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

Diff of /code/trunk/ChangeLog

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

revision 600 by ph10, Mon May 9 08:54:11 2011 UTC revision 1517 by ph10, Fri Feb 6 12:10:13 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 8.13 30-Apr-2011  Version 8.37 xx-xxx-2015
5    ------------------------
6    
7    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8        for those parentheses to be closed with whatever has been captured so far.
9        However, it was failing to mark any other groups between the hightest
10        capture so far and the currrent group as "unset". Thus, the ovector for
11        those groups contained whatever was previously there. An example is the
12        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13    
14    2.  If an assertion condition was quantified with a minimum of zero (an odd
15        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16    
17    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
18        unrecognized modifier, a crash could occur.
19    
20    4.  An attempt to do global matching in pcretest with a zero-length ovector
21        caused a crash.
22    
23    5.  Fixed a memory leak during matching that could occur for a subpattern
24        subroutine call (recursive or otherwise) if the number of captured groups
25        that had to be saved was greater than ten.
26    
27    
28    Version 8.36 26-September-2014
29    ------------------------------
30    
31    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
32        -Wmissing-field-initializers and -Wunused-parameter.
33    
34    2.  The tests for quantifiers being too big (greater than 65535) were being
35        applied after reading the number, and stupidly assuming that integer
36        overflow would give a negative number. The tests are now applied as the
37        numbers are read.
38    
39    3.  Tidy code in pcre_exec.c where two branches that used to be different are
40        now the same.
41    
42    4.  The JIT compiler did not generate match limit checks for certain
43        bracketed expressions with quantifiers. This may lead to exponential
44        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
45        issue should be resolved now.
46    
47    5.  Fixed an issue, which occures when nested alternatives are optimized
48        with table jumps.
49    
50    6.  Inserted two casts and changed some ints to size_t in the light of some
51        reported 64-bit compiler warnings (Bugzilla 1477).
52    
53    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
54        an empty string, which sometimes were behaving incorrectly in the
55        interpreter (though correctly in the JIT matcher). This pcretest input is
56        an example:
57    
58          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
59          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
60    
61        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
62        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
63        for an empty string was breaking the inner loop and carrying on at a lower
64        level, when possessive repeated groups should always return to a higher
65        level as they have no backtrack points in them. The empty string test now
66        occurs at the outer level.
67    
68    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
69        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
70    
71    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
72        doesn't).
73    
74    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
75        to hold a pattern to be incorrectly computed (too small) when there were
76        named back references to duplicated names. This could cause "internal
77        error: code overflow" or "double free or corruption" or other memory
78        handling errors.
79    
80    11. When named subpatterns had the same prefixes, back references could be
81        confused. For example, in this pattern:
82    
83          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
84    
85        the reference to 'Name' was incorrectly treated as a reference to a
86        duplicate name.
87    
88    12. A pattern such as /^s?c/mi8 where the optional character has more than
89        one "other case" was incorrectly compiled such that it would only try to
90        match starting at "c".
91    
92    13. When a pattern starting with \s was studied, VT was not included in the
93        list of possible starting characters; this should have been part of the
94        8.34/18 patch.
95    
96    14. If a character class started [\Qx]... where x is any character, the class
97        was incorrectly terminated at the ].
98    
99    15. If a pattern that started with a caseless match for a character with more
100        than one "other case" was studied, PCRE did not set up the starting code
101        unit bit map for the list of possible characters. Now it does. This is an
102        optimization improvement, not a bug fix.
103    
104    16. The Unicode data tables have been updated to Unicode 7.0.0.
105    
106    17. Fixed a number of memory leaks in pcregrep.
107    
108    18. Avoid a compiler warning (from some compilers) for a function call with
109        a cast that removes "const" from an lvalue by using an intermediate
110        variable (to which the compiler does not object).
111    
112    19. Incorrect code was compiled if a group that contained an internal recursive
113        back reference was optional (had quantifier with a minimum of zero). This
114        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
115        caused segmentation faults because of stack overflows at compile time.
116    
117    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
118        group that is quantified with an indefinite repeat, caused a compile-time
119        loop which used up all the system stack and provoked a segmentation fault.
120        This was not the same bug as 19 above.
121    
122    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
123        Patch by Mike Frysinger.
124    
125    
126    Version 8.35 04-April-2014
127    --------------------------
128    
129    1.  A new flag is set, when property checks are present in an XCLASS.
130        When this flag is not set, PCRE can perform certain optimizations
131        such as studying these XCLASS-es.
132    
133    2.  The auto-possessification of character sets were improved: a normal
134        and an extended character set can be compared now. Furthermore
135        the JIT compiler optimizes more character set checks.
136    
137    3.  Got rid of some compiler warnings for potentially uninitialized variables
138        that show up only when compiled with -O2.
139    
140    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
141        of a match later then the end of the match. The pcretest program was not
142        handling the case sensibly - it was outputting from the start to the next
143        binary zero. It now reports this situation in a message, and outputs the
144        text from the end to the start.
145    
146    5.  Fast forward search is improved in JIT. Instead of the first three
147        characters, any three characters with fixed position can be searched.
148        Search order: first, last, middle.
149    
150    6.  Improve character range checks in JIT. Characters are read by an inprecise
151        function now, which returns with an unknown value if the character code is
152        above a certain threshold (e.g: 256). The only limitation is that the value
153        must be bigger than the threshold as well. This function is useful when
154        the characters above the threshold are handled in the same way.
155    
156    7.  The macros whose names start with RAWUCHAR are placeholders for a future
157        mode in which only the bottom 21 bits of 32-bit data items are used. To
158        make this more memorable for those maintaining the code, the names have
159        been changed to start with UCHAR21, and an extensive comment has been added
160        to their definition.
161    
162    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
163        to the export list in Makefile.am (they were accidentally omitted from the
164        8.34 tarball).
165    
166    9.  The informational output from pcretest used the phrase "starting byte set"
167        which is inappropriate for the 16-bit and 32-bit libraries. As the output
168        for "first char" and "need char" really means "non-UTF-char", I've changed
169        "byte" to "char", and slightly reworded the output. The documentation about
170        these values has also been (I hope) clarified.
171    
172    10. Another JIT related optimization: use table jumps for selecting the correct
173        backtracking path, when more than four alternatives are present inside a
174        bracket.
175    
176    11. Empty match is not possible, when the minimum length is greater than zero,
177        and there is no \K in the pattern. JIT should avoid empty match checks in
178        such cases.
179    
180    12. In a caseless character class with UCP support, when a character with more
181        than one alternative case was not the first character of a range, not all
182        the alternative cases were added to the class. For example, s and \x{17f}
183        are both alternative cases for S: the class [RST] was handled correctly,
184        but [R-T] was not.
