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

Diff of /code/trunk/ChangeLog

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

revision 214 by ph10, Wed Aug 15 14:08:10 2007 UTC revision 1522 by ph10, Sun Feb 8 17:02:05 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 09-Aug-07  Version 8.37 xx-xxx-2015
5    ------------------------
6    
7    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8        for those parentheses to be closed with whatever has been captured so far.
9        However, it was failing to mark any other groups between the hightest
10        capture so far and the currrent group as "unset". Thus, the ovector for
11        those groups contained whatever was previously there. An example is the
12        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13    
14    2.  If an assertion condition was quantified with a minimum of zero (an odd
15        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16    
17    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
18        unrecognized modifier, a crash could occur.
19    
20    4.  An attempt to do global matching in pcretest with a zero-length ovector
21        caused a crash.
22    
23    5.  Fixed a memory leak during matching that could occur for a subpattern
24        subroutine call (recursive or otherwise) if the number of captured groups
25        that had to be saved was greater than ten.
26    
27    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
28        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
29        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
30    
31    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
32        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
33    
34    8. When a pattern is compiled, it remembers the highest back reference so that
35       when matching, if the ovector is too small, extra memory can be obtained to
36       use instead. A conditional subpattern whose condition is a check on a
37       capture having happened, such as, for example in the pattern
38       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
39       setting the highest backreference number. This mattered only if pcre_exec()
40       was called with an ovector that was too small to hold the capture, and there
41       was no other kind of back reference (a situation which is probably quite
42       rare). The effect of the bug was that the condition was always treated as
43       FALSE when the capture could not be consulted, leading to a incorrect
44       behaviour by pcre2_match(). This bug has been fixed.
45    
46    9. A reference to a duplicated named group (either a back reference or a test
47       for being set in a conditional) that occurred in a part of the pattern where
48       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
49       to be incorrectly calculated, leading to overwriting.
50    
51    10. A mutually recursive set of back references such as (\2)(\1) caused a
52        segfault at study time (while trying to find the minimum matching length).
53        The infinite loop is now broken (with the minimum length unset, that is,
54        zero).
55    
56    
57    Version 8.36 26-September-2014
58    ------------------------------
59    
60    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
61        -Wmissing-field-initializers and -Wunused-parameter.
62    
63    2.  The tests for quantifiers being too big (greater than 65535) were being
64        applied after reading the number, and stupidly assuming that integer
65        overflow would give a negative number. The tests are now applied as the
66        numbers are read.
67    
68    3.  Tidy code in pcre_exec.c where two branches that used to be different are
69        now the same.
70    
71    4.  The JIT compiler did not generate match limit checks for certain
72        bracketed expressions with quantifiers. This may lead to exponential
73        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
74        issue should be resolved now.
75    
76    5.  Fixed an issue, which occures when nested alternatives are optimized
77        with table jumps.
78    
79    6.  Inserted two casts and changed some ints to size_t in the light of some
80        reported 64-bit compiler warnings (Bugzilla 1477).
81    
82    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
83        an empty string, which sometimes were behaving incorrectly in the
84        interpreter (though correctly in the JIT matcher). This pcretest input is
85        an example:
86    
87          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
88          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
89    
90        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
91        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
92        for an empty string was breaking the inner loop and carrying on at a lower
93        level, when possessive repeated groups should always return to a higher
94        level as they have no backtrack points in them. The empty string test now
95        occurs at the outer level.
96    
97    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
98        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
99    
100    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
101        doesn't).
102    
103    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
104        to hold a pattern to be incorrectly computed (too small) when there were
105        named back references to duplicated names. This could cause "internal
106        error: code overflow" or "double free or corruption" or other memory
107        handling errors.
108    
109    11. When named subpatterns had the same prefixes, back references could be
110        confused. For example, in this pattern:
111    
112          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
113    
114        the reference to 'Name' was incorrectly treated as a reference to a
115        duplicate name.
116    
117    12. A pattern such as /^s?c/mi8 where the optional character has more than
118        one "other case" was incorrectly compiled such that it would only try to
119        match starting at "c".
120    
121    13. When a pattern starting with \s was studied, VT was not included in the
122        list of possible starting characters; this should have been part of the
123        8.34/18 patch.
124    
125    14. If a character class started [\Qx]... where x is any character, the class
126        was incorrectly terminated at the ].
127    
128    15. If a pattern that started with a caseless match for a character with more
129        than one "other case" was studied, PCRE did not set up the starting code
130        unit bit map for the list of possible characters. Now it does. This is an
131        optimization improvement, not a bug fix.
132    
133    16. The Unicode data tables have been updated to Unicode 7.0.0.
134    
135    17. Fixed a number of memory leaks in pcregrep.
136    
137    18. Avoid a compiler warning (from some compilers) for a function call with
138        a cast that removes "const" from an lvalue by using an intermediate
139        variable (to which the compiler does not object).
140    
141    19. Incorrect code was compiled if a group that contained an internal recursive
142        back reference was optional (had quantifier with a minimum of zero). This
143        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
144        caused segmentation faults because of stack overflows at compile time.
145    
146    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
147        group that is quantified with an indefinite repeat, caused a compile-time
148        loop which used up all the system stack and provoked a segmentation fault.
149        This was not the same bug as 19 above.
150    
151    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
152        Patch by Mike Frysinger.
153    
154    
155    Version 8.35 04-April-2014
156    --------------------------
157    
158    1.  A new flag is set, when property checks are present in an XCLASS.
159        When this flag is not set, PCRE can perform certain optimizations
160        such as studying these XCLASS-es.
161    
162    2.  The auto-possessification of character sets were improved: a normal
163        and an extended character set can be compared now. Furthermore
164        the JIT compiler optimizes more character set checks.
165    
166    3.  Got rid of some compiler warnings for potentially uninitialized variables
167        that show up only when compiled with -O2.
168    
169    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
170        of a match later then the end of the match. The pcretest program was not
171        handling the case sensibly - it was outputting from the start to the next
172        binary zero. It now reports this situation in a message, and outputs the
173        text from the end to the start.
174    
175    5.  Fast forward search is improved in JIT. Instead of the first three
176        characters, any three characters with fixed position can be searched.
177        Search order: first, last, middle.
178    
179    6.  Improve character range checks in JIT. Characters are read by an inprecise
180        function now, which returns with an unknown value if the character code is
181        above a certain threshold (e.g: 256). The only limitation is that the value
182        must be bigger than the threshold as well. This function is useful when
183        the characters above the threshold are handled in the same way.
184    
185    7.  The macros whose names start with RAWUCHAR are placeholders for a future
186        mode in which only the bottom 21 bits of 32-bit data items are used. To
187        make this more memorable for those maintaining the code, the names have
188        been changed to start with UCHAR21, and an extensive comment has been added
189        to their definition.
190    
191    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
192        to the export list in Makefile.am (they were accidentally omitted from the
193        8.34 tarball).
194    
195    9.  The informational output from pcretest used the phrase "starting byte set"
196        which is inappropriate for the 16-bit and 32-bit libraries. As the output
197        for "first char" and "need char" really means "non-UTF-char", I've changed
198        "byte" to "char", and slightly reworded the output. The documentation about
199        these values has also been (I hope) clarified.
200    
201    10. Another JIT related optimization: use table jumps for selecting the correct
202        backtracking path, when more than four alternatives are present inside a
203        bracket.
204    
205    11. Empty match is not possible, when the minimum length is greater than zero,
206        and there is no \K in the pattern. JIT should avoid empty match checks in
207        such cases.
208    
209    12. In a caseless character class with UCP support, when a character with more
210        than one alternative case was not the first character of a range, not all
211        the alternative cases were added to the class. For example, s and \x{17f}
212        are both alternative cases for S: the class [RST] was handled correctly,
213        but [R-T] was not.
214    
215    13. The configure.ac file always checked for pthread support when JIT was
216        enabled. This is not used in Windows, so I have put this test inside a
217        check for the presence of windows.h (which was already tested for).
218    
219    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
220        The algorithm provides a way to skip certain starting offsets, and usually
221        faster than linear prefix searches.
222    
223    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
224        as for 'fr_FR' and 'french'. For some reason, however, it then used the
225        Windows-specific input and output files, which have 'french' screwed in.
226        So this could never have worked. One of the problems with locales is that
227        they aren't always the same. I have now updated RunTest so that it checks
228        the output of the locale test (test 3) against three different output
229        files, and it allows the test to pass if any one of them matches. With luck
230        this should make the test pass on some versions of Solaris where it was
231        failing. Because of the uncertainty, the script did not used to stop if
232        test 3 failed; it now does. If further versions of a French locale ever
233        come to light, they can now easily be added.
234    
235    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
236        there was a message during ./configure, but it did not stop. This now
237        provokes an error. The invalid example in README has been corrected.
238        If a value less than the minimum is given, the minimum value has always
239        been used, but now a warning is given.
240    
241    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
242        was a bug in the test system, which is now fixed. Also, the list of various
243        configurations that are tested for each release did not have one with both
244        16/32 bits and --enable-bar-anycrlf. It now does.
245    
246    18. pcretest was missing "-C bsr" for displaying the \R default setting.
247    
248    19. Little endian PowerPC systems are supported now by the JIT compiler.
249    
250    20. The fast forward newline mechanism could enter to an infinite loop on
251        certain invalid UTF-8 input. Although we don't support these cases
252        this issue can be fixed by a performance optimization.
253    
254    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
255        not take account if existing stack usage. There is now a new global
256        variable called pcre_stack_guard that can be set to point to an external
257        function to check stack availability. It is called at the start of
258        processing every parenthesized group.
259    
260    22. A typo in the code meant that in ungreedy mode the max/min qualifier
261        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
262        not match "ab".
263    
264    23. When UTF was disabled, the JIT program reported some incorrect compile
265        errors. These messages are silenced now.
266    
267    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
268        compiler.
269    
270    25. Change all the temporary files used in RunGrepTest to be different to those
271        used by RunTest so that the tests can be run simultaneously, for example by
272        "make -j check".
273    
274    
275    Version 8.34 15-December-2013
276    -----------------------------
277    
278    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
279        executable memory. Patch inspired by Carsten Klein.
280    
281    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
282        this macro is never tested and has no effect, because the work to support
283        coverage involves only compiling and linking options and special targets in
284        the Makefile. The comment in config.h implied that defining the macro would
285        enable coverage support, which is totally false. There was also support for
286        setting this macro in the CMake files (my fault, I just copied it from
287        configure). SUPPORT_GCOV has now been removed.
288    
289    3.  Make a small performance improvement in strlen16() and strlen32() in
290        pcretest.
291    
292    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
293        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
294        these cases). There was also one in pcretest.c.
295    
296    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
297    
298    6.  In UTF mode, the code for checking whether a group could match an empty
299        string (which is used for indefinitely repeated groups to allow for
300        breaking an infinite loop) was broken when the group contained a repeated
301        negated single-character class with a character that occupied more than one
302        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
303        UTF-8 mode). The effect was undefined: the group might or might not be
304        deemed as matching an empty string, or the program might have crashed.
305    
306    7.  The code for checking whether a group could match an empty string was not
307        recognizing that \h, \H, \v, \V, and \R must match a character.
308    
309    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
310        an empty string. If it can, pcretest shows this in its information output.
311    
312    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
313        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
314        matched by pcre_exec() without using JIT:
315    
316        (a) If the rest of the pattern did not match after a maximal run of
317            grapheme clusters, the code for backing up to try with fewer of them
318            did not always back up over a full grapheme when characters that do not
319            have the modifier quality were involved, e.g. Hangul syllables.
320    
321        (b) If the match point in a subject started with modifier character, and
322            there was no match, the code could incorrectly back up beyond the match
323            point, and potentially beyond the first character in the subject,
324            leading to a segfault or an incorrect match result.
325    
326    10. A conditional group with an assertion condition could lead to PCRE
327        recording an incorrect first data item for a match if no other first data
328        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
329        first data item, and therefore matched "ca" after "c" instead of at the
330        start.
