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

Diff of /code/trunk/ChangeLog

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

revision 167 by ph10, Wed May 9 15:53:54 2007 UTC revision 1515 by ph10, Tue Feb 3 17:08:54 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-07  Version 8.37 xx-xxx-201x
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    
24    
25    Version 8.36 26-September-2014
26    ------------------------------
27    
28    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
29        -Wmissing-field-initializers and -Wunused-parameter.
30    
31    2.  The tests for quantifiers being too big (greater than 65535) were being
32        applied after reading the number, and stupidly assuming that integer
33        overflow would give a negative number. The tests are now applied as the
34        numbers are read.
35    
36    3.  Tidy code in pcre_exec.c where two branches that used to be different are
37        now the same.
38    
39    4.  The JIT compiler did not generate match limit checks for certain
40        bracketed expressions with quantifiers. This may lead to exponential
41        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
42        issue should be resolved now.
43    
44    5.  Fixed an issue, which occures when nested alternatives are optimized
45        with table jumps.
46    
47    6.  Inserted two casts and changed some ints to size_t in the light of some
48        reported 64-bit compiler warnings (Bugzilla 1477).
49    
50    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
51        an empty string, which sometimes were behaving incorrectly in the
52        interpreter (though correctly in the JIT matcher). This pcretest input is
53        an example:
54    
55          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
56          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
57    
58        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
59        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
60        for an empty string was breaking the inner loop and carrying on at a lower
61        level, when possessive repeated groups should always return to a higher
62        level as they have no backtrack points in them. The empty string test now
63        occurs at the outer level.
64    
65    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
66        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
67    
68    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
69        doesn't).
70    
71    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
72        to hold a pattern to be incorrectly computed (too small) when there were
73        named back references to duplicated names. This could cause "internal
74        error: code overflow" or "double free or corruption" or other memory
75        handling errors.
76    
77    11. When named subpatterns had the same prefixes, back references could be
78        confused. For example, in this pattern:
79    
80          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
81    
82        the reference to 'Name' was incorrectly treated as a reference to a
83        duplicate name.
84    
85    12. A pattern such as /^s?c/mi8 where the optional character has more than
86        one "other case" was incorrectly compiled such that it would only try to
87        match starting at "c".
88    
89    13. When a pattern starting with \s was studied, VT was not included in the
90        list of possible starting characters; this should have been part of the
91        8.34/18 patch.
92    
93    14. If a character class started [\Qx]... where x is any character, the class
94        was incorrectly terminated at the ].
95    
96    15. If a pattern that started with a caseless match for a character with more
97        than one "other case" was studied, PCRE did not set up the starting code
98        unit bit map for the list of possible characters. Now it does. This is an
99        optimization improvement, not a bug fix.
100    
101    16. The Unicode data tables have been updated to Unicode 7.0.0.
102    
103    17. Fixed a number of memory leaks in pcregrep.
104    
105    18. Avoid a compiler warning (from some compilers) for a function call with
106        a cast that removes "const" from an lvalue by using an intermediate
107        variable (to which the compiler does not object).
108    
109    19. Incorrect code was compiled if a group that contained an internal recursive
110        back reference was optional (had quantifier with a minimum of zero). This
111        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
112        caused segmentation faults because of stack overflows at compile time.
113    
114    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
115        group that is quantified with an indefinite repeat, caused a compile-time
116        loop which used up all the system stack and provoked a segmentation fault.
117        This was not the same bug as 19 above.
118    
119    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
120        Patch by Mike Frysinger.
121    
122    
123    Version 8.35 04-April-2014
124    --------------------------
125    
126    1.  A new flag is set, when property checks are present in an XCLASS.
127        When this flag is not set, PCRE can perform certain optimizations
128        such as studying these XCLASS-es.
129    
130    2.  The auto-possessification of character sets were improved: a normal
131        and an extended character set can be compared now. Furthermore
132        the JIT compiler optimizes more character set checks.
133    
134    3.  Got rid of some compiler warnings for potentially uninitialized variables
135        that show up only when compiled with -O2.
136    
137    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
138        of a match later then the end of the match. The pcretest program was not
139        handling the case sensibly - it was outputting from the start to the next
140        binary zero. It now reports this situation in a message, and outputs the
141        text from the end to the start.
142    
143    5.  Fast forward search is improved in JIT. Instead of the first three
144        characters, any three characters with fixed position can be searched.
145        Search order: first, last, middle.
146    
147    6.  Improve character range checks in JIT. Characters are read by an inprecise
148        function now, which returns with an unknown value if the character code is
149        above a certain threshold (e.g: 256). The only limitation is that the value
150        must be bigger than the threshold as well. This function is useful when
151        the characters above the threshold are handled in the same way.
152    
153    7.  The macros whose names start with RAWUCHAR are placeholders for a future
154        mode in which only the bottom 21 bits of 32-bit data items are used. To
155        make this more memorable for those maintaining the code, the names have
156        been changed to start with UCHAR21, and an extensive comment has been added
157        to their definition.
158    
159    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
160        to the export list in Makefile.am (they were accidentally omitted from the
161        8.34 tarball).
162    
163    9.  The informational output from pcretest used the phrase "starting byte set"
164        which is inappropriate for the 16-bit and 32-bit libraries. As the output
165        for "first char" and "need char" really means "non-UTF-char", I've changed
166        "byte" to "char", and slightly reworded the output. The documentation about
167        these values has also been (I hope) clarified.
168    
169    10. Another JIT related optimization: use table jumps for selecting the correct
170        backtracking path, when more than four alternatives are present inside a
171        bracket.
172    
173    11. Empty match is not possible, when the minimum length is greater than zero,
174        and there is no \K in the pattern. JIT should avoid empty match checks in
175        such cases.
176    
177    12. In a caseless character class with UCP support, when a character with more
178        than one alternative case was not the first character of a range, not all
179        the alternative cases were added to the class. For example, s and \x{17f}
180        are both alternative cases for S: the class [RST] was handled correctly,
181        but [R-T] was not.
182    
183    13. The configure.ac file always checked for pthread support when JIT was
184        enabled. This is not used in Windows, so I have put this test inside a
185        check for the presence of windows.h (which was already tested for).
186    
187    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
188        The algorithm provides a way to skip certain starting offsets, and usually
189        faster than linear prefix searches.
190    
191    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
192        as for 'fr_FR' and 'french'. For some reason, however, it then used the
193        Windows-specific input and output files, which have 'french' screwed in.
194        So this could never have worked. One of the problems with locales is that
195        they aren't always the same. I have now updated RunTest so that it checks
196        the output of the locale test (test 3) against three different output
197        files, and it allows the test to pass if any one of them matches. With luck
198        this should make the test pass on some versions of Solaris where it was
199        failing. Because of the uncertainty, the script did not used to stop if
200        test 3 failed; it now does. If further versions of a French locale ever
201        come to light, they can now easily be added.
202    
203    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
204        there was a message during ./configure, but it did not stop. This now
205        provokes an error. The invalid example in README has been corrected.
206        If a value less than the minimum is given, the minimum value has always
207        been used, but now a warning is given.
208    
209    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
210        was a bug in the test system, which is now fixed. Also, the list of various
211        configurations that are tested for each release did not have one with both
212        16/32 bits and --enable-bar-anycrlf. It now does.
213    
214    18. pcretest was missing "-C bsr" for displaying the \R default setting.
215    
216    19. Little endian PowerPC systems are supported now by the JIT compiler.
217    
218    20. The fast forward newline mechanism could enter to an infinite loop on
219        certain invalid UTF-8 input. Although we don't support these cases
220        this issue can be fixed by a performance optimization.
221    
222    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
223        not take account if existing stack usage. There is now a new global
224        variable called pcre_stack_guard that can be set to point to an external
225        function to check stack availability. It is called at the start of
226        processing every parenthesized group.
227    
228    22. A typo in the code meant that in ungreedy mode the max/min qualifier
229        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
230        not match "ab".
231    
232    23. When UTF was disabled, the JIT program reported some incorrect compile
233        errors. These messages are silenced now.
234    
235    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
236        compiler.
237    
238    25. Change all the temporary files used in RunGrepTest to be different to those
239        used by RunTest so that the tests can be run simultaneously, for example by
240        "make -j check".
241    
242    
243    Version 8.34 15-December-2013
244    -----------------------------
245    
246    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
247        executable memory. Patch inspired by Carsten Klein.
248    
249    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
250        this macro is never tested and has no effect, because the work to support
251        coverage involves only compiling and linking options and special targets in
252        the Makefile. The comment in config.h implied that defining the macro would
253        enable coverage support, which is totally false. There was also support for
254        setting this macro in the CMake files (my fault, I just copied it from
255        configure). SUPPORT_GCOV has now been removed.
256    
257    3.  Make a small performance improvement in strlen16() and strlen32() in
258        pcretest.
259    
260    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
261        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
262        these cases). There was also one in pcretest.c.
263    
264    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
265    
266    6.  In UTF mode, the code for checking whether a group could match an empty
267        string (which is used for indefinitely repeated groups to allow for
268        breaking an infinite loop) was broken when the group contained a repeated
269        negated single-character class with a character that occupied more than one
270        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
271        UTF-8 mode). The effect was undefined: the group might or might not be
272        deemed as matching an empty string, or the program might have crashed.
273    
274    7.  The code for checking whether a group could match an empty string was not
275        recognizing that \h, \H, \v, \V, and \R must match a character.
276    
277    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
278        an empty string. If it can, pcretest shows this in its information output.
279    
280    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
281        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
282        matched by pcre_exec() without using JIT:
283    
284        (a) If the rest of the pattern did not match after a maximal run of
285            grapheme clusters, the code for backing up to try with fewer of them
286            did not always back up over a full grapheme when characters that do not
287            have the modifier quality were involved, e.g. Hangul syllables.
288    
289        (b) If the match point in a subject started with modifier character, and
290            there was no match, the code could incorrectly back up beyond the match
291            point, and potentially beyond the first character in the subject,
292            leading to a segfault or an incorrect match result.
293    
294    10. A conditional group with an assertion condition could lead to PCRE
295        recording an incorrect first data item for a match if no other first data
296        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
297        first data item, and therefore matched "ca" after "c" instead of at the
298        start.
299    
300    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
301        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
302    
303    12. The source of pcregrep now includes z/OS-specific code so that it can be
304        compiled for z/OS as part of the special z/OS distribution.
305    
306    13. Added the -T and -TM options to pcretest.
307    
308    14. The code in pcre_compile.c for creating the table of named capturing groups
309        has been refactored. Instead of creating the table dynamically during the
310        actual compiling pass, the information is remembered during the pre-compile
311        pass (on the stack unless there are more than 20 named groups, in which
312        case malloc() is used) and the whole table is created before the actual
313        compile happens. This has simplified the code (it is now nearly 150 lines
314        shorter) and prepared the way for better handling of references to groups
315        with duplicate names.
316    
317    15. A back reference to a named subpattern when there is more than one of the
318        same name now checks them in the order in which they appear in the pattern.
319        The first one that is set is used for the reference. Previously only the
320        first one was inspected. This change makes PCRE more compatible with Perl.
321    
322    16. Unicode character properties were updated from Unicode 6.3.0.
323    
324    17. The compile-time code for auto-possessification has been refactored, based
325        on a patch by Zoltan Herczeg. It now happens after instead of during
326        compilation. The code is cleaner, and more cases are handled. The option
327        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
328        options in pcretest are provided to set it. It can also be set by
329        (*NO_AUTO_POSSESS) at the start of a pattern.
330    
331    18. The character VT has been added to the default ("C" locale) set of
332        characters that match \s and are generally treated as white space,
333        following this same change in Perl 5.18. There is now no difference between
334        "Perl space" and "POSIX space". Whether VT is treated as white space in
335        other locales depends on the locale.
336    
337    19. The code for checking named groups as conditions, either for being set or
338        for being recursed, has been refactored (this is related to 14 and 15
339        above). Processing unduplicated named groups should now be as fast at
340        numerical groups, and processing duplicated groups should be faster than
341        before.
342    
343    20. Two patches to the CMake build system, by Alexander Barkov:
344    
345          (1) Replace the "source" command by "." in CMakeLists.txt because
346              "source" is a bash-ism.
347    
348          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
349              without these the CMake build does not work on Solaris.
