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

Diff of /code/trunk/ChangeLog

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

revision 127 by ph10, Mon Mar 19 11:44:45 2007 UTC revision 1514 by ph10, Tue Feb 3 14:34:44 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.1 12-Mar-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    
21    
22    Version 8.36 26-September-2014
23    ------------------------------
24    
25    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
26        -Wmissing-field-initializers and -Wunused-parameter.
27    
28    2.  The tests for quantifiers being too big (greater than 65535) were being
29        applied after reading the number, and stupidly assuming that integer
30        overflow would give a negative number. The tests are now applied as the
31        numbers are read.
32    
33    3.  Tidy code in pcre_exec.c where two branches that used to be different are
34        now the same.
35    
36    4.  The JIT compiler did not generate match limit checks for certain
37        bracketed expressions with quantifiers. This may lead to exponential
38        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
39        issue should be resolved now.
40    
41    5.  Fixed an issue, which occures when nested alternatives are optimized
42        with table jumps.
43    
44    6.  Inserted two casts and changed some ints to size_t in the light of some
45        reported 64-bit compiler warnings (Bugzilla 1477).
46    
47    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
48        an empty string, which sometimes were behaving incorrectly in the
49        interpreter (though correctly in the JIT matcher). This pcretest input is
50        an example:
51    
52          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
53          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
54    
55        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
56        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
57        for an empty string was breaking the inner loop and carrying on at a lower
58        level, when possessive repeated groups should always return to a higher
59        level as they have no backtrack points in them. The empty string test now
60        occurs at the outer level.
61    
62    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
63        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
64    
65    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
66        doesn't).
67    
68    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
69        to hold a pattern to be incorrectly computed (too small) when there were
70        named back references to duplicated names. This could cause "internal
71        error: code overflow" or "double free or corruption" or other memory
72        handling errors.
73    
74    11. When named subpatterns had the same prefixes, back references could be
75        confused. For example, in this pattern:
76    
77          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
78    
79        the reference to 'Name' was incorrectly treated as a reference to a
80        duplicate name.
81    
82    12. A pattern such as /^s?c/mi8 where the optional character has more than
83        one "other case" was incorrectly compiled such that it would only try to
84        match starting at "c".
85    
86    13. When a pattern starting with \s was studied, VT was not included in the
87        list of possible starting characters; this should have been part of the
88        8.34/18 patch.
89    
90    14. If a character class started [\Qx]... where x is any character, the class
91        was incorrectly terminated at the ].
92    
93    15. If a pattern that started with a caseless match for a character with more
94        than one "other case" was studied, PCRE did not set up the starting code
95        unit bit map for the list of possible characters. Now it does. This is an
96        optimization improvement, not a bug fix.
97    
98    16. The Unicode data tables have been updated to Unicode 7.0.0.
99    
100    17. Fixed a number of memory leaks in pcregrep.
101    
102    18. Avoid a compiler warning (from some compilers) for a function call with
103        a cast that removes "const" from an lvalue by using an intermediate
104        variable (to which the compiler does not object).
105    
106    19. Incorrect code was compiled if a group that contained an internal recursive
107        back reference was optional (had quantifier with a minimum of zero). This
108        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
109        caused segmentation faults because of stack overflows at compile time.
110    
111    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
112        group that is quantified with an indefinite repeat, caused a compile-time
113        loop which used up all the system stack and provoked a segmentation fault.
114        This was not the same bug as 19 above.
115    
116    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
117        Patch by Mike Frysinger.
118    
119    
120    Version 8.35 04-April-2014
121    --------------------------
122    
123    1.  A new flag is set, when property checks are present in an XCLASS.
124        When this flag is not set, PCRE can perform certain optimizations
125        such as studying these XCLASS-es.
126    
127    2.  The auto-possessification of character sets were improved: a normal
128        and an extended character set can be compared now. Furthermore
129        the JIT compiler optimizes more character set checks.
130    
131    3.  Got rid of some compiler warnings for potentially uninitialized variables
132        that show up only when compiled with -O2.
133    
134    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
135        of a match later then the end of the match. The pcretest program was not
136        handling the case sensibly - it was outputting from the start to the next
137        binary zero. It now reports this situation in a message, and outputs the
138        text from the end to the start.
139    
140    5.  Fast forward search is improved in JIT. Instead of the first three
141        characters, any three characters with fixed position can be searched.
142        Search order: first, last, middle.
143    
144    6.  Improve character range checks in JIT. Characters are read by an inprecise
145        function now, which returns with an unknown value if the character code is
146        above a certain threshold (e.g: 256). The only limitation is that the value
147        must be bigger than the threshold as well. This function is useful when
148        the characters above the threshold are handled in the same way.
149    
150    7.  The macros whose names start with RAWUCHAR are placeholders for a future
151        mode in which only the bottom 21 bits of 32-bit data items are used. To
152        make this more memorable for those maintaining the code, the names have
153        been changed to start with UCHAR21, and an extensive comment has been added
154        to their definition.
155    
156    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
157        to the export list in Makefile.am (they were accidentally omitted from the
158        8.34 tarball).
159    
160    9.  The informational output from pcretest used the phrase "starting byte set"
161        which is inappropriate for the 16-bit and 32-bit libraries. As the output
162        for "first char" and "need char" really means "non-UTF-char", I've changed
163        "byte" to "char", and slightly reworded the output. The documentation about
164        these values has also been (I hope) clarified.
165    
166    10. Another JIT related optimization: use table jumps for selecting the correct
167        backtracking path, when more than four alternatives are present inside a
168        bracket.
169    
170    11. Empty match is not possible, when the minimum length is greater than zero,
171        and there is no \K in the pattern. JIT should avoid empty match checks in
172        such cases.
173    
174    12. In a caseless character class with UCP support, when a character with more
175        than one alternative case was not the first character of a range, not all
176        the alternative cases were added to the class. For example, s and \x{17f}
177        are both alternative cases for S: the class [RST] was handled correctly,
178        but [R-T] was not.
179    
180    13. The configure.ac file always checked for pthread support when JIT was
181        enabled. This is not used in Windows, so I have put this test inside a
182        check for the presence of windows.h (which was already tested for).
183    
184    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
185        The algorithm provides a way to skip certain starting offsets, and usually
186        faster than linear prefix searches.
187    
188    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
189        as for 'fr_FR' and 'french'. For some reason, however, it then used the
190        Windows-specific input and output files, which have 'french' screwed in.
191        So this could never have worked. One of the problems with locales is that
192        they aren't always the same. I have now updated RunTest so that it checks
193        the output of the locale test (test 3) against three different output
194        files, and it allows the test to pass if any one of them matches. With luck
195        this should make the test pass on some versions of Solaris where it was
196        failing. Because of the uncertainty, the script did not used to stop if
197        test 3 failed; it now does. If further versions of a French locale ever
198        come to light, they can now easily be added.
199    
200    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
201        there was a message during ./configure, but it did not stop. This now
202        provokes an error. The invalid example in README has been corrected.
203        If a value less than the minimum is given, the minimum value has always
204        been used, but now a warning is given.
205    
206    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
207        was a bug in the test system, which is now fixed. Also, the list of various
208        configurations that are tested for each release did not have one with both
209        16/32 bits and --enable-bar-anycrlf. It now does.
210    
211    18. pcretest was missing "-C bsr" for displaying the \R default setting.
212    
213    19. Little endian PowerPC systems are supported now by the JIT compiler.
214    
215    20. The fast forward newline mechanism could enter to an infinite loop on
216        certain invalid UTF-8 input. Although we don't support these cases
217        this issue can be fixed by a performance optimization.
218    
219    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
220        not take account if existing stack usage. There is now a new global
221        variable called pcre_stack_guard that can be set to point to an external
222        function to check stack availability. It is called at the start of
223        processing every parenthesized group.
224    
225    22. A typo in the code meant that in ungreedy mode the max/min qualifier
226        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
227        not match "ab".
228    
229    23. When UTF was disabled, the JIT program reported some incorrect compile
230        errors. These messages are silenced now.
231    
232    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
233        compiler.
234    
235    25. Change all the temporary files used in RunGrepTest to be different to those
236        used by RunTest so that the tests can be run simultaneously, for example by
237        "make -j check".
238    
239    
240    Version 8.34 15-December-2013
241    -----------------------------
242    
243    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
244        executable memory. Patch inspired by Carsten Klein.
245    
246    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
247        this macro is never tested and has no effect, because the work to support
248        coverage involves only compiling and linking options and special targets in
249        the Makefile. The comment in config.h implied that defining the macro would
250        enable coverage support, which is totally false. There was also support for
251        setting this macro in the CMake files (my fault, I just copied it from
252        configure). SUPPORT_GCOV has now been removed.
253    
254    3.  Make a small performance improvement in strlen16() and strlen32() in
255        pcretest.
256    
257    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
258        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
259        these cases). There was also one in pcretest.c.
260    
261    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
262    
263    6.  In UTF mode, the code for checking whether a group could match an empty
264        string (which is used for indefinitely repeated groups to allow for
265        breaking an infinite loop) was broken when the group contained a repeated
266        negated single-character class with a character that occupied more than one
267        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
268        UTF-8 mode). The effect was undefined: the group might or might not be
269        deemed as matching an empty string, or the program might have crashed.
270    
271    7.  The code for checking whether a group could match an empty string was not
272        recognizing that \h, \H, \v, \V, and \R must match a character.
273    
274    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
275        an empty string. If it can, pcretest shows this in its information output.
276    
277    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
278        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
279        matched by pcre_exec() without using JIT:
280    
281        (a) If the rest of the pattern did not match after a maximal run of
282            grapheme clusters, the code for backing up to try with fewer of them
283            did not always back up over a full grapheme when characters that do not
284            have the modifier quality were involved, e.g. Hangul syllables.
285    
286        (b) If the match point in a subject started with modifier character, and
287            there was no match, the code could incorrectly back up beyond the match
288            point, and potentially beyond the first character in the subject,
289            leading to a segfault or an incorrect match result.
290    
291    10. A conditional group with an assertion condition could lead to PCRE
292        recording an incorrect first data item for a match if no other first data
293        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
294        first data item, and therefore matched "ca" after "c" instead of at the
295        start.
296    
297    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
298        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
299    
300    12. The source of pcregrep now includes z/OS-specific code so that it can be
301        compiled for z/OS as part of the special z/OS distribution.
302    
303    13. Added the -T and -TM options to pcretest.
304    
305    14. The code in pcre_compile.c for creating the table of named capturing groups
306        has been refactored. Instead of creating the table dynamically during the
307        actual compiling pass, the information is remembered during the pre-compile
308        pass (on the stack unless there are more than 20 named groups, in which
309        case malloc() is used) and the whole table is created before the actual
310        compile happens. This has simplified the code (it is now nearly 150 lines
311        shorter) and prepared the way for better handling of references to groups
312        with duplicate names.
313    
314    15. A back reference to a named subpattern when there is more than one of the
315        same name now checks them in the order in which they appear in the pattern.
316        The first one that is set is used for the reference. Previously only the
317        first one was inspected. This change makes PCRE more compatible with Perl.
318    
319    16. Unicode character properties were updated from Unicode 6.3.0.
320    
321    17. The compile-time code for auto-possessification has been refactored, based
322        on a patch by Zoltan Herczeg. It now happens after instead of during
323        compilation. The code is cleaner, and more cases are handled. The option
324        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
325        options in pcretest are provided to set it. It can also be set by
326        (*NO_AUTO_POSSESS) at the start of a pattern.
327    
328    18. The character VT has been added to the default ("C" locale) set of
329        characters that match \s and are generally treated as white space,
330        following this same change in Perl 5.18. There is now no difference between
331        "Perl space" and "POSIX space". Whether VT is treated as white space in
332        other locales depends on the locale.
333    
334    19. The code for checking named groups as conditions, either for being set or
335        for being recursed, has been refactored (this is related to 14 and 15
336        above). Processing unduplicated named groups should now be as fast at
337        numerical groups, and processing duplicated groups should be faster than
338        before.
339    
340    20. Two patches to the CMake build system, by Alexander Barkov:
341    
342          (1) Replace the "source" command by "." in CMakeLists.txt because
343              "source" is a bash-ism.
344    
345          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
346              without these the CMake build does not work on Solaris.
347    
348    21. Perl has changed its handling of \8 and \9. If there is no previously
349        encountered capturing group of those numbers, they are treated as the
350        literal characters 8 and 9 instead of a binary zero followed by the
351        literals. PCRE now does the same.
352    
353    22. Following Perl, added \o{} to specify codepoints in octal, making it
354        possible to specify values greater than 0777 and also making them
355        unambiguous.
