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

Diff of /code/trunk/ChangeLog

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

revision 131 by ph10, Mon Mar 26 15:10:12 2007 UTC revision 1488 by ph10, Wed Jun 18 18:38:00 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    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
30        an empty string, which sometimes were behaving incorrectly in the
31        interpreter (though correctly in the JIT matcher). This pcretest input is
32        an example:
33    
34          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
35          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
36    
37        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
38        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
39        for an empty string was breaking the inner loop and carrying on at a lower
40        level, when possessive repeated groups should always return to a higher
41        level as they have no backtrack points in them. The empty string test now
42        occurs at the outer level.
43    
44    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
45        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
46    
47    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
48        doesn't).
49    
50    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
51        to hold a pattern to be incorrectly computed (too small) when there were
52        named back references to duplicated names. This could cause "internal
53        error: code overflow" or "double free or corruption" or other memory
54        handling errors.
55    
56    11. When named subpatterns had the same prefixes, back references could be
57        confused. For example, in this pattern:
58    
59          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
60    
61        the reference to 'Name' was incorrectly treated as a reference to a
62        duplicate name.
63    
64    12. A pattern such as /^s?c/mi8 where the optional character has more than
65        one "other case" was incorrectly compiled such that it would only try to
66        match starting at "c".
67    
68    13. When a pattern starting with \s was studied, VT was not included in the
69        list of possible starting characters; this should have been part of the
70        8.34/18 patch.
71    
72    14. If a character class started [\Qx]... where x is any character, the class
73        was incorrectly terminated at the ].
74    
75    15. If a pattern that started with a caseless match for a character with more
76        than one "other case" was studied, PCRE did not set up the starting code
77        unit bit map for the list of possible characters. Now it does. This is an
78        optimization improvement, not a bug fix.
79    
80    
81    Version 8.35 04-April-2014
82    --------------------------
83    
84    1.  A new flag is set, when property checks are present in an XCLASS.
85        When this flag is not set, PCRE can perform certain optimizations
86        such as studying these XCLASS-es.
87    
88    2.  The auto-possessification of character sets were improved: a normal
89        and an extended character set can be compared now. Furthermore
90        the JIT compiler optimizes more character set checks.
91    
92    3.  Got rid of some compiler warnings for potentially uninitialized variables
93        that show up only when compiled with -O2.
94    
95    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
96        of a match later then the end of the match. The pcretest program was not
97        handling the case sensibly - it was outputting from the start to the next
98        binary zero. It now reports this situation in a message, and outputs the
99        text from the end to the start.
100    
101    5.  Fast forward search is improved in JIT. Instead of the first three
102        characters, any three characters with fixed position can be searched.
103        Search order: first, last, middle.
104    
105    6.  Improve character range checks in JIT. Characters are read by an inprecise
106        function now, which returns with an unknown value if the character code is
107        above a certain treshold (e.g: 256). The only limitation is that the value
108        must be bigger than the treshold as well. This function is useful, when
109        the characters above the treshold are handled in the same way.
110    
111    7.  The macros whose names start with RAWUCHAR are placeholders for a future
112        mode in which only the bottom 21 bits of 32-bit data items are used. To
113        make this more memorable for those maintaining the code, the names have
114        been changed to start with UCHAR21, and an extensive comment has been added
115        to their definition.
116    
117    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
118        to the export list in Makefile.am (they were accidentally omitted from the
119        8.34 tarball).
120    
121    9.  The informational output from pcretest used the phrase "starting byte set"
122        which is inappropriate for the 16-bit and 32-bit libraries. As the output
123        for "first char" and "need char" really means "non-UTF-char", I've changed
124        "byte" to "char", and slightly reworded the output. The documentation about
125        these values has also been (I hope) clarified.
126    
127    10. Another JIT related optimization: use table jumps for selecting the correct
128        backtracking path, when more than four alternatives are present inside a
129        bracket.
130    
131    11. Empty match is not possible, when the minimum length is greater than zero,
132        and there is no \K in the pattern. JIT should avoid empty match checks in
133        such cases.
134    
135    12. In a caseless character class with UCP support, when a character with more
136        than one alternative case was not the first character of a range, not all
137        the alternative cases were added to the class. For example, s and \x{17f}
138        are both alternative cases for S: the class [RST] was handled correctly,
139        but [R-T] was not.
140    
141    13. The configure.ac file always checked for pthread support when JIT was
142        enabled. This is not used in Windows, so I have put this test inside a
143        check for the presence of windows.h (which was already tested for).
144    
145    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
146        The algorithm provides a way to skip certain starting offsets, and usually
147        faster than linear prefix searches.
148    
149    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
150        as for 'fr_FR' and 'french'. For some reason, however, it then used the
151        Windows-specific input and output files, which have 'french' screwed in.
152        So this could never have worked. One of the problems with locales is that
153        they aren't always the same. I have now updated RunTest so that it checks
154        the output of the locale test (test 3) against three different output
155        files, and it allows the test to pass if any one of them matches. With luck
156        this should make the test pass on some versions of Solaris where it was
157        failing. Because of the uncertainty, the script did not used to stop if
158        test 3 failed; it now does. If further versions of a French locale ever
159        come to light, they can now easily be added.
160    
161    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
162        there was a message during ./configure, but it did not stop. This now
163        provokes an error. The invalid example in README has been corrected.
164        If a value less than the minimum is given, the minimum value has always
165        been used, but now a warning is given.
166    
167    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
168        was a bug in the test system, which is now fixed. Also, the list of various
169        configurations that are tested for each release did not have one with both
170        16/32 bits and --enable-bar-anycrlf. It now does.
171    
172    18. pcretest was missing "-C bsr" for displaying the \R default setting.
173    
174    19. Little endian PowerPC systems are supported now by the JIT compiler.
175    
176    20. The fast forward newline mechanism could enter to an infinite loop on
177        certain invalid UTF-8 input. Although we don't support these cases
178        this issue can be fixed by a performance optimization.
179    
180    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
181        not take account if existing stack usage. There is now a new global
182        variable called pcre_stack_guard that can be set to point to an external
183        function to check stack availability. It is called at the start of
184        processing every parenthesized group.
185    
186    22. A typo in the code meant that in ungreedy mode the max/min qualifier
187        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
188        not match "ab".
189    
190    23. When UTF was disabled, the JIT program reported some incorrect compile
191        errors. These messages are silenced now.
192    
193    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
194        compiler.
195    
196    25. Change all the temporary files used in RunGrepTest to be different to those
197        used by RunTest so that the tests can be run simultaneously, for example by
198        "make -j check".
199    
200    
201    Version 8.34 15-December-2013
202    -----------------------------
203    
204    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
205        executable memory. Patch inspired by Carsten Klein.
206    
207    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
208        this macro is never tested and has no effect, because the work to support
209        coverage involves only compiling and linking options and special targets in
210        the Makefile. The comment in config.h implied that defining the macro would
211        enable coverage support, which is totally false. There was also support for
212        setting this macro in the CMake files (my fault, I just copied it from
213        configure). SUPPORT_GCOV has now been removed.
214    
215    3.  Make a small performance improvement in strlen16() and strlen32() in
216        pcretest.
217    
218    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
219        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
220        these cases). There was also one in pcretest.c.
221    
222    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
223    
224    6.  In UTF mode, the code for checking whether a group could match an empty
225        string (which is used for indefinitely repeated groups to allow for
226        breaking an infinite loop) was broken when the group contained a repeated
227        negated single-character class with a character that occupied more than one
228        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
229        UTF-8 mode). The effect was undefined: the group might or might not be
230        deemed as matching an empty string, or the program might have crashed.
231    
232    7.  The code for checking whether a group could match an empty string was not
233        recognizing that \h, \H, \v, \V, and \R must match a character.
234    
235    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
236        an empty string. If it can, pcretest shows this in its information output.
237    
238    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
239        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
240        matched by pcre_exec() without using JIT:
241    
242        (a) If the rest of the pattern did not match after a maximal run of
243            grapheme clusters, the code for backing up to try with fewer of them
244            did not always back up over a full grapheme when characters that do not
245            have the modifier quality were involved, e.g. Hangul syllables.
246    
247        (b) If the match point in a subject started with modifier character, and
248            there was no match, the code could incorrectly back up beyond the match
249            point, and potentially beyond the first character in the subject,
250            leading to a segfault or an incorrect match result.
251    
252    10. A conditional group with an assertion condition could lead to PCRE
253        recording an incorrect first data item for a match if no other first data
254        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
255        first data item, and therefore matched "ca" after "c" instead of at the
256        start.
257    
258    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
259        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
260    
261    12. The source of pcregrep now includes z/OS-specific code so that it can be
262        compiled for z/OS as part of the special z/OS distribution.
263    
264    13. Added the -T and -TM options to pcretest.
265    
266    14. The code in pcre_compile.c for creating the table of named capturing groups
267        has been refactored. Instead of creating the table dynamically during the
268        actual compiling pass, the information is remembered during the pre-compile
269        pass (on the stack unless there are more than 20 named groups, in which
270        case malloc() is used) and the whole table is created before the actual
271        compile happens. This has simplified the code (it is now nearly 150 lines
272        shorter) and prepared the way for better handling of references to groups
273        with duplicate names.
274    
275    15. A back reference to a named subpattern when there is more than one of the
276        same name now checks them in the order in which they appear in the pattern.
277        The first one that is set is used for the reference. Previously only the
278        first one was inspected. This change makes PCRE more compatible with Perl.
279    
280    16. Unicode character properties were updated from Unicode 6.3.0.
281    
282    17. The compile-time code for auto-possessification has been refactored, based
283        on a patch by Zoltan Herczeg. It now happens after instead of during
284        compilation. The code is cleaner, and more cases are handled. The option
285        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
286        options in pcretest are provided to set it. It can also be set by
287        (*NO_AUTO_POSSESS) at the start of a pattern.
288    
289    18. The character VT has been added to the default ("C" locale) set of
290        characters that match \s and are generally treated as white space,
291        following this same change in Perl 5.18. There is now no difference between
292        "Perl space" and "POSIX space". Whether VT is treated as white space in
293        other locales depends on the locale.
294    
295    19. The code for checking named groups as conditions, either for being set or
296        for being recursed, has been refactored (this is related to 14 and 15
297        above). Processing unduplicated named groups should now be as fast at
298        numerical groups, and processing duplicated groups should be faster than
299        before.
300    
301    20. Two patches to the CMake build system, by Alexander Barkov:
302    
303          (1) Replace the "source" command by "." in CMakeLists.txt because
304              "source" is a bash-ism.
305    
306          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
307              without these the CMake build does not work on Solaris.
308    
309    21. Perl has changed its handling of \8 and \9. If there is no previously
310        encountered capturing group of those numbers, they are treated as the
311        literal characters 8 and 9 instead of a binary zero followed by the
312        literals. PCRE now does the same.
313    
314    22. Following Perl, added \o{} to specify codepoints in octal, making it
315        possible to specify values greater than 0777 and also making them
316        unambiguous.
317    
318    23. Perl now gives an error for missing closing braces after \x{... instead of
319        treating the string as literal. PCRE now does the same.
320    
321    24. RunTest used to grumble if an inappropriate test was selected explicitly,
322        but just skip it when running all tests. This make it awkward to run ranges
323        of tests when one of them was inappropriate. Now it just skips any
324        inappropriate tests, as it always did when running all tests.
325    
326    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
327        character types such as \d or \w, too many callouts were inserted, and the
328        data that they returned was rubbish.
329    
330    26. In UCP mode, \s was not matching two of the characters that Perl matches,
331        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
332        were matched by \h. The code has now been refactored so that the lists of
333        the horizontal and vertical whitespace characters used for \h and \v (which
334        are defined only in one place) are now also used for \s.
335    
336    27. Add JIT support for the 64 bit TileGX architecture.
337        Patch by Jiong Wang (Tilera Corporation).
338    
339    28. Possessive quantifiers for classes (both explicit and automatically
340        generated) now use special opcodes instead of wrapping in ONCE brackets.
341    
342    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
343        (because it's meaningless), this was not happening when PCRE_CASELESS was
344        set. Not wrong, but inefficient.
345    
346    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
347        properties for \w, \d, etc) is present in a test regex. Otherwise if the
348        test contains no characters greater than 255, Perl doesn't realise it
349        should be using Unicode semantics.
350    
351    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
352        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
353        does in Unicode mode.
354    
355    32. Added the "forbid" facility to pcretest so that putting tests into the
356        wrong test files can sometimes be quickly detected.
