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

Diff of /code/trunk/ChangeLog

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

revision 142 by ph10, Fri Mar 30 15:55:18 2007 UTC revision 1477 by ph10, Wed May 21 17:53:49 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.1 12-Mar-07  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    4.  The JIT compiler did not generate match limit checks for certain
19        bracketed expressions with quantifiers. This may lead to exponential
20        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
21        issue should be resolved now.
22    
23    5.  Fixed an issue, which occures when nested alternatives are optimized
24        with table jumps.
25    
26    6.  Inserted two casts and changed some ints to size_t in the light of some
27        reported 64-bit compiler warnings (Bugzilla 1477).
28    
29    
30    Version 8.35 04-April-2014
31    --------------------------
32    
33    1.  A new flag is set, when property checks are present in an XCLASS.
34        When this flag is not set, PCRE can perform certain optimizations
35        such as studying these XCLASS-es.
36    
37    2.  The auto-possessification of character sets were improved: a normal
38        and an extended character set can be compared now. Furthermore
39        the JIT compiler optimizes more character set checks.
40    
41    3.  Got rid of some compiler warnings for potentially uninitialized variables
42        that show up only when compiled with -O2.
43    
44    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
45        of a match later then the end of the match. The pcretest program was not
46        handling the case sensibly - it was outputting from the start to the next
47        binary zero. It now reports this situation in a message, and outputs the
48        text from the end to the start.
49    
50    5.  Fast forward search is improved in JIT. Instead of the first three
51        characters, any three characters with fixed position can be searched.
52        Search order: first, last, middle.
53    
54    6.  Improve character range checks in JIT. Characters are read by an inprecise
55        function now, which returns with an unknown value if the character code is
56        above a certain treshold (e.g: 256). The only limitation is that the value
57        must be bigger than the treshold as well. This function is useful, when
58        the characters above the treshold are handled in the same way.
59    
60    7.  The macros whose names start with RAWUCHAR are placeholders for a future
61        mode in which only the bottom 21 bits of 32-bit data items are used. To
62        make this more memorable for those maintaining the code, the names have
63        been changed to start with UCHAR21, and an extensive comment has been added
64        to their definition.
65    
66    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
67        to the export list in Makefile.am (they were accidentally omitted from the
68        8.34 tarball).
69    
70    9.  The informational output from pcretest used the phrase "starting byte set"
71        which is inappropriate for the 16-bit and 32-bit libraries. As the output
72        for "first char" and "need char" really means "non-UTF-char", I've changed
73        "byte" to "char", and slightly reworded the output. The documentation about
74        these values has also been (I hope) clarified.
75    
76    10. Another JIT related optimization: use table jumps for selecting the correct
77        backtracking path, when more than four alternatives are present inside a
78        bracket.
79    
80    11. Empty match is not possible, when the minimum length is greater than zero,
81        and there is no \K in the pattern. JIT should avoid empty match checks in
82        such cases.
83    
84    12. In a caseless character class with UCP support, when a character with more
85        than one alternative case was not the first character of a range, not all
86        the alternative cases were added to the class. For example, s and \x{17f}
87        are both alternative cases for S: the class [RST] was handled correctly,
88        but [R-T] was not.
89    
90    13. The configure.ac file always checked for pthread support when JIT was
91        enabled. This is not used in Windows, so I have put this test inside a
92        check for the presence of windows.h (which was already tested for).
93    
94    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
95        The algorithm provides a way to skip certain starting offsets, and usually
96        faster than linear prefix searches.
97    
98    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
99        as for 'fr_FR' and 'french'. For some reason, however, it then used the
100        Windows-specific input and output files, which have 'french' screwed in.
101        So this could never have worked. One of the problems with locales is that
102        they aren't always the same. I have now updated RunTest so that it checks
103        the output of the locale test (test 3) against three different output
104        files, and it allows the test to pass if any one of them matches. With luck
105        this should make the test pass on some versions of Solaris where it was
106        failing. Because of the uncertainty, the script did not used to stop if
107        test 3 failed; it now does. If further versions of a French locale ever
108        come to light, they can now easily be added.
109    
110    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
111        there was a message during ./configure, but it did not stop. This now
112        provokes an error. The invalid example in README has been corrected.
113        If a value less than the minimum is given, the minimum value has always
114        been used, but now a warning is given.
115    
116    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
117        was a bug in the test system, which is now fixed. Also, the list of various
118        configurations that are tested for each release did not have one with both
119        16/32 bits and --enable-bar-anycrlf. It now does.
120    
121    18. pcretest was missing "-C bsr" for displaying the \R default setting.
122    
123    19. Little endian PowerPC systems are supported now by the JIT compiler.
124    
125    20. The fast forward newline mechanism could enter to an infinite loop on
126        certain invalid UTF-8 input. Although we don't support these cases
127        this issue can be fixed by a performance optimization.
128    
129    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
130        not take account if existing stack usage. There is now a new global
131        variable called pcre_stack_guard that can be set to point to an external
132        function to check stack availability. It is called at the start of
133        processing every parenthesized group.
134    
135    22. A typo in the code meant that in ungreedy mode the max/min qualifier
136        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
137        not match "ab".
138    
139    23. When UTF was disabled, the JIT program reported some incorrect compile
140        errors. These messages are silenced now.
141    
142    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
143        compiler.
144    
145    25. Change all the temporary files used in RunGrepTest to be different to those
146        used by RunTest so that the tests can be run simultaneously, for example by
147        "make -j check".
148    
149    
150    Version 8.34 15-December-2013
151    -----------------------------
152    
153    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
154        executable memory. Patch inspired by Carsten Klein.
155    
156    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
157        this macro is never tested and has no effect, because the work to support
158        coverage involves only compiling and linking options and special targets in
159        the Makefile. The comment in config.h implied that defining the macro would
160        enable coverage support, which is totally false. There was also support for
161        setting this macro in the CMake files (my fault, I just copied it from
162        configure). SUPPORT_GCOV has now been removed.
163    
164    3.  Make a small performance improvement in strlen16() and strlen32() in
165        pcretest.
166    
167    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
168        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
169        these cases). There was also one in pcretest.c.
170    
171    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
172    
173    6.  In UTF mode, the code for checking whether a group could match an empty
174        string (which is used for indefinitely repeated groups to allow for
175        breaking an infinite loop) was broken when the group contained a repeated
176        negated single-character class with a character that occupied more than one
177        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
178        UTF-8 mode). The effect was undefined: the group might or might not be
179        deemed as matching an empty string, or the program might have crashed.
180    
181    7.  The code for checking whether a group could match an empty string was not
182        recognizing that \h, \H, \v, \V, and \R must match a character.
183    
184    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
185        an empty string. If it can, pcretest shows this in its information output.
186    
187    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
188        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
189        matched by pcre_exec() without using JIT:
190    
191        (a) If the rest of the pattern did not match after a maximal run of
192            grapheme clusters, the code for backing up to try with fewer of them
193            did not always back up over a full grapheme when characters that do not
194            have the modifier quality were involved, e.g. Hangul syllables.
195    
196        (b) If the match point in a subject started with modifier character, and
197            there was no match, the code could incorrectly back up beyond the match
198            point, and potentially beyond the first character in the subject,
199            leading to a segfault or an incorrect match result.
200    
201    10. A conditional group with an assertion condition could lead to PCRE
202        recording an incorrect first data item for a match if no other first data
203        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
204        first data item, and therefore matched "ca" after "c" instead of at the
205        start.
206    
207    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
208        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
209    
210    12. The source of pcregrep now includes z/OS-specific code so that it can be
211        compiled for z/OS as part of the special z/OS distribution.
212    
213    13. Added the -T and -TM options to pcretest.
214    
215    14. The code in pcre_compile.c for creating the table of named capturing groups
216        has been refactored. Instead of creating the table dynamically during the
217        actual compiling pass, the information is remembered during the pre-compile
218        pass (on the stack unless there are more than 20 named groups, in which
219        case malloc() is used) and the whole table is created before the actual
220        compile happens. This has simplified the code (it is now nearly 150 lines
221        shorter) and prepared the way for better handling of references to groups
222        with duplicate names.
223    
224    15. A back reference to a named subpattern when there is more than one of the
225        same name now checks them in the order in which they appear in the pattern.
226        The first one that is set is used for the reference. Previously only the
227        first one was inspected. This change makes PCRE more compatible with Perl.
228    
229    16. Unicode character properties were updated from Unicode 6.3.0.
230    
231    17. The compile-time code for auto-possessification has been refactored, based
232        on a patch by Zoltan Herczeg. It now happens after instead of during
233        compilation. The code is cleaner, and more cases are handled. The option
234        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
235        options in pcretest are provided to set it. It can also be set by
236        (*NO_AUTO_POSSESS) at the start of a pattern.
237    
238    18. The character VT has been added to the default ("C" locale) set of
239        characters that match \s and are generally treated as white space,
240        following this same change in Perl 5.18. There is now no difference between
241        "Perl space" and "POSIX space". Whether VT is treated as white space in
242        other locales depends on the locale.
243    
244    19. The code for checking named groups as conditions, either for being set or
245        for being recursed, has been refactored (this is related to 14 and 15
246        above). Processing unduplicated named groups should now be as fast at
247        numerical groups, and processing duplicated groups should be faster than
248        before.
249    
250    20. Two patches to the CMake build system, by Alexander Barkov:
251    
252          (1) Replace the "source" command by "." in CMakeLists.txt because
253              "source" is a bash-ism.
254    
255          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
256              without these the CMake build does not work on Solaris.
257    
258    21. Perl has changed its handling of \8 and \9. If there is no previously
259        encountered capturing group of those numbers, they are treated as the
260        literal characters 8 and 9 instead of a binary zero followed by the
261        literals. PCRE now does the same.
262    
263    22. Following Perl, added \o{} to specify codepoints in octal, making it
264        possible to specify values greater than 0777 and also making them
265        unambiguous.
266    
267    23. Perl now gives an error for missing closing braces after \x{... instead of
268        treating the string as literal. PCRE now does the same.
269    
270    24. RunTest used to grumble if an inappropriate test was selected explicitly,
271        but just skip it when running all tests. This make it awkward to run ranges
272        of tests when one of them was inappropriate. Now it just skips any
273        inappropriate tests, as it always did when running all tests.
274    
275    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
276        character types such as \d or \w, too many callouts were inserted, and the
277        data that they returned was rubbish.
278    
279    26. In UCP mode, \s was not matching two of the characters that Perl matches,
280        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
281        were matched by \h. The code has now been refactored so that the lists of
282        the horizontal and vertical whitespace characters used for \h and \v (which
283        are defined only in one place) are now also used for \s.
284    
285    27. Add JIT support for the 64 bit TileGX architecture.
286        Patch by Jiong Wang (Tilera Corporation).
287    
288    28. Possessive quantifiers for classes (both explicit and automatically
289        generated) now use special opcodes instead of wrapping in ONCE brackets.
290    
291    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
292        (because it's meaningless), this was not happening when PCRE_CASELESS was
293        set. Not wrong, but inefficient.
294    
295    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
296        properties for \w, \d, etc) is present in a test regex. Otherwise if the
297        test contains no characters greater than 255, Perl doesn't realise it
298        should be using Unicode semantics.
299    
300    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
301        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
302        does in Unicode mode.
303    
304    32. Added the "forbid" facility to pcretest so that putting tests into the
305        wrong test files can sometimes be quickly detected.
306    
307    33. There is now a limit (default 250) on the depth of nesting of parentheses.
308        This limit is imposed to control the amount of system stack used at compile
309        time. It can be changed at build time by --with-parens-nest-limit=xxx or
310        the equivalent in CMake.
311    
312    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
313        errors. Perl warns for these when in warning mode, but PCRE has no facility
314        for giving warnings.
315    
316    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
317        However, this was not working for (?!) because it is optimized to (*FAIL),
318        for which PCRE does not allow quantifiers. The optimization is now disabled
319        when a quantifier follows (?!). I can't see any use for this, but it makes
320        things uniform.
321    
322    36. Perl no longer allows group names to start with digits, so I have made this
323        change also in PCRE. It simplifies the code a bit.