356    
357    23. Perl now gives an error for missing closing braces after \x{... instead of
358        treating the string as literal. PCRE now does the same.
359    
360    24. RunTest used to grumble if an inappropriate test was selected explicitly,
361        but just skip it when running all tests. This make it awkward to run ranges
362        of tests when one of them was inappropriate. Now it just skips any
363        inappropriate tests, as it always did when running all tests.
364    
365    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
366        character types such as \d or \w, too many callouts were inserted, and the
367        data that they returned was rubbish.
368    
369    26. In UCP mode, \s was not matching two of the characters that Perl matches,
370        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
371        were matched by \h. The code has now been refactored so that the lists of
372        the horizontal and vertical whitespace characters used for \h and \v (which
373        are defined only in one place) are now also used for \s.
374    
375    27. Add JIT support for the 64 bit TileGX architecture.
376        Patch by Jiong Wang (Tilera Corporation).
377    
378    28. Possessive quantifiers for classes (both explicit and automatically
379        generated) now use special opcodes instead of wrapping in ONCE brackets.
380    
381    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
382        (because it's meaningless), this was not happening when PCRE_CASELESS was
383        set. Not wrong, but inefficient.
384    
385    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
386        properties for \w, \d, etc) is present in a test regex. Otherwise if the
387        test contains no characters greater than 255, Perl doesn't realise it
388        should be using Unicode semantics.
389    
390    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
391        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
392        does in Unicode mode.
393    
394    32. Added the "forbid" facility to pcretest so that putting tests into the
395        wrong test files can sometimes be quickly detected.
396    
397    33. There is now a limit (default 250) on the depth of nesting of parentheses.
398        This limit is imposed to control the amount of system stack used at compile
399        time. It can be changed at build time by --with-parens-nest-limit=xxx or
400        the equivalent in CMake.
401    
402    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
403        errors. Perl warns for these when in warning mode, but PCRE has no facility
404        for giving warnings.
405    
406    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
407        However, this was not working for (?!) because it is optimized to (*FAIL),
408        for which PCRE does not allow quantifiers. The optimization is now disabled
409        when a quantifier follows (?!). I can't see any use for this, but it makes
410        things uniform.
411    
412    36. Perl no longer allows group names to start with digits, so I have made this
413        change also in PCRE. It simplifies the code a bit.
414    
415    37. In extended mode, Perl ignores spaces before a + that indicates a
416        possessive quantifier. PCRE allowed a space before the quantifier, but not
417        before the possessive +. It now does.
418    
419    38. The use of \K (reset reported match start) within a repeated possessive
420        group such as (a\Kb)*+ was not working.
421    
422    40. Document that the same character tables must be used at compile time and
423        run time, and that the facility to pass tables to pcre_exec() and
424        pcre_dfa_exec() is for use only with saved/restored patterns.
425    
426    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
427        features for Builds with MSVC:
428    
429        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
430           builds). The .rc files can be used to set FileDescription and many other
431           attributes.
432    
433        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
434           This allows higher-level build scripts which want .pdb files to avoid
435           hard-coding the exact files needed."
436    
437    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
438        mean "start of word" and "end of word", respectively, as a transition aid.
439    
440    43. A minimizing repeat of a class containing codepoints greater than 255 in
441        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
442        compiled to use the heap for recursion.
443    
444    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
445        is configured.
446    
447    
448    Version 8.33 28-May-2013
449    ------------------------
450    
451    1.  Added 'U' to some constants that are compared to unsigned integers, to
452        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
453        variables that are added to signed variables, to ensure the result is
454        signed and can be negated.
455    
456    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
457        CMake config files.
458    
459    3.  Revise the creation of config.h.generic so that all boolean macros are
460        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
461        overriding via -D on the command line possible.
462    
463    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
464        to unsigned int is reported to make a quite noticeable speed difference in
465        a specific Windows environment. Testing on Linux did also appear to show
466        some benefit (and it is clearly not harmful). Also fixed the definition of
467        Xop which should be unsigned.
468    
469    5.  Related to (4), changing the definition of the intermediate variable cc
470        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
471        improvements.
472    
473    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
474        unnecessary spaces.
475    
476    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
477        and later.
478    
479    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
480    
481    9.  Optimizing fast_forward_start_bits in JIT.
482    
483    10. Adding support for callouts in JIT, and fixing some issues revealed
484        during this work. Namely:
485    
486        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
487    
488        (b) Minimum length was not checked before the matching is started.
489    
490    11. The value of capture_last that is passed to callouts was incorrect in some
491        cases when there was a capture on one path that was subsequently abandoned
492        after a backtrack. Also, the capture_last value is now reset after a
493        recursion, since all captures are also reset in this case.
494    
495    12. The interpreter no longer returns the "too many substrings" error in the
496        case when an overflowing capture is in a branch that is subsequently
497        abandoned after a backtrack.
498    
499    13. In the pathological case when an offset vector of size 2 is used, pcretest
500        now prints out the matched string after a yield of 0 or 1.
501    
502    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
503        Only supported by the JIT compiler at the moment.
504    
505    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
506    
507    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
508        offset of the starting point of the matching process, provided the offsets
509        vector is large enough.
510    
511    17. The \A escape now records a lookbehind value of 1, though its execution
512        does not actually inspect the previous character. This is to ensure that,
513        in partial multi-segment matching, at least one character from the old
514        segment is retained when a new segment is processed. Otherwise, if there
515        are no lookbehinds in the pattern, \A might match incorrectly at the start
516        of a new segment.
517    
518    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
519    
520    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
521        gives some modest performance improvement in 8-bit mode.
522    
523    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
524        be expressed in certain programming languages using Universal Character
525        Names.
526    
527    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
528        which points out that "non characters" are not "characters that may not
529        appear in Unicode strings" but rather "characters that are reserved for
530        internal use and have only local meaning".
531    
532    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
533        there was a conditional group that depended on an assertion, if the
534        assertion was false, the callout that immediately followed the alternation
535        in the condition was skipped when pcre_exec() was used for matching.
536    
537    23. Allow an explicit callout to be inserted before an assertion that is the
538        condition for a conditional group, for compatibility with automatic
539        callouts, which always insert a callout at this point.
540    
541    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
542        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
543    
544    25. (*PRUNE) is now supported by the JIT compiler.
545    
546    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
547    
548    27. Fix the case where there are two or more SKIPs with arguments that may be
549        ignored.
550    
551    28. (*SKIP) is now supported by the JIT compiler.
552    
553    29. (*THEN) is now supported by the JIT compiler.
554    
555    30. Update RunTest with additional test selector options.
556    
557    31. The way PCRE handles backtracking verbs has been changed in two ways.
558    
559        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
560        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
561        backtracking. In some cases this makes it more Perl-compatible, but Perl's
562        rather obscure rules do not always do the same thing.
563    
564        (2) Previously, backtracking verbs were confined within assertions. This is
565        no longer the case for positive assertions, except for (*ACCEPT). Again,
566        this sometimes improves Perl compatibility, and sometimes does not.
567    
568    32. A number of tests that were in test 2 because Perl did things differently
569        have been moved to test 1, because either Perl or PCRE has changed, and
570        these tests are now compatible.
571    
572    32. Backtracking control verbs are now handled in the same way in JIT and
573        interpreter.
574    
575    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
576        contained a forward subroutine reference caused a compile error.
577    
578    34. Auto-detect and optimize limited repetitions in JIT.
579    
580    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
581        blocking (*UTF) etc.
582    
583    36. In the interpreter, maximizing pattern repetitions for characters and
584        character types now use tail recursion, which reduces stack usage.
585    
586    37. The value of the max lookbehind was not correctly preserved if a compiled
587        and saved regex was reloaded on a host of different endianness.
588    
589    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
590        of the compiled pattern block, expand the flags field from 16 to 32 bits
591        because it was almost full.
592    
593    39. Try madvise first before posix_madvise.
594    
595    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
596        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
597        original change locked out all matches of empty strings. This has been
598        changed so that one match of an empty string per line is recognized.
599        Subsequent searches on the same line (for colouring or for --only-matching,
600        for example) do not recognize empty strings.
601    
602    41. Applied a user patch to fix a number of spelling mistakes in comments.
603    
604    42. Data lines longer than 65536 caused pcretest to crash.
605    
606    43. Clarified the data type for length and startoffset arguments for pcre_exec
607        and pcre_dfa_exec in the function-specific man pages, where they were
608        explicitly stated to be in bytes, never having been updated. I also added
609        some clarification to the pcreapi man page.
610    
611    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
612        a segmentation fault.
613    
614    
615    Version 8.32 30-November-2012
616    -----------------------------
617    
618    1.  Improved JIT compiler optimizations for first character search and single
619        character iterators.
620    
621    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
622        Patch by Daniel Richard G.
623    
624    3.  Single character iterator optimizations in the JIT compiler.
625    
626    4.  Improved JIT compiler optimizations for character ranges.
627    
628    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
629        Reported by Giuseppe D'Angelo.
630    
631    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
632        of a line, was being set incorrectly in cases where .* appeared inside
633        atomic brackets at the start of a pattern, or where there was a subsequent
634        *PRUNE or *SKIP.
635    
636    7.  Improved instruction cache flush for POWER/PowerPC.
637        Patch by Daniel Richard G.
638    
639    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
640        grep:
641    
642        (a) There is now no limit to the number of patterns to be matched.
643    
644        (b) An error is given if a pattern is too long.
645    
646        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
647            are now supported.
648    
649        (d) --exclude-from and --include-from (multiple use) have been added.
650    
651        (e) Exclusions and inclusions now apply to all files and directories, not
652            just to those obtained from scanning a directory recursively.
653    
654        (f) Multiple uses of -f and --file-list are now supported.
655    
656        (g) In a Windows environment, the default for -d has been changed from
657            "read" (the GNU grep default) to "skip", because otherwise the presence
658            of a directory in the file list provokes an error.
659    
660        (h) The documentation has been revised and clarified in places.
661    
662    9.  Improve the matching speed of capturing brackets.
663    
664    10. Changed the meaning of \X so that it now matches a Unicode extended
665        grapheme cluster.
666    
667    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
668        out POSIX threads when JIT support is configured.
669    
670    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
671    
672    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
673        structure could go wrong in environments where size_t is not the same size
674        as int.
675    
676    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
677        set.
678    
679    15. The EBCDIC support had decayed; later updates to the code had included
680        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
681        general tidy up of EBCDIC-related issues, and the documentation was also
682        not quite right. There is now a test that can be run on ASCII systems to
683        check some of the EBCDIC-related things (but is it not a full test).
684    
685    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
686        in a small tidy to the code.
687    
688    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
689    
690    18. If the --only-matching (-o) option in pcregrep is specified multiple
691        times, each one causes appropriate output. For example, -o1 -o2 outputs the
692        substrings matched by the 1st and 2nd capturing parentheses. A separating
693        string can be specified by --om-separator (default empty).
694    
695    19. Improving the first n character searches.
696    
697    20. Turn case lists for horizontal and vertical white space into macros so that
698        they are defined only once.
699    
700    21. This set of changes together give more compatible Unicode case-folding
701        behaviour for characters that have more than one other case when UCP
702        support is available.
703    
704        (a) The Unicode property table now has offsets into a new table of sets of
705            three or more characters that are case-equivalent. The MultiStage2.py
706            script that generates these tables (the pcre_ucd.c file) now scans
707            CaseFolding.txt instead of UnicodeData.txt for character case
708            information.
709    
710        (b) The code for adding characters or ranges of characters to a character
711            class has been abstracted into a generalized function that also handles
712            case-independence. In UTF-mode with UCP support, this uses the new data
713            to handle characters with more than one other case.
714    
715        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
716            whose other case is greater than 256 are now correctly matched
717            caselessly. Previously, the high codepoint matched the low one, but not
718            vice versa.
719    
720        (d) The processing of \h, \H, \v, and \ in character classes now makes use
721            of the new class addition function, using character lists defined as
722            macros alongside the case definitions of 20 above.
723    
724        (e) Caseless back references now work with characters that have more than
725            one other case.
726    
727        (f) General caseless matching of characters with more than one other case
728            is supported.
729    
730    22. Unicode character properties were updated from Unicode 6.2.0
731    
732    23. Improved CMake support under Windows. Patch by Daniel Richard G.
733    
734    24. Add support for 32-bit character strings, and UTF-32
735    
736    25. Major JIT compiler update (code refactoring and bugfixing).
737        Experimental Sparc 32 support is added.
738    
739    26. Applied a modified version of Daniel Richard G's patch to create
740        pcre.h.generic and config.h.generic by "make" instead of in the
741        PrepareRelease script.
742    
743    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
744        pcre_compile.c when checking for a zero character.