357    
358    33. There is now a limit (default 250) on the depth of nesting of parentheses.
359        This limit is imposed to control the amount of system stack used at compile
360        time. It can be changed at build time by --with-parens-nest-limit=xxx or
361        the equivalent in CMake.
362    
363    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
364        errors. Perl warns for these when in warning mode, but PCRE has no facility
365        for giving warnings.
366    
367    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
368        However, this was not working for (?!) because it is optimized to (*FAIL),
369        for which PCRE does not allow quantifiers. The optimization is now disabled
370        when a quantifier follows (?!). I can't see any use for this, but it makes
371        things uniform.
372    
373    36. Perl no longer allows group names to start with digits, so I have made this
374        change also in PCRE. It simplifies the code a bit.
375    
376    37. In extended mode, Perl ignores spaces before a + that indicates a
377        possessive quantifier. PCRE allowed a space before the quantifier, but not
378        before the possessive +. It now does.
379    
380    38. The use of \K (reset reported match start) within a repeated possessive
381        group such as (a\Kb)*+ was not working.
382    
383    40. Document that the same character tables must be used at compile time and
384        run time, and that the facility to pass tables to pcre_exec() and
385        pcre_dfa_exec() is for use only with saved/restored patterns.
386    
387    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
388        features for Builds with MSVC:
389    
390        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
391           builds). The .rc files can be used to set FileDescription and many other
392           attributes.
393    
394        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
395           This allows higher-level build scripts which want .pdb files to avoid
396           hard-coding the exact files needed."
397    
398    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
399        mean "start of word" and "end of word", respectively, as a transition aid.
400    
401    43. A minimizing repeat of a class containing codepoints greater than 255 in
402        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
403        compiled to use the heap for recursion.
404    
405    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
406        is configured.
407    
408    
409    Version 8.33 28-May-2013
410    ------------------------
411    
412    1.  Added 'U' to some constants that are compared to unsigned integers, to
413        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
414        variables that are added to signed variables, to ensure the result is
415        signed and can be negated.
416    
417    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
418        CMake config files.
419    
420    3.  Revise the creation of config.h.generic so that all boolean macros are
421        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
422        overriding via -D on the command line possible.
423    
424    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
425        to unsigned int is reported to make a quite noticeable speed difference in
426        a specific Windows environment. Testing on Linux did also appear to show
427        some benefit (and it is clearly not harmful). Also fixed the definition of
428        Xop which should be unsigned.
429    
430    5.  Related to (4), changing the definition of the intermediate variable cc
431        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
432        improvements.
433    
434    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
435        unnecessary spaces.
436    
437    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
438        and later.
439    
440    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
441    
442    9.  Optimizing fast_forward_start_bits in JIT.
443    
444    10. Adding support for callouts in JIT, and fixing some issues revealed
445        during this work. Namely:
446    
447        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
448    
449        (b) Minimum length was not checked before the matching is started.
450    
451    11. The value of capture_last that is passed to callouts was incorrect in some
452        cases when there was a capture on one path that was subsequently abandoned
453        after a backtrack. Also, the capture_last value is now reset after a
454        recursion, since all captures are also reset in this case.
455    
456    12. The interpreter no longer returns the "too many substrings" error in the
457        case when an overflowing capture is in a branch that is subsequently
458        abandoned after a backtrack.
459    
460    13. In the pathological case when an offset vector of size 2 is used, pcretest
461        now prints out the matched string after a yield of 0 or 1.
462    
463    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
464        Only supported by the JIT compiler at the moment.
465    
466    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
467    
468    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
469        offset of the starting point of the matching process, provided the offsets
470        vector is large enough.
471    
472    17. The \A escape now records a lookbehind value of 1, though its execution
473        does not actually inspect the previous character. This is to ensure that,
474        in partial multi-segment matching, at least one character from the old
475        segment is retained when a new segment is processed. Otherwise, if there
476        are no lookbehinds in the pattern, \A might match incorrectly at the start
477        of a new segment.
478    
479    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
480    
481    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
482        gives some modest performance improvement in 8-bit mode.
483    
484    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
485        be expressed in certain programming languages using Universal Character
486        Names.
487    
488    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
489        which points out that "non characters" are not "characters that may not
490        appear in Unicode strings" but rather "characters that are reserved for
491        internal use and have only local meaning".
492    
493    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
494        there was a conditional group that depended on an assertion, if the
495        assertion was false, the callout that immediately followed the alternation
496        in the condition was skipped when pcre_exec() was used for matching.
497    
498    23. Allow an explicit callout to be inserted before an assertion that is the
499        condition for a conditional group, for compatibility with automatic
500        callouts, which always insert a callout at this point.
501    
502    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
503        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
504    
505    25. (*PRUNE) is now supported by the JIT compiler.
506    
507    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
508    
509    27. Fix the case where there are two or more SKIPs with arguments that may be
510        ignored.
511    
512    28. (*SKIP) is now supported by the JIT compiler.
513    
514    29. (*THEN) is now supported by the JIT compiler.
515    
516    30. Update RunTest with additional test selector options.
517    
518    31. The way PCRE handles backtracking verbs has been changed in two ways.
519    
520        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
521        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
522        backtracking. In some cases this makes it more Perl-compatible, but Perl's
523        rather obscure rules do not always do the same thing.
524    
525        (2) Previously, backtracking verbs were confined within assertions. This is
526        no longer the case for positive assertions, except for (*ACCEPT). Again,
527        this sometimes improves Perl compatibility, and sometimes does not.
528    
529    32. A number of tests that were in test 2 because Perl did things differently
530        have been moved to test 1, because either Perl or PCRE has changed, and
531        these tests are now compatible.
532    
533    32. Backtracking control verbs are now handled in the same way in JIT and
534        interpreter.
535    
536    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
537        contained a forward subroutine reference caused a compile error.
538    
539    34. Auto-detect and optimize limited repetitions in JIT.
540    
541    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
542        blocking (*UTF) etc.
543    
544    36. In the interpreter, maximizing pattern repetitions for characters and
545        character types now use tail recursion, which reduces stack usage.
546    
547    37. The value of the max lookbehind was not correctly preserved if a compiled
548        and saved regex was reloaded on a host of different endianness.
549    
550    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
551        of the compiled pattern block, expand the flags field from 16 to 32 bits
552        because it was almost full.
553    
554    39. Try madvise first before posix_madvise.
555    
556    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
557        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
558        original change locked out all matches of empty strings. This has been
559        changed so that one match of an empty string per line is recognized.
560        Subsequent searches on the same line (for colouring or for --only-matching,
561        for example) do not recognize empty strings.
562    
563    41. Applied a user patch to fix a number of spelling mistakes in comments.
564    
565    42. Data lines longer than 65536 caused pcretest to crash.
566    
567    43. Clarified the data type for length and startoffset arguments for pcre_exec
568        and pcre_dfa_exec in the function-specific man pages, where they were
569        explicitly stated to be in bytes, never having been updated. I also added
570        some clarification to the pcreapi man page.
571    
572    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
573        a segmentation fault.
574    
575    
576    Version 8.32 30-November-2012
577    -----------------------------
578    
579    1.  Improved JIT compiler optimizations for first character search and single
580        character iterators.
581    
582    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
583        Patch by Daniel Richard G.
584    
585    3.  Single character iterator optimizations in the JIT compiler.
586    
587    4.  Improved JIT compiler optimizations for character ranges.
588    
589    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
590        Reported by Giuseppe D'Angelo.
591    
592    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
593        of a line, was being set incorrectly in cases where .* appeared inside
594        atomic brackets at the start of a pattern, or where there was a subsequent
595        *PRUNE or *SKIP.
596    
597    7.  Improved instruction cache flush for POWER/PowerPC.
598        Patch by Daniel Richard G.
599    
600    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
601        grep:
602    
603        (a) There is now no limit to the number of patterns to be matched.
604    
605        (b) An error is given if a pattern is too long.
606    
607        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
608            are now supported.
609    
610        (d) --exclude-from and --include-from (multiple use) have been added.
611    
612        (e) Exclusions and inclusions now apply to all files and directories, not
613            just to those obtained from scanning a directory recursively.
614    
615        (f) Multiple uses of -f and --file-list are now supported.
616    
617        (g) In a Windows environment, the default for -d has been changed from
618            "read" (the GNU grep default) to "skip", because otherwise the presence
619            of a directory in the file list provokes an error.
620    
621        (h) The documentation has been revised and clarified in places.
622    
623    9.  Improve the matching speed of capturing brackets.
624    
625    10. Changed the meaning of \X so that it now matches a Unicode extended
626        grapheme cluster.
627    
628    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
629        out POSIX threads when JIT support is configured.
630    
631    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
632    
633    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
634        structure could go wrong in environments where size_t is not the same size
635        as int.
636    
637    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
638        set.
639    
640    15. The EBCDIC support had decayed; later updates to the code had included
641        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
642        general tidy up of EBCDIC-related issues, and the documentation was also
643        not quite right. There is now a test that can be run on ASCII systems to
644        check some of the EBCDIC-related things (but is it not a full test).
645    
646    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
647        in a small tidy to the code.
648    
649    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
650    
651    18. If the --only-matching (-o) option in pcregrep is specified multiple
652        times, each one causes appropriate output. For example, -o1 -o2 outputs the
653        substrings matched by the 1st and 2nd capturing parentheses. A separating
654        string can be specified by --om-separator (default empty).
655    
656    19. Improving the first n character searches.
657    
658    20. Turn case lists for horizontal and vertical white space into macros so that
659        they are defined only once.
660    
661    21. This set of changes together give more compatible Unicode case-folding
662        behaviour for characters that have more than one other case when UCP
663        support is available.
664    
665        (a) The Unicode property table now has offsets into a new table of sets of
666            three or more characters that are case-equivalent. The MultiStage2.py
667            script that generates these tables (the pcre_ucd.c file) now scans
668            CaseFolding.txt instead of UnicodeData.txt for character case
669            information.
670    
671        (b) The code for adding characters or ranges of characters to a character
672            class has been abstracted into a generalized function that also handles
673            case-independence. In UTF-mode with UCP support, this uses the new data
674            to handle characters with more than one other case.
675    
676        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
677            whose other case is greater than 256 are now correctly matched
678            caselessly. Previously, the high codepoint matched the low one, but not
679            vice versa.
680    
681        (d) The processing of \h, \H, \v, and \ in character classes now makes use
682            of the new class addition function, using character lists defined as
683            macros alongside the case definitions of 20 above.
684    
685        (e) Caseless back references now work with characters that have more than
686            one other case.
687    
688        (f) General caseless matching of characters with more than one other case
689            is supported.
690    
691    22. Unicode character properties were updated from Unicode 6.2.0
692    
693    23. Improved CMake support under Windows. Patch by Daniel Richard G.
694    
695    24. Add support for 32-bit character strings, and UTF-32
696    
697    25. Major JIT compiler update (code refactoring and bugfixing).
698        Experimental Sparc 32 support is added.
699    
700    26. Applied a modified version of Daniel Richard G's patch to create
701        pcre.h.generic and config.h.generic by "make" instead of in the
702        PrepareRelease script.
703    
704    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
705        pcre_compile.c when checking for a zero character.
706    
707    28. Introducing a native interface for JIT. Through this interface, the compiled
708        machine code can be directly executed. The purpose of this interface is to
709        provide fast pattern matching, so several sanity checks are not performed.
710        However, feature tests are still performed. The new interface provides
711        1.4x speedup compared to the old one.
712    
713    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
714        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
715        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
716    
717    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
718        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
719        the "old" RFC 2279). Instead, it ended up passing a negative length to
720        pcre_exec().
721    
722    31. Add support for GCC's visibility feature to hide internal functions.
723    
724    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
725        "unknown -C option" after outputting 0 or 1.
726    
727    33. There is now support for generating a code coverage report for the test
728        suite in environments where gcc is the compiler and lcov is installed. This
729        is mainly for the benefit of the developers.
730    
731    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
732        unaddressable using valgrind annotations, allowing valgrind to detect
733        invalid memory accesses. This is mainly for the benefit of the developers.
734    
735    25. (*UTF) can now be used to start a pattern in any of the three libraries.
736    
737    26. Give configure error if --enable-cpp but no C++ compiler found.
738    
739    
740    Version 8.31 06-July-2012
741    -------------------------
742    
743    1.  Fixing a wrong JIT test case and some compiler warnings.