324    
325    37. In extended mode, Perl ignores spaces before a + that indicates a
326        possessive quantifier. PCRE allowed a space before the quantifier, but not
327        before the possessive +. It now does.
328    
329    38. The use of \K (reset reported match start) within a repeated possessive
330        group such as (a\Kb)*+ was not working.
331    
332    40. Document that the same character tables must be used at compile time and
333        run time, and that the facility to pass tables to pcre_exec() and
334        pcre_dfa_exec() is for use only with saved/restored patterns.
335    
336    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
337        features for Builds with MSVC:
338    
339        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
340           builds). The .rc files can be used to set FileDescription and many other
341           attributes.
342    
343        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
344           This allows higher-level build scripts which want .pdb files to avoid
345           hard-coding the exact files needed."
346    
347    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
348        mean "start of word" and "end of word", respectively, as a transition aid.
349    
350    43. A minimizing repeat of a class containing codepoints greater than 255 in
351        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
352        compiled to use the heap for recursion.
353    
354    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
355        is configured.
356    
357    
358    Version 8.33 28-May-2013
359    ------------------------
360    
361    1.  Added 'U' to some constants that are compared to unsigned integers, to
362        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
363        variables that are added to signed variables, to ensure the result is
364        signed and can be negated.
365    
366    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
367        CMake config files.
368    
369    3.  Revise the creation of config.h.generic so that all boolean macros are
370        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
371        overriding via -D on the command line possible.
372    
373    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
374        to unsigned int is reported to make a quite noticeable speed difference in
375        a specific Windows environment. Testing on Linux did also appear to show
376        some benefit (and it is clearly not harmful). Also fixed the definition of
377        Xop which should be unsigned.
378    
379    5.  Related to (4), changing the definition of the intermediate variable cc
380        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
381        improvements.
382    
383    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
384        unnecessary spaces.
385    
386    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
387        and later.
388    
389    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
390    
391    9.  Optimizing fast_forward_start_bits in JIT.
392    
393    10. Adding support for callouts in JIT, and fixing some issues revealed
394        during this work. Namely:
395    
396        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
397    
398        (b) Minimum length was not checked before the matching is started.
399    
400    11. The value of capture_last that is passed to callouts was incorrect in some
401        cases when there was a capture on one path that was subsequently abandoned
402        after a backtrack. Also, the capture_last value is now reset after a
403        recursion, since all captures are also reset in this case.
404    
405    12. The interpreter no longer returns the "too many substrings" error in the
406        case when an overflowing capture is in a branch that is subsequently
407        abandoned after a backtrack.
408    
409    13. In the pathological case when an offset vector of size 2 is used, pcretest
410        now prints out the matched string after a yield of 0 or 1.
411    
412    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
413        Only supported by the JIT compiler at the moment.
414    
415    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
416    
417    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
418        offset of the starting point of the matching process, provided the offsets
419        vector is large enough.
420    
421    17. The \A escape now records a lookbehind value of 1, though its execution
422        does not actually inspect the previous character. This is to ensure that,
423        in partial multi-segment matching, at least one character from the old
424        segment is retained when a new segment is processed. Otherwise, if there
425        are no lookbehinds in the pattern, \A might match incorrectly at the start
426        of a new segment.
427    
428    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
429    
430    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
431        gives some modest performance improvement in 8-bit mode.
432    
433    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
434        be expressed in certain programming languages using Universal Character
435        Names.
436    
437    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
438        which points out that "non characters" are not "characters that may not
439        appear in Unicode strings" but rather "characters that are reserved for
440        internal use and have only local meaning".
441    
442    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
443        there was a conditional group that depended on an assertion, if the
444        assertion was false, the callout that immediately followed the alternation
445        in the condition was skipped when pcre_exec() was used for matching.
446    
447    23. Allow an explicit callout to be inserted before an assertion that is the
448        condition for a conditional group, for compatibility with automatic
449        callouts, which always insert a callout at this point.
450    
451    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
452        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
453    
454    25. (*PRUNE) is now supported by the JIT compiler.
455    
456    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
457    
458    27. Fix the case where there are two or more SKIPs with arguments that may be
459        ignored.
460    
461    28. (*SKIP) is now supported by the JIT compiler.
462    
463    29. (*THEN) is now supported by the JIT compiler.
464    
465    30. Update RunTest with additional test selector options.
466    
467    31. The way PCRE handles backtracking verbs has been changed in two ways.
468    
469        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
470        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
471        backtracking. In some cases this makes it more Perl-compatible, but Perl's
472        rather obscure rules do not always do the same thing.
473    
474        (2) Previously, backtracking verbs were confined within assertions. This is
475        no longer the case for positive assertions, except for (*ACCEPT). Again,
476        this sometimes improves Perl compatibility, and sometimes does not.
477    
478    32. A number of tests that were in test 2 because Perl did things differently
479        have been moved to test 1, because either Perl or PCRE has changed, and
480        these tests are now compatible.
481    
482    32. Backtracking control verbs are now handled in the same way in JIT and
483        interpreter.
484    
485    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
486        contained a forward subroutine reference caused a compile error.
487    
488    34. Auto-detect and optimize limited repetitions in JIT.
489    
490    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
491        blocking (*UTF) etc.
492    
493    36. In the interpreter, maximizing pattern repetitions for characters and
494        character types now use tail recursion, which reduces stack usage.
495    
496    37. The value of the max lookbehind was not correctly preserved if a compiled
497        and saved regex was reloaded on a host of different endianness.
498    
499    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
500        of the compiled pattern block, expand the flags field from 16 to 32 bits
501        because it was almost full.
502    
503    39. Try madvise first before posix_madvise.
504    
505    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
506        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
507        original change locked out all matches of empty strings. This has been
508        changed so that one match of an empty string per line is recognized.
509        Subsequent searches on the same line (for colouring or for --only-matching,
510        for example) do not recognize empty strings.
511    
512    41. Applied a user patch to fix a number of spelling mistakes in comments.
513    
514    42. Data lines longer than 65536 caused pcretest to crash.
515    
516    43. Clarified the data type for length and startoffset arguments for pcre_exec
517        and pcre_dfa_exec in the function-specific man pages, where they were
518        explicitly stated to be in bytes, never having been updated. I also added
519        some clarification to the pcreapi man page.
520    
521    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
522        a segmentation fault.
523    
524    
525    Version 8.32 30-November-2012
526    -----------------------------
527    
528    1.  Improved JIT compiler optimizations for first character search and single
529        character iterators.
530    
531    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
532        Patch by Daniel Richard G.
533    
534    3.  Single character iterator optimizations in the JIT compiler.
535    
536    4.  Improved JIT compiler optimizations for character ranges.
537    
538    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
539        Reported by Giuseppe D'Angelo.
540    
541    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
542        of a line, was being set incorrectly in cases where .* appeared inside
543        atomic brackets at the start of a pattern, or where there was a subsequent
544        *PRUNE or *SKIP.
545    
546    7.  Improved instruction cache flush for POWER/PowerPC.
547        Patch by Daniel Richard G.
548    
549    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
550        grep:
551    
552        (a) There is now no limit to the number of patterns to be matched.
553    
554        (b) An error is given if a pattern is too long.
555    
556        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
557            are now supported.
558    
559        (d) --exclude-from and --include-from (multiple use) have been added.
560    
561        (e) Exclusions and inclusions now apply to all files and directories, not
562            just to those obtained from scanning a directory recursively.
563    
564        (f) Multiple uses of -f and --file-list are now supported.
565    
566        (g) In a Windows environment, the default for -d has been changed from
567            "read" (the GNU grep default) to "skip", because otherwise the presence
568            of a directory in the file list provokes an error.
569    
570        (h) The documentation has been revised and clarified in places.
571    
572    9.  Improve the matching speed of capturing brackets.
573    
574    10. Changed the meaning of \X so that it now matches a Unicode extended
575        grapheme cluster.
576    
577    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
578        out POSIX threads when JIT support is configured.
579    
580    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
581    
582    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
583        structure could go wrong in environments where size_t is not the same size
584        as int.
585    
586    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
587        set.
588    
589    15. The EBCDIC support had decayed; later updates to the code had included
590        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
591        general tidy up of EBCDIC-related issues, and the documentation was also
592        not quite right. There is now a test that can be run on ASCII systems to
593        check some of the EBCDIC-related things (but is it not a full test).
594    
595    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
596        in a small tidy to the code.
597    
598    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
599    
600    18. If the --only-matching (-o) option in pcregrep is specified multiple
601        times, each one causes appropriate output. For example, -o1 -o2 outputs the
602        substrings matched by the 1st and 2nd capturing parentheses. A separating
603        string can be specified by --om-separator (default empty).
604    
605    19. Improving the first n character searches.
606    
607    20. Turn case lists for horizontal and vertical white space into macros so that
608        they are defined only once.
609    
610    21. This set of changes together give more compatible Unicode case-folding
611        behaviour for characters that have more than one other case when UCP
612        support is available.
613    
614        (a) The Unicode property table now has offsets into a new table of sets of
615            three or more characters that are case-equivalent. The MultiStage2.py
616            script that generates these tables (the pcre_ucd.c file) now scans
617            CaseFolding.txt instead of UnicodeData.txt for character case
618            information.
619    
620        (b) The code for adding characters or ranges of characters to a character
621            class has been abstracted into a generalized function that also handles
622            case-independence. In UTF-mode with UCP support, this uses the new data
623            to handle characters with more than one other case.
624    
625        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
626            whose other case is greater than 256 are now correctly matched
627            caselessly. Previously, the high codepoint matched the low one, but not
628            vice versa.
629    
630        (d) The processing of \h, \H, \v, and \ in character classes now makes use
631            of the new class addition function, using character lists defined as
632            macros alongside the case definitions of 20 above.
633    
634        (e) Caseless back references now work with characters that have more than
635            one other case.
636    
637        (f) General caseless matching of characters with more than one other case
638            is supported.
639    
640    22. Unicode character properties were updated from Unicode 6.2.0
641    
642    23. Improved CMake support under Windows. Patch by Daniel Richard G.
643    
644    24. Add support for 32-bit character strings, and UTF-32
645    
646    25. Major JIT compiler update (code refactoring and bugfixing).
647        Experimental Sparc 32 support is added.
648    
649    26. Applied a modified version of Daniel Richard G's patch to create
650        pcre.h.generic and config.h.generic by "make" instead of in the
651        PrepareRelease script.
652    
653    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
654        pcre_compile.c when checking for a zero character.
655    
656    28. Introducing a native interface for JIT. Through this interface, the compiled
657        machine code can be directly executed. The purpose of this interface is to
658        provide fast pattern matching, so several sanity checks are not performed.
659        However, feature tests are still performed. The new interface provides
660        1.4x speedup compared to the old one.
661    
662    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
663        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
664        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
665    
666    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
667        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
668        the "old" RFC 2279). Instead, it ended up passing a negative length to
669        pcre_exec().
670    
671    31. Add support for GCC's visibility feature to hide internal functions.
672    
673    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
674        "unknown -C option" after outputting 0 or 1.
675    
676    33. There is now support for generating a code coverage report for the test
677        suite in environments where gcc is the compiler and lcov is installed. This
678        is mainly for the benefit of the developers.
679    
680    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
681        unaddressable using valgrind annotations, allowing valgrind to detect
682        invalid memory accesses. This is mainly for the benefit of the developers.
683    
684    25. (*UTF) can now be used to start a pattern in any of the three libraries.
685    
686    26. Give configure error if --enable-cpp but no C++ compiler found.
687    
688    
689    Version 8.31 06-July-2012
690    -------------------------
691    
692    1.  Fixing a wrong JIT test case and some compiler warnings.
693    
694    2.  Removed a bashism from the RunTest script.
695    
696    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
697        to unsigned type, result still unsigned" that was given by an MS compiler
698        on encountering the code "-sizeof(xxx)".
699    
700    4.  Partial matching support is added to the JIT compiler.
701    
702    5.  Fixed several bugs concerned with partial matching of items that consist
703        of more than one character:
704    
705        (a) /^(..)\1/ did not partially match "aba" because checking references was
706            done on an "all or nothing" basis. This also applied to repeated
707            references.