185    
186    13. The configure.ac file always checked for pthread support when JIT was
187        enabled. This is not used in Windows, so I have put this test inside a
188        check for the presence of windows.h (which was already tested for).
189    
190    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
191        The algorithm provides a way to skip certain starting offsets, and usually
192        faster than linear prefix searches.
193    
194    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
195        as for 'fr_FR' and 'french'. For some reason, however, it then used the
196        Windows-specific input and output files, which have 'french' screwed in.
197        So this could never have worked. One of the problems with locales is that
198        they aren't always the same. I have now updated RunTest so that it checks
199        the output of the locale test (test 3) against three different output
200        files, and it allows the test to pass if any one of them matches. With luck
201        this should make the test pass on some versions of Solaris where it was
202        failing. Because of the uncertainty, the script did not used to stop if
203        test 3 failed; it now does. If further versions of a French locale ever
204        come to light, they can now easily be added.
205    
206    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
207        there was a message during ./configure, but it did not stop. This now
208        provokes an error. The invalid example in README has been corrected.
209        If a value less than the minimum is given, the minimum value has always
210        been used, but now a warning is given.
211    
212    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
213        was a bug in the test system, which is now fixed. Also, the list of various
214        configurations that are tested for each release did not have one with both
215        16/32 bits and --enable-bar-anycrlf. It now does.
216    
217    18. pcretest was missing "-C bsr" for displaying the \R default setting.
218    
219    19. Little endian PowerPC systems are supported now by the JIT compiler.
220    
221    20. The fast forward newline mechanism could enter to an infinite loop on
222        certain invalid UTF-8 input. Although we don't support these cases
223        this issue can be fixed by a performance optimization.
224    
225    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
226        not take account if existing stack usage. There is now a new global
227        variable called pcre_stack_guard that can be set to point to an external
228        function to check stack availability. It is called at the start of
229        processing every parenthesized group.
230    
231    22. A typo in the code meant that in ungreedy mode the max/min qualifier
232        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
233        not match "ab".
234    
235    23. When UTF was disabled, the JIT program reported some incorrect compile
236        errors. These messages are silenced now.
237    
238    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
239        compiler.
240    
241    25. Change all the temporary files used in RunGrepTest to be different to those
242        used by RunTest so that the tests can be run simultaneously, for example by
243        "make -j check".
244    
245    
246    Version 8.34 15-December-2013
247    -----------------------------
248    
249    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
250        executable memory. Patch inspired by Carsten Klein.
251    
252    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
253        this macro is never tested and has no effect, because the work to support
254        coverage involves only compiling and linking options and special targets in
255        the Makefile. The comment in config.h implied that defining the macro would
256        enable coverage support, which is totally false. There was also support for
257        setting this macro in the CMake files (my fault, I just copied it from
258        configure). SUPPORT_GCOV has now been removed.
259    
260    3.  Make a small performance improvement in strlen16() and strlen32() in
261        pcretest.
262    
263    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
264        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
265        these cases). There was also one in pcretest.c.
266    
267    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
268    
269    6.  In UTF mode, the code for checking whether a group could match an empty
270        string (which is used for indefinitely repeated groups to allow for
271        breaking an infinite loop) was broken when the group contained a repeated
272        negated single-character class with a character that occupied more than one
273        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
274        UTF-8 mode). The effect was undefined: the group might or might not be
275        deemed as matching an empty string, or the program might have crashed.
276    
277    7.  The code for checking whether a group could match an empty string was not
278        recognizing that \h, \H, \v, \V, and \R must match a character.
279    
280    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
281        an empty string. If it can, pcretest shows this in its information output.
282    
283    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
284        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
285        matched by pcre_exec() without using JIT:
286    
287        (a) If the rest of the pattern did not match after a maximal run of
288            grapheme clusters, the code for backing up to try with fewer of them
289            did not always back up over a full grapheme when characters that do not
290            have the modifier quality were involved, e.g. Hangul syllables.
291    
292        (b) If the match point in a subject started with modifier character, and
293            there was no match, the code could incorrectly back up beyond the match
294            point, and potentially beyond the first character in the subject,
295            leading to a segfault or an incorrect match result.
296    
297    10. A conditional group with an assertion condition could lead to PCRE
298        recording an incorrect first data item for a match if no other first data
299        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
300        first data item, and therefore matched "ca" after "c" instead of at the
301        start.
302    
303    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
304        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
305    
306    12. The source of pcregrep now includes z/OS-specific code so that it can be
307        compiled for z/OS as part of the special z/OS distribution.
308    
309    13. Added the -T and -TM options to pcretest.
310    
311    14. The code in pcre_compile.c for creating the table of named capturing groups
312        has been refactored. Instead of creating the table dynamically during the
313        actual compiling pass, the information is remembered during the pre-compile
314        pass (on the stack unless there are more than 20 named groups, in which
315        case malloc() is used) and the whole table is created before the actual
316        compile happens. This has simplified the code (it is now nearly 150 lines
317        shorter) and prepared the way for better handling of references to groups
318        with duplicate names.
319    
320    15. A back reference to a named subpattern when there is more than one of the
321        same name now checks them in the order in which they appear in the pattern.
322        The first one that is set is used for the reference. Previously only the
323        first one was inspected. This change makes PCRE more compatible with Perl.
324    
325    16. Unicode character properties were updated from Unicode 6.3.0.
326    
327    17. The compile-time code for auto-possessification has been refactored, based
328        on a patch by Zoltan Herczeg. It now happens after instead of during
329        compilation. The code is cleaner, and more cases are handled. The option
330        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
331        options in pcretest are provided to set it. It can also be set by
332        (*NO_AUTO_POSSESS) at the start of a pattern.
333    
334    18. The character VT has been added to the default ("C" locale) set of
335        characters that match \s and are generally treated as white space,
336        following this same change in Perl 5.18. There is now no difference between
337        "Perl space" and "POSIX space". Whether VT is treated as white space in
338        other locales depends on the locale.
339    
340    19. The code for checking named groups as conditions, either for being set or
341        for being recursed, has been refactored (this is related to 14 and 15
342        above). Processing unduplicated named groups should now be as fast at
343        numerical groups, and processing duplicated groups should be faster than
344        before.
345    
346    20. Two patches to the CMake build system, by Alexander Barkov:
347    
348          (1) Replace the "source" command by "." in CMakeLists.txt because
349              "source" is a bash-ism.
350    
351          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
352              without these the CMake build does not work on Solaris.
353    
354    21. Perl has changed its handling of \8 and \9. If there is no previously
355        encountered capturing group of those numbers, they are treated as the
356        literal characters 8 and 9 instead of a binary zero followed by the
357        literals. PCRE now does the same.
358    
359    22. Following Perl, added \o{} to specify codepoints in octal, making it
360        possible to specify values greater than 0777 and also making them
361        unambiguous.
