configured.
568    
569    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
570    
571    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
572        structure could go wrong in environments where size_t is not the same size
573        as int.
574    
575    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
576        set.
577    
578    15. The EBCDIC support had decayed; later updates to the code had included
579        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
580        general tidy up of EBCDIC-related issues, and the documentation was also
581        not quite right. There is now a test that can be run on ASCII systems to
582        check some of the EBCDIC-related things (but is it not a full test).
583    
584    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
585        in a small tidy to the code.
586    
587    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
588    
589    18. If the --only-matching (-o) option in pcregrep is specified multiple
590        times, each one causes appropriate output. For example, -o1 -o2 outputs the
591        substrings matched by the 1st and 2nd capturing parentheses. A separating
592        string can be specified by --om-separator (default empty).
593    
594    19. Improving the first n character searches.
595    
596    20. Turn case lists for horizontal and vertical white space into macros so that
597        they are defined only once.
598    
599    21. This set of changes together give more compatible Unicode case-folding
600        behaviour for characters that have more than one other case when UCP
601        support is available.
602    
603        (a) The Unicode property table now has offsets into a new table of sets of
604            three or more characters that are case-equivalent. The MultiStage2.py
605            script that generates these tables (the pcre_ucd.c file) now scans
606            CaseFolding.txt instead of UnicodeData.txt for character case
607            information.
608    
609        (b) The code for adding characters or ranges of characters to a character
610            class has been abstracted into a generalized function that also handles
611            case-independence. In UTF-mode with UCP support, this uses the new data
612            to handle characters with more than one other case.
613    
614        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
615            whose other case is greater than 256 are now correctly matched
616            caselessly. Previously, the high codepoint matched the low one, but not
617            vice versa.
618    
619        (d) The processing of \h, \H, \v, and \ in character classes now makes use
620            of the new class addition function, using character lists defined as
621            macros alongside the case definitions of 20 above.
622    
623        (e) Caseless back references now work with characters that have more than
624            one other case.
625    
626        (f) General caseless matching of characters with more than one other case
627            is supported.
628    
629    22. Unicode character properties were updated from Unicode 6.2.0
630    
631    23. Improved CMake support under Windows. Patch by Daniel Richard G.
632    
633    24. Add support for 32-bit character strings, and UTF-32
634    
635    25. Major JIT compiler update (code refactoring and bugfixing).
636        Experimental Sparc 32 support is added.
637    
638    26. Applied a modified version of Daniel Richard G's patch to create
639        pcre.h.generic and config.h.generic by "make" instead of in the
640        PrepareRelease script.
641    
642    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
643        pcre_compile.c when checking for a zero character.
644    
645    28. Introducing a native interface for JIT. Through this interface, the compiled
646        machine code can be directly executed. The purpose of this interface is to
647        provide fast pattern matching, so several sanity checks are not performed.
648        However, feature tests are still performed. The new interface provides
649        1.4x speedup compared to the old one.
650    
651    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
652        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
653        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
654    
655    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
656        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
657        the "old" RFC 2279). Instead, it ended up passing a negative length to
658        pcre_exec().
659    
660    31. Add support for GCC's visibility feature to hide internal functions.
661    
662    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
663        "unknown -C option" after outputting 0 or 1.
664    
665    33. There is now support for generating a code coverage report for the test
666        suite in environments where gcc is the compiler and lcov is installed. This
667        is mainly for the benefit of the developers.
668    
669    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
670        unaddressable using valgrind annotations, allowing valgrind to detect
671        invalid memory accesses. This is mainly for the benefit of the developers.
672    
673    25. (*UTF) can now be used to start a pattern in any of the three libraries.
674    
675    26. Give configure error if --enable-cpp but no C++ compiler found.
676    
677    
678    Version 8.31 06-July-2012
679    -------------------------
680    
681    1.  Fixing a wrong JIT test case and some compiler warnings.
682    
683    2.  Removed a bashism from the RunTest script.
684    
685    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
686        to unsigned type, result still unsigned" that was given by an MS compiler
687        on encountering the code "-sizeof(xxx)".
688    
689    4.  Partial matching support is added to the JIT compiler.
690    
691    5.  Fixed several bugs concerned with partial matching of items that consist
692        of more than one character:
693    
694        (a) /^(..)\1/ did not partially match "aba" because checking references was
695            done on an "all or nothing" basis. This also applied to repeated
696            references.
697    
698        (b) \R did not give a hard partial match if \r was found at the end of the
699            subject.
700    
701        (c) \X did not give a hard partial match after matching one or more
702            characters at the end of the subject.
703    
704        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
705            a partial match for the string "\r".
706    
707        (e) When newline was set to CRLF, the metacharacter "." did not recognize
708            a partial match for a CR character at the end of the subject string.
709    
710    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
711        running pcretest, the text "(JIT)" added to the output whenever JIT is
712        actually used to run the match.
713    
714    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
715        or /S+[+] with a digit between 1 and 7.
716    
717    8.  OP_NOT now supports any UTF character not just single-byte ones.
718    
719    9.  (*MARK) control verb is now supported by the JIT compiler.
720    
721    10. The command "./RunTest list" lists the available tests without actually
722        running any of them. (Because I keep forgetting what they all are.)
723    
724    11. Add PCRE_INFO_MAXLOOKBEHIND.
725    
726    12. Applied a (slightly modified) user-supplied patch that improves performance
727        when the heap is used for recursion (compiled with --disable-stack-for-
728        recursion). Instead of malloc and free for each heap frame each time a
729        logical recursion happens, frames are retained on a chain and re-used where
730        possible. This sometimes gives as much as 30% improvement.
731    
732    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
733        call.
734    
735    14. As documented, (*COMMIT) is now confined to within a positive assertion.
736    
737    15. It is now possible to link pcretest with libedit as an alternative to
738        libreadline.
739    
740    16. (*COMMIT) control verb is now supported by the JIT compiler.
741    
742    17. The Unicode data tables have been updated to Unicode 6.1.0.
743    
744    18. Added --file-list option to pcregrep.
745    
746    19. Added binary file support to pcregrep, including the -a, --binary-files,
747        -I, and --text options.
748    
749    20. The madvise function is renamed for posix_madvise for QNX compatibility
750        reasons. Fixed by Giuseppe D'Angelo.
751    
752    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
753    
754    22. Changed the option for creating symbolic links for 16-bit man pages from
755        -s to -sf so that re-installing does not cause issues.
756    
757    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
758    
759    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
760        matches in certain environments (the workspace was not being correctly
761        retained). Also added to pcre_dfa_exec() a simple plausibility check on
762        some of the workspace data at the beginning of a restart.
763    
764    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
765        was not doing so when it should - probably a typo introduced by SVN 528
766        (change 8.10/14).