708    
709        (b) \R did not give a hard partial match if \r was found at the end of the
710            subject.
711    
712        (c) \X did not give a hard partial match after matching one or more
713            characters at the end of the subject.
714    
715        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
716            a partial match for the string "\r".
717    
718        (e) When newline was set to CRLF, the metacharacter "." did not recognize
719            a partial match for a CR character at the end of the subject string.
720    
721    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
722        running pcretest, the text "(JIT)" added to the output whenever JIT is
723        actually used to run the match.
724    
725    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
726        or /S+[+] with a digit between 1 and 7.
727    
728    8.  OP_NOT now supports any UTF character not just single-byte ones.
729    
730    9.  (*MARK) control verb is now supported by the JIT compiler.
731    
732    10. The command "./RunTest list" lists the available tests without actually
733        running any of them. (Because I keep forgetting what they all are.)
734    
735    11. Add PCRE_INFO_MAXLOOKBEHIND.
736    
737    12. Applied a (slightly modified) user-supplied patch that improves performance
738        when the heap is used for recursion (compiled with --disable-stack-for-
739        recursion). Instead of malloc and free for each heap frame each time a
740        logical recursion happens, frames are retained on a chain and re-used where
741        possible. This sometimes gives as much as 30% improvement.
742    
743    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
744        call.
745    
746    14. As documented, (*COMMIT) is now confined to within a positive assertion.
747    
748    15. It is now possible to link pcretest with libedit as an alternative to
749        libreadline.
750    
751    16. (*COMMIT) control verb is now supported by the JIT compiler.
752    
753    17. The Unicode data tables have been updated to Unicode 6.1.0.
754    
755    18. Added --file-list option to pcregrep.
756    
757    19. Added binary file support to pcregrep, including the -a, --binary-files,
758        -I, and --text options.
759    
760    20. The madvise function is renamed for posix_madvise for QNX compatibility
761        reasons. Fixed by Giuseppe D'Angelo.
762    
763    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
764    
765    22. Changed the option for creating symbolic links for 16-bit man pages from
766        -s to -sf so that re-installing does not cause issues.
767    
768    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
769    
770    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
771        matches in certain environments (the workspace was not being correctly
772        retained). Also added to pcre_dfa_exec() a simple plausibility check on
773        some of the workspace data at the beginning of a restart.
774    
775    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
776        was not doing so when it should - probably a typo introduced by SVN 528
777        (change 8.10/14).
778    
779    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
780        \w+ when the character tables indicated that \x{c4} was a word character.
781        There were several related cases, all because the tests for doing a table
782        lookup were testing for characters less than 127 instead of 255.
783    
784    27. If a pattern contains capturing parentheses that are not used in a match,
785        their slots in the ovector are set to -1. For those that are higher than
786        any matched groups, this happens at the end of processing. In the case when
787        there were back references that the ovector was too small to contain
788        (causing temporary malloc'd memory to be used during matching), and the
789        highest capturing number was not used, memory off the end of the ovector
790        was incorrectly being set to -1. (It was using the size of the temporary
791        memory instead of the true size.)
792    
793    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
794        ovector size, it uses memory at the end of the block that it has got.
795    
796    29. Check for an overlong MARK name and give an error at compile time. The
797        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
798    
799    30. JIT compiler update.
800    
801    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
802        Rill for the patch.
803    
804    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
805    
806    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
807    
808    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
809        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
810        was enabled.
811    
812    35. Improve JIT code generation for greedy plus quantifier.
813    
814    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
815        1 to "aa" instead of to an empty string. The bug affected repeated groups
816        that could potentially match an empty string.
817    
818    37. Optimizing single character iterators in JIT.
819    
820    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
821        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
822        codepoints that are too big for the mode are faulted, and in a UTF mode,
823        disallowed codepoints are also faulted.
824    
825    39. If PCRE was compiled with UTF support, in three places in the DFA
826        matcher there was code that should only have been obeyed in UTF mode, but
827        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
828        processing when bytes with values greater than 127 were present. In 16-bit
829        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
830        both cases the values are those that cannot be the first data item in a UTF
831        character. The three items that might have provoked this were recursions,
832        possessively repeated groups, and atomic groups.
833    
834    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
835        and pcregrep, because some OS require shared objects to be explicitly
836        passed to ld, causing the link step to fail if they are not.
837    
838    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
839        mode, patterns that started with \h* or \R* might be incorrectly matched.
840    
841    
842    Version 8.30 04-February-2012
843    -----------------------------
844    
845    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
846        name is defined in ctype.h.
847    
848    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
849        only in quite long subpatterns.
850    
851    3.  Removed the function pcre_info(), which has been obsolete and deprecated
852        since it was replaced by pcre_fullinfo() in February 2000.
853    
854    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
855        match a (*MARK), and the match failed at the start of the subject, a
856        reference to memory before the start of the subject could occur. This bug
857        was introduced by fix 17 of release 8.21.
858    
859    5.  A reference to an unset group with zero minimum repetition was giving
860        totally wrong answers (in non-JavaScript-compatibility mode). For example,
861        /(another)?(\1?)test/ matched against "hello world test". This bug was
862        introduced in release 8.13.
863    
864    6.  Add support for 16-bit character strings (a large amount of work involving
865        many changes and refactorings).
866    
867    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
868        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
869        from a file.
870    
871    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
872        rounding is not applied in this particular case).
873    
874    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
875        if they appear, or are escaped, in patterns.
876    
877    10. Get rid of a number of -Wunused-but-set-variable warnings.
878    
879    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
880        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
881        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
882        also returns the mark "x". This bug applied to capturing parentheses,
883        non-capturing parentheses, and atomic parentheses. It also applied to some
884        assertions.
885    
886    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
887        information out of configure.ac instead of relying on pcre.h.generic, which
888        is not stored in the repository.
889    
890    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
891        -lreadline.
892    
893    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
894    
895    15. Applied Graycode's patch to put the top-level frame on the stack rather
896        than the heap when not using the stack for recursion. This gives a
897        performance improvement in many cases when recursion is not deep.
898    
899    16. Experimental code added to "pcretest -C" to output the stack frame size.
900    
901    
902    Version 8.21 12-Dec-2011
903    ------------------------
904    
905    1.  Updating the JIT compiler.
906    
907    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
908        are added as well.
909    
910    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
911        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
912        calling _pcre_jit_exec. Some extra comments are added.
913    
914    4.  (*MARK) settings inside atomic groups that do not contain any capturing
915        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
916        was introduced by change 18 for 8.20.
917    
918    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
919        ECMA-262 standard.
920    
921    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
922        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
923        This bug was probably introduced by change 9 of 8.13.
924    
925    7.  While fixing 6 above, I noticed that a number of other items were being
926        incorrectly rejected as "not fixed length". This arose partly because newer
927        opcodes had not been added to the fixed-length checking code. I have (a)
928        corrected the bug and added tests for these items, and (b) arranged for an
929        error to occur if an unknown opcode is encountered while checking for fixed
930        length instead of just assuming "not fixed length". The items that were
931        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
932        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
933        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
934    
935    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
936        being incorrectly compiled and would have given unpredicatble results.
937    
938    9.  A possessively repeated subpattern with minimum repeat count greater than
939        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
940        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
941        the first (A) could occur when it should not.
942    
943    10. Add a cast and remove a redundant test from the code.
944    
945    11. JIT should use pcre_malloc/pcre_free for allocation.
946    
947    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
948        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
949        is anything other than /usr, -L is still shown).
950    
951    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
952    
953    14. Perl does not support \N without a following name in a [] class; PCRE now
954        also gives an error.
955    
956    15. If a forward reference was repeated with an upper limit of around 2000,
957        it caused the error "internal error: overran compiling workspace". The
958        maximum number of forward references (including repeats) was limited by the
959        internal workspace, and dependent on the LINK_SIZE. The code has been
960        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
961        the default depends on LINK_SIZE. There is a new upper limit (for safety)
962        of around 200,000 forward references. While doing this, I also speeded up
963        the filling in of repeated forward references.
964    
965    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
966        incorrectly expecting the subject to contain another "a" after the start.
967    
968    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
969        in the match, the SKIP should be ignored. This was not happening; instead
970        the SKIP was being treated as NOMATCH. For patterns such as
971        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
972        tested.
973    
974    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
975        now much more compatible with Perl, in particular in cases where the result
976        is a non-match for a non-anchored pattern. For example, if
977        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
978        "m", where previously it did not return a name. A side effect of this
979        change is that for partial matches, the last encountered mark name is
980        returned, as for non matches. A number of tests that were previously not
981        Perl-compatible have been moved into the Perl-compatible test files. The
982        refactoring has had the pleasing side effect of removing one argument from
983        the match() function, thus reducing its stack requirements.
984    
985    19. If the /S+ option was used in pcretest to study a pattern using JIT,
986        subsequent uses of /S (without +) incorrectly behaved like /S+.
987    
988    21. Retrieve executable code size support for the JIT compiler and fixing
989        some warnings.
990    
991    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
992        not work when the shorter character appeared right at the end of the
993        subject string.
994    
995    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
996        systems.
997    
998    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
999        output it when the /M option is used in pcretest.
1000    
1001    25. The CheckMan script was not being included in the distribution. Also, added
1002        an explicit "perl" to run Perl scripts from the PrepareRelease script
1003        because this is reportedly needed in Windows.
1004    
1005    26. If study data was being save in a file and studying had not found a set of
1006        "starts with" bytes for the pattern, the data written to the file (though
1007        never used) was taken from uninitialized memory and so caused valgrind to
1008        complain.
1009    
1010    27. Updated RunTest.bat as provided by Sheri Pierce.
1011    
1012    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1013    
1014    29. Computation of memory usage for the table of capturing group names was
1015        giving an unnecessarily large value.
1016    
1017    
1018    Version 8.20 21-Oct-2011
1019    ------------------------
1020    
1021    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1022        a POSIX class. After further experiments with Perl, which convinced me that
1023        Perl has bugs and confusions, a closing square bracket is no longer allowed
1024        in a POSIX name. This bug also affected patterns with classes that started
1025        with full stops.
1026    
1027    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1028        captured substring, but while checking the failing first alternative,
1029        substring 1 is temporarily captured. If the output vector supplied to
1030        pcre_exec() was not big enough for this capture, the yield of the function
1031        was still zero ("insufficient space for captured substrings"). This cannot
1032        be totally fixed without adding another stack variable, which seems a lot
1033        of expense for a edge case. However, I have improved the situation in cases
1034        such as /(a)(b)x|abc/ matched against "abc", where the return code
1035        indicates that fewer than the maximum number of slots in the ovector have
1036        been set.
1037    
1038    3.  Related to (2) above: when there are more back references in a pattern than
1039        slots in the output vector, pcre_exec() uses temporary memory during
1040        matching, and copies in the captures as far as possible afterwards. It was
1041        using the entire output vector, but this conflicts with the specification
1042        that only 2/3 is used for passing back captured substrings. Now it uses
1043        only the first 2/3, for compatibility. This is, of course, another edge
1044        case.
1045    
1046    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1047        main code base, and can be used by building with --enable-jit. When this is
1048        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1049        runtime --no-jit option is given.
1050    
1051    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1052        ovector, the return from the function was zero, implying that there were
1053        other matches that did not fit. The correct "exactly full" value is now
1054        returned.
1055    
1056    6.  If a subpattern that was called recursively or as a subroutine contained
1057        (*PRUNE) or any other control that caused it to give a non-standard return,
1058        invalid errors such as "Error -26 (nested recursion at the same subject
1059        position)" or even infinite loops could occur.
1060    
1061    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1062        computing the minimum length on reaching *ACCEPT, and so ended up with the
1063        wrong value of 1 rather than 0. Further investigation indicates that
1064        computing a minimum subject length in the presence of *ACCEPT is difficult
1065        (think back references, subroutine calls), and so I have changed the code
1066        so that no minimum is registered for a pattern that contains *ACCEPT.
1067    
1068    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1069        it was not handled as intended. [But see 16 below.]
1070    
1071    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1072        Sheri Pierce.