350    
351    21. Perl has changed its handling of \8 and \9. If there is no previously
352        encountered capturing group of those numbers, they are treated as the
353        literal characters 8 and 9 instead of a binary zero followed by the
354        literals. PCRE now does the same.
355    
356    22. Following Perl, added \o{} to specify codepoints in octal, making it
357        possible to specify values greater than 0777 and also making them
358        unambiguous.
359    
360    23. Perl now gives an error for missing closing braces after \x{... instead of
361        treating the string as literal. PCRE now does the same.
362    
363    24. RunTest used to grumble if an inappropriate test was selected explicitly,
364        but just skip it when running all tests. This make it awkward to run ranges
365        of tests when one of them was inappropriate. Now it just skips any
366        inappropriate tests, as it always did when running all tests.
367    
368    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
369        character types such as \d or \w, too many callouts were inserted, and the
370        data that they returned was rubbish.
371    
372    26. In UCP mode, \s was not matching two of the characters that Perl matches,
373        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
374        were matched by \h. The code has now been refactored so that the lists of
375        the horizontal and vertical whitespace characters used for \h and \v (which
376        are defined only in one place) are now also used for \s.
377    
378    27. Add JIT support for the 64 bit TileGX architecture.
379        Patch by Jiong Wang (Tilera Corporation).
380    
381    28. Possessive quantifiers for classes (both explicit and automatically
382        generated) now use special opcodes instead of wrapping in ONCE brackets.
383    
384    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
385        (because it's meaningless), this was not happening when PCRE_CASELESS was
386        set. Not wrong, but inefficient.
387    
388    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
389        properties for \w, \d, etc) is present in a test regex. Otherwise if the
390        test contains no characters greater than 255, Perl doesn't realise it
391        should be using Unicode semantics.
392    
393    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
394        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
395        does in Unicode mode.
396    
397    32. Added the "forbid" facility to pcretest so that putting tests into the
398        wrong test files can sometimes be quickly detected.
399    
400    33. There is now a limit (default 250) on the depth of nesting of parentheses.
401        This limit is imposed to control the amount of system stack used at compile
402        time. It can be changed at build time by --with-parens-nest-limit=xxx or
403        the equivalent in CMake.
404    
405    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
406        errors. Perl warns for these when in warning mode, but PCRE has no facility
407        for giving warnings.
408    
409    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
410        However, this was not working for (?!) because it is optimized to (*FAIL),
411        for which PCRE does not allow quantifiers. The optimization is now disabled
412        when a quantifier follows (?!). I can't see any use for this, but it makes
413        things uniform.
414    
415    36. Perl no longer allows group names to start with digits, so I have made this
416        change also in PCRE. It simplifies the code a bit.
417    
418    37. In extended mode, Perl ignores spaces before a + that indicates a
419        possessive quantifier. PCRE allowed a space before the quantifier, but not
420        before the possessive +. It now does.
421    
422    38. The use of \K (reset reported match start) within a repeated possessive
423        group such as (a\Kb)*+ was not working.
424    
425    40. Document that the same character tables must be used at compile time and
426        run time, and that the facility to pass tables to pcre_exec() and
427        pcre_dfa_exec() is for use only with saved/restored patterns.
428    
429    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
430        features for Builds with MSVC:
431    
432        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
433           builds). The .rc files can be used to set FileDescription and many other
434           attributes.
435    
436        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
437           This allows higher-level build scripts which want .pdb files to avoid
438           hard-coding the exact files needed."
439    
440    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
441        mean "start of word" and "end of word", respectively, as a transition aid.
442    
443    43. A minimizing repeat of a class containing codepoints greater than 255 in
444        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
445        compiled to use the heap for recursion.
446    
447    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
448        is configured.
449    
450    
451    Version 8.33 28-May-2013
452    ------------------------
453    
454    1.  Added 'U' to some constants that are compared to unsigned integers, to
455        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
456        variables that are added to signed variables, to ensure the result is
457        signed and can be negated.
458    
459    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
460        CMake config files.
461    
462    3.  Revise the creation of config.h.generic so that all boolean macros are
463        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
464        overriding via -D on the command line possible.
465    
466    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
467        to unsigned int is reported to make a quite noticeable speed difference in
468        a specific Windows environment. Testing on Linux did also appear to show
469        some benefit (and it is clearly not harmful). Also fixed the definition of
470        Xop which should be unsigned.
471    
472    5.  Related to (4), changing the definition of the intermediate variable cc
473        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
474        improvements.
475    
476    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
477        unnecessary spaces.
478    
479    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
480        and later.
481    
482    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
483    
484    9.  Optimizing fast_forward_start_bits in JIT.
485    
486    10. Adding support for callouts in JIT, and fixing some issues revealed
487        during this work. Namely:
488    
489        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
490    
491        (b) Minimum length was not checked before the matching is started.
492    
493    11. The value of capture_last that is passed to callouts was incorrect in some
494        cases when there was a capture on one path that was subsequently abandoned
495        after a backtrack. Also, the capture_last value is now reset after a
496        recursion, since all captures are also reset in this case.
497    
498    12. The interpreter no longer returns the "too many substrings" error in the
499        case when an overflowing capture is in a branch that is subsequently
500        abandoned after a backtrack.
501    
502    13. In the pathological case when an offset vector of size 2 is used, pcretest
503        now prints out the matched string after a yield of 0 or 1.
504    
505    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
506        Only supported by the JIT compiler at the moment.
507    
508    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
509    
510    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
511        offset of the starting point of the matching process, provided the offsets
512        vector is large enough.
513    
514    17. The \A escape now records a lookbehind value of 1, though its execution
515        does not actually inspect the previous character. This is to ensure that,
516        in partial multi-segment matching, at least one character from the old
517        segment is retained when a new segment is processed. Otherwise, if there
518        are no lookbehinds in the pattern, \A might match incorrectly at the start
519        of a new segment.
520    
521    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
522    
523    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
524        gives some modest performance improvement in 8-bit mode.
525    
526    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
527        be expressed in certain programming languages using Universal Character
528        Names.
529    
530    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
531        which points out that "non characters" are not "characters that may not
532        appear in Unicode strings" but rather "characters that are reserved for
533        internal use and have only local meaning".
534    
535    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
536        there was a conditional group that depended on an assertion, if the
537        assertion was false, the callout that immediately followed the alternation
538        in the condition was skipped when pcre_exec() was used for matching.
539    
540    23. Allow an explicit callout to be inserted before an assertion that is the
541        condition for a conditional group, for compatibility with automatic
542        callouts, which always insert a callout at this point.
543    
544    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
545        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
546    
547    25. (*PRUNE) is now supported by the JIT compiler.
548    
549    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
550    
551    27. Fix the case where there are two or more SKIPs with arguments that may be
552        ignored.
553    
554    28. (*SKIP) is now supported by the JIT compiler.
555    
556    29. (*THEN) is now supported by the JIT compiler.
557    
558    30. Update RunTest with additional test selector options.
559    
560    31. The way PCRE handles backtracking verbs has been changed in two ways.
561    
562        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
563        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
564        backtracking. In some cases this makes it more Perl-compatible, but Perl's
565        rather obscure rules do not always do the same thing.
566    
567        (2) Previously, backtracking verbs were confined within assertions. This is
568        no longer the case for positive assertions, except for (*ACCEPT). Again,
569        this sometimes improves Perl compatibility, and sometimes does not.
570    
571    32. A number of tests that were in test 2 because Perl did things differently
572        have been moved to test 1, because either Perl or PCRE has changed, and
573        these tests are now compatible.
574    
575    32. Backtracking control verbs are now handled in the same way in JIT and
576        interpreter.
577    
578    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
579        contained a forward subroutine reference caused a compile error.
580    
581    34. Auto-detect and optimize limited repetitions in JIT.
582    
583    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
584        blocking (*UTF) etc.
585    
586    36. In the interpreter, maximizing pattern repetitions for characters and
587        character types now use tail recursion, which reduces stack usage.
588    
589    37. The value of the max lookbehind was not correctly preserved if a compiled
590        and saved regex was reloaded on a host of different endianness.
591    
592    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
593        of the compiled pattern block, expand the flags field from 16 to 32 bits
594        because it was almost full.
595    
596    39. Try madvise first before posix_madvise.
597    
598    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
599        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
600        original change locked out all matches of empty strings. This has been
601        changed so that one match of an empty string per line is recognized.
602        Subsequent searches on the same line (for colouring or for --only-matching,
603        for example) do not recognize empty strings.
604    
605    41. Applied a user patch to fix a number of spelling mistakes in comments.
606    
607    42. Data lines longer than 65536 caused pcretest to crash.
608    
609    43. Clarified the data type for length and startoffset arguments for pcre_exec
610        and pcre_dfa_exec in the function-specific man pages, where they were
611        explicitly stated to be in bytes, never having been updated. I also added
612        some clarification to the pcreapi man page.
613    
614    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
615        a segmentation fault.
616    
617    
618    Version 8.32 30-November-2012
619    -----------------------------
620    
621    1.  Improved JIT compiler optimizations for first character search and single
622        character iterators.
623    
624    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
625        Patch by Daniel Richard G.
626    
627    3.  Single character iterator optimizations in the JIT compiler.
628    
629    4.  Improved JIT compiler optimizations for character ranges.
630    
631    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
632        Reported by Giuseppe D'Angelo.
633    
634    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
635        of a line, was being set incorrectly in cases where .* appeared inside
636        atomic brackets at the start of a pattern, or where there was a subsequent
637        *PRUNE or *SKIP.
638    
639    7.  Improved instruction cache flush for POWER/PowerPC.
640        Patch by Daniel Richard G.
641    
642    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
643        grep:
644    
645        (a) There is now no limit to the number of patterns to be matched.
646    
647        (b) An error is given if a pattern is too long.
648    
649        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
650            are now supported.
651    
652        (d) --exclude-from and --include-from (multiple use) have been added.
653    
654        (e) Exclusions and inclusions now apply to all files and directories, not
655            just to those obtained from scanning a directory recursively.
656    
657        (f) Multiple uses of -f and --file-list are now supported.
658    
659        (g) In a Windows environment, the default for -d has been changed from
660            "read" (the GNU grep default) to "skip", because otherwise the presence
661            of a directory in the file list provokes an error.
662    
663        (h) The documentation has been revised and clarified in places.
664    
665    9.  Improve the matching speed of capturing brackets.
666    
667    10. Changed the meaning of \X so that it now matches a Unicode extended
668        grapheme cluster.
669    
670    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
671        out POSIX threads when JIT support is configured.
672    
673    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
674    
675    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
676        structure could go wrong in environments where size_t is not the same size
677        as int.
678    
679    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
680        set.
681    
682    15. The EBCDIC support had decayed; later updates to the code had included
683        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
684        general tidy up of EBCDIC-related issues, and the documentation was also
685        not quite right. There is now a test that can be run on ASCII systems to
686        check some of the EBCDIC-related things (but is it not a full test).
687    
688    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
689        in a small tidy to the code.
690    
691    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
692    
693    18. If the --only-matching (-o) option in pcregrep is specified multiple
694        times, each one causes appropriate output. For example, -o1 -o2 outputs the
695        substrings matched by the 1st and 2nd capturing parentheses. A separating
696        string can be specified by --om-separator (default empty).
697    
698    19. Improving the first n character searches.
699    
700    20. Turn case lists for horizontal and vertical white space into macros so that
701        they are defined only once.
702    
703    21. This set of changes together give more compatible Unicode case-folding
704        behaviour for characters that have more than one other case when UCP
705        support is available.
706    
707        (a) The Unicode property table now has offsets into a new table of sets of
708            three or more characters that are case-equivalent. The MultiStage2.py
709            script that generates these tables (the pcre_ucd.c file) now scans
710            CaseFolding.txt instead of UnicodeData.txt for character case
711            information.
712    
713        (b) The code for adding characters or ranges of characters to a character
714            class has been abstracted into a generalized function that also handles
715            case-independence. In UTF-mode with UCP support, this uses the new data
716            to handle characters with more than one other case.
717    
718        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
719            whose other case is greater than 256 are now correctly matched
720            caselessly. Previously, the high codepoint matched the low one, but not
721            vice versa.
722    
723        (d) The processing of \h, \H, \v, and \ in character classes now makes use
724            of the new class addition function, using character lists defined as
725            macros alongside the case definitions of 20 above.
726    
727        (e) Caseless back references now work with characters that have more than
728            one other case.