745    
746    28. Introducing a native interface for JIT. Through this interface, the compiled
747        machine code can be directly executed. The purpose of this interface is to
748        provide fast pattern matching, so several sanity checks are not performed.
749        However, feature tests are still performed. The new interface provides
750        1.4x speedup compared to the old one.
751    
752    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
753        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
754        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
755    
756    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
757        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
758        the "old" RFC 2279). Instead, it ended up passing a negative length to
759        pcre_exec().
760    
761    31. Add support for GCC's visibility feature to hide internal functions.
762    
763    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
764        "unknown -C option" after outputting 0 or 1.
765    
766    33. There is now support for generating a code coverage report for the test
767        suite in environments where gcc is the compiler and lcov is installed. This
768        is mainly for the benefit of the developers.
769    
770    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
771        unaddressable using valgrind annotations, allowing valgrind to detect
772        invalid memory accesses. This is mainly for the benefit of the developers.
773    
774    25. (*UTF) can now be used to start a pattern in any of the three libraries.
775    
776    26. Give configure error if --enable-cpp but no C++ compiler found.
777    
778    
779    Version 8.31 06-July-2012
780    -------------------------
781    
782    1.  Fixing a wrong JIT test case and some compiler warnings.
783    
784    2.  Removed a bashism from the RunTest script.
785    
786    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
787        to unsigned type, result still unsigned" that was given by an MS compiler
788        on encountering the code "-sizeof(xxx)".
789    
790    4.  Partial matching support is added to the JIT compiler.
791    
792    5.  Fixed several bugs concerned with partial matching of items that consist
793        of more than one character:
794    
795        (a) /^(..)\1/ did not partially match "aba" because checking references was
796            done on an "all or nothing" basis. This also applied to repeated
797            references.
798    
799        (b) \R did not give a hard partial match if \r was found at the end of the
800            subject.
801    
802        (c) \X did not give a hard partial match after matching one or more
803            characters at the end of the subject.
804    
805        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
806            a partial match for the string "\r".
807    
808        (e) When newline was set to CRLF, the metacharacter "." did not recognize
809            a partial match for a CR character at the end of the subject string.
810    
811    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
812        running pcretest, the text "(JIT)" added to the output whenever JIT is
813        actually used to run the match.
814    
815    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
816        or /S+[+] with a digit between 1 and 7.
817    
818    8.  OP_NOT now supports any UTF character not just single-byte ones.
819    
820    9.  (*MARK) control verb is now supported by the JIT compiler.
821    
822    10. The command "./RunTest list" lists the available tests without actually
823        running any of them. (Because I keep forgetting what they all are.)
824    
825    11. Add PCRE_INFO_MAXLOOKBEHIND.
826    
827    12. Applied a (slightly modified) user-supplied patch that improves performance
828        when the heap is used for recursion (compiled with --disable-stack-for-
829        recursion). Instead of malloc and free for each heap frame each time a
830        logical recursion happens, frames are retained on a chain and re-used where
831        possible. This sometimes gives as much as 30% improvement.
832    
833    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
834        call.
835    
836    14. As documented, (*COMMIT) is now confined to within a positive assertion.
837    
838    15. It is now possible to link pcretest with libedit as an alternative to
839        libreadline.
840    
841    16. (*COMMIT) control verb is now supported by the JIT compiler.
842    
843    17. The Unicode data tables have been updated to Unicode 6.1.0.
844    
845    18. Added --file-list option to pcregrep.
846    
847    19. Added binary file support to pcregrep, including the -a, --binary-files,
848        -I, and --text options.
849    
850    20. The madvise function is renamed for posix_madvise for QNX compatibility
851        reasons. Fixed by Giuseppe D'Angelo.
852    
853    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
854    
855    22. Changed the option for creating symbolic links for 16-bit man pages from
856        -s to -sf so that re-installing does not cause issues.
857    
858    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
859    
860    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
861        matches in certain environments (the workspace was not being correctly
862        retained). Also added to pcre_dfa_exec() a simple plausibility check on
863        some of the workspace data at the beginning of a restart.
864    
865    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
866        was not doing so when it should - probably a typo introduced by SVN 528
867        (change 8.10/14).
868    
869    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
870        \w+ when the character tables indicated that \x{c4} was a word character.
871        There were several related cases, all because the tests for doing a table
872        lookup were testing for characters less than 127 instead of 255.
873    
874    27. If a pattern contains capturing parentheses that are not used in a match,
875        their slots in the ovector are set to -1. For those that are higher than
876        any matched groups, this happens at the end of processing. In the case when
877        there were back references that the ovector was too small to contain
878        (causing temporary malloc'd memory to be used during matching), and the
879        highest capturing number was not used, memory off the end of the ovector
880        was incorrectly being set to -1. (It was using the size of the temporary
881        memory instead of the true size.)
882    
883    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
884        ovector size, it uses memory at the end of the block that it has got.
885    
886    29. Check for an overlong MARK name and give an error at compile time. The
887        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
888    
889    30. JIT compiler update.
890    
891    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
892        Rill for the patch.
893    
894    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
895    
896    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
897    
898    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
899        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
900        was enabled.
901    
902    35. Improve JIT code generation for greedy plus quantifier.
903    
904    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
905        1 to "aa" instead of to an empty string. The bug affected repeated groups
906        that could potentially match an empty string.
907    
908    37. Optimizing single character iterators in JIT.
909    
910    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
911        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
912        codepoints that are too big for the mode are faulted, and in a UTF mode,
913        disallowed codepoints are also faulted.
914    
915    39. If PCRE was compiled with UTF support, in three places in the DFA
916        matcher there was code that should only have been obeyed in UTF mode, but
917        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
918        processing when bytes with values greater than 127 were present. In 16-bit
919        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
920        both cases the values are those that cannot be the first data item in a UTF
921        character. The three items that might have provoked this were recursions,
922        possessively repeated groups, and atomic groups.
923    
924    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
925        and pcregrep, because some OS require shared objects to be explicitly
926        passed to ld, causing the link step to fail if they are not.
927    
928    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
929        mode, patterns that started with \h* or \R* might be incorrectly matched.
930    
931    
932    Version 8.30 04-February-2012
933    -----------------------------
934    
935    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
936        name is defined in ctype.h.
937    
938    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
939        only in quite long subpatterns.
940    
941    3.  Removed the function pcre_info(), which has been obsolete and deprecated
942        since it was replaced by pcre_fullinfo() in February 2000.
943    
944    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
945        match a (*MARK), and the match failed at the start of the subject, a
946        reference to memory before the start of the subject could occur. This bug
947        was introduced by fix 17 of release 8.21.
948    
949    5.  A reference to an unset group with zero minimum repetition was giving
950        totally wrong answers (in non-JavaScript-compatibility mode). For example,
951        /(another)?(\1?)test/ matched against "hello world test". This bug was
952        introduced in release 8.13.
953    
954    6.  Add support for 16-bit character strings (a large amount of work involving
955        many changes and refactorings).
956    
957    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
958        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
959        from a file.
960    
961    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
962        rounding is not applied in this particular case).
963    
964    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
965        if they appear, or are escaped, in patterns.
966    
967    10. Get rid of a number of -Wunused-but-set-variable warnings.
968    
969    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
970        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
971        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
972        also returns the mark "x". This bug applied to capturing parentheses,
973        non-capturing parentheses, and atomic parentheses. It also applied to some
974        assertions.
975    
976    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
977        information out of configure.ac instead of relying on pcre.h.generic, which
978        is not stored in the repository.
979    
980    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
981        -lreadline.
982    
983    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
984    
985    15. Applied Graycode's patch to put the top-level frame on the stack rather
986        than the heap when not using the stack for recursion. This gives a
987        performance improvement in many cases when recursion is not deep.
988    
989    16. Experimental code added to "pcretest -C" to output the stack frame size.
990    
991    
992    Version 8.21 12-Dec-2011
993    ------------------------
994    
995    1.  Updating the JIT compiler.
996    
997    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
998        are added as well.
999    
1000    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1001        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1002        calling _pcre_jit_exec. Some extra comments are added.
1003    
1004    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1005        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1006        was introduced by change 18 for 8.20.
1007    
1008    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1009        ECMA-262 standard.
1010    
1011    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1012        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1013        This bug was probably introduced by change 9 of 8.13.
1014    
1015    7.  While fixing 6 above, I noticed that a number of other items were being
1016        incorrectly rejected as "not fixed length". This arose partly because newer
1017        opcodes had not been added to the fixed-length checking code. I have (a)
1018        corrected the bug and added tests for these items, and (b) arranged for an
1019        error to occur if an unknown opcode is encountered while checking for fixed
1020        length instead of just assuming "not fixed length". The items that were
1021        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1022        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1023        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1024    
1025    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1026        being incorrectly compiled and would have given unpredicatble results.
1027    
1028    9.  A possessively repeated subpattern with minimum repeat count greater than
1029        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1030        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1031        the first (A) could occur when it should not.
1032    
1033    10. Add a cast and remove a redundant test from the code.
1034    
1035    11. JIT should use pcre_malloc/pcre_free for allocation.
1036    
1037    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1038        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1039        is anything other than /usr, -L is still shown).
1040    
1041    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1042    
1043    14. Perl does not support \N without a following name in a [] class; PCRE now
1044        also gives an error.
1045    
1046    15. If a forward reference was repeated with an upper limit of around 2000,
1047        it caused the error "internal error: overran compiling workspace". The
1048        maximum number of forward references (including repeats) was limited by the
1049        internal workspace, and dependent on the LINK_SIZE. The code has been
1050        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1051        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1052        of around 200,000 forward references. While doing this, I also speeded up
1053        the filling in of repeated forward references.
1054    
1055    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1056        incorrectly expecting the subject to contain another "a" after the start.
1057    
1058    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1059        in the match, the SKIP should be ignored. This was not happening; instead
1060        the SKIP was being treated as NOMATCH. For patterns such as
1061        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1062        tested.
1063    
1064    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1065        now much more compatible with Perl, in particular in cases where the result
1066        is a non-match for a non-anchored pattern. For example, if
1067        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1068        "m", where previously it did not return a name. A side effect of this
1069        change is that for partial matches, the last encountered mark name is
1070        returned, as for non matches. A number of tests that were previously not
1071        Perl-compatible have been moved into the Perl-compatible test files. The
1072        refactoring has had the pleasing side effect of removing one argument from
1073        the match() function, thus reducing its stack requirements.
1074    
1075    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1076        subsequent uses of /S (without +) incorrectly behaved like /S+.
1077    
1078    21. Retrieve executable code size support for the JIT compiler and fixing
1079        some warnings.
1080    
1081    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1082        not work when the shorter character appeared right at the end of the
1083        subject string.
1084    
1085    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1086        systems.
1087    
1088    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1089        output it when the /M option is used in pcretest.
1090    
1091    25. The CheckMan script was not being included in the distribution. Also, added
1092        an explicit "perl" to run Perl scripts from the PrepareRelease script
1093        because this is reportedly needed in Windows.
1094    
1095    26. If study data was being save in a file and studying had not found a set of
1096        "starts with" bytes for the pattern, the data written to the file (though
1097        never used) was taken from uninitialized memory and so caused valgrind to
1098        complain.
1099    
1100    27. Updated RunTest.bat as provided by Sheri Pierce.
1101    
1102    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1103    
1104    29. Computation of memory usage for the table of capturing group names was
1105        giving an unnecessarily large value.
1106    
1107    
1108    Version 8.20 21-Oct-2011
1109    ------------------------
1110    
1111    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1112        a POSIX class. After further experiments with Perl, which convinced me that
1113        Perl has bugs and confusions, a closing square bracket is no longer allowed
1114        in a POSIX name. This bug also affected patterns with classes that started
1115        with full stops.
1116    
1117    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1118        captured substring, but while checking the failing first alternative,
1119        substring 1 is temporarily captured. If the output vector supplied to
1120        pcre_exec() was not big enough for this capture, the yield of the function
1121        was still zero ("insufficient space for captured substrings"). This cannot
1122        be totally fixed without adding another stack variable, which seems a lot
1123        of expense for a edge case. However, I have improved the situation in cases
1124        such as /(a)(b)x|abc/ matched against "abc", where the return code
1125        indicates that fewer than the maximum number of slots in the ovector have
1126        been set.