362    
363    23. Perl now gives an error for missing closing braces after \x{... instead of
364        treating the string as literal. PCRE now does the same.
365    
366    24. RunTest used to grumble if an inappropriate test was selected explicitly,
367        but just skip it when running all tests. This make it awkward to run ranges
368        of tests when one of them was inappropriate. Now it just skips any
369        inappropriate tests, as it always did when running all tests.
370    
371    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
372        character types such as \d or \w, too many callouts were inserted, and the
373        data that they returned was rubbish.
374    
375    26. In UCP mode, \s was not matching two of the characters that Perl matches,
376        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
377        were matched by \h. The code has now been refactored so that the lists of
378        the horizontal and vertical whitespace characters used for \h and \v (which
379        are defined only in one place) are now also used for \s.
380    
381    27. Add JIT support for the 64 bit TileGX architecture.
382        Patch by Jiong Wang (Tilera Corporation).
383    
384    28. Possessive quantifiers for classes (both explicit and automatically
385        generated) now use special opcodes instead of wrapping in ONCE brackets.
386    
387    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
388        (because it's meaningless), this was not happening when PCRE_CASELESS was
389        set. Not wrong, but inefficient.
390    
391    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
392        properties for \w, \d, etc) is present in a test regex. Otherwise if the
393        test contains no characters greater than 255, Perl doesn't realise it
394        should be using Unicode semantics.
395    
396    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
397        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
398        does in Unicode mode.
399    
400    32. Added the "forbid" facility to pcretest so that putting tests into the
401        wrong test files can sometimes be quickly detected.
402    
403    33. There is now a limit (default 250) on the depth of nesting of parentheses.
404        This limit is imposed to control the amount of system stack used at compile
405        time. It can be changed at build time by --with-parens-nest-limit=xxx or
406        the equivalent in CMake.
407    
408    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
409        errors. Perl warns for these when in warning mode, but PCRE has no facility
410        for giving warnings.
411    
412    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
413        However, this was not working for (?!) because it is optimized to (*FAIL),
414        for which PCRE does not allow quantifiers. The optimization is now disabled
415        when a quantifier follows (?!). I can't see any use for this, but it makes
416        things uniform.
417    
418    36. Perl no longer allows group names to start with digits, so I have made this
419        change also in PCRE. It simplifies the code a bit.
420    
421    37. In extended mode, Perl ignores spaces before a + that indicates a
422        possessive quantifier. PCRE allowed a space before the quantifier, but not
423        before the possessive +. It now does.
424    
425    38. The use of \K (reset reported match start) within a repeated possessive
426        group such as (a\Kb)*+ was not working.
427    
428    40. Document that the same character tables must be used at compile time and
429        run time, and that the facility to pass tables to pcre_exec() and
430        pcre_dfa_exec() is for use only with saved/restored patterns.
431    
432    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
433        features for Builds with MSVC:
434    
435        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
436           builds). The .rc files can be used to set FileDescription and many other
437           attributes.
438    
439        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
440           This allows higher-level build scripts which want .pdb files to avoid
441           hard-coding the exact files needed."
442    
443    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
444        mean "start of word" and "end of word", respectively, as a transition aid.
445    
446    43. A minimizing repeat of a class containing codepoints greater than 255 in
447        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
448        compiled to use the heap for recursion.
449    
450    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
451        is configured.
452    
453    
454    Version 8.33 28-May-2013
455    ------------------------
456    
457    1.  Added 'U' to some constants that are compared to unsigned integers, to
458        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
459        variables that are added to signed variables, to ensure the result is
460        signed and can be negated.
461    
462    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
463        CMake config files.
464    
465    3.  Revise the creation of config.h.generic so that all boolean macros are
466        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
467        overriding via -D on the command line possible.
468    
469    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
470        to unsigned int is reported to make a quite noticeable speed difference in
471        a specific Windows environment. Testing on Linux did also appear to show
472        some benefit (and it is clearly not harmful). Also fixed the definition of
473        Xop which should be unsigned.
474    
475    5.  Related to (4), changing the definition of the intermediate variable cc
476        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
477        improvements.
478    
479    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
480        unnecessary spaces.
481    
482    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
483        and later.
484    
485    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
486    
487    9.  Optimizing fast_forward_start_bits in JIT.
488    
489    10. Adding support for callouts in JIT, and fixing some issues revealed
490        during this work. Namely:
491    
492        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
493    
494        (b) Minimum length was not checked before the matching is started.
495    
496    11. The value of capture_last that is passed to callouts was incorrect in some
497        cases when there was a capture on one path that was subsequently abandoned
498        after a backtrack. Also, the capture_last value is now reset after a
499        recursion, since all captures are also reset in this case.
500    
501    12. The interpreter no longer returns the "too many substrings" error in the
502        case when an overflowing capture is in a branch that is subsequently
503        abandoned after a backtrack.
504    
505    13. In the pathological case when an offset vector of size 2 is used, pcretest
506        now prints out the matched string after a yield of 0 or 1.
507    
508    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
509        Only supported by the JIT compiler at the moment.
510    
511    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
512    
513    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
514        offset of the starting point of the matching process, provided the offsets
515        vector is large enough.
516    
517    17. The \A escape now records a lookbehind value of 1, though its execution
518        does not actually inspect the previous character. This is to ensure that,
519        in partial multi-segment matching, at least one character from the old
520        segment is retained when a new segment is processed. Otherwise, if there
521        are no lookbehinds in the pattern, \A might match incorrectly at the start
522        of a new segment.
523    
524    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
525    
526    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
527        gives some modest performance improvement in 8-bit mode.
528    
529    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
530        be expressed in certain programming languages using Universal Character
531        Names.
532    
533    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
534        which points out that "non characters" are not "characters that may not
535        appear in Unicode strings" but rather "characters that are reserved for
536        internal use and have only local meaning".
537    
538    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
539        there was a conditional group that depended on an assertion, if the
540        assertion was false, the callout that immediately followed the alternation
541        in the condition was skipped when pcre_exec() was used for matching.
542    
543    23. Allow an explicit callout to be inserted before an assertion that is the
544        condition for a conditional group, for compatibility with automatic
545        callouts, which always insert a callout at this point.
546    
547    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
548        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
549    
550    25. (*PRUNE) is now supported by the JIT compiler.
551    
552    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
553    
554    27. Fix the case where there are two or more SKIPs with arguments that may be
555        ignored.
556    
557    28. (*SKIP) is now supported by the JIT compiler.
558    
559    29. (*THEN) is now supported by the JIT compiler.
560    
561    30. Update RunTest with additional test selector options.
562    
563    31. The way PCRE handles backtracking verbs has been changed in two ways.
564    
565        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
566        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
567        backtracking. In some cases this makes it more Perl-compatible, but Perl's
568        rather obscure rules do not always do the same thing.