331    
332    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
333        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
334    
335    12. The source of pcregrep now includes z/OS-specific code so that it can be
336        compiled for z/OS as part of the special z/OS distribution.
337    
338    13. Added the -T and -TM options to pcretest.
339    
340    14. The code in pcre_compile.c for creating the table of named capturing groups
341        has been refactored. Instead of creating the table dynamically during the
342        actual compiling pass, the information is remembered during the pre-compile
343        pass (on the stack unless there are more than 20 named groups, in which
344        case malloc() is used) and the whole table is created before the actual
345        compile happens. This has simplified the code (it is now nearly 150 lines
346        shorter) and prepared the way for better handling of references to groups
347        with duplicate names.
348    
349    15. A back reference to a named subpattern when there is more than one of the
350        same name now checks them in the order in which they appear in the pattern.
351        The first one that is set is used for the reference. Previously only the
352        first one was inspected. This change makes PCRE more compatible with Perl.
353    
354    16. Unicode character properties were updated from Unicode 6.3.0.
355    
356    17. The compile-time code for auto-possessification has been refactored, based
357        on a patch by Zoltan Herczeg. It now happens after instead of during
358        compilation. The code is cleaner, and more cases are handled. The option
359        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
360        options in pcretest are provided to set it. It can also be set by
361        (*NO_AUTO_POSSESS) at the start of a pattern.
362    
363    18. The character VT has been added to the default ("C" locale) set of
364        characters that match \s and are generally treated as white space,
365        following this same change in Perl 5.18. There is now no difference between
366        "Perl space" and "POSIX space". Whether VT is treated as white space in
367        other locales depends on the locale.
368    
369    19. The code for checking named groups as conditions, either for being set or
370        for being recursed, has been refactored (this is related to 14 and 15
371        above). Processing unduplicated named groups should now be as fast at
372        numerical groups, and processing duplicated groups should be faster than
373        before.
374    
375    20. Two patches to the CMake build system, by Alexander Barkov:
376    
377          (1) Replace the "source" command by "." in CMakeLists.txt because
378              "source" is a bash-ism.
379    
380          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
381              without these the CMake build does not work on Solaris.
382    
383    21. Perl has changed its handling of \8 and \9. If there is no previously
384        encountered capturing group of those numbers, they are treated as the
385        literal characters 8 and 9 instead of a binary zero followed by the
386        literals. PCRE now does the same.
387    
388    22. Following Perl, added \o{} to specify codepoints in octal, making it
389        possible to specify values greater than 0777 and also making them
390        unambiguous.
391    
392    23. Perl now gives an error for missing closing braces after \x{... instead of
393        treating the string as literal. PCRE now does the same.
394    
395    24. RunTest used to grumble if an inappropriate test was selected explicitly,
396        but just skip it when running all tests. This make it awkward to run ranges
397        of tests when one of them was inappropriate. Now it just skips any
398        inappropriate tests, as it always did when running all tests.
399    
400    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
401        character types such as \d or \w, too many callouts were inserted, and the
402        data that they returned was rubbish.
403    
404    26. In UCP mode, \s was not matching two of the characters that Perl matches,
405        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
406        were matched by \h. The code has now been refactored so that the lists of
407        the horizontal and vertical whitespace characters used for \h and \v (which
408        are defined only in one place) are now also used for \s.
409    
410    27. Add JIT support for the 64 bit TileGX architecture.
411        Patch by Jiong Wang (Tilera Corporation).
412    
413    28. Possessive quantifiers for classes (both explicit and automatically
414        generated) now use special opcodes instead of wrapping in ONCE brackets.
415    
416    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
417        (because it's meaningless), this was not happening when PCRE_CASELESS was
418        set. Not wrong, but inefficient.
419    
420    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
421        properties for \w, \d, etc) is present in a test regex. Otherwise if the
422        test contains no characters greater than 255, Perl doesn't realise it
423        should be using Unicode semantics.
424    
425    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
426        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
427        does in Unicode mode.
428    
429    32. Added the "forbid" facility to pcretest so that putting tests into the
430        wrong test files can sometimes be quickly detected.
431    
432    33. There is now a limit (default 250) on the depth of nesting of parentheses.
433        This limit is imposed to control the amount of system stack used at compile
434        time. It can be changed at build time by --with-parens-nest-limit=xxx or
435        the equivalent in CMake.
436    
437    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
438        errors. Perl warns for these when in warning mode, but PCRE has no facility
439        for giving warnings.
440    
441    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
442        However, this was not working for (?!) because it is optimized to (*FAIL),
443        for which PCRE does not allow quantifiers. The optimization is now disabled
444        when a quantifier follows (?!). I can't see any use for this, but it makes
445        things uniform.
446    
447    36. Perl no longer allows group names to start with digits, so I have made this
448        change also in PCRE. It simplifies the code a bit.
449    
450    37. In extended mode, Perl ignores spaces before a + that indicates a
451        possessive quantifier. PCRE allowed a space before the quantifier, but not
452        before the possessive +. It now does.
453    
454    38. The use of \K (reset reported match start) within a repeated possessive
455        group such as (a\Kb)*+ was not working.
456    
457    40. Document that the same character tables must be used at compile time and
458        run time, and that the facility to pass tables to pcre_exec() and
459        pcre_dfa_exec() is for use only with saved/restored patterns.
460    
461    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
462        features for Builds with MSVC:
463    
464        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
465           builds). The .rc files can be used to set FileDescription and many other
466           attributes.
467    
468        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
469           This allows higher-level build scripts which want .pdb files to avoid
470           hard-coding the exact files needed."
471    
472    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
473        mean "start of word" and "end of word", respectively, as a transition aid.
474    
475    43. A minimizing repeat of a class containing codepoints greater than 255 in
476        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
477        compiled to use the heap for recursion.
478    
479    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
480        is configured.
481    
482    
483    Version 8.33 28-May-2013
484    ------------------------
485    
486    1.  Added 'U' to some constants that are compared to unsigned integers, to
487        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
488        variables that are added to signed variables, to ensure the result is
489        signed and can be negated.
490    
491    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
492        CMake config files.
493    
494    3.  Revise the creation of config.h.generic so that all boolean macros are
495        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
496        overriding via -D on the command line possible.
497    
498    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
499        to unsigned int is reported to make a quite noticeable speed difference in
500        a specific Windows environment. Testing on Linux did also appear to show
501        some benefit (and it is clearly not harmful). Also fixed the definition of
502        Xop which should be unsigned.
503    
504    5.  Related to (4), changing the definition of the intermediate variable cc
505        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
506        improvements.
507    
508    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
509        unnecessary spaces.
510    
511    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
512        and later.
513    
514    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
515    
516    9.  Optimizing fast_forward_start_bits in JIT.
517    
518    10. Adding support for callouts in JIT, and fixing some issues revealed
519        during this work. Namely:
520    
521        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
522    
523        (b) Minimum length was not checked before the matching is started.
524    
525    11. The value of capture_last that is passed to callouts was incorrect in some
526        cases when there was a capture on one path that was subsequently abandoned
527        after a backtrack. Also, the capture_last value is now reset after a
528        recursion, since all captures are also reset in this case.
529    
530    12. The interpreter no longer returns the "too many substrings" error in the
531        case when an overflowing capture is in a branch that is subsequently
532        abandoned after a backtrack.
533    
534    13. In the pathological case when an offset vector of size 2 is used, pcretest
535        now prints out the matched string after a yield of 0 or 1.
536    
537    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
538        Only supported by the JIT compiler at the moment.
539    
540    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
541    
542    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
543        offset of the starting point of the matching process, provided the offsets
544        vector is large enough.
545    
546    17. The \A escape now records a lookbehind value of 1, though its execution
547        does not actually inspect the previous character. This is to ensure that,
548        in partial multi-segment matching, at least one character from the old
549        segment is retained when a new segment is processed. Otherwise, if there
550        are no lookbehinds in the pattern, \A might match incorrectly at the start
551        of a new segment.
552    
553    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
554    
555    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
556        gives some modest performance improvement in 8-bit mode.
557    
558    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
559        be expressed in certain programming languages using Universal Character
560        Names.
561    
562    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
563        which points out that "non characters" are not "characters that may not
564        appear in Unicode strings" but rather "characters that are reserved for
565        internal use and have only local meaning".
566    
567    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
568        there was a conditional group that depended on an assertion, if the
569        assertion was false, the callout that immediately followed the alternation
570        in the condition was skipped when pcre_exec() was used for matching.
571    
572    23. Allow an explicit callout to be inserted before an assertion that is the
573        condition for a conditional group, for compatibility with automatic
574        callouts, which always insert a callout at this point.
575    
576    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
577        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
578    
579    25. (*PRUNE) is now supported by the JIT compiler.
580    
581    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
582    
583    27. Fix the case where there are two or more SKIPs with arguments that may be
584        ignored.
585    
586    28. (*SKIP) is now supported by the JIT compiler.
587    
588    29. (*THEN) is now supported by the JIT compiler.
589    
590    30. Update RunTest with additional test selector options.
591    
592    31. The way PCRE handles backtracking verbs has been changed in two ways.
593    
594        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
595        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
596        backtracking. In some cases this makes it more Perl-compatible, but Perl's
597        rather obscure rules do not always do the same thing.
598    
599        (2) Previously, backtracking verbs were confined within assertions. This is
600        no longer the case for positive assertions, except for (*ACCEPT). Again,
601        this sometimes improves Perl compatibility, and sometimes does not.
602    
603    32. A number of tests that were in test 2 because Perl did things differently
604        have been moved to test 1, because either Perl or PCRE has changed, and
605        these tests are now compatible.
606    
607    32. Backtracking control verbs are now handled in the same way in JIT and
608        interpreter.
609    
610    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
611        contained a forward subroutine reference caused a compile error.
612    
613    34. Auto-detect and optimize limited repetitions in JIT.
614    
615    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
616        blocking (*UTF) etc.
617    
618    36. In the interpreter, maximizing pattern repetitions for characters and
619        character types now use tail recursion, which reduces stack usage.
620    
621    37. The value of the max lookbehind was not correctly preserved if a compiled
622        and saved regex was reloaded on a host of different endianness.
623    
624    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
625        of the compiled pattern block, expand the flags field from 16 to 32 bits
626        because it was almost full.
627    
628    39. Try madvise first before posix_madvise.
629    
630    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
631        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
632        original change locked out all matches of empty strings. This has been
633        changed so that one match of an empty string per line is recognized.
634        Subsequent searches on the same line (for colouring or for --only-matching,
635        for example) do not recognize empty strings.
636    
637    41. Applied a user patch to fix a number of spelling mistakes in comments.
638    
639    42. Data lines longer than 65536 caused pcretest to crash.
640    
641    43. Clarified the data type for length and startoffset arguments for pcre_exec
642        and pcre_dfa_exec in the function-specific man pages, where they were
643        explicitly stated to be in bytes, never having been updated. I also added
644        some clarification to the pcreapi man page.
645    
646    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
647        a segmentation fault.
648    
649    
650    Version 8.32 30-November-2012
651    -----------------------------
652    
653    1.  Improved JIT compiler optimizations for first character search and single
654        character iterators.
655    
656    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
657        Patch by Daniel Richard G.
658    
659    3.  Single character iterator optimizations in the JIT compiler.
660    
661    4.  Improved JIT compiler optimizations for character ranges.
662    
663    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
664        Reported by Giuseppe D'Angelo.
665    
666    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
667        of a line, was being set incorrectly in cases where .* appeared inside
668        atomic brackets at the start of a pattern, or where there was a subsequent
669        *PRUNE or *SKIP.
670    
671    7.  Improved instruction cache flush for POWER/PowerPC.
672        Patch by Daniel Richard G.
673    
674    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
675        grep:
676    
677        (a) There is now no limit to the number of patterns to be matched.
678    
679        (b) An error is given if a pattern is too long.
680    
681        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
682            are now supported.