767    
768    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
769        \w+ when the character tables indicated that \x{c4} was a word character.
770        There were several related cases, all because the tests for doing a table
771        lookup were testing for characters less than 127 instead of 255.
772    
773    27. If a pattern contains capturing parentheses that are not used in a match,
774        their slots in the ovector are set to -1. For those that are higher than
775        any matched groups, this happens at the end of processing. In the case when
776        there were back references that the ovector was too small to contain
777        (causing temporary malloc'd memory to be used during matching), and the
778        highest capturing number was not used, memory off the end of the ovector
779        was incorrectly being set to -1. (It was using the size of the temporary
780        memory instead of the true size.)
781    
782    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
783        ovector size, it uses memory at the end of the block that it has got.
784    
785    29. Check for an overlong MARK name and give an error at compile time. The
786        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
787    
788    30. JIT compiler update.
789    
790    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
791        Rill for the patch.
792    
793    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
794    
795    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
796    
797    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
798        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
799        was enabled.
800    
801    35. Improve JIT code generation for greedy plus quantifier.
802    
803    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
804        1 to "aa" instead of to an empty string. The bug affected repeated groups
805        that could potentially match an empty string.
806    
807    37. Optimizing single character iterators in JIT.
808    
809    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
810        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
811        codepoints that are too big for the mode are faulted, and in a UTF mode,
812        disallowed codepoints are also faulted.
813    
814    39. If PCRE was compiled with UTF support, in three places in the DFA
815        matcher there was code that should only have been obeyed in UTF mode, but
816        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
817        processing when bytes with values greater than 127 were present. In 16-bit
818        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
819        both cases the values are those that cannot be the first data item in a UTF
820        character. The three items that might have provoked this were recursions,
821        possessively repeated groups, and atomic groups.
822    
823    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
824        and pcregrep, because some OS require shared objects to be explicitly
825        passed to ld, causing the link step to fail if they are not.
826    
827    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
828        mode, patterns that started with \h* or \R* might be incorrectly matched.
829    
830    
831    Version 8.30 04-February-2012
832    -----------------------------
833    
834    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
835        name is defined in ctype.h.
836    
837    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
838        only in quite long subpatterns.
839    
840    3.  Removed the function pcre_info(), which has been obsolete and deprecated
841        since it was replaced by pcre_fullinfo() in February 2000.
842    
843    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
844        match a (*MARK), and the match failed at the start of the subject, a
845        reference to memory before the start of the subject could occur. This bug
846        was introduced by fix 17 of release 8.21.
847    
848    5.  A reference to an unset group with zero minimum repetition was giving
849        totally wrong answers (in non-JavaScript-compatibility mode). For example,
850        /(another)?(\1?)test/ matched against "hello world test". This bug was
851        introduced in release 8.13.
852    
853    6.  Add support for 16-bit character strings (a large amount of work involving
854        many changes and refactorings).
855    
856    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
857        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
858        from a file.
859    
860    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
861        rounding is not applied in this particular case).
862    
863    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
864        if they appear, or are escaped, in patterns.
865    
866    10. Get rid of a number of -Wunused-but-set-variable warnings.
867    
868    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
869        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
870        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
871        also returns the mark "x". This bug applied to capturing parentheses,
872        non-capturing parentheses, and atomic parentheses. It also applied to some
873        assertions.
874    
875    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
876        information out of configure.ac instead of relying on pcre.h.generic, which
877        is not stored in the repository.
878    
879    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
880        -lreadline.
881    
882    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
883    
884    15. Applied Graycode's patch to put the top-level frame on the stack rather
885        than the heap when not using the stack for recursion. This gives a
886        performance improvement in many cases when recursion is not deep.
887    
888    16. Experimental code added to "pcretest -C" to output the stack frame size.
889    
890    
891    Version 8.21 12-Dec-2011
892    ------------------------
893    
894    1.  Updating the JIT compiler.
895    
896    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
897        are added as well.
898    
899    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
900        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
901        calling _pcre_jit_exec. Some extra comments are added.
902    
903    4.  (*MARK) settings inside atomic groups that do not contain any capturing
904        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
905        was introduced by change 18 for 8.20.
906    
907    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
908        ECMA-262 standard.
909    
910    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
911        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
912        This bug was probably introduced by change 9 of 8.13.
913    
914    7.  While fixing 6 above, I noticed that a number of other items were being
915        incorrectly rejected as "not fixed length". This arose partly because newer
916        opcodes had not been added to the fixed-length checking code. I have (a)
917        corrected the bug and added tests for these items, and (b) arranged for an
918        error to occur if an unknown opcode is encountered while checking for fixed
919        length instead of just assuming "not fixed length". The items that were
920        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
921        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
922        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
923    
924    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
925        being incorrectly compiled and would have given unpredicatble results.
926    
927    9.  A possessively repeated subpattern with minimum repeat count greater than
928        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
929        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
930        the first (A) could occur when it should not.
931    
932    10. Add a cast and remove a redundant test from the code.
933    
934    11. JIT should use pcre_malloc/pcre_free for allocation.
935    
936    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
937        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
938        is anything other than /usr, -L is still shown).
939    
940    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
941    
942    14. Perl does not support \N without a following name in a [] class; PCRE now
943        also gives an error.
944    
945    15. If a forward reference was repeated with an upper limit of around 2000,
946        it caused the error "internal error: overran compiling workspace". The
947        maximum number of forward references (including repeats) was limited by the
948        internal workspace, and dependent on the LINK_SIZE. The code has been
949        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
950        the default depends on LINK_SIZE. There is a new upper limit (for safety)
951        of around 200,000 forward references. While doing this, I also speeded up
952        the filling in of repeated forward references.
953    
954    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
955        incorrectly expecting the subject to contain another "a" after the start.
956    
957    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
958        in the match, the SKIP should be ignored. This was not happening; instead
959        the SKIP was being treated as NOMATCH. For patterns such as
960        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
961        tested.
962    
963    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
964        now much more compatible with Perl, in particular in cases where the result
965        is a non-match for a non-anchored pattern. For example, if
966        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
967        "m", where previously it did not return a name. A side effect of this
968        change is that for partial matches, the last encountered mark name is
969        returned, as for non matches. A number of tests that were previously not
970        Perl-compatible have been moved into the Perl-compatible test files. The
971        refactoring has had the pleasing side effect of removing one argument from
972        the match() function, thus reducing its stack requirements.
973    
974    19. If the /S+ option was used in pcretest to study a pattern using JIT,
975        subsequent uses of /S (without +) incorrectly behaved like /S+.
976    
977    21. Retrieve executable code size support for the JIT compiler and fixing
978        some warnings.