569    
570        (2) Previously, backtracking verbs were confined within assertions. This is
571        no longer the case for positive assertions, except for (*ACCEPT). Again,
572        this sometimes improves Perl compatibility, and sometimes does not.
573    
574    32. A number of tests that were in test 2 because Perl did things differently
575        have been moved to test 1, because either Perl or PCRE has changed, and
576        these tests are now compatible.
577    
578    32. Backtracking control verbs are now handled in the same way in JIT and
579        interpreter.
580    
581    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
582        contained a forward subroutine reference caused a compile error.
583    
584    34. Auto-detect and optimize limited repetitions in JIT.
585    
586    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
587        blocking (*UTF) etc.
588    
589    36. In the interpreter, maximizing pattern repetitions for characters and
590        character types now use tail recursion, which reduces stack usage.
591    
592    37. The value of the max lookbehind was not correctly preserved if a compiled
593        and saved regex was reloaded on a host of different endianness.
594    
595    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
596        of the compiled pattern block, expand the flags field from 16 to 32 bits
597        because it was almost full.
598    
599    39. Try madvise first before posix_madvise.
600    
601    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
602        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
603        original change locked out all matches of empty strings. This has been
604        changed so that one match of an empty string per line is recognized.
605        Subsequent searches on the same line (for colouring or for --only-matching,
606        for example) do not recognize empty strings.
607    
608    41. Applied a user patch to fix a number of spelling mistakes in comments.
609    
610    42. Data lines longer than 65536 caused pcretest to crash.
611    
612    43. Clarified the data type for length and startoffset arguments for pcre_exec
613        and pcre_dfa_exec in the function-specific man pages, where they were
614        explicitly stated to be in bytes, never having been updated. I also added
615        some clarification to the pcreapi man page.
616    
617    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
618        a segmentation fault.
619    
620    
621    Version 8.32 30-November-2012
622    -----------------------------
623    
624    1.  Improved JIT compiler optimizations for first character search and single
625        character iterators.
626    
627    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
628        Patch by Daniel Richard G.
629    
630    3.  Single character iterator optimizations in the JIT compiler.
631    
632    4.  Improved JIT compiler optimizations for character ranges.
633    
634    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
635        Reported by Giuseppe D'Angelo.
636    
637    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
638        of a line, was being set incorrectly in cases where .* appeared inside
639        atomic brackets at the start of a pattern, or where there was a subsequent
640        *PRUNE or *SKIP.
641    
642    7.  Improved instruction cache flush for POWER/PowerPC.
643        Patch by Daniel Richard G.
644    
645    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
646        grep:
647    
648        (a) There is now no limit to the number of patterns to be matched.
649    
650        (b) An error is given if a pattern is too long.
651    
652        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
653            are now supported.
654    
655        (d) --exclude-from and --include-from (multiple use) have been added.
656    
657        (e) Exclusions and inclusions now apply to all files and directories, not
658            just to those obtained from scanning a directory recursively.
659    
660        (f) Multiple uses of -f and --file-list are now supported.
661    
662        (g) In a Windows environment, the default for -d has been changed from
663            "read" (the GNU grep default) to "skip", because otherwise the presence
664            of a directory in the file list provokes an error.
665    
666        (h) The documentation has been revised and clarified in places.
667    
668    9.  Improve the matching speed of capturing brackets.
669    
670    10. Changed the meaning of \X so that it now matches a Unicode extended
671        grapheme cluster.
672    
673    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
674        out POSIX threads when JIT support is configured.
675    
676    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
677    
678    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
679        structure could go wrong in environments where size_t is not the same size
680        as int.
681    
682    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
683        set.
684    
685    15. The EBCDIC support had decayed; later updates to the code had included
686        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
687        general tidy up of EBCDIC-related issues, and the documentation was also
688        not quite right. There is now a test that can be run on ASCII systems to
689        check some of the EBCDIC-related things (but is it not a full test).
690    
691    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
692        in a small tidy to the code.
693    
694    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
695    
696    18. If the --only-matching (-o) option in pcregrep is specified multiple
697        times, each one causes appropriate output. For example, -o1 -o2 outputs the
698        substrings matched by the 1st and 2nd capturing parentheses. A separating
699        string can be specified by --om-separator (default empty).
700    
701    19. Improving the first n character searches.
702    
703    20. Turn case lists for horizontal and vertical white space into macros so that
704        they are defined only once.
705    
706    21. This set of changes together give more compatible Unicode case-folding
707        behaviour for characters that have more than one other case when UCP
708        support is available.
709    
710        (a) The Unicode property table now has offsets into a new table of sets of
711            three or more characters that are case-equivalent. The MultiStage2.py
712            script that generates these tables (the pcre_ucd.c file) now scans
713            CaseFolding.txt instead of UnicodeData.txt for character case
714            information.
715    
716        (b) The code for adding characters or ranges of characters to a character
717            class has been abstracted into a generalized function that also handles
718            case-independence. In UTF-mode with UCP support, this uses the new data
719            to handle characters with more than one other case.
720    
721        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
722            whose other case is greater than 256 are now correctly matched
723            caselessly. Previously, the high codepoint matched the low one, but not
724            vice versa.
725    
726        (d) The processing of \h, \H, \v, and \ in character classes now makes use
727            of the new class addition function, using character lists defined as
728            macros alongside the case definitions of 20 above.
729    
730        (e) Caseless back references now work with characters that have more than
731            one other case.
732    
733        (f) General caseless matching of characters with more than one other case
734            is supported.
735    
736    22. Unicode character properties were updated from Unicode 6.2.0
737    
738    23. Improved CMake support under Windows. Patch by Daniel Richard G.
739    
740    24. Add support for 32-bit character strings, and UTF-32
741    
742    25. Major JIT compiler update (code refactoring and bugfixing).
743        Experimental Sparc 32 support is added.
744    
745    26. Applied a modified version of Daniel Richard G's patch to create
746        pcre.h.generic and config.h.generic by "make" instead of in the
747        PrepareRelease script.
748    
749    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
750        pcre_compile.c when checking for a zero character.
751    
752    28. Introducing a native interface for JIT. Through this interface, the compiled
753        machine code can be directly executed. The purpose of this interface is to
754        provide fast pattern matching, so several sanity checks are not performed.
755        However, feature tests are still performed. The new interface provides
756        1.4x speedup compared to the old one.
757    
758    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
759        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
760        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
761    
762    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
763        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
764        the "old" RFC 2279). Instead, it ended up passing a negative length to
765        pcre_exec().
766    
767    31. Add support for GCC's visibility feature to hide internal functions.
768    
769    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
770        "unknown -C option" after outputting 0 or 1.
771    
772    33. There is now support for generating a code coverage report for the test
773        suite in environments where gcc is the compiler and lcov is installed. This
774        is mainly for the benefit of the developers.