683    
684        (d) --exclude-from and --include-from (multiple use) have been added.
685    
686        (e) Exclusions and inclusions now apply to all files and directories, not
687            just to those obtained from scanning a directory recursively.
688    
689        (f) Multiple uses of -f and --file-list are now supported.
690    
691        (g) In a Windows environment, the default for -d has been changed from
692            "read" (the GNU grep default) to "skip", because otherwise the presence
693            of a directory in the file list provokes an error.
694    
695        (h) The documentation has been revised and clarified in places.
696    
697    9.  Improve the matching speed of capturing brackets.
698    
699    10. Changed the meaning of \X so that it now matches a Unicode extended
700        grapheme cluster.
701    
702    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
703        out POSIX threads when JIT support is configured.
704    
705    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
706    
707    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
708        structure could go wrong in environments where size_t is not the same size
709        as int.
710    
711    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
712        set.
713    
714    15. The EBCDIC support had decayed; later updates to the code had included
715        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
716        general tidy up of EBCDIC-related issues, and the documentation was also
717        not quite right. There is now a test that can be run on ASCII systems to
718        check some of the EBCDIC-related things (but is it not a full test).
719    
720    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
721        in a small tidy to the code.
722    
723    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
724    
725    18. If the --only-matching (-o) option in pcregrep is specified multiple
726        times, each one causes appropriate output. For example, -o1 -o2 outputs the
727        substrings matched by the 1st and 2nd capturing parentheses. A separating
728        string can be specified by --om-separator (default empty).
729    
730    19. Improving the first n character searches.
731    
732    20. Turn case lists for horizontal and vertical white space into macros so that
733        they are defined only once.
734    
735    21. This set of changes together give more compatible Unicode case-folding
736        behaviour for characters that have more than one other case when UCP
737        support is available.
738    
739        (a) The Unicode property table now has offsets into a new table of sets of
740            three or more characters that are case-equivalent. The MultiStage2.py
741            script that generates these tables (the pcre_ucd.c file) now scans
742            CaseFolding.txt instead of UnicodeData.txt for character case
743            information.
744    
745        (b) The code for adding characters or ranges of characters to a character
746            class has been abstracted into a generalized function that also handles
747            case-independence. In UTF-mode with UCP support, this uses the new data
748            to handle characters with more than one other case.
749    
750        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
751            whose other case is greater than 256 are now correctly matched
752            caselessly. Previously, the high codepoint matched the low one, but not
753            vice versa.
754    
755        (d) The processing of \h, \H, \v, and \ in character classes now makes use
756            of the new class addition function, using character lists defined as
757            macros alongside the case definitions of 20 above.
758    
759        (e) Caseless back references now work with characters that have more than
760            one other case.
761    
762        (f) General caseless matching of characters with more than one other case
763            is supported.
764    
765    22. Unicode character properties were updated from Unicode 6.2.0
766    
767    23. Improved CMake support under Windows. Patch by Daniel Richard G.
768    
769    24. Add support for 32-bit character strings, and UTF-32
770    
771    25. Major JIT compiler update (code refactoring and bugfixing).
772        Experimental Sparc 32 support is added.
773    
774    26. Applied a modified version of Daniel Richard G's patch to create
775        pcre.h.generic and config.h.generic by "make" instead of in the
776        PrepareRelease script.
777    
778    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
779        pcre_compile.c when checking for a zero character.
780    
781    28. Introducing a native interface for JIT. Through this interface, the compiled
782        machine code can be directly executed. The purpose of this interface is to
783        provide fast pattern matching, so several sanity checks are not performed.
784        However, feature tests are still performed. The new interface provides
785        1.4x speedup compared to the old one.
786    
787    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
788        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
789        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
790    
791    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
792        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
793        the "old" RFC 2279). Instead, it ended up passing a negative length to
794        pcre_exec().
795    
796    31. Add support for GCC's visibility feature to hide internal functions.
797    
798    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
799        "unknown -C option" after outputting 0 or 1.
800    
801    33. There is now support for generating a code coverage report for the test
802        suite in environments where gcc is the compiler and lcov is installed. This
803        is mainly for the benefit of the developers.
804    
805    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
806        unaddressable using valgrind annotations, allowing valgrind to detect
807        invalid memory accesses. This is mainly for the benefit of the developers.
808    
809    25. (*UTF) can now be used to start a pattern in any of the three libraries.
810    
811    26. Give configure error if --enable-cpp but no C++ compiler found.
812    
813    
814    Version 8.31 06-July-2012
815    -------------------------
816    
817    1.  Fixing a wrong JIT test case and some compiler warnings.
818    
819    2.  Removed a bashism from the RunTest script.
820    
821    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
822        to unsigned type, result still unsigned" that was given by an MS compiler
823        on encountering the code "-sizeof(xxx)".
824    
825    4.  Partial matching support is added to the JIT compiler.
826    
827    5.  Fixed several bugs concerned with partial matching of items that consist
828        of more than one character:
829    
830        (a) /^(..)\1/ did not partially match "aba" because checking references was
831            done on an "all or nothing" basis. This also applied to repeated
832            references.
833    
834        (b) \R did not give a hard partial match if \r was found at the end of the
835            subject.
836    
837        (c) \X did not give a hard partial match after matching one or more
838            characters at the end of the subject.
839    
840        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
841            a partial match for the string "\r".
842    
843        (e) When newline was set to CRLF, the metacharacter "." did not recognize
844            a partial match for a CR character at the end of the subject string.
845    
846    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
847        running pcretest, the text "(JIT)" added to the output whenever JIT is
848        actually used to run the match.
849    
850    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
851        or /S+[+] with a digit between 1 and 7.
852    
853    8.  OP_NOT now supports any UTF character not just single-byte ones.
854    
855    9.  (*MARK) control verb is now supported by the JIT compiler.
856    
857    10. The command "./RunTest list" lists the available tests without actually
858        running any of them. (Because I keep forgetting what they all are.)
859    
860    11. Add PCRE_INFO_MAXLOOKBEHIND.
861    
862    12. Applied a (slightly modified) user-supplied patch that improves performance
863        when the heap is used for recursion (compiled with --disable-stack-for-
864        recursion). Instead of malloc and free for each heap frame each time a
865        logical recursion happens, frames are retained on a chain and re-used where
866        possible. This sometimes gives as much as 30% improvement.
867    
868    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
869        call.
870    
871    14. As documented, (*COMMIT) is now confined to within a positive assertion.
872    
873    15. It is now possible to link pcretest with libedit as an alternative to
874        libreadline.
875    
876    16. (*COMMIT) control verb is now supported by the JIT compiler.
877    
878    17. The Unicode data tables have been updated to Unicode 6.1.0.
879    
880    18. Added --file-list option to pcregrep.
881    
882    19. Added binary file support to pcregrep, including the -a, --binary-files,
883        -I, and --text options.
884    
885    20. The madvise function is renamed for posix_madvise for QNX compatibility
886        reasons. Fixed by Giuseppe D'Angelo.
887    
888    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
889    
890    22. Changed the option for creating symbolic links for 16-bit man pages from
891        -s to -sf so that re-installing does not cause issues.
892    
893    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
894    
895    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
896        matches in certain environments (the workspace was not being correctly
897        retained). Also added to pcre_dfa_exec() a simple plausibility check on
898        some of the workspace data at the beginning of a restart.
899    
900    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
901        was not doing so when it should - probably a typo introduced by SVN 528
902        (change 8.10/14).
903    
904    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
905        \w+ when the character tables indicated that \x{c4} was a word character.
906        There were several related cases, all because the tests for doing a table
907        lookup were testing for characters less than 127 instead of 255.
908    
909    27. If a pattern contains capturing parentheses that are not used in a match,
910        their slots in the ovector are set to -1. For those that are higher than
911        any matched groups, this happens at the end of processing. In the case when
912        there were back references that the ovector was too small to contain
913        (causing temporary malloc'd memory to be used during matching), and the
914        highest capturing number was not used, memory off the end of the ovector
915        was incorrectly being set to -1. (It was using the size of the temporary
916        memory instead of the true size.)
917    
918    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
919        ovector size, it uses memory at the end of the block that it has got.
920    
921    29. Check for an overlong MARK name and give an error at compile time. The
922        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
923    
924    30. JIT compiler update.
925    
926    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
927        Rill for the patch.
928    
929    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
930    
931    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
932    
933    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
934        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
935        was enabled.
936    
937    35. Improve JIT code generation for greedy plus quantifier.
938    
939    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
940        1 to "aa" instead of to an empty string. The bug affected repeated groups
941        that could potentially match an empty string.
942    
943    37. Optimizing single character iterators in JIT.
944    
945    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
946        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
947        codepoints that are too big for the mode are faulted, and in a UTF mode,
948        disallowed codepoints are also faulted.
949    
950    39. If PCRE was compiled with UTF support, in three places in the DFA
951        matcher there was code that should only have been obeyed in UTF mode, but
952        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
953        processing when bytes with values greater than 127 were present. In 16-bit
954        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
955        both cases the values are those that cannot be the first data item in a UTF
956        character. The three items that might have provoked this were recursions,
957        possessively repeated groups, and atomic groups.
958    
959    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
960        and pcregrep, because some OS require shared objects to be explicitly
961        passed to ld, causing the link step to fail if they are not.
962    
963    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
964        mode, patterns that started with \h* or \R* might be incorrectly matched.
965    
966    
967    Version 8.30 04-February-2012
968    -----------------------------
969    
970    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
971        name is defined in ctype.h.
972    
973    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
974        only in quite long subpatterns.
975    
976    3.  Removed the function pcre_info(), which has been obsolete and deprecated
977        since it was replaced by pcre_fullinfo() in February 2000.
978    
979    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
980        match a (*MARK), and the match failed at the start of the subject, a
981        reference to memory before the start of the subject could occur. This bug
982        was introduced by fix 17 of release 8.21.
983    
984    5.  A reference to an unset group with zero minimum repetition was giving
985        totally wrong answers (in non-JavaScript-compatibility mode). For example,
986        /(another)?(\1?)test/ matched against "hello world test". This bug was
987        introduced in release 8.13.
988    
989    6.  Add support for 16-bit character strings (a large amount of work involving
990        many changes and refactorings).
991    
992    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
993        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
994        from a file.
995    
996    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
997        rounding is not applied in this particular case).
998    
999    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1000        if they appear, or are escaped, in patterns.
1001    
1002    10. Get rid of a number of -Wunused-but-set-variable warnings.
1003    
1004    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1005        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1006        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1007        also returns the mark "x". This bug applied to capturing parentheses,
1008        non-capturing parentheses, and atomic parentheses. It also applied to some
1009        assertions.
1010    
1011    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1012        information out of configure.ac instead of relying on pcre.h.generic, which
1013        is not stored in the repository.
1014    
1015    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1016        -lreadline.
1017    
1018    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1019    
1020    15. Applied Graycode's patch to put the top-level frame on the stack rather
1021        than the heap when not using the stack for recursion. This gives a
1022        performance improvement in many cases when recursion is not deep.
1023    
1024    16. Experimental code added to "pcretest -C" to output the stack frame size.
1025    
1026    
1027    Version 8.21 12-Dec-2011
1028    ------------------------
1029    
1030    1.  Updating the JIT compiler.
1031    
1032    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1033        are added as well.
1034    
1035    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1036        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1037        calling _pcre_jit_exec. Some extra comments are added.
1038    
1039    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1040        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1041        was introduced by change 18 for 8.20.
1042    
1043    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1044        ECMA-262 standard.
1045    
1046    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1047        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1048        This bug was probably introduced by change 9 of 8.13.
1049    
1050    7.  While fixing 6 above, I noticed that a number of other items were being
1051        incorrectly rejected as "not fixed length". This arose partly because newer
1052        opcodes had not been added to the fixed-length checking code. I have (a)
1053        corrected the bug and added tests for these items, and (b) arranged for an
1054        error to occur if an unknown opcode is encountered while checking for fixed
1055        length instead of just assuming "not fixed length". The items that were
1056        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1057        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1058        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1059    
1060    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1061        being incorrectly compiled and would have given unpredicatble results.