1127    
1128    3.  Related to (2) above: when there are more back references in a pattern than
1129        slots in the output vector, pcre_exec() uses temporary memory during
1130        matching, and copies in the captures as far as possible afterwards. It was
1131        using the entire output vector, but this conflicts with the specification
1132        that only 2/3 is used for passing back captured substrings. Now it uses
1133        only the first 2/3, for compatibility. This is, of course, another edge
1134        case.
1135    
1136    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1137        main code base, and can be used by building with --enable-jit. When this is
1138        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1139        runtime --no-jit option is given.
1140    
1141    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1142        ovector, the return from the function was zero, implying that there were
1143        other matches that did not fit. The correct "exactly full" value is now
1144        returned.
1145    
1146    6.  If a subpattern that was called recursively or as a subroutine contained
1147        (*PRUNE) or any other control that caused it to give a non-standard return,
1148        invalid errors such as "Error -26 (nested recursion at the same subject
1149        position)" or even infinite loops could occur.
1150    
1151    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1152        computing the minimum length on reaching *ACCEPT, and so ended up with the
1153        wrong value of 1 rather than 0. Further investigation indicates that
1154        computing a minimum subject length in the presence of *ACCEPT is difficult
1155        (think back references, subroutine calls), and so I have changed the code
1156        so that no minimum is registered for a pattern that contains *ACCEPT.
1157    
1158    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1159        it was not handled as intended. [But see 16 below.]
1160    
1161    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1162        Sheri Pierce.
1163    
1164    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1165        the first byte in a match must be "a".
1166    
1167    11. Change 17 for 8.13 increased the recursion depth for patterns like
1168        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1169        pattern contains any instances of (*THEN). If it does not, the old
1170        optimizations are restored. It would be nice to do this on a per-group
1171        basis, but at the moment that is not feasible.
1172    
1173    12. In some environments, the output of pcretest -C is CRLF terminated. This
1174        broke RunTest's code that checks for the link size. A single white space
1175        character after the value is now allowed for.
1176    
1177    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1178        For "fr", it uses the Windows-specific input and output files.
1179    
1180    14. If (*THEN) appeared in a group that was called recursively or as a
1181        subroutine, it did not work as intended. [But see next item.]
1182    
1183    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1184        pattern fragments (but not containing any | characters). If A and B are
1185        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1186        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1187        D. In other words, Perl considers parentheses that do not contain any |
1188        characters to be part of a surrounding alternative, whereas PCRE was
1189        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1190        differently. PCRE now behaves in the same way as Perl, except in the case
1191        of subroutine/recursion calls such as (?1) which have in any case always
1192        been different (but PCRE had them first :-).
1193    
1194    16. Related to 15 above: Perl does not treat the | in a conditional group as
1195        creating alternatives. Such a group is treated in the same way as an
1196        ordinary group without any | characters when processing (*THEN). PCRE has
1197        been changed to match Perl's behaviour.
1198    
1199    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1200        RunGrepTest script failed.
1201    
1202    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1203        inevitable for groups that contain captures, but it can lead to a lot of
1204        stack use in large patterns. The old behaviour has been restored for atomic
1205        groups that do not contain any capturing parentheses.
1206    
1207    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1208        suppress the check for a minimum subject length at run time. (If it was
1209        given to pcre_exec() or pcre_dfa_exec() it did work.)
1210    
1211    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1212        fail to work when decoding hex characters in data strings in EBCDIC
1213        environments.
1214    
1215    21. It appears that in at least one Mac OS environment, the isxdigit() function
1216        is implemented as a macro that evaluates to its argument more than once,
1217        contravening the C 90 Standard (I haven't checked a later standard). There
1218        was an instance in pcretest which caused it to go wrong when processing
1219        \x{...} escapes in subject strings. The has been rewritten to avoid using
1220        things like p++ in the argument of isxdigit().
1221    
1222    
1223    Version 8.13 16-Aug-2011
1224    ------------------------
1225    
1226    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1227    
1228    2.  Two minor typos in pcre_internal.h have been fixed.
1229    
1230    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1231        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1232        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1233    
1234    4.  There were a number of related bugs in the code for matching backrefences
1235        caselessly in UTF-8 mode when codes for the characters concerned were
1236        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1237        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1238        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1239        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1240        2-byte code at the end of the subject (it thought there wasn't enough data
1241        left).
1242    
1243    5.  Comprehensive information about what went wrong is now returned by
1244        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1245        as the output vector has at least 2 elements. The offset of the start of
1246        the failing character and a reason code are placed in the vector.
1247    
1248    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1249        now returned is for the first byte of the failing character, instead of the
1250        last byte inspected. This is an incompatible change, but I hope it is small
1251        enough not to be a problem. It makes the returned offset consistent with
1252        pcre_exec() and pcre_dfa_exec().
1253    
1254    7.  pcretest now gives a text phrase as well as the error number when
1255        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1256        failure, the offset and reason code are output.
1257    
1258    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1259        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1260        back over a single character (\n). This seems wrong (because it treated the
1261        two characters as a single entity when going forwards), conflicts with the
1262        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1263        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1264        has been changed.
1265    
1266    9.  Some internal refactoring has changed the processing so that the handling
1267        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1268        time (the PCRE_DOTALL option was changed this way some time ago: version
1269        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1270        which was always a bit of a fudge. It also means that there is one less
1271        argument for the match() function, which reduces its stack requirements
1272        slightly. This change also fixes an incompatibility with Perl: the pattern
1273        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1274    
1275    10. More internal refactoring has drastically reduced the number of recursive
1276        calls to match() for possessively repeated groups such as (abc)++ when
1277        using pcre_exec().
1278    
1279    11. While implementing 10, a number of bugs in the handling of groups were
1280        discovered and fixed:
1281    
1282        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1283        (a|)*(?1) gave a compile-time internal error.
1284        ((a|)+)+  did not notice that the outer group could match an empty string.
1285        (^a|^)+   was not marked as anchored.
1286        (.*a|.*)+ was not marked as matching at start or after a newline.
1287    
1288    12. Yet more internal refactoring has removed another argument from the match()
1289        function. Special calls to this function are now indicated by setting a
1290        value in a variable in the "match data" data block.
1291    
1292    13. Be more explicit in pcre_study() instead of relying on "default" for
1293        opcodes that mean there is no starting character; this means that when new
1294        ones are added and accidentally left out of pcre_study(), testing should
1295        pick them up.
1296    
1297    14. The -s option of pcretest has been documented for ages as being an old
1298        synonym of -m (show memory usage). I have changed it to mean "force study
1299        for every regex", that is, assume /S for every regex. This is similar to -i
1300        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1301        using it. It makes it easier to run collections of tests with and without
1302        study enabled, and thereby test pcre_study() more easily. All the standard
1303        tests are now run with and without -s (but some patterns can be marked as
1304        "never study" - see 20 below).
1305    
1306    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1307        restoration of the capturing data to the outer values was not happening
1308        correctly.
1309    
1310    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1311        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1312        pattern had matched an empty string, and so incorrectly returned a no
1313        match.
1314    
1315    17. There was optimizing code for the last branch of non-capturing parentheses,
1316        and also for the obeyed branch of a conditional subexpression, which used
1317        tail recursion to cut down on stack usage. Unfortunately, now that there is
1318        the possibility of (*THEN) occurring in these branches, tail recursion is
1319        no longer possible because the return has to be checked for (*THEN). These
1320        two optimizations have therefore been removed. [But see 8.20/11 above.]
1321    
1322    18. If a pattern containing \R was studied, it was assumed that \R always
1323        matched two bytes, thus causing the minimum subject length to be
1324        incorrectly computed because \R can also match just one byte.
1325    
1326    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1327        was incorrectly computed.
1328    
1329    20. If /S is present twice on a test pattern in pcretest input, it now
1330        *disables* studying, thereby overriding the use of -s on the command line
1331        (see 14 above). This is necessary for one or two tests to keep the output
1332        identical in both cases.
1333    
1334    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1335        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1336    
1337    22. When an atomic group that contained a capturing parenthesis was
1338        successfully matched, but the branch in which it appeared failed, the
1339        capturing was not being forgotten if a higher numbered group was later
1340        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1341        group 1 to "a", when in fact it should be unset. This applied to multi-
1342        branched capturing and non-capturing groups, repeated or not, and also to
1343        positive assertions (capturing in negative assertions does not happen
1344        in PCRE) and also to nested atomic groups.
1345    
1346    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1347        subject after a captured substring, to make it easier to tell which of a
1348        number of identical substrings has been captured.
1349    
1350    24. The way atomic groups are processed by pcre_exec() has been changed so that
1351        if they are repeated, backtracking one repetition now resets captured
1352        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1353        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1354        "aaa". Previously, it would have been "a". As part of this code
1355        refactoring, the way recursive calls are handled has also been changed.
1356    
1357    25. If an assertion condition captured any substrings, they were not passed
1358        back unless some other capturing happened later. For example, if
1359        (?(?=(a))a) was matched against "a", no capturing was returned.
1360    
1361    26. When studying a pattern that contained subroutine calls or assertions,
1362        the code for finding the minimum length of a possible match was handling
1363        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1364        group 1 called group 2 while simultaneously a separate group 2 called group
1365        1). A stack overflow occurred in this case. I have fixed this by limiting
1366        the recursion depth to 10.
1367    
1368    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1369        Fortmann. This supports explicit test numbers on the command line, and has
1370        argument validation and error reporting.
1371    
1372    28. An instance of \X with an unlimited repeat could fail if at any point the
1373        first character it looked at was a mark character.
1374    
1375    29. Some minor code refactoring concerning Unicode properties and scripts
1376        should reduce the stack requirement of match() slightly.
1377    
1378    30. Added the '=' option to pcretest to check the setting of unused capturing
1379        slots at the end of the pattern, which are documented as being -1, but are
1380        not included in the return count.
1381    
1382    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1383        compiled something random. Now it gives a compile-time error (as does
1384        Perl).
1385    
1386    32. A *MARK encountered during the processing of a positive assertion is now
1387        recorded and passed back (compatible with Perl).
1388    
1389    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1390        had alternative anchored branches, the search for a second match in a line
1391        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1392        matched the line "0102" twice. The same bug affected patterns that started
1393        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1394        twice.
1395    
1396    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1397        does, and because of capturing effects, quantifying parenthesized
1398        assertions may at times be useful. Quantifiers are now allowed for
1399        parenthesized assertions.
1400    
1401    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1402    
1403    36. \g was being checked for fancy things in a character class, when it should
1404        just be a literal "g".
1405    
1406    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1407        appearance of a nested POSIX class supersedes an apparent external class.
1408        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1409        unescaped square brackets may also appear as part of class names. For
1410        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1411        more like Perl. (But see 8.20/1 above.)
1412    
1413    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1414        was because it thought it was \N{name}, which is not supported).
1415    
1416    39. Add minix to OS list not supporting the -S option in pcretest.
1417    
1418    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1419        cannot analyze patterns in sufficient detail to catch mutual recursions
1420        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1421        subgroup is called recursively as a subpattern for a second time at the
1422        same position in the subject string. In previous releases this might have
1423        been caught by the recursion limit, or it might have run out of stack.
1424    
1425    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1426        happen only once. PCRE was, however incorrectly giving a compile time error
1427        "recursive call could loop indefinitely" because it cannot analyze the
1428        pattern in sufficient detail. The compile time test no longer happens when
1429        PCRE is compiling a conditional subpattern, but actual runaway loops are
1430        now caught at runtime (see 40 above).
1431    
1432    42. It seems that Perl allows any characters other than a closing parenthesis
1433        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1434        has been changed to be the same.
1435    
1436    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1437        as not to get warnings when autogen.sh is called. Also changed
1438        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1439    
1440    44. To help people who use pcregrep to scan files containing exceedingly long
1441        lines, the following changes have been made:
1442    
1443        (a) The default value of the buffer size parameter has been increased from
1444            8K to 20K. (The actual buffer used is three times this size.)
1445    
1446        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1447            PCRE is built.
1448    
1449        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1450            to be set at run time.
1451    
1452        (d) Numerical values in pcregrep options can be followed by K or M, for
1453            example --buffer-size=50K.
1454    
1455        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1456            given and the return code is set to 2.
1457    
1458    45. Add a pointer to the latest mark to the callout data block.
1459    
1460    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1461        partial match of an empty string instead of no match. This was specific to
1462        the use of ".".
1463    
1464    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1465        complete match instead of a partial match. This bug was dependent on both
1466        the PCRE_UTF8 and PCRE_DOTALL options being set.
1467    
1468    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1469        starting byte set, because \b was not being ignored.
1470    
1471    
1472    Version 8.12 15-Jan-2011
1473    ------------------------
1474    
1475    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1476        checks for such things as part of the documentation building process.