744    
745    2.  Removed a bashism from the RunTest script.
746    
747    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
748        to unsigned type, result still unsigned" that was given by an MS compiler
749        on encountering the code "-sizeof(xxx)".
750    
751    4.  Partial matching support is added to the JIT compiler.
752    
753    5.  Fixed several bugs concerned with partial matching of items that consist
754        of more than one character:
755    
756        (a) /^(..)\1/ did not partially match "aba" because checking references was
757            done on an "all or nothing" basis. This also applied to repeated
758            references.
759    
760        (b) \R did not give a hard partial match if \r was found at the end of the
761            subject.
762    
763        (c) \X did not give a hard partial match after matching one or more
764            characters at the end of the subject.
765    
766        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
767            a partial match for the string "\r".
768    
769        (e) When newline was set to CRLF, the metacharacter "." did not recognize
770            a partial match for a CR character at the end of the subject string.
771    
772    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
773        running pcretest, the text "(JIT)" added to the output whenever JIT is
774        actually used to run the match.
775    
776    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
777        or /S+[+] with a digit between 1 and 7.
778    
779    8.  OP_NOT now supports any UTF character not just single-byte ones.
780    
781    9.  (*MARK) control verb is now supported by the JIT compiler.
782    
783    10. The command "./RunTest list" lists the available tests without actually
784        running any of them. (Because I keep forgetting what they all are.)
785    
786    11. Add PCRE_INFO_MAXLOOKBEHIND.
787    
788    12. Applied a (slightly modified) user-supplied patch that improves performance
789        when the heap is used for recursion (compiled with --disable-stack-for-
790        recursion). Instead of malloc and free for each heap frame each time a
791        logical recursion happens, frames are retained on a chain and re-used where
792        possible. This sometimes gives as much as 30% improvement.
793    
794    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
795        call.
796    
797    14. As documented, (*COMMIT) is now confined to within a positive assertion.
798    
799    15. It is now possible to link pcretest with libedit as an alternative to
800        libreadline.
801    
802    16. (*COMMIT) control verb is now supported by the JIT compiler.
803    
804    17. The Unicode data tables have been updated to Unicode 6.1.0.
805    
806    18. Added --file-list option to pcregrep.
807    
808    19. Added binary file support to pcregrep, including the -a, --binary-files,
809        -I, and --text options.
810    
811    20. The madvise function is renamed for posix_madvise for QNX compatibility
812        reasons. Fixed by Giuseppe D'Angelo.
813    
814    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
815    
816    22. Changed the option for creating symbolic links for 16-bit man pages from
817        -s to -sf so that re-installing does not cause issues.
818    
819    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
820    
821    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
822        matches in certain environments (the workspace was not being correctly
823        retained). Also added to pcre_dfa_exec() a simple plausibility check on
824        some of the workspace data at the beginning of a restart.
825    
826    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
827        was not doing so when it should - probably a typo introduced by SVN 528
828        (change 8.10/14).
829    
830    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
831        \w+ when the character tables indicated that \x{c4} was a word character.
832        There were several related cases, all because the tests for doing a table
833        lookup were testing for characters less than 127 instead of 255.
834    
835    27. If a pattern contains capturing parentheses that are not used in a match,
836        their slots in the ovector are set to -1. For those that are higher than
837        any matched groups, this happens at the end of processing. In the case when
838        there were back references that the ovector was too small to contain
839        (causing temporary malloc'd memory to be used during matching), and the
840        highest capturing number was not used, memory off the end of the ovector
841        was incorrectly being set to -1. (It was using the size of the temporary
842        memory instead of the true size.)
843    
844    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
845        ovector size, it uses memory at the end of the block that it has got.
846    
847    29. Check for an overlong MARK name and give an error at compile time. The
848        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
849    
850    30. JIT compiler update.
851    
852    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
853        Rill for the patch.
854    
855    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
856    
857    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
858    
859    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
860        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
861        was enabled.
862    
863    35. Improve JIT code generation for greedy plus quantifier.
864    
865    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
866        1 to "aa" instead of to an empty string. The bug affected repeated groups
867        that could potentially match an empty string.
868    
869    37. Optimizing single character iterators in JIT.
870    
871    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
872        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
873        codepoints that are too big for the mode are faulted, and in a UTF mode,
874        disallowed codepoints are also faulted.
875    
876    39. If PCRE was compiled with UTF support, in three places in the DFA
877        matcher there was code that should only have been obeyed in UTF mode, but
878        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
879        processing when bytes with values greater than 127 were present. In 16-bit
880        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
881        both cases the values are those that cannot be the first data item in a UTF
882        character. The three items that might have provoked this were recursions,
883        possessively repeated groups, and atomic groups.
884    
885    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
886        and pcregrep, because some OS require shared objects to be explicitly
887        passed to ld, causing the link step to fail if they are not.
888    
889    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
890        mode, patterns that started with \h* or \R* might be incorrectly matched.
891    
892    
893    Version 8.30 04-February-2012
894    -----------------------------
895    
896    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
897        name is defined in ctype.h.
898    
899    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
900        only in quite long subpatterns.
901    
902    3.  Removed the function pcre_info(), which has been obsolete and deprecated
903        since it was replaced by pcre_fullinfo() in February 2000.
904    
905    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
906        match a (*MARK), and the match failed at the start of the subject, a
907        reference to memory before the start of the subject could occur. This bug
908        was introduced by fix 17 of release 8.21.
909    
910    5.  A reference to an unset group with zero minimum repetition was giving
911        totally wrong answers (in non-JavaScript-compatibility mode). For example,
912        /(another)?(\1?)test/ matched against "hello world test". This bug was
913        introduced in release 8.13.
914    
915    6.  Add support for 16-bit character strings (a large amount of work involving
916        many changes and refactorings).
917    
918    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
919        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
920        from a file.
921    
922    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
923        rounding is not applied in this particular case).
924    
925    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
926        if they appear, or are escaped, in patterns.
927    
928    10. Get rid of a number of -Wunused-but-set-variable warnings.
929    
930    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
931        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
932        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
933        also returns the mark "x". This bug applied to capturing parentheses,
934        non-capturing parentheses, and atomic parentheses. It also applied to some
935        assertions.
936    
937    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
938        information out of configure.ac instead of relying on pcre.h.generic, which
939        is not stored in the repository.
940    
941    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
942        -lreadline.
943    
944    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
945    
946    15. Applied Graycode's patch to put the top-level frame on the stack rather
947        than the heap when not using the stack for recursion. This gives a
948        performance improvement in many cases when recursion is not deep.
949    
950    16. Experimental code added to "pcretest -C" to output the stack frame size.
951    
952    
953    Version 8.21 12-Dec-2011
954    ------------------------
955    
956    1.  Updating the JIT compiler.
957    
958    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
959        are added as well.
960    
961    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
962        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
963        calling _pcre_jit_exec. Some extra comments are added.
964    
965    4.  (*MARK) settings inside atomic groups that do not contain any capturing
966        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
967        was introduced by change 18 for 8.20.
968    
969    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
970        ECMA-262 standard.
971    
972    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
973        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
974        This bug was probably introduced by change 9 of 8.13.
975    
976    7.  While fixing 6 above, I noticed that a number of other items were being
977        incorrectly rejected as "not fixed length". This arose partly because newer
978        opcodes had not been added to the fixed-length checking code. I have (a)
979        corrected the bug and added tests for these items, and (b) arranged for an
980        error to occur if an unknown opcode is encountered while checking for fixed
981        length instead of just assuming "not fixed length". The items that were
982        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
983        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
984        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
985    
986    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
987        being incorrectly compiled and would have given unpredicatble results.
988    
989    9.  A possessively repeated subpattern with minimum repeat count greater than
990        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
991        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
992        the first (A) could occur when it should not.
993    
994    10. Add a cast and remove a redundant test from the code.
995    
996    11. JIT should use pcre_malloc/pcre_free for allocation.
997    
998    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
999        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1000        is anything other than /usr, -L is still shown).
1001    
1002    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1003    
1004    14. Perl does not support \N without a following name in a [] class; PCRE now
1005        also gives an error.
1006    
1007    15. If a forward reference was repeated with an upper limit of around 2000,
1008        it caused the error "internal error: overran compiling workspace". The
1009        maximum number of forward references (including repeats) was limited by the
1010        internal workspace, and dependent on the LINK_SIZE. The code has been
1011        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1012        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1013        of around 200,000 forward references. While doing this, I also speeded up
1014        the filling in of repeated forward references.
1015    
1016    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1017        incorrectly expecting the subject to contain another "a" after the start.
1018    
1019    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1020        in the match, the SKIP should be ignored. This was not happening; instead
1021        the SKIP was being treated as NOMATCH. For patterns such as
1022        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1023        tested.
1024    
1025    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1026        now much more compatible with Perl, in particular in cases where the result
1027        is a non-match for a non-anchored pattern. For example, if
1028        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1029        "m", where previously it did not return a name. A side effect of this
1030        change is that for partial matches, the last encountered mark name is
1031        returned, as for non matches. A number of tests that were previously not
1032        Perl-compatible have been moved into the Perl-compatible test files. The
1033        refactoring has had the pleasing side effect of removing one argument from
1034        the match() function, thus reducing its stack requirements.
1035    
1036    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1037        subsequent uses of /S (without +) incorrectly behaved like /S+.
1038    
1039    21. Retrieve executable code size support for the JIT compiler and fixing
1040        some warnings.
1041    
1042    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1043        not work when the shorter character appeared right at the end of the
1044        subject string.
1045    
1046    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1047        systems.
1048    
1049    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1050        output it when the /M option is used in pcretest.
1051    
1052    25. The CheckMan script was not being included in the distribution. Also, added
1053        an explicit "perl" to run Perl scripts from the PrepareRelease script
1054        because this is reportedly needed in Windows.
1055    
1056    26. If study data was being save in a file and studying had not found a set of
1057        "starts with" bytes for the pattern, the data written to the file (though
1058        never used) was taken from uninitialized memory and so caused valgrind to
1059        complain.
1060    
1061    27. Updated RunTest.bat as provided by Sheri Pierce.
1062    
1063    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1064    
1065    29. Computation of memory usage for the table of capturing group names was
1066        giving an unnecessarily large value.
1067    
1068    
1069    Version 8.20 21-Oct-2011
1070    ------------------------
1071    
1072    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1073        a POSIX class. After further experiments with Perl, which convinced me that
1074        Perl has bugs and confusions, a closing square bracket is no longer allowed
1075        in a POSIX name. This bug also affected patterns with classes that started
1076        with full stops.
1077    
1078    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1079        captured substring, but while checking the failing first alternative,
1080        substring 1 is temporarily captured. If the output vector supplied to
1081        pcre_exec() was not big enough for this capture, the yield of the function
1082        was still zero ("insufficient space for captured substrings"). This cannot
1083        be totally fixed without adding another stack variable, which seems a lot
1084        of expense for a edge case. However, I have improved the situation in cases
1085        such as /(a)(b)x|abc/ matched against "abc", where the return code
1086        indicates that fewer than the maximum number of slots in the ovector have
1087        been set.
1088    
1089    3.  Related to (2) above: when there are more back references in a pattern than
1090        slots in the output vector, pcre_exec() uses temporary memory during
1091        matching, and copies in the captures as far as possible afterwards. It was
1092        using the entire output vector, but this conflicts with the specification
1093        that only 2/3 is used for passing back captured substrings. Now it uses
1094        only the first 2/3, for compatibility. This is, of course, another edge
1095        case.
1096    
1097    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1098        main code base, and can be used by building with --enable-jit. When this is
1099        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1100        runtime --no-jit option is given.
1101    
1102    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1103        ovector, the return from the function was zero, implying that there were
1104        other matches that did not fit. The correct "exactly full" value is now
1105        returned.
1106    
1107    6.  If a subpattern that was called recursively or as a subroutine contained
1108        (*PRUNE) or any other control that caused it to give a non-standard return,
1109        invalid errors such as "Error -26 (nested recursion at the same subject
1110        position)" or even infinite loops could occur.
1111    
1112    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1113        computing the minimum length on reaching *ACCEPT, and so ended up with the
1114        wrong value of 1 rather than 0. Further investigation indicates that
1115        computing a minimum subject length in the presence of *ACCEPT is difficult
1116        (think back references, subroutine calls), and so I have changed the code
1117        so that no minimum is registered for a pattern that contains *ACCEPT.
1118    
1119    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1120        it was not handled as intended. [But see 16 below.]