1062    
1063    9.  A possessively repeated subpattern with minimum repeat count greater than
1064        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1065        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1066        the first (A) could occur when it should not.
1067    
1068    10. Add a cast and remove a redundant test from the code.
1069    
1070    11. JIT should use pcre_malloc/pcre_free for allocation.
1071    
1072    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1073        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1074        is anything other than /usr, -L is still shown).
1075    
1076    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1077    
1078    14. Perl does not support \N without a following name in a [] class; PCRE now
1079        also gives an error.
1080    
1081    15. If a forward reference was repeated with an upper limit of around 2000,
1082        it caused the error "internal error: overran compiling workspace". The
1083        maximum number of forward references (including repeats) was limited by the
1084        internal workspace, and dependent on the LINK_SIZE. The code has been
1085        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1086        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1087        of around 200,000 forward references. While doing this, I also speeded up
1088        the filling in of repeated forward references.
1089    
1090    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1091        incorrectly expecting the subject to contain another "a" after the start.
1092    
1093    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1094        in the match, the SKIP should be ignored. This was not happening; instead
1095        the SKIP was being treated as NOMATCH. For patterns such as
1096        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1097        tested.
1098    
1099    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1100        now much more compatible with Perl, in particular in cases where the result
1101        is a non-match for a non-anchored pattern. For example, if
1102        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1103        "m", where previously it did not return a name. A side effect of this
1104        change is that for partial matches, the last encountered mark name is
1105        returned, as for non matches. A number of tests that were previously not
1106        Perl-compatible have been moved into the Perl-compatible test files. The
1107        refactoring has had the pleasing side effect of removing one argument from
1108        the match() function, thus reducing its stack requirements.
1109    
1110    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1111        subsequent uses of /S (without +) incorrectly behaved like /S+.
1112    
1113    21. Retrieve executable code size support for the JIT compiler and fixing
1114        some warnings.
1115    
1116    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1117        not work when the shorter character appeared right at the end of the
1118        subject string.
1119    
1120    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1121        systems.
1122    
1123    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1124        output it when the /M option is used in pcretest.
1125    
1126    25. The CheckMan script was not being included in the distribution. Also, added
1127        an explicit "perl" to run Perl scripts from the PrepareRelease script
1128        because this is reportedly needed in Windows.
1129    
1130    26. If study data was being save in a file and studying had not found a set of
1131        "starts with" bytes for the pattern, the data written to the file (though
1132        never used) was taken from uninitialized memory and so caused valgrind to
1133        complain.
1134    
1135    27. Updated RunTest.bat as provided by Sheri Pierce.
1136    
1137    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1138    
1139    29. Computation of memory usage for the table of capturing group names was
1140        giving an unnecessarily large value.
1141    
1142    
1143    Version 8.20 21-Oct-2011
1144    ------------------------
1145    
1146    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1147        a POSIX class. After further experiments with Perl, which convinced me that
1148        Perl has bugs and confusions, a closing square bracket is no longer allowed
1149        in a POSIX name. This bug also affected patterns with classes that started
1150        with full stops.
1151    
1152    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1153        captured substring, but while checking the failing first alternative,
1154        substring 1 is temporarily captured. If the output vector supplied to
1155        pcre_exec() was not big enough for this capture, the yield of the function
1156        was still zero ("insufficient space for captured substrings"). This cannot
1157        be totally fixed without adding another stack variable, which seems a lot
1158        of expense for a edge case. However, I have improved the situation in cases
1159        such as /(a)(b)x|abc/ matched against "abc", where the return code
1160        indicates that fewer than the maximum number of slots in the ovector have
1161        been set.
1162    
1163    3.  Related to (2) above: when there are more back references in a pattern than
1164        slots in the output vector, pcre_exec() uses temporary memory during
1165        matching, and copies in the captures as far as possible afterwards. It was
1166        using the entire output vector, but this conflicts with the specification
1167        that only 2/3 is used for passing back captured substrings. Now it uses
1168        only the first 2/3, for compatibility. This is, of course, another edge
1169        case.
1170    
1171    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1172        main code base, and can be used by building with --enable-jit. When this is
1173        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1174        runtime --no-jit option is given.
1175    
1176    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1177        ovector, the return from the function was zero, implying that there were
1178        other matches that did not fit. The correct "exactly full" value is now
1179        returned.
1180    
1181    6.  If a subpattern that was called recursively or as a subroutine contained
1182        (*PRUNE) or any other control that caused it to give a non-standard return,
1183        invalid errors such as "Error -26 (nested recursion at the same subject
1184        position)" or even infinite loops could occur.
1185    
1186    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1187        computing the minimum length on reaching *ACCEPT, and so ended up with the
1188        wrong value of 1 rather than 0. Further investigation indicates that
1189        computing a minimum subject length in the presence of *ACCEPT is difficult
1190        (think back references, subroutine calls), and so I have changed the code
1191        so that no minimum is registered for a pattern that contains *ACCEPT.
1192    
1193    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1194        it was not handled as intended. [But see 16 below.]
1195    
1196    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1197        Sheri Pierce.
1198    
1199    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1200        the first byte in a match must be "a".
1201    
1202    11. Change 17 for 8.13 increased the recursion depth for patterns like
1203        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1204        pattern contains any instances of (*THEN). If it does not, the old
1205        optimizations are restored. It would be nice to do this on a per-group
1206        basis, but at the moment that is not feasible.
1207    
1208    12. In some environments, the output of pcretest -C is CRLF terminated. This
1209        broke RunTest's code that checks for the link size. A single white space
1210        character after the value is now allowed for.
1211    
1212    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1213        For "fr", it uses the Windows-specific input and output files.
1214    
1215    14. If (*THEN) appeared in a group that was called recursively or as a
1216        subroutine, it did not work as intended. [But see next item.]
1217    
1218    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1219        pattern fragments (but not containing any | characters). If A and B are
1220        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1221        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1222        D. In other words, Perl considers parentheses that do not contain any |
1223        characters to be part of a surrounding alternative, whereas PCRE was
1224        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1225        differently. PCRE now behaves in the same way as Perl, except in the case
1226        of subroutine/recursion calls such as (?1) which have in any case always
1227        been different (but PCRE had them first :-).
1228    
1229    16. Related to 15 above: Perl does not treat the | in a conditional group as
1230        creating alternatives. Such a group is treated in the same way as an
1231        ordinary group without any | characters when processing (*THEN). PCRE has
1232        been changed to match Perl's behaviour.
1233    
1234    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1235        RunGrepTest script failed.
1236    
1237    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1238        inevitable for groups that contain captures, but it can lead to a lot of
1239        stack use in large patterns. The old behaviour has been restored for atomic
1240        groups that do not contain any capturing parentheses.
1241    
1242    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1243        suppress the check for a minimum subject length at run time. (If it was
1244        given to pcre_exec() or pcre_dfa_exec() it did work.)
1245    
1246    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1247        fail to work when decoding hex characters in data strings in EBCDIC
1248        environments.
1249    
1250    21. It appears that in at least one Mac OS environment, the isxdigit() function
1251        is implemented as a macro that evaluates to its argument more than once,
1252        contravening the C 90 Standard (I haven't checked a later standard). There
1253        was an instance in pcretest which caused it to go wrong when processing
1254        \x{...} escapes in subject strings. The has been rewritten to avoid using
1255        things like p++ in the argument of isxdigit().
1256    
1257    
1258    Version 8.13 16-Aug-2011
1259    ------------------------
1260    
1261    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1262    
1263    2.  Two minor typos in pcre_internal.h have been fixed.
1264    
1265    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1266        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1267        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1268    
1269    4.  There were a number of related bugs in the code for matching backrefences
1270        caselessly in UTF-8 mode when codes for the characters concerned were
1271        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1272        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1273        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1274        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1275        2-byte code at the end of the subject (it thought there wasn't enough data
1276        left).
1277    
1278    5.  Comprehensive information about what went wrong is now returned by
1279        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1280        as the output vector has at least 2 elements. The offset of the start of
1281        the failing character and a reason code are placed in the vector.
1282    
1283    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1284        now returned is for the first byte of the failing character, instead of the
1285        last byte inspected. This is an incompatible change, but I hope it is small
1286        enough not to be a problem. It makes the returned offset consistent with
1287        pcre_exec() and pcre_dfa_exec().
1288    
1289    7.  pcretest now gives a text phrase as well as the error number when
1290        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1291        failure, the offset and reason code are output.
1292    
1293    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1294        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1295        back over a single character (\n). This seems wrong (because it treated the
1296        two characters as a single entity when going forwards), conflicts with the
1297        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1298        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1299        has been changed.
1300    
1301    9.  Some internal refactoring has changed the processing so that the handling
1302        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1303        time (the PCRE_DOTALL option was changed this way some time ago: version
1304        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1305        which was always a bit of a fudge. It also means that there is one less
1306        argument for the match() function, which reduces its stack requirements
1307        slightly. This change also fixes an incompatibility with Perl: the pattern
1308        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1309    
1310    10. More internal refactoring has drastically reduced the number of recursive
1311        calls to match() for possessively repeated groups such as (abc)++ when
1312        using pcre_exec().
1313    
1314    11. While implementing 10, a number of bugs in the handling of groups were
1315        discovered and fixed:
1316    
1317        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1318        (a|)*(?1) gave a compile-time internal error.
1319        ((a|)+)+  did not notice that the outer group could match an empty string.
1320        (^a|^)+   was not marked as anchored.
1321        (.*a|.*)+ was not marked as matching at start or after a newline.
1322    
1323    12. Yet more internal refactoring has removed another argument from the match()
1324        function. Special calls to this function are now indicated by setting a
1325        value in a variable in the "match data" data block.
1326    
1327    13. Be more explicit in pcre_study() instead of relying on "default" for
1328        opcodes that mean there is no starting character; this means that when new
1329        ones are added and accidentally left out of pcre_study(), testing should
1330        pick them up.
1331    
1332    14. The -s option of pcretest has been documented for ages as being an old
1333        synonym of -m (show memory usage). I have changed it to mean "force study
1334        for every regex", that is, assume /S for every regex. This is similar to -i
1335        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1336        using it. It makes it easier to run collections of tests with and without
1337        study enabled, and thereby test pcre_study() more easily. All the standard
1338        tests are now run with and without -s (but some patterns can be marked as
1339        "never study" - see 20 below).
1340    
1341    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1342        restoration of the capturing data to the outer values was not happening
1343        correctly.
1344    
1345    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1346        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1347        pattern had matched an empty string, and so incorrectly returned a no
1348        match.
1349    
1350    17. There was optimizing code for the last branch of non-capturing parentheses,
1351        and also for the obeyed branch of a conditional subexpression, which used
1352        tail recursion to cut down on stack usage. Unfortunately, now that there is
1353        the possibility of (*THEN) occurring in these branches, tail recursion is
1354        no longer possible because the return has to be checked for (*THEN). These
1355        two optimizations have therefore been removed. [But see 8.20/11 above.]
1356    
1357    18. If a pattern containing \R was studied, it was assumed that \R always
1358        matched two bytes, thus causing the minimum subject length to be
1359        incorrectly computed because \R can also match just one byte.
1360    
1361    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1362        was incorrectly computed.
1363    
1364    20. If /S is present twice on a test pattern in pcretest input, it now
1365        *disables* studying, thereby overriding the use of -s on the command line
1366        (see 14 above). This is necessary for one or two tests to keep the output
1367        identical in both cases.
1368    
1369    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1370        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1371    
1372    22. When an atomic group that contained a capturing parenthesis was
1373        successfully matched, but the branch in which it appeared failed, the
1374        capturing was not being forgotten if a higher numbered group was later
1375        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1376        group 1 to "a", when in fact it should be unset. This applied to multi-
1377        branched capturing and non-capturing groups, repeated or not, and also to
1378        positive assertions (capturing in negative assertions does not happen
1379        in PCRE) and also to nested atomic groups.