729    
730        (f) General caseless matching of characters with more than one other case
731            is supported.
732    
733    22. Unicode character properties were updated from Unicode 6.2.0
734    
735    23. Improved CMake support under Windows. Patch by Daniel Richard G.
736    
737    24. Add support for 32-bit character strings, and UTF-32
738    
739    25. Major JIT compiler update (code refactoring and bugfixing).
740        Experimental Sparc 32 support is added.
741    
742    26. Applied a modified version of Daniel Richard G's patch to create
743        pcre.h.generic and config.h.generic by "make" instead of in the
744        PrepareRelease script.
745    
746    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
747        pcre_compile.c when checking for a zero character.
748    
749    28. Introducing a native interface for JIT. Through this interface, the compiled
750        machine code can be directly executed. The purpose of this interface is to
751        provide fast pattern matching, so several sanity checks are not performed.
752        However, feature tests are still performed. The new interface provides
753        1.4x speedup compared to the old one.
754    
755    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
756        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
757        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
758    
759    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
760        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
761        the "old" RFC 2279). Instead, it ended up passing a negative length to
762        pcre_exec().
763    
764    31. Add support for GCC's visibility feature to hide internal functions.
765    
766    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
767        "unknown -C option" after outputting 0 or 1.
768    
769    33. There is now support for generating a code coverage report for the test
770        suite in environments where gcc is the compiler and lcov is installed. This
771        is mainly for the benefit of the developers.
772    
773    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
774        unaddressable using valgrind annotations, allowing valgrind to detect
775        invalid memory accesses. This is mainly for the benefit of the developers.
776    
777    25. (*UTF) can now be used to start a pattern in any of the three libraries.
778    
779    26. Give configure error if --enable-cpp but no C++ compiler found.
780    
781    
782    Version 8.31 06-July-2012
783    -------------------------
784    
785    1.  Fixing a wrong JIT test case and some compiler warnings.
786    
787    2.  Removed a bashism from the RunTest script.
788    
789    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
790        to unsigned type, result still unsigned" that was given by an MS compiler
791        on encountering the code "-sizeof(xxx)".
792    
793    4.  Partial matching support is added to the JIT compiler.
794    
795    5.  Fixed several bugs concerned with partial matching of items that consist
796        of more than one character:
797    
798        (a) /^(..)\1/ did not partially match "aba" because checking references was
799            done on an "all or nothing" basis. This also applied to repeated
800            references.
801    
802        (b) \R did not give a hard partial match if \r was found at the end of the
803            subject.
804    
805        (c) \X did not give a hard partial match after matching one or more
806            characters at the end of the subject.
807    
808        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
809            a partial match for the string "\r".
810    
811        (e) When newline was set to CRLF, the metacharacter "." did not recognize
812            a partial match for a CR character at the end of the subject string.
813    
814    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
815        running pcretest, the text "(JIT)" added to the output whenever JIT is
816        actually used to run the match.
817    
818    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
819        or /S+[+] with a digit between 1 and 7.
820    
821    8.  OP_NOT now supports any UTF character not just single-byte ones.
822    
823    9.  (*MARK) control verb is now supported by the JIT compiler.
824    
825    10. The command "./RunTest list" lists the available tests without actually
826        running any of them. (Because I keep forgetting what they all are.)
827    
828    11. Add PCRE_INFO_MAXLOOKBEHIND.
829    
830    12. Applied a (slightly modified) user-supplied patch that improves performance
831        when the heap is used for recursion (compiled with --disable-stack-for-
832        recursion). Instead of malloc and free for each heap frame each time a
833        logical recursion happens, frames are retained on a chain and re-used where
834        possible. This sometimes gives as much as 30% improvement.
835    
836    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
837        call.
838    
839    14. As documented, (*COMMIT) is now confined to within a positive assertion.
840    
841    15. It is now possible to link pcretest with libedit as an alternative to
842        libreadline.
843    
844    16. (*COMMIT) control verb is now supported by the JIT compiler.
845    
846    17. The Unicode data tables have been updated to Unicode 6.1.0.
847    
848    18. Added --file-list option to pcregrep.
849    
850    19. Added binary file support to pcregrep, including the -a, --binary-files,
851        -I, and --text options.
852    
853    20. The madvise function is renamed for posix_madvise for QNX compatibility
854        reasons. Fixed by Giuseppe D'Angelo.
855    
856    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
857    
858    22. Changed the option for creating symbolic links for 16-bit man pages from
859        -s to -sf so that re-installing does not cause issues.
860    
861    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
862    
863    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
864        matches in certain environments (the workspace was not being correctly
865        retained). Also added to pcre_dfa_exec() a simple plausibility check on
866        some of the workspace data at the beginning of a restart.
867    
868    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
869        was not doing so when it should - probably a typo introduced by SVN 528
870        (change 8.10/14).
871    
872    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
873        \w+ when the character tables indicated that \x{c4} was a word character.
874        There were several related cases, all because the tests for doing a table
875        lookup were testing for characters less than 127 instead of 255.
876    
877    27. If a pattern contains capturing parentheses that are not used in a match,
878        their slots in the ovector are set to -1. For those that are higher than
879        any matched groups, this happens at the end of processing. In the case when
880        there were back references that the ovector was too small to contain
881        (causing temporary malloc'd memory to be used during matching), and the
882        highest capturing number was not used, memory off the end of the ovector
883        was incorrectly being set to -1. (It was using the size of the temporary
884        memory instead of the true size.)
885    
886    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
887        ovector size, it uses memory at the end of the block that it has got.
888    
889    29. Check for an overlong MARK name and give an error at compile time. The
890        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
891    
892    30. JIT compiler update.
893    
894    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
895        Rill for the patch.
896    
897    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
898    
899    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
900    
901    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
902        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
903        was enabled.
904    
905    35. Improve JIT code generation for greedy plus quantifier.
906    
907    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
908        1 to "aa" instead of to an empty string. The bug affected repeated groups
909        that could potentially match an empty string.
910    
911    37. Optimizing single character iterators in JIT.
912    
913    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
914        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
915        codepoints that are too big for the mode are faulted, and in a UTF mode,
916        disallowed codepoints are also faulted.
917    
918    39. If PCRE was compiled with UTF support, in three places in the DFA
919        matcher there was code that should only have been obeyed in UTF mode, but
920        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
921        processing when bytes with values greater than 127 were present. In 16-bit
922        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
923        both cases the values are those that cannot be the first data item in a UTF
924        character. The three items that might have provoked this were recursions,
925        possessively repeated groups, and atomic groups.
926    
927    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
928        and pcregrep, because some OS require shared objects to be explicitly
929        passed to ld, causing the link step to fail if they are not.
930    
931    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
932        mode, patterns that started with \h* or \R* might be incorrectly matched.
933    
934    
935    Version 8.30 04-February-2012
936    -----------------------------
937    
938    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
939        name is defined in ctype.h.
940    
941    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
942        only in quite long subpatterns.
943    
944    3.  Removed the function pcre_info(), which has been obsolete and deprecated
945        since it was replaced by pcre_fullinfo() in February 2000.
946    
947    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
948        match a (*MARK), and the match failed at the start of the subject, a
949        reference to memory before the start of the subject could occur. This bug
950        was introduced by fix 17 of release 8.21.
951    
952    5.  A reference to an unset group with zero minimum repetition was giving
953        totally wrong answers (in non-JavaScript-compatibility mode). For example,
954        /(another)?(\1?)test/ matched against "hello world test". This bug was
955        introduced in release 8.13.
956    
957    6.  Add support for 16-bit character strings (a large amount of work involving
958        many changes and refactorings).
959    
960    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
961        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
962        from a file.
963    
964    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
965        rounding is not applied in this particular case).
966    
967    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
968        if they appear, or are escaped, in patterns.
969    
970    10. Get rid of a number of -Wunused-but-set-variable warnings.
971    
972    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
973        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
974        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
975        also returns the mark "x". This bug applied to capturing parentheses,
976        non-capturing parentheses, and atomic parentheses. It also applied to some
977        assertions.
978    
979    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
980        information out of configure.ac instead of relying on pcre.h.generic, which
981        is not stored in the repository.
982    
983    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
984        -lreadline.
985    
986    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
987    
988    15. Applied Graycode's patch to put the top-level frame on the stack rather
989        than the heap when not using the stack for recursion. This gives a
990        performance improvement in many cases when recursion is not deep.
991    
992    16. Experimental code added to "pcretest -C" to output the stack frame size.
993    
994    
995    Version 8.21 12-Dec-2011
996    ------------------------
997    
998    1.  Updating the JIT compiler.
999    
1000    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1001        are added as well.
1002    
1003    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1004        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1005        calling _pcre_jit_exec. Some extra comments are added.
1006    
1007    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1008        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1009        was introduced by change 18 for 8.20.
1010    
1011    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1012        ECMA-262 standard.
1013    
1014    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1015        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1016        This bug was probably introduced by change 9 of 8.13.
1017    
1018    7.  While fixing 6 above, I noticed that a number of other items were being
1019        incorrectly rejected as "not fixed length". This arose partly because newer
1020        opcodes had not been added to the fixed-length checking code. I have (a)
1021        corrected the bug and added tests for these items, and (b) arranged for an
1022        error to occur if an unknown opcode is encountered while checking for fixed
1023        length instead of just assuming "not fixed length". The items that were
1024        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1025        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1026        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1027    
1028    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1029        being incorrectly compiled and would have given unpredicatble results.
1030    
1031    9.  A possessively repeated subpattern with minimum repeat count greater than
1032        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1033        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1034        the first (A) could occur when it should not.
1035    
1036    10. Add a cast and remove a redundant test from the code.
1037    
1038    11. JIT should use pcre_malloc/pcre_free for allocation.
1039    
1040    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1041        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1042        is anything other than /usr, -L is still shown).
1043    
1044    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1045    
1046    14. Perl does not support \N without a following name in a [] class; PCRE now
1047        also gives an error.
1048    
1049    15. If a forward reference was repeated with an upper limit of around 2000,
1050        it caused the error "internal error: overran compiling workspace". The
1051        maximum number of forward references (including repeats) was limited by the
1052        internal workspace, and dependent on the LINK_SIZE. The code has been
1053        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1054        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1055        of around 200,000 forward references. While doing this, I also speeded up
1056        the filling in of repeated forward references.
1057    
1058    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1059        incorrectly expecting the subject to contain another "a" after the start.
1060    
1061    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1062        in the match, the SKIP should be ignored. This was not happening; instead
1063        the SKIP was being treated as NOMATCH. For patterns such as
1064        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1065        tested.
1066    
1067    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1068        now much more compatible with Perl, in particular in cases where the result
1069        is a non-match for a non-anchored pattern. For example, if
1070        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1071        "m", where previously it did not return a name. A side effect of this
1072        change is that for partial matches, the last encountered mark name is
1073        returned, as for non matches. A number of tests that were previously not
1074        Perl-compatible have been moved into the Perl-compatible test files. The
1075        refactoring has had the pleasing side effect of removing one argument from
1076        the match() function, thus reducing its stack requirements.
1077    
1078    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1079        subsequent uses of /S (without +) incorrectly behaved like /S+.
1080    
1081    21. Retrieve executable code size support for the JIT compiler and fixing
1082        some warnings.
1083    
1084    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1085        not work when the shorter character appeared right at the end of the
1086        subject string.
1087    
1088    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1089        systems.
1090    
1091    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1092        output it when the /M option is used in pcretest.
1093    
1094    25. The CheckMan script was not being included in the distribution. Also, added
1095        an explicit "perl" to run Perl scripts from the PrepareRelease script
1096        because this is reportedly needed in Windows.
1097    
1098    26. If study data was being save in a file and studying had not found a set of
1099        "starts with" bytes for the pattern, the data written to the file (though
1100        never used) was taken from uninitialized memory and so caused valgrind to
1101        complain.
1102    
1103    27. Updated RunTest.bat as provided by Sheri Pierce.
1104    
1105    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1106    
1107    29. Computation of memory usage for the table of capturing group names was
1108        giving an unnecessarily large value.
1109    
1110    
1111    Version 8.20 21-Oct-2011
1112    ------------------------
1113    
1114    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1115        a POSIX class. After further experiments with Perl, which convinced me that
1116        Perl has bugs and confusions, a closing square bracket is no longer allowed
1117        in a POSIX name. This bug also affected patterns with classes that started
1118        with full stops.