1121    
1122    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1123        Sheri Pierce.
1124    
1125    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1126        the first byte in a match must be "a".
1127    
1128    11. Change 17 for 8.13 increased the recursion depth for patterns like
1129        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1130        pattern contains any instances of (*THEN). If it does not, the old
1131        optimizations are restored. It would be nice to do this on a per-group
1132        basis, but at the moment that is not feasible.
1133    
1134    12. In some environments, the output of pcretest -C is CRLF terminated. This
1135        broke RunTest's code that checks for the link size. A single white space
1136        character after the value is now allowed for.
1137    
1138    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1139        For "fr", it uses the Windows-specific input and output files.
1140    
1141    14. If (*THEN) appeared in a group that was called recursively or as a
1142        subroutine, it did not work as intended. [But see next item.]
1143    
1144    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1145        pattern fragments (but not containing any | characters). If A and B are
1146        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1147        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1148        D. In other words, Perl considers parentheses that do not contain any |
1149        characters to be part of a surrounding alternative, whereas PCRE was
1150        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1151        differently. PCRE now behaves in the same way as Perl, except in the case
1152        of subroutine/recursion calls such as (?1) which have in any case always
1153        been different (but PCRE had them first :-).
1154    
1155    16. Related to 15 above: Perl does not treat the | in a conditional group as
1156        creating alternatives. Such a group is treated in the same way as an
1157        ordinary group without any | characters when processing (*THEN). PCRE has
1158        been changed to match Perl's behaviour.
1159    
1160    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1161        RunGrepTest script failed.
1162    
1163    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1164        inevitable for groups that contain captures, but it can lead to a lot of
1165        stack use in large patterns. The old behaviour has been restored for atomic
1166        groups that do not contain any capturing parentheses.
1167    
1168    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1169        suppress the check for a minimum subject length at run time. (If it was
1170        given to pcre_exec() or pcre_dfa_exec() it did work.)
1171    
1172    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1173        fail to work when decoding hex characters in data strings in EBCDIC
1174        environments.
1175    
1176    21. It appears that in at least one Mac OS environment, the isxdigit() function
1177        is implemented as a macro that evaluates to its argument more than once,
1178        contravening the C 90 Standard (I haven't checked a later standard). There
1179        was an instance in pcretest which caused it to go wrong when processing
1180        \x{...} escapes in subject strings. The has been rewritten to avoid using
1181        things like p++ in the argument of isxdigit().
1182    
1183    
1184    Version 8.13 16-Aug-2011
1185    ------------------------
1186    
1187    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1188    
1189    2.  Two minor typos in pcre_internal.h have been fixed.
1190    
1191    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1192        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1193        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1194    
1195    4.  There were a number of related bugs in the code for matching backrefences
1196        caselessly in UTF-8 mode when codes for the characters concerned were
1197        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1198        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1199        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1200        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1201        2-byte code at the end of the subject (it thought there wasn't enough data
1202        left).
1203    
1204    5.  Comprehensive information about what went wrong is now returned by
1205        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1206        as the output vector has at least 2 elements. The offset of the start of
1207        the failing character and a reason code are placed in the vector.
1208    
1209    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1210        now returned is for the first byte of the failing character, instead of the
1211        last byte inspected. This is an incompatible change, but I hope it is small
1212        enough not to be a problem. It makes the returned offset consistent with
1213        pcre_exec() and pcre_dfa_exec().
1214    
1215    7.  pcretest now gives a text phrase as well as the error number when
1216        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1217        failure, the offset and reason code are output.
1218    
1219    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1220        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1221        back over a single character (\n). This seems wrong (because it treated the
1222        two characters as a single entity when going forwards), conflicts with the
1223        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1224        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1225        has been changed.
1226    
1227    9.  Some internal refactoring has changed the processing so that the handling
1228        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1229        time (the PCRE_DOTALL option was changed this way some time ago: version
1230        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1231        which was always a bit of a fudge. It also means that there is one less
1232        argument for the match() function, which reduces its stack requirements
1233        slightly. This change also fixes an incompatibility with Perl: the pattern
1234        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1235    
1236    10. More internal refactoring has drastically reduced the number of recursive
1237        calls to match() for possessively repeated groups such as (abc)++ when
1238        using pcre_exec().
1239    
1240    11. While implementing 10, a number of bugs in the handling of groups were
1241        discovered and fixed:
1242    
1243        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1244        (a|)*(?1) gave a compile-time internal error.
1245        ((a|)+)+  did not notice that the outer group could match an empty string.
1246        (^a|^)+   was not marked as anchored.
1247        (.*a|.*)+ was not marked as matching at start or after a newline.
1248    
1249    12. Yet more internal refactoring has removed another argument from the match()
1250        function. Special calls to this function are now indicated by setting a
1251        value in a variable in the "match data" data block.
1252    
1253    13. Be more explicit in pcre_study() instead of relying on "default" for
1254        opcodes that mean there is no starting character; this means that when new
1255        ones are added and accidentally left out of pcre_study(), testing should
1256        pick them up.
1257    
1258    14. The -s option of pcretest has been documented for ages as being an old
1259        synonym of -m (show memory usage). I have changed it to mean "force study
1260        for every regex", that is, assume /S for every regex. This is similar to -i
1261        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1262        using it. It makes it easier to run collections of tests with and without
1263        study enabled, and thereby test pcre_study() more easily. All the standard
1264        tests are now run with and without -s (but some patterns can be marked as
1265        "never study" - see 20 below).
1266    
1267    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1268        restoration of the capturing data to the outer values was not happening
1269        correctly.
1270    
1271    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1272        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1273        pattern had matched an empty string, and so incorrectly returned a no
1274        match.
1275    
1276    17. There was optimizing code for the last branch of non-capturing parentheses,
1277        and also for the obeyed branch of a conditional subexpression, which used
1278        tail recursion to cut down on stack usage. Unfortunately, now that there is
1279        the possibility of (*THEN) occurring in these branches, tail recursion is
1280        no longer possible because the return has to be checked for (*THEN). These
1281        two optimizations have therefore been removed. [But see 8.20/11 above.]
1282    
1283    18. If a pattern containing \R was studied, it was assumed that \R always
1284        matched two bytes, thus causing the minimum subject length to be
1285        incorrectly computed because \R can also match just one byte.
1286    
1287    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1288        was incorrectly computed.
1289    
1290    20. If /S is present twice on a test pattern in pcretest input, it now
1291        *disables* studying, thereby overriding the use of -s on the command line
1292        (see 14 above). This is necessary for one or two tests to keep the output
1293        identical in both cases.
1294    
1295    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1296        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1297    
1298    22. When an atomic group that contained a capturing parenthesis was
1299        successfully matched, but the branch in which it appeared failed, the
1300        capturing was not being forgotten if a higher numbered group was later
1301        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1302        group 1 to "a", when in fact it should be unset. This applied to multi-
1303        branched capturing and non-capturing groups, repeated or not, and also to
1304        positive assertions (capturing in negative assertions does not happen
1305        in PCRE) and also to nested atomic groups.
1306    
1307    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1308        subject after a captured substring, to make it easier to tell which of a
1309        number of identical substrings has been captured.
1310    
1311    24. The way atomic groups are processed by pcre_exec() has been changed so that
1312        if they are repeated, backtracking one repetition now resets captured
1313        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1314        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1315        "aaa". Previously, it would have been "a". As part of this code
1316        refactoring, the way recursive calls are handled has also been changed.
1317    
1318    25. If an assertion condition captured any substrings, they were not passed
1319        back unless some other capturing happened later. For example, if
1320        (?(?=(a))a) was matched against "a", no capturing was returned.
1321    
1322    26. When studying a pattern that contained subroutine calls or assertions,
1323        the code for finding the minimum length of a possible match was handling
1324        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1325        group 1 called group 2 while simultaneously a separate group 2 called group
1326        1). A stack overflow occurred in this case. I have fixed this by limiting
1327        the recursion depth to 10.
1328    
1329    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1330        Fortmann. This supports explicit test numbers on the command line, and has
1331        argument validation and error reporting.
1332    
1333    28. An instance of \X with an unlimited repeat could fail if at any point the
1334        first character it looked at was a mark character.
1335    
1336    29. Some minor code refactoring concerning Unicode properties and scripts
1337        should reduce the stack requirement of match() slightly.
1338    
1339    30. Added the '=' option to pcretest to check the setting of unused capturing
1340        slots at the end of the pattern, which are documented as being -1, but are
1341        not included in the return count.
1342    
1343    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1344        compiled something random. Now it gives a compile-time error (as does
1345        Perl).
1346    
1347    32. A *MARK encountered during the processing of a positive assertion is now
1348        recorded and passed back (compatible with Perl).
1349    
1350    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1351        had alternative anchored branches, the search for a second match in a line
1352        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1353        matched the line "0102" twice. The same bug affected patterns that started
1354        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1355        twice.
1356    
1357    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1358        does, and because of capturing effects, quantifying parenthesized
1359        assertions may at times be useful. Quantifiers are now allowed for
1360        parenthesized assertions.
1361    
1362    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1363    
1364    36. \g was being checked for fancy things in a character class, when it should
1365        just be a literal "g".
1366    
1367    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1368        appearance of a nested POSIX class supersedes an apparent external class.
1369        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1370        unescaped square brackets may also appear as part of class names. For
1371        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1372        more like Perl. (But see 8.20/1 above.)
1373    
1374    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1375        was because it thought it was \N{name}, which is not supported).
1376    
1377    39. Add minix to OS list not supporting the -S option in pcretest.
1378    
1379    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1380        cannot analyze patterns in sufficient detail to catch mutual recursions
1381        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1382        subgroup is called recursively as a subpattern for a second time at the
1383        same position in the subject string. In previous releases this might have
1384        been caught by the recursion limit, or it might have run out of stack.
1385    
1386    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1387        happen only once. PCRE was, however incorrectly giving a compile time error
1388        "recursive call could loop indefinitely" because it cannot analyze the
1389        pattern in sufficient detail. The compile time test no longer happens when
1390        PCRE is compiling a conditional subpattern, but actual runaway loops are
1391        now caught at runtime (see 40 above).
1392    
1393    42. It seems that Perl allows any characters other than a closing parenthesis
1394        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1395        has been changed to be the same.
1396    
1397    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1398        as not to get warnings when autogen.sh is called. Also changed
1399        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1400    
1401    44. To help people who use pcregrep to scan files containing exceedingly long
1402        lines, the following changes have been made:
1403    
1404        (a) The default value of the buffer size parameter has been increased from
1405            8K to 20K. (The actual buffer used is three times this size.)
1406    
1407        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1408            PCRE is built.
1409    
1410        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1411            to be set at run time.
1412    
1413        (d) Numerical values in pcregrep options can be followed by K or M, for
1414            example --buffer-size=50K.
1415    
1416        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1417            given and the return code is set to 2.
1418    
1419    45. Add a pointer to the latest mark to the callout data block.
1420    
1421    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1422        partial match of an empty string instead of no match. This was specific to
1423        the use of ".".
1424    
1425    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1426        complete match instead of a partial match. This bug was dependent on both
1427        the PCRE_UTF8 and PCRE_DOTALL options being set.
1428    
1429    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1430        starting byte set, because \b was not being ignored.
1431    
1432    
1433    Version 8.12 15-Jan-2011
1434    ------------------------
1435    
1436    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1437        checks for such things as part of the documentation building process.
1438    
1439    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1440        --match-limit and --recursion-limit options (added for 8.11). In
1441        particular, this made one of the standard tests fail. (The integer value
1442        went into the wrong half of a long int.)
1443    
1444    3.  If the --colour option was given to pcregrep with -v (invert match), it
1445        did strange things, either producing crazy output, or crashing. It should,
1446        of course, ignore a request for colour when reporting lines that do not
1447        match.
1448    
1449    4.  Another pcregrep bug caused similar problems if --colour was specified with
1450        -M (multiline) and the pattern match finished with a line ending.
1451    
1452    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1453        matched in multiline mode, the following line was shown as part of the
1454        match. This seems wrong, so I have changed it.
1455    
1456    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1457        the check for further matches in the same line (so they could be coloured)
1458        to overrun the end of the current line. If another match was found, it was
1459        incorrectly shown (and then shown again when found in the next line).
1460    
1461    7.  If pcregrep was compiled under Windows, there was a reference to the
1462        function pcregrep_exit() before it was defined. I am assuming this was
1463        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1464        reported by a user. I've moved the definition above the reference.