1380    
1381    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1382        subject after a captured substring, to make it easier to tell which of a
1383        number of identical substrings has been captured.
1384    
1385    24. The way atomic groups are processed by pcre_exec() has been changed so that
1386        if they are repeated, backtracking one repetition now resets captured
1387        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1388        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1389        "aaa". Previously, it would have been "a". As part of this code
1390        refactoring, the way recursive calls are handled has also been changed.
1391    
1392    25. If an assertion condition captured any substrings, they were not passed
1393        back unless some other capturing happened later. For example, if
1394        (?(?=(a))a) was matched against "a", no capturing was returned.
1395    
1396    26. When studying a pattern that contained subroutine calls or assertions,
1397        the code for finding the minimum length of a possible match was handling
1398        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1399        group 1 called group 2 while simultaneously a separate group 2 called group
1400        1). A stack overflow occurred in this case. I have fixed this by limiting
1401        the recursion depth to 10.
1402    
1403    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1404        Fortmann. This supports explicit test numbers on the command line, and has
1405        argument validation and error reporting.
1406    
1407    28. An instance of \X with an unlimited repeat could fail if at any point the
1408        first character it looked at was a mark character.
1409    
1410    29. Some minor code refactoring concerning Unicode properties and scripts
1411        should reduce the stack requirement of match() slightly.
1412    
1413    30. Added the '=' option to pcretest to check the setting of unused capturing
1414        slots at the end of the pattern, which are documented as being -1, but are
1415        not included in the return count.
1416    
1417    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1418        compiled something random. Now it gives a compile-time error (as does
1419        Perl).
1420    
1421    32. A *MARK encountered during the processing of a positive assertion is now
1422        recorded and passed back (compatible with Perl).
1423    
1424    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1425        had alternative anchored branches, the search for a second match in a line
1426        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1427        matched the line "0102" twice. The same bug affected patterns that started
1428        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1429        twice.
1430    
1431    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1432        does, and because of capturing effects, quantifying parenthesized
1433        assertions may at times be useful. Quantifiers are now allowed for
1434        parenthesized assertions.
1435    
1436    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1437    
1438    36. \g was being checked for fancy things in a character class, when it should
1439        just be a literal "g".
1440    
1441    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1442        appearance of a nested POSIX class supersedes an apparent external class.
1443        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1444        unescaped square brackets may also appear as part of class names. For
1445        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1446        more like Perl. (But see 8.20/1 above.)
1447    
1448    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1449        was because it thought it was \N{name}, which is not supported).
1450    
1451    39. Add minix to OS list not supporting the -S option in pcretest.
1452    
1453    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1454        cannot analyze patterns in sufficient detail to catch mutual recursions
1455        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1456        subgroup is called recursively as a subpattern for a second time at the
1457        same position in the subject string. In previous releases this might have
1458        been caught by the recursion limit, or it might have run out of stack.
1459    
1460    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1461        happen only once. PCRE was, however incorrectly giving a compile time error
1462        "recursive call could loop indefinitely" because it cannot analyze the
1463        pattern in sufficient detail. The compile time test no longer happens when
1464        PCRE is compiling a conditional subpattern, but actual runaway loops are
1465        now caught at runtime (see 40 above).
1466    
1467    42. It seems that Perl allows any characters other than a closing parenthesis
1468        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1469        has been changed to be the same.
1470    
1471    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1472        as not to get warnings when autogen.sh is called. Also changed
1473        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1474    
1475    44. To help people who use pcregrep to scan files containing exceedingly long
1476        lines, the following changes have been made:
1477    
1478        (a) The default value of the buffer size parameter has been increased from
1479            8K to 20K. (The actual buffer used is three times this size.)
1480    
1481        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1482            PCRE is built.
1483    
1484        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1485            to be set at run time.
1486    
1487        (d) Numerical values in pcregrep options can be followed by K or M, for
1488            example --buffer-size=50K.
1489    
1490        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1491            given and the return code is set to 2.
1492    
1493    45. Add a pointer to the latest mark to the callout data block.
1494    
1495    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1496        partial match of an empty string instead of no match. This was specific to
1497        the use of ".".
1498    
1499    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1500        complete match instead of a partial match. This bug was dependent on both
1501        the PCRE_UTF8 and PCRE_DOTALL options being set.
1502    
1503    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1504        starting byte set, because \b was not being ignored.
1505    
1506    
1507    Version 8.12 15-Jan-2011
1508    ------------------------
1509    
1510    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1511        checks for such things as part of the documentation building process.
1512    
1513    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1514        --match-limit and --recursion-limit options (added for 8.11). In
1515        particular, this made one of the standard tests fail. (The integer value
1516        went into the wrong half of a long int.)
1517    
1518    3.  If the --colour option was given to pcregrep with -v (invert match), it
1519        did strange things, either producing crazy output, or crashing. It should,
1520        of course, ignore a request for colour when reporting lines that do not
1521        match.
1522    
1523    4.  Another pcregrep bug caused similar problems if --colour was specified with
1524        -M (multiline) and the pattern match finished with a line ending.
1525    
1526    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1527        matched in multiline mode, the following line was shown as part of the
1528        match. This seems wrong, so I have changed it.
1529    
1530    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1531        the check for further matches in the same line (so they could be coloured)
1532        to overrun the end of the current line. If another match was found, it was
1533        incorrectly shown (and then shown again when found in the next line).
1534    
1535    7.  If pcregrep was compiled under Windows, there was a reference to the
1536        function pcregrep_exit() before it was defined. I am assuming this was
1537        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1538        reported by a user. I've moved the definition above the reference.
1539    
1540    
1541    Version 8.11 10-Dec-2010
1542    ------------------------
1543    
1544    1.  (*THEN) was not working properly if there were untried alternatives prior
1545        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1546        backtracked to try for "b" instead of moving to the next alternative branch
1547        at the same level (in this case, to look for "c"). The Perl documentation
1548        is clear that when (*THEN) is backtracked onto, it goes to the "next
1549        alternative in the innermost enclosing group".
1550    
1551    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1552        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1553        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1554        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1555        (*THEN).
1556    
1557    3.  If \s appeared in a character class, it removed the VT character from
1558        the class, even if it had been included by some previous item, for example
1559        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1560        of \s, but is part of the POSIX "space" class.)
1561    
1562    4.  A partial match never returns an empty string (because you can always
1563        match an empty string at the end of the subject); however the checking for
1564        an empty string was starting at the "start of match" point. This has been
1565        changed to the "earliest inspected character" point, because the returned
1566        data for a partial match starts at this character. This means that, for
1567        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1568        (previously it gave "no match").
1569    
1570    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1571        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1572        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1573        has an implication that the given string is incomplete (because a partial
1574        match is preferred over a full match). For this reason, these items now
1575        give a partial match in this situation. [Aside: previously, the one case
1576        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1577        match rather than a full match, which was wrong by the old rules, but is
1578        now correct.]
1579    
1580    6.  There was a bug in the handling of #-introduced comments, recognized when
1581        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1582        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1583        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1584        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1585        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1586        places in pcre_compile().
1587    
1588    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1589        comments when looking ahead for named forward references to subpatterns,
1590        the only newline sequence it recognized was NL. It now handles newlines
1591        according to the set newline convention.
1592    
1593    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1594        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1595        cater for a lack of strerror(). These oversights have been fixed.
1596    
1597    9.  Added --match-limit and --recursion-limit to pcregrep.
1598    
1599    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1600    
1601    11. When the -o option was used, pcregrep was setting a return code of 1, even
1602        when matches were found, and --line-buffered was not being honoured.
1603    
1604    12. Added an optional parentheses number to the -o and --only-matching options
1605        of pcregrep.
1606    
1607    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1608        can match an empty string. The code to do it in pcretest and pcredemo
1609        needed fixing:
1610    
1611        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1612            only one byte after an empty string match just before CRLF (this case
1613            just got forgotten; "any" and "anycrlf" were OK).
1614    
1615        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1616            mode when an empty string match preceded an ASCII character followed by
1617            a non-ASCII character. (The code for advancing by one character rather
1618            than one byte was nonsense.)
1619    
1620        (c) The pcredemo.c sample program did not have any code at all to handle
1621            the cases when CRLF is a valid newline sequence.
1622    
1623    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1624        as a starting offset was within the subject string. There is now a new
1625        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1626        negative or greater than the length of the string. In order to test this,
1627        pcretest is extended to allow the setting of negative starting offsets.
1628    
1629    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1630        starting offset points to the beginning of a UTF-8 character was
1631        unnecessarily clumsy. I tidied it up.
1632    
1633    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1634        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1635    
1636    17. Nobody had reported that the --include_dir option, which was added in
1637        release 7.7 should have been called --include-dir (hyphen, not underscore)
1638        for compatibility with GNU grep. I have changed it to --include-dir, but
1639        left --include_dir as an undocumented synonym, and the same for
1640        --exclude-dir, though that is not available in GNU grep, at least as of
1641        release 2.5.4.
1642    
1643    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1644        characters from a string of bytes) have been redefined so as not to use
1645        loops, in order to improve performance in some environments. At the same
1646        time, I abstracted some of the common code into auxiliary macros to save
1647        repetition (this should not affect the compiled code).
1648    
1649    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1650        compile-time error is now given if \c is not followed by an ASCII
1651        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1652        different, and any byte value is allowed.)
1653    
1654    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1655        START_OPTIMIZE option, which is now allowed at compile time - but just
1656        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1657        to pcregrep and other applications that have no direct access to PCRE
1658        options. The new /Y option in pcretest sets this option when calling
1659        pcre_compile().
1660    
1661    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1662        back references. Groups containing recursive back references were forced to
1663        be atomic by that change, but in the case of named groups, the amount of
1664        memory required was incorrectly computed, leading to "Failed: internal
1665        error: code overflow". This has been fixed.
1666    
1667    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1668        pcretest.c, to avoid build problems in some Borland environments.
1669    
1670    
1671    Version 8.10 25-Jun-2010
1672    ------------------------
1673    
1674    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1675        THEN.
1676    
1677    2.  (*ACCEPT) was not working when inside an atomic group.
1678    
1679    3.  Inside a character class, \B is treated as a literal by default, but
1680        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1681        causes the error). The code is unchanged, but I tidied the documentation.
1682    
1683    4.  Inside a character class, PCRE always treated \R and \X as literals,
1684        whereas Perl faults them if its -w option is set. I have changed PCRE so
1685        that it faults them when PCRE_EXTRA is set.
1686    
1687    5.  Added support for \N, which always matches any character other than
1688        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1689    
1690    6.  When compiling pcregrep with newer versions of gcc which may have
1691        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1692        declared with attribute warn_unused_result" were given. Just casting the
1693        result to (void) does not stop the warnings; a more elaborate fudge is
1694        needed. I've used a macro to implement this.
1695    
1696    7.  Minor change to pcretest.c to avoid a compiler warning.
1697    
1698    8.  Added four artifical Unicode properties to help with an option to make
1699        \s etc use properties (see next item). The new properties are: Xan
1700        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1701    
1702    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1703        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1704        this option. Modified pcretest to add /W to test this facility. Added
1705        REG_UCP to make it available via the POSIX interface.
1706    
1707    10. Added --line-buffered to pcregrep.
1708    
1709    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1710        studied, and the match started with a letter with a code point greater than
1711        127 whose first byte was different to the first byte of the other case of
1712        the letter, the other case of this starting letter was not recognized
1713        (#976).
1714    
1715    12. If a pattern that was studied started with a repeated Unicode property
1716        test, for example, \p{Nd}+, there was the theoretical possibility of
1717        setting up an incorrect bitmap of starting bytes, but fortunately it could
1718        not have actually happened in practice until change 8 above was made (it
1719        added property types that matched character-matching opcodes).
1720    
1721    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1722        possible starting bytes for non-anchored patterns.