1119    
1120    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1121        captured substring, but while checking the failing first alternative,
1122        substring 1 is temporarily captured. If the output vector supplied to
1123        pcre_exec() was not big enough for this capture, the yield of the function
1124        was still zero ("insufficient space for captured substrings"). This cannot
1125        be totally fixed without adding another stack variable, which seems a lot
1126        of expense for a edge case. However, I have improved the situation in cases
1127        such as /(a)(b)x|abc/ matched against "abc", where the return code
1128        indicates that fewer than the maximum number of slots in the ovector have
1129        been set.
1130    
1131    3.  Related to (2) above: when there are more back references in a pattern than
1132        slots in the output vector, pcre_exec() uses temporary memory during
1133        matching, and copies in the captures as far as possible afterwards. It was
1134        using the entire output vector, but this conflicts with the specification
1135        that only 2/3 is used for passing back captured substrings. Now it uses
1136        only the first 2/3, for compatibility. This is, of course, another edge
1137        case.
1138    
1139    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1140        main code base, and can be used by building with --enable-jit. When this is
1141        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1142        runtime --no-jit option is given.
1143    
1144    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1145        ovector, the return from the function was zero, implying that there were
1146        other matches that did not fit. The correct "exactly full" value is now
1147        returned.
1148    
1149    6.  If a subpattern that was called recursively or as a subroutine contained
1150        (*PRUNE) or any other control that caused it to give a non-standard return,
1151        invalid errors such as "Error -26 (nested recursion at the same subject
1152        position)" or even infinite loops could occur.
1153    
1154    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1155        computing the minimum length on reaching *ACCEPT, and so ended up with the
1156        wrong value of 1 rather than 0. Further investigation indicates that
1157        computing a minimum subject length in the presence of *ACCEPT is difficult
1158        (think back references, subroutine calls), and so I have changed the code
1159        so that no minimum is registered for a pattern that contains *ACCEPT.
1160    
1161    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1162        it was not handled as intended. [But see 16 below.]
1163    
1164    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1165        Sheri Pierce.
1166    
1167    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1168        the first byte in a match must be "a".
1169    
1170    11. Change 17 for 8.13 increased the recursion depth for patterns like
1171        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1172        pattern contains any instances of (*THEN). If it does not, the old
1173        optimizations are restored. It would be nice to do this on a per-group
1174        basis, but at the moment that is not feasible.
1175    
1176    12. In some environments, the output of pcretest -C is CRLF terminated. This
1177        broke RunTest's code that checks for the link size. A single white space
1178        character after the value is now allowed for.
1179    
1180    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1181        For "fr", it uses the Windows-specific input and output files.
1182    
1183    14. If (*THEN) appeared in a group that was called recursively or as a
1184        subroutine, it did not work as intended. [But see next item.]
1185    
1186    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1187        pattern fragments (but not containing any | characters). If A and B are
1188        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1189        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1190        D. In other words, Perl considers parentheses that do not contain any |
1191        characters to be part of a surrounding alternative, whereas PCRE was
1192        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1193        differently. PCRE now behaves in the same way as Perl, except in the case
1194        of subroutine/recursion calls such as (?1) which have in any case always
1195        been different (but PCRE had them first :-).
1196    
1197    16. Related to 15 above: Perl does not treat the | in a conditional group as
1198        creating alternatives. Such a group is treated in the same way as an
1199        ordinary group without any | characters when processing (*THEN). PCRE has
1200        been changed to match Perl's behaviour.
1201    
1202    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1203        RunGrepTest script failed.
1204    
1205    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1206        inevitable for groups that contain captures, but it can lead to a lot of
1207        stack use in large patterns. The old behaviour has been restored for atomic
1208        groups that do not contain any capturing parentheses.
1209    
1210    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1211        suppress the check for a minimum subject length at run time. (If it was
1212        given to pcre_exec() or pcre_dfa_exec() it did work.)
1213    
1214    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1215        fail to work when decoding hex characters in data strings in EBCDIC
1216        environments.
1217    
1218    21. It appears that in at least one Mac OS environment, the isxdigit() function
1219        is implemented as a macro that evaluates to its argument more than once,
1220        contravening the C 90 Standard (I haven't checked a later standard). There
1221        was an instance in pcretest which caused it to go wrong when processing
1222        \x{...} escapes in subject strings. The has been rewritten to avoid using
1223        things like p++ in the argument of isxdigit().
1224    
1225    
1226    Version 8.13 16-Aug-2011
1227    ------------------------
1228    
1229    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1230    
1231    2.  Two minor typos in pcre_internal.h have been fixed.
1232    
1233    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1234        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1235        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1236    
1237    4.  There were a number of related bugs in the code for matching backrefences
1238        caselessly in UTF-8 mode when codes for the characters concerned were
1239        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1240        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1241        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1242        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1243        2-byte code at the end of the subject (it thought there wasn't enough data
1244        left).
1245    
1246    5.  Comprehensive information about what went wrong is now returned by
1247        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1248        as the output vector has at least 2 elements. The offset of the start of
1249        the failing character and a reason code are placed in the vector.
1250    
1251    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1252        now returned is for the first byte of the failing character, instead of the
1253        last byte inspected. This is an incompatible change, but I hope it is small
1254        enough not to be a problem. It makes the returned offset consistent with
1255        pcre_exec() and pcre_dfa_exec().
1256    
1257    7.  pcretest now gives a text phrase as well as the error number when
1258        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1259        failure, the offset and reason code are output.
1260    
1261    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1262        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1263        back over a single character (\n). This seems wrong (because it treated the
1264        two characters as a single entity when going forwards), conflicts with the
1265        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1266        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1267        has been changed.
1268    
1269    9.  Some internal refactoring has changed the processing so that the handling
1270        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1271        time (the PCRE_DOTALL option was changed this way some time ago: version
1272        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1273        which was always a bit of a fudge. It also means that there is one less
1274        argument for the match() function, which reduces its stack requirements
1275        slightly. This change also fixes an incompatibility with Perl: the pattern
1276        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1277    
1278    10. More internal refactoring has drastically reduced the number of recursive
1279        calls to match() for possessively repeated groups such as (abc)++ when
1280        using pcre_exec().
1281    
1282    11. While implementing 10, a number of bugs in the handling of groups were
1283        discovered and fixed:
1284    
1285        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1286        (a|)*(?1) gave a compile-time internal error.
1287        ((a|)+)+  did not notice that the outer group could match an empty string.
1288        (^a|^)+   was not marked as anchored.
1289        (.*a|.*)+ was not marked as matching at start or after a newline.
1290    
1291    12. Yet more internal refactoring has removed another argument from the match()
1292        function. Special calls to this function are now indicated by setting a
1293        value in a variable in the "match data" data block.
1294    
1295    13. Be more explicit in pcre_study() instead of relying on "default" for
1296        opcodes that mean there is no starting character; this means that when new
1297        ones are added and accidentally left out of pcre_study(), testing should
1298        pick them up.
1299    
1300    14. The -s option of pcretest has been documented for ages as being an old
1301        synonym of -m (show memory usage). I have changed it to mean "force study
1302        for every regex", that is, assume /S for every regex. This is similar to -i
1303        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1304        using it. It makes it easier to run collections of tests with and without
1305        study enabled, and thereby test pcre_study() more easily. All the standard
1306        tests are now run with and without -s (but some patterns can be marked as
1307        "never study" - see 20 below).
1308    
1309    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1310        restoration of the capturing data to the outer values was not happening
1311        correctly.
1312    
1313    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1314        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1315        pattern had matched an empty string, and so incorrectly returned a no
1316        match.
1317    
1318    17. There was optimizing code for the last branch of non-capturing parentheses,
1319        and also for the obeyed branch of a conditional subexpression, which used
1320        tail recursion to cut down on stack usage. Unfortunately, now that there is
1321        the possibility of (*THEN) occurring in these branches, tail recursion is
1322        no longer possible because the return has to be checked for (*THEN). These
1323        two optimizations have therefore been removed. [But see 8.20/11 above.]
1324    
1325    18. If a pattern containing \R was studied, it was assumed that \R always
1326        matched two bytes, thus causing the minimum subject length to be
1327        incorrectly computed because \R can also match just one byte.
1328    
1329    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1330        was incorrectly computed.
1331    
1332    20. If /S is present twice on a test pattern in pcretest input, it now
1333        *disables* studying, thereby overriding the use of -s on the command line
1334        (see 14 above). This is necessary for one or two tests to keep the output
1335        identical in both cases.
1336    
1337    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1338        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1339    
1340    22. When an atomic group that contained a capturing parenthesis was
1341        successfully matched, but the branch in which it appeared failed, the
1342        capturing was not being forgotten if a higher numbered group was later
1343        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1344        group 1 to "a", when in fact it should be unset. This applied to multi-
1345        branched capturing and non-capturing groups, repeated or not, and also to
1346        positive assertions (capturing in negative assertions does not happen
1347        in PCRE) and also to nested atomic groups.
1348    
1349    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1350        subject after a captured substring, to make it easier to tell which of a
1351        number of identical substrings has been captured.
1352    
1353    24. The way atomic groups are processed by pcre_exec() has been changed so that
1354        if they are repeated, backtracking one repetition now resets captured
1355        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1356        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1357        "aaa". Previously, it would have been "a". As part of this code
1358        refactoring, the way recursive calls are handled has also been changed.
1359    
1360    25. If an assertion condition captured any substrings, they were not passed
1361        back unless some other capturing happened later. For example, if
1362        (?(?=(a))a) was matched against "a", no capturing was returned.
1363    
1364    26. When studying a pattern that contained subroutine calls or assertions,
1365        the code for finding the minimum length of a possible match was handling
1366        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1367        group 1 called group 2 while simultaneously a separate group 2 called group
1368        1). A stack overflow occurred in this case. I have fixed this by limiting
1369        the recursion depth to 10.
1370    
1371    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1372        Fortmann. This supports explicit test numbers on the command line, and has
1373        argument validation and error reporting.
1374    
1375    28. An instance of \X with an unlimited repeat could fail if at any point the
1376        first character it looked at was a mark character.
1377    
1378    29. Some minor code refactoring concerning Unicode properties and scripts
1379        should reduce the stack requirement of match() slightly.
1380    
1381    30. Added the '=' option to pcretest to check the setting of unused capturing
1382        slots at the end of the pattern, which are documented as being -1, but are
1383        not included in the return count.
1384    
1385    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1386        compiled something random. Now it gives a compile-time error (as does
1387        Perl).
1388    
1389    32. A *MARK encountered during the processing of a positive assertion is now
1390        recorded and passed back (compatible with Perl).
1391    
1392    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1393        had alternative anchored branches, the search for a second match in a line
1394        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1395        matched the line "0102" twice. The same bug affected patterns that started
1396        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1397        twice.
1398    
1399    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1400        does, and because of capturing effects, quantifying parenthesized
1401        assertions may at times be useful. Quantifiers are now allowed for
1402        parenthesized assertions.
1403    
1404    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1405    
1406    36. \g was being checked for fancy things in a character class, when it should
1407        just be a literal "g".
1408    
1409    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1410        appearance of a nested POSIX class supersedes an apparent external class.
1411        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1412        unescaped square brackets may also appear as part of class names. For
1413        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1414        more like Perl. (But see 8.20/1 above.)
1415    
1416    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1417        was because it thought it was \N{name}, which is not supported).
1418    
1419    39. Add minix to OS list not supporting the -S option in pcretest.
1420    
1421    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1422        cannot analyze patterns in sufficient detail to catch mutual recursions
1423        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1424        subgroup is called recursively as a subpattern for a second time at the
1425        same position in the subject string. In previous releases this might have
1426        been caught by the recursion limit, or it might have run out of stack.
1427    
1428    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1429        happen only once. PCRE was, however incorrectly giving a compile time error
1430        "recursive call could loop indefinitely" because it cannot analyze the
1431        pattern in sufficient detail. The compile time test no longer happens when
1432        PCRE is compiling a conditional subpattern, but actual runaway loops are
1433        now caught at runtime (see 40 above).
1434    
1435    42. It seems that Perl allows any characters other than a closing parenthesis
1436        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1437        has been changed to be the same.
1438    
1439    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1440        as not to get warnings when autogen.sh is called. Also changed
1441        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1442    
1443    44. To help people who use pcregrep to scan files containing exceedingly long
1444        lines, the following changes have been made:
1445    
1446        (a) The default value of the buffer size parameter has been increased from
1447            8K to 20K. (The actual buffer used is three times this size.)
