ViewVC logotype

Diff of /code/trunk/ChangeLog

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

revision 45 by nigel, Sat Feb 24 21:39:25 2007 UTC revision 1521 by ph10, Sun Feb 8 16:43:13 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
4    Version 8.37 xx-xxx-2015
5    ------------------------
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".
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.
17    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
18        unrecognized modifier, a crash could occur.
20    4.  An attempt to do global matching in pcretest with a zero-length ovector
21        caused a crash.
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.
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.
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))/.
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.
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.
52    Version 8.36 26-September-2014
53    ------------------------------
55    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
56        -Wmissing-field-initializers and -Wunused-parameter.
58    2.  The tests for quantifiers being too big (greater than 65535) were being
59        applied after reading the number, and stupidly assuming that integer
60        overflow would give a negative number. The tests are now applied as the
61        numbers are read.
63    3.  Tidy code in pcre_exec.c where two branches that used to be different are
64        now the same.
66    4.  The JIT compiler did not generate match limit checks for certain
67        bracketed expressions with quantifiers. This may lead to exponential
68        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
69        issue should be resolved now.
71    5.  Fixed an issue, which occures when nested alternatives are optimized
72        with table jumps.
74    6.  Inserted two casts and changed some ints to size_t in the light of some
75        reported 64-bit compiler warnings (Bugzilla 1477).
77    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
78        an empty string, which sometimes were behaving incorrectly in the
79        interpreter (though correctly in the JIT matcher). This pcretest input is
80        an example:
82          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
85        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
86        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
87        for an empty string was breaking the inner loop and carrying on at a lower
88        level, when possessive repeated groups should always return to a higher
89        level as they have no backtrack points in them. The empty string test now
90        occurs at the outer level.
92    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
93        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
95    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
96        doesn't).
98    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
99        to hold a pattern to be incorrectly computed (too small) when there were
100        named back references to duplicated names. This could cause "internal
101        error: code overflow" or "double free or corruption" or other memory
102        handling errors.
104    11. When named subpatterns had the same prefixes, back references could be
105        confused. For example, in this pattern:
107          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
109        the reference to 'Name' was incorrectly treated as a reference to a
110        duplicate name.
112    12. A pattern such as /^s?c/mi8 where the optional character has more than
113        one "other case" was incorrectly compiled such that it would only try to
114        match starting at "c".
116    13. When a pattern starting with \s was studied, VT was not included in the
117        list of possible starting characters; this should have been part of the
118        8.34/18 patch.
120    14. If a character class started [\Qx]... where x is any character, the class
121        was incorrectly terminated at the ].
123    15. If a pattern that started with a caseless match for a character with more
124        than one "other case" was studied, PCRE did not set up the starting code
125        unit bit map for the list of possible characters. Now it does. This is an
126        optimization improvement, not a bug fix.
128    16. The Unicode data tables have been updated to Unicode 7.0.0.
130    17. Fixed a number of memory leaks in pcregrep.
132    18. Avoid a compiler warning (from some compilers) for a function call with
133        a cast that removes "const" from an lvalue by using an intermediate
134        variable (to which the compiler does not object).
136    19. Incorrect code was compiled if a group that contained an internal recursive
137        back reference was optional (had quantifier with a minimum of zero). This
138        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
139        caused segmentation faults because of stack overflows at compile time.
141    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
142        group that is quantified with an indefinite repeat, caused a compile-time
143        loop which used up all the system stack and provoked a segmentation fault.
144        This was not the same bug as 19 above.
146    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
147        Patch by Mike Frysinger.
150    Version 8.35 04-April-2014
151    --------------------------
153    1.  A new flag is set, when property checks are present in an XCLASS.
154        When this flag is not set, PCRE can perform certain optimizations
155        such as studying these XCLASS-es.
157    2.  The auto-possessification of character sets were improved: a normal
158        and an extended character set can be compared now. Furthermore
159        the JIT compiler optimizes more character set checks.
161    3.  Got rid of some compiler warnings for potentially uninitialized variables
162        that show up only when compiled with -O2.
164    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
165        of a match later then the end of the match. The pcretest program was not
166        handling the case sensibly - it was outputting from the start to the next
167        binary zero. It now reports this situation in a message, and outputs the
168        text from the end to the start.
170    5.  Fast forward search is improved in JIT. Instead of the first three
171        characters, any three characters with fixed position can be searched.
172        Search order: first, last, middle.
174    6.  Improve character range checks in JIT. Characters are read by an inprecise
175        function now, which returns with an unknown value if the character code is
176        above a certain threshold (e.g: 256). The only limitation is that the value
177        must be bigger than the threshold as well. This function is useful when
178        the characters above the threshold are handled in the same way.
180    7.  The macros whose names start with RAWUCHAR are placeholders for a future
181        mode in which only the bottom 21 bits of 32-bit data items are used. To
182        make this more memorable for those maintaining the code, the names have
183        been changed to start with UCHAR21, and an extensive comment has been added
184        to their definition.
186    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
187        to the export list in Makefile.am (they were accidentally omitted from the
188        8.34 tarball).
190    9.  The informational output from pcretest used the phrase "starting byte set"
191        which is inappropriate for the 16-bit and 32-bit libraries. As the output
192        for "first char" and "need char" really means "non-UTF-char", I've changed
193        "byte" to "char", and slightly reworded the output. The documentation about
194        these values has also been (I hope) clarified.
196    10. Another JIT related optimization: use table jumps for selecting the correct
197        backtracking path, when more than four alternatives are present inside a
198        bracket.
200    11. Empty match is not possible, when the minimum length is greater than zero,
201        and there is no \K in the pattern. JIT should avoid empty match checks in
202        such cases.
204    12. In a caseless character class with UCP support, when a character with more
205        than one alternative case was not the first character of a range, not all
206        the alternative cases were added to the class. For example, s and \x{17f}
207        are both alternative cases for S: the class [RST] was handled correctly,
208        but [R-T] was not.