979    
980    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
981        not work when the shorter character appeared right at the end of the
982        subject string.
983    
984    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
985        systems.
986    
987    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
988        output it when the /M option is used in pcretest.
989    
990    25. The CheckMan script was not being included in the distribution. Also, added
991        an explicit "perl" to run Perl scripts from the PrepareRelease script
992        because this is reportedly needed in Windows.
993    
994    26. If study data was being save in a file and studying had not found a set of
995        "starts with" bytes for the pattern, the data written to the file (though
996        never used) was taken from uninitialized memory and so caused valgrind to
997        complain.
998    
999    27. Updated RunTest.bat as provided by Sheri Pierce.
1000    
1001    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1002    
1003    29. Computation of memory usage for the table of capturing group names was
1004        giving an unnecessarily large value.
1005    
1006    
1007    Version 8.20 21-Oct-2011
1008    ------------------------
1009    
1010    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1011        a POSIX class. After further experiments with Perl, which convinced me that
1012        Perl has bugs and confusions, a closing square bracket is no longer allowed
1013        in a POSIX name. This bug also affected patterns with classes that started
1014        with full stops.
1015    
1016    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1017        captured substring, but while checking the failing first alternative,
1018        substring 1 is temporarily captured. If the output vector supplied to
1019        pcre_exec() was not big enough for this capture, the yield of the function
1020        was still zero ("insufficient space for captured substrings"). This cannot
1021        be totally fixed without adding another stack variable, which seems a lot
1022        of expense for a edge case. However, I have improved the situation in cases
1023        such as /(a)(b)x|abc/ matched against "abc", where the return code
1024        indicates that fewer than the maximum number of slots in the ovector have
1025        been set.
1026    
1027    3.  Related to (2) above: when there are more back references in a pattern than
1028        slots in the output vector, pcre_exec() uses temporary memory during
1029        matching, and copies in the captures as far as possible afterwards. It was
1030        using the entire output vector, but this conflicts with the specification
1031        that only 2/3 is used for passing back captured substrings. Now it uses
1032        only the first 2/3, for compatibility. This is, of course, another edge
1033        case.
1034    
1035    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1036        main code base, and can be used by building with --enable-jit. When this is
1037        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1038        runtime --no-jit option is given.
1039    
1040    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1041        ovector, the return from the function was zero, implying that there were
1042        other matches that did not fit. The correct "exactly full" value is now
1043        returned.
1044    
1045    6.  If a subpattern that was called recursively or as a subroutine contained
1046        (*PRUNE) or any other control that caused it to give a non-standard return,
1047        invalid errors such as "Error -26 (nested recursion at the same subject
1048        position)" or even infinite loops could occur.
1049    
1050    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1051        computing the minimum length on reaching *ACCEPT, and so ended up with the
1052        wrong value of 1 rather than 0. Further investigation indicates that
1053        computing a minimum subject length in the presence of *ACCEPT is difficult
1054        (think back references, subroutine calls), and so I have changed the code
1055        so that no minimum is registered for a pattern that contains *ACCEPT.
1056    
1057    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1058        it was not handled as intended. [But see 16 below.]
1059    
1060    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1061        Sheri Pierce.
1062    
1063    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1064        the first byte in a match must be "a".
1065    
1066    11. Change 17 for 8.13 increased the recursion depth for patterns like
1067        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1068        pattern contains any instances of (*THEN). If it does not, the old
1069        optimizations are restored. It would be nice to do this on a per-group
1070        basis, but at the moment that is not feasible.
1071    
1072    12. In some environments, the output of pcretest -C is CRLF terminated. This
1073        broke RunTest's code that checks for the link size. A single white space
1074        character after the value is now allowed for.
1075    
1076    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1077        For "fr", it uses the Windows-specific input and output files.
1078    
1079    14. If (*THEN) appeared in a group that was called recursively or as a
1080        subroutine, it did not work as intended. [But see next item.]
1081    
1082    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1083        pattern fragments (but not containing any | characters). If A and B are
1084        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1085        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1086        D. In other words, Perl considers parentheses that do not contain any |
1087        characters to be part of a surrounding alternative, whereas PCRE was
1088        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1089        differently. PCRE now behaves in the same way as Perl, except in the case
1090        of subroutine/recursion calls such as (?1) which have in any case always
1091        been different (but PCRE had them first :-).
1092    
1093    16. Related to 15 above: Perl does not treat the | in a conditional group as
1094        creating alternatives. Such a group is treated in the same way as an
1095        ordinary group without any | characters when processing (*THEN). PCRE has
1096        been changed to match Perl's behaviour.
1097    
1098    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1099        RunGrepTest script failed.
1100    
1101    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1102        inevitable for groups that contain captures, but it can lead to a lot of
1103        stack use in large patterns. The old behaviour has been restored for atomic
1104        groups that do not contain any capturing parentheses.
1105    
1106    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1107        suppress the check for a minimum subject length at run time. (If it was
1108        given to pcre_exec() or pcre_dfa_exec() it did work.)
1109    
1110    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1111        fail to work when decoding hex characters in data strings in EBCDIC
1112        environments.
1113    
1114    21. It appears that in at least one Mac OS environment, the isxdigit() function
1115        is implemented as a macro that evaluates to its argument more than once,
1116        contravening the C 90 Standard (I haven't checked a later standard). There
1117        was an instance in pcretest which caused it to go wrong when processing
1118        \x{...} escapes in subject strings. The has been rewritten to avoid using
1119        things like p++ in the argument of isxdigit().
1120    
1121    
1122    Version 8.13 16-Aug-2011
1123    ------------------------
1124    
1125    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1126    
1127    2.  Two minor typos in pcre_internal.h have been fixed.
1128    
1129    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1130        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1131        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1132    
1133    4.  There were a number of related bugs in the code for matching backrefences
1134        caselessly in UTF-8 mode when codes for the characters concerned were
1135        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1136        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1137        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1138        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1139        2-byte code at the end of the subject (it thought there wasn't enough data
1140        left).
1141    
1142    5.  Comprehensive information about what went wrong is now returned by
1143        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1144        as the output vector has at least 2 elements. The offset of the start of
1145        the failing character and a reason code are placed in the vector.
1146    
1147    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1148        now returned is for the first byte of the failing character, instead of the
1149        last byte inspected. This is an incompatible change, but I hope it is small
1150        enough not to be a problem. It makes the returned offset consistent with
1151        pcre_exec() and pcre_dfa_exec().
1152    
1153    7.  pcretest now gives a text phrase as well as the error number when
1154        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1155        failure, the offset and reason code are output.
1156    
1157    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1158        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1159        back over a single character (\n). This seems wrong (because it treated the
1160        two characters as a single entity when going forwards), conflicts with the
1161        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1162        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1163        has been changed.