1448    
1449        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1450            PCRE is built.
1451    
1452        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1453            to be set at run time.
1454    
1455        (d) Numerical values in pcregrep options can be followed by K or M, for
1456            example --buffer-size=50K.
1457    
1458        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1459            given and the return code is set to 2.
1460    
1461    45. Add a pointer to the latest mark to the callout data block.
1462    
1463    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1464        partial match of an empty string instead of no match. This was specific to
1465        the use of ".".
1466    
1467    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1468        complete match instead of a partial match. This bug was dependent on both
1469        the PCRE_UTF8 and PCRE_DOTALL options being set.
1470    
1471    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1472        starting byte set, because \b was not being ignored.
1473    
1474    
1475    Version 8.12 15-Jan-2011
1476    ------------------------
1477    
1478    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1479        checks for such things as part of the documentation building process.
1480    
1481    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1482        --match-limit and --recursion-limit options (added for 8.11). In
1483        particular, this made one of the standard tests fail. (The integer value
1484        went into the wrong half of a long int.)
1485    
1486    3.  If the --colour option was given to pcregrep with -v (invert match), it
1487        did strange things, either producing crazy output, or crashing. It should,
1488        of course, ignore a request for colour when reporting lines that do not
1489        match.
1490    
1491    4.  Another pcregrep bug caused similar problems if --colour was specified with
1492        -M (multiline) and the pattern match finished with a line ending.
1493    
1494    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1495        matched in multiline mode, the following line was shown as part of the
1496        match. This seems wrong, so I have changed it.
1497    
1498    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1499        the check for further matches in the same line (so they could be coloured)
1500        to overrun the end of the current line. If another match was found, it was
1501        incorrectly shown (and then shown again when found in the next line).
1502    
1503    7.  If pcregrep was compiled under Windows, there was a reference to the
1504        function pcregrep_exit() before it was defined. I am assuming this was
1505        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1506        reported by a user. I've moved the definition above the reference.
1507    
1508    
1509    Version 8.11 10-Dec-2010
1510    ------------------------
1511    
1512    1.  (*THEN) was not working properly if there were untried alternatives prior
1513        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1514        backtracked to try for "b" instead of moving to the next alternative branch
1515        at the same level (in this case, to look for "c"). The Perl documentation
1516        is clear that when (*THEN) is backtracked onto, it goes to the "next
1517        alternative in the innermost enclosing group".
1518    
1519    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1520        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1521        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1522        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1523        (*THEN).
1524    
1525    3.  If \s appeared in a character class, it removed the VT character from
1526        the class, even if it had been included by some previous item, for example
1527        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1528        of \s, but is part of the POSIX "space" class.)
1529    
1530    4.  A partial match never returns an empty string (because you can always
1531        match an empty string at the end of the subject); however the checking for
1532        an empty string was starting at the "start of match" point. This has been
1533        changed to the "earliest inspected character" point, because the returned
1534        data for a partial match starts at this character. This means that, for
1535        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1536        (previously it gave "no match").
1537    
1538    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1539        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1540        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1541        has an implication that the given string is incomplete (because a partial
1542        match is preferred over a full match). For this reason, these items now
1543        give a partial match in this situation. [Aside: previously, the one case
1544        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1545        match rather than a full match, which was wrong by the old rules, but is
1546        now correct.]
1547    
1548    6.  There was a bug in the handling of #-introduced comments, recognized when
1549        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1550        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1551        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1552        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1553        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1554        places in pcre_compile().
1555    
1556    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1557        comments when looking ahead for named forward references to subpatterns,
1558        the only newline sequence it recognized was NL. It now handles newlines
1559        according to the set newline convention.
1560    
1561    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1562        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1563        cater for a lack of strerror(). These oversights have been fixed.
1564    
1565    9.  Added --match-limit and --recursion-limit to pcregrep.
1566    
1567    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1568    
1569    11. When the -o option was used, pcregrep was setting a return code of 1, even
1570        when matches were found, and --line-buffered was not being honoured.
1571    
1572    12. Added an optional parentheses number to the -o and --only-matching options
1573        of pcregrep.
1574    
1575    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1576        can match an empty string. The code to do it in pcretest and pcredemo
1577        needed fixing:
1578    
1579        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1580            only one byte after an empty string match just before CRLF (this case
1581            just got forgotten; "any" and "anycrlf" were OK).
1582    
1583        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1584            mode when an empty string match preceded an ASCII character followed by
1585            a non-ASCII character. (The code for advancing by one character rather
1586            than one byte was nonsense.)
1587    
1588        (c) The pcredemo.c sample program did not have any code at all to handle
1589            the cases when CRLF is a valid newline sequence.
1590    
1591    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1592        as a starting offset was within the subject string. There is now a new
1593        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1594        negative or greater than the length of the string. In order to test this,
1595        pcretest is extended to allow the setting of negative starting offsets.
1596    
1597    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1598        starting offset points to the beginning of a UTF-8 character was
1599        unnecessarily clumsy. I tidied it up.
1600    
1601    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1602        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1603    
1604    17. Nobody had reported that the --include_dir option, which was added in
1605        release 7.7 should have been called --include-dir (hyphen, not underscore)
1606        for compatibility with GNU grep. I have changed it to --include-dir, but
1607        left --include_dir as an undocumented synonym, and the same for
1608        --exclude-dir, though that is not available in GNU grep, at least as of
1609        release 2.5.4.
1610    
1611    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1612        characters from a string of bytes) have been redefined so as not to use
1613        loops, in order to improve performance in some environments. At the same
1614        time, I abstracted some of the common code into auxiliary macros to save
1615        repetition (this should not affect the compiled code).
1616    
1617    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1618        compile-time error is now given if \c is not followed by an ASCII
1619        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1620        different, and any byte value is allowed.)
1621    
1622    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1623        START_OPTIMIZE option, which is now allowed at compile time - but just
1624        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1625        to pcregrep and other applications that have no direct access to PCRE
1626        options. The new /Y option in pcretest sets this option when calling
1627        pcre_compile().
1628    
1629    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1630        back references. Groups containing recursive back references were forced to
1631        be atomic by that change, but in the case of named groups, the amount of
1632        memory required was incorrectly computed, leading to "Failed: internal
1633        error: code overflow". This has been fixed.
1634    
1635    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1636        pcretest.c, to avoid build problems in some Borland environments.
1637    
1638    
1639    Version 8.10 25-Jun-2010
1640    ------------------------
1641    
1642    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1643        THEN.
1644    
1645    2.  (*ACCEPT) was not working when inside an atomic group.
1646    
1647    3.  Inside a character class, \B is treated as a literal by default, but
1648        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1649        causes the error). The code is unchanged, but I tidied the documentation.
1650    
1651    4.  Inside a character class, PCRE always treated \R and \X as literals,
1652        whereas Perl faults them if its -w option is set. I have changed PCRE so
1653        that it faults them when PCRE_EXTRA is set.
1654    
1655    5.  Added support for \N, which always matches any character other than
1656        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1657    
1658    6.  When compiling pcregrep with newer versions of gcc which may have
1659        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1660        declared with attribute warn_unused_result" were given. Just casting the
1661        result to (void) does not stop the warnings; a more elaborate fudge is
1662        needed. I've used a macro to implement this.
1663    
1664    7.  Minor change to pcretest.c to avoid a compiler warning.
1665    
1666    8.  Added four artifical Unicode properties to help with an option to make
1667        \s etc use properties (see next item). The new properties are: Xan
1668        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1669    
1670    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1671        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1672        this option. Modified pcretest to add /W to test this facility. Added
1673        REG_UCP to make it available via the POSIX interface.
1674    
1675    10. Added --line-buffered to pcregrep.
1676    
1677    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1678        studied, and the match started with a letter with a code point greater than
1679        127 whose first byte was different to the first byte of the other case of
1680        the letter, the other case of this starting letter was not recognized
1681        (#976).
1682    
1683    12. If a pattern that was studied started with a repeated Unicode property
1684        test, for example, \p{Nd}+, there was the theoretical possibility of
1685        setting up an incorrect bitmap of starting bytes, but fortunately it could
1686        not have actually happened in practice until change 8 above was made (it
1687        added property types that matched character-matching opcodes).
1688    
1689    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1690        possible starting bytes for non-anchored patterns.
1691    
1692    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1693        \R, and also a number of cases that involve Unicode properties, both
1694        explicit and implicit when PCRE_UCP is set.
1695    
1696    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1697        input, it could crash or give wrong results if characters with values
1698        greater than 0xc0 were present in the subject string. (Detail: it assumed
1699        UTF-8 input when processing these items.)
1700    
1701    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1702        size_t is 64-bit (#991).
1703    
1704    17. Added a check for running out of memory when PCRE is compiled with
1705        --disable-stack-for-recursion (#990).
1706    
1707    18. If the last data line in a file for pcretest does not have a newline on
1708        the end, a newline was missing in the output.
1709    
1710    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1711        less than 128) in its various bitmaps. However, there is a facility for
1712        generating tables according to the current locale when PCRE is compiled. It
1713        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1714        characters, are recognized by isspace() and therefore were getting set in
1715        these tables, and indeed these tables seem to approximate to ISO 8859. This
1716        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1717        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1718        which of course cannot start UTF-8 characters. I have changed the code so
1719        that only real ASCII characters (less than 128) and the correct starting
1720        bytes for UTF-8 encodings are set for characters greater than 127 when in
1721        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1722        altogether.)
1723    
1724    20. Added the /T option to pcretest so as to be able to run tests with non-
1725        standard character tables, thus making it possible to include the tests
1726        used for 19 above in the standard set of tests.
1727    
1728    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1729        reference to a subpattern the other side of a comment that contains an
1730        opening parenthesis caused either an internal compiling error, or a
1731        reference to the wrong subpattern.
1732    
1733    
1734    Version 8.02 19-Mar-2010
1735    ------------------------
1736    
1737    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1738    
1739    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1740        configured.
1741    
1742    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1743        original author of that file, following a query about its status.
1744    
1745    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1746        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1747    
1748    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1749        quantifier applied to a forward-referencing subroutine call, could compile
1750        incorrect code or give the error "internal error: previously-checked
1751        referenced subpattern not found".
1752    
1753    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1754        variables to point to external functions. For these systems, therefore,
1755        pcre_malloc etc. are now initialized to local functions that call the
1756        relevant global functions.
1757    
1758    7.  There were two entries missing in the vectors called coptable and poptable
1759        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1760        I've fixed the data, and added a kludgy way of testing at compile time that
1761        the lengths are correct (equal to the number of opcodes).
1762    
1763    8.  Following on from 7, I added a similar kludge to check the length of the
1764        eint vector in pcreposix.c.
1765    
1766    9.  Error texts for pcre_compile() are held as one long string to avoid too
1767        much relocation at load time. To find a text, the string is searched,
1768        counting zeros. There was no check for running off the end of the string,
1769        which could happen if a new error number was added without updating the
1770        string.
1771    
1772    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1773    
1774    11. \K was not working if it appeared in an atomic group or in a group that
1775        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1776        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1777        the assertion is positive, but not if it is negative.
1778    
1779    12. Change 11 fortuitously reduced the size of the stack frame used in the
1780        "match()" function of pcre_exec.c by one pointer. Forthcoming
1781        implementation of support for (*MARK) will need an extra pointer on the
1782        stack; I have reserved it now, so that the stack frame size does not
1783        decrease.
1784    
1785    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1786        item in branch that calls a recursion is a subroutine call - as in the
1787        second branch in the above example - was incorrectly given the compile-
1788        time error "recursive call could loop indefinitely" because pcre_compile()
1789        was not correctly checking the subroutine for matching a non-empty string.
1790    
1791    14. The checks for overrunning compiling workspace could trigger after an
1792        overrun had occurred. This is a "should never occur" error, but it can be
1793        triggered by pathological patterns such as hundreds of nested parentheses.
1794        The checks now trigger 100 bytes before the end of the workspace.
1795    
1796    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1797    
1798    
1799    Version 8.01 19-Jan-2010
1800    ------------------------
1801    
1802    1.  If a pattern contained a conditional subpattern with only one branch (in
1803        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1804        computed the wrong minimum data length (which is of course zero for such
1805        subpatterns). This could cause incorrect "no match" results.