1073    
1074    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1075        the first byte in a match must be "a".
1076    
1077    11. Change 17 for 8.13 increased the recursion depth for patterns like
1078        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1079        pattern contains any instances of (*THEN). If it does not, the old
1080        optimizations are restored. It would be nice to do this on a per-group
1081        basis, but at the moment that is not feasible.
1082    
1083    12. In some environments, the output of pcretest -C is CRLF terminated. This
1084        broke RunTest's code that checks for the link size. A single white space
1085        character after the value is now allowed for.
1086    
1087    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1088        For "fr", it uses the Windows-specific input and output files.
1089    
1090    14. If (*THEN) appeared in a group that was called recursively or as a
1091        subroutine, it did not work as intended. [But see next item.]
1092    
1093    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1094        pattern fragments (but not containing any | characters). If A and B are
1095        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1096        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1097        D. In other words, Perl considers parentheses that do not contain any |
1098        characters to be part of a surrounding alternative, whereas PCRE was
1099        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1100        differently. PCRE now behaves in the same way as Perl, except in the case
1101        of subroutine/recursion calls such as (?1) which have in any case always
1102        been different (but PCRE had them first :-).
1103    
1104    16. Related to 15 above: Perl does not treat the | in a conditional group as
1105        creating alternatives. Such a group is treated in the same way as an
1106        ordinary group without any | characters when processing (*THEN). PCRE has
1107        been changed to match Perl's behaviour.
1108    
1109    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1110        RunGrepTest script failed.
1111    
1112    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1113        inevitable for groups that contain captures, but it can lead to a lot of
1114        stack use in large patterns. The old behaviour has been restored for atomic
1115        groups that do not contain any capturing parentheses.
1116    
1117    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1118        suppress the check for a minimum subject length at run time. (If it was
1119        given to pcre_exec() or pcre_dfa_exec() it did work.)
1120    
1121    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1122        fail to work when decoding hex characters in data strings in EBCDIC
1123        environments.
1124    
1125    21. It appears that in at least one Mac OS environment, the isxdigit() function
1126        is implemented as a macro that evaluates to its argument more than once,
1127        contravening the C 90 Standard (I haven't checked a later standard). There
1128        was an instance in pcretest which caused it to go wrong when processing
1129        \x{...} escapes in subject strings. The has been rewritten to avoid using
1130        things like p++ in the argument of isxdigit().
1131    
1132    
1133    Version 8.13 16-Aug-2011
1134    ------------------------
1135    
1136    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1137    
1138    2.  Two minor typos in pcre_internal.h have been fixed.
1139    
1140    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1141        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1142        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1143    
1144    4.  There were a number of related bugs in the code for matching backrefences
1145        caselessly in UTF-8 mode when codes for the characters concerned were
1146        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1147        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1148        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1149        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1150        2-byte code at the end of the subject (it thought there wasn't enough data
1151        left).
1152    
1153    5.  Comprehensive information about what went wrong is now returned by
1154        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1155        as the output vector has at least 2 elements. The offset of the start of
1156        the failing character and a reason code are placed in the vector.
1157    
1158    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1159        now returned is for the first byte of the failing character, instead of the
1160        last byte inspected. This is an incompatible change, but I hope it is small
1161        enough not to be a problem. It makes the returned offset consistent with
1162        pcre_exec() and pcre_dfa_exec().
1163    
1164    7.  pcretest now gives a text phrase as well as the error number when
1165        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1166        failure, the offset and reason code are output.
1167    
1168    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1169        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1170        back over a single character (\n). This seems wrong (because it treated the
1171        two characters as a single entity when going forwards), conflicts with the
1172        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1173        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1174        has been changed.
1175    
1176    9.  Some internal refactoring has changed the processing so that the handling
1177        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1178        time (the PCRE_DOTALL option was changed this way some time ago: version
1179        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1180        which was always a bit of a fudge. It also means that there is one less
1181        argument for the match() function, which reduces its stack requirements
1182        slightly. This change also fixes an incompatibility with Perl: the pattern
1183        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1184    
1185    10. More internal refactoring has drastically reduced the number of recursive
1186        calls to match() for possessively repeated groups such as (abc)++ when
1187        using pcre_exec().
1188    
1189    11. While implementing 10, a number of bugs in the handling of groups were
1190        discovered and fixed:
1191    
1192        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1193        (a|)*(?1) gave a compile-time internal error.
1194        ((a|)+)+  did not notice that the outer group could match an empty string.
1195        (^a|^)+   was not marked as anchored.
1196        (.*a|.*)+ was not marked as matching at start or after a newline.
1197    
1198    12. Yet more internal refactoring has removed another argument from the match()
1199        function. Special calls to this function are now indicated by setting a
1200        value in a variable in the "match data" data block.
1201    
1202    13. Be more explicit in pcre_study() instead of relying on "default" for
1203        opcodes that mean there is no starting character; this means that when new
1204        ones are added and accidentally left out of pcre_study(), testing should
1205        pick them up.
1206    
1207    14. The -s option of pcretest has been documented for ages as being an old
1208        synonym of -m (show memory usage). I have changed it to mean "force study
1209        for every regex", that is, assume /S for every regex. This is similar to -i
1210        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1211        using it. It makes it easier to run collections of tests with and without
1212        study enabled, and thereby test pcre_study() more easily. All the standard
1213        tests are now run with and without -s (but some patterns can be marked as
1214        "never study" - see 20 below).
1215    
1216    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1217        restoration of the capturing data to the outer values was not happening
1218        correctly.
1219    
1220    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1221        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1222        pattern had matched an empty string, and so incorrectly returned a no
1223        match.
1224    
1225    17. There was optimizing code for the last branch of non-capturing parentheses,
1226        and also for the obeyed branch of a conditional subexpression, which used
1227        tail recursion to cut down on stack usage. Unfortunately, now that there is
1228        the possibility of (*THEN) occurring in these branches, tail recursion is
1229        no longer possible because the return has to be checked for (*THEN). These
1230        two optimizations have therefore been removed. [But see 8.20/11 above.]
1231    
1232    18. If a pattern containing \R was studied, it was assumed that \R always
1233        matched two bytes, thus causing the minimum subject length to be
1234        incorrectly computed because \R can also match just one byte.
1235    
1236    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1237        was incorrectly computed.
1238    
1239    20. If /S is present twice on a test pattern in pcretest input, it now
1240        *disables* studying, thereby overriding the use of -s on the command line
1241        (see 14 above). This is necessary for one or two tests to keep the output
1242        identical in both cases.
1243    
1244    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1245        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1246    
1247    22. When an atomic group that contained a capturing parenthesis was
1248        successfully matched, but the branch in which it appeared failed, the
1249        capturing was not being forgotten if a higher numbered group was later
1250        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1251        group 1 to "a", when in fact it should be unset. This applied to multi-
1252        branched capturing and non-capturing groups, repeated or not, and also to
1253        positive assertions (capturing in negative assertions does not happen
1254        in PCRE) and also to nested atomic groups.
1255    
1256    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1257        subject after a captured substring, to make it easier to tell which of a
1258        number of identical substrings has been captured.
1259    
1260    24. The way atomic groups are processed by pcre_exec() has been changed so that
1261        if they are repeated, backtracking one repetition now resets captured
1262        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1263        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1264        "aaa". Previously, it would have been "a". As part of this code
1265        refactoring, the way recursive calls are handled has also been changed.
1266    
1267    25. If an assertion condition captured any substrings, they were not passed
1268        back unless some other capturing happened later. For example, if
1269        (?(?=(a))a) was matched against "a", no capturing was returned.
1270    
1271    26. When studying a pattern that contained subroutine calls or assertions,
1272        the code for finding the minimum length of a possible match was handling
1273        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1274        group 1 called group 2 while simultaneously a separate group 2 called group
1275        1). A stack overflow occurred in this case. I have fixed this by limiting
1276        the recursion depth to 10.
1277    
1278    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1279        Fortmann. This supports explicit test numbers on the command line, and has
1280        argument validation and error reporting.
1281    
1282    28. An instance of \X with an unlimited repeat could fail if at any point the
1283        first character it looked at was a mark character.
1284    
1285    29. Some minor code refactoring concerning Unicode properties and scripts
1286        should reduce the stack requirement of match() slightly.
1287    
1288    30. Added the '=' option to pcretest to check the setting of unused capturing
1289        slots at the end of the pattern, which are documented as being -1, but are
1290        not included in the return count.
1291    
1292    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1293        compiled something random. Now it gives a compile-time error (as does
1294        Perl).
1295    
1296    32. A *MARK encountered during the processing of a positive assertion is now
1297        recorded and passed back (compatible with Perl).
1298    
1299    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1300        had alternative anchored branches, the search for a second match in a line
1301        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1302        matched the line "0102" twice. The same bug affected patterns that started
1303        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1304        twice.
1305    
1306    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1307        does, and because of capturing effects, quantifying parenthesized
1308        assertions may at times be useful. Quantifiers are now allowed for
1309        parenthesized assertions.
1310    
1311    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1312    
1313    36. \g was being checked for fancy things in a character class, when it should
1314        just be a literal "g".
1315    
1316    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1317        appearance of a nested POSIX class supersedes an apparent external class.
1318        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1319        unescaped square brackets may also appear as part of class names. For
1320        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1321        more like Perl. (But see 8.20/1 above.)
1322    
1323    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1324        was because it thought it was \N{name}, which is not supported).
1325    
1326    39. Add minix to OS list not supporting the -S option in pcretest.
1327    
1328    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1329        cannot analyze patterns in sufficient detail to catch mutual recursions
1330        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1331        subgroup is called recursively as a subpattern for a second time at the
1332        same position in the subject string. In previous releases this might have
1333        been caught by the recursion limit, or it might have run out of stack.
1334    
1335    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1336        happen only once. PCRE was, however incorrectly giving a compile time error
1337        "recursive call could loop indefinitely" because it cannot analyze the
1338        pattern in sufficient detail. The compile time test no longer happens when
1339        PCRE is compiling a conditional subpattern, but actual runaway loops are
1340        now caught at runtime (see 40 above).
1341    
1342    42. It seems that Perl allows any characters other than a closing parenthesis
1343        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1344        has been changed to be the same.
1345    
1346    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1347        as not to get warnings when autogen.sh is called. Also changed
1348        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1349    
1350    44. To help people who use pcregrep to scan files containing exceedingly long
1351        lines, the following changes have been made:
1352    
1353        (a) The default value of the buffer size parameter has been increased from
1354            8K to 20K. (The actual buffer used is three times this size.)
1355    
1356        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1357            PCRE is built.
1358    
1359        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1360            to be set at run time.
1361    
1362        (d) Numerical values in pcregrep options can be followed by K or M, for
1363            example --buffer-size=50K.
1364    
1365        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1366            given and the return code is set to 2.
1367    
1368    45. Add a pointer to the latest mark to the callout data block.
1369    
1370    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1371        partial match of an empty string instead of no match. This was specific to
1372        the use of ".".
1373    
1374    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1375        complete match instead of a partial match. This bug was dependent on both
1376        the PCRE_UTF8 and PCRE_DOTALL options being set.
1377    
1378    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1379        starting byte set, because \b was not being ignored.
1380    
1381    
1382    Version 8.12 15-Jan-2011
1383    ------------------------
1384    
1385    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1386        checks for such things as part of the documentation building process.
1387    
1388    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1389        --match-limit and --recursion-limit options (added for 8.11). In
1390        particular, this made one of the standard tests fail. (The integer value
1391        went into the wrong half of a long int.)
1392    
1393    3.  If the --colour option was given to pcregrep with -v (invert match), it
1394        did strange things, either producing crazy output, or crashing. It should,
1395        of course, ignore a request for colour when reporting lines that do not
1396        match.
1397    
1398    4.  Another pcregrep bug caused similar problems if --colour was specified with
1399        -M (multiline) and the pattern match finished with a line ending.
1400    
1401    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1402        matched in multiline mode, the following line was shown as part of the
1403        match. This seems wrong, so I have changed it.