1164    
1165    9.  Some internal refactoring has changed the processing so that the handling
1166        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1167        time (the PCRE_DOTALL option was changed this way some time ago: version
1168        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1169        which was always a bit of a fudge. It also means that there is one less
1170        argument for the match() function, which reduces its stack requirements
1171        slightly. This change also fixes an incompatibility with Perl: the pattern
1172        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1173    
1174    10. More internal refactoring has drastically reduced the number of recursive
1175        calls to match() for possessively repeated groups such as (abc)++ when
1176        using pcre_exec().
1177    
1178    11. While implementing 10, a number of bugs in the handling of groups were
1179        discovered and fixed:
1180    
1181        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1182        (a|)*(?1) gave a compile-time internal error.
1183        ((a|)+)+  did not notice that the outer group could match an empty string.
1184        (^a|^)+   was not marked as anchored.
1185        (.*a|.*)+ was not marked as matching at start or after a newline.
1186    
1187    12. Yet more internal refactoring has removed another argument from the match()
1188        function. Special calls to this function are now indicated by setting a
1189        value in a variable in the "match data" data block.
1190    
1191    13. Be more explicit in pcre_study() instead of relying on "default" for
1192        opcodes that mean there is no starting character; this means that when new
1193        ones are added and accidentally left out of pcre_study(), testing should
1194        pick them up.
1195    
1196    14. The -s option of pcretest has been documented for ages as being an old
1197        synonym of -m (show memory usage). I have changed it to mean "force study
1198        for every regex", that is, assume /S for every regex. This is similar to -i
1199        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1200        using it. It makes it easier to run collections of tests with and without
1201        study enabled, and thereby test pcre_study() more easily. All the standard
1202        tests are now run with and without -s (but some patterns can be marked as
1203        "never study" - see 20 below).
1204    
1205    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1206        restoration of the capturing data to the outer values was not happening
1207        correctly.
1208    
1209    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1210        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1211        pattern had matched an empty string, and so incorrectly returned a no
1212        match.
1213    
1214    17. There was optimizing code for the last branch of non-capturing parentheses,
1215        and also for the obeyed branch of a conditional subexpression, which used
1216        tail recursion to cut down on stack usage. Unfortunately, now that there is
1217        the possibility of (*THEN) occurring in these branches, tail recursion is
1218        no longer possible because the return has to be checked for (*THEN). These
1219        two optimizations have therefore been removed. [But see 8.20/11 above.]
1220    
1221    18. If a pattern containing \R was studied, it was assumed that \R always
1222        matched two bytes, thus causing the minimum subject length to be
1223        incorrectly computed because \R can also match just one byte.
1224    
1225    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1226        was incorrectly computed.
1227    
1228    20. If /S is present twice on a test pattern in pcretest input, it now
1229        *disables* studying, thereby overriding the use of -s on the command line
1230        (see 14 above). This is necessary for one or two tests to keep the output
1231        identical in both cases.
1232    
1233    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1234        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1235    
1236    22. When an atomic group that contained a capturing parenthesis was
1237        successfully matched, but the branch in which it appeared failed, the
1238        capturing was not being forgotten if a higher numbered group was later
1239        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1240        group 1 to "a", when in fact it should be unset. This applied to multi-
1241        branched capturing and non-capturing groups, repeated or not, and also to
1242        positive assertions (capturing in negative assertions does not happen
1243        in PCRE) and also to nested atomic groups.
1244    
1245    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1246        subject after a captured substring, to make it easier to tell which of a
1247        number of identical substrings has been captured.
1248    
1249    24. The way atomic groups are processed by pcre_exec() has been changed so that
1250        if they are repeated, backtracking one repetition now resets captured
1251        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1252        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1253        "aaa". Previously, it would have been "a". As part of this code
1254        refactoring, the way recursive calls are handled has also been changed.
1255    
1256    25. If an assertion condition captured any substrings, they were not passed
1257        back unless some other capturing happened later. For example, if
1258        (?(?=(a))a) was matched against "a", no capturing was returned.
1259    
1260    26. When studying a pattern that contained subroutine calls or assertions,
1261        the code for finding the minimum length of a possible match was handling
1262        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1263        group 1 called group 2 while simultaneously a separate group 2 called group
1264        1). A stack overflow occurred in this case. I have fixed this by limiting
1265        the recursion depth to 10.
1266    
1267    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1268        Fortmann. This supports explicit test numbers on the command line, and has
1269        argument validation and error reporting.
1270    
1271    28. An instance of \X with an unlimited repeat could fail if at any point the
1272        first character it looked at was a mark character.
1273    
1274    29. Some minor code refactoring concerning Unicode properties and scripts
1275        should reduce the stack requirement of match() slightly.
1276    
1277    30. Added the '=' option to pcretest to check the setting of unused capturing
1278        slots at the end of the pattern, which are documented as being -1, but are
1279        not included in the return count.
1280    
1281    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1282        compiled something random. Now it gives a compile-time error (as does
1283        Perl).
1284    
1285    32. A *MARK encountered during the processing of a positive assertion is now
1286        recorded and passed back (compatible with Perl).
1287    
1288    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1289        had alternative anchored branches, the search for a second match in a line
1290        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1291        matched the line "0102" twice. The same bug affected patterns that started
1292        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1293        twice.
1294    
1295    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1296        does, and because of capturing effects, quantifying parenthesized
1297        assertions may at times be useful. Quantifiers are now allowed for
1298        parenthesized assertions.
1299    
1300    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1301    
1302    36. \g was being checked for fancy things in a character class, when it should
1303        just be a literal "g".
1304    
1305    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1306        appearance of a nested POSIX class supersedes an apparent external class.
1307        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1308        unescaped square brackets may also appear as part of class names. For
1309        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1310        more like Perl. (But see 8.20/1 above.)
1311    
1312    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1313        was because it thought it was \N{name}, which is not supported).
1314    
1315    39. Add minix to OS list not supporting the -S option in pcretest.
1316    
1317    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1318        cannot analyze patterns in sufficient detail to catch mutual recursions
1319        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1320        subgroup is called recursively as a subpattern for a second time at the
1321        same position in the subject string. In previous releases this might have
1322        been caught by the recursion limit, or it might have run out of stack.
1323    
1324    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1325        happen only once. PCRE was, however incorrectly giving a compile time error
1326        "recursive call could loop indefinitely" because it cannot analyze the
1327        pattern in sufficient detail. The compile time test no longer happens when
1328        PCRE is compiling a conditional subpattern, but actual runaway loops are
1329        now caught at runtime (see 40 above).
1330    
1331    42. It seems that Perl allows any characters other than a closing parenthesis
1332        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1333        has been changed to be the same.