1806    
1807    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1808        the pattern is reset in the first branch, pcre_compile() failed with
1809        "internal error: code overflow at offset...". This happened only when
1810        the reset was to the original external option setting. (An optimization
1811        abstracts leading options settings into an external setting, which was the
1812        cause of this.)
1813    
1814    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1815        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1816        assertion pattern did not match (meaning that the assertion was true), it
1817        was incorrectly treated as false if the SKIP had been reached during the
1818        matching. This also applied to assertions used as conditions.
1819    
1820    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1821        assertion subpattern, including such a pattern used as a condition,
1822        unpredictable results occurred, instead of the error return
1823        PCRE_ERROR_DFA_UITEM.
1824    
1825    5.  The C++ GlobalReplace function was not working like Perl for the special
1826        situation when an empty string is matched. It now does the fancy magic
1827        stuff that is necessary.
1828    
1829    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1830        removed. (These were left over from very, very early versions of PCRE.)
1831    
1832    7.  Some cosmetic changes to the code to make life easier when compiling it
1833        as part of something else:
1834    
1835        (a) Change DEBUG to PCRE_DEBUG.
1836    
1837        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1838            called "current" as "current_branch", to prevent a collision with the
1839            Linux macro when compiled as a kernel module.
1840    
1841        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1842            prevent a collision with the Linux macro when compiled as a kernel
1843            module.
1844    
1845    8.  In pcre_compile() there are some checks for integer overflows that used to
1846        cast potentially large values to (double). This has been changed to that
1847        when building, a check for int64_t is made, and if it is found, it is used
1848        instead, thus avoiding the use of floating point arithmetic. (There is no
1849        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1850        double.
1851    
1852    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1853        2005 (difference between two addresses compared to an unsigned value).
1854    
1855    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1856        custom one, because of the following reported problem in Windows:
1857    
1858          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1859              under Win32.
1860          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1861              therefore missing the function definition.
1862          - The compiler thus generates a "C" signature for the test function.
1863          - The linker fails to find the "C" function.
1864          - PCRE fails to configure if asked to do so against libbz2.
1865    
1866    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1867        messages were output:
1868    
1869          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1870          rerunning libtoolize, to keep the correct libtool macros in-tree.
1871          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1872    
1873        I have done both of these things.
1874    
1875    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1876        most of the time, it *can* run out if it is given a pattern that contains a
1877        runaway infinite recursion. I updated the discussion in the pcrestack man
1878        page.
1879    
1880    13. Now that we have gone to the x.xx style of version numbers, the minor
1881        version may start with zero. Using 08 or 09 is a bad idea because users
1882        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1883        interpreted as invalid octal numbers. I've updated the previous comment in
1884        configure.ac, and also added a check that gives an error if 08 or 09 are
1885        used.
1886    
1887    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1888        causing partial matching to fail when the end of the subject matched \W
1889        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1890    
1891    15. There were some discrepancies between the declarations in pcre_internal.h
1892        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1893        their definitions. The declarations used "const uschar *" and the
1894        definitions used USPTR. Even though USPTR is normally defined as "const
1895        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1896        reported that: "This difference in casting confuses some C++ compilers, for
1897        example, SunCC recognizes above declarations as different functions and
1898        generates broken code for hbpcre." I have changed the declarations to use
1899        USPTR.
1900    
1901    16. GNU libtool is named differently on some systems. The autogen.sh script now
1902        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1903        (FreeBSD).
1904    
1905    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1906        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1907        comment: "Figure out how to create a longlong from a string: strtoll and
1908        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1909        instance, but it only takes 2 args instead of 3!"
1910    
1911    18. A subtle bug concerned with back references has been fixed by a change of
1912        specification, with a corresponding code fix. A pattern such as
1913        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1914        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1915        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1916        same bug. Such groups have to be quantified to be useful, or contained
1917        inside another quantified group. (If there's no repetition, the reference
1918        can never match.) The problem arises because, having left the group and
1919        moved on to the rest of the pattern, a later failure that backtracks into
1920        the group uses the captured value from the final iteration of the group
1921        rather than the correct earlier one. I have fixed this in PCRE by forcing
1922        any group that contains a reference to itself to be an atomic group; that
1923        is, there cannot be any backtracking into it once it has completed. This is
1924        similar to recursive and subroutine calls.
1925    
1926    
1927    Version 8.00 19-Oct-09
1928    ----------------------
1929    
1930    1.  The table for translating pcre_compile() error codes into POSIX error codes
1931        was out-of-date, and there was no check on the pcre_compile() error code
1932        being within the table. This could lead to an OK return being given in
1933        error.
1934    
1935    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1936        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1937        in a Windows environment.
1938    
1939    3.  The pcregrep --count option prints the count for each file even when it is
1940        zero, as does GNU grep. However, pcregrep was also printing all files when
1941        --files-with-matches was added. Now, when both options are given, it prints
1942        counts only for those files that have at least one match. (GNU grep just
1943        prints the file name in this circumstance, but including the count seems
1944        more useful - otherwise, why use --count?) Also ensured that the
1945        combination -clh just lists non-zero counts, with no names.
1946    
1947    4.  The long form of the pcregrep -F option was incorrectly implemented as
1948        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1949        but it seems right to fix it, and I didn't think it was worth preserving
1950        the old behaviour.
1951    
1952    5.  The command line items --regex=pattern and --regexp=pattern were not
1953        recognized by pcregrep, which required --regex pattern or --regexp pattern
1954        (with a space rather than an '='). The man page documented the '=' forms,
1955        which are compatible with GNU grep; these now work.
1956    
1957    6.  No libpcreposix.pc file was created for pkg-config; there was just
1958        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1959    
1960    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1961        when UCP support is not needed, by modifying the Python script that
1962        generates it from Unicode data files. This should not matter if the module
1963        is correctly used as a library, but I received one complaint about 50K of
1964        unwanted data. My guess is that the person linked everything into his
1965        program rather than using a library. Anyway, it does no harm.
1966    
1967    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1968        was a minimum greater than 1 for a wide character in a possessive
1969        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1970        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1971        character. Chaos in the form of incorrect output or a compiling loop could
1972        result.
1973    
1974    9.  The restrictions on what a pattern can contain when partial matching is
1975        requested for pcre_exec() have been removed. All patterns can now be
1976        partially matched by this function. In addition, if there are at least two
1977        slots in the offset vector, the offset of the earliest inspected character
1978        for the match and the offset of the end of the subject are set in them when
1979        PCRE_ERROR_PARTIAL is returned.
1980    
1981    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1982        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1983        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1984        and may be more useful for multi-segment matching.
1985    
1986    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1987        used to be given if ever the end of the subject was reached; now it is
1988        given only if matching could not proceed because another character was
1989        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1990        string "Z", which now yields "no match" instead of "partial match". In the
1991        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1992        final character ended with (*FAIL).
1993    
1994    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1995        if the pattern had a "must contain" character that was already found in the
1996        earlier partial match, unless partial matching was again requested. For
1997        example, with the pattern /dog.(body)?/, the "must contain" character is
1998        "g". If the first part-match was for the string "dog", restarting with
1999        "sbody" failed. This bug has been fixed.
2000    
2001    13. The string returned by pcre_dfa_exec() after a partial match has been
2002        changed so that it starts at the first inspected character rather than the
2003        first character of the match. This makes a difference only if the pattern
2004        starts with a lookbehind assertion or \b or \B (\K is not supported by
2005        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2006        matching functions compatible, and I think it's the right thing to do.
2007    
2008    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2009        so that the demonstration program is easily available in environments where
2010        PCRE has not been installed from source.
2011    
2012    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2013        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2014        library.
2015    
2016    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2017        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2018        is not the first non-POSIX option to be added. Clearly some people find
2019        these options useful.
2020    
2021    17. If a caller to the POSIX matching function regexec() passes a non-zero
2022        value for nmatch with a NULL value for pmatch, the value of
2023        nmatch is forced to zero.
2024    
2025    18. RunGrepTest did not have a test for the availability of the -u option of
2026        the diff command, as RunTest does. It now checks in the same way as
2027        RunTest, and also checks for the -b option.
2028    
2029    19. If an odd number of negated classes containing just a single character
2030        interposed, within parentheses, between a forward reference to a named
2031        subpattern and the definition of the subpattern, compilation crashed with
2032        an internal error, complaining that it could not find the referenced
2033        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2034        [The bug was that it was starting one character too far in when skipping
2035        over the character class, thus treating the ] as data rather than
2036        terminating the class. This meant it could skip too much.]
2037    
2038    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2039        /g option in pcretest when the pattern contains \K, which makes it possible
2040        to have an empty string match not at the start, even when the pattern is
2041        anchored. Updated pcretest and pcredemo to use this option.
2042    
2043    21. If the maximum number of capturing subpatterns in a recursion was greater
2044        than the maximum at the outer level, the higher number was returned, but
2045        with unset values at the outer level. The correct (outer level) value is
2046        now given.
2047    
2048    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2049        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2050        make it do so. The string so far is captured, making this feature
2051        compatible with Perl.
2052    
2053    23. The tests have been re-organized, adding tests 11 and 12, to make it
2054        possible to check the Perl 5.10 features against Perl 5.10.
2055    
2056    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2057        pattern matches a fixed length string. PCRE did not allow this; now it
2058        does. Neither allows recursion.
2059    
2060    25. I finally figured out how to implement a request to provide the minimum
2061        length of subject string that was needed in order to match a given pattern.
2062        (It was back references and recursion that I had previously got hung up
2063        on.) This code has now been added to pcre_study(); it finds a lower bound
2064        to the length of subject needed. It is not necessarily the greatest lower
2065        bound, but using it to avoid searching strings that are too short does give
2066        some useful speed-ups. The value is available to calling programs via
2067        pcre_fullinfo().
2068    
2069    26. While implementing 25, I discovered to my embarrassment that pcretest had
2070        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2071        study optimizations had never been tested with that matching function.
2072        Oops. What is worse, even when it was passed study data, there was a bug in
2073        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2074        were also very few tests of studied patterns with pcre_dfa_exec().
2075    
2076    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2077        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2078        on the other side of the coin, they are no longer allowed to have different
2079        names, because these cannot be distinguished in PCRE, and this has caused
2080        confusion. (This is a difference from Perl.)
2081    
2082    28. When duplicate subpattern names are present (necessarily with different
2083        numbers, as required by 27 above), and a test is made by name in a
2084        conditional pattern, either for a subpattern having been matched, or for
2085        recursion in such a pattern, all the associated numbered subpatterns are
2086        tested, and the overall condition is true if the condition is true for any
2087        one of them. This is the way Perl works, and is also more like the way
2088        testing by number works.
2089    
2090    
2091    Version 7.9 11-Apr-09
2092    ---------------------
2093    
2094    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2095        (pcretest), all targets were linked against these libraries. This included
2096        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2097        libraries. This caused unwanted dependencies to be created. This problem
2098        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2099        pcretest is linked with readline.
2100    
2101    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2102        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2103        moved outside it again, because FALSE and TRUE are already defined in AIX,
2104        but BOOL is not.
2105    
2106    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2107        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2108    
2109    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2110        hyphens) following file names and line numbers when outputting matching
2111        lines. This is not true; no spaces are inserted. I have also clarified the
2112        wording for the --colour (or --color) option.
2113    
2114    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2115        was not coloured; this is different to GNU grep, so I have changed it to be
2116        the same.
2117    
2118    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2119        each matching line was coloured. Now it goes on to look for further matches
2120        of any of the test patterns, which is the same behaviour as GNU grep.
2121    
2122    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2123        doesn't make sense to accept an empty string match in pcregrep, so I have
2124        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2125        seems to be how GNU grep behaves. [But see later change 40 for release
2126        8.33.]
2127    
2128    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2129        start or after a newline", because the conditional assertion was not being
2130        correctly handled. The rule now is that both the assertion and what follows
2131        in the first alternative must satisfy the test.
2132    
2133    9.  If auto-callout was enabled in a pattern with a conditional group whose
2134        condition was an assertion, PCRE could crash during matching, both with
2135        pcre_exec() and pcre_dfa_exec().
2136    
2137    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2138        used for matching.
2139    
2140    11. Unicode property support in character classes was not working for
2141        characters (bytes) greater than 127 when not in UTF-8 mode.
2142    
2143    12. Added the -M command line option to pcretest.
2144    
2145    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2146    
2147    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2148    
2149    16. Added comments and documentation about mis-use of no_arg in the C++
2150        wrapper.