210    13. The configure.ac file always checked for pthread support when JIT was
211        enabled. This is not used in Windows, so I have put this test inside a
212        check for the presence of windows.h (which was already tested for).
214    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
215        The algorithm provides a way to skip certain starting offsets, and usually
216        faster than linear prefix searches.
218    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
219        as for 'fr_FR' and 'french'. For some reason, however, it then used the
220        Windows-specific input and output files, which have 'french' screwed in.
221        So this could never have worked. One of the problems with locales is that
222        they aren't always the same. I have now updated RunTest so that it checks
223        the output of the locale test (test 3) against three different output
224        files, and it allows the test to pass if any one of them matches. With luck
225        this should make the test pass on some versions of Solaris where it was
226        failing. Because of the uncertainty, the script did not used to stop if
227        test 3 failed; it now does. If further versions of a French locale ever
228        come to light, they can now easily be added.
230    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
231        there was a message during ./configure, but it did not stop. This now
232        provokes an error. The invalid example in README has been corrected.
233        If a value less than the minimum is given, the minimum value has always
234        been used, but now a warning is given.
236    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
237        was a bug in the test system, which is now fixed. Also, the list of various
238        configurations that are tested for each release did not have one with both
239        16/32 bits and --enable-bar-anycrlf. It now does.
241    18. pcretest was missing "-C bsr" for displaying the \R default setting.
243    19. Little endian PowerPC systems are supported now by the JIT compiler.
245    20. The fast forward newline mechanism could enter to an infinite loop on
246        certain invalid UTF-8 input. Although we don't support these cases
247        this issue can be fixed by a performance optimization.
249    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
250        not take account if existing stack usage. There is now a new global
251        variable called pcre_stack_guard that can be set to point to an external
252        function to check stack availability. It is called at the start of
253        processing every parenthesized group.
255    22. A typo in the code meant that in ungreedy mode the max/min qualifier
256        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
257        not match "ab".
259    23. When UTF was disabled, the JIT program reported some incorrect compile
260        errors. These messages are silenced now.
262    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
263        compiler.
265    25. Change all the temporary files used in RunGrepTest to be different to those
266        used by RunTest so that the tests can be run simultaneously, for example by
267        "make -j check".
270    Version 8.34 15-December-2013
271    -----------------------------
273    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
274        executable memory. Patch inspired by Carsten Klein.
276    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
277        this macro is never tested and has no effect, because the work to support
278        coverage involves only compiling and linking options and special targets in
279        the Makefile. The comment in config.h implied that defining the macro would
280        enable coverage support, which is totally false. There was also support for
281        setting this macro in the CMake files (my fault, I just copied it from
282        configure). SUPPORT_GCOV has now been removed.
284    3.  Make a small performance improvement in strlen16() and strlen32() in
285        pcretest.
287    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
288        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
289        these cases). There was also one in pcretest.c.
291    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
293    6.  In UTF mode, the code for checking whether a group could match an empty
294        string (which is used for indefinitely repeated groups to allow for
295        breaking an infinite loop) was broken when the group contained a repeated
296        negated single-character class with a character that occupied more than one
297        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
298        UTF-8 mode). The effect was undefined: the group might or might not be
299        deemed as matching an empty string, or the program might have crashed.
301    7.  The code for checking whether a group could match an empty string was not
302        recognizing that \h, \H, \v, \V, and \R must match a character.
304    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
305        an empty string. If it can, pcretest shows this in its information output.
307    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
308        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
309        matched by pcre_exec() without using JIT:
311        (a) If the rest of the pattern did not match after a maximal run of
312            grapheme clusters, the code for backing up to try with fewer of them
313            did not always back up over a full grapheme when characters that do not
314            have the modifier quality were involved, e.g. Hangul syllables.
316        (b) If the match point in a subject started with modifier character, and
317            there was no match, the code could incorrectly back up beyond the match
318            point, and potentially beyond the first character in the subject,
319            leading to a segfault or an incorrect match result.
321    10. A conditional group with an assertion condition could lead to PCRE
322        recording an incorrect first data item for a match if no other first data
323        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
324        first data item, and therefore matched "ca" after "c" instead of at the
325        start.
327    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
328        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
330    12. The source of pcregrep now includes z/OS-specific code so that it can be
331        compiled for z/OS as part of the special z/OS distribution.
333    13. Added the -T and -TM options to pcretest.
335    14. The code in pcre_compile.c for creating the table of named capturing groups
336        has been refactored. Instead of creating the table dynamically during the
337        actual compiling pass, the information is remembered during the pre-compile
338        pass (on the stack unless there are more than 20 named groups, in which
339        case malloc() is used) and the whole table is created before the actual
340        compile happens. This has simplified the code (it is now nearly 150 lines
341        shorter) and prepared the way for better handling of references to groups
342        with duplicate names.
344    15. A back reference to a named subpattern when there is more than one of the
345        same name now checks them in the order in which they appear in the pattern.
346        The first one that is set is used for the reference. Previously only the
347        first one was inspected. This change makes PCRE more compatible with Perl.
349    16. Unicode character properties were updated from Unicode 6.3.0.
351    17. The compile-time code for auto-possessification has been refactored, based
352        on a patch by Zoltan Herczeg. It now happens after instead of during
353        compilation. The code is cleaner, and more cases are handled. The option
354        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
355        options in pcretest are provided to set it. It can also be set by
356        (*NO_AUTO_POSSESS) at the start of a pattern.
358    18. The character VT has been added to the default ("C" locale) set of
359        characters that match \s and are generally treated as white space,
360        following this same change in Perl 5.18. There is now no difference between
361        "Perl space" and "POSIX space". Whether VT is treated as white space in
362        other locales depends on the locale.
364    19. The code for checking named groups as conditions, either for being set or
365        for being recursed, has been refactored (this is related to 14 and 15
366        above). Processing unduplicated named groups should now be as fast at
367        numerical groups, and processing duplicated groups should be faster than
368        before.