1404    
1405    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1406        the check for further matches in the same line (so they could be coloured)
1407        to overrun the end of the current line. If another match was found, it was
1408        incorrectly shown (and then shown again when found in the next line).
1409    
1410    7.  If pcregrep was compiled under Windows, there was a reference to the
1411        function pcregrep_exit() before it was defined. I am assuming this was
1412        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1413        reported by a user. I've moved the definition above the reference.
1414    
1415    
1416    Version 8.11 10-Dec-2010
1417    ------------------------
1418    
1419    1.  (*THEN) was not working properly if there were untried alternatives prior
1420        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1421        backtracked to try for "b" instead of moving to the next alternative branch
1422        at the same level (in this case, to look for "c"). The Perl documentation
1423        is clear that when (*THEN) is backtracked onto, it goes to the "next
1424        alternative in the innermost enclosing group".
1425    
1426    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1427        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1428        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1429        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1430        (*THEN).
1431    
1432    3.  If \s appeared in a character class, it removed the VT character from
1433        the class, even if it had been included by some previous item, for example
1434        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1435        of \s, but is part of the POSIX "space" class.)
1436    
1437    4.  A partial match never returns an empty string (because you can always
1438        match an empty string at the end of the subject); however the checking for
1439        an empty string was starting at the "start of match" point. This has been
1440        changed to the "earliest inspected character" point, because the returned
1441        data for a partial match starts at this character. This means that, for
1442        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1443        (previously it gave "no match").
1444    
1445    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1446        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1447        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1448        has an implication that the given string is incomplete (because a partial
1449        match is preferred over a full match). For this reason, these items now
1450        give a partial match in this situation. [Aside: previously, the one case
1451        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1452        match rather than a full match, which was wrong by the old rules, but is
1453        now correct.]
1454    
1455    6.  There was a bug in the handling of #-introduced comments, recognized when
1456        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1457        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1458        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1459        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1460        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1461        places in pcre_compile().
1462    
1463    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1464        comments when looking ahead for named forward references to subpatterns,
1465        the only newline sequence it recognized was NL. It now handles newlines
1466        according to the set newline convention.
1467    
1468    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1469        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1470        cater for a lack of strerror(). These oversights have been fixed.
1471    
1472    9.  Added --match-limit and --recursion-limit to pcregrep.
1473    
1474    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1475    
1476    11. When the -o option was used, pcregrep was setting a return code of 1, even
1477        when matches were found, and --line-buffered was not being honoured.
1478    
1479    12. Added an optional parentheses number to the -o and --only-matching options
1480        of pcregrep.
1481    
1482    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1483        can match an empty string. The code to do it in pcretest and pcredemo
1484        needed fixing:
1485    
1486        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1487            only one byte after an empty string match just before CRLF (this case
1488            just got forgotten; "any" and "anycrlf" were OK).
1489    
1490        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1491            mode when an empty string match preceded an ASCII character followed by
1492            a non-ASCII character. (The code for advancing by one character rather
1493            than one byte was nonsense.)
1494    
1495        (c) The pcredemo.c sample program did not have any code at all to handle
1496            the cases when CRLF is a valid newline sequence.
1497    
1498    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1499        as a starting offset was within the subject string. There is now a new
1500        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1501        negative or greater than the length of the string. In order to test this,
1502        pcretest is extended to allow the setting of negative starting offsets.
1503    
1504    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1505        starting offset points to the beginning of a UTF-8 character was
1506        unnecessarily clumsy. I tidied it up.
1507    
1508    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1509        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1510    
1511    17. Nobody had reported that the --include_dir option, which was added in
1512        release 7.7 should have been called --include-dir (hyphen, not underscore)
1513        for compatibility with GNU grep. I have changed it to --include-dir, but
1514        left --include_dir as an undocumented synonym, and the same for
1515        --exclude-dir, though that is not available in GNU grep, at least as of
1516        release 2.5.4.
1517    
1518    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1519        characters from a string of bytes) have been redefined so as not to use
1520        loops, in order to improve performance in some environments. At the same
1521        time, I abstracted some of the common code into auxiliary macros to save
1522        repetition (this should not affect the compiled code).
1523    
1524    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1525        compile-time error is now given if \c is not followed by an ASCII
1526        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1527        different, and any byte value is allowed.)
1528    
1529    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1530        START_OPTIMIZE option, which is now allowed at compile time - but just
1531        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1532        to pcregrep and other applications that have no direct access to PCRE
1533        options. The new /Y option in pcretest sets this option when calling
1534        pcre_compile().
1535    
1536    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1537        back references. Groups containing recursive back references were forced to
1538        be atomic by that change, but in the case of named groups, the amount of
1539        memory required was incorrectly computed, leading to "Failed: internal
1540        error: code overflow". This has been fixed.
1541    
1542    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1543        pcretest.c, to avoid build problems in some Borland environments.
1544    
1545    
1546    Version 8.10 25-Jun-2010
1547    ------------------------
1548    
1549    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1550        THEN.
1551    
1552    2.  (*ACCEPT) was not working when inside an atomic group.
1553    
1554    3.  Inside a character class, \B is treated as a literal by default, but
1555        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1556        causes the error). The code is unchanged, but I tidied the documentation.
1557    
1558    4.  Inside a character class, PCRE always treated \R and \X as literals,
1559        whereas Perl faults them if its -w option is set. I have changed PCRE so
1560        that it faults them when PCRE_EXTRA is set.
1561    
1562    5.  Added support for \N, which always matches any character other than
1563        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1564    
1565    6.  When compiling pcregrep with newer versions of gcc which may have
1566        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1567        declared with attribute warn_unused_result" were given. Just casting the
1568        result to (void) does not stop the warnings; a more elaborate fudge is
1569        needed. I've used a macro to implement this.
1570    
1571    7.  Minor change to pcretest.c to avoid a compiler warning.
1572    
1573    8.  Added four artifical Unicode properties to help with an option to make
1574        \s etc use properties (see next item). The new properties are: Xan
1575        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1576    
1577    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1578        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1579        this option. Modified pcretest to add /W to test this facility. Added
1580        REG_UCP to make it available via the POSIX interface.
1581    
1582    10. Added --line-buffered to pcregrep.
1583    
1584    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1585        studied, and the match started with a letter with a code point greater than
1586        127 whose first byte was different to the first byte of the other case of
1587        the letter, the other case of this starting letter was not recognized
1588        (#976).
1589    
1590    12. If a pattern that was studied started with a repeated Unicode property
1591        test, for example, \p{Nd}+, there was the theoretical possibility of
1592        setting up an incorrect bitmap of starting bytes, but fortunately it could
1593        not have actually happened in practice until change 8 above was made (it
1594        added property types that matched character-matching opcodes).
1595    
1596    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1597        possible starting bytes for non-anchored patterns.
1598    
1599    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1600        \R, and also a number of cases that involve Unicode properties, both
1601        explicit and implicit when PCRE_UCP is set.
1602    
1603    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1604        input, it could crash or give wrong results if characters with values
1605        greater than 0xc0 were present in the subject string. (Detail: it assumed
1606        UTF-8 input when processing these items.)
1607    
1608    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1609        size_t is 64-bit (#991).
1610    
1611    17. Added a check for running out of memory when PCRE is compiled with
1612        --disable-stack-for-recursion (#990).
1613    
1614    18. If the last data line in a file for pcretest does not have a newline on
1615        the end, a newline was missing in the output.
1616    
1617    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1618        less than 128) in its various bitmaps. However, there is a facility for
1619        generating tables according to the current locale when PCRE is compiled. It
1620        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1621        characters, are recognized by isspace() and therefore were getting set in
1622        these tables, and indeed these tables seem to approximate to ISO 8859. This
1623        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1624        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1625        which of course cannot start UTF-8 characters. I have changed the code so
1626        that only real ASCII characters (less than 128) and the correct starting
1627        bytes for UTF-8 encodings are set for characters greater than 127 when in
1628        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1629        altogether.)
1630    
1631    20. Added the /T option to pcretest so as to be able to run tests with non-
1632        standard character tables, thus making it possible to include the tests
1633        used for 19 above in the standard set of tests.
1634    
1635    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1636        reference to a subpattern the other side of a comment that contains an
1637        opening parenthesis caused either an internal compiling error, or a
1638        reference to the wrong subpattern.
1639    
1640    
1641    Version 8.02 19-Mar-2010
1642    ------------------------
1643    
1644    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1645    
1646    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1647        configured.
1648    
1649    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1650        original author of that file, following a query about its status.
1651    
1652    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1653        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1654    
1655    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1656        quantifier applied to a forward-referencing subroutine call, could compile
1657        incorrect code or give the error "internal error: previously-checked
1658        referenced subpattern not found".
1659    
1660    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1661        variables to point to external functions. For these systems, therefore,
1662        pcre_malloc etc. are now initialized to local functions that call the
1663        relevant global functions.
1664    
1665    7.  There were two entries missing in the vectors called coptable and poptable
1666        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1667        I've fixed the data, and added a kludgy way of testing at compile time that
1668        the lengths are correct (equal to the number of opcodes).
1669    
1670    8.  Following on from 7, I added a similar kludge to check the length of the
1671        eint vector in pcreposix.c.
1672    
1673    9.  Error texts for pcre_compile() are held as one long string to avoid too
1674        much relocation at load time. To find a text, the string is searched,
1675        counting zeros. There was no check for running off the end of the string,
1676        which could happen if a new error number was added without updating the
1677        string.
1678    
1679    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1680    
1681    11. \K was not working if it appeared in an atomic group or in a group that
1682        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1683        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1684        the assertion is positive, but not if it is negative.
1685    
1686    12. Change 11 fortuitously reduced the size of the stack frame used in the
1687        "match()" function of pcre_exec.c by one pointer. Forthcoming
1688        implementation of support for (*MARK) will need an extra pointer on the
1689        stack; I have reserved it now, so that the stack frame size does not
1690        decrease.
1691    
1692    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1693        item in branch that calls a recursion is a subroutine call - as in the
1694        second branch in the above example - was incorrectly given the compile-
1695        time error "recursive call could loop indefinitely" because pcre_compile()
1696        was not correctly checking the subroutine for matching a non-empty string.
1697    
1698    14. The checks for overrunning compiling workspace could trigger after an
1699        overrun had occurred. This is a "should never occur" error, but it can be
1700        triggered by pathological patterns such as hundreds of nested parentheses.
1701        The checks now trigger 100 bytes before the end of the workspace.
1702    
1703    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1704    
1705    
1706    Version 8.01 19-Jan-2010
1707    ------------------------
1708    
1709    1.  If a pattern contained a conditional subpattern with only one branch (in
1710        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1711        computed the wrong minimum data length (which is of course zero for such
1712        subpatterns). This could cause incorrect "no match" results.
1713    
1714    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1715        the pattern is reset in the first branch, pcre_compile() failed with
1716        "internal error: code overflow at offset...". This happened only when
1717        the reset was to the original external option setting. (An optimization
1718        abstracts leading options settings into an external setting, which was the
1719        cause of this.)
1720    
1721    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1722        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1723        assertion pattern did not match (meaning that the assertion was true), it
1724        was incorrectly treated as false if the SKIP had been reached during the
1725        matching. This also applied to assertions used as conditions.
1726    
1727    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1728        assertion subpattern, including such a pattern used as a condition,
1729        unpredictable results occurred, instead of the error return
1730        PCRE_ERROR_DFA_UITEM.
1731    
1732    5.  The C++ GlobalReplace function was not working like Perl for the special
1733        situation when an empty string is matched. It now does the fancy magic
1734        stuff that is necessary.
1735    
1736    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1737        removed. (These were left over from very, very early versions of PCRE.)
1738    
1739    7.  Some cosmetic changes to the code to make life easier when compiling it
1740        as part of something else:
1741    
1742        (a) Change DEBUG to PCRE_DEBUG.
1743    
1744        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1745            called "current" as "current_branch", to prevent a collision with the
1746            Linux macro when compiled as a kernel module.
1747    
1748        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1749            prevent a collision with the Linux macro when compiled as a kernel
1750            module.