2151    
2152    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2153        from Martin Jerabek that uses macro names for all relevant character and
2154        string constants.
2155    
2156    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2157        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2158        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2159        these, but not everybody uses configure.
2160    
2161    19. A conditional group that had only one branch was not being correctly
2162        recognized as an item that could match an empty string. This meant that an
2163        enclosing group might also not be so recognized, causing infinite looping
2164        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2165        with the subject "ab", where knowledge that the repeated group can match
2166        nothing is needed in order to break the loop.
2167    
2168    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2169        exec(), but without supplying a callout function, matching went wrong.
2170    
2171    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2172        leak if the size of the offset vector was greater than 30. When the vector
2173        is smaller, the saved offsets during recursion go onto a local stack
2174        vector, but for larger vectors malloc() is used. It was failing to free
2175        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2176        error, in fact).
2177    
2178    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2179        heapframe that is used only when UTF-8 support is enabled. This caused no
2180        problem, but was untidy.
2181    
2182    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2183        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2184        included within another project.
2185    
2186    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2187        slightly modified by me:
2188    
2189          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2190              not building pcregrep.
2191    
2192          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2193              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2194    
2195    25. Forward references, both numeric and by name, in patterns that made use of
2196        duplicate group numbers, could behave incorrectly or give incorrect errors,
2197        because when scanning forward to find the reference group, PCRE was not
2198        taking into account the duplicate group numbers. A pattern such as
2199        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2200    
2201    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2202        the feature of a custom pointer more persuasive (as requested by a user).
2203    
2204    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2205        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2206        pre-defined.
2207    
2208    28. Added support for (*UTF8) at the start of a pattern.
2209    
2210    29. Arrange for flags added by the "release type" setting in CMake to be shown
2211        in the configuration summary.
2212    
2213    
2214    Version 7.8 05-Sep-08
2215    ---------------------
2216    
2217    1.  Replaced UCP searching code with optimized version as implemented for Ad
2218        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2219        stage table and inline lookup instead of a function, giving speed ups of 2
2220        to 5 times on some simple patterns that I tested. Permission was given to
2221        distribute the MultiStage2.py script that generates the tables (it's not in
2222        the tarball, but is in the Subversion repository).
2223    
2224    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2225        scripts.
2226    
2227    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2228        a group with a zero qualifier. The result of the study could be incorrect,
2229        or the function might crash, depending on the pattern.
2230    
2231    4.  Caseless matching was not working for non-ASCII characters in back
2232        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2233        It now works when Unicode Property Support is available.
2234    
2235    5.  In pcretest, an escape such as \x{de} in the data was always generating
2236        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2237        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2238        truncation.
2239    
2240    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2241    
2242    7.  Added two (int) casts to pcregrep when printing the difference of two
2243        pointers, in case they are 64-bit values.
2244    
2245    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2246        test 2 if it fails.
2247    
2248    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2249        and a #define of that name to empty if it is not externally set. This is to
2250        allow users of MSVC to set it if necessary.
2251    
2252    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2253        the convenience functions in the pcre_get.c source file.
2254    
2255    11. An option change at the start of a pattern that had top-level alternatives
2256        could cause overwriting and/or a crash. This command provoked a crash in
2257        some environments:
2258    
2259          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2260    
2261        This potential security problem was recorded as CVE-2008-2371.
2262    
2263    12. For a pattern where the match had to start at the beginning or immediately
2264        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2265        pcre_dfa_exec() could read past the end of the passed subject if there was
2266        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2267        pcretest so that it places the subject at the end of its malloc-ed buffer.
2268    
2269    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2270        exec() might read past the end of the data buffer in UTF-8 mode.
2271    
2272    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2273        the data contained the byte 0x85 as part of a UTF-8 character within its
2274        first line. This applied both to normal and DFA matching.
2275    
2276    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2277        /^[^d]*?$/8 failed to match "abc".
2278    
2279    16. Added a missing copyright notice to pcrecpp_internal.h.
2280    
2281    17. Make it more clear in the documentation that values returned from
2282        pcre_exec() in ovector are byte offsets, not character counts.
2283    
2284    18. Tidied a few places to stop certain compilers from issuing warnings.
2285    
2286    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2287        supplied by Stefan Weber. I made a further small update for 7.8 because
2288        there is a change of source arrangements: the pcre_searchfuncs.c module is
2289        replaced by pcre_ucd.c.
2290    
2291    
2292    Version 7.7 07-May-08
2293    ---------------------
2294    
2295    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2296        a string to a long long, pretend we don't even have a long long." This is
2297        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2298    
2299    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2300        pre-7.6 versions, which defined a global no_arg variable instead of putting
2301        it in the RE class. (See also #8 below.)
2302    
2303    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2304        Lopes.
2305    
2306    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2307    
2308        (1) The include/exclude patterns were being applied to the whole pathnames
2309            of files, instead of just to the final components.
2310    
2311        (2) If there was more than one level of directory, the subdirectories were
2312            skipped unless they satisfied the include/exclude conditions. This is
2313            inconsistent with GNU grep (and could even be seen as contrary to the
2314            pcregrep specification - which I improved to make it absolutely clear).
2315            The action now is always to scan all levels of directory, and just
2316            apply the include/exclude patterns to regular files.
2317    
2318    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2319        --exclude_dir in the tests to avoid scanning .svn directories.
2320    
2321    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2322        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2323        doesn't support NULs in patterns.
2324    
2325    7.  Added some missing "const"s to declarations of static tables in
2326        pcre_compile.c and pcre_dfa_exec.c.
2327    
2328    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2329        caused by fix #2  above. (Subsequently also a second patch to fix the
2330        first patch. And a third patch - this was a messy problem.)
2331    
2332    9.  Applied Craig's patch to remove the use of push_back().
2333    
2334    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2335        matching function regexec().
2336    
2337    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2338        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2339        references. PCRE supports relative numbers with this syntax (I don't think
2340        Oniguruma does).
2341    
2342    12. Previously, a group with a zero repeat such as (...){0} was completely
2343        omitted from the compiled regex. However, this means that if the group
2344        was called as a subroutine from elsewhere in the pattern, things went wrong
2345        (an internal error was given). Such groups are now left in the compiled
2346        pattern, with a new opcode that causes them to be skipped at execution
2347        time.
2348    
2349    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2350        to the way PCRE behaves:
2351    
2352        (a) A lone ] character is dis-allowed (Perl treats it as data).
2353    
2354        (b) A back reference to an unmatched subpattern matches an empty string
2355            (Perl fails the current match path).
2356    
2357        (c) A data ] in a character class must be notated as \] because if the
2358            first data character in a class is ], it defines an empty class. (In
2359            Perl it is not possible to have an empty class.) The empty class []
2360            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2361            The negative empty class [^] matches any one character, independently
2362            of the DOTALL setting.
2363    
2364    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2365        non-existent subpattern following a character class starting with ']' and
2366        containing () gave an internal compiling error instead of "reference to
2367        non-existent subpattern". Fortunately, when the pattern did exist, the
2368        compiled code was correct. (When scanning forwards to check for the
2369        existence of the subpattern, it was treating the data ']' as terminating
2370        the class, so got the count wrong. When actually compiling, the reference
2371        was subsequently set up correctly.)
2372    
2373    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2374        it was being rejected as not supported by pcre_dfa_exec(), even though
2375        other assertions are supported. I have made pcre_dfa_exec() support
2376        (*FAIL).
2377    
2378    16. The implementation of 13c above involved the invention of a new opcode,
2379        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2380        cannot be changed at match time, I realized I could make a small
2381        improvement to matching performance by compiling OP_ALLANY instead of
2382        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2383        on the OP_ANY path.
2384    
2385    17. Compiling pcretest on Windows with readline support failed without the
2386        following two fixes: (1) Make the unistd.h include conditional on
2387        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2388    
2389    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2390        ncurses library to be included for pcretest when ReadLine support is
2391        requested, but also to allow for it to be overridden. This patch came from
2392        Daniel Bergström.
2393    
2394    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2395        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2396        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2397        spotting this.
2398    
2399    
2400    Version 7.6 28-Jan-08
2401    ---------------------
2402    
2403    1.  A character class containing a very large number of characters with
2404        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2405        overflow.
2406    
2407    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2408        HAVE_LONG_LONG is not defined.
2409    
2410    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2411        bring them up to date and include new features. This patch includes:
2412    
2413        - Fixed PH's badly added libz and libbz2 support.
2414        - Fixed a problem with static linking.
2415        - Added pcredemo. [But later removed - see 7 below.]
2416        - Fixed dftables problem and added an option.
2417        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2418            HAVE_LONG_LONG.
2419        - Added readline support for pcretest.
2420        - Added an listing of the option settings after cmake has run.
2421    
2422    4.  A user submitted a patch to Makefile that makes it easy to create
2423        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2424        Makefile.am that cause it to include this special target, without
2425        affecting anything else. Note that the same mingw target plus all
2426        the other distribution libraries and programs are now supported
2427        when configuring with CMake (see 6 below) instead of with
2428        Configure/Make.
2429    
2430    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2431        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2432        exported in the Windows port". It has not yet been confirmed that the patch
2433        solves the problem, but it does no harm.
2434    
2435    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2436        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2437        with CMake, and also correct the comment about stack recursion.
2438    
2439    7.  Remove the automatic building of pcredemo from the ./configure system and
2440        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2441        of a program that users should build themselves after PCRE is installed, so
2442        building it automatically is not really right. What is more, it gave
2443        trouble in some build environments.
2444    
2445    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2446    
2447    
2448    Version 7.5 10-Jan-08
2449    ---------------------
2450    
2451    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2452        values in parens when parsing an RE using the C++ wrapper."
2453    
2454    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2455        Characters greater than 255 were excluded from the class instead of being
2456        included.
2457    
2458    3.  The same bug as (2) above applied to negated POSIX classes such as
2459        [:^space:].
2460    
2461    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2462        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2463        I have changed it.
2464    
2465    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2466        first named subpattern) and a construct such as (?&a) would reference the
2467        first named subpattern whose name started with "a" (in other words, the
2468        length check was missing). Both these problems are fixed. "Subpattern name
2469        expected" is now given for (?&) (a zero-length name), and this patch also
2470        makes it give the same error for \k'' (previously it complained that that
2471        was a reference to a non-existent subpattern).
2472    
2473    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2474        this is right because (?- can be followed by option settings as well as by
2475        digits. I have, however, made the messages clearer.
2476    
2477    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2478        than the number used in the conditional) now cause a compile-time error.
2479        This is actually not compatible with Perl, which accepts such patterns, but
2480        treats the conditional as always being FALSE (as PCRE used to), but it
2481        seems to me that giving a diagnostic is better.
2482    
2483    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2484        and messages.
2485    
2486    9.  Fix two occurrences of "backslash" in comments that should have been
2487        "backspace".
2488    
2489    10. Remove two redundant lines of code that can never be obeyed (their function
2490        was moved elsewhere).
2491    
2492    11. The program that makes PCRE's Unicode character property table had a bug
2493        which caused it to generate incorrect table entries for sequences of
2494        characters that have the same character type, but are in different scripts.
2495        It amalgamated them into a single range, with the script of the first of
2496        them. In other words, some characters were in the wrong script. There were
2497        thirteen such cases, affecting characters in the following ranges:
2498    
2499          U+002b0 - U+002c1
2500          U+0060c - U+0060d
2501          U+0061e - U+00612
2502          U+0064b - U+0065e
2503          U+0074d - U+0076d
2504          U+01800 - U+01805
2505          U+01d00 - U+01d77
2506          U+01d9b - U+01dbf
2507          U+0200b - U+0200f
2508          U+030fc - U+030fe
2509          U+03260 - U+0327f
2510          U+0fb46 - U+0fbb1
2511          U+10450 - U+1049d
2512    
2513    12. The -o option (show only the matching part of a line) for pcregrep was not
2514        compatible with GNU grep in that, if there was more than one match in a
2515        line, it showed only the first of them. It now behaves in the same way as
2516        GNU grep.
2517    
2518    13. If the -o and -v options were combined for pcregrep, it printed a blank
2519        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2520        does the same. The return code can be used to tell if there were any
2521        non-matching lines.
2522    
2523    14. Added --file-offsets and --line-offsets to pcregrep.