370    20. Two patches to the CMake build system, by Alexander Barkov:
372          (1) Replace the "source" command by "." in CMakeLists.txt because
373              "source" is a bash-ism.
375          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
376              without these the CMake build does not work on Solaris.
378    21. Perl has changed its handling of \8 and \9. If there is no previously
379        encountered capturing group of those numbers, they are treated as the
380        literal characters 8 and 9 instead of a binary zero followed by the
381        literals. PCRE now does the same.
383    22. Following Perl, added \o{} to specify codepoints in octal, making it
384        possible to specify values greater than 0777 and also making them
385        unambiguous.
387    23. Perl now gives an error for missing closing braces after \x{... instead of
388        treating the string as literal. PCRE now does the same.
390    24. RunTest used to grumble if an inappropriate test was selected explicitly,
391        but just skip it when running all tests. This make it awkward to run ranges
392        of tests when one of them was inappropriate. Now it just skips any
393        inappropriate tests, as it always did when running all tests.
395    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
396        character types such as \d or \w, too many callouts were inserted, and the
397        data that they returned was rubbish.
399    26. In UCP mode, \s was not matching two of the characters that Perl matches,
400        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
401        were matched by \h. The code has now been refactored so that the lists of
402        the horizontal and vertical whitespace characters used for \h and \v (which
403        are defined only in one place) are now also used for \s.
405    27. Add JIT support for the 64 bit TileGX architecture.
406        Patch by Jiong Wang (Tilera Corporation).
408    28. Possessive quantifiers for classes (both explicit and automatically
409        generated) now use special opcodes instead of wrapping in ONCE brackets.
411    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
412        (because it's meaningless), this was not happening when PCRE_CASELESS was
413        set. Not wrong, but inefficient.
415    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
416        properties for \w, \d, etc) is present in a test regex. Otherwise if the
417        test contains no characters greater than 255, Perl doesn't realise it
418        should be using Unicode semantics.
420    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
421        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
422        does in Unicode mode.
424    32. Added the "forbid" facility to pcretest so that putting tests into the
425        wrong test files can sometimes be quickly detected.
427    33. There is now a limit (default 250) on the depth of nesting of parentheses.
428        This limit is imposed to control the amount of system stack used at compile
429        time. It can be changed at build time by --with-parens-nest-limit=xxx or
430        the equivalent in CMake.
432    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
433        errors. Perl warns for these when in warning mode, but PCRE has no facility
434        for giving warnings.
436    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
437        However, this was not working for (?!) because it is optimized to (*FAIL),
438        for which PCRE does not allow quantifiers. The optimization is now disabled
439        when a quantifier follows (?!). I can't see any use for this, but it makes
440        things uniform.
442    36. Perl no longer allows group names to start with digits, so I have made this
443        change also in PCRE. It simplifies the code a bit.
445    37. In extended mode, Perl ignores spaces before a + that indicates a
446        possessive quantifier. PCRE allowed a space before the quantifier, but not
447        before the possessive +. It now does.
449    38. The use of \K (reset reported match start) within a repeated possessive
450        group such as (a\Kb)*+ was not working.
452    40. Document that the same character tables must be used at compile time and
453        run time, and that the facility to pass tables to pcre_exec() and
454        pcre_dfa_exec() is for use only with saved/restored patterns.
456    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
457        features for Builds with MSVC:
459        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
460           builds). The .rc files can be used to set FileDescription and many other
461           attributes.
463        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
464           This allows higher-level build scripts which want .pdb files to avoid
465           hard-coding the exact files needed."
467    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
468        mean "start of word" and "end of word", respectively, as a transition aid.
470    43. A minimizing repeat of a class containing codepoints greater than 255 in
471        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
472        compiled to use the heap for recursion.
474    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
475        is configured.
478    Version 8.33 28-May-2013
479    ------------------------
481    1.  Added 'U' to some constants that are compared to unsigned integers, to
482        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
483        variables that are added to signed variables, to ensure the result is
484        signed and can be negated.
486    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
487        CMake config files.
489    3.  Revise the creation of config.h.generic so that all boolean macros are
490        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
491        overriding via -D on the command line possible.
493    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
494        to unsigned int is reported to make a quite noticeable speed difference in
495        a specific Windows environment. Testing on Linux did also appear to show
496        some benefit (and it is clearly not harmful). Also fixed the definition of
497        Xop which should be unsigned.
499    5.  Related to (4), changing the definition of the intermediate variable cc
500        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
501        improvements.
503    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
504        unnecessary spaces.
506    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
507        and later.
509    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
511    9.  Optimizing fast_forward_start_bits in JIT.
513    10. Adding support for callouts in JIT, and fixing some issues revealed
514        during this work. Namely:
516        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
518        (b) Minimum length was not checked before the matching is started.
520    11. The value of capture_last that is passed to callouts was incorrect in some
521        cases when there was a capture on one path that was subsequently abandoned
522        after a backtrack. Also, the capture_last value is now reset after a
523        recursion, since all captures are also reset in this case.
525    12. The interpreter no longer returns the "too many substrings" error in the
526        case when an overflowing capture is in a branch that is subsequently
527        abandoned after a backtrack.
529    13. In the pathological case when an offset vector of size 2 is used, pcretest
530        now prints out the matched string after a yield of 0 or 1.
532    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
533        Only supported by the JIT compiler at the moment.
535    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
537    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
538        offset of the starting point of the matching process, provided the offsets
539        vector is large enough.
541    17. The \A escape now records a lookbehind value of 1, though its execution
542        does not actually inspect the previous character. This is to ensure that,
543        in partial multi-segment matching, at least one character from the old
544        segment is retained when a new segment is processed. Otherwise, if there
545        are no lookbehinds in the pattern, \A might match incorrectly at the start
546        of a new segment.
548    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
550    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
551        gives some modest performance improvement in 8-bit mode.