1751    
1752    8.  In pcre_compile() there are some checks for integer overflows that used to
1753        cast potentially large values to (double). This has been changed to that
1754        when building, a check for int64_t is made, and if it is found, it is used
1755        instead, thus avoiding the use of floating point arithmetic. (There is no
1756        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1757        double.
1758    
1759    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1760        2005 (difference between two addresses compared to an unsigned value).
1761    
1762    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1763        custom one, because of the following reported problem in Windows:
1764    
1765          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1766              under Win32.
1767          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1768              therefore missing the function definition.
1769          - The compiler thus generates a "C" signature for the test function.
1770          - The linker fails to find the "C" function.
1771          - PCRE fails to configure if asked to do so against libbz2.
1772    
1773    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1774        messages were output:
1775    
1776          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1777          rerunning libtoolize, to keep the correct libtool macros in-tree.
1778          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1779    
1780        I have done both of these things.
1781    
1782    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1783        most of the time, it *can* run out if it is given a pattern that contains a
1784        runaway infinite recursion. I updated the discussion in the pcrestack man
1785        page.
1786    
1787    13. Now that we have gone to the x.xx style of version numbers, the minor
1788        version may start with zero. Using 08 or 09 is a bad idea because users
1789        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1790        interpreted as invalid octal numbers. I've updated the previous comment in
1791        configure.ac, and also added a check that gives an error if 08 or 09 are
1792        used.
1793    
1794    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1795        causing partial matching to fail when the end of the subject matched \W
1796        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1797    
1798    15. There were some discrepancies between the declarations in pcre_internal.h
1799        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1800        their definitions. The declarations used "const uschar *" and the
1801        definitions used USPTR. Even though USPTR is normally defined as "const
1802        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1803        reported that: "This difference in casting confuses some C++ compilers, for
1804        example, SunCC recognizes above declarations as different functions and
1805        generates broken code for hbpcre." I have changed the declarations to use
1806        USPTR.
1807    
1808    16. GNU libtool is named differently on some systems. The autogen.sh script now
1809        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1810        (FreeBSD).
1811    
1812    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1813        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1814        comment: "Figure out how to create a longlong from a string: strtoll and
1815        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1816        instance, but it only takes 2 args instead of 3!"
1817    
1818    18. A subtle bug concerned with back references has been fixed by a change of
1819        specification, with a corresponding code fix. A pattern such as
1820        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1821        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1822        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1823        same bug. Such groups have to be quantified to be useful, or contained
1824        inside another quantified group. (If there's no repetition, the reference
1825        can never match.) The problem arises because, having left the group and
1826        moved on to the rest of the pattern, a later failure that backtracks into
1827        the group uses the captured value from the final iteration of the group
1828        rather than the correct earlier one. I have fixed this in PCRE by forcing
1829        any group that contains a reference to itself to be an atomic group; that
1830        is, there cannot be any backtracking into it once it has completed. This is
1831        similar to recursive and subroutine calls.
1832    
1833    
1834    Version 8.00 19-Oct-09
1835    ----------------------
1836    
1837    1.  The table for translating pcre_compile() error codes into POSIX error codes
1838        was out-of-date, and there was no check on the pcre_compile() error code
1839        being within the table. This could lead to an OK return being given in
1840        error.
1841    
1842    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1843        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1844        in a Windows environment.
1845    
1846    3.  The pcregrep --count option prints the count for each file even when it is
1847        zero, as does GNU grep. However, pcregrep was also printing all files when
1848        --files-with-matches was added. Now, when both options are given, it prints
1849        counts only for those files that have at least one match. (GNU grep just
1850        prints the file name in this circumstance, but including the count seems
1851        more useful - otherwise, why use --count?) Also ensured that the
1852        combination -clh just lists non-zero counts, with no names.
1853    
1854    4.  The long form of the pcregrep -F option was incorrectly implemented as
1855        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1856        but it seems right to fix it, and I didn't think it was worth preserving
1857        the old behaviour.
1858    
1859    5.  The command line items --regex=pattern and --regexp=pattern were not
1860        recognized by pcregrep, which required --regex pattern or --regexp pattern
1861        (with a space rather than an '='). The man page documented the '=' forms,
1862        which are compatible with GNU grep; these now work.
1863    
1864    6.  No libpcreposix.pc file was created for pkg-config; there was just
1865        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1866    
1867    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1868        when UCP support is not needed, by modifying the Python script that
1869        generates it from Unicode data files. This should not matter if the module
1870        is correctly used as a library, but I received one complaint about 50K of
1871        unwanted data. My guess is that the person linked everything into his
1872        program rather than using a library. Anyway, it does no harm.
1873    
1874    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1875        was a minimum greater than 1 for a wide character in a possessive
1876        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1877        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1878        character. Chaos in the form of incorrect output or a compiling loop could
1879        result.
1880    
1881    9.  The restrictions on what a pattern can contain when partial matching is
1882        requested for pcre_exec() have been removed. All patterns can now be
1883        partially matched by this function. In addition, if there are at least two
1884        slots in the offset vector, the offset of the earliest inspected character
1885        for the match and the offset of the end of the subject are set in them when
1886        PCRE_ERROR_PARTIAL is returned.
1887    
1888    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1889        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1890        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1891        and may be more useful for multi-segment matching.
1892    
1893    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1894        used to be given if ever the end of the subject was reached; now it is
1895        given only if matching could not proceed because another character was
1896        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1897        string "Z", which now yields "no match" instead of "partial match". In the
1898        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1899        final character ended with (*FAIL).
1900    
1901    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1902        if the pattern had a "must contain" character that was already found in the
1903        earlier partial match, unless partial matching was again requested. For
1904        example, with the pattern /dog.(body)?/, the "must contain" character is
1905        "g". If the first part-match was for the string "dog", restarting with
1906        "sbody" failed. This bug has been fixed.
1907    
1908    13. The string returned by pcre_dfa_exec() after a partial match has been
1909        changed so that it starts at the first inspected character rather than the
1910        first character of the match. This makes a difference only if the pattern
1911        starts with a lookbehind assertion or \b or \B (\K is not supported by
1912        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1913        matching functions compatible, and I think it's the right thing to do.
1914    
1915    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1916        so that the demonstration program is easily available in environments where
1917        PCRE has not been installed from source.
1918    
1919    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1920        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1921        library.
1922    
1923    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1924        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1925        is not the first non-POSIX option to be added. Clearly some people find
1926        these options useful.
1927    
1928    17. If a caller to the POSIX matching function regexec() passes a non-zero
1929        value for nmatch with a NULL value for pmatch, the value of
1930        nmatch is forced to zero.
1931    
1932    18. RunGrepTest did not have a test for the availability of the -u option of
1933        the diff command, as RunTest does. It now checks in the same way as
1934        RunTest, and also checks for the -b option.
1935    
1936    19. If an odd number of negated classes containing just a single character
1937        interposed, within parentheses, between a forward reference to a named
1938        subpattern and the definition of the subpattern, compilation crashed with
1939        an internal error, complaining that it could not find the referenced
1940        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1941        [The bug was that it was starting one character too far in when skipping
1942        over the character class, thus treating the ] as data rather than
1943        terminating the class. This meant it could skip too much.]
1944    
1945    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1946        /g option in pcretest when the pattern contains \K, which makes it possible
1947        to have an empty string match not at the start, even when the pattern is
1948        anchored. Updated pcretest and pcredemo to use this option.
1949    
1950    21. If the maximum number of capturing subpatterns in a recursion was greater
1951        than the maximum at the outer level, the higher number was returned, but
1952        with unset values at the outer level. The correct (outer level) value is
1953        now given.
1954    
1955    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1956        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1957        make it do so. The string so far is captured, making this feature
1958        compatible with Perl.
1959    
1960    23. The tests have been re-organized, adding tests 11 and 12, to make it
1961        possible to check the Perl 5.10 features against Perl 5.10.
1962    
1963    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1964        pattern matches a fixed length string. PCRE did not allow this; now it
1965        does. Neither allows recursion.
1966    
1967    25. I finally figured out how to implement a request to provide the minimum
1968        length of subject string that was needed in order to match a given pattern.
1969        (It was back references and recursion that I had previously got hung up
1970        on.) This code has now been added to pcre_study(); it finds a lower bound
1971        to the length of subject needed. It is not necessarily the greatest lower
1972        bound, but using it to avoid searching strings that are too short does give
1973        some useful speed-ups. The value is available to calling programs via
1974        pcre_fullinfo().
1975    
1976    26. While implementing 25, I discovered to my embarrassment that pcretest had
1977        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1978        study optimizations had never been tested with that matching function.
1979        Oops. What is worse, even when it was passed study data, there was a bug in
1980        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1981        were also very few tests of studied patterns with pcre_dfa_exec().
1982    
1983    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1984        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1985        on the other side of the coin, they are no longer allowed to have different
1986        names, because these cannot be distinguished in PCRE, and this has caused
1987        confusion. (This is a difference from Perl.)
1988    
1989    28. When duplicate subpattern names are present (necessarily with different
1990        numbers, as required by 27 above), and a test is made by name in a
1991        conditional pattern, either for a subpattern having been matched, or for
1992        recursion in such a pattern, all the associated numbered subpatterns are
1993        tested, and the overall condition is true if the condition is true for any
1994        one of them. This is the way Perl works, and is also more like the way
1995        testing by number works.
1996    
1997    
1998    Version 7.9 11-Apr-09
1999    ---------------------
2000    
2001    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2002        (pcretest), all targets were linked against these libraries. This included
2003        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2004        libraries. This caused unwanted dependencies to be created. This problem
2005        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2006        pcretest is linked with readline.
2007    
2008    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2009        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2010        moved outside it again, because FALSE and TRUE are already defined in AIX,
2011        but BOOL is not.
2012    
2013    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2014        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2015    
2016    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2017        hyphens) following file names and line numbers when outputting matching
2018        lines. This is not true; no spaces are inserted. I have also clarified the
2019        wording for the --colour (or --color) option.
2020    
2021    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2022        was not coloured; this is different to GNU grep, so I have changed it to be
2023        the same.
2024    
2025    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2026        each matching line was coloured. Now it goes on to look for further matches
2027        of any of the test patterns, which is the same behaviour as GNU grep.
2028    
2029    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2030        doesn't make sense to accept an empty string match in pcregrep, so I have
2031        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2032        seems to be how GNU grep behaves. [But see later change 40 for release
2033        8.33.]
2034    
2035    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2036        start or after a newline", because the conditional assertion was not being
2037        correctly handled. The rule now is that both the assertion and what follows
2038        in the first alternative must satisfy the test.
2039    
2040    9.  If auto-callout was enabled in a pattern with a conditional group whose
2041        condition was an assertion, PCRE could crash during matching, both with
2042        pcre_exec() and pcre_dfa_exec().
2043    
2044    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2045        used for matching.
2046    
2047    11. Unicode property support in character classes was not working for
2048        characters (bytes) greater than 127 when not in UTF-8 mode.
2049    
2050    12. Added the -M command line option to pcretest.
2051    
2052    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2053    
2054    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2055    
2056    16. Added comments and documentation about mis-use of no_arg in the C++
2057        wrapper.
2058    
2059    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2060        from Martin Jerabek that uses macro names for all relevant character and
2061        string constants.
2062    
2063    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2064        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2065        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2066        these, but not everybody uses configure.
2067    
2068    19. A conditional group that had only one branch was not being correctly
2069        recognized as an item that could match an empty string. This meant that an
2070        enclosing group might also not be so recognized, causing infinite looping
2071        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2072        with the subject "ab", where knowledge that the repeated group can match
2073        nothing is needed in order to break the loop.
2074    
2075    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2076        exec(), but without supplying a callout function, matching went wrong.
2077    
2078    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2079        leak if the size of the offset vector was greater than 30. When the vector
2080        is smaller, the saved offsets during recursion go onto a local stack
2081        vector, but for larger vectors malloc() is used. It was failing to free
2082        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2083        error, in fact).
2084    
2085    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2086        heapframe that is used only when UTF-8 support is enabled. This caused no
2087        problem, but was untidy.
2088    
2089    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2090        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2091        included within another project.
2092    
2093    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2094        slightly modified by me:
2095    
2096          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2097              not building pcregrep.