1465    
1466    
1467    Version 8.11 10-Dec-2010
1468    ------------------------
1469    
1470    1.  (*THEN) was not working properly if there were untried alternatives prior
1471        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1472        backtracked to try for "b" instead of moving to the next alternative branch
1473        at the same level (in this case, to look for "c"). The Perl documentation
1474        is clear that when (*THEN) is backtracked onto, it goes to the "next
1475        alternative in the innermost enclosing group".
1476    
1477    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1478        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1479        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1480        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1481        (*THEN).
1482    
1483    3.  If \s appeared in a character class, it removed the VT character from
1484        the class, even if it had been included by some previous item, for example
1485        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1486        of \s, but is part of the POSIX "space" class.)
1487    
1488    4.  A partial match never returns an empty string (because you can always
1489        match an empty string at the end of the subject); however the checking for
1490        an empty string was starting at the "start of match" point. This has been
1491        changed to the "earliest inspected character" point, because the returned
1492        data for a partial match starts at this character. This means that, for
1493        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1494        (previously it gave "no match").
1495    
1496    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1497        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1498        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1499        has an implication that the given string is incomplete (because a partial
1500        match is preferred over a full match). For this reason, these items now
1501        give a partial match in this situation. [Aside: previously, the one case
1502        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1503        match rather than a full match, which was wrong by the old rules, but is
1504        now correct.]
1505    
1506    6.  There was a bug in the handling of #-introduced comments, recognized when
1507        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1508        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1509        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1510        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1511        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1512        places in pcre_compile().
1513    
1514    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1515        comments when looking ahead for named forward references to subpatterns,
1516        the only newline sequence it recognized was NL. It now handles newlines
1517        according to the set newline convention.
1518    
1519    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1520        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1521        cater for a lack of strerror(). These oversights have been fixed.
1522    
1523    9.  Added --match-limit and --recursion-limit to pcregrep.
1524    
1525    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1526    
1527    11. When the -o option was used, pcregrep was setting a return code of 1, even
1528        when matches were found, and --line-buffered was not being honoured.
1529    
1530    12. Added an optional parentheses number to the -o and --only-matching options
1531        of pcregrep.
1532    
1533    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1534        can match an empty string. The code to do it in pcretest and pcredemo
1535        needed fixing:
1536    
1537        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1538            only one byte after an empty string match just before CRLF (this case
1539            just got forgotten; "any" and "anycrlf" were OK).
1540    
1541        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1542            mode when an empty string match preceded an ASCII character followed by
1543            a non-ASCII character. (The code for advancing by one character rather
1544            than one byte was nonsense.)
1545    
1546        (c) The pcredemo.c sample program did not have any code at all to handle
1547            the cases when CRLF is a valid newline sequence.
1548    
1549    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1550        as a starting offset was within the subject string. There is now a new
1551        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1552        negative or greater than the length of the string. In order to test this,
1553        pcretest is extended to allow the setting of negative starting offsets.
1554    
1555    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1556        starting offset points to the beginning of a UTF-8 character was
1557        unnecessarily clumsy. I tidied it up.
1558    
1559    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1560        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1561    
1562    17. Nobody had reported that the --include_dir option, which was added in
1563        release 7.7 should have been called --include-dir (hyphen, not underscore)
1564        for compatibility with GNU grep. I have changed it to --include-dir, but
1565        left --include_dir as an undocumented synonym, and the same for
1566        --exclude-dir, though that is not available in GNU grep, at least as of
1567        release 2.5.4.
1568    
1569    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1570        characters from a string of bytes) have been redefined so as not to use
1571        loops, in order to improve performance in some environments. At the same
1572        time, I abstracted some of the common code into auxiliary macros to save
1573        repetition (this should not affect the compiled code).
1574    
1575    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1576        compile-time error is now given if \c is not followed by an ASCII
1577        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1578        different, and any byte value is allowed.)
1579    
1580    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1581        START_OPTIMIZE option, which is now allowed at compile time - but just
1582        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1583        to pcregrep and other applications that have no direct access to PCRE
1584        options. The new /Y option in pcretest sets this option when calling
1585        pcre_compile().
1586    
1587    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1588        back references. Groups containing recursive back references were forced to
1589        be atomic by that change, but in the case of named groups, the amount of
1590        memory required was incorrectly computed, leading to "Failed: internal
1591        error: code overflow". This has been fixed.
1592    
1593    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1594        pcretest.c, to avoid build problems in some Borland environments.
1595    
1596    
1597    Version 8.10 25-Jun-2010
1598    ------------------------
1599    
1600    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1601        THEN.
1602    
1603    2.  (*ACCEPT) was not working when inside an atomic group.
1604    
1605    3.  Inside a character class, \B is treated as a literal by default, but
1606        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1607        causes the error). The code is unchanged, but I tidied the documentation.
1608    
1609    4.  Inside a character class, PCRE always treated \R and \X as literals,
1610        whereas Perl faults them if its -w option is set. I have changed PCRE so
1611        that it faults them when PCRE_EXTRA is set.
1612    
1613    5.  Added support for \N, which always matches any character other than
1614        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1615    
1616    6.  When compiling pcregrep with newer versions of gcc which may have
1617        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1618        declared with attribute warn_unused_result" were given. Just casting the
1619        result to (void) does not stop the warnings; a more elaborate fudge is
1620        needed. I've used a macro to implement this.
1621    
1622    7.  Minor change to pcretest.c to avoid a compiler warning.
1623    
1624    8.  Added four artifical Unicode properties to help with an option to make
1625        \s etc use properties (see next item). The new properties are: Xan
1626        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1627    
1628    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1629        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1630        this option. Modified pcretest to add /W to test this facility. Added
1631        REG_UCP to make it available via the POSIX interface.
1632    
1633    10. Added --line-buffered to pcregrep.
1634    
1635    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1636        studied, and the match started with a letter with a code point greater than
1637        127 whose first byte was different to the first byte of the other case of
1638        the letter, the other case of this starting letter was not recognized
1639        (#976).
1640    
1641    12. If a pattern that was studied started with a repeated Unicode property
1642        test, for example, \p{Nd}+, there was the theoretical possibility of
1643        setting up an incorrect bitmap of starting bytes, but fortunately it could
1644        not have actually happened in practice until change 8 above was made (it
1645        added property types that matched character-matching opcodes).
1646    
1647    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1648        possible starting bytes for non-anchored patterns.
1649    
1650    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1651        \R, and also a number of cases that involve Unicode properties, both
1652        explicit and implicit when PCRE_UCP is set.
1653    
1654    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1655        input, it could crash or give wrong results if characters with values
1656        greater than 0xc0 were present in the subject string. (Detail: it assumed
1657        UTF-8 input when processing these items.)
1658    
1659    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1660        size_t is 64-bit (#991).
1661    
1662    17. Added a check for running out of memory when PCRE is compiled with
1663        --disable-stack-for-recursion (#990).
1664    
1665    18. If the last data line in a file for pcretest does not have a newline on
1666        the end, a newline was missing in the output.
1667    
1668    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1669        less than 128) in its various bitmaps. However, there is a facility for
1670        generating tables according to the current locale when PCRE is compiled. It
1671        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1672        characters, are recognized by isspace() and therefore were getting set in
1673        these tables, and indeed these tables seem to approximate to ISO 8859. This
1674        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1675        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1676        which of course cannot start UTF-8 characters. I have changed the code so
1677        that only real ASCII characters (less than 128) and the correct starting
1678        bytes for UTF-8 encodings are set for characters greater than 127 when in
1679        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1680        altogether.)
1681    
1682    20. Added the /T option to pcretest so as to be able to run tests with non-
1683        standard character tables, thus making it possible to include the tests
1684        used for 19 above in the standard set of tests.
1685    
1686    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1687        reference to a subpattern the other side of a comment that contains an
1688        opening parenthesis caused either an internal compiling error, or a
1689        reference to the wrong subpattern.
1690    
1691    
1692    Version 8.02 19-Mar-2010
1693    ------------------------
1694    
1695    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1696    
1697    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1698        configured.
1699    
1700    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1701        original author of that file, following a query about its status.
1702    
1703    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1704        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1705    
1706    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1707        quantifier applied to a forward-referencing subroutine call, could compile
1708        incorrect code or give the error "internal error: previously-checked
1709        referenced subpattern not found".
1710    
1711    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1712        variables to point to external functions. For these systems, therefore,
1713        pcre_malloc etc. are now initialized to local functions that call the
1714        relevant global functions.
1715    
1716    7.  There were two entries missing in the vectors called coptable and poptable
1717        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1718        I've fixed the data, and added a kludgy way of testing at compile time that
1719        the lengths are correct (equal to the number of opcodes).
1720    
1721    8.  Following on from 7, I added a similar kludge to check the length of the
1722        eint vector in pcreposix.c.
1723    
1724    9.  Error texts for pcre_compile() are held as one long string to avoid too
1725        much relocation at load time. To find a text, the string is searched,
1726        counting zeros. There was no check for running off the end of the string,
1727        which could happen if a new error number was added without updating the
1728        string.
1729    
1730    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1731    
1732    11. \K was not working if it appeared in an atomic group or in a group that
1733        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1734        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1735        the assertion is positive, but not if it is negative.
1736    
1737    12. Change 11 fortuitously reduced the size of the stack frame used in the
1738        "match()" function of pcre_exec.c by one pointer. Forthcoming
1739        implementation of support for (*MARK) will need an extra pointer on the
1740        stack; I have reserved it now, so that the stack frame size does not
1741        decrease.
1742    
1743    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1744        item in branch that calls a recursion is a subroutine call - as in the
1745        second branch in the above example - was incorrectly given the compile-
1746        time error "recursive call could loop indefinitely" because pcre_compile()
1747        was not correctly checking the subroutine for matching a non-empty string.
1748    
1749    14. The checks for overrunning compiling workspace could trigger after an
1750        overrun had occurred. This is a "should never occur" error, but it can be
1751        triggered by pathological patterns such as hundreds of nested parentheses.
1752        The checks now trigger 100 bytes before the end of the workspace.
1753    
1754    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1755    
1756    
1757    Version 8.01 19-Jan-2010
1758    ------------------------
1759    
1760    1.  If a pattern contained a conditional subpattern with only one branch (in
1761        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1762        computed the wrong minimum data length (which is of course zero for such
1763        subpatterns). This could cause incorrect "no match" results.
1764    
1765    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1766        the pattern is reset in the first branch, pcre_compile() failed with
1767        "internal error: code overflow at offset...". This happened only when
1768        the reset was to the original external option setting. (An optimization
1769        abstracts leading options settings into an external setting, which was the
1770        cause of this.)
1771    
1772    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1773        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1774        assertion pattern did not match (meaning that the assertion was true), it
1775        was incorrectly treated as false if the SKIP had been reached during the
1776        matching. This also applied to assertions used as conditions.
1777    
1778    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1779        assertion subpattern, including such a pattern used as a condition,
1780        unpredictable results occurred, instead of the error return
1781        PCRE_ERROR_DFA_UITEM.
1782    
1783    5.  The C++ GlobalReplace function was not working like Perl for the special
1784        situation when an empty string is matched. It now does the fancy magic
1785        stuff that is necessary.
1786    
1787    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1788        removed. (These were left over from very, very early versions of PCRE.)
1789    
1790    7.  Some cosmetic changes to the code to make life easier when compiling it
1791        as part of something else:
1792    
1793        (a) Change DEBUG to PCRE_DEBUG.
1794    
1795        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1796            called "current" as "current_branch", to prevent a collision with the
1797            Linux macro when compiled as a kernel module.
1798    
1799        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1800            prevent a collision with the Linux macro when compiled as a kernel
1801            module.
1802    
1803    8.  In pcre_compile() there are some checks for integer overflows that used to
1804        cast potentially large values to (double). This has been changed to that
1805        when building, a check for int64_t is made, and if it is found, it is used
1806        instead, thus avoiding the use of floating point arithmetic. (There is no
1807        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1808        double.
1809    
1810    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1811        2005 (difference between two addresses compared to an unsigned value).
1812    
1813    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1814        custom one, because of the following reported problem in Windows:
1815    
1816          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1817              under Win32.
1818          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1819              therefore missing the function definition.
1820          - The compiler thus generates a "C" signature for the test function.
1821          - The linker fails to find the "C" function.
1822          - PCRE fails to configure if asked to do so against libbz2.
1823    
1824    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1825        messages were output:
1826    
1827          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1828          rerunning libtoolize, to keep the correct libtool macros in-tree.