775    
776    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
777        unaddressable using valgrind annotations, allowing valgrind to detect
778        invalid memory accesses. This is mainly for the benefit of the developers.
779    
780    25. (*UTF) can now be used to start a pattern in any of the three libraries.
781    
782    26. Give configure error if --enable-cpp but no C++ compiler found.
783    
784    
785    Version 8.31 06-July-2012
786    -------------------------
787    
788    1.  Fixing a wrong JIT test case and some compiler warnings.
789    
790    2.  Removed a bashism from the RunTest script.
791    
792    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
793        to unsigned type, result still unsigned" that was given by an MS compiler
794        on encountering the code "-sizeof(xxx)".
795    
796    4.  Partial matching support is added to the JIT compiler.
797    
798    5.  Fixed several bugs concerned with partial matching of items that consist
799        of more than one character:
800    
801        (a) /^(..)\1/ did not partially match "aba" because checking references was
802            done on an "all or nothing" basis. This also applied to repeated
803            references.
804    
805        (b) \R did not give a hard partial match if \r was found at the end of the
806            subject.
807    
808        (c) \X did not give a hard partial match after matching one or more
809            characters at the end of the subject.
810    
811        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
812            a partial match for the string "\r".
813    
814        (e) When newline was set to CRLF, the metacharacter "." did not recognize
815            a partial match for a CR character at the end of the subject string.
816    
817    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
818        running pcretest, the text "(JIT)" added to the output whenever JIT is
819        actually used to run the match.
820    
821    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
822        or /S+[+] with a digit between 1 and 7.
823    
824    8.  OP_NOT now supports any UTF character not just single-byte ones.
825    
826    9.  (*MARK) control verb is now supported by the JIT compiler.
827    
828    10. The command "./RunTest list" lists the available tests without actually
829        running any of them. (Because I keep forgetting what they all are.)
830    
831    11. Add PCRE_INFO_MAXLOOKBEHIND.
832    
833    12. Applied a (slightly modified) user-supplied patch that improves performance
834        when the heap is used for recursion (compiled with --disable-stack-for-
835        recursion). Instead of malloc and free for each heap frame each time a
836        logical recursion happens, frames are retained on a chain and re-used where
837        possible. This sometimes gives as much as 30% improvement.
838    
839    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
840        call.
841    
842    14. As documented, (*COMMIT) is now confined to within a positive assertion.
843    
844    15. It is now possible to link pcretest with libedit as an alternative to
845        libreadline.
846    
847    16. (*COMMIT) control verb is now supported by the JIT compiler.
848    
849    17. The Unicode data tables have been updated to Unicode 6.1.0.
850    
851    18. Added --file-list option to pcregrep.
852    
853    19. Added binary file support to pcregrep, including the -a, --binary-files,
854        -I, and --text options.
855    
856    20. The madvise function is renamed for posix_madvise for QNX compatibility
857        reasons. Fixed by Giuseppe D'Angelo.
858    
859    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
860    
861    22. Changed the option for creating symbolic links for 16-bit man pages from
862        -s to -sf so that re-installing does not cause issues.
863    
864    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
865    
866    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
867        matches in certain environments (the workspace was not being correctly
868        retained). Also added to pcre_dfa_exec() a simple plausibility check on
869        some of the workspace data at the beginning of a restart.
870    
871    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
872        was not doing so when it should - probably a typo introduced by SVN 528
873        (change 8.10/14).
874    
875    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
876        \w+ when the character tables indicated that \x{c4} was a word character.
877        There were several related cases, all because the tests for doing a table
878        lookup were testing for characters less than 127 instead of 255.
879    
880    27. If a pattern contains capturing parentheses that are not used in a match,
881        their slots in the ovector are set to -1. For those that are higher than
882        any matched groups, this happens at the end of processing. In the case when
883        there were back references that the ovector was too small to contain
884        (causing temporary malloc'd memory to be used during matching), and the
885        highest capturing number was not used, memory off the end of the ovector
886        was incorrectly being set to -1. (It was using the size of the temporary
887        memory instead of the true size.)
888    
889    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
890        ovector size, it uses memory at the end of the block that it has got.
891    
892    29. Check for an overlong MARK name and give an error at compile time. The
893        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
894    
895    30. JIT compiler update.
896    
897    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
898        Rill for the patch.
899    
900    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
901    
902    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
903    
904    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
905        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
906        was enabled.
907    
908    35. Improve JIT code generation for greedy plus quantifier.
909    
910    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
911        1 to "aa" instead of to an empty string. The bug affected repeated groups
912        that could potentially match an empty string.
913    
914    37. Optimizing single character iterators in JIT.
915    
916    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
917        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
918        codepoints that are too big for the mode are faulted, and in a UTF mode,
919        disallowed codepoints are also faulted.
920    
921    39. If PCRE was compiled with UTF support, in three places in the DFA
922        matcher there was code that should only have been obeyed in UTF mode, but
923        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
924        processing when bytes with values greater than 127 were present. In 16-bit
925        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
926        both cases the values are those that cannot be the first data item in a UTF
927        character. The three items that might have provoked this were recursions,
928        possessively repeated groups, and atomic groups.
929    
930    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
931        and pcregrep, because some OS require shared objects to be explicitly
932        passed to ld, causing the link step to fail if they are not.
933    
934    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
935        mode, patterns that started with \h* or \R* might be incorrectly matched.
936    
937    
938    Version 8.30 04-February-2012
939    -----------------------------
940    
941    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
942        name is defined in ctype.h.
943    
944    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
945        only in quite long subpatterns.
946    
947    3.  Removed the function pcre_info(), which has been obsolete and deprecated
948        since it was replaced by pcre_fullinfo() in February 2000.
949    
950    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
951        match a (*MARK), and the match failed at the start of the subject, a
952        reference to memory before the start of the subject could occur. This bug
953        was introduced by fix 17 of release 8.21.
954    
955    5.  A reference to an unset group with zero minimum repetition was giving
956        totally wrong answers (in non-JavaScript-compatibility mode). For example,
957        /(another)?(\1?)test/ matched against "hello world test". This bug was
958        introduced in release 8.13.
959    
960    6.  Add support for 16-bit character strings (a large amount of work involving
961        many changes and refactorings).
962    
963    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
964        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
965        from a file.
966    
967    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
968        rounding is not applied in this particular case).
969    
970    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
971        if they appear, or are escaped, in patterns.
972    
973    10. Get rid of a number of -Wunused-but-set-variable warnings.
974    
975    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
976        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
977        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
978        also returns the mark "x". This bug applied to capturing parentheses,
979        non-capturing parentheses, and atomic parentheses. It also applied to some
980        assertions.
981    
982    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
983        information out of configure.ac instead of relying on pcre.h.generic, which
984        is not stored in the repository.
985    
986    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
987        -lreadline.