1477    
1478    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1479        --match-limit and --recursion-limit options (added for 8.11). In
1480        particular, this made one of the standard tests fail. (The integer value
1481        went into the wrong half of a long int.)
1482    
1483    3.  If the --colour option was given to pcregrep with -v (invert match), it
1484        did strange things, either producing crazy output, or crashing. It should,
1485        of course, ignore a request for colour when reporting lines that do not
1486        match.
1487    
1488    4.  Another pcregrep bug caused similar problems if --colour was specified with
1489        -M (multiline) and the pattern match finished with a line ending.
1490    
1491    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1492        matched in multiline mode, the following line was shown as part of the
1493        match. This seems wrong, so I have changed it.
1494    
1495    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1496        the check for further matches in the same line (so they could be coloured)
1497        to overrun the end of the current line. If another match was found, it was
1498        incorrectly shown (and then shown again when found in the next line).
1499    
1500    7.  If pcregrep was compiled under Windows, there was a reference to the
1501        function pcregrep_exit() before it was defined. I am assuming this was
1502        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1503        reported by a user. I've moved the definition above the reference.
1504    
1505    
1506    Version 8.11 10-Dec-2010
1507    ------------------------
1508    
1509    1.  (*THEN) was not working properly if there were untried alternatives prior
1510        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1511        backtracked to try for "b" instead of moving to the next alternative branch
1512        at the same level (in this case, to look for "c"). The Perl documentation
1513        is clear that when (*THEN) is backtracked onto, it goes to the "next
1514        alternative in the innermost enclosing group".
1515    
1516    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1517        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1518        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1519        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1520        (*THEN).
1521    
1522    3.  If \s appeared in a character class, it removed the VT character from
1523        the class, even if it had been included by some previous item, for example
1524        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1525        of \s, but is part of the POSIX "space" class.)
1526    
1527    4.  A partial match never returns an empty string (because you can always
1528        match an empty string at the end of the subject); however the checking for
1529        an empty string was starting at the "start of match" point. This has been
1530        changed to the "earliest inspected character" point, because the returned
1531        data for a partial match starts at this character. This means that, for
1532        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1533        (previously it gave "no match").
1534    
1535    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1536        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1537        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1538        has an implication that the given string is incomplete (because a partial
1539        match is preferred over a full match). For this reason, these items now
1540        give a partial match in this situation. [Aside: previously, the one case
1541        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1542        match rather than a full match, which was wrong by the old rules, but is
1543        now correct.]
1544    
1545    6.  There was a bug in the handling of #-introduced comments, recognized when
1546        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1547        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1548        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1549        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1550        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1551        places in pcre_compile().
1552    
1553    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1554        comments when looking ahead for named forward references to subpatterns,
1555        the only newline sequence it recognized was NL. It now handles newlines
1556        according to the set newline convention.
1557    
1558    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1559        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1560        cater for a lack of strerror(). These oversights have been fixed.
1561    
1562    9.  Added --match-limit and --recursion-limit to pcregrep.
1563    
1564    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1565    
1566    11. When the -o option was used, pcregrep was setting a return code of 1, even
1567        when matches were found, and --line-buffered was not being honoured.
1568    
1569    12. Added an optional parentheses number to the -o and --only-matching options
1570        of pcregrep.
1571    
1572    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1573        can match an empty string. The code to do it in pcretest and pcredemo
1574        needed fixing:
1575    
1576        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1577            only one byte after an empty string match just before CRLF (this case
1578            just got forgotten; "any" and "anycrlf" were OK).
1579    
1580        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1581            mode when an empty string match preceded an ASCII character followed by
1582            a non-ASCII character. (The code for advancing by one character rather
1583            than one byte was nonsense.)
1584    
1585        (c) The pcredemo.c sample program did not have any code at all to handle
1586            the cases when CRLF is a valid newline sequence.
1587    
1588    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1589        as a starting offset was within the subject string. There is now a new
1590        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1591        negative or greater than the length of the string. In order to test this,
1592        pcretest is extended to allow the setting of negative starting offsets.
1593    
1594    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1595        starting offset points to the beginning of a UTF-8 character was
1596        unnecessarily clumsy. I tidied it up.
1597    
1598    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1599        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1600    
1601    17. Nobody had reported that the --include_dir option, which was added in
1602        release 7.7 should have been called --include-dir (hyphen, not underscore)
1603        for compatibility with GNU grep. I have changed it to --include-dir, but
1604        left --include_dir as an undocumented synonym, and the same for
1605        --exclude-dir, though that is not available in GNU grep, at least as of
1606        release 2.5.4.
1607    
1608    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1609        characters from a string of bytes) have been redefined so as not to use
1610        loops, in order to improve performance in some environments. At the same
1611        time, I abstracted some of the common code into auxiliary macros to save
1612        repetition (this should not affect the compiled code).
1613    
1614    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1615        compile-time error is now given if \c is not followed by an ASCII
1616        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1617        different, and any byte value is allowed.)
1618    
1619    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1620        START_OPTIMIZE option, which is now allowed at compile time - but just
1621        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1622        to pcregrep and other applications that have no direct access to PCRE
1623        options. The new /Y option in pcretest sets this option when calling
1624        pcre_compile().
1625    
1626    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1627        back references. Groups containing recursive back references were forced to
1628        be atomic by that change, but in the case of named groups, the amount of
1629        memory required was incorrectly computed, leading to "Failed: internal
1630        error: code overflow". This has been fixed.
1631    
1632    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1633        pcretest.c, to avoid build problems in some Borland environments.
1634    
1635    
1636    Version 8.10 25-Jun-2010
1637    ------------------------
1638    
1639    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1640        THEN.
1641    
1642    2.  (*ACCEPT) was not working when inside an atomic group.
1643    
1644    3.  Inside a character class, \B is treated as a literal by default, but
1645        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1646        causes the error). The code is unchanged, but I tidied the documentation.
1647    
1648    4.  Inside a character class, PCRE always treated \R and \X as literals,
1649        whereas Perl faults them if its -w option is set. I have changed PCRE so
1650        that it faults them when PCRE_EXTRA is set.
1651    
1652    5.  Added support for \N, which always matches any character other than
1653        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1654    
1655    6.  When compiling pcregrep with newer versions of gcc which may have
1656        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1657        declared with attribute warn_unused_result" were given. Just casting the
1658        result to (void) does not stop the warnings; a more elaborate fudge is
1659        needed. I've used a macro to implement this.
1660    
1661    7.  Minor change to pcretest.c to avoid a compiler warning.
1662    
1663    8.  Added four artifical Unicode properties to help with an option to make
1664        \s etc use properties (see next item). The new properties are: Xan
1665        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1666    
1667    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1668        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1669        this option. Modified pcretest to add /W to test this facility. Added
1670        REG_UCP to make it available via the POSIX interface.
1671    
1672    10. Added --line-buffered to pcregrep.
1673    
1674    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1675        studied, and the match started with a letter with a code point greater than
1676        127 whose first byte was different to the first byte of the other case of
1677        the letter, the other case of this starting letter was not recognized
1678        (#976).
1679    
1680    12. If a pattern that was studied started with a repeated Unicode property
1681        test, for example, \p{Nd}+, there was the theoretical possibility of
1682        setting up an incorrect bitmap of starting bytes, but fortunately it could
1683        not have actually happened in practice until change 8 above was made (it
1684        added property types that matched character-matching opcodes).
1685    
1686    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1687        possible starting bytes for non-anchored patterns.
1688    
1689    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1690        \R, and also a number of cases that involve Unicode properties, both
1691        explicit and implicit when PCRE_UCP is set.
1692    
1693    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1694        input, it could crash or give wrong results if characters with values
1695        greater than 0xc0 were present in the subject string. (Detail: it assumed
1696        UTF-8 input when processing these items.)
1697    
1698    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1699        size_t is 64-bit (#991).
1700    
1701    17. Added a check for running out of memory when PCRE is compiled with
1702        --disable-stack-for-recursion (#990).
1703    
1704    18. If the last data line in a file for pcretest does not have a newline on
1705        the end, a newline was missing in the output.
1706    
1707    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1708        less than 128) in its various bitmaps. However, there is a facility for
1709        generating tables according to the current locale when PCRE is compiled. It
1710        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1711        characters, are recognized by isspace() and therefore were getting set in
1712        these tables, and indeed these tables seem to approximate to ISO 8859. This
1713        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1714        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1715        which of course cannot start UTF-8 characters. I have changed the code so
1716        that only real ASCII characters (less than 128) and the correct starting
1717        bytes for UTF-8 encodings are set for characters greater than 127 when in
1718        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1719        altogether.)
1720    
1721    20. Added the /T option to pcretest so as to be able to run tests with non-
1722        standard character tables, thus making it possible to include the tests
1723        used for 19 above in the standard set of tests.
1724    
1725    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1726        reference to a subpattern the other side of a comment that contains an
1727        opening parenthesis caused either an internal compiling error, or a
1728        reference to the wrong subpattern.
1729    
1730    
1731    Version 8.02 19-Mar-2010
1732    ------------------------
1733    
1734    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1735    
1736    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1737        configured.
1738    
1739    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1740        original author of that file, following a query about its status.
1741    
1742    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1743        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1744    
1745    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1746        quantifier applied to a forward-referencing subroutine call, could compile
1747        incorrect code or give the error "internal error: previously-checked
1748        referenced subpattern not found".
1749    
1750    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1751        variables to point to external functions. For these systems, therefore,
1752        pcre_malloc etc. are now initialized to local functions that call the
1753        relevant global functions.
1754    
1755    7.  There were two entries missing in the vectors called coptable and poptable
1756        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1757        I've fixed the data, and added a kludgy way of testing at compile time that
1758        the lengths are correct (equal to the number of opcodes).
1759    
1760    8.  Following on from 7, I added a similar kludge to check the length of the
1761        eint vector in pcreposix.c.
1762    
1763    9.  Error texts for pcre_compile() are held as one long string to avoid too
1764        much relocation at load time. To find a text, the string is searched,
1765        counting zeros. There was no check for running off the end of the string,
1766        which could happen if a new error number was added without updating the
1767        string.
1768    
1769    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1770    
1771    11. \K was not working if it appeared in an atomic group or in a group that
1772        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1773        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1774        the assertion is positive, but not if it is negative.
1775    
1776    12. Change 11 fortuitously reduced the size of the stack frame used in the
1777        "match()" function of pcre_exec.c by one pointer. Forthcoming
1778        implementation of support for (*MARK) will need an extra pointer on the
1779        stack; I have reserved it now, so that the stack frame size does not
1780        decrease.
1781    
1782    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1783        item in branch that calls a recursion is a subroutine call - as in the
1784        second branch in the above example - was incorrectly given the compile-
1785        time error "recursive call could loop indefinitely" because pcre_compile()
1786        was not correctly checking the subroutine for matching a non-empty string.
1787    
1788    14. The checks for overrunning compiling workspace could trigger after an
1789        overrun had occurred. This is a "should never occur" error, but it can be
1790        triggered by pathological patterns such as hundreds of nested parentheses.
1791        The checks now trigger 100 bytes before the end of the workspace.
1792    
1793    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1794    
1795    
1796    Version 8.01 19-Jan-2010
1797    ------------------------
1798    
1799    1.  If a pattern contained a conditional subpattern with only one branch (in
1800        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1801        computed the wrong minimum data length (which is of course zero for such
1802        subpatterns). This could cause incorrect "no match" results.
1803    
1804    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1805        the pattern is reset in the first branch, pcre_compile() failed with
1806        "internal error: code overflow at offset...". This happened only when
1807        the reset was to the original external option setting. (An optimization
1808        abstracts leading options settings into an external setting, which was the
1809        cause of this.)
1810    
1811    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1812        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1813        assertion pattern did not match (meaning that the assertion was true), it
1814        was incorrectly treated as false if the SKIP had been reached during the
1815        matching. This also applied to assertions used as conditions.
1816    
1817    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1818        assertion subpattern, including such a pattern used as a condition,
1819        unpredictable results occurred, instead of the error return
1820        PCRE_ERROR_DFA_UITEM.
1821    
1822    5.  The C++ GlobalReplace function was not working like Perl for the special
1823        situation when an empty string is matched. It now does the fancy magic
1824        stuff that is necessary.
1825    
1826    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1827        removed. (These were left over from very, very early versions of PCRE.)
1828    
1829    7.  Some cosmetic changes to the code to make life easier when compiling it
1830        as part of something else:
1831    
1832        (a) Change DEBUG to PCRE_DEBUG.
1833    
1834        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1835            called "current" as "current_branch", to prevent a collision with the
1836            Linux macro when compiled as a kernel module.
1837    
1838        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1839            prevent a collision with the Linux macro when compiled as a kernel
1840            module.