553    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
554        be expressed in certain programming languages using Universal Character
555        Names.
557    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
558        which points out that "non characters" are not "characters that may not
559        appear in Unicode strings" but rather "characters that are reserved for
560        internal use and have only local meaning".
562    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
563        there was a conditional group that depended on an assertion, if the
564        assertion was false, the callout that immediately followed the alternation
565        in the condition was skipped when pcre_exec() was used for matching.
567    23. Allow an explicit callout to be inserted before an assertion that is the
568        condition for a conditional group, for compatibility with automatic
569        callouts, which always insert a callout at this point.
571    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
572        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
574    25. (*PRUNE) is now supported by the JIT compiler.
576    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
578    27. Fix the case where there are two or more SKIPs with arguments that may be
579        ignored.
581    28. (*SKIP) is now supported by the JIT compiler.
583    29. (*THEN) is now supported by the JIT compiler.
585    30. Update RunTest with additional test selector options.
587    31. The way PCRE handles backtracking verbs has been changed in two ways.
589        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
590        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
591        backtracking. In some cases this makes it more Perl-compatible, but Perl's
592        rather obscure rules do not always do the same thing.
594        (2) Previously, backtracking verbs were confined within assertions. This is
595        no longer the case for positive assertions, except for (*ACCEPT). Again,
596        this sometimes improves Perl compatibility, and sometimes does not.
598    32. A number of tests that were in test 2 because Perl did things differently
599        have been moved to test 1, because either Perl or PCRE has changed, and
600        these tests are now compatible.
602    32. Backtracking control verbs are now handled in the same way in JIT and
603        interpreter.
605    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
606        contained a forward subroutine reference caused a compile error.
608    34. Auto-detect and optimize limited repetitions in JIT.
610    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
611        blocking (*UTF) etc.
613    36. In the interpreter, maximizing pattern repetitions for characters and
614        character types now use tail recursion, which reduces stack usage.
616    37. The value of the max lookbehind was not correctly preserved if a compiled
617        and saved regex was reloaded on a host of different endianness.
619    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
620        of the compiled pattern block, expand the flags field from 16 to 32 bits
621        because it was almost full.
623    39. Try madvise first before posix_madvise.
625    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
626        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
627        original change locked out all matches of empty strings. This has been
628        changed so that one match of an empty string per line is recognized.
629        Subsequent searches on the same line (for colouring or for --only-matching,
630        for example) do not recognize empty strings.
632    41. Applied a user patch to fix a number of spelling mistakes in comments.
634    42. Data lines longer than 65536 caused pcretest to crash.
636    43. Clarified the data type for length and startoffset arguments for pcre_exec
637        and pcre_dfa_exec in the function-specific man pages, where they were
638        explicitly stated to be in bytes, never having been updated. I also added
639        some clarification to the pcreapi man page.
641    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
642        a segmentation fault.
645    Version 8.32 30-November-2012
646    -----------------------------
648    1.  Improved JIT compiler optimizations for first character search and single
649        character iterators.
651    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
652        Patch by Daniel Richard G.
654    3.  Single character iterator optimizations in the JIT compiler.
656    4.  Improved JIT compiler optimizations for character ranges.
658    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
659        Reported by Giuseppe D'Angelo.
661    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
662        of a line, was being set incorrectly in cases where .* appeared inside
663        atomic brackets at the start of a pattern, or where there was a subsequent
664        *PRUNE or *SKIP.
666    7.  Improved instruction cache flush for POWER/PowerPC.
667        Patch by Daniel Richard G.
669    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
670        grep:
672        (a) There is now no limit to the number of patterns to be matched.
674        (b) An error is given if a pattern is too long.
676        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
677            are now supported.
679        (d) --exclude-from and --include-from (multiple use) have been added.
681        (e) Exclusions and inclusions now apply to all files and directories, not
682            just to those obtained from scanning a directory recursively.
684        (f) Multiple uses of -f and --file-list are now supported.
686        (g) In a Windows environment, the default for -d has been changed from
687            "read" (the GNU grep default) to "skip", because otherwise the presence
688            of a directory in the file list provokes an error.
690        (h) The documentation has been revised and clarified in places.
692    9.  Improve the matching speed of capturing brackets.
694    10. Changed the meaning of \X so that it now matches a Unicode extended
695        grapheme cluster.
697    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
698        out POSIX threads when JIT support is configured.
700    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
702    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
703        structure could go wrong in environments where size_t is not the same size
704        as int.
706    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
707        set.
709    15. The EBCDIC support had decayed; later updates to the code had included
710        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
711        general tidy up of EBCDIC-related issues, and the documentation was also
712        not quite right. There is now a test that can be run on ASCII systems to
713        check some of the EBCDIC-related things (but is it not a full test).
715    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
716        in a small tidy to the code.
718    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
720    18. If the --only-matching (-o) option in pcregrep is specified multiple
721        times, each one causes appropriate output. For example, -o1 -o2 outputs the
722        substrings matched by the 1st and 2nd capturing parentheses. A separating
723        string can be specified by --om-separator (default empty).
725    19. Improving the first n character searches.
727    20. Turn case lists for horizontal and vertical white space into macros so that
728        they are defined only once.
730    21. This set of changes together give more compatible Unicode case-folding
731        behaviour for characters that have more than one other case when UCP
732        support is available.
734        (a) The Unicode property table now has offsets into a new table of sets of
735            three or more characters that are case-equivalent. The MultiStage2.py
736            script that generates these tables (the pcre_ucd.c file) now scans
737            CaseFolding.txt instead of UnicodeData.txt for character case
738            information.
740        (b) The code for adding characters or ranges of characters to a character
741            class has been abstracted into a generalized function that also handles
742            case-independence. In UTF-mode with UCP support, this uses the new data
743            to handle characters with more than one other case.
745        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
746            whose other case is greater than 256 are now correctly matched
747            caselessly. Previously, the high codepoint matched the low one, but not
748            vice versa.