988    
989    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
990    
991    15. Applied Graycode's patch to put the top-level frame on the stack rather
992        than the heap when not using the stack for recursion. This gives a
993        performance improvement in many cases when recursion is not deep.
994    
995    16. Experimental code added to "pcretest -C" to output the stack frame size.
996    
997    
998    Version 8.21 12-Dec-2011
999    ------------------------
1000    
1001    1.  Updating the JIT compiler.
1002    
1003    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1004        are added as well.
1005    
1006    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1007        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1008        calling _pcre_jit_exec. Some extra comments are added.
1009    
1010    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1011        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1012        was introduced by change 18 for 8.20.
1013    
1014    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1015        ECMA-262 standard.
1016    
1017    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1018        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1019        This bug was probably introduced by change 9 of 8.13.
1020    
1021    7.  While fixing 6 above, I noticed that a number of other items were being
1022        incorrectly rejected as "not fixed length". This arose partly because newer
1023        opcodes had not been added to the fixed-length checking code. I have (a)
1024        corrected the bug and added tests for these items, and (b) arranged for an
1025        error to occur if an unknown opcode is encountered while checking for fixed
1026        length instead of just assuming "not fixed length". The items that were
1027        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1028        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1029        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1030    
1031    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1032        being incorrectly compiled and would have given unpredicatble results.
1033    
1034    9.  A possessively repeated subpattern with minimum repeat count greater than
1035        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1036        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1037        the first (A) could occur when it should not.
1038    
1039    10. Add a cast and remove a redundant test from the code.
1040    
1041    11. JIT should use pcre_malloc/pcre_free for allocation.
1042    
1043    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1044        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1045        is anything other than /usr, -L is still shown).
1046    
1047    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1048    
1049    14. Perl does not support \N without a following name in a [] class; PCRE now
1050        also gives an error.
1051    
1052    15. If a forward reference was repeated with an upper limit of around 2000,
1053        it caused the error "internal error: overran compiling workspace". The
1054        maximum number of forward references (including repeats) was limited by the
1055        internal workspace, and dependent on the LINK_SIZE. The code has been
1056        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1057        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1058        of around 200,000 forward references. While doing this, I also speeded up
1059        the filling in of repeated forward references.
1060    
1061    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1062        incorrectly expecting the subject to contain another "a" after the start.
1063    
1064    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1065        in the match, the SKIP should be ignored. This was not happening; instead
1066        the SKIP was being treated as NOMATCH. For patterns such as
1067        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1068        tested.
1069    
1070    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1071        now much more compatible with Perl, in particular in cases where the result
1072        is a non-match for a non-anchored pattern. For example, if
1073        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1074        "m", where previously it did not return a name. A side effect of this
1075        change is that for partial matches, the last encountered mark name is
1076        returned, as for non matches. A number of tests that were previously not
1077        Perl-compatible have been moved into the Perl-compatible test files. The
1078        refactoring has had the pleasing side effect of removing one argument from
1079        the match() function, thus reducing its stack requirements.
1080    
1081    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1082        subsequent uses of /S (without +) incorrectly behaved like /S+.
1083    
1084    21. Retrieve executable code size support for the JIT compiler and fixing
1085        some warnings.
1086    
1087    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1088        not work when the shorter character appeared right at the end of the
1089        subject string.
1090    
1091    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1092        systems.
1093    
1094    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1095        output it when the /M option is used in pcretest.
1096    
1097    25. The CheckMan script was not being included in the distribution. Also, added
1098        an explicit "perl" to run Perl scripts from the PrepareRelease script
1099        because this is reportedly needed in Windows.
1100    
1101    26. If study data was being save in a file and studying had not found a set of
1102        "starts with" bytes for the pattern, the data written to the file (though
1103        never used) was taken from uninitialized memory and so caused valgrind to
1104        complain.
1105    
1106    27. Updated RunTest.bat as provided by Sheri Pierce.
1107    
1108    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1109    
1110    29. Computation of memory usage for the table of capturing group names was
1111        giving an unnecessarily large value.
1112    
1113    
1114    Version 8.20 21-Oct-2011
1115    ------------------------
1116    
1117    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1118        a POSIX class. After further experiments with Perl, which convinced me that
1119        Perl has bugs and confusions, a closing square bracket is no longer allowed
1120        in a POSIX name. This bug also affected patterns with classes that started
1121        with full stops.
1122    
1123    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1124        captured substring, but while checking the failing first alternative,
1125        substring 1 is temporarily captured. If the output vector supplied to
1126        pcre_exec() was not big enough for this capture, the yield of the function
1127        was still zero ("insufficient space for captured substrings"). This cannot
1128        be totally fixed without adding another stack variable, which seems a lot
1129        of expense for a edge case. However, I have improved the situation in cases
1130        such as /(a)(b)x|abc/ matched against "abc", where the return code
1131        indicates that fewer than the maximum number of slots in the ovector have
1132        been set.
1133    
1134    3.  Related to (2) above: when there are more back references in a pattern than
1135        slots in the output vector, pcre_exec() uses temporary memory during
1136        matching, and copies in the captures as far as possible afterwards. It was
1137        using the entire output vector, but this conflicts with the specification
1138        that only 2/3 is used for passing back captured substrings. Now it uses
1139        only the first 2/3, for compatibility. This is, of course, another edge
1140        case.
1141    
1142    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1143        main code base, and can be used by building with --enable-jit. When this is
1144        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1145        runtime --no-jit option is given.
1146    
1147    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1148        ovector, the return from the function was zero, implying that there were
1149        other matches that did not fit. The correct "exactly full" value is now
1150        returned.
1151    
1152    6.  If a subpattern that was called recursively or as a subroutine contained
1153        (*PRUNE) or any other control that caused it to give a non-standard return,
1154        invalid errors such as "Error -26 (nested recursion at the same subject
1155        position)" or even infinite loops could occur.
1156    
1157    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1158        computing the minimum length on reaching *ACCEPT, and so ended up with the
1159        wrong value of 1 rather than 0. Further investigation indicates that
1160        computing a minimum subject length in the presence of *ACCEPT is difficult
1161        (think back references, subroutine calls), and so I have changed the code
1162        so that no minimum is registered for a pattern that contains *ACCEPT.
1163    
1164    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1165        it was not handled as intended. [But see 16 below.]
1166    
1167    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1168        Sheri Pierce.
1169    
1170    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1171        the first byte in a match must be "a".
1172    
1173    11. Change 17 for 8.13 increased the recursion depth for patterns like
1174        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1175        pattern contains any instances of (*THEN). If it does not, the old
1176        optimizations are restored. It would be nice to do this on a per-group
1177        basis, but at the moment that is not feasible.
1178    
1179    12. In some environments, the output of pcretest -C is CRLF terminated. This
1180        broke RunTest's code that checks for the link size. A single white space
1181        character after the value is now allowed for.