1841    
1842    8.  In pcre_compile() there are some checks for integer overflows that used to
1843        cast potentially large values to (double). This has been changed to that
1844        when building, a check for int64_t is made, and if it is found, it is used
1845        instead, thus avoiding the use of floating point arithmetic. (There is no
1846        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1847        double.
1848    
1849    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1850        2005 (difference between two addresses compared to an unsigned value).
1851    
1852    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1853        custom one, because of the following reported problem in Windows:
1854    
1855          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1856              under Win32.
1857          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1858              therefore missing the function definition.
1859          - The compiler thus generates a "C" signature for the test function.
1860          - The linker fails to find the "C" function.
1861          - PCRE fails to configure if asked to do so against libbz2.
1862    
1863    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1864        messages were output:
1865    
1866          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1867          rerunning libtoolize, to keep the correct libtool macros in-tree.
1868          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1869    
1870        I have done both of these things.
1871    
1872    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1873        most of the time, it *can* run out if it is given a pattern that contains a
1874        runaway infinite recursion. I updated the discussion in the pcrestack man
1875        page.
1876    
1877    13. Now that we have gone to the x.xx style of version numbers, the minor
1878        version may start with zero. Using 08 or 09 is a bad idea because users
1879        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1880        interpreted as invalid octal numbers. I've updated the previous comment in
1881        configure.ac, and also added a check that gives an error if 08 or 09 are
1882        used.
1883    
1884    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1885        causing partial matching to fail when the end of the subject matched \W
1886        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1887    
1888    15. There were some discrepancies between the declarations in pcre_internal.h
1889        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1890        their definitions. The declarations used "const uschar *" and the
1891        definitions used USPTR. Even though USPTR is normally defined as "const
1892        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1893        reported that: "This difference in casting confuses some C++ compilers, for
1894        example, SunCC recognizes above declarations as different functions and
1895        generates broken code for hbpcre." I have changed the declarations to use
1896        USPTR.
1897    
1898    16. GNU libtool is named differently on some systems. The autogen.sh script now
1899        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1900        (FreeBSD).
1901    
1902    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1903        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1904        comment: "Figure out how to create a longlong from a string: strtoll and
1905        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1906        instance, but it only takes 2 args instead of 3!"
1907    
1908    18. A subtle bug concerned with back references has been fixed by a change of
1909        specification, with a corresponding code fix. A pattern such as
1910        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1911        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1912        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1913        same bug. Such groups have to be quantified to be useful, or contained
1914        inside another quantified group. (If there's no repetition, the reference
1915        can never match.) The problem arises because, having left the group and
1916        moved on to the rest of the pattern, a later failure that backtracks into
1917        the group uses the captured value from the final iteration of the group
1918        rather than the correct earlier one. I have fixed this in PCRE by forcing
1919        any group that contains a reference to itself to be an atomic group; that
1920        is, there cannot be any backtracking into it once it has completed. This is
1921        similar to recursive and subroutine calls.
1922    
1923    
1924    Version 8.00 19-Oct-09
1925    ----------------------
1926    
1927    1.  The table for translating pcre_compile() error codes into POSIX error codes
1928        was out-of-date, and there was no check on the pcre_compile() error code
1929        being within the table. This could lead to an OK return being given in
1930        error.
1931    
1932    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1933        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1934        in a Windows environment.
1935    
1936    3.  The pcregrep --count option prints the count for each file even when it is
1937        zero, as does GNU grep. However, pcregrep was also printing all files when
1938        --files-with-matches was added. Now, when both options are given, it prints
1939        counts only for those files that have at least one match. (GNU grep just
1940        prints the file name in this circumstance, but including the count seems
1941        more useful - otherwise, why use --count?) Also ensured that the
1942        combination -clh just lists non-zero counts, with no names.
1943    
1944    4.  The long form of the pcregrep -F option was incorrectly implemented as
1945        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1946        but it seems right to fix it, and I didn't think it was worth preserving
1947        the old behaviour.
1948    
1949    5.  The command line items --regex=pattern and --regexp=pattern were not
1950        recognized by pcregrep, which required --regex pattern or --regexp pattern
1951        (with a space rather than an '='). The man page documented the '=' forms,
1952        which are compatible with GNU grep; these now work.
1953    
1954    6.  No libpcreposix.pc file was created for pkg-config; there was just
1955        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1956    
1957    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1958        when UCP support is not needed, by modifying the Python script that
1959        generates it from Unicode data files. This should not matter if the module
1960        is correctly used as a library, but I received one complaint about 50K of
1961        unwanted data. My guess is that the person linked everything into his
1962        program rather than using a library. Anyway, it does no harm.
1963    
1964    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1965        was a minimum greater than 1 for a wide character in a possessive
1966        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1967        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1968        character. Chaos in the form of incorrect output or a compiling loop could
1969        result.
1970    
1971    9.  The restrictions on what a pattern can contain when partial matching is
1972        requested for pcre_exec() have been removed. All patterns can now be
1973        partially matched by this function. In addition, if there are at least two
1974        slots in the offset vector, the offset of the earliest inspected character
1975        for the match and the offset of the end of the subject are set in them when
1976        PCRE_ERROR_PARTIAL is returned.
1977    
1978    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1979        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1980        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1981        and may be more useful for multi-segment matching.
1982    
1983    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1984        used to be given if ever the end of the subject was reached; now it is
1985        given only if matching could not proceed because another character was
1986        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1987        string "Z", which now yields "no match" instead of "partial match". In the
1988        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1989        final character ended with (*FAIL).
1990    
1991    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1992        if the pattern had a "must contain" character that was already found in the
1993        earlier partial match, unless partial matching was again requested. For
1994        example, with the pattern /dog.(body)?/, the "must contain" character is
1995        "g". If the first part-match was for the string "dog", restarting with
1996        "sbody" failed. This bug has been fixed.
1997    
1998    13. The string returned by pcre_dfa_exec() after a partial match has been
1999        changed so that it starts at the first inspected character rather than the
2000        first character of the match. This makes a difference only if the pattern
2001        starts with a lookbehind assertion or \b or \B (\K is not supported by
2002        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2003        matching functions compatible, and I think it's the right thing to do.
2004    
2005    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2006        so that the demonstration program is easily available in environments where
2007        PCRE has not been installed from source.
2008    
2009    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2010        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2011        library.
2012    
2013    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2014        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2015        is not the first non-POSIX option to be added. Clearly some people find
2016        these options useful.
2017    
2018    17. If a caller to the POSIX matching function regexec() passes a non-zero
2019        value for nmatch with a NULL value for pmatch, the value of
2020        nmatch is forced to zero.
2021    
2022    18. RunGrepTest did not have a test for the availability of the -u option of
2023        the diff command, as RunTest does. It now checks in the same way as
2024        RunTest, and also checks for the -b option.
2025    
2026    19. If an odd number of negated classes containing just a single character
2027        interposed, within parentheses, between a forward reference to a named
2028        subpattern and the definition of the subpattern, compilation crashed with
2029        an internal error, complaining that it could not find the referenced
2030        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2031        [The bug was that it was starting one character too far in when skipping
2032        over the character class, thus treating the ] as data rather than
2033        terminating the class. This meant it could skip too much.]
2034    
2035    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2036        /g option in pcretest when the pattern contains \K, which makes it possible
2037        to have an empty string match not at the start, even when the pattern is
2038        anchored. Updated pcretest and pcredemo to use this option.
2039    
2040    21. If the maximum number of capturing subpatterns in a recursion was greater
2041        than the maximum at the outer level, the higher number was returned, but
2042        with unset values at the outer level. The correct (outer level) value is
2043        now given.
2044    
2045    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2046        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2047        make it do so. The string so far is captured, making this feature
2048        compatible with Perl.
2049    
2050    23. The tests have been re-organized, adding tests 11 and 12, to make it
2051        possible to check the Perl 5.10 features against Perl 5.10.
2052    
2053    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2054        pattern matches a fixed length string. PCRE did not allow this; now it
2055        does. Neither allows recursion.
2056    
2057    25. I finally figured out how to implement a request to provide the minimum
2058        length of subject string that was needed in order to match a given pattern.
2059        (It was back references and recursion that I had previously got hung up
2060        on.) This code has now been added to pcre_study(); it finds a lower bound
2061        to the length of subject needed. It is not necessarily the greatest lower
2062        bound, but using it to avoid searching strings that are too short does give
2063        some useful speed-ups. The value is available to calling programs via
2064        pcre_fullinfo().
2065    
2066    26. While implementing 25, I discovered to my embarrassment that pcretest had
2067        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2068        study optimizations had never been tested with that matching function.
2069        Oops. What is worse, even when it was passed study data, there was a bug in
2070        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2071        were also very few tests of studied patterns with pcre_dfa_exec().
2072    
2073    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2074        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2075        on the other side of the coin, they are no longer allowed to have different
2076        names, because these cannot be distinguished in PCRE, and this has caused
2077        confusion. (This is a difference from Perl.)
2078    
2079    28. When duplicate subpattern names are present (necessarily with different
2080        numbers, as required by 27 above), and a test is made by name in a
2081        conditional pattern, either for a subpattern having been matched, or for
2082        recursion in such a pattern, all the associated numbered subpatterns are
2083        tested, and the overall condition is true if the condition is true for any
2084        one of them. This is the way Perl works, and is also more like the way
2085        testing by number works.
2086    
2087    
2088    Version 7.9 11-Apr-09
2089    ---------------------
2090    
2091    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2092        (pcretest), all targets were linked against these libraries. This included
2093        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2094        libraries. This caused unwanted dependencies to be created. This problem
2095        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2096        pcretest is linked with readline.
2097    
2098    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2099        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2100        moved outside it again, because FALSE and TRUE are already defined in AIX,
2101        but BOOL is not.
2102    
2103    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2104        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2105    
2106    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2107        hyphens) following file names and line numbers when outputting matching
2108        lines. This is not true; no spaces are inserted. I have also clarified the
2109        wording for the --colour (or --color) option.
2110    
2111    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2112        was not coloured; this is different to GNU grep, so I have changed it to be
2113        the same.
2114    
2115    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2116        each matching line was coloured. Now it goes on to look for further matches
2117        of any of the test patterns, which is the same behaviour as GNU grep.
2118    
2119    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2120        doesn't make sense to accept an empty string match in pcregrep, so I have
2121        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2122        seems to be how GNU grep behaves. [But see later change 40 for release
2123        8.33.]
2124    
2125    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2126        start or after a newline", because the conditional assertion was not being
2127        correctly handled. The rule now is that both the assertion and what follows
2128        in the first alternative must satisfy the test.
2129    
2130    9.  If auto-callout was enabled in a pattern with a conditional group whose
2131        condition was an assertion, PCRE could crash during matching, both with
2132        pcre_exec() and pcre_dfa_exec().
2133    
2134    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2135        used for matching.
2136    
2137    11. Unicode property support in character classes was not working for
2138        characters (bytes) greater than 127 when not in UTF-8 mode.
2139    
2140    12. Added the -M command line option to pcretest.
2141    
2142    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2143    
2144    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2145    
2146    16. Added comments and documentation about mis-use of no_arg in the C++
2147        wrapper.
2148    
2149    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2150        from Martin Jerabek that uses macro names for all relevant character and
2151        string constants.
2152    
2153    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2154        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2155        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2156        these, but not everybody uses configure.
2157    
2158    19. A conditional group that had only one branch was not being correctly
2159        recognized as an item that could match an empty string. This meant that an
2160        enclosing group might also not be so recognized, causing infinite looping
2161        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2162        with the subject "ab", where knowledge that the repeated group can match
2163        nothing is needed in order to break the loop.
2164    
2165    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2166        exec(), but without supplying a callout function, matching went wrong.
2167    
2168    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2169        leak if the size of the offset vector was greater than 30. When the vector
2170        is smaller, the saved offsets during recursion go onto a local stack
2171        vector, but for larger vectors malloc() is used. It was failing to free
2172        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2173        error, in fact).
2174    
2175    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2176        heapframe that is used only when UTF-8 support is enabled. This caused no
2177        problem, but was untidy.
2178    
2179    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2180        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2181        included within another project.
2182    
2183    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2184        slightly modified by me:
2185    
2186          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2187              not building pcregrep.
2188    
2189          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2190              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2191    
2192    25. Forward references, both numeric and by name, in patterns that made use of
2193        duplicate group numbers, could behave incorrectly or give incorrect errors,
2194        because when scanning forward to find the reference group, PCRE was not
2195        taking into account the duplicate group numbers. A pattern such as
2196        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2197    
2198    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2199        the feature of a custom pointer more persuasive (as requested by a user).