1723    
1724    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1725        \R, and also a number of cases that involve Unicode properties, both
1726        explicit and implicit when PCRE_UCP is set.
1727    
1728    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1729        input, it could crash or give wrong results if characters with values
1730        greater than 0xc0 were present in the subject string. (Detail: it assumed
1731        UTF-8 input when processing these items.)
1732    
1733    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1734        size_t is 64-bit (#991).
1735    
1736    17. Added a check for running out of memory when PCRE is compiled with
1737        --disable-stack-for-recursion (#990).
1738    
1739    18. If the last data line in a file for pcretest does not have a newline on
1740        the end, a newline was missing in the output.
1741    
1742    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1743        less than 128) in its various bitmaps. However, there is a facility for
1744        generating tables according to the current locale when PCRE is compiled. It
1745        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1746        characters, are recognized by isspace() and therefore were getting set in
1747        these tables, and indeed these tables seem to approximate to ISO 8859. This
1748        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1749        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1750        which of course cannot start UTF-8 characters. I have changed the code so
1751        that only real ASCII characters (less than 128) and the correct starting
1752        bytes for UTF-8 encodings are set for characters greater than 127 when in
1753        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1754        altogether.)
1755    
1756    20. Added the /T option to pcretest so as to be able to run tests with non-
1757        standard character tables, thus making it possible to include the tests
1758        used for 19 above in the standard set of tests.
1759    
1760    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1761        reference to a subpattern the other side of a comment that contains an
1762        opening parenthesis caused either an internal compiling error, or a
1763        reference to the wrong subpattern.
1764    
1765    
1766    Version 8.02 19-Mar-2010
1767    ------------------------
1768    
1769    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1770    
1771    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1772        configured.
1773    
1774    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1775        original author of that file, following a query about its status.
1776    
1777    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1778        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1779    
1780    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1781        quantifier applied to a forward-referencing subroutine call, could compile
1782        incorrect code or give the error "internal error: previously-checked
1783        referenced subpattern not found".
1784    
1785    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1786        variables to point to external functions. For these systems, therefore,
1787        pcre_malloc etc. are now initialized to local functions that call the
1788        relevant global functions.
1789    
1790    7.  There were two entries missing in the vectors called coptable and poptable
1791        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1792        I've fixed the data, and added a kludgy way of testing at compile time that
1793        the lengths are correct (equal to the number of opcodes).
1794    
1795    8.  Following on from 7, I added a similar kludge to check the length of the
1796        eint vector in pcreposix.c.
1797    
1798    9.  Error texts for pcre_compile() are held as one long string to avoid too
1799        much relocation at load time. To find a text, the string is searched,
1800        counting zeros. There was no check for running off the end of the string,
1801        which could happen if a new error number was added without updating the
1802        string.
1803    
1804    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1805    
1806    11. \K was not working if it appeared in an atomic group or in a group that
1807        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1808        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1809        the assertion is positive, but not if it is negative.
1810    
1811    12. Change 11 fortuitously reduced the size of the stack frame used in the
1812        "match()" function of pcre_exec.c by one pointer. Forthcoming
1813        implementation of support for (*MARK) will need an extra pointer on the
1814        stack; I have reserved it now, so that the stack frame size does not
1815        decrease.
1816    
1817    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1818        item in branch that calls a recursion is a subroutine call - as in the
1819        second branch in the above example - was incorrectly given the compile-
1820        time error "recursive call could loop indefinitely" because pcre_compile()
1821        was not correctly checking the subroutine for matching a non-empty string.
1822    
1823    14. The checks for overrunning compiling workspace could trigger after an
1824        overrun had occurred. This is a "should never occur" error, but it can be
1825        triggered by pathological patterns such as hundreds of nested parentheses.
1826        The checks now trigger 100 bytes before the end of the workspace.
1827    
1828    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1829    
1830    
1831    Version 8.01 19-Jan-2010
1832    ------------------------
1833    
1834    1.  If a pattern contained a conditional subpattern with only one branch (in
1835        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1836        computed the wrong minimum data length (which is of course zero for such
1837        subpatterns). This could cause incorrect "no match" results.
1838    
1839    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1840        the pattern is reset in the first branch, pcre_compile() failed with
1841        "internal error: code overflow at offset...". This happened only when
1842        the reset was to the original external option setting. (An optimization
1843        abstracts leading options settings into an external setting, which was the
1844        cause of this.)
1845    
1846    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1847        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1848        assertion pattern did not match (meaning that the assertion was true), it
1849        was incorrectly treated as false if the SKIP had been reached during the
1850        matching. This also applied to assertions used as conditions.
1851    
1852    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1853        assertion subpattern, including such a pattern used as a condition,
1854        unpredictable results occurred, instead of the error return
1855        PCRE_ERROR_DFA_UITEM.
1856    
1857    5.  The C++ GlobalReplace function was not working like Perl for the special
1858        situation when an empty string is matched. It now does the fancy magic
1859        stuff that is necessary.
1860    
1861    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1862        removed. (These were left over from very, very early versions of PCRE.)
1863    
1864    7.  Some cosmetic changes to the code to make life easier when compiling it
1865        as part of something else:
1866    
1867        (a) Change DEBUG to PCRE_DEBUG.
1868    
1869        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1870            called "current" as "current_branch", to prevent a collision with the
1871            Linux macro when compiled as a kernel module.
1872    
1873        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1874            prevent a collision with the Linux macro when compiled as a kernel
1875            module.
1876    
1877    8.  In pcre_compile() there are some checks for integer overflows that used to
1878        cast potentially large values to (double). This has been changed to that
1879        when building, a check for int64_t is made, and if it is found, it is used
1880        instead, thus avoiding the use of floating point arithmetic. (There is no
1881        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1882        double.
1883    
1884    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1885        2005 (difference between two addresses compared to an unsigned value).
1886    
1887    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1888        custom one, because of the following reported problem in Windows:
1889    
1890          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1891              under Win32.
1892          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1893              therefore missing the function definition.
1894          - The compiler thus generates a "C" signature for the test function.
1895          - The linker fails to find the "C" function.
1896          - PCRE fails to configure if asked to do so against libbz2.
1897    
1898    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1899        messages were output:
1900    
1901          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1902          rerunning libtoolize, to keep the correct libtool macros in-tree.
1903          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1904    
1905        I have done both of these things.
1906    
1907    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1908        most of the time, it *can* run out if it is given a pattern that contains a
1909        runaway infinite recursion. I updated the discussion in the pcrestack man
1910        page.
1911    
1912    13. Now that we have gone to the x.xx style of version numbers, the minor
1913        version may start with zero. Using 08 or 09 is a bad idea because users
1914        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1915        interpreted as invalid octal numbers. I've updated the previous comment in
1916        configure.ac, and also added a check that gives an error if 08 or 09 are
1917        used.
1918    
1919    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1920        causing partial matching to fail when the end of the subject matched \W
1921        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1922    
1923    15. There were some discrepancies between the declarations in pcre_internal.h
1924        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1925        their definitions. The declarations used "const uschar *" and the
1926        definitions used USPTR. Even though USPTR is normally defined as "const
1927        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1928        reported that: "This difference in casting confuses some C++ compilers, for
1929        example, SunCC recognizes above declarations as different functions and
1930        generates broken code for hbpcre." I have changed the declarations to use
1931        USPTR.
1932    
1933    16. GNU libtool is named differently on some systems. The autogen.sh script now
1934        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1935        (FreeBSD).
1936    
1937    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1938        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1939        comment: "Figure out how to create a longlong from a string: strtoll and
1940        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1941        instance, but it only takes 2 args instead of 3!"
1942    
1943    18. A subtle bug concerned with back references has been fixed by a change of
1944        specification, with a corresponding code fix. A pattern such as
1945        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1946        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1947        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1948        same bug. Such groups have to be quantified to be useful, or contained
1949        inside another quantified group. (If there's no repetition, the reference
1950        can never match.) The problem arises because, having left the group and
1951        moved on to the rest of the pattern, a later failure that backtracks into
1952        the group uses the captured value from the final iteration of the group
1953        rather than the correct earlier one. I have fixed this in PCRE by forcing
1954        any group that contains a reference to itself to be an atomic group; that
1955        is, there cannot be any backtracking into it once it has completed. This is
1956        similar to recursive and subroutine calls.
1957    
1958    
1959    Version 8.00 19-Oct-09
1960    ----------------------
1961    
1962    1.  The table for translating pcre_compile() error codes into POSIX error codes
1963        was out-of-date, and there was no check on the pcre_compile() error code
1964        being within the table. This could lead to an OK return being given in
1965        error.
1966    
1967    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1968        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1969        in a Windows environment.
1970    
1971    3.  The pcregrep --count option prints the count for each file even when it is
1972        zero, as does GNU grep. However, pcregrep was also printing all files when
1973        --files-with-matches was added. Now, when both options are given, it prints
1974        counts only for those files that have at least one match. (GNU grep just
1975        prints the file name in this circumstance, but including the count seems
1976        more useful - otherwise, why use --count?) Also ensured that the
1977        combination -clh just lists non-zero counts, with no names.
1978    
1979    4.  The long form of the pcregrep -F option was incorrectly implemented as
1980        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1981        but it seems right to fix it, and I didn't think it was worth preserving
1982        the old behaviour.
1983    
1984    5.  The command line items --regex=pattern and --regexp=pattern were not
1985        recognized by pcregrep, which required --regex pattern or --regexp pattern
1986        (with a space rather than an '='). The man page documented the '=' forms,
1987        which are compatible with GNU grep; these now work.
1988    
1989    6.  No libpcreposix.pc file was created for pkg-config; there was just
1990        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1991    
1992    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1993        when UCP support is not needed, by modifying the Python script that
1994        generates it from Unicode data files. This should not matter if the module
1995        is correctly used as a library, but I received one complaint about 50K of
1996        unwanted data. My guess is that the person linked everything into his
1997        program rather than using a library. Anyway, it does no harm.
1998    
1999    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2000        was a minimum greater than 1 for a wide character in a possessive
2001        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2002        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2003        character. Chaos in the form of incorrect output or a compiling loop could
2004        result.
2005    
2006    9.  The restrictions on what a pattern can contain when partial matching is
2007        requested for pcre_exec() have been removed. All patterns can now be
2008        partially matched by this function. In addition, if there are at least two
2009        slots in the offset vector, the offset of the earliest inspected character
2010        for the match and the offset of the end of the subject are set in them when
2011        PCRE_ERROR_PARTIAL is returned.
2012    
2013    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2014        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2015        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2016        and may be more useful for multi-segment matching.
2017    
2018    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2019        used to be given if ever the end of the subject was reached; now it is
2020        given only if matching could not proceed because another character was
2021        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2022        string "Z", which now yields "no match" instead of "partial match". In the
2023        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2024        final character ended with (*FAIL).
2025    
2026    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2027        if the pattern had a "must contain" character that was already found in the
2028        earlier partial match, unless partial matching was again requested. For
2029        example, with the pattern /dog.(body)?/, the "must contain" character is
2030        "g". If the first part-match was for the string "dog", restarting with
2031        "sbody" failed. This bug has been fixed.
2032    
2033    13. The string returned by pcre_dfa_exec() after a partial match has been
2034        changed so that it starts at the first inspected character rather than the
2035        first character of the match. This makes a difference only if the pattern
2036        starts with a lookbehind assertion or \b or \B (\K is not supported by
2037        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2038        matching functions compatible, and I think it's the right thing to do.
2039    
2040    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2041        so that the demonstration program is easily available in environments where
2042        PCRE has not been installed from source.
2043    
2044    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2045        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2046        library.
2047    
2048    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2049        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2050        is not the first non-POSIX option to be added. Clearly some people find
2051        these options useful.
2052    
2053    17. If a caller to the POSIX matching function regexec() passes a non-zero
2054        value for nmatch with a NULL value for pmatch, the value of
2055        nmatch is forced to zero.