1182    
1183    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1184        For "fr", it uses the Windows-specific input and output files.
1185    
1186    14. If (*THEN) appeared in a group that was called recursively or as a
1187        subroutine, it did not work as intended. [But see next item.]
1188    
1189    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1190        pattern fragments (but not containing any | characters). If A and B are
1191        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1192        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1193        D. In other words, Perl considers parentheses that do not contain any |
1194        characters to be part of a surrounding alternative, whereas PCRE was
1195        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1196        differently. PCRE now behaves in the same way as Perl, except in the case
1197        of subroutine/recursion calls such as (?1) which have in any case always
1198        been different (but PCRE had them first :-).
1199    
1200    16. Related to 15 above: Perl does not treat the | in a conditional group as
1201        creating alternatives. Such a group is treated in the same way as an
1202        ordinary group without any | characters when processing (*THEN). PCRE has
1203        been changed to match Perl's behaviour.
1204    
1205    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1206        RunGrepTest script failed.
1207    
1208    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1209        inevitable for groups that contain captures, but it can lead to a lot of
1210        stack use in large patterns. The old behaviour has been restored for atomic
1211        groups that do not contain any capturing parentheses.
1212    
1213    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1214        suppress the check for a minimum subject length at run time. (If it was
1215        given to pcre_exec() or pcre_dfa_exec() it did work.)
1216    
1217    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1218        fail to work when decoding hex characters in data strings in EBCDIC
1219        environments.
1220    
1221    21. It appears that in at least one Mac OS environment, the isxdigit() function
1222        is implemented as a macro that evaluates to its argument more than once,
1223        contravening the C 90 Standard (I haven't checked a later standard). There
1224        was an instance in pcretest which caused it to go wrong when processing
1225        \x{...} escapes in subject strings. The has been rewritten to avoid using
1226        things like p++ in the argument of isxdigit().
1227    
1228    
1229    Version 8.13 16-Aug-2011
1230  ------------------------  ------------------------
1231    
1232  1.  The Unicode data tables have been updated to Unicode 6.0.0.  1.  The Unicode data tables have been updated to Unicode 6.0.0.
# Line 20  Version 8.13 30-Apr-2011 Line 1245  Version 8.13 30-Apr-2011
1245      code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a      code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1246      2-byte code at the end of the subject (it thought there wasn't enough data      2-byte code at the end of the subject (it thought there wasn't enough data
1247      left).      left).
1248    
1249  5.  Comprehensive information about what went wrong is now returned by  5.  Comprehensive information about what went wrong is now returned by
1250      pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long      pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1251      as the output vector has at least 2 elements. The offset of the start of      as the output vector has at least 2 elements. The offset of the start of
1252      the failing character and a reason code are placed in the vector.      the failing character and a reason code are placed in the vector.
1253    
1254  6.  When the UTF-8 string check fails for pcre_compile(), the offset that is  6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1255      now returned is for the first byte of the failing character, instead of the      now returned is for the first byte of the failing character, instead of the
1256      last byte inspected. This is an incompatible change, but I hope it is small      last byte inspected. This is an incompatible change, but I hope it is small
1257      enough not to be a problem. It makes the returned offset consistent with      enough not to be a problem. It makes the returned offset consistent with
1258      pcre_exec() and pcre_dfa_exec().      pcre_exec() and pcre_dfa_exec().
1259    
1260  7.  pcretest now gives a text phrase as well as the error number when  7.  pcretest now gives a text phrase as well as the error number when
1261      pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check      pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1262      failure, the offset and reason code are output.      failure, the offset and reason code are output.
1263    
1264  8.  When \R was used with a maximizing quantifier it failed to skip backwards  8.  When \R was used with a maximizing quantifier it failed to skip backwards
1265      over a \r\n pair if the subsequent match failed. Instead, it just skipped      over a \r\n pair if the subsequent match failed. Instead, it just skipped
1266      back over a single character (\n). This seems wrong (because it treated the      back over a single character (\n). This seems wrong (because it treated the
1267      two characters as a single entity when going forwards), conflicts with the      two characters as a single entity when going forwards), conflicts with the
1268      documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the      documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1269      behaviour of \R* different to (\R)*, which also seems wrong. The behaviour      behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1270      has been changed.      has been changed.
1271    
1272    9.  Some internal refactoring has changed the processing so that the handling
1273        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1274        time (the PCRE_DOTALL option was changed this way some time ago: version
1275        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1276        which was always a bit of a fudge. It also means that there is one less
1277        argument for the match() function, which reduces its stack requirements
1278        slightly. This change also fixes an incompatibility with Perl: the pattern
1279        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1280    
1281    10. More internal refactoring has drastically reduced the number of recursive
1282        calls to match() for possessively repeated groups such as (abc)++ when
1283        using pcre_exec().
1284    
1285    11. While implementing 10, a number of bugs in the handling of groups were
1286        discovered and fixed:
1287    
1288        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1289        (a|)*(?1) gave a compile-time internal error.
1290        ((a|)+)+  did not notice that the outer group could match an empty string.
1291        (^a|^)+   was not marked as anchored.
1292        (.*a|.*)+ was not marked as matching at start or after a newline.
1293    
1294    12. Yet more internal refactoring has removed another argument from the match()
1295        function. Special calls to this function are now indicated by setting a
1296        value in a variable in the "match data" data block.
1297    
1298    13. Be more explicit in pcre_study() instead of relying on "default" for
1299        opcodes that mean there is no starting character; this means that when new
1300        ones are added and accidentally left out of pcre_study(), testing should
1301        pick them up.
1302    
1303    14. The -s option of pcretest has been documented for ages as being an old
1304        synonym of -m (show memory usage). I have changed it to mean "force study
1305        for every regex", that is, assume /S for every regex. This is similar to -i
1306        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1307        using it. It makes it easier to run collections of tests with and without
1308        study enabled, and thereby test pcre_study() more easily. All the standard
1309        tests are now run with and without -s (but some patterns can be marked as
1310        "never study" - see 20 below).
1311    
1312    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1313        restoration of the capturing data to the outer values was not happening
1314        correctly.
1315    
1316    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1317        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1318        pattern had matched an empty string, and so incorrectly returned a no
1319        match.
1320    
1321    17. There was optimizing code for the last branch of non-capturing parentheses,
1322        and also for the obeyed branch of a conditional subexpression, which used
1323        tail recursion to cut down on stack usage. Unfortunately, now that there is
1324        the possibility of (*THEN) occurring in these branches, tail recursion is
1325        no longer possible because the return has to be checked for (*THEN). These
1326        two optimizations have therefore been removed. [But see 8.20/11 above.]
1327    
1328    18. If a pattern containing \R was studied, it was assumed that \R always
1329        matched two bytes, thus causing the minimum subject length to be
1330        incorrectly computed because \R can also match just one byte.