1829          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1830    
1831        I have done both of these things.
1832    
1833    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1834        most of the time, it *can* run out if it is given a pattern that contains a
1835        runaway infinite recursion. I updated the discussion in the pcrestack man
1836        page.
1837    
1838    13. Now that we have gone to the x.xx style of version numbers, the minor
1839        version may start with zero. Using 08 or 09 is a bad idea because users
1840        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1841        interpreted as invalid octal numbers. I've updated the previous comment in
1842        configure.ac, and also added a check that gives an error if 08 or 09 are
1843        used.
1844    
1845    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1846        causing partial matching to fail when the end of the subject matched \W
1847        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1848    
1849    15. There were some discrepancies between the declarations in pcre_internal.h
1850        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1851        their definitions. The declarations used "const uschar *" and the
1852        definitions used USPTR. Even though USPTR is normally defined as "const
1853        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1854        reported that: "This difference in casting confuses some C++ compilers, for
1855        example, SunCC recognizes above declarations as different functions and
1856        generates broken code for hbpcre." I have changed the declarations to use
1857        USPTR.
1858    
1859    16. GNU libtool is named differently on some systems. The autogen.sh script now
1860        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1861        (FreeBSD).
1862    
1863    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1864        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1865        comment: "Figure out how to create a longlong from a string: strtoll and
1866        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1867        instance, but it only takes 2 args instead of 3!"
1868    
1869    18. A subtle bug concerned with back references has been fixed by a change of
1870        specification, with a corresponding code fix. A pattern such as
1871        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1872        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1873        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1874        same bug. Such groups have to be quantified to be useful, or contained
1875        inside another quantified group. (If there's no repetition, the reference
1876        can never match.) The problem arises because, having left the group and
1877        moved on to the rest of the pattern, a later failure that backtracks into
1878        the group uses the captured value from the final iteration of the group
1879        rather than the correct earlier one. I have fixed this in PCRE by forcing
1880        any group that contains a reference to itself to be an atomic group; that
1881        is, there cannot be any backtracking into it once it has completed. This is
1882        similar to recursive and subroutine calls.
1883    
1884    
1885    Version 8.00 19-Oct-09
1886    ----------------------
1887    
1888    1.  The table for translating pcre_compile() error codes into POSIX error codes
1889        was out-of-date, and there was no check on the pcre_compile() error code
1890        being within the table. This could lead to an OK return being given in
1891        error.
1892    
1893    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1894        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1895        in a Windows environment.
1896    
1897    3.  The pcregrep --count option prints the count for each file even when it is
1898        zero, as does GNU grep. However, pcregrep was also printing all files when
1899        --files-with-matches was added. Now, when both options are given, it prints
1900        counts only for those files that have at least one match. (GNU grep just
1901        prints the file name in this circumstance, but including the count seems
1902        more useful - otherwise, why use --count?) Also ensured that the
1903        combination -clh just lists non-zero counts, with no names.
1904    
1905    4.  The long form of the pcregrep -F option was incorrectly implemented as
1906        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1907        but it seems right to fix it, and I didn't think it was worth preserving
1908        the old behaviour.
1909    
1910    5.  The command line items --regex=pattern and --regexp=pattern were not
1911        recognized by pcregrep, which required --regex pattern or --regexp pattern
1912        (with a space rather than an '='). The man page documented the '=' forms,
1913        which are compatible with GNU grep; these now work.
1914    
1915    6.  No libpcreposix.pc file was created for pkg-config; there was just
1916        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1917    
1918    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1919        when UCP support is not needed, by modifying the Python script that
1920        generates it from Unicode data files. This should not matter if the module
1921        is correctly used as a library, but I received one complaint about 50K of
1922        unwanted data. My guess is that the person linked everything into his
1923        program rather than using a library. Anyway, it does no harm.
1924    
1925    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1926        was a minimum greater than 1 for a wide character in a possessive
1927        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1928        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1929        character. Chaos in the form of incorrect output or a compiling loop could
1930        result.
1931    
1932    9.  The restrictions on what a pattern can contain when partial matching is
1933        requested for pcre_exec() have been removed. All patterns can now be
1934        partially matched by this function. In addition, if there are at least two
1935        slots in the offset vector, the offset of the earliest inspected character
1936        for the match and the offset of the end of the subject are set in them when
1937        PCRE_ERROR_PARTIAL is returned.
1938    
1939    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1940        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1941        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1942        and may be more useful for multi-segment matching.
1943    
1944    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1945        used to be given if ever the end of the subject was reached; now it is
1946        given only if matching could not proceed because another character was
1947        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1948        string "Z", which now yields "no match" instead of "partial match". In the
1949        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1950        final character ended with (*FAIL).
1951    
1952    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1953        if the pattern had a "must contain" character that was already found in the
1954        earlier partial match, unless partial matching was again requested. For
1955        example, with the pattern /dog.(body)?/, the "must contain" character is
1956        "g". If the first part-match was for the string "dog", restarting with
1957        "sbody" failed. This bug has been fixed.
1958    
1959    13. The string returned by pcre_dfa_exec() after a partial match has been
1960        changed so that it starts at the first inspected character rather than the
1961        first character of the match. This makes a difference only if the pattern
1962        starts with a lookbehind assertion or \b or \B (\K is not supported by
1963        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1964        matching functions compatible, and I think it's the right thing to do.
1965    
1966    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1967        so that the demonstration program is easily available in environments where
1968        PCRE has not been installed from source.
1969    
1970    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1971        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1972        library.
1973    
1974    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1975        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1976        is not the first non-POSIX option to be added. Clearly some people find
1977        these options useful.
1978    
1979    17. If a caller to the POSIX matching function regexec() passes a non-zero
1980        value for nmatch with a NULL value for pmatch, the value of
1981        nmatch is forced to zero.
1982    
1983    18. RunGrepTest did not have a test for the availability of the -u option of
1984        the diff command, as RunTest does. It now checks in the same way as
1985        RunTest, and also checks for the -b option.
1986    
1987    19. If an odd number of negated classes containing just a single character
1988        interposed, within parentheses, between a forward reference to a named
1989        subpattern and the definition of the subpattern, compilation crashed with
1990        an internal error, complaining that it could not find the referenced
1991        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1992        [The bug was that it was starting one character too far in when skipping
1993        over the character class, thus treating the ] as data rather than
1994        terminating the class. This meant it could skip too much.]
1995    
1996    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1997        /g option in pcretest when the pattern contains \K, which makes it possible
1998        to have an empty string match not at the start, even when the pattern is
1999        anchored. Updated pcretest and pcredemo to use this option.
2000    
2001    21. If the maximum number of capturing subpatterns in a recursion was greater
2002        than the maximum at the outer level, the higher number was returned, but
2003        with unset values at the outer level. The correct (outer level) value is
2004        now given.
2005    
2006    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2007        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2008        make it do so. The string so far is captured, making this feature
2009        compatible with Perl.
2010    
2011    23. The tests have been re-organized, adding tests 11 and 12, to make it
2012        possible to check the Perl 5.10 features against Perl 5.10.
2013    
2014    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2015        pattern matches a fixed length string. PCRE did not allow this; now it
2016        does. Neither allows recursion.
2017    
2018    25. I finally figured out how to implement a request to provide the minimum
2019        length of subject string that was needed in order to match a given pattern.
2020        (It was back references and recursion that I had previously got hung up
2021        on.) This code has now been added to pcre_study(); it finds a lower bound
2022        to the length of subject needed. It is not necessarily the greatest lower
2023        bound, but using it to avoid searching strings that are too short does give
2024        some useful speed-ups. The value is available to calling programs via
2025        pcre_fullinfo().
2026    
2027    26. While implementing 25, I discovered to my embarrassment that pcretest had
2028        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2029        study optimizations had never been tested with that matching function.
2030        Oops. What is worse, even when it was passed study data, there was a bug in
2031        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2032        were also very few tests of studied patterns with pcre_dfa_exec().
2033    
2034    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2035        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2036        on the other side of the coin, they are no longer allowed to have different
2037        names, because these cannot be distinguished in PCRE, and this has caused
2038        confusion. (This is a difference from Perl.)
2039    
2040    28. When duplicate subpattern names are present (necessarily with different
2041        numbers, as required by 27 above), and a test is made by name in a
2042        conditional pattern, either for a subpattern having been matched, or for
2043        recursion in such a pattern, all the associated numbered subpatterns are
2044        tested, and the overall condition is true if the condition is true for any
2045        one of them. This is the way Perl works, and is also more like the way
2046        testing by number works.
2047    
2048    
2049    Version 7.9 11-Apr-09
2050    ---------------------
2051    
2052    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2053        (pcretest), all targets were linked against these libraries. This included
2054        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2055        libraries. This caused unwanted dependencies to be created. This problem
2056        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2057        pcretest is linked with readline.
2058    
2059    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2060        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2061        moved outside it again, because FALSE and TRUE are already defined in AIX,
2062        but BOOL is not.
2063    
2064    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2065        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2066    
2067    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2068        hyphens) following file names and line numbers when outputting matching
2069        lines. This is not true; no spaces are inserted. I have also clarified the
2070        wording for the --colour (or --color) option.
2071    
2072    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2073        was not coloured; this is different to GNU grep, so I have changed it to be
2074        the same.
2075    
2076    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2077        each matching line was coloured. Now it goes on to look for further matches
2078        of any of the test patterns, which is the same behaviour as GNU grep.
2079    
2080    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2081        doesn't make sense to accept an empty string match in pcregrep, so I have
2082        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2083        seems to be how GNU grep behaves. [But see later change 40 for release
2084        8.33.]
2085    
2086    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2087        start or after a newline", because the conditional assertion was not being
2088        correctly handled. The rule now is that both the assertion and what follows
2089        in the first alternative must satisfy the test.
2090    
2091    9.  If auto-callout was enabled in a pattern with a conditional group whose
2092        condition was an assertion, PCRE could crash during matching, both with
2093        pcre_exec() and pcre_dfa_exec().
2094    
2095    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2096        used for matching.
2097    
2098    11. Unicode property support in character classes was not working for
2099        characters (bytes) greater than 127 when not in UTF-8 mode.
2100    
2101    12. Added the -M command line option to pcretest.
2102    
2103    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2104    
2105    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2106    
2107    16. Added comments and documentation about mis-use of no_arg in the C++
2108        wrapper.
2109    
2110    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2111        from Martin Jerabek that uses macro names for all relevant character and
2112        string constants.
2113    
2114    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2115        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2116        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2117        these, but not everybody uses configure.
2118    
2119    19. A conditional group that had only one branch was not being correctly
2120        recognized as an item that could match an empty string. This meant that an
2121        enclosing group might also not be so recognized, causing infinite looping
2122        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2123        with the subject "ab", where knowledge that the repeated group can match
2124        nothing is needed in order to break the loop.
2125    
2126    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2127        exec(), but without supplying a callout function, matching went wrong.
2128    
2129    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2130        leak if the size of the offset vector was greater than 30. When the vector
2131        is smaller, the saved offsets during recursion go onto a local stack
2132        vector, but for larger vectors malloc() is used. It was failing to free
2133        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2134        error, in fact).
2135    
2136    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2137        heapframe that is used only when UTF-8 support is enabled. This caused no
2138        problem, but was untidy.
2139    
2140    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2141        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2142        included within another project.
2143    
2144    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2145        slightly modified by me:
2146    
2147          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2148              not building pcregrep.
2149    
2150          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2151              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2152    
2153    25. Forward references, both numeric and by name, in patterns that made use of
2154        duplicate group numbers, could behave incorrectly or give incorrect errors,
2155        because when scanning forward to find the reference group, PCRE was not
2156        taking into account the duplicate group numbers. A pattern such as
2157        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2158    
2159    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2160        the feature of a custom pointer more persuasive (as requested by a user).
2161    
2162    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2163        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2164        pre-defined.
2165    
2166    28. Added support for (*UTF8) at the start of a pattern.
2167    
2168    29. Arrange for flags added by the "release type" setting in CMake to be shown
2169        in the configuration summary.
2170    
2171    
2172    Version 7.8 05-Sep-08
2173    ---------------------
2174    
2175    1.  Replaced UCP searching code with optimized version as implemented for Ad
2176        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2177        stage table and inline lookup instead of a function, giving speed ups of 2
2178        to 5 times on some simple patterns that I tested. Permission was given to
2179        distribute the MultiStage2.py script that generates the tables (it's not in
2180        the tarball, but is in the Subversion repository).