2056    
2057    18. RunGrepTest did not have a test for the availability of the -u option of
2058        the diff command, as RunTest does. It now checks in the same way as
2059        RunTest, and also checks for the -b option.
2060    
2061    19. If an odd number of negated classes containing just a single character
2062        interposed, within parentheses, between a forward reference to a named
2063        subpattern and the definition of the subpattern, compilation crashed with
2064        an internal error, complaining that it could not find the referenced
2065        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2066        [The bug was that it was starting one character too far in when skipping
2067        over the character class, thus treating the ] as data rather than
2068        terminating the class. This meant it could skip too much.]
2069    
2070    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2071        /g option in pcretest when the pattern contains \K, which makes it possible
2072        to have an empty string match not at the start, even when the pattern is
2073        anchored. Updated pcretest and pcredemo to use this option.
2074    
2075    21. If the maximum number of capturing subpatterns in a recursion was greater
2076        than the maximum at the outer level, the higher number was returned, but
2077        with unset values at the outer level. The correct (outer level) value is
2078        now given.
2079    
2080    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2081        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2082        make it do so. The string so far is captured, making this feature
2083        compatible with Perl.
2084    
2085    23. The tests have been re-organized, adding tests 11 and 12, to make it
2086        possible to check the Perl 5.10 features against Perl 5.10.
2087    
2088    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2089        pattern matches a fixed length string. PCRE did not allow this; now it
2090        does. Neither allows recursion.
2091    
2092    25. I finally figured out how to implement a request to provide the minimum
2093        length of subject string that was needed in order to match a given pattern.
2094        (It was back references and recursion that I had previously got hung up
2095        on.) This code has now been added to pcre_study(); it finds a lower bound
2096        to the length of subject needed. It is not necessarily the greatest lower
2097        bound, but using it to avoid searching strings that are too short does give
2098        some useful speed-ups. The value is available to calling programs via
2099        pcre_fullinfo().
2100    
2101    26. While implementing 25, I discovered to my embarrassment that pcretest had
2102        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2103        study optimizations had never been tested with that matching function.
2104        Oops. What is worse, even when it was passed study data, there was a bug in
2105        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2106        were also very few tests of studied patterns with pcre_dfa_exec().
2107    
2108    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2109        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2110        on the other side of the coin, they are no longer allowed to have different
2111        names, because these cannot be distinguished in PCRE, and this has caused
2112        confusion. (This is a difference from Perl.)
2113    
2114    28. When duplicate subpattern names are present (necessarily with different
2115        numbers, as required by 27 above), and a test is made by name in a
2116        conditional pattern, either for a subpattern having been matched, or for
2117        recursion in such a pattern, all the associated numbered subpatterns are
2118        tested, and the overall condition is true if the condition is true for any
2119        one of them. This is the way Perl works, and is also more like the way
2120        testing by number works.
2121    
2122    
2123    Version 7.9 11-Apr-09
2124    ---------------------
2125    
2126    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2127        (pcretest), all targets were linked against these libraries. This included
2128        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2129        libraries. This caused unwanted dependencies to be created. This problem
2130        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2131        pcretest is linked with readline.
2132    
2133    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2134        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2135        moved outside it again, because FALSE and TRUE are already defined in AIX,
2136        but BOOL is not.
2137    
2138    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2139        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2140    
2141    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2142        hyphens) following file names and line numbers when outputting matching
2143        lines. This is not true; no spaces are inserted. I have also clarified the
2144        wording for the --colour (or --color) option.
2145    
2146    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2147        was not coloured; this is different to GNU grep, so I have changed it to be
2148        the same.
2149    
2150    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2151        each matching line was coloured. Now it goes on to look for further matches
2152        of any of the test patterns, which is the same behaviour as GNU grep.
2153    
2154    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2155        doesn't make sense to accept an empty string match in pcregrep, so I have
2156        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2157        seems to be how GNU grep behaves. [But see later change 40 for release
2158        8.33.]
2159    
2160    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2161        start or after a newline", because the conditional assertion was not being
2162        correctly handled. The rule now is that both the assertion and what follows
2163        in the first alternative must satisfy the test.
2164    
2165    9.  If auto-callout was enabled in a pattern with a conditional group whose
2166        condition was an assertion, PCRE could crash during matching, both with
2167        pcre_exec() and pcre_dfa_exec().
2168    
2169    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2170        used for matching.
2171    
2172    11. Unicode property support in character classes was not working for
2173        characters (bytes) greater than 127 when not in UTF-8 mode.
2174    
2175    12. Added the -M command line option to pcretest.
2176    
2177    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2178    
2179    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2180    
2181    16. Added comments and documentation about mis-use of no_arg in the C++
2182        wrapper.
2183    
2184    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2185        from Martin Jerabek that uses macro names for all relevant character and
2186        string constants.
2187    
2188    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2189        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2190        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2191        these, but not everybody uses configure.
2192    
2193    19. A conditional group that had only one branch was not being correctly
2194        recognized as an item that could match an empty string. This meant that an
2195        enclosing group might also not be so recognized, causing infinite looping
2196        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2197        with the subject "ab", where knowledge that the repeated group can match
2198        nothing is needed in order to break the loop.
2199    
2200    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2201        exec(), but without supplying a callout function, matching went wrong.
2202    
2203    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2204        leak if the size of the offset vector was greater than 30. When the vector
2205        is smaller, the saved offsets during recursion go onto a local stack
2206        vector, but for larger vectors malloc() is used. It was failing to free
2207        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2208        error, in fact).
2209    
2210    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2211        heapframe that is used only when UTF-8 support is enabled. This caused no
2212        problem, but was untidy.
2213    
2214    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2215        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2216        included within another project.
2217    
2218    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2219        slightly modified by me:
2220    
2221          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2222              not building pcregrep.
2223    
2224          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2225              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2226    
2227    25. Forward references, both numeric and by name, in patterns that made use of
2228        duplicate group numbers, could behave incorrectly or give incorrect errors,
2229        because when scanning forward to find the reference group, PCRE was not
2230        taking into account the duplicate group numbers. A pattern such as
2231        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2232    
2233    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2234        the feature of a custom pointer more persuasive (as requested by a user).
2235    
2236    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2237        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2238        pre-defined.
2239    
2240    28. Added support for (*UTF8) at the start of a pattern.
2241    
2242    29. Arrange for flags added by the "release type" setting in CMake to be shown
2243        in the configuration summary.
2244    
2245    
2246    Version 7.8 05-Sep-08
2247    ---------------------
2248    
2249    1.  Replaced UCP searching code with optimized version as implemented for Ad
2250        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2251        stage table and inline lookup instead of a function, giving speed ups of 2
2252        to 5 times on some simple patterns that I tested. Permission was given to
2253        distribute the MultiStage2.py script that generates the tables (it's not in
2254        the tarball, but is in the Subversion repository).
2255    
2256    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2257        scripts.
2258    
2259    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2260        a group with a zero qualifier. The result of the study could be incorrect,
2261        or the function might crash, depending on the pattern.
2262    
2263    4.  Caseless matching was not working for non-ASCII characters in back
2264        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2265        It now works when Unicode Property Support is available.
2266    
2267    5.  In pcretest, an escape such as \x{de} in the data was always generating
2268        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2269        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2270        truncation.
2271    
2272    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2273    
2274    7.  Added two (int) casts to pcregrep when printing the difference of two
2275        pointers, in case they are 64-bit values.
2276    
2277    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2278        test 2 if it fails.
2279    
2280    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2281        and a #define of that name to empty if it is not externally set. This is to
2282        allow users of MSVC to set it if necessary.
2283    
2284    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2285        the convenience functions in the pcre_get.c source file.
2286    
2287    11. An option change at the start of a pattern that had top-level alternatives
2288        could cause overwriting and/or a crash. This command provoked a crash in
2289        some environments:
2290    
2291          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2292    
2293        This potential security problem was recorded as CVE-2008-2371.
2294    
2295    12. For a pattern where the match had to start at the beginning or immediately
2296        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2297        pcre_dfa_exec() could read past the end of the passed subject if there was
2298        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2299        pcretest so that it places the subject at the end of its malloc-ed buffer.
2300    
2301    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2302        exec() might read past the end of the data buffer in UTF-8 mode.
2303    
2304    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2305        the data contained the byte 0x85 as part of a UTF-8 character within its
2306        first line. This applied both to normal and DFA matching.
2307    
2308    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2309        /^[^d]*?$/8 failed to match "abc".
2310    
2311    16. Added a missing copyright notice to pcrecpp_internal.h.
2312    
2313    17. Make it more clear in the documentation that values returned from
2314        pcre_exec() in ovector are byte offsets, not character counts.
2315    
2316    18. Tidied a few places to stop certain compilers from issuing warnings.
2317    
2318    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2319        supplied by Stefan Weber. I made a further small update for 7.8 because
2320        there is a change of source arrangements: the pcre_searchfuncs.c module is
2321        replaced by pcre_ucd.c.
2322    
2323    
2324    Version 7.7 07-May-08
2325    ---------------------
2326    
2327    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2328        a string to a long long, pretend we don't even have a long long." This is
2329        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2330    
2331    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2332        pre-7.6 versions, which defined a global no_arg variable instead of putting
2333        it in the RE class. (See also #8 below.)
2334    
2335    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2336        Lopes.
2337    
2338    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2339    
2340        (1) The include/exclude patterns were being applied to the whole pathnames
2341            of files, instead of just to the final components.
2342    
2343        (2) If there was more than one level of directory, the subdirectories were
2344            skipped unless they satisfied the include/exclude conditions. This is
2345            inconsistent with GNU grep (and could even be seen as contrary to the
2346            pcregrep specification - which I improved to make it absolutely clear).
2347            The action now is always to scan all levels of directory, and just
2348            apply the include/exclude patterns to regular files.
2349    
2350    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2351        --exclude_dir in the tests to avoid scanning .svn directories.
2352    
2353    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2354        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2355        doesn't support NULs in patterns.
2356    
2357    7.  Added some missing "const"s to declarations of static tables in
2358        pcre_compile.c and pcre_dfa_exec.c.
2359    
2360    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2361        caused by fix #2  above. (Subsequently also a second patch to fix the
2362        first patch. And a third patch - this was a messy problem.)
2363    
2364    9.  Applied Craig's patch to remove the use of push_back().
2365    
2366    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2367        matching function regexec().
2368    
2369    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2370        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2371        references. PCRE supports relative numbers with this syntax (I don't think
2372        Oniguruma does).
2373    
2374    12. Previously, a group with a zero repeat such as (...){0} was completely
2375        omitted from the compiled regex. However, this means that if the group
2376        was called as a subroutine from elsewhere in the pattern, things went wrong
2377        (an internal error was given). Such groups are now left in the compiled
2378        pattern, with a new opcode that causes them to be skipped at execution
2379        time.
2380    
2381    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2382        to the way PCRE behaves:
2383    
2384        (a) A lone ] character is dis-allowed (Perl treats it as data).
2385    
2386        (b) A back reference to an unmatched subpattern matches an empty string
2387            (Perl fails the current match path).
2388    
2389        (c) A data ] in a character class must be notated as \] because if the
2390            first data character in a class is ], it defines an empty class. (In
2391            Perl it is not possible to have an empty class.) The empty class []
2392            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2393            The negative empty class [^] matches any one character, independently
2394            of the DOTALL setting.
2395    
2396    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2397        non-existent subpattern following a character class starting with ']' and
2398        containing () gave an internal compiling error instead of "reference to
2399        non-existent subpattern". Fortunately, when the pattern did exist, the
2400        compiled code was correct. (When scanning forwards to check for the
2401        existence of the subpattern, it was treating the data ']' as terminating
2402        the class, so got the count wrong. When actually compiling, the reference
2403        was subsequently set up correctly.)
2404    
2405    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2406        it was being rejected as not supported by pcre_dfa_exec(), even though
2407        other assertions are supported. I have made pcre_dfa_exec() support
2408        (*FAIL).