1331    
1332    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1333        was incorrectly computed.
1334    
1335    20. If /S is present twice on a test pattern in pcretest input, it now
1336        *disables* studying, thereby overriding the use of -s on the command line
1337        (see 14 above). This is necessary for one or two tests to keep the output
1338        identical in both cases.
1339    
1340    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1341        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1342    
1343    22. When an atomic group that contained a capturing parenthesis was
1344        successfully matched, but the branch in which it appeared failed, the
1345        capturing was not being forgotten if a higher numbered group was later
1346        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1347        group 1 to "a", when in fact it should be unset. This applied to multi-
1348        branched capturing and non-capturing groups, repeated or not, and also to
1349        positive assertions (capturing in negative assertions does not happen
1350        in PCRE) and also to nested atomic groups.
1351    
1352    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1353        subject after a captured substring, to make it easier to tell which of a
1354        number of identical substrings has been captured.
1355    
1356    24. The way atomic groups are processed by pcre_exec() has been changed so that
1357        if they are repeated, backtracking one repetition now resets captured
1358        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1359        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1360        "aaa". Previously, it would have been "a". As part of this code
1361        refactoring, the way recursive calls are handled has also been changed.
1362    
1363    25. If an assertion condition captured any substrings, they were not passed
1364        back unless some other capturing happened later. For example, if
1365        (?(?=(a))a) was matched against "a", no capturing was returned.
1366    
1367    26. When studying a pattern that contained subroutine calls or assertions,
1368        the code for finding the minimum length of a possible match was handling
1369        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1370        group 1 called group 2 while simultaneously a separate group 2 called group
1371        1). A stack overflow occurred in this case. I have fixed this by limiting
1372        the recursion depth to 10.
1373    
1374    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1375        Fortmann. This supports explicit test numbers on the command line, and has
1376        argument validation and error reporting.
1377    
1378    28. An instance of \X with an unlimited repeat could fail if at any point the
1379        first character it looked at was a mark character.
1380    
1381    29. Some minor code refactoring concerning Unicode properties and scripts
1382        should reduce the stack requirement of match() slightly.
1383    
1384    30. Added the '=' option to pcretest to check the setting of unused capturing
1385        slots at the end of the pattern, which are documented as being -1, but are
1386        not included in the return count.
1387    
1388    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1389        compiled something random. Now it gives a compile-time error (as does
1390        Perl).
1391    
1392    32. A *MARK encountered during the processing of a positive assertion is now
1393        recorded and passed back (compatible with Perl).
1394    
1395    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1396        had alternative anchored branches, the search for a second match in a line
1397        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1398        matched the line "0102" twice. The same bug affected patterns that started
1399        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1400        twice.
1401    
1402    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1403        does, and because of capturing effects, quantifying parenthesized
1404        assertions may at times be useful. Quantifiers are now allowed for
1405        parenthesized assertions.
1406    
1407    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1408    
1409    36. \g was being checked for fancy things in a character class, when it should
1410        just be a literal "g".
1411    
1412    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1413        appearance of a nested POSIX class supersedes an apparent external class.
1414        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1415        unescaped square brackets may also appear as part of class names. For
1416        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1417        more like Perl. (But see 8.20/1 above.)
1418    
1419    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1420        was because it thought it was \N{name}, which is not supported).
1421    
1422    39. Add minix to OS list not supporting the -S option in pcretest.
1423    
1424    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1425        cannot analyze patterns in sufficient detail to catch mutual recursions
1426        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1427        subgroup is called recursively as a subpattern for a second time at the
1428        same position in the subject string. In previous releases this might have
1429        been caught by the recursion limit, or it might have run out of stack.
1430    
1431    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1432        happen only once. PCRE was, however incorrectly giving a compile time error
1433        "recursive call could loop indefinitely" because it cannot analyze the
1434        pattern in sufficient detail. The compile time test no longer happens when
1435        PCRE is compiling a conditional subpattern, but actual runaway loops are
1436        now caught at runtime (see 40 above).
1437    
1438    42. It seems that Perl allows any characters other than a closing parenthesis
1439        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1440        has been changed to be the same.
1441    
1442    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1443        as not to get warnings when autogen.sh is called. Also changed
1444        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1445    
1446    44. To help people who use pcregrep to scan files containing exceedingly long
1447        lines, the following changes have been made:
1448    
1449        (a) The default value of the buffer size parameter has been increased from
1450            8K to 20K. (The actual buffer used is three times this size.)
1451    
1452        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1453            PCRE is built.
1454    
1455        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1456            to be set at run time.
1457    
1458        (d) Numerical values in pcregrep options can be followed by K or M, for
1459            example --buffer-size=50K.
1460    
1461        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1462            given and the return code is set to 2.
1463    
1464    45. Add a pointer to the latest mark to the callout data block.
1465    
1466    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1467        partial match of an empty string instead of no match. This was specific to
1468        the use of ".".
1469    
1470    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1471        complete match instead of a partial match. This bug was dependent on both
1472        the PCRE_UTF8 and PCRE_DOTALL options being set.
1473    
1474    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1475        starting byte set, because \b was not being ignored.
1476    
1477    
1478  Version 8.12 15-Jan-2011  Version 8.12 15-Jan-2011
1479  ------------------------  ------------------------
# Line 695  Version 7.9 11-Apr-09 Line 2125  Version 7.9 11-Apr-09
2125  7.  A pattern that could match an empty string could cause pcregrep to loop; it  7.  A pattern that could match an empty string could cause pcregrep to loop; it
2126      doesn't make sense to accept an empty string match in pcregrep, so I have      doesn't make sense to accept an empty string match in pcregrep, so I have
2127      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2128      seems to be how GNU grep behaves.      seems to be how GNU grep behaves. [But see later change 40 for release
2129        8.33.]
2130    
2131  8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at  8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2132      start or after a newline", because the conditional assertion was not being      start or after a newline", because the conditional assertion was not being
# Line 938  Version 7.7 07-May-08 Line 2369  Version 7.7 07-May-08
2369      containing () gave an internal compiling error instead of "reference to      containing () gave an internal compiling error instead of "reference to
2370      non-existent subpattern". Fortunately, when the pattern did exist, the      non-existent subpattern". Fortunately, when the pattern did exist, the
2371      compiled code was correct. (When scanning forwards to check for the      compiled code was correct. (When scanning forwards to check for the
2372      existencd of the subpattern, it was treating the data ']' as terminating      existence of the subpattern, it was treating the data ']' as terminating
2373      the class, so got the count wrong. When actually compiling, the reference      the class, so got the count wrong. When actually compiling, the reference
2374      was subsequently set up correctly.)      was subsequently set up correctly.)
2375    

Legend:
Removed from v.600  
changed lines
  Added in v.1517

  ViewVC Help
Powered by ViewVC 1.1.5