2200    
2201    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2202        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2203        pre-defined.
2204    
2205    28. Added support for (*UTF8) at the start of a pattern.
2206    
2207    29. Arrange for flags added by the "release type" setting in CMake to be shown
2208        in the configuration summary.
2209    
2210    
2211    Version 7.8 05-Sep-08
2212    ---------------------
2213    
2214    1.  Replaced UCP searching code with optimized version as implemented for Ad
2215        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2216        stage table and inline lookup instead of a function, giving speed ups of 2
2217        to 5 times on some simple patterns that I tested. Permission was given to
2218        distribute the MultiStage2.py script that generates the tables (it's not in
2219        the tarball, but is in the Subversion repository).
2220    
2221    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2222        scripts.
2223    
2224    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2225        a group with a zero qualifier. The result of the study could be incorrect,
2226        or the function might crash, depending on the pattern.
2227    
2228    4.  Caseless matching was not working for non-ASCII characters in back
2229        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2230        It now works when Unicode Property Support is available.
2231    
2232    5.  In pcretest, an escape such as \x{de} in the data was always generating
2233        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2234        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2235        truncation.
2236    
2237    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2238    
2239    7.  Added two (int) casts to pcregrep when printing the difference of two
2240        pointers, in case they are 64-bit values.
2241    
2242    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2243        test 2 if it fails.
2244    
2245    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2246        and a #define of that name to empty if it is not externally set. This is to
2247        allow users of MSVC to set it if necessary.
2248    
2249    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2250        the convenience functions in the pcre_get.c source file.
2251    
2252    11. An option change at the start of a pattern that had top-level alternatives
2253        could cause overwriting and/or a crash. This command provoked a crash in
2254        some environments:
2255    
2256          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2257    
2258        This potential security problem was recorded as CVE-2008-2371.
2259    
2260    12. For a pattern where the match had to start at the beginning or immediately
2261        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2262        pcre_dfa_exec() could read past the end of the passed subject if there was
2263        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2264        pcretest so that it places the subject at the end of its malloc-ed buffer.
2265    
2266    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2267        exec() might read past the end of the data buffer in UTF-8 mode.
2268    
2269    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2270        the data contained the byte 0x85 as part of a UTF-8 character within its
2271        first line. This applied both to normal and DFA matching.
2272    
2273    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2274        /^[^d]*?$/8 failed to match "abc".
2275    
2276    16. Added a missing copyright notice to pcrecpp_internal.h.
2277    
2278    17. Make it more clear in the documentation that values returned from
2279        pcre_exec() in ovector are byte offsets, not character counts.
2280    
2281    18. Tidied a few places to stop certain compilers from issuing warnings.
2282    
2283    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2284        supplied by Stefan Weber. I made a further small update for 7.8 because
2285        there is a change of source arrangements: the pcre_searchfuncs.c module is
2286        replaced by pcre_ucd.c.
2287    
2288    
2289    Version 7.7 07-May-08
2290    ---------------------
2291    
2292    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2293        a string to a long long, pretend we don't even have a long long." This is
2294        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2295    
2296    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2297        pre-7.6 versions, which defined a global no_arg variable instead of putting
2298        it in the RE class. (See also #8 below.)
2299    
2300    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2301        Lopes.
2302    
2303    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2304    
2305        (1) The include/exclude patterns were being applied to the whole pathnames
2306            of files, instead of just to the final components.
2307    
2308        (2) If there was more than one level of directory, the subdirectories were
2309            skipped unless they satisfied the include/exclude conditions. This is
2310            inconsistent with GNU grep (and could even be seen as contrary to the
2311            pcregrep specification - which I improved to make it absolutely clear).
2312            The action now is always to scan all levels of directory, and just
2313            apply the include/exclude patterns to regular files.
2314    
2315    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2316        --exclude_dir in the tests to avoid scanning .svn directories.
2317    
2318    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2319        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2320        doesn't support NULs in patterns.
2321    
2322    7.  Added some missing "const"s to declarations of static tables in
2323        pcre_compile.c and pcre_dfa_exec.c.
2324    
2325    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2326        caused by fix #2  above. (Subsequently also a second patch to fix the
2327        first patch. And a third patch - this was a messy problem.)
2328    
2329    9.  Applied Craig's patch to remove the use of push_back().
2330    
2331    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2332        matching function regexec().
2333    
2334    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2335        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2336        references. PCRE supports relative numbers with this syntax (I don't think
2337        Oniguruma does).
2338    
2339    12. Previously, a group with a zero repeat such as (...){0} was completely
2340        omitted from the compiled regex. However, this means that if the group
2341        was called as a subroutine from elsewhere in the pattern, things went wrong
2342        (an internal error was given). Such groups are now left in the compiled
2343        pattern, with a new opcode that causes them to be skipped at execution
2344        time.
2345    
2346    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2347        to the way PCRE behaves:
2348    
2349        (a) A lone ] character is dis-allowed (Perl treats it as data).
2350    
2351        (b) A back reference to an unmatched subpattern matches an empty string
2352            (Perl fails the current match path).
2353    
2354        (c) A data ] in a character class must be notated as \] because if the
2355            first data character in a class is ], it defines an empty class. (In
2356            Perl it is not possible to have an empty class.) The empty class []
2357            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2358            The negative empty class [^] matches any one character, independently
2359            of the DOTALL setting.
2360    
2361    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2362        non-existent subpattern following a character class starting with ']' and
2363        containing () gave an internal compiling error instead of "reference to
2364        non-existent subpattern". Fortunately, when the pattern did exist, the
2365        compiled code was correct. (When scanning forwards to check for the
2366        existence of the subpattern, it was treating the data ']' as terminating
2367        the class, so got the count wrong. When actually compiling, the reference
2368        was subsequently set up correctly.)
2369    
2370    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2371        it was being rejected as not supported by pcre_dfa_exec(), even though
2372        other assertions are supported. I have made pcre_dfa_exec() support
2373        (*FAIL).
2374    
2375    16. The implementation of 13c above involved the invention of a new opcode,
2376        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2377        cannot be changed at match time, I realized I could make a small
2378        improvement to matching performance by compiling OP_ALLANY instead of
2379        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2380        on the OP_ANY path.
2381    
2382    17. Compiling pcretest on Windows with readline support failed without the
2383        following two fixes: (1) Make the unistd.h include conditional on
2384        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2385    
2386    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2387        ncurses library to be included for pcretest when ReadLine support is
2388        requested, but also to allow for it to be overridden. This patch came from
2389        Daniel Bergström.
2390    
2391    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2392        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2393        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2394        spotting this.
2395    
2396    
2397    Version 7.6 28-Jan-08
2398    ---------------------
2399    
2400    1.  A character class containing a very large number of characters with
2401        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2402        overflow.
2403    
2404    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2405        HAVE_LONG_LONG is not defined.
2406    
2407    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2408        bring them up to date and include new features. This patch includes:
2409    
2410        - Fixed PH's badly added libz and libbz2 support.
2411        - Fixed a problem with static linking.
2412        - Added pcredemo. [But later removed - see 7 below.]
2413        - Fixed dftables problem and added an option.
2414        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2415            HAVE_LONG_LONG.
2416        - Added readline support for pcretest.
2417        - Added an listing of the option settings after cmake has run.
2418    
2419    4.  A user submitted a patch to Makefile that makes it easy to create
2420        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2421        Makefile.am that cause it to include this special target, without
2422        affecting anything else. Note that the same mingw target plus all
2423        the other distribution libraries and programs are now supported
2424        when configuring with CMake (see 6 below) instead of with
2425        Configure/Make.
2426    
2427    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2428        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2429        exported in the Windows port". It has not yet been confirmed that the patch
2430        solves the problem, but it does no harm.
2431    
2432    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2433        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2434        with CMake, and also correct the comment about stack recursion.
2435    
2436    7.  Remove the automatic building of pcredemo from the ./configure system and
2437        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2438        of a program that users should build themselves after PCRE is installed, so
2439        building it automatically is not really right. What is more, it gave
2440        trouble in some build environments.
2441    
2442    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2443    
2444    
2445    Version 7.5 10-Jan-08
2446    ---------------------
2447    
2448    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2449        values in parens when parsing an RE using the C++ wrapper."
2450    
2451    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2452        Characters greater than 255 were excluded from the class instead of being
2453        included.
2454    
2455    3.  The same bug as (2) above applied to negated POSIX classes such as
2456        [:^space:].
2457    
2458    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2459        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2460        I have changed it.
2461    
2462    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2463        first named subpattern) and a construct such as (?&a) would reference the
2464        first named subpattern whose name started with "a" (in other words, the
2465        length check was missing). Both these problems are fixed. "Subpattern name
2466        expected" is now given for (?&) (a zero-length name), and this patch also
2467        makes it give the same error for \k'' (previously it complained that that
2468        was a reference to a non-existent subpattern).
2469    
2470    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2471        this is right because (?- can be followed by option settings as well as by
2472        digits. I have, however, made the messages clearer.
2473    
2474    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2475        than the number used in the conditional) now cause a compile-time error.
2476        This is actually not compatible with Perl, which accepts such patterns, but
2477        treats the conditional as always being FALSE (as PCRE used to), but it
2478        seems to me that giving a diagnostic is better.
2479    
2480    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2481        and messages.
2482    
2483    9.  Fix two occurrences of "backslash" in comments that should have been
2484        "backspace".
2485    
2486    10. Remove two redundant lines of code that can never be obeyed (their function
2487        was moved elsewhere).
2488    
2489    11. The program that makes PCRE's Unicode character property table had a bug
2490        which caused it to generate incorrect table entries for sequences of
2491        characters that have the same character type, but are in different scripts.
2492        It amalgamated them into a single range, with the script of the first of
2493        them. In other words, some characters were in the wrong script. There were
2494        thirteen such cases, affecting characters in the following ranges:
2495    
2496          U+002b0 - U+002c1
2497          U+0060c - U+0060d
2498          U+0061e - U+00612
2499          U+0064b - U+0065e
2500          U+0074d - U+0076d
2501          U+01800 - U+01805
2502          U+01d00 - U+01d77
2503          U+01d9b - U+01dbf
2504          U+0200b - U+0200f
2505          U+030fc - U+030fe
2506          U+03260 - U+0327f
2507          U+0fb46 - U+0fbb1
2508          U+10450 - U+1049d
2509    
2510    12. The -o option (show only the matching part of a line) for pcregrep was not
2511        compatible with GNU grep in that, if there was more than one match in a
2512        line, it showed only the first of them. It now behaves in the same way as
2513        GNU grep.
2514    
2515    13. If the -o and -v options were combined for pcregrep, it printed a blank
2516        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2517        does the same. The return code can be used to tell if there were any
2518        non-matching lines.
2519    
2520    14. Added --file-offsets and --line-offsets to pcregrep.
2521    
2522    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2523        infinitely looping recursion. The bug was that positive lookaheads were not
2524        being skipped when checking for a possible empty match (negative lookaheads
2525        and both kinds of lookbehind were skipped).
2526    
2527    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2528        inclusion of <windows.h> to before rather than after the definition of
2529        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2530    
2531    17. Specifying a possessive quantifier with a specific limit for a Unicode
2532        character property caused pcre_compile() to compile bad code, which led at
2533        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2534        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2535        caused the error; without that there was no problem.
2536    
2537    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2538    
2539    19. Added --enable-pcretest-libreadline.
2540    
2541    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2542        RE::GlobalReplace(). As a result, the number of replacements returned was
2543        double what it should be. I removed one of the increments, but Craig sent a
2544        later patch that removed the other one (the right fix) and added unit tests
2545        that check the return values (which was not done before).
2546    
2547    21. Several CMake things:
2548    
2549        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2550            the names libpcre and libpcreposix, not just pcre and pcreposix.
2551    
2552        (2) The above change means that pcretest and pcregrep are now correctly
2553            linked with the newly-built libraries, not previously installed ones.
2554    
2555        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2556    
2557    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2558        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2559        UTF-8 newline character). The key issue is that the pattern starts .*;
2560        this means that the match must be either at the beginning, or after a
2561        newline. The bug was in the code for advancing after a failed match and
2562        checking that the new position followed a newline. It was not taking
2563        account of UTF-8 characters correctly.
2564    
2565    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2566        character classes. PCRE was not treating the sequence [:...:] as a
2567        character class unless the ... were all letters. Perl, however, seems to
2568        allow any characters between [: and :], though of course it rejects as
2569        unknown any "names" that contain non-letters, because all the known class
2570        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2571        for example, whereas PCRE did not - it did not recognize a POSIX character
2572        class. This seemed a bit dangerous, so the code has been changed to be
2573        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2574        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2575        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2576        Perl does, and where it didn't before.