750        (d) The processing of \h, \H, \v, and \ in character classes now makes use
751            of the new class addition function, using character lists defined as
752            macros alongside the case definitions of 20 above.
754        (e) Caseless back references now work with characters that have more than
755            one other case.
757        (f) General caseless matching of characters with more than one other case
758            is supported.
760    22. Unicode character properties were updated from Unicode 6.2.0
762    23. Improved CMake support under Windows. Patch by Daniel Richard G.
764    24. Add support for 32-bit character strings, and UTF-32
766    25. Major JIT compiler update (code refactoring and bugfixing).
767        Experimental Sparc 32 support is added.
769    26. Applied a modified version of Daniel Richard G's patch to create
770        pcre.h.generic and config.h.generic by "make" instead of in the
771        PrepareRelease script.
773    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
774        pcre_compile.c when checking for a zero character.
776    28. Introducing a native interface for JIT. Through this interface, the compiled
777        machine code can be directly executed. The purpose of this interface is to
778        provide fast pattern matching, so several sanity checks are not performed.
779        However, feature tests are still performed. The new interface provides
780        1.4x speedup compared to the old one.
782    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
783        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
784        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
786    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
787        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
788        the "old" RFC 2279). Instead, it ended up passing a negative length to
789        pcre_exec().
791    31. Add support for GCC's visibility feature to hide internal functions.
793    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
794        "unknown -C option" after outputting 0 or 1.
796    33. There is now support for generating a code coverage report for the test
797        suite in environments where gcc is the compiler and lcov is installed. This
798        is mainly for the benefit of the developers.
800    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
801        unaddressable using valgrind annotations, allowing valgrind to detect
802        invalid memory accesses. This is mainly for the benefit of the developers.
804    25. (*UTF) can now be used to start a pattern in any of the three libraries.
806    26. Give configure error if --enable-cpp but no C++ compiler found.
809    Version 8.31 06-July-2012
810    -------------------------
812    1.  Fixing a wrong JIT test case and some compiler warnings.
814    2.  Removed a bashism from the RunTest script.
816    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
817        to unsigned type, result still unsigned" that was given by an MS compiler
818        on encountering the code "-sizeof(xxx)".
820    4.  Partial matching support is added to the JIT compiler.
822    5.  Fixed several bugs concerned with partial matching of items that consist
823        of more than one character:
825        (a) /^(..)\1/ did not partially match "aba" because checking references was
826            done on an "all or nothing" basis. This also applied to repeated
827            references.
829        (b) \R did not give a hard partial match if \r was found at the end of the
830            subject.
832        (c) \X did not give a hard partial match after matching one or more
833            characters at the end of the subject.
835        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
836            a partial match for the string "\r".
838        (e) When newline was set to CRLF, the metacharacter "." did not recognize
839            a partial match for a CR character at the end of the subject string.
841    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
842        running pcretest, the text "(JIT)" added to the output whenever JIT is
843        actually used to run the match.
845    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
846        or /S+[+] with a digit between 1 and 7.
848    8.  OP_NOT now supports any UTF character not just single-byte ones.
850    9.  (*MARK) control verb is now supported by the JIT compiler.
852    10. The command "./RunTest list" lists the available tests without actually
853        running any of them. (Because I keep forgetting what they all are.)
857    12. Applied a (slightly modified) user-supplied patch that improves performance
858        when the heap is used for recursion (compiled with --disable-stack-for-
859        recursion). Instead of malloc and free for each heap frame each time a
860        logical recursion happens, frames are retained on a chain and re-used where
861        possible. This sometimes gives as much as 30% improvement.
863    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
864        call.
866    14. As documented, (*COMMIT) is now confined to within a positive assertion.
868    15. It is now possible to link pcretest with libedit as an alternative to
869        libreadline.
871    16. (*COMMIT) control verb is now supported by the JIT compiler.
873    17. The Unicode data tables have been updated to Unicode 6.1.0.
875    18. Added --file-list option to pcregrep.
877    19. Added binary file support to pcregrep, including the -a, --binary-files,
878        -I, and --text options.
880    20. The madvise function is renamed for posix_madvise for QNX compatibility
881        reasons. Fixed by Giuseppe D'Angelo.
883    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
885    22. Changed the option for creating symbolic links for 16-bit man pages from
886        -s to -sf so that re-installing does not cause issues.
888    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
890    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
891        matches in certain environments (the workspace was not being correctly
892        retained). Also added to pcre_dfa_exec() a simple plausibility check on
893        some of the workspace data at the beginning of a restart.
895    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
896        was not doing so when it should - probably a typo introduced by SVN 528
897        (change 8.10/14).
899    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
900        \w+ when the character tables indicated that \x{c4} was a word character.
901        There were several related cases, all because the tests for doing a table
902        lookup were testing for characters less than 127 instead of 255.
904    27. If a pattern contains capturing parentheses that are not used in a match,
905        their slots in the ovector are set to -1. For those that are higher than
906        any matched groups, this happens at the end of processing. In the case when
907        there were back references that the ovector was too small to contain
908        (causing temporary malloc'd memory to be used during matching), and the
909        highest capturing number was not used, memory off the end of the ovector
910        was incorrectly being set to -1. (It was using the size of the temporary
911        memory instead of the true size.)
913    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
914        ovector size, it uses memory at the end of the block that it has got.
916    29. Check for an overlong MARK name and give an error at compile time. The
917        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
919    30. JIT compiler update.
921    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
922        Rill for the patch.
924    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
926    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
928    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
929        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
930        was enabled.
932    35. Improve JIT code generation for greedy plus quantifier.
934    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
935        1 to "aa" instead of to an empty string. The bug affected repeated groups
936        that could potentially match an empty string.
938    37. Optimizing single character iterators in JIT.
940    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
941        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
942        codepoints that are too big for the mode are faulted, and in a UTF mode,
943        disallowed codepoints are also faulted.