1334    
1335    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1336        as not to get warnings when autogen.sh is called. Also changed
1337        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1338    
1339    44. To help people who use pcregrep to scan files containing exceedingly long
1340        lines, the following changes have been made:
1341    
1342        (a) The default value of the buffer size parameter has been increased from
1343            8K to 20K. (The actual buffer used is three times this size.)
1344    
1345        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1346            PCRE is built.
1347    
1348        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1349            to be set at run time.
1350    
1351        (d) Numerical values in pcregrep options can be followed by K or M, for
1352            example --buffer-size=50K.
1353    
1354        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1355            given and the return code is set to 2.
1356    
1357    45. Add a pointer to the latest mark to the callout data block.
1358    
1359    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1360        partial match of an empty string instead of no match. This was specific to
1361        the use of ".".
1362    
1363    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1364        complete match instead of a partial match. This bug was dependent on both
1365        the PCRE_UTF8 and PCRE_DOTALL options being set.
1366    
1367    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1368        starting byte set, because \b was not being ignored.
1369    
1370    
1371    Version 8.12 15-Jan-2011
1372    ------------------------
1373    
1374    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1375        checks for such things as part of the documentation building process.
1376    
1377    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1378        --match-limit and --recursion-limit options (added for 8.11). In
1379        particular, this made one of the standard tests fail. (The integer value
1380        went into the wrong half of a long int.)
1381    
1382    3.  If the --colour option was given to pcregrep with -v (invert match), it
1383        did strange things, either producing crazy output, or crashing. It should,
1384        of course, ignore a request for colour when reporting lines that do not
1385        match.
1386    
1387    4.  Another pcregrep bug caused similar problems if --colour was specified with
1388        -M (multiline) and the pattern match finished with a line ending.
1389    
1390    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1391        matched in multiline mode, the following line was shown as part of the
1392        match. This seems wrong, so I have changed it.
1393    
1394    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1395        the check for further matches in the same line (so they could be coloured)
1396        to overrun the end of the current line. If another match was found, it was
1397        incorrectly shown (and then shown again when found in the next line).
1398    
1399    7.  If pcregrep was compiled under Windows, there was a reference to the
1400        function pcregrep_exit() before it was defined. I am assuming this was
1401        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1402        reported by a user. I've moved the definition above the reference.
1403    
1404    
1405    Version 8.11 10-Dec-2010
1406    ------------------------
1407    
1408    1.  (*THEN) was not working properly if there were untried alternatives prior
1409        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1410        backtracked to try for "b" instead of moving to the next alternative branch
1411        at the same level (in this case, to look for "c"). The Perl documentation
1412        is clear that when (*THEN) is backtracked onto, it goes to the "next
1413        alternative in the innermost enclosing group".
1414    
1415    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1416        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1417        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1418        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1419        (*THEN).
1420    
1421    3.  If \s appeared in a character class, it removed the VT character from
1422        the class, even if it had been included by some previous item, for example
1423        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1424        of \s, but is part of the POSIX "space" class.)
1425    
1426    4.  A partial match never returns an empty string (because you can always
1427        match an empty string at the end of the subject); however the checking for
1428        an empty string was starting at the "start of match" point. This has been
1429        changed to the "earliest inspected character" point, because the returned
1430        data for a partial match starts at this character. This means that, for
1431        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1432        (previously it gave "no match").
1433    
1434    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1435        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1436        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1437        has an implication that the given string is incomplete (because a partial
1438        match is preferred over a full match). For this reason, these items now
1439        give a partial match in this situation. [Aside: previously, the one case
1440        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1441        match rather than a full match, which was wrong by the old rules, but is
1442        now correct.]
1443    
1444    6.  There was a bug in the handling of #-introduced comments, recognized when
1445        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1446        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1447        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1448        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1449        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1450        places in pcre_compile().
1451    
1452    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1453        comments when looking ahead for named forward references to subpatterns,
1454        the only newline sequence it recognized was NL. It now handles newlines
1455        according to the set newline convention.
1456    
1457    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1458        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1459        cater for a lack of strerror(). These oversights have been fixed.
1460    
1461    9.  Added --match-limit and --recursion-limit to pcregrep.
1462    
1463    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1464    
1465    11. When the -o option was used, pcregrep was setting a return code of 1, even
1466        when matches were found, and --line-buffered was not being honoured.
1467    
1468    12. Added an optional parentheses number to the -o and --only-matching options
1469        of pcregrep.
1470    
1471    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1472        can match an empty string. The code to do it in pcretest and pcredemo
1473        needed fixing:
1474    
1475        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1476            only one byte after an empty string match just before CRLF (this case
1477            just got forgotten; "any" and "anycrlf" were OK).
1478    
1479        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1480            mode when an empty string match preceded an ASCII character followed by
1481            a non-ASCII character. (The code for advancing by one character rather
1482            than one byte was nonsense.)
1483    
1484        (c) The pcredemo.c sample program did not have any code at all to handle
1485            the cases when CRLF is a valid newline sequence.
1486    
1487    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1488        as a starting offset was within the subject string. There is now a new
1489        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1490        negative or greater than the length of the string. In order to test this,
1491        pcretest is extended to allow the setting of negative starting offsets.
1492    
1493    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1494        starting offset points to the beginning of a UTF-8 character was
1495        unnecessarily clumsy. I tidied it up.
1496    
1497    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1498        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1499    
1500    17. Nobody had reported that the --include_dir option, which was added in
1501        release 7.7 should have been called --include-dir (hyphen, not underscore)
1502        for compatibility with GNU grep. I have changed it to --include-dir, but
1503        left --include_dir as an undocumented synonym, and the same for
1504        --exclude-dir, though that is not available in GNU grep, at least as of
1505        release 2.5.4.
1506    
1507    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1508        characters from a string of bytes) have been redefined so as not to use
1509        loops, in order to improve performance in some environments. At the same
1510        time, I abstracted some of the common code into auxiliary macros to save
1511        repetition (this should not affect the compiled code).
1512    
1513    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1514        compile-time error is now given if \c is not followed by an ASCII
1515        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1516        different, and any byte value is allowed.)
1517    
1518    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1519        START_OPTIMIZE option, which is now allowed at compile time - but just
1520        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1521        to pcregrep and other applications that have no direct access to PCRE
1522        options. The new /Y option in pcretest sets this option when calling
1523        pcre_compile().
1524    
1525    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1526        back references. Groups containing recursive back references were forced to
1527        be atomic by that change, but in the case of named groups, the amount of
1528        memory required was incorrectly computed, leading to "Failed: internal
1529        error: code overflow". This has been fixed.
1530    
1531    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1532        pcretest.c, to avoid build problems in some Borland environments.