2181    
2182    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2183        scripts.
2184    
2185    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2186        a group with a zero qualifier. The result of the study could be incorrect,
2187        or the function might crash, depending on the pattern.
2188    
2189    4.  Caseless matching was not working for non-ASCII characters in back
2190        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2191        It now works when Unicode Property Support is available.
2192    
2193    5.  In pcretest, an escape such as \x{de} in the data was always generating
2194        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2195        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2196        truncation.
2197    
2198    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2199    
2200    7.  Added two (int) casts to pcregrep when printing the difference of two
2201        pointers, in case they are 64-bit values.
2202    
2203    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2204        test 2 if it fails.
2205    
2206    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2207        and a #define of that name to empty if it is not externally set. This is to
2208        allow users of MSVC to set it if necessary.
2209    
2210    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2211        the convenience functions in the pcre_get.c source file.
2212    
2213    11. An option change at the start of a pattern that had top-level alternatives
2214        could cause overwriting and/or a crash. This command provoked a crash in
2215        some environments:
2216    
2217          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2218    
2219        This potential security problem was recorded as CVE-2008-2371.
2220    
2221    12. For a pattern where the match had to start at the beginning or immediately
2222        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2223        pcre_dfa_exec() could read past the end of the passed subject if there was
2224        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2225        pcretest so that it places the subject at the end of its malloc-ed buffer.
2226    
2227    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2228        exec() might read past the end of the data buffer in UTF-8 mode.
2229    
2230    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2231        the data contained the byte 0x85 as part of a UTF-8 character within its
2232        first line. This applied both to normal and DFA matching.
2233    
2234    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2235        /^[^d]*?$/8 failed to match "abc".
2236    
2237    16. Added a missing copyright notice to pcrecpp_internal.h.
2238    
2239    17. Make it more clear in the documentation that values returned from
2240        pcre_exec() in ovector are byte offsets, not character counts.
2241    
2242    18. Tidied a few places to stop certain compilers from issuing warnings.
2243    
2244    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2245        supplied by Stefan Weber. I made a further small update for 7.8 because
2246        there is a change of source arrangements: the pcre_searchfuncs.c module is
2247        replaced by pcre_ucd.c.
2248    
2249    
2250    Version 7.7 07-May-08
2251    ---------------------
2252    
2253    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2254        a string to a long long, pretend we don't even have a long long." This is
2255        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2256    
2257    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2258        pre-7.6 versions, which defined a global no_arg variable instead of putting
2259        it in the RE class. (See also #8 below.)
2260    
2261    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2262        Lopes.
2263    
2264    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2265    
2266        (1) The include/exclude patterns were being applied to the whole pathnames
2267            of files, instead of just to the final components.
2268    
2269        (2) If there was more than one level of directory, the subdirectories were
2270            skipped unless they satisfied the include/exclude conditions. This is
2271            inconsistent with GNU grep (and could even be seen as contrary to the
2272            pcregrep specification - which I improved to make it absolutely clear).
2273            The action now is always to scan all levels of directory, and just
2274            apply the include/exclude patterns to regular files.
2275    
2276    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2277        --exclude_dir in the tests to avoid scanning .svn directories.
2278    
2279    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2280        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2281        doesn't support NULs in patterns.
2282    
2283    7.  Added some missing "const"s to declarations of static tables in
2284        pcre_compile.c and pcre_dfa_exec.c.
2285    
2286    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2287        caused by fix #2  above. (Subsequently also a second patch to fix the
2288        first patch. And a third patch - this was a messy problem.)
2289    
2290    9.  Applied Craig's patch to remove the use of push_back().
2291    
2292    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2293        matching function regexec().
2294    
2295    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2296        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2297        references. PCRE supports relative numbers with this syntax (I don't think
2298        Oniguruma does).
2299    
2300    12. Previously, a group with a zero repeat such as (...){0} was completely
2301        omitted from the compiled regex. However, this means that if the group
2302        was called as a subroutine from elsewhere in the pattern, things went wrong
2303        (an internal error was given). Such groups are now left in the compiled
2304        pattern, with a new opcode that causes them to be skipped at execution
2305        time.
2306    
2307    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2308        to the way PCRE behaves:
2309    
2310        (a) A lone ] character is dis-allowed (Perl treats it as data).
2311    
2312        (b) A back reference to an unmatched subpattern matches an empty string
2313            (Perl fails the current match path).
2314    
2315        (c) A data ] in a character class must be notated as \] because if the
2316            first data character in a class is ], it defines an empty class. (In
2317            Perl it is not possible to have an empty class.) The empty class []
2318            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2319            The negative empty class [^] matches any one character, independently
2320            of the DOTALL setting.
2321    
2322    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2323        non-existent subpattern following a character class starting with ']' and
2324        containing () gave an internal compiling error instead of "reference to
2325        non-existent subpattern". Fortunately, when the pattern did exist, the
2326        compiled code was correct. (When scanning forwards to check for the
2327        existence of the subpattern, it was treating the data ']' as terminating
2328        the class, so got the count wrong. When actually compiling, the reference
2329        was subsequently set up correctly.)
2330    
2331    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2332        it was being rejected as not supported by pcre_dfa_exec(), even though
2333        other assertions are supported. I have made pcre_dfa_exec() support
2334        (*FAIL).
2335    
2336    16. The implementation of 13c above involved the invention of a new opcode,
2337        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2338        cannot be changed at match time, I realized I could make a small
2339        improvement to matching performance by compiling OP_ALLANY instead of
2340        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2341        on the OP_ANY path.
2342    
2343    17. Compiling pcretest on Windows with readline support failed without the
2344        following two fixes: (1) Make the unistd.h include conditional on
2345        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2346    
2347    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2348        ncurses library to be included for pcretest when ReadLine support is
2349        requested, but also to allow for it to be overridden. This patch came from
2350        Daniel Bergström.
2351    
2352    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2353        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2354        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2355        spotting this.
2356    
2357    
2358    Version 7.6 28-Jan-08
2359    ---------------------
2360    
2361    1.  A character class containing a very large number of characters with
2362        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2363        overflow.
2364    
2365    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2366        HAVE_LONG_LONG is not defined.
2367    
2368    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2369        bring them up to date and include new features. This patch includes:
2370    
2371        - Fixed PH's badly added libz and libbz2 support.
2372        - Fixed a problem with static linking.
2373        - Added pcredemo. [But later removed - see 7 below.]
2374        - Fixed dftables problem and added an option.
2375        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2376            HAVE_LONG_LONG.
2377        - Added readline support for pcretest.
2378        - Added an listing of the option settings after cmake has run.
2379    
2380    4.  A user submitted a patch to Makefile that makes it easy to create
2381        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2382        Makefile.am that cause it to include this special target, without
2383        affecting anything else. Note that the same mingw target plus all
2384        the other distribution libraries and programs are now supported
2385        when configuring with CMake (see 6 below) instead of with
2386        Configure/Make.
2387    
2388    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2389        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2390        exported in the Windows port". It has not yet been confirmed that the patch
2391        solves the problem, but it does no harm.
2392    
2393    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2394        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2395        with CMake, and also correct the comment about stack recursion.
2396    
2397    7.  Remove the automatic building of pcredemo from the ./configure system and
2398        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2399        of a program that users should build themselves after PCRE is installed, so
2400        building it automatically is not really right. What is more, it gave
2401        trouble in some build environments.
2402    
2403    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2404    
2405    
2406    Version 7.5 10-Jan-08
2407    ---------------------
2408    
2409    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2410        values in parens when parsing an RE using the C++ wrapper."
2411    
2412    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2413        Characters greater than 255 were excluded from the class instead of being
2414        included.
2415    
2416    3.  The same bug as (2) above applied to negated POSIX classes such as
2417        [:^space:].
2418    
2419    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2420        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2421        I have changed it.
2422    
2423    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2424        first named subpattern) and a construct such as (?&a) would reference the
2425        first named subpattern whose name started with "a" (in other words, the
2426        length check was missing). Both these problems are fixed. "Subpattern name
2427        expected" is now given for (?&) (a zero-length name), and this patch also
2428        makes it give the same error for \k'' (previously it complained that that
2429        was a reference to a non-existent subpattern).
2430    
2431    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2432        this is right because (?- can be followed by option settings as well as by
2433        digits. I have, however, made the messages clearer.
2434    
2435    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2436        than the number used in the conditional) now cause a compile-time error.
2437        This is actually not compatible with Perl, which accepts such patterns, but
2438        treats the conditional as always being FALSE (as PCRE used to), but it
2439        seems to me that giving a diagnostic is better.
2440    
2441    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2442        and messages.
2443    
2444    9.  Fix two occurrences of "backslash" in comments that should have been
2445        "backspace".
2446    
2447    10. Remove two redundant lines of code that can never be obeyed (their function
2448        was moved elsewhere).
2449    
2450    11. The program that makes PCRE's Unicode character property table had a bug
2451        which caused it to generate incorrect table entries for sequences of
2452        characters that have the same character type, but are in different scripts.
2453        It amalgamated them into a single range, with the script of the first of
2454        them. In other words, some characters were in the wrong script. There were
2455        thirteen such cases, affecting characters in the following ranges:
2456    
2457          U+002b0 - U+002c1
2458          U+0060c - U+0060d
2459          U+0061e - U+00612
2460          U+0064b - U+0065e
2461          U+0074d - U+0076d
2462          U+01800 - U+01805
2463          U+01d00 - U+01d77
2464          U+01d9b - U+01dbf
2465          U+0200b - U+0200f
2466          U+030fc - U+030fe
2467          U+03260 - U+0327f
2468          U+0fb46 - U+0fbb1
2469          U+10450 - U+1049d
2470    
2471    12. The -o option (show only the matching part of a line) for pcregrep was not
2472        compatible with GNU grep in that, if there was more than one match in a
2473        line, it showed only the first of them. It now behaves in the same way as
2474        GNU grep.
2475    
2476    13. If the -o and -v options were combined for pcregrep, it printed a blank
2477        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2478        does the same. The return code can be used to tell if there were any
2479        non-matching lines.
2480    
2481    14. Added --file-offsets and --line-offsets to pcregrep.
2482    
2483    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2484        infinitely looping recursion. The bug was that positive lookaheads were not
2485        being skipped when checking for a possible empty match (negative lookaheads
2486        and both kinds of lookbehind were skipped).
2487    
2488    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2489        inclusion of <windows.h> to before rather than after the definition of
2490        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2491    
2492    17. Specifying a possessive quantifier with a specific limit for a Unicode
2493        character property caused pcre_compile() to compile bad code, which led at
2494        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2495        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2496        caused the error; without that there was no problem.
2497    
2498    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2499    
2500    19. Added --enable-pcretest-libreadline.
2501    
2502    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2503        RE::GlobalReplace(). As a result, the number of replacements returned was
2504        double what it should be. I removed one of the increments, but Craig sent a
2505        later patch that removed the other one (the right fix) and added unit tests
2506        that check the return values (which was not done before).
2507    
2508    21. Several CMake things:
2509    
2510        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2511            the names libpcre and libpcreposix, not just pcre and pcreposix.
2512    
2513        (2) The above change means that pcretest and pcregrep are now correctly
2514            linked with the newly-built libraries, not previously installed ones.
2515    
2516        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2517    
2518    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2519        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2520        UTF-8 newline character). The key issue is that the pattern starts .*;
2521        this means that the match must be either at the beginning, or after a
2522        newline. The bug was in the code for advancing after a failed match and
2523        checking that the new position followed a newline. It was not taking
2524        account of UTF-8 characters correctly.
2525    
2526    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2527        character classes. PCRE was not treating the sequence [:...:] as a
2528        character class unless the ... were all letters. Perl, however, seems to
2529        allow any characters between [: and :], though of course it rejects as
2530        unknown any "names" that contain non-letters, because all the known class
2531        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2532        for example, whereas PCRE did not - it did not recognize a POSIX character
2533        class. This seemed a bit dangerous, so the code has been changed to be
2534        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2535        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2536        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2537        Perl does, and where it didn't before.
2538    
2539    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2540        Windows environments %n is disabled by default.
2541    
2542    
2543    Version 7.4 21-Sep-07
2544    ---------------------
2545    
2546    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2547        means that a class such as [\s] counted as "explicit reference to CR or
2548        LF". That isn't really right - the whole point of the change was to try to
2549        help when there was an actual mention of one of the two characters. So now
2550        the change happens only if \r or \n (or a literal CR or LF) character is
2551        encountered.