945    39. If PCRE was compiled with UTF support, in three places in the DFA
946        matcher there was code that should only have been obeyed in UTF mode, but
947        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
948        processing when bytes with values greater than 127 were present. In 16-bit
949        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
950        both cases the values are those that cannot be the first data item in a UTF
951        character. The three items that might have provoked this were recursions,
952        possessively repeated groups, and atomic groups.
954    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
955        and pcregrep, because some OS require shared objects to be explicitly
956        passed to ld, causing the link step to fail if they are not.
958    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
959        mode, patterns that started with \h* or \R* might be incorrectly matched.
962    Version 8.30 04-February-2012
963    -----------------------------
965    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
966        name is defined in ctype.h.
968    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
969        only in quite long subpatterns.
971    3.  Removed the function pcre_info(), which has been obsolete and deprecated
972        since it was replaced by pcre_fullinfo() in February 2000.
974    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
975        match a (*MARK), and the match failed at the start of the subject, a
976        reference to memory before the start of the subject could occur. This bug
977        was introduced by fix 17 of release 8.21.
979    5.  A reference to an unset group with zero minimum repetition was giving
980        totally wrong answers (in non-JavaScript-compatibility mode). For example,
981        /(another)?(\1?)test/ matched against "hello world test". This bug was
982        introduced in release 8.13.
984    6.  Add support for 16-bit character strings (a large amount of work involving
985        many changes and refactorings).
987    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
988        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
989        from a file.
991    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
992        rounding is not applied in this particular case).
994    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
995        if they appear, or are escaped, in patterns.
997    10. Get rid of a number of -Wunused-but-set-variable warnings.
999    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1000        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1001        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1002        also returns the mark "x". This bug applied to capturing parentheses,
1003        non-capturing parentheses, and atomic parentheses. It also applied to some
1004        assertions.
1006    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1007        information out of configure.ac instead of relying on pcre.h.generic, which
1008        is not stored in the repository.
1010    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1011        -lreadline.
1013    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1015    15. Applied Graycode's patch to put the top-level frame on the stack rather
1016        than the heap when not using the stack for recursion. This gives a
1017        performance improvement in many cases when recursion is not deep.
1019    16. Experimental code added to "pcretest -C" to output the stack frame size.
1022    Version 8.21 12-Dec-2011
1023    ------------------------
1025    1.  Updating the JIT compiler.
1027    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1028        are added as well.
1030    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1031        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1032        calling _pcre_jit_exec. Some extra comments are added.
1034    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1035        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1036        was introduced by change 18 for 8.20.
1038    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1039        ECMA-262 standard.
1041    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1042        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1043        This bug was probably introduced by change 9 of 8.13.
1045    7.  While fixing 6 above, I noticed that a number of other items were being
1046        incorrectly rejected as "not fixed length". This arose partly because newer
1047        opcodes had not been added to the fixed-length checking code. I have (a)
1048        corrected the bug and added tests for these items, and (b) arranged for an
1049        error to occur if an unknown opcode is encountered while checking for fixed
1050        length instead of just assuming "not fixed length". The items that were
1051        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1052        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1053        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1055    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1056        being incorrectly compiled and would have given unpredicatble results.
1058    9.  A possessively repeated subpattern with minimum repeat count greater than
1059        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1060        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1061        the first (A) could occur when it should not.
1063    10. Add a cast and remove a redundant test from the code.
1065    11. JIT should use pcre_malloc/pcre_free for allocation.
1067    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1068        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1069        is anything other than /usr, -L is still shown).
1071    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1073    14. Perl does not support \N without a following name in a [] class; PCRE now
1074        also gives an error.
1076    15. If a forward reference was repeated with an upper limit of around 2000,
1077        it caused the error "internal error: overran compiling workspace". The
1078        maximum number of forward references (including repeats) was limited by the
1079        internal workspace, and dependent on the LINK_SIZE. The code has been
1080        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1081        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1082        of around 200,000 forward references. While doing this, I also speeded up
1083        the filling in of repeated forward references.
1085    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1086        incorrectly expecting the subject to contain another "a" after the start.
1088    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1089        in the match, the SKIP should be ignored. This was not happening; instead
1090        the SKIP was being treated as NOMATCH. For patterns such as
1091        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1092        tested.
1094    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1095        now much more compatible with Perl, in particular in cases where the result
1096        is a non-match for a non-anchored pattern. For example, if
1097        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1098        "m", where previously it did not return a name. A side effect of this
1099        change is that for partial matches, the last encountered mark name is
1100        returned, as for non matches. A number of tests that were previously not
1101        Perl-compatible have been moved into the Perl-compatible test files. The
1102        refactoring has had the pleasing side effect of removing one argument from
1103        the match() function, thus reducing its stack requirements.
1105    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1106        subsequent uses of /S (without +) incorrectly behaved like /S+.
1108    21. Retrieve executable code size support for the JIT compiler and fixing
1109        some warnings.
1111    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1112        not work when the shorter character appeared right at the end of the
1113        subject string.
1115    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1116        systems.
1118    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1119        output it when the /M option is used in pcretest.
1121    25. The CheckMan script was not being included in the distribution. Also, added
1122        an explicit "perl" to run Perl scripts from the PrepareRelease script
1123        because this is reportedly needed in Windows.
1125    26. If study data was being save in a file and studying had not found a set of
1126        "starts with" bytes for the pattern, the data written to the file (though
1127        never used) was taken from uninitialized memory and so caused valgrind to
1128        complain.
1130    27. Updated RunTest.bat as provided by Sheri Pierce.
1132    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1134    29. Computation of memory usage for the table of capturing group names was
1135        giving an unnecessarily large value.
1138    Version 8.20 21-Oct-2011
1139    ------------------------
1141    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1142        a POSIX class. After further experiments with Perl, which convinced me that
1143        Perl has bugs and confusions, a closing square bracket is no longer allowed
1144        in a POSIX name. This bug also affected patterns with classes that started
1145        with full stops.