1533    
1534    
1535    Version 8.10 25-Jun-2010
1536    ------------------------
1537    
1538    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1539        THEN.
1540    
1541    2.  (*ACCEPT) was not working when inside an atomic group.
1542    
1543    3.  Inside a character class, \B is treated as a literal by default, but
1544        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1545        causes the error). The code is unchanged, but I tidied the documentation.
1546    
1547    4.  Inside a character class, PCRE always treated \R and \X as literals,
1548        whereas Perl faults them if its -w option is set. I have changed PCRE so
1549        that it faults them when PCRE_EXTRA is set.
1550    
1551    5.  Added support for \N, which always matches any character other than
1552        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1553    
1554    6.  When compiling pcregrep with newer versions of gcc which may have
1555        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1556        declared with attribute warn_unused_result" were given. Just casting the
1557        result to (void) does not stop the warnings; a more elaborate fudge is
1558        needed. I've used a macro to implement this.
1559    
1560    7.  Minor change to pcretest.c to avoid a compiler warning.
1561    
1562    8.  Added four artifical Unicode properties to help with an option to make
1563        \s etc use properties (see next item). The new properties are: Xan
1564        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1565    
1566    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1567        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1568        this option. Modified pcretest to add /W to test this facility. Added
1569        REG_UCP to make it available via the POSIX interface.
1570    
1571    10. Added --line-buffered to pcregrep.
1572    
1573    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1574        studied, and the match started with a letter with a code point greater than
1575        127 whose first byte was different to the first byte of the other case of
1576        the letter, the other case of this starting letter was not recognized
1577        (#976).
1578    
1579    12. If a pattern that was studied started with a repeated Unicode property
1580        test, for example, \p{Nd}+, there was the theoretical possibility of
1581        setting up an incorrect bitmap of starting bytes, but fortunately it could
1582        not have actually happened in practice until change 8 above was made (it
1583        added property types that matched character-matching opcodes).
1584    
1585    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1586        possible starting bytes for non-anchored patterns.
1587    
1588    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1589        \R, and also a number of cases that involve Unicode properties, both
1590        explicit and implicit when PCRE_UCP is set.
1591    
1592    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1593        input, it could crash or give wrong results if characters with values
1594        greater than 0xc0 were present in the subject string. (Detail: it assumed
1595        UTF-8 input when processing these items.)
1596    
1597    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1598        size_t is 64-bit (#991).
1599    
1600    17. Added a check for running out of memory when PCRE is compiled with
1601        --disable-stack-for-recursion (#990).
1602    
1603    18. If the last data line in a file for pcretest does not have a newline on
1604        the end, a newline was missing in the output.
1605    
1606    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1607        less than 128) in its various bitmaps. However, there is a facility for
1608        generating tables according to the current locale when PCRE is compiled. It
1609        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1610        characters, are recognized by isspace() and therefore were getting set in
1611        these tables, and indeed these tables seem to approximate to ISO 8859. This
1612        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1613        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1614        which of course cannot start UTF-8 characters. I have changed the code so
1615        that only real ASCII characters (less than 128) and the correct starting
1616        bytes for UTF-8 encodings are set for characters greater than 127 when in
1617        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1618        altogether.)
1619    
1620    20. Added the /T option to pcretest so as to be able to run tests with non-
1621        standard character tables, thus making it possible to include the tests
1622        used for 19 above in the standard set of tests.
1623    
1624    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1625        reference to a subpattern the other side of a comment that contains an
1626        opening parenthesis caused either an internal compiling error, or a
1627        reference to the wrong subpattern.
1628    
1629    
1630    Version 8.02 19-Mar-2010
1631    ------------------------
1632    
1633    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1634    
1635    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1636        configured.
1637    
1638    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1639        original author of that file, following a query about its status.
1640    
1641    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1642        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1643    
1644    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1645        quantifier applied to a forward-referencing subroutine call, could compile
1646        incorrect code or give the error "internal error: previously-checked
1647        referenced subpattern not found".
1648    
1649    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1650        variables to point to external functions. For these systems, therefore,
1651        pcre_malloc etc. are now initialized to local functions that call the
1652        relevant global functions.
1653    
1654    7.  There were two entries missing in the vectors called coptable and poptable
1655        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1656        I've fixed the data, and added a kludgy way of testing at compile time that
1657        the lengths are correct (equal to the number of opcodes).
1658    
1659    8.  Following on from 7, I added a similar kludge to check the length of the
1660        eint vector in pcreposix.c.
1661    
1662    9.  Error texts for pcre_compile() are held as one long string to avoid too
1663        much relocation at load time. To find a text, the string is searched,
1664        counting zeros. There was no check for running off the end of the string,
1665        which could happen if a new error number was added without updating the
1666        string.
1667    
1668    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1669    
1670    11. \K was not working if it appeared in an atomic group or in a group that
1671        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1672        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1673        the assertion is positive, but not if it is negative.
1674    
1675    12. Change 11 fortuitously reduced the size of the stack frame used in the
1676        "match()" function of pcre_exec.c by one pointer. Forthcoming
1677        implementation of support for (*MARK) will need an extra pointer on the
1678        stack; I have reserved it now, so that the stack frame size does not
1679        decrease.
1680    
1681    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1682        item in branch that calls a recursion is a subroutine call - as in the
1683        second branch in the above example - was incorrectly given the compile-
1684        time error "recursive call could loop indefinitely" because pcre_compile()
1685        was not correctly checking the subroutine for matching a non-empty string.
1686    
1687    14. The checks for overrunning compiling workspace could trigger after an
1688        overrun had occurred. This is a "should never occur" error, but it can be
1689        triggered by pathological patterns such as hundreds of nested parentheses.
1690        The checks now trigger 100 bytes before the end of the workspace.
1691    
1692    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1693    
1694    
1695    Version 8.01 19-Jan-2010
1696    ------------------------
1697    
1698    1.  If a pattern contained a conditional subpattern with only one branch (in
1699        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1700        computed the wrong minimum data length (which is of course zero for such
1701        subpatterns). This could cause incorrect "no match" results.
1702    
1703    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1704        the pattern is reset in the first branch, pcre_compile() failed with
1705        "internal error: code overflow at offset...". This happened only when
1706        the reset was to the original external option setting. (An optimization
1707        abstracts leading options settings into an external setting, which was the
1708        cause of this.)
1709    
1710    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1711        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1712        assertion pattern did not match (meaning that the assertion was true), it
1713        was incorrectly treated as false if the SKIP had been reached during the
1714        matching. This also applied to assertions used as conditions.
1715    
1716    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1717        assertion subpattern, including such a pattern used as a condition,
1718        unpredictable results occurred, instead of the error return
1719        PCRE_ERROR_DFA_UITEM.
1720    
1721    5.  The C++ GlobalReplace function was not working like Perl for the special
1722        situation when an empty string is matched. It now does the fancy magic
1723        stuff that is necessary.