2552    
2553    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2554        of both had grown to the point where there were only 3 bits left.
2555        Fortunately, there was spare space in the data structure, and so I have
2556        moved the internal flags into a new 16-bit field to free up more option
2557        bits.
2558    
2559    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2560        but did not set the internal JCHANGED flag - either of these is enough to
2561        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2562        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2563        start sets both bits.
2564    
2565    4.  Added options (at build time, compile time, exec time) to change \R from
2566        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2567    
2568    5.  doc/pcresyntax.html was missing from the distribution.
2569    
2570    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2571        compatibility, even though it is no longer used.
2572    
2573    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2574        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2575        windows.h file is present (where different names are used). [This was
2576        reversed later after testing - see 16 below.]
2577    
2578    8.  Changed all #include <config.h> to #include "config.h". There were also
2579        some further <pcre.h> cases that I changed to "pcre.h".
2580    
2581    9.  When pcregrep was used with the --colour option, it missed the line ending
2582        sequence off the lines that it output.
2583    
2584    10. It was pointed out to me that arrays of string pointers cause lots of
2585        relocations when a shared library is dynamically loaded. A technique of
2586        using a single long string with a table of offsets can drastically reduce
2587        these. I have refactored PCRE in four places to do this. The result is
2588        dramatic:
2589    
2590          Originally:                          290
2591          After changing UCP table:            187
2592          After changing error message table:   43
2593          After changing table of "verbs"       36
2594          After changing table of Posix names   22
2595    
2596        Thanks to the folks working on Gregex for glib for this insight.
2597    
2598    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2599        unicode-properties was also set.
2600    
2601    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2602    
2603    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2604        checked only for CRLF.
2605    
2606    14. Added casts to pcretest.c to avoid compiler warnings.
2607    
2608    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2609    
2610    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2611        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2612        entirely. This removes changes made in 7 above.
2613    
2614    17. The CMake files have been updated, and there is now more information about
2615        building with CMake in the NON-UNIX-USE document.
2616    
2617    
2618    Version 7.3 28-Aug-07
2619    ---------------------
2620    
2621     1. In the rejigging of the build system that eventually resulted in 7.1, the
2622        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2623        brackets there is not right, since it causes compilers to look for an
2624        installed pcre.h, not the version that is in the source that is being
2625        compiled (which of course may be different). I have changed it back to:
2626    
2627          #include "pcre.h"
2628    
2629        I have a vague recollection that the change was concerned with compiling in
2630        different directories, but in the new build system, that is taken care of
2631        by the VPATH setting the Makefile.
2632    
2633     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2634        when the subject happened to end in the byte 0x85 (e.g. if the last
2635        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2636        characters but of course it shouldn't be taken as a newline when it is part
2637        of another character. The bug was that, for an unlimited repeat of . in
2638        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2639        characters when looking for a newline.
2640    
2641     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2642    
2643     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2644        in debug output.
2645    
2646     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2647        long printing in the pcrecpp unittest when running under MinGW.
2648    
2649     6. ESC_K was left out of the EBCDIC table.
2650    
2651     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2652        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2653        limit also applies to "virtual nesting" when a pattern is recursive, and in
2654        this case 1000 isn't so big. I have been able to remove this limit at the
2655        expense of backing off one optimization in certain circumstances. Normally,
2656        when pcre_exec() would call its internal match() function recursively and
2657        immediately return the result unconditionally, it uses a "tail recursion"
2658        feature to save stack. However, when a subpattern that can match an empty
2659        string has an unlimited repetition quantifier, it no longer makes this
2660        optimization. That gives it a stack frame in which to save the data for
2661        checking that an empty string has been matched. Previously this was taken
2662        from the 1000-entry workspace that had been reserved. So now there is no
2663        explicit limit, but more stack is used.
2664    
2665     8. Applied Daniel's patches to solve problems with the import/export magic
2666        syntax that is required for Windows, and which was going wrong for the
2667        pcreposix and pcrecpp parts of the library. These were overlooked when this
2668        problem was solved for the main library.
2669    
2670     9. There were some crude static tests to avoid integer overflow when computing
2671        the size of patterns that contain repeated groups with explicit upper
2672        limits. As the maximum quantifier is 65535, the maximum group length was
2673        set at 30,000 so that the product of these two numbers did not overflow a
2674        32-bit integer. However, it turns out that people want to use groups that
2675        are longer than 30,000 bytes (though not repeat them that many times).
2676        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2677        made it possible to implement the integer overflow checks in a much more
2678        dynamic way, which I have now done. The artificial limitation on group
2679        length has been removed - we now have only the limit on the total length of
2680        the compiled pattern, which depends on the LINK_SIZE setting.
2681    
2682    10. Fixed a bug in the documentation for get/copy named substring when
2683        duplicate names are permitted. If none of the named substrings are set, the
2684        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2685        empty string.
2686    
2687    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2688        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2689        because the ] is interpreted as the first data character and the
2690        terminating ] is not found. PCRE has been made compatible with Perl in this
2691        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2692        cause memory overwriting.
2693    
2694    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2695        string has been matched (to stop an infinite loop). It was not recognizing
2696        a conditional subpattern that could match an empty string if that
2697        subpattern was within another subpattern. For example, it looped when
2698        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2699        condition was not nested. This bug has been fixed.
2700    
2701    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2702        past the start of the subject in the presence of bytes with the top bit
2703        set, for example "\x8aBCD".
2704    
2705    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2706        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2707    
2708    14. Optimized (?!) to (*FAIL).
2709    
2710    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2711        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2712        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2713        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2714        does: it's just the validity check that is more restrictive.
2715    
2716    16. Inserted checks for integer overflows during escape sequence (backslash)
2717        processing, and also fixed erroneous offset values for syntax errors during
2718        backslash processing.
2719    
2720    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2721        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2722    
2723    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2724        caused an overrun.
2725    
2726    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2727        something other than just ASCII characters) inside a group that had an
2728        unlimited repeat caused a loop at compile time (while checking to see
2729        whether the group could match an empty string).
2730    
2731    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2732        [\P{Any}] did so. (Error in the code for printing property names.)
2733    
2734    21. An orphan \E inside a character class could cause a crash.
2735    
2736    22. A repeated capturing bracket such as (A)? could cause a wild memory
2737        reference during compilation.
2738    
2739    23. There are several functions in pcre_compile() that scan along a compiled
2740        expression for various reasons (e.g. to see if it's fixed length for look
2741        behind). There were bugs in these functions when a repeated \p or \P was
2742        present in the pattern. These operators have additional parameters compared
2743        with \d, etc, and these were not being taken into account when moving along
2744        the compiled data. Specifically:
2745    
2746        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2747            length.
2748    
2749        (b) An item such as \pL+ within a repeated group could cause crashes or
2750            loops.
2751    
2752        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2753            "reference to non-existent subpattern" error.
2754    
2755        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2756    
2757    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2758        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2759    
2760    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2761    
2762    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2763        character were causing crashes (broken optimization).
2764    
2765    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2766        \p or \P) caused a compile-time loop.
2767    
2768    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2769        break. For example, the unstudied pattern [\r\n]A does not match the string
2770        "\r\nA" because change 7.0/46 below moves the current point on by two
2771        characters after failing to match at the start. However, the pattern \nA
2772        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2773        the same is true. There doesn't seem any very clean way out of this, but
2774        what I have chosen to do makes the common cases work: PCRE now takes note
2775        of whether there can be an explicit match for \r or \n anywhere in the
2776        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2777        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2778        pattern has explicit CR or LF references.
2779    
2780    29. Added (*CR) etc for changing newline setting at start of pattern.
2781    
2782    
2783    Version 7.2 19-Jun-07
2784    ---------------------
2785    
2786     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2787        which is apparently normally available under Windows.
2788    
2789     2. Re-jig the pcregrep tests with different newline settings in an attempt
2790        to make them independent of the local environment's newline setting.
2791    
2792     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2793    
2794     4. Some of the "internals" tests were previously cut out when the link size
2795        was not 2, because the output contained actual offsets. The recent new
2796        "Z" feature of pcretest means that these can be cut out, making the tests
2797        usable with all link sizes.
2798    
2799     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2800        stack recursion. This gives a massive performance boost under BSD, but just
2801        a small improvement under Linux. However, it saves one field in the frame
2802        in all cases.
2803    
2804     6. Added more features from the forthcoming Perl 5.10:
2805    
2806        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2807            recursion call. It refers to the nth most recently opened parentheses.
2808    
2809        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2810            to be opened parentheses.
2811    
2812        (c) Conditions that refer to capturing parentheses can be specified
2813            relatively, for example, (?(-2)... or (?(+3)...
2814    
2815        (d) \K resets the start of the current match so that everything before
2816            is not part of it.
2817    
2818        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2819    
2820        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2821            reference syntax.
2822    
2823        (g) (?| introduces a group in which the numbering of parentheses in each
2824            alternative starts with the same number.
2825    
2826        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2827    
2828     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2829        PCRE_INFO_JCHANGED.
2830    
2831     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2832        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2833        for detecting groups that can match an empty string.
2834    
2835     9. A pattern with a very large number of alternatives (more than several
2836        hundred) was running out of internal workspace during the pre-compile
2837        phase, where pcre_compile() figures out how much memory will be needed. A
2838        bit of new cunning has reduced the workspace needed for groups with
2839        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2840        workspace instead of running out of the 4096 that are available.
2841    
2842    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2843    
2844    11. Applied patch from Google to remove an optimization that didn't quite work.
2845        The report of the bug said:
2846    
2847          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2848          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2849          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2850    
2851    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2852        it matched the wrong number of bytes.
2853    
2854    
2855    Version 7.1 24-Apr-07
2856  ---------------------  ---------------------
2857    
2858   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
# Line 33  Version 7.1 12-Mar-07 Line 2884  Version 7.1 12-Mar-07
2884   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2885      Weber: (1) pcre_internal.h was missing some function renames; (2) updated      Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2886      makevp.bat for the current PCRE, using the additional files      makevp.bat for the current PCRE, using the additional files
2887      makevp-c.txt, makevp-l.txt, and pcregexp.pas.      makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2888    
2889   6. A Windows user reported a minor discrepancy with test 2, which turned out   6. A Windows user reported a minor discrepancy with test 2, which turned out
2890      to be caused by a trailing space on an input line that had got lost in his      to be caused by a trailing space on an input line that had got lost in his
# Line 115  Version 7.1 12-Mar-07 Line 2966  Version 7.1 12-Mar-07
2966    
2967  14. If erroroffset was passed as NULL to pcre_compile, it provoked a  14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2968      segmentation fault instead of returning the appropriate error message.      segmentation fault instead of returning the appropriate error message.
2969    
2970  15. In multiline mode when the newline sequence was set to "any", the pattern  15. In multiline mode when the newline sequence was set to "any", the pattern
2971      ^$ would give a match between the \r and \n of a subject such as "A\r\nB".      ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2972      This doesn't seem right; it now treats the CRLF combination as the line      This doesn't seem right; it now treats the CRLF combination as the line
2973      ending, and so does not match in that case. It's only a pattern such as ^$      ending, and so does not match in that case. It's only a pattern such as ^$
2974      that would hit this one: something like ^ABC$ would have failed after \r      that would hit this one: something like ^ABC$ would have failed after \r
2975      and then tried again after \r\n.      and then tried again after \r\n.
2976    
2977  16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"  16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2978      in an attempt to make files that differ only in their line terminators      in an attempt to make files that differ only in their line terminators
2979      compare equal. This works on Linux.      compare equal. This works on Linux.
2980    
2981    17. Under certain error circumstances pcregrep might try to free random memory
2982        as it exited. This is now fixed, thanks to valgrind.
2983    
2984    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2985        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2986        was because its rules for how to advance for /g after matching an empty
2987        string at the end of a line did not allow for this case. They now check for
2988        it specially.
2989    
2990    20. pcretest is supposed to handle patterns and data of any length, by
2991        extending its buffers when necessary. It was getting this wrong when the
2992        buffer for a data line had to be extended.
2993    
2994    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2995        CRLF as a newline sequence.
2996    
2997    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2998        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2999        I have nevertheless tidied it up.
3000    
3001    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3002    
3003    24. Added a man page for pcre-config.
3004    
3005    
3006  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06

Legend:
Removed from v.131  
changed lines
  Added in v.1488

  ViewVC Help
Powered by ViewVC 1.1.5