1147    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1148        captured substring, but while checking the failing first alternative,
1149        substring 1 is temporarily captured. If the output vector supplied to
1150        pcre_exec() was not big enough for this capture, the yield of the function
1151        was still zero ("insufficient space for captured substrings"). This cannot
1152        be totally fixed without adding another stack variable, which seems a lot
1153        of expense for a edge case. However, I have improved the situation in cases
1154        such as /(a)(b)x|abc/ matched against "abc", where the return code
1155        indicates that fewer than the maximum number of slots in the ovector have
1156        been set.
1158    3.  Related to (2) above: when there are more back references in a pattern than
1159        slots in the output vector, pcre_exec() uses temporary memory during
1160        matching, and copies in the captures as far as possible afterwards. It was
1161        using the entire output vector, but this conflicts with the specification
1162        that only 2/3 is used for passing back captured substrings. Now it uses
1163        only the first 2/3, for compatibility. This is, of course, another edge
1164        case.
1166    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1167        main code base, and can be used by building with --enable-jit. When this is
1168        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1169        runtime --no-jit option is given.
1171    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1172        ovector, the return from the function was zero, implying that there were
1173        other matches that did not fit. The correct "exactly full" value is now
1174        returned.
1176    6.  If a subpattern that was called recursively or as a subroutine contained
1177        (*PRUNE) or any other control that caused it to give a non-standard return,
1178        invalid errors such as "Error -26 (nested recursion at the same subject
1179        position)" or even infinite loops could occur.
1181    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1182        computing the minimum length on reaching *ACCEPT, and so ended up with the
1183        wrong value of 1 rather than 0. Further investigation indicates that
1184        computing a minimum subject length in the presence of *ACCEPT is difficult
1185        (think back references, subroutine calls), and so I have changed the code
1186        so that no minimum is registered for a pattern that contains *ACCEPT.
1188    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1189        it was not handled as intended. [But see 16 below.]
1191    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1192        Sheri Pierce.
1194    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1195        the first byte in a match must be "a".
1197    11. Change 17 for 8.13 increased the recursion depth for patterns like
1198        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1199        pattern contains any instances of (*THEN). If it does not, the old
1200        optimizations are restored. It would be nice to do this on a per-group
1201        basis, but at the moment that is not feasible.
1203    12. In some environments, the output of pcretest -C is CRLF terminated. This
1204        broke RunTest's code that checks for the link size. A single white space
1205        character after the value is now allowed for.
1207    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1208        For "fr", it uses the Windows-specific input and output files.
1210    14. If (*THEN) appeared in a group that was called recursively or as a
1211        subroutine, it did not work as intended. [But see next item.]
1213    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1214        pattern fragments (but not containing any | characters). If A and B are
1215        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1216        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1217        D. In other words, Perl considers parentheses that do not contain any |
1218        characters to be part of a surrounding alternative, whereas PCRE was
1219        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1220        differently. PCRE now behaves in the same way as Perl, except in the case
1221        of subroutine/recursion calls such as (?1) which have in any case always
1222        been different (but PCRE had them first :-).
1224    16. Related to 15 above: Perl does not treat the | in a conditional group as
1225        creating alternatives. Such a group is treated in the same way as an
1226        ordinary group without any | characters when processing (*THEN). PCRE has
1227        been changed to match Perl's behaviour.
1229    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1230        RunGrepTest script failed.
1232    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1233        inevitable for groups that contain captures, but it can lead to a lot of
1234        stack use in large patterns. The old behaviour has been restored for atomic
1235        groups that do not contain any capturing parentheses.
1237    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1238        suppress the check for a minimum subject length at run time. (If it was
1239        given to pcre_exec() or pcre_dfa_exec() it did work.)
1241    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1242        fail to work when decoding hex characters in data strings in EBCDIC
1243        environments.
1245    21. It appears that in at least one Mac OS environment, the isxdigit() function
1246        is implemented as a macro that evaluates to its argument more than once,
1247        contravening the C 90 Standard (I haven't checked a later standard). There
1248        was an instance in pcretest which caused it to go wrong when processing
1249        \x{...} escapes in subject strings. The has been rewritten to avoid using
1250        things like p++ in the argument of isxdigit().
1253    Version 8.13 16-Aug-2011
1254    ------------------------
1256    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1258    2.  Two minor typos in pcre_internal.h have been fixed.
1260    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1261        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1262        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1264    4.  There were a number of related bugs in the code for matching backrefences
1265        caselessly in UTF-8 mode when codes for the characters concerned were
1266        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1267        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1268        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1269        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1270        2-byte code at the end of the subject (it thought there wasn't enough data
1271        left).
1273    5.  Comprehensive information about what went wrong is now returned by
1274        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1275        as the output vector has at least 2 elements. The offset of the start of
1276        the failing character and a reason code are placed in the vector.
1278    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1279        now returned is for the first byte of the failing character, instead of the
1280        last byte inspected. This is an incompatible change, but I hope it is small
1281        enough not to be a problem. It makes the returned offset consistent with
1282        pcre_exec() and pcre_dfa_exec().
1284    7.  pcretest now gives a text phrase as well as the error number when
1285        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1286        failure, the offset and reason code are output.
1288    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1289        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1290        back over a single character (\n). This seems wrong (because it treated the
1291        two characters as a single entity when going forwards), conflicts with the
1292        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1293        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1294        has been changed.
1296    9.  Some internal refactoring has changed the processing so that the handling
1297        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1298        time (the PCRE_DOTALL option was changed this way some time ago: version
1299        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1300        which was always a bit of a fudge. It also means that there is one less
1301        argument for the match() function, which reduces its stack requirements
1302        slightly. This change also fixes an incompatibility with Perl: the pattern
1303        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1305    10. More internal refactoring has drastically reduced the number of recursive
1306        calls to match() for possessively repeated groups such as (abc)++ when
1307        using pcre_exec().