1724    
1725    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1726        removed. (These were left over from very, very early versions of PCRE.)
1727    
1728    7.  Some cosmetic changes to the code to make life easier when compiling it
1729        as part of something else:
1730    
1731        (a) Change DEBUG to PCRE_DEBUG.
1732    
1733        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1734            called "current" as "current_branch", to prevent a collision with the
1735            Linux macro when compiled as a kernel module.
1736    
1737        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1738            prevent a collision with the Linux macro when compiled as a kernel
1739            module.
1740    
1741    8.  In pcre_compile() there are some checks for integer overflows that used to
1742        cast potentially large values to (double). This has been changed to that
1743        when building, a check for int64_t is made, and if it is found, it is used
1744        instead, thus avoiding the use of floating point arithmetic. (There is no
1745        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1746        double.
1747    
1748    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1749        2005 (difference between two addresses compared to an unsigned value).
1750    
1751    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1752        custom one, because of the following reported problem in Windows:
1753    
1754          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1755              under Win32.
1756          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1757              therefore missing the function definition.
1758          - The compiler thus generates a "C" signature for the test function.
1759          - The linker fails to find the "C" function.
1760          - PCRE fails to configure if asked to do so against libbz2.
1761    
1762    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1763        messages were output:
1764    
1765          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1766          rerunning libtoolize, to keep the correct libtool macros in-tree.
1767          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1768    
1769        I have done both of these things.
1770    
1771    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1772        most of the time, it *can* run out if it is given a pattern that contains a
1773        runaway infinite recursion. I updated the discussion in the pcrestack man
1774        page.
1775    
1776    13. Now that we have gone to the x.xx style of version numbers, the minor
1777        version may start with zero. Using 08 or 09 is a bad idea because users
1778        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1779        interpreted as invalid octal numbers. I've updated the previous comment in
1780        configure.ac, and also added a check that gives an error if 08 or 09 are
1781        used.
1782    
1783    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1784        causing partial matching to fail when the end of the subject matched \W
1785        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1786    
1787    15. There were some discrepancies between the declarations in pcre_internal.h
1788        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1789        their definitions. The declarations used "const uschar *" and the
1790        definitions used USPTR. Even though USPTR is normally defined as "const
1791        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1792        reported that: "This difference in casting confuses some C++ compilers, for
1793        example, SunCC recognizes above declarations as different functions and
1794        generates broken code for hbpcre." I have changed the declarations to use
1795        USPTR.
1796    
1797    16. GNU libtool is named differently on some systems. The autogen.sh script now
1798        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1799        (FreeBSD).
1800    
1801    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1802        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1803        comment: "Figure out how to create a longlong from a string: strtoll and
1804        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1805        instance, but it only takes 2 args instead of 3!"
1806    
1807    18. A subtle bug concerned with back references has been fixed by a change of
1808        specification, with a corresponding code fix. A pattern such as
1809        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1810        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1811        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1812        same bug. Such groups have to be quantified to be useful, or contained
1813        inside another quantified group. (If there's no repetition, the reference
1814        can never match.) The problem arises because, having left the group and
1815        moved on to the rest of the pattern, a later failure that backtracks into
1816        the group uses the captured value from the final iteration of the group
1817        rather than the correct earlier one. I have fixed this in PCRE by forcing
1818        any group that contains a reference to itself to be an atomic group; that
1819        is, there cannot be any backtracking into it once it has completed. This is
1820        similar to recursive and subroutine calls.
1821    
1822    
1823    Version 8.00 19-Oct-09
1824  ----------------------  ----------------------
1825    
1826  1.  The table for translating pcre_compile() error codes into POSIX error codes  1.  The table for translating pcre_compile() error codes into POSIX error codes
1827      was out-of-date, and there was no check on the pcre_compile() error code      was out-of-date, and there was no check on the pcre_compile() error code
1828      being within the table. This could lead to an OK return being given in      being within the table. This could lead to an OK return being given in
1829      error.      error.
1830    
1831  2.  Changed the call to open a subject file in pcregrep from fopen(pathname,  2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1832      "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests      "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1833      in a Windows environment.      in a Windows environment.
1834    
1835  3.  The pcregrep --count option prints the count for each file even when it is  3.  The pcregrep --count option prints the count for each file even when it is
1836      zero, as does GNU grep. However, pcregrep was also printing all files when      zero, as does GNU grep. However, pcregrep was also printing all files when
1837      --files-with-matches was added. Now, when both options are given, it prints      --files-with-matches was added. Now, when both options are given, it prints
1838      counts only for those files that have at least one match. (GNU grep just      counts only for those files that have at least one match. (GNU grep just
1839      prints the file name in this circumstance, but including the count seems      prints the file name in this circumstance, but including the count seems
1840      more useful - otherwise, why use --count?) Also ensured that the      more useful - otherwise, why use --count?) Also ensured that the
1841      combination -clh just lists non-zero counts, with no names.      combination -clh just lists non-zero counts, with no names.
1842    
1843  4.  The long form of the pcregrep -F option was incorrectly implemented as  4.  The long form of the pcregrep -F option was incorrectly implemented as
1844      --fixed_strings instead of --fixed-strings. This is an incompatible change,      --fixed_strings instead of --fixed-strings. This is an incompatible change,
1845      but it seems right to fix it, and I didn't think it was worth preserving      but it seems right to fix it, and I didn't think it was worth preserving
1846      the old behaviour.      the old behaviour.
1847    
1848  5.  The command line items --regex=pattern and --regexp=pattern were not  5.  The command line items --regex=pattern and --regexp=pattern were not
1849      recognized by pcregrep, which required --regex pattern or --regexp pattern      recognized by pcregrep, which required --regex pattern or --regexp pattern
1850      (with a space rather than an '='). The man page documented the '=' forms,      (with a space rather than an '='). The man page documented the '=' forms,
1851      which are compatible with GNU grep; these now work.      which are compatible with GNU grep; these now work.
1852    
1853  6.  No libpcreposix.pc file was created for pkg-config; there was just  6.  No libpcreposix.pc file was created for pkg-config; there was just
1854      libpcre.pc and libpcrecpp.pc. The omission has been rectified.      libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1855    
1856  7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size  7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1857      when UCP support is not needed, by modifying the Python script that      when UCP support is not needed, by modifying the Python script that
1858      generates it from Unicode data files. This should not matter if the module      generates it from Unicode data files. This should not matter if the module
1859      is correctly used as a library, but I received one complaint about 50K of      is correctly used as a library, but I received one complaint about 50K of
1860      unwanted data. My guess is that the person linked everything into his      unwanted data. My guess is that the person linked everything into his
1861      program rather than using a library. Anyway, it does no harm.      program rather than using a library. Anyway, it does no harm.