2098    
2099          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2100              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2101    
2102    25. Forward references, both numeric and by name, in patterns that made use of
2103        duplicate group numbers, could behave incorrectly or give incorrect errors,
2104        because when scanning forward to find the reference group, PCRE was not
2105        taking into account the duplicate group numbers. A pattern such as
2106        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2107    
2108    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2109        the feature of a custom pointer more persuasive (as requested by a user).
2110    
2111    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2112        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2113        pre-defined.
2114    
2115    28. Added support for (*UTF8) at the start of a pattern.
2116    
2117    29. Arrange for flags added by the "release type" setting in CMake to be shown
2118        in the configuration summary.
2119    
2120    
2121    Version 7.8 05-Sep-08
2122    ---------------------
2123    
2124    1.  Replaced UCP searching code with optimized version as implemented for Ad
2125        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2126        stage table and inline lookup instead of a function, giving speed ups of 2
2127        to 5 times on some simple patterns that I tested. Permission was given to
2128        distribute the MultiStage2.py script that generates the tables (it's not in
2129        the tarball, but is in the Subversion repository).
2130    
2131    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2132        scripts.
2133    
2134    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2135        a group with a zero qualifier. The result of the study could be incorrect,
2136        or the function might crash, depending on the pattern.
2137    
2138    4.  Caseless matching was not working for non-ASCII characters in back
2139        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2140        It now works when Unicode Property Support is available.
2141    
2142    5.  In pcretest, an escape such as \x{de} in the data was always generating
2143        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2144        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2145        truncation.
2146    
2147    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2148    
2149    7.  Added two (int) casts to pcregrep when printing the difference of two
2150        pointers, in case they are 64-bit values.
2151    
2152    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2153        test 2 if it fails.
2154    
2155    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2156        and a #define of that name to empty if it is not externally set. This is to
2157        allow users of MSVC to set it if necessary.
2158    
2159    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2160        the convenience functions in the pcre_get.c source file.
2161    
2162    11. An option change at the start of a pattern that had top-level alternatives
2163        could cause overwriting and/or a crash. This command provoked a crash in
2164        some environments:
2165    
2166          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2167    
2168        This potential security problem was recorded as CVE-2008-2371.
2169    
2170    12. For a pattern where the match had to start at the beginning or immediately
2171        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2172        pcre_dfa_exec() could read past the end of the passed subject if there was
2173        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2174        pcretest so that it places the subject at the end of its malloc-ed buffer.
2175    
2176    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2177        exec() might read past the end of the data buffer in UTF-8 mode.
2178    
2179    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2180        the data contained the byte 0x85 as part of a UTF-8 character within its
2181        first line. This applied both to normal and DFA matching.
2182    
2183    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2184        /^[^d]*?$/8 failed to match "abc".
2185    
2186    16. Added a missing copyright notice to pcrecpp_internal.h.
2187    
2188    17. Make it more clear in the documentation that values returned from
2189        pcre_exec() in ovector are byte offsets, not character counts.
2190    
2191    18. Tidied a few places to stop certain compilers from issuing warnings.
2192    
2193    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2194        supplied by Stefan Weber. I made a further small update for 7.8 because
2195        there is a change of source arrangements: the pcre_searchfuncs.c module is
2196        replaced by pcre_ucd.c.
2197    
2198    
2199    Version 7.7 07-May-08
2200    ---------------------
2201    
2202    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2203        a string to a long long, pretend we don't even have a long long." This is
2204        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2205    
2206    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2207        pre-7.6 versions, which defined a global no_arg variable instead of putting
2208        it in the RE class. (See also #8 below.)
2209    
2210    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2211        Lopes.
2212    
2213    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2214    
2215        (1) The include/exclude patterns were being applied to the whole pathnames
2216            of files, instead of just to the final components.
2217    
2218        (2) If there was more than one level of directory, the subdirectories were
2219            skipped unless they satisfied the include/exclude conditions. This is
2220            inconsistent with GNU grep (and could even be seen as contrary to the
2221            pcregrep specification - which I improved to make it absolutely clear).
2222            The action now is always to scan all levels of directory, and just
2223            apply the include/exclude patterns to regular files.
2224    
2225    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2226        --exclude_dir in the tests to avoid scanning .svn directories.
2227    
2228    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2229        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2230        doesn't support NULs in patterns.
2231    
2232    7.  Added some missing "const"s to declarations of static tables in
2233        pcre_compile.c and pcre_dfa_exec.c.
2234    
2235    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2236        caused by fix #2  above. (Subsequently also a second patch to fix the
2237        first patch. And a third patch - this was a messy problem.)
2238    
2239    9.  Applied Craig's patch to remove the use of push_back().
2240    
2241    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2242        matching function regexec().
2243    
2244    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2245        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2246        references. PCRE supports relative numbers with this syntax (I don't think
2247        Oniguruma does).
2248    
2249    12. Previously, a group with a zero repeat such as (...){0} was completely
2250        omitted from the compiled regex. However, this means that if the group
2251        was called as a subroutine from elsewhere in the pattern, things went wrong
2252        (an internal error was given). Such groups are now left in the compiled
2253        pattern, with a new opcode that causes them to be skipped at execution
2254        time.
2255    
2256    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2257        to the way PCRE behaves:
2258    
2259        (a) A lone ] character is dis-allowed (Perl treats it as data).
2260    
2261        (b) A back reference to an unmatched subpattern matches an empty string
2262            (Perl fails the current match path).
2263    
2264        (c) A data ] in a character class must be notated as \] because if the
2265            first data character in a class is ], it defines an empty class. (In
2266            Perl it is not possible to have an empty class.) The empty class []
2267            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2268            The negative empty class [^] matches any one character, independently
2269            of the DOTALL setting.
2270    
2271    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2272        non-existent subpattern following a character class starting with ']' and
2273        containing () gave an internal compiling error instead of "reference to
2274        non-existent subpattern". Fortunately, when the pattern did exist, the
2275        compiled code was correct. (When scanning forwards to check for the
2276        existence of the subpattern, it was treating the data ']' as terminating
2277        the class, so got the count wrong. When actually compiling, the reference
2278        was subsequently set up correctly.)
2279    
2280    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2281        it was being rejected as not supported by pcre_dfa_exec(), even though
2282        other assertions are supported. I have made pcre_dfa_exec() support
2283        (*FAIL).
2284    
2285    16. The implementation of 13c above involved the invention of a new opcode,
2286        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2287        cannot be changed at match time, I realized I could make a small
2288        improvement to matching performance by compiling OP_ALLANY instead of
2289        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2290        on the OP_ANY path.
2291    
2292    17. Compiling pcretest on Windows with readline support failed without the
2293        following two fixes: (1) Make the unistd.h include conditional on
2294        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2295    
2296    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2297        ncurses library to be included for pcretest when ReadLine support is
2298        requested, but also to allow for it to be overridden. This patch came from
2299        Daniel Bergström.
2300    
2301    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2302        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2303        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2304        spotting this.
2305    
2306    
2307    Version 7.6 28-Jan-08
2308    ---------------------
2309    
2310    1.  A character class containing a very large number of characters with
2311        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2312        overflow.
2313    
2314    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2315        HAVE_LONG_LONG is not defined.
2316    
2317    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2318        bring them up to date and include new features. This patch includes:
2319    
2320        - Fixed PH's badly added libz and libbz2 support.
2321        - Fixed a problem with static linking.
2322        - Added pcredemo. [But later removed - see 7 below.]
2323        - Fixed dftables problem and added an option.
2324        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2325            HAVE_LONG_LONG.
2326        - Added readline support for pcretest.
2327        - Added an listing of the option settings after cmake has run.
2328    
2329    4.  A user submitted a patch to Makefile that makes it easy to create
2330        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2331        Makefile.am that cause it to include this special target, without
2332        affecting anything else. Note that the same mingw target plus all
2333        the other distribution libraries and programs are now supported
2334        when configuring with CMake (see 6 below) instead of with
2335        Configure/Make.
2336    
2337    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2338        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2339        exported in the Windows port". It has not yet been confirmed that the patch
2340        solves the problem, but it does no harm.
2341    
2342    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2343        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2344        with CMake, and also correct the comment about stack recursion.
2345    
2346    7.  Remove the automatic building of pcredemo from the ./configure system and
2347        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2348        of a program that users should build themselves after PCRE is installed, so
2349        building it automatically is not really right. What is more, it gave
2350        trouble in some build environments.
2351    
2352    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2353    
2354    
2355    Version 7.5 10-Jan-08
2356    ---------------------
2357    
2358    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2359        values in parens when parsing an RE using the C++ wrapper."
2360    
2361    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2362        Characters greater than 255 were excluded from the class instead of being
2363        included.
2364    
2365    3.  The same bug as (2) above applied to negated POSIX classes such as
2366        [:^space:].
2367    
2368    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2369        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2370        I have changed it.
2371    
2372    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2373        first named subpattern) and a construct such as (?&a) would reference the
2374        first named subpattern whose name started with "a" (in other words, the
2375        length check was missing). Both these problems are fixed. "Subpattern name
2376        expected" is now given for (?&) (a zero-length name), and this patch also
2377        makes it give the same error for \k'' (previously it complained that that
2378        was a reference to a non-existent subpattern).
2379    
2380    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2381        this is right because (?- can be followed by option settings as well as by
2382        digits. I have, however, made the messages clearer.
2383    
2384    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2385        than the number used in the conditional) now cause a compile-time error.
2386        This is actually not compatible with Perl, which accepts such patterns, but
2387        treats the conditional as always being FALSE (as PCRE used to), but it
2388        seems to me that giving a diagnostic is better.
2389    
2390    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2391        and messages.
2392    
2393    9.  Fix two occurrences of "backslash" in comments that should have been
2394        "backspace".
2395    
2396    10. Remove two redundant lines of code that can never be obeyed (their function
2397        was moved elsewhere).
2398    
2399    11. The program that makes PCRE's Unicode character property table had a bug
2400        which caused it to generate incorrect table entries for sequences of
2401        characters that have the same character type, but are in different scripts.
2402        It amalgamated them into a single range, with the script of the first of
2403        them. In other words, some characters were in the wrong script. There were
2404        thirteen such cases, affecting characters in the following ranges:
2405    
2406          U+002b0 - U+002c1
2407          U+0060c - U+0060d
2408          U+0061e - U+00612
2409          U+0064b - U+0065e
2410          U+0074d - U+0076d
2411          U+01800 - U+01805
2412          U+01d00 - U+01d77
2413          U+01d9b - U+01dbf
2414          U+0200b - U+0200f
2415          U+030fc - U+030fe
2416          U+03260 - U+0327f
2417          U+0fb46 - U+0fbb1
2418          U+10450 - U+1049d
2419    
2420    12. The -o option (show only the matching part of a line) for pcregrep was not
2421        compatible with GNU grep in that, if there was more than one match in a
2422        line, it showed only the first of them. It now behaves in the same way as
2423        GNU grep.
2424    
2425    13. If the -o and -v options were combined for pcregrep, it printed a blank
2426        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2427        does the same. The return code can be used to tell if there were any
2428        non-matching lines.
2429    
2430    14. Added --file-offsets and --line-offsets to pcregrep.
2431    
2432    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2433        infinitely looping recursion. The bug was that positive lookaheads were not
2434        being skipped when checking for a possible empty match (negative lookaheads
2435        and both kinds of lookbehind were skipped).
2436    
2437    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2438        inclusion of <windows.h> to before rather than after the definition of
2439        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2440    
2441    17. Specifying a possessive quantifier with a specific limit for a Unicode
2442        character property caused pcre_compile() to compile bad code, which led at
2443        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2444        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2445        caused the error; without that there was no problem.
2446    
2447    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2448    
2449    19. Added --enable-pcretest-libreadline.
2450    
2451    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2452        RE::GlobalReplace(). As a result, the number of replacements returned was
2453        double what it should be. I removed one of the increments, but Craig sent a
2454        later patch that removed the other one (the right fix) and added unit tests
2455        that check the return values (which was not done before).
2456    
2457    21. Several CMake things:
2458    
2459        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2460            the names libpcre and libpcreposix, not just pcre and pcreposix.