2577    
2578    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2579        Windows environments %n is disabled by default.
2580    
2581    
2582    Version 7.4 21-Sep-07
2583    ---------------------
2584    
2585    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2586        means that a class such as [\s] counted as "explicit reference to CR or
2587        LF". That isn't really right - the whole point of the change was to try to
2588        help when there was an actual mention of one of the two characters. So now
2589        the change happens only if \r or \n (or a literal CR or LF) character is
2590        encountered.
2591    
2592    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2593        of both had grown to the point where there were only 3 bits left.
2594        Fortunately, there was spare space in the data structure, and so I have
2595        moved the internal flags into a new 16-bit field to free up more option
2596        bits.
2597    
2598    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2599        but did not set the internal JCHANGED flag - either of these is enough to
2600        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2601        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2602        start sets both bits.
2603    
2604    4.  Added options (at build time, compile time, exec time) to change \R from
2605        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2606    
2607    5.  doc/pcresyntax.html was missing from the distribution.
2608    
2609    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2610        compatibility, even though it is no longer used.
2611    
2612    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2613        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2614        windows.h file is present (where different names are used). [This was
2615        reversed later after testing - see 16 below.]
2616    
2617    8.  Changed all #include <config.h> to #include "config.h". There were also
2618        some further <pcre.h> cases that I changed to "pcre.h".
2619    
2620    9.  When pcregrep was used with the --colour option, it missed the line ending
2621        sequence off the lines that it output.
2622    
2623    10. It was pointed out to me that arrays of string pointers cause lots of
2624        relocations when a shared library is dynamically loaded. A technique of
2625        using a single long string with a table of offsets can drastically reduce
2626        these. I have refactored PCRE in four places to do this. The result is
2627        dramatic:
2628    
2629          Originally:                          290
2630          After changing UCP table:            187
2631          After changing error message table:   43
2632          After changing table of "verbs"       36
2633          After changing table of Posix names   22
2634    
2635        Thanks to the folks working on Gregex for glib for this insight.
2636    
2637    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2638        unicode-properties was also set.
2639    
2640    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2641    
2642    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2643        checked only for CRLF.
2644    
2645    14. Added casts to pcretest.c to avoid compiler warnings.
2646    
2647    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2648    
2649    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2650        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2651        entirely. This removes changes made in 7 above.
2652    
2653    17. The CMake files have been updated, and there is now more information about
2654        building with CMake in the NON-UNIX-USE document.
2655    
2656    
2657    Version 7.3 28-Aug-07
2658    ---------------------
2659    
2660     1. In the rejigging of the build system that eventually resulted in 7.1, the
2661        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2662        brackets there is not right, since it causes compilers to look for an
2663        installed pcre.h, not the version that is in the source that is being
2664        compiled (which of course may be different). I have changed it back to:
2665    
2666          #include "pcre.h"
2667    
2668        I have a vague recollection that the change was concerned with compiling in
2669        different directories, but in the new build system, that is taken care of
2670        by the VPATH setting the Makefile.
2671    
2672     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2673        when the subject happened to end in the byte 0x85 (e.g. if the last
2674        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2675        characters but of course it shouldn't be taken as a newline when it is part
2676        of another character. The bug was that, for an unlimited repeat of . in
2677        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2678        characters when looking for a newline.
2679    
2680     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2681    
2682     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2683        in debug output.
2684    
2685     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2686        long printing in the pcrecpp unittest when running under MinGW.
2687    
2688     6. ESC_K was left out of the EBCDIC table.
2689    
2690     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2691        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2692        limit also applies to "virtual nesting" when a pattern is recursive, and in
2693        this case 1000 isn't so big. I have been able to remove this limit at the
2694        expense of backing off one optimization in certain circumstances. Normally,
2695        when pcre_exec() would call its internal match() function recursively and
2696        immediately return the result unconditionally, it uses a "tail recursion"
2697        feature to save stack. However, when a subpattern that can match an empty
2698        string has an unlimited repetition quantifier, it no longer makes this
2699        optimization. That gives it a stack frame in which to save the data for
2700        checking that an empty string has been matched. Previously this was taken
2701        from the 1000-entry workspace that had been reserved. So now there is no
2702        explicit limit, but more stack is used.
2703    
2704     8. Applied Daniel's patches to solve problems with the import/export magic
2705        syntax that is required for Windows, and which was going wrong for the
2706        pcreposix and pcrecpp parts of the library. These were overlooked when this
2707        problem was solved for the main library.
2708    
2709     9. There were some crude static tests to avoid integer overflow when computing
2710        the size of patterns that contain repeated groups with explicit upper
2711        limits. As the maximum quantifier is 65535, the maximum group length was
2712        set at 30,000 so that the product of these two numbers did not overflow a
2713        32-bit integer. However, it turns out that people want to use groups that
2714        are longer than 30,000 bytes (though not repeat them that many times).
2715        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2716        made it possible to implement the integer overflow checks in a much more
2717        dynamic way, which I have now done. The artificial limitation on group
2718        length has been removed - we now have only the limit on the total length of
2719        the compiled pattern, which depends on the LINK_SIZE setting.
2720    
2721    10. Fixed a bug in the documentation for get/copy named substring when
2722        duplicate names are permitted. If none of the named substrings are set, the
2723        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2724        empty string.
2725    
2726    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2727        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2728        because the ] is interpreted as the first data character and the
2729        terminating ] is not found. PCRE has been made compatible with Perl in this
2730        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2731        cause memory overwriting.
2732    
2733    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2734        string has been matched (to stop an infinite loop). It was not recognizing
2735        a conditional subpattern that could match an empty string if that
2736        subpattern was within another subpattern. For example, it looped when
2737        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2738        condition was not nested. This bug has been fixed.
2739    
2740    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2741        past the start of the subject in the presence of bytes with the top bit
2742        set, for example "\x8aBCD".
2743    
2744    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2745        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2746    
2747    14. Optimized (?!) to (*FAIL).
2748    
2749    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2750        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2751        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2752        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2753        does: it's just the validity check that is more restrictive.
2754    
2755    16. Inserted checks for integer overflows during escape sequence (backslash)
2756        processing, and also fixed erroneous offset values for syntax errors during
2757        backslash processing.
2758    
2759    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2760        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2761    
2762    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2763        caused an overrun.
2764    
2765    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2766        something other than just ASCII characters) inside a group that had an
2767        unlimited repeat caused a loop at compile time (while checking to see
2768        whether the group could match an empty string).
2769    
2770    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2771        [\P{Any}] did so. (Error in the code for printing property names.)
2772    
2773    21. An orphan \E inside a character class could cause a crash.
2774    
2775    22. A repeated capturing bracket such as (A)? could cause a wild memory
2776        reference during compilation.
2777    
2778    23. There are several functions in pcre_compile() that scan along a compiled
2779        expression for various reasons (e.g. to see if it's fixed length for look
2780        behind). There were bugs in these functions when a repeated \p or \P was
2781        present in the pattern. These operators have additional parameters compared
2782        with \d, etc, and these were not being taken into account when moving along
2783        the compiled data. Specifically:
2784    
2785        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2786            length.
2787    
2788        (b) An item such as \pL+ within a repeated group could cause crashes or
2789            loops.
2790    
2791        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2792            "reference to non-existent subpattern" error.
2793    
2794        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2795    
2796    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2797        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2798    
2799    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2800    
2801    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2802        character were causing crashes (broken optimization).
2803    
2804    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2805        \p or \P) caused a compile-time loop.
2806    
2807    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2808        break. For example, the unstudied pattern [\r\n]A does not match the string
2809        "\r\nA" because change 7.0/46 below moves the current point on by two
2810        characters after failing to match at the start. However, the pattern \nA
2811        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2812        the same is true. There doesn't seem any very clean way out of this, but
2813        what I have chosen to do makes the common cases work: PCRE now takes note
2814        of whether there can be an explicit match for \r or \n anywhere in the
2815        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2816        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2817        pattern has explicit CR or LF references.
2818    
2819    29. Added (*CR) etc for changing newline setting at start of pattern.
2820    
2821    
2822    Version 7.2 19-Jun-07
2823    ---------------------
2824    
2825     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2826        which is apparently normally available under Windows.
2827    
2828     2. Re-jig the pcregrep tests with different newline settings in an attempt
2829        to make them independent of the local environment's newline setting.
2830    
2831     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2832    
2833     4. Some of the "internals" tests were previously cut out when the link size
2834        was not 2, because the output contained actual offsets. The recent new
2835        "Z" feature of pcretest means that these can be cut out, making the tests
2836        usable with all link sizes.
2837    
2838     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2839        stack recursion. This gives a massive performance boost under BSD, but just
2840        a small improvement under Linux. However, it saves one field in the frame
2841        in all cases.
2842    
2843     6. Added more features from the forthcoming Perl 5.10:
2844    
2845        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2846            recursion call. It refers to the nth most recently opened parentheses.
2847    
2848        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2849            to be opened parentheses.
2850    
2851        (c) Conditions that refer to capturing parentheses can be specified
2852            relatively, for example, (?(-2)... or (?(+3)...
2853    
2854        (d) \K resets the start of the current match so that everything before
2855            is not part of it.
2856    
2857        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2858    
2859        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2860            reference syntax.
2861    
2862        (g) (?| introduces a group in which the numbering of parentheses in each
2863            alternative starts with the same number.
2864    
2865        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2866    
2867     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2868        PCRE_INFO_JCHANGED.
2869    
2870     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2871        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2872        for detecting groups that can match an empty string.
2873    
2874     9. A pattern with a very large number of alternatives (more than several
2875        hundred) was running out of internal workspace during the pre-compile
2876        phase, where pcre_compile() figures out how much memory will be needed. A
2877        bit of new cunning has reduced the workspace needed for groups with
2878        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2879        workspace instead of running out of the 4096 that are available.
2880    
2881    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2882    
2883    11. Applied patch from Google to remove an optimization that didn't quite work.
2884        The report of the bug said:
2885    
2886          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2887          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2888          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2889    
2890    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2891        it matched the wrong number of bytes.
2892    
2893    
2894    Version 7.1 24-Apr-07
2895  ---------------------  ---------------------
2896    
2897   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
# Line 33  Version 7.1 12-Mar-07 Line 2923  Version 7.1 12-Mar-07
2923   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2924      Weber: (1) pcre_internal.h was missing some function renames; (2) updated      Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2925      makevp.bat for the current PCRE, using the additional files      makevp.bat for the current PCRE, using the additional files
2926      makevp-compile.txt, makevp-linklib.txt, and pcregexp.pas.      makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2927    
2928   6. A Windows user reported a minor discrepancy with test 2, which turned out   6. A Windows user reported a minor discrepancy with test 2, which turned out
2929      to be caused by a trailing space on an input line that had got lost in his      to be caused by a trailing space on an input line that had got lost in his
# Line 116  Version 7.1 12-Mar-07 Line 3006  Version 7.1 12-Mar-07
3006  14. If erroroffset was passed as NULL to pcre_compile, it provoked a  14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3007      segmentation fault instead of returning the appropriate error message.      segmentation fault instead of returning the appropriate error message.
3008    
3009    15. In multiline mode when the newline sequence was set to "any", the pattern
3010        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3011        This doesn't seem right; it now treats the CRLF combination as the line
3012        ending, and so does not match in that case. It's only a pattern such as ^$
3013        that would hit this one: something like ^ABC$ would have failed after \r
3014        and then tried again after \r\n.
3015    
3016    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3017        in an attempt to make files that differ only in their line terminators
3018        compare equal. This works on Linux.
3019    
3020    17. Under certain error circumstances pcregrep might try to free random memory
3021        as it exited. This is now fixed, thanks to valgrind.
3022    
3023    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3024        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3025        was because its rules for how to advance for /g after matching an empty
3026        string at the end of a line did not allow for this case. They now check for
3027        it specially.
3028    
3029    20. pcretest is supposed to handle patterns and data of any length, by
3030        extending its buffers when necessary. It was getting this wrong when the
3031        buffer for a data line had to be extended.
3032    
3033    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3034        CRLF as a newline sequence.
3035    
3036    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3037        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3038        I have nevertheless tidied it up.
3039    
3040    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3041    
3042    24. Added a man page for pcre-config.
3043    
3044    
3045  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06
3046  ---------------------  ---------------------

Legend:
Removed from v.127  
changed lines
  Added in v.1514

  ViewVC Help
Powered by ViewVC 1.1.5