2524    
2525    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2526        infinitely looping recursion. The bug was that positive lookaheads were not
2527        being skipped when checking for a possible empty match (negative lookaheads
2528        and both kinds of lookbehind were skipped).
2529    
2530    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2531        inclusion of <windows.h> to before rather than after the definition of
2532        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2533    
2534    17. Specifying a possessive quantifier with a specific limit for a Unicode
2535        character property caused pcre_compile() to compile bad code, which led at
2536        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2537        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2538        caused the error; without that there was no problem.
2539    
2540    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2541    
2542    19. Added --enable-pcretest-libreadline.
2543    
2544    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2545        RE::GlobalReplace(). As a result, the number of replacements returned was
2546        double what it should be. I removed one of the increments, but Craig sent a
2547        later patch that removed the other one (the right fix) and added unit tests
2548        that check the return values (which was not done before).
2549    
2550    21. Several CMake things:
2551    
2552        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2553            the names libpcre and libpcreposix, not just pcre and pcreposix.
2554    
2555        (2) The above change means that pcretest and pcregrep are now correctly
2556            linked with the newly-built libraries, not previously installed ones.
2557    
2558        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2559    
2560    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2561        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2562        UTF-8 newline character). The key issue is that the pattern starts .*;
2563        this means that the match must be either at the beginning, or after a
2564        newline. The bug was in the code for advancing after a failed match and
2565        checking that the new position followed a newline. It was not taking
2566        account of UTF-8 characters correctly.
2567    
2568    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2569        character classes. PCRE was not treating the sequence [:...:] as a
2570        character class unless the ... were all letters. Perl, however, seems to
2571        allow any characters between [: and :], though of course it rejects as
2572        unknown any "names" that contain non-letters, because all the known class
2573        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2574        for example, whereas PCRE did not - it did not recognize a POSIX character
2575        class. This seemed a bit dangerous, so the code has been changed to be
2576        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2577        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2578        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2579        Perl does, and where it didn't before.
2580    
2581    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2582        Windows environments %n is disabled by default.
2583    
2584    
2585    Version 7.4 21-Sep-07
2586    ---------------------
2587    
2588    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2589        means that a class such as [\s] counted as "explicit reference to CR or
2590        LF". That isn't really right - the whole point of the change was to try to
2591        help when there was an actual mention of one of the two characters. So now
2592        the change happens only if \r or \n (or a literal CR or LF) character is
2593        encountered.
2594    
2595    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2596        of both had grown to the point where there were only 3 bits left.
2597        Fortunately, there was spare space in the data structure, and so I have
2598        moved the internal flags into a new 16-bit field to free up more option
2599        bits.
2600    
2601    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2602        but did not set the internal JCHANGED flag - either of these is enough to
2603        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2604        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2605        start sets both bits.
2606    
2607    4.  Added options (at build time, compile time, exec time) to change \R from
2608        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2609    
2610    5.  doc/pcresyntax.html was missing from the distribution.
2611    
2612    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2613        compatibility, even though it is no longer used.
2614    
2615    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2616        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2617        windows.h file is present (where different names are used). [This was
2618        reversed later after testing - see 16 below.]
2619    
2620    8.  Changed all #include <config.h> to #include "config.h". There were also
2621        some further <pcre.h> cases that I changed to "pcre.h".
2622    
2623    9.  When pcregrep was used with the --colour option, it missed the line ending
2624        sequence off the lines that it output.
2625    
2626    10. It was pointed out to me that arrays of string pointers cause lots of
2627        relocations when a shared library is dynamically loaded. A technique of
2628        using a single long string with a table of offsets can drastically reduce
2629        these. I have refactored PCRE in four places to do this. The result is
2630        dramatic:
2631    
2632          Originally:                          290
2633          After changing UCP table:            187
2634          After changing error message table:   43
2635          After changing table of "verbs"       36
2636          After changing table of Posix names   22
2637    
2638        Thanks to the folks working on Gregex for glib for this insight.
2639    
2640    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2641        unicode-properties was also set.
2642    
2643    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2644    
2645    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2646        checked only for CRLF.
2647    
2648    14. Added casts to pcretest.c to avoid compiler warnings.
2649    
2650    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2651    
2652    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2653        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2654        entirely. This removes changes made in 7 above.
2655    
2656    17. The CMake files have been updated, and there is now more information about
2657        building with CMake in the NON-UNIX-USE document.
2658    
2659    
2660    Version 7.3 28-Aug-07
2661    ---------------------
2662    
2663     1. In the rejigging of the build system that eventually resulted in 7.1, the
2664        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2665        brackets there is not right, since it causes compilers to look for an
2666        installed pcre.h, not the version that is in the source that is being
2667        compiled (which of course may be different). I have changed it back to:
2668    
2669          #include "pcre.h"
2670    
2671        I have a vague recollection that the change was concerned with compiling in
2672        different directories, but in the new build system, that is taken care of
2673        by the VPATH setting the Makefile.
2674    
2675     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2676        when the subject happened to end in the byte 0x85 (e.g. if the last
2677        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2678        characters but of course it shouldn't be taken as a newline when it is part
2679        of another character. The bug was that, for an unlimited repeat of . in
2680        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2681        characters when looking for a newline.
2682    
2683     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2684    
2685     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2686        in debug output.
2687    
2688     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2689        long printing in the pcrecpp unittest when running under MinGW.
2690    
2691     6. ESC_K was left out of the EBCDIC table.
2692    
2693     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2694        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2695        limit also applies to "virtual nesting" when a pattern is recursive, and in
2696        this case 1000 isn't so big. I have been able to remove this limit at the
2697        expense of backing off one optimization in certain circumstances. Normally,
2698        when pcre_exec() would call its internal match() function recursively and
2699        immediately return the result unconditionally, it uses a "tail recursion"
2700        feature to save stack. However, when a subpattern that can match an empty
2701        string has an unlimited repetition quantifier, it no longer makes this
2702        optimization. That gives it a stack frame in which to save the data for
2703        checking that an empty string has been matched. Previously this was taken
2704        from the 1000-entry workspace that had been reserved. So now there is no
2705        explicit limit, but more stack is used.
2706    
2707     8. Applied Daniel's patches to solve problems with the import/export magic
2708        syntax that is required for Windows, and which was going wrong for the
2709        pcreposix and pcrecpp parts of the library. These were overlooked when this
2710        problem was solved for the main library.
2711    
2712     9. There were some crude static tests to avoid integer overflow when computing
2713        the size of patterns that contain repeated groups with explicit upper
2714        limits. As the maximum quantifier is 65535, the maximum group length was
2715        set at 30,000 so that the product of these two numbers did not overflow a
2716        32-bit integer. However, it turns out that people want to use groups that
2717        are longer than 30,000 bytes (though not repeat them that many times).
2718        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2719        made it possible to implement the integer overflow checks in a much more
2720        dynamic way, which I have now done. The artificial limitation on group
2721        length has been removed - we now have only the limit on the total length of
2722        the compiled pattern, which depends on the LINK_SIZE setting.
2723    
2724    10. Fixed a bug in the documentation for get/copy named substring when
2725        duplicate names are permitted. If none of the named substrings are set, the
2726        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2727        empty string.
2728    
2729    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2730        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2731        because the ] is interpreted as the first data character and the
2732        terminating ] is not found. PCRE has been made compatible with Perl in this
2733        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2734        cause memory overwriting.
2735    
2736    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2737        string has been matched (to stop an infinite loop). It was not recognizing
2738        a conditional subpattern that could match an empty string if that
2739        subpattern was within another subpattern. For example, it looped when
2740        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2741        condition was not nested. This bug has been fixed.
2742    
2743    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2744        past the start of the subject in the presence of bytes with the top bit
2745        set, for example "\x8aBCD".
2746    
2747    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2748        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2749    
2750    14. Optimized (?!) to (*FAIL).
2751    
2752    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2753        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2754        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2755        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2756        does: it's just the validity check that is more restrictive.
2757    
2758    16. Inserted checks for integer overflows during escape sequence (backslash)
2759        processing, and also fixed erroneous offset values for syntax errors during
2760        backslash processing.
2761    
2762    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2763        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2764    
2765    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2766        caused an overrun.
2767    
2768    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2769        something other than just ASCII characters) inside a group that had an
2770        unlimited repeat caused a loop at compile time (while checking to see
2771        whether the group could match an empty string).
2772    
2773    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2774        [\P{Any}] did so. (Error in the code for printing property names.)
2775    
2776    21. An orphan \E inside a character class could cause a crash.
2777    
2778    22. A repeated capturing bracket such as (A)? could cause a wild memory
2779        reference during compilation.
2780    
2781    23. There are several functions in pcre_compile() that scan along a compiled
2782        expression for various reasons (e.g. to see if it's fixed length for look
2783        behind). There were bugs in these functions when a repeated \p or \P was
2784        present in the pattern. These operators have additional parameters compared
2785        with \d, etc, and these were not being taken into account when moving along
2786        the compiled data. Specifically:
2787    
2788        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2789            length.
2790    
2791        (b) An item such as \pL+ within a repeated group could cause crashes or
2792            loops.
2793    
2794        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2795            "reference to non-existent subpattern" error.
2796    
2797        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2798    
2799    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2800        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2801    
2802    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2803    
2804    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2805        character were causing crashes (broken optimization).
2806    
2807    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2808        \p or \P) caused a compile-time loop.
2809    
2810    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2811        break. For example, the unstudied pattern [\r\n]A does not match the string
2812        "\r\nA" because change 7.0/46 below moves the current point on by two
2813        characters after failing to match at the start. However, the pattern \nA
2814        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2815        the same is true. There doesn't seem any very clean way out of this, but
2816        what I have chosen to do makes the common cases work: PCRE now takes note
2817        of whether there can be an explicit match for \r or \n anywhere in the
2818        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2819        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2820        pattern has explicit CR or LF references.
2821    
2822    29. Added (*CR) etc for changing newline setting at start of pattern.
2823    
2824    
2825    Version 7.2 19-Jun-07
2826  ---------------------  ---------------------
2827    
2828   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
# Line 21  Version 7.2 01-May-07 Line 2842  Version 7.2 01-May-07
2842      stack recursion. This gives a massive performance boost under BSD, but just      stack recursion. This gives a massive performance boost under BSD, but just
2843      a small improvement under Linux. However, it saves one field in the frame      a small improvement under Linux. However, it saves one field in the frame
2844      in all cases.      in all cases.
2845    
2846   6. Added more features from the forthcoming Perl 5.10:   6. Added more features from the forthcoming Perl 5.10:
2847    
2848      (a) (?-n) (where n is a string of digits) is a relative subroutine or      (a) (?-n) (where n is a string of digits) is a relative subroutine or
2849          recursion call. It refers to the nth most recently opened parentheses.          recursion call. It refers to the nth most recently opened parentheses.
2850    
2851      (b) (?+n) is also a relative subroutine call; it refers to the nth next      (b) (?+n) is also a relative subroutine call; it refers to the nth next
2852          to be opened parentheses.          to be opened parentheses.
2853    
2854      (c) Conditions that refer to capturing parentheses can be specified      (c) Conditions that refer to capturing parentheses can be specified
2855          relatively, for example, (?(-2)... or (?(+3)...          relatively, for example, (?(-2)... or (?(+3)...
2856    
2857        (d) \K resets the start of the current match so that everything before
2858            is not part of it.
2859    
2860        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2861    
2862        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2863            reference syntax.
2864    
2865        (g) (?| introduces a group in which the numbering of parentheses in each
2866            alternative starts with the same number.
2867    
2868        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2869    
2870     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2871        PCRE_INFO_JCHANGED.
2872    
2873     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2874        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2875        for detecting groups that can match an empty string.
2876    
2877     9. A pattern with a very large number of alternatives (more than several
2878        hundred) was running out of internal workspace during the pre-compile
2879        phase, where pcre_compile() figures out how much memory will be needed. A
2880        bit of new cunning has reduced the workspace needed for groups with
2881        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2882        workspace instead of running out of the 4096 that are available.
2883    
2884    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2885    
2886    11. Applied patch from Google to remove an optimization that didn't quite work.
2887        The report of the bug said:
2888    
2889          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2890          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2891          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2892    
2893    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2894        it matched the wrong number of bytes.
2895    
2896    
2897  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
2898  ---------------------  ---------------------

Legend:
Removed from v.167  
changed lines
  Added in v.1515

  ViewVC Help
Powered by ViewVC 1.1.5