2461    
2462        (2) The above change means that pcretest and pcregrep are now correctly
2463            linked with the newly-built libraries, not previously installed ones.
2464    
2465        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2466    
2467    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2468        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2469        UTF-8 newline character). The key issue is that the pattern starts .*;
2470        this means that the match must be either at the beginning, or after a
2471        newline. The bug was in the code for advancing after a failed match and
2472        checking that the new position followed a newline. It was not taking
2473        account of UTF-8 characters correctly.
2474    
2475    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2476        character classes. PCRE was not treating the sequence [:...:] as a
2477        character class unless the ... were all letters. Perl, however, seems to
2478        allow any characters between [: and :], though of course it rejects as
2479        unknown any "names" that contain non-letters, because all the known class
2480        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2481        for example, whereas PCRE did not - it did not recognize a POSIX character
2482        class. This seemed a bit dangerous, so the code has been changed to be
2483        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2484        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2485        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2486        Perl does, and where it didn't before.
2487    
2488    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2489        Windows environments %n is disabled by default.
2490    
2491    
2492    Version 7.4 21-Sep-07
2493    ---------------------
2494    
2495    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2496        means that a class such as [\s] counted as "explicit reference to CR or
2497        LF". That isn't really right - the whole point of the change was to try to
2498        help when there was an actual mention of one of the two characters. So now
2499        the change happens only if \r or \n (or a literal CR or LF) character is
2500        encountered.
2501    
2502    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2503        of both had grown to the point where there were only 3 bits left.
2504        Fortunately, there was spare space in the data structure, and so I have
2505        moved the internal flags into a new 16-bit field to free up more option
2506        bits.
2507    
2508    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2509        but did not set the internal JCHANGED flag - either of these is enough to
2510        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2511        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2512        start sets both bits.
2513    
2514    4.  Added options (at build time, compile time, exec time) to change \R from
2515        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2516    
2517    5.  doc/pcresyntax.html was missing from the distribution.
2518    
2519    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2520        compatibility, even though it is no longer used.
2521    
2522    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2523        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2524        windows.h file is present (where different names are used). [This was
2525        reversed later after testing - see 16 below.]
2526    
2527    8.  Changed all #include <config.h> to #include "config.h". There were also
2528        some further <pcre.h> cases that I changed to "pcre.h".
2529    
2530    9.  When pcregrep was used with the --colour option, it missed the line ending
2531        sequence off the lines that it output.
2532    
2533    10. It was pointed out to me that arrays of string pointers cause lots of
2534        relocations when a shared library is dynamically loaded. A technique of
2535        using a single long string with a table of offsets can drastically reduce
2536        these. I have refactored PCRE in four places to do this. The result is
2537        dramatic:
2538    
2539          Originally:                          290
2540          After changing UCP table:            187
2541          After changing error message table:   43
2542          After changing table of "verbs"       36
2543          After changing table of Posix names   22
2544    
2545        Thanks to the folks working on Gregex for glib for this insight.
2546    
2547    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2548        unicode-properties was also set.
2549    
2550    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2551    
2552    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2553        checked only for CRLF.
2554    
2555    14. Added casts to pcretest.c to avoid compiler warnings.
2556    
2557    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2558    
2559    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2560        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2561        entirely. This removes changes made in 7 above.
2562    
2563    17. The CMake files have been updated, and there is now more information about
2564        building with CMake in the NON-UNIX-USE document.
2565    
2566    
2567    Version 7.3 28-Aug-07
2568    ---------------------
2569    
2570     1. In the rejigging of the build system that eventually resulted in 7.1, the
2571        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2572        brackets there is not right, since it causes compilers to look for an
2573        installed pcre.h, not the version that is in the source that is being
2574        compiled (which of course may be different). I have changed it back to:
2575    
2576          #include "pcre.h"
2577    
2578        I have a vague recollection that the change was concerned with compiling in
2579        different directories, but in the new build system, that is taken care of
2580        by the VPATH setting the Makefile.
2581    
2582     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2583        when the subject happened to end in the byte 0x85 (e.g. if the last
2584        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2585        characters but of course it shouldn't be taken as a newline when it is part
2586        of another character. The bug was that, for an unlimited repeat of . in
2587        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2588        characters when looking for a newline.
2589    
2590     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2591    
2592     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2593        in debug output.
2594    
2595     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2596        long printing in the pcrecpp unittest when running under MinGW.
2597    
2598     6. ESC_K was left out of the EBCDIC table.
2599    
2600     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2601        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2602        limit also applies to "virtual nesting" when a pattern is recursive, and in
2603        this case 1000 isn't so big. I have been able to remove this limit at the
2604        expense of backing off one optimization in certain circumstances. Normally,
2605        when pcre_exec() would call its internal match() function recursively and
2606        immediately return the result unconditionally, it uses a "tail recursion"
2607        feature to save stack. However, when a subpattern that can match an empty
2608        string has an unlimited repetition quantifier, it no longer makes this
2609        optimization. That gives it a stack frame in which to save the data for
2610        checking that an empty string has been matched. Previously this was taken
2611        from the 1000-entry workspace that had been reserved. So now there is no
2612        explicit limit, but more stack is used.
2613    
2614     8. Applied Daniel's patches to solve problems with the import/export magic
2615        syntax that is required for Windows, and which was going wrong for the
2616        pcreposix and pcrecpp parts of the library. These were overlooked when this
2617        problem was solved for the main library.
2618    
2619     9. There were some crude static tests to avoid integer overflow when computing
2620        the size of patterns that contain repeated groups with explicit upper
2621        limits. As the maximum quantifier is 65535, the maximum group length was
2622        set at 30,000 so that the product of these two numbers did not overflow a
2623        32-bit integer. However, it turns out that people want to use groups that
2624        are longer than 30,000 bytes (though not repeat them that many times).
2625        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2626        made it possible to implement the integer overflow checks in a much more
2627        dynamic way, which I have now done. The artificial limitation on group
2628        length has been removed - we now have only the limit on the total length of
2629        the compiled pattern, which depends on the LINK_SIZE setting.
2630    
2631    10. Fixed a bug in the documentation for get/copy named substring when
2632        duplicate names are permitted. If none of the named substrings are set, the
2633        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2634        empty string.
2635    
2636    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2637        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2638        because the ] is interpreted as the first data character and the
2639        terminating ] is not found. PCRE has been made compatible with Perl in this
2640        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2641        cause memory overwriting.
2642    
2643    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2644        string has been matched (to stop an infinite loop). It was not recognizing
2645        a conditional subpattern that could match an empty string if that
2646        subpattern was within another subpattern. For example, it looped when
2647        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2648        condition was not nested. This bug has been fixed.
2649    
2650    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2651        past the start of the subject in the presence of bytes with the top bit
2652        set, for example "\x8aBCD".
2653    
2654    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2655        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2656    
2657    14. Optimized (?!) to (*FAIL).
2658    
2659    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2660        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2661        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2662        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2663        does: it's just the validity check that is more restrictive.
2664    
2665    16. Inserted checks for integer overflows during escape sequence (backslash)
2666        processing, and also fixed erroneous offset values for syntax errors during
2667        backslash processing.
2668    
2669    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2670        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2671    
2672    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2673        caused an overrun.
2674    
2675    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2676        something other than just ASCII characters) inside a group that had an
2677        unlimited repeat caused a loop at compile time (while checking to see
2678        whether the group could match an empty string).
2679    
2680    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2681        [\P{Any}] did so. (Error in the code for printing property names.)
2682    
2683    21. An orphan \E inside a character class could cause a crash.
2684    
2685    22. A repeated capturing bracket such as (A)? could cause a wild memory
2686        reference during compilation.
2687    
2688    23. There are several functions in pcre_compile() that scan along a compiled
2689        expression for various reasons (e.g. to see if it's fixed length for look
2690        behind). There were bugs in these functions when a repeated \p or \P was
2691        present in the pattern. These operators have additional parameters compared
2692        with \d, etc, and these were not being taken into account when moving along
2693        the compiled data. Specifically:
2694    
2695        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2696            length.
2697    
2698        (b) An item such as \pL+ within a repeated group could cause crashes or
2699            loops.
2700    
2701        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2702            "reference to non-existent subpattern" error.
2703    
2704        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2705    
2706    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2707        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2708    
2709    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2710    
2711    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2712        character were causing crashes (broken optimization).
2713    
2714    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2715        \p or \P) caused a compile-time loop.
2716    
2717    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2718        break. For example, the unstudied pattern [\r\n]A does not match the string
2719        "\r\nA" because change 7.0/46 below moves the current point on by two
2720        characters after failing to match at the start. However, the pattern \nA
2721        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2722        the same is true. There doesn't seem any very clean way out of this, but
2723        what I have chosen to do makes the common cases work: PCRE now takes note
2724        of whether there can be an explicit match for \r or \n anywhere in the
2725        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2726        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2727        pattern has explicit CR or LF references.
2728    
2729    29. Added (*CR) etc for changing newline setting at start of pattern.
2730    
2731    
2732    Version 7.2 19-Jun-07
2733    ---------------------
2734    
2735     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2736        which is apparently normally available under Windows.
2737    
2738     2. Re-jig the pcregrep tests with different newline settings in an attempt
2739        to make them independent of the local environment's newline setting.
2740    
2741     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2742    
2743     4. Some of the "internals" tests were previously cut out when the link size
2744        was not 2, because the output contained actual offsets. The recent new
2745        "Z" feature of pcretest means that these can be cut out, making the tests
2746        usable with all link sizes.
2747    
2748     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2749        stack recursion. This gives a massive performance boost under BSD, but just
2750        a small improvement under Linux. However, it saves one field in the frame
2751        in all cases.
2752    
2753     6. Added more features from the forthcoming Perl 5.10:
2754    
2755        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2756            recursion call. It refers to the nth most recently opened parentheses.
2757    
2758        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2759            to be opened parentheses.
2760    
2761        (c) Conditions that refer to capturing parentheses can be specified
2762            relatively, for example, (?(-2)... or (?(+3)...
2763    
2764        (d) \K resets the start of the current match so that everything before
2765            is not part of it.
2766    
2767        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2768    
2769        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2770            reference syntax.
2771    
2772        (g) (?| introduces a group in which the numbering of parentheses in each
2773            alternative starts with the same number.
2774    
2775        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2776    
2777     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2778        PCRE_INFO_JCHANGED.
2779    
2780     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2781        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2782        for detecting groups that can match an empty string.
2783    
2784     9. A pattern with a very large number of alternatives (more than several
2785        hundred) was running out of internal workspace during the pre-compile
2786        phase, where pcre_compile() figures out how much memory will be needed. A
2787        bit of new cunning has reduced the workspace needed for groups with
2788        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2789        workspace instead of running out of the 4096 that are available.
2790    
2791    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2792    
2793    11. Applied patch from Google to remove an optimization that didn't quite work.
2794        The report of the bug said:
2795    
2796          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2797          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2798          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2799    
2800    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2801        it matched the wrong number of bytes.
2802    
2803    
2804    Version 7.1 24-Apr-07
2805  ---------------------  ---------------------
2806    
2807   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
# Line 133  Version 7.1 12-Mar-07 Line 2933  Version 7.1 12-Mar-07
2933  19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string  19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2934      "abc\r\n\r\n", it found an unwanted second match after the second \r. This      "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2935      was because its rules for how to advance for /g after matching an empty      was because its rules for how to advance for /g after matching an empty
2936      string did not allow for this case. They now check for it specially.      string at the end of a line did not allow for this case. They now check for
2937        it specially.
2938    
2939    20. pcretest is supposed to handle patterns and data of any length, by
2940        extending its buffers when necessary. It was getting this wrong when the
2941        buffer for a data line had to be extended.
2942    
2943    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2944        CRLF as a newline sequence.
2945    
2946    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2947        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2948        I have nevertheless tidied it up.
2949    
2950    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2951    
2952    24. Added a man page for pcre-config.
2953    
2954    
2955  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06

Legend:
Removed from v.142  
changed lines
  Added in v.1477

  ViewVC Help
Powered by ViewVC 1.1.5