1862    
1863    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1864        was a minimum greater than 1 for a wide character in a possessive
1865        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1866        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1867        character. Chaos in the form of incorrect output or a compiling loop could
1868        result.
1869    
1870    9.  The restrictions on what a pattern can contain when partial matching is
1871        requested for pcre_exec() have been removed. All patterns can now be
1872        partially matched by this function. In addition, if there are at least two
1873        slots in the offset vector, the offset of the earliest inspected character
1874        for the match and the offset of the end of the subject are set in them when
1875        PCRE_ERROR_PARTIAL is returned.
1876    
1877    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1878        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1879        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1880        and may be more useful for multi-segment matching.
1881    
1882    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1883        used to be given if ever the end of the subject was reached; now it is
1884        given only if matching could not proceed because another character was
1885        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1886        string "Z", which now yields "no match" instead of "partial match". In the
1887        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1888        final character ended with (*FAIL).
1889    
1890    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1891        if the pattern had a "must contain" character that was already found in the
1892        earlier partial match, unless partial matching was again requested. For
1893        example, with the pattern /dog.(body)?/, the "must contain" character is
1894        "g". If the first part-match was for the string "dog", restarting with
1895        "sbody" failed. This bug has been fixed.
1896    
1897    13. The string returned by pcre_dfa_exec() after a partial match has been
1898        changed so that it starts at the first inspected character rather than the
1899        first character of the match. This makes a difference only if the pattern
1900        starts with a lookbehind assertion or \b or \B (\K is not supported by
1901        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1902        matching functions compatible, and I think it's the right thing to do.
1903    
1904    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1905        so that the demonstration program is easily available in environments where
1906        PCRE has not been installed from source.
1907    
1908    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1909        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1910        library.
1911    
1912    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1913        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1914        is not the first non-POSIX option to be added. Clearly some people find
1915        these options useful.
1916    
1917    17. If a caller to the POSIX matching function regexec() passes a non-zero
1918        value for nmatch with a NULL value for pmatch, the value of
1919        nmatch is forced to zero.
1920    
1921    18. RunGrepTest did not have a test for the availability of the -u option of
1922        the diff command, as RunTest does. It now checks in the same way as
1923        RunTest, and also checks for the -b option.
1924    
1925    19. If an odd number of negated classes containing just a single character
1926        interposed, within parentheses, between a forward reference to a named
1927        subpattern and the definition of the subpattern, compilation crashed with
1928        an internal error, complaining that it could not find the referenced
1929        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1930        [The bug was that it was starting one character too far in when skipping
1931        over the character class, thus treating the ] as data rather than
1932        terminating the class. This meant it could skip too much.]
1933    
1934    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1935        /g option in pcretest when the pattern contains \K, which makes it possible
1936        to have an empty string match not at the start, even when the pattern is
1937        anchored. Updated pcretest and pcredemo to use this option.
1938    
1939    21. If the maximum number of capturing subpatterns in a recursion was greater
1940        than the maximum at the outer level, the higher number was returned, but
1941        with unset values at the outer level. The correct (outer level) value is
1942        now given.
1943    
1944    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1945        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1946        make it do so. The string so far is captured, making this feature
1947        compatible with Perl.
1948    
1949    23. The tests have been re-organized, adding tests 11 and 12, to make it
1950        possible to check the Perl 5.10 features against Perl 5.10.
1951    
1952    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1953        pattern matches a fixed length string. PCRE did not allow this; now it
1954        does. Neither allows recursion.
1955    
1956    25. I finally figured out how to implement a request to provide the minimum
1957        length of subject string that was needed in order to match a given pattern.
1958        (It was back references and recursion that I had previously got hung up
1959        on.) This code has now been added to pcre_study(); it finds a lower bound
1960        to the length of subject needed. It is not necessarily the greatest lower
1961        bound, but using it to avoid searching strings that are too short does give
1962        some useful speed-ups. The value is available to calling programs via
1963        pcre_fullinfo().
1964    
1965    26. While implementing 25, I discovered to my embarrassment that pcretest had
1966        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1967        study optimizations had never been tested with that matching function.
1968        Oops. What is worse, even when it was passed study data, there was a bug in
1969        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1970        were also very few tests of studied patterns with pcre_dfa_exec().
1971    
1972    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1973        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1974        on the other side of the coin, they are no longer allowed to have different
1975        names, because these cannot be distinguished in PCRE, and this has caused
1976        confusion. (This is a difference from Perl.)
1977    
1978    28. When duplicate subpattern names are present (necessarily with different
1979        numbers, as required by 27 above), and a test is made by name in a
1980        conditional pattern, either for a subpattern having been matched, or for
1981        recursion in such a pattern, all the associated numbered subpatterns are
1982        tested, and the overall condition is true if the condition is true for any
1983        one of them. This is the way Perl works, and is also more like the way
1984        testing by number works.
1985    
1986    
1987  Version 7.9 11-Apr-09  Version 7.9 11-Apr-09
1988  ---------------------  ---------------------
# Line 76  Version 7.9 11-Apr-09 Line 2018  Version 7.9 11-Apr-09
2018  7.  A pattern that could match an empty string could cause pcregrep to loop; it  7.  A pattern that could match an empty string could cause pcregrep to loop; it
2019      doesn't make sense to accept an empty string match in pcregrep, so I have      doesn't make sense to accept an empty string match in pcregrep, so I have
2020      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2021      seems to be how GNU grep behaves.      seems to be how GNU grep behaves. [But see later change 40 for release
2022        8.33.]
2023    
2024  8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at  8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2025      start or after a newline", because the conditional assertion was not being      start or after a newline", because the conditional assertion was not being
# Line 319  Version 7.7 07-May-08 Line 2262  Version 7.7 07-May-08
2262      containing () gave an internal compiling error instead of "reference to      containing () gave an internal compiling error instead of "reference to
2263      non-existent subpattern". Fortunately, when the pattern did exist, the      non-existent subpattern". Fortunately, when the pattern did exist, the
2264      compiled code was correct. (When scanning forwards to check for the      compiled code was correct. (When scanning forwards to check for the
2265      existencd of the subpattern, it was treating the data ']' as terminating      existence of the subpattern, it was treating the data ']' as terminating
2266      the class, so got the count wrong. When actually compiling, the reference      the class, so got the count wrong. When actually compiling, the reference
2267      was subsequently set up correctly.)      was subsequently set up correctly.)
2268    

Legend:
Removed from v.425  
changed lines
  Added in v.1473

  ViewVC Help
Powered by ViewVC 1.1.5