2409    
2410    16. The implementation of 13c above involved the invention of a new opcode,
2411        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2412        cannot be changed at match time, I realized I could make a small
2413        improvement to matching performance by compiling OP_ALLANY instead of
2414        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2415        on the OP_ANY path.
2416    
2417    17. Compiling pcretest on Windows with readline support failed without the
2418        following two fixes: (1) Make the unistd.h include conditional on
2419        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2420    
2421    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2422        ncurses library to be included for pcretest when ReadLine support is
2423        requested, but also to allow for it to be overridden. This patch came from
2424        Daniel Bergström.
2425    
2426    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2427        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2428        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2429        spotting this.
2430    
2431    
2432    Version 7.6 28-Jan-08
2433    ---------------------
2434    
2435    1.  A character class containing a very large number of characters with
2436        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2437        overflow.
2438    
2439    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2440        HAVE_LONG_LONG is not defined.
2441    
2442    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2443        bring them up to date and include new features. This patch includes:
2444    
2445        - Fixed PH's badly added libz and libbz2 support.
2446        - Fixed a problem with static linking.
2447        - Added pcredemo. [But later removed - see 7 below.]
2448        - Fixed dftables problem and added an option.
2449        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2450            HAVE_LONG_LONG.
2451        - Added readline support for pcretest.
2452        - Added an listing of the option settings after cmake has run.
2453    
2454    4.  A user submitted a patch to Makefile that makes it easy to create
2455        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2456        Makefile.am that cause it to include this special target, without
2457        affecting anything else. Note that the same mingw target plus all
2458        the other distribution libraries and programs are now supported
2459        when configuring with CMake (see 6 below) instead of with
2460        Configure/Make.
2461    
2462    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2463        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2464        exported in the Windows port". It has not yet been confirmed that the patch
2465        solves the problem, but it does no harm.
2466    
2467    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2468        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2469        with CMake, and also correct the comment about stack recursion.
2470    
2471    7.  Remove the automatic building of pcredemo from the ./configure system and
2472        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2473        of a program that users should build themselves after PCRE is installed, so
2474        building it automatically is not really right. What is more, it gave
2475        trouble in some build environments.
2476    
2477    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2478    
2479    
2480    Version 7.5 10-Jan-08
2481    ---------------------
2482    
2483    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2484        values in parens when parsing an RE using the C++ wrapper."
2485    
2486    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2487        Characters greater than 255 were excluded from the class instead of being
2488        included.
2489    
2490    3.  The same bug as (2) above applied to negated POSIX classes such as
2491        [:^space:].
2492    
2493    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2494        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2495        I have changed it.
2496    
2497    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2498        first named subpattern) and a construct such as (?&a) would reference the
2499        first named subpattern whose name started with "a" (in other words, the
2500        length check was missing). Both these problems are fixed. "Subpattern name
2501        expected" is now given for (?&) (a zero-length name), and this patch also
2502        makes it give the same error for \k'' (previously it complained that that
2503        was a reference to a non-existent subpattern).
2504    
2505    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2506        this is right because (?- can be followed by option settings as well as by
2507        digits. I have, however, made the messages clearer.
2508    
2509    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2510        than the number used in the conditional) now cause a compile-time error.
2511        This is actually not compatible with Perl, which accepts such patterns, but
2512        treats the conditional as always being FALSE (as PCRE used to), but it
2513        seems to me that giving a diagnostic is better.
2514    
2515    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2516        and messages.
2517    
2518    9.  Fix two occurrences of "backslash" in comments that should have been
2519        "backspace".
2520    
2521    10. Remove two redundant lines of code that can never be obeyed (their function
2522        was moved elsewhere).
2523    
2524    11. The program that makes PCRE's Unicode character property table had a bug
2525        which caused it to generate incorrect table entries for sequences of
2526        characters that have the same character type, but are in different scripts.
2527        It amalgamated them into a single range, with the script of the first of
2528        them. In other words, some characters were in the wrong script. There were
2529        thirteen such cases, affecting characters in the following ranges:
2530    
2531          U+002b0 - U+002c1
2532          U+0060c - U+0060d
2533          U+0061e - U+00612
2534          U+0064b - U+0065e
2535          U+0074d - U+0076d
2536          U+01800 - U+01805
2537          U+01d00 - U+01d77
2538          U+01d9b - U+01dbf
2539          U+0200b - U+0200f
2540          U+030fc - U+030fe
2541          U+03260 - U+0327f
2542          U+0fb46 - U+0fbb1
2543          U+10450 - U+1049d
2544    
2545    12. The -o option (show only the matching part of a line) for pcregrep was not
2546        compatible with GNU grep in that, if there was more than one match in a
2547        line, it showed only the first of them. It now behaves in the same way as
2548        GNU grep.
2549    
2550    13. If the -o and -v options were combined for pcregrep, it printed a blank
2551        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2552        does the same. The return code can be used to tell if there were any
2553        non-matching lines.
2554    
2555    14. Added --file-offsets and --line-offsets to pcregrep.
2556    
2557    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2558        infinitely looping recursion. The bug was that positive lookaheads were not
2559        being skipped when checking for a possible empty match (negative lookaheads
2560        and both kinds of lookbehind were skipped).
2561    
2562    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2563        inclusion of <windows.h> to before rather than after the definition of
2564        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2565    
2566    17. Specifying a possessive quantifier with a specific limit for a Unicode
2567        character property caused pcre_compile() to compile bad code, which led at
2568        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2569        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2570        caused the error; without that there was no problem.
2571    
2572    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2573    
2574    19. Added --enable-pcretest-libreadline.
2575    
2576    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2577        RE::GlobalReplace(). As a result, the number of replacements returned was
2578        double what it should be. I removed one of the increments, but Craig sent a
2579        later patch that removed the other one (the right fix) and added unit tests
2580        that check the return values (which was not done before).
2581    
2582    21. Several CMake things:
2583    
2584        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2585            the names libpcre and libpcreposix, not just pcre and pcreposix.
2586    
2587        (2) The above change means that pcretest and pcregrep are now correctly
2588            linked with the newly-built libraries, not previously installed ones.
2589    
2590        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2591    
2592    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2593        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2594        UTF-8 newline character). The key issue is that the pattern starts .*;
2595        this means that the match must be either at the beginning, or after a
2596        newline. The bug was in the code for advancing after a failed match and
2597        checking that the new position followed a newline. It was not taking
2598        account of UTF-8 characters correctly.
2599    
2600    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2601        character classes. PCRE was not treating the sequence [:...:] as a
2602        character class unless the ... were all letters. Perl, however, seems to
2603        allow any characters between [: and :], though of course it rejects as
2604        unknown any "names" that contain non-letters, because all the known class
2605        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2606        for example, whereas PCRE did not - it did not recognize a POSIX character
2607        class. This seemed a bit dangerous, so the code has been changed to be
2608        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2609        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2610        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2611        Perl does, and where it didn't before.
2612    
2613    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2614        Windows environments %n is disabled by default.
2615    
2616    
2617    Version 7.4 21-Sep-07
2618    ---------------------
2619    
2620    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2621        means that a class such as [\s] counted as "explicit reference to CR or
2622        LF". That isn't really right - the whole point of the change was to try to
2623        help when there was an actual mention of one of the two characters. So now
2624        the change happens only if \r or \n (or a literal CR or LF) character is
2625        encountered.
2626    
2627    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2628        of both had grown to the point where there were only 3 bits left.
2629        Fortunately, there was spare space in the data structure, and so I have
2630        moved the internal flags into a new 16-bit field to free up more option
2631        bits.
2632    
2633    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2634        but did not set the internal JCHANGED flag - either of these is enough to
2635        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2636        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2637        start sets both bits.
2638    
2639    4.  Added options (at build time, compile time, exec time) to change \R from
2640        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2641    
2642    5.  doc/pcresyntax.html was missing from the distribution.
2643    
2644    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2645        compatibility, even though it is no longer used.
2646    
2647    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2648        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2649        windows.h file is present (where different names are used). [This was
2650        reversed later after testing - see 16 below.]
2651    
2652    8.  Changed all #include <config.h> to #include "config.h". There were also
2653        some further <pcre.h> cases that I changed to "pcre.h".
2654    
2655    9.  When pcregrep was used with the --colour option, it missed the line ending
2656        sequence off the lines that it output.
2657    
2658    10. It was pointed out to me that arrays of string pointers cause lots of
2659        relocations when a shared library is dynamically loaded. A technique of
2660        using a single long string with a table of offsets can drastically reduce
2661        these. I have refactored PCRE in four places to do this. The result is
2662        dramatic:
2663    
2664          Originally:                          290
2665          After changing UCP table:            187
2666          After changing error message table:   43
2667          After changing table of "verbs"       36
2668          After changing table of Posix names   22
2669    
2670        Thanks to the folks working on Gregex for glib for this insight.
2671    
2672    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2673        unicode-properties was also set.
2674    
2675    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2676    
2677    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2678        checked only for CRLF.
2679    
2680    14. Added casts to pcretest.c to avoid compiler warnings.
2681    
2682    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2683    
2684    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2685        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2686        entirely. This removes changes made in 7 above.
2687    
2688    17. The CMake files have been updated, and there is now more information about
2689        building with CMake in the NON-UNIX-USE document.
2690    
2691    
2692    Version 7.3 28-Aug-07
2693  ---------------------  ---------------------
2694    
2695   1. In the rejigging of the build system that eventually resulted in 7.1, the   1. In the rejigging of the build system that eventually resulted in 7.1, the
# Line 98  Version 7.3 09-Aug-07 Line 2786  Version 7.3 09-Aug-07
2786      the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the      the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2787      full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still      full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2788      does: it's just the validity check that is more restrictive.      does: it's just the validity check that is more restrictive.
2789    
2790  16. Inserted checks for integer overflows during escape sequence (backslash)  16. Inserted checks for integer overflows during escape sequence (backslash)
2791      processing, and also fixed erroneous offset values for syntax errors during      processing, and also fixed erroneous offset values for syntax errors during
2792      backslash processing.      backslash processing.
2793    
2794  17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)  17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2795      for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".      for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2796    
2797    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2798        caused an overrun.
2799    
2800    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2801        something other than just ASCII characters) inside a group that had an
2802        unlimited repeat caused a loop at compile time (while checking to see
2803        whether the group could match an empty string).
2804    
2805    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2806        [\P{Any}] did so. (Error in the code for printing property names.)
2807    
2808    21. An orphan \E inside a character class could cause a crash.
2809    
2810    22. A repeated capturing bracket such as (A)? could cause a wild memory
2811        reference during compilation.
2812    
2813    23. There are several functions in pcre_compile() that scan along a compiled
2814        expression for various reasons (e.g. to see if it's fixed length for look
2815        behind). There were bugs in these functions when a repeated \p or \P was
2816        present in the pattern. These operators have additional parameters compared
2817        with \d, etc, and these were not being taken into account when moving along
2818        the compiled data. Specifically:
2819    
2820        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2821            length.
2822    
2823        (b) An item such as \pL+ within a repeated group could cause crashes or
2824            loops.
2825    
2826        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2827            "reference to non-existent subpattern" error.
2828    
2829        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2830    
2831    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2832        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2833    
2834    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2835    
2836    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2837        character were causing crashes (broken optimization).
2838    
2839    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2840        \p or \P) caused a compile-time loop.
2841    
2842    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2843        break. For example, the unstudied pattern [\r\n]A does not match the string
2844        "\r\nA" because change 7.0/46 below moves the current point on by two
2845        characters after failing to match at the start. However, the pattern \nA
2846        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2847        the same is true. There doesn't seem any very clean way out of this, but
2848        what I have chosen to do makes the common cases work: PCRE now takes note
2849        of whether there can be an explicit match for \r or \n anywhere in the
2850        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2851        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2852        pattern has explicit CR or LF references.
2853    
2854    29. Added (*CR) etc for changing newline setting at start of pattern.
2855    
2856    
2857  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07

Legend:
Removed from v.214  
changed lines
  Added in v.1522

  ViewVC Help
Powered by ViewVC 1.1.5