1309    11. While implementing 10, a number of bugs in the handling of groups were
1310        discovered and fixed:
1312        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1313        (a|)*(?1) gave a compile-time internal error.
1314        ((a|)+)+  did not notice that the outer group could match an empty string.
1315        (^a|^)+   was not marked as anchored.
1316        (.*a|.*)+ was not marked as matching at start or after a newline.
1318    12. Yet more internal refactoring has removed another argument from the match()
1319        function. Special calls to this function are now indicated by setting a
1320        value in a variable in the "match data" data block.
1322    13. Be more explicit in pcre_study() instead of relying on "default" for
1323        opcodes that mean there is no starting character; this means that when new
1324        ones are added and accidentally left out of pcre_study(), testing should
1325        pick them up.
1327    14. The -s option of pcretest has been documented for ages as being an old
1328        synonym of -m (show memory usage). I have changed it to mean "force study
1329        for every regex", that is, assume /S for every regex. This is similar to -i
1330        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1331        using it. It makes it easier to run collections of tests with and without
1332        study enabled, and thereby test pcre_study() more easily. All the standard
1333        tests are now run with and without -s (but some patterns can be marked as
1334        "never study" - see 20 below).
1336    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1337        restoration of the capturing data to the outer values was not happening
1338        correctly.
1340    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1341        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1342        pattern had matched an empty string, and so incorrectly returned a no
1343        match.
1345    17. There was optimizing code for the last branch of non-capturing parentheses,
1346        and also for the obeyed branch of a conditional subexpression, which used
1347        tail recursion to cut down on stack usage. Unfortunately, now that there is
1348        the possibility of (*THEN) occurring in these branches, tail recursion is
1349        no longer possible because the return has to be checked for (*THEN). These
1350        two optimizations have therefore been removed. [But see 8.20/11 above.]
1352    18. If a pattern containing \R was studied, it was assumed that \R always
1353        matched two bytes, thus causing the minimum subject length to be
1354        incorrectly computed because \R can also match just one byte.
1356    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1357        was incorrectly computed.
1359    20. If /S is present twice on a test pattern in pcretest input, it now
1360        *disables* studying, thereby overriding the use of -s on the command line
1361        (see 14 above). This is necessary for one or two tests to keep the output
1362        identical in both cases.
1364    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1365        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1367    22. When an atomic group that contained a capturing parenthesis was
1368        successfully matched, but the branch in which it appeared failed, the
1369        capturing was not being forgotten if a higher numbered group was later
1370        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1371        group 1 to "a", when in fact it should be unset. This applied to multi-
1372        branched capturing and non-capturing groups, repeated or not, and also to
1373        positive assertions (capturing in negative assertions does not happen
1374        in PCRE) and also to nested atomic groups.
1376    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1377        subject after a captured substring, to make it easier to tell which of a
1378        number of identical substrings has been captured.
1380    24. The way atomic groups are processed by pcre_exec() has been changed so that
1381        if they are repeated, backtracking one repetition now resets captured
1382        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1383        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1384        "aaa". Previously, it would have been "a". As part of this code
1385        refactoring, the way recursive calls are handled has also been changed.
1387    25. If an assertion condition captured any substrings, they were not passed
1388        back unless some other capturing happened later. For example, if
1389        (?(?=(a))a) was matched against "a", no capturing was returned.
1391    26. When studying a pattern that contained subroutine calls or assertions,
1392        the code for finding the minimum length of a possible match was handling
1393        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1394        group 1 called group 2 while simultaneously a separate group 2 called group
1395        1). A stack overflow occurred in this case. I have fixed this by limiting
1396        the recursion depth to 10.
1398    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1399        Fortmann. This supports explicit test numbers on the command line, and has
1400        argument validation and error reporting.
1402    28. An instance of \X with an unlimited repeat could fail if at any point the
1403        first character it looked at was a mark character.
1405    29. Some minor code refactoring concerning Unicode properties and scripts
1406        should reduce the stack requirement of match() slightly.
1408    30. Added the '=' option to pcretest to check the setting of unused capturing
1409        slots at the end of the pattern, which are documented as being -1, but are
1410        not included in the return count.
1412    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1413        compiled something random. Now it gives a compile-time error (as does
1414        Perl).
1416    32. A *MARK encountered during the processing of a positive assertion is now
1417        recorded and passed back (compatible with Perl).
1419    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1420        had alternative anchored branches, the search for a second match in a line
1421        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1422        matched the line "0102" twice. The same bug affected patterns that started
1423        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1424        twice.
1426    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1427        does, and because of capturing effects, quantifying parenthesized
1428        assertions may at times be useful. Quantifiers are now allowed for
1429        parenthesized assertions.
1431    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1433    36. \g was being checked for fancy things in a character class, when it should
1434        just be a literal "g".
1436    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1437        appearance of a nested POSIX class supersedes an apparent external class.
1438        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1439        unescaped square brackets may also appear as part of class names. For
1440        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1441        more like Perl. (But see 8.20/1 above.)
1443    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1444        was because it thought it was \N{name}, which is not supported).
1446    39. Add minix to OS list not supporting the -S option in pcretest.
1448    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1449        cannot analyze patterns in sufficient detail to catch mutual recursions
1450        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1451        subgroup is called recursively as a subpattern for a second time at the
1452        same position in the subject string. In previous releases this might have
1453        been caught by the recursion limit, or it might have run out of stack.
1455    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1456        happen only once. PCRE was, however incorrectly giving a compile time error
1457        "recursive call could loop indefinitely" because it cannot analyze the
1458        pattern in sufficient detail. The compile time test no longer happens when
1459        PCRE is compiling a conditional subpattern, but actual runaway loops are
1460        now caught at runtime (see 40 above).
1462    42. It seems that Perl allows any characters other than a closing parenthesis
1463        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1464        has been changed to be the same.
1466    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1467        as not to get warnings when autogen.sh is called. Also changed
1468        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1470    44. To help people who use pcregrep to scan files containing exceedingly long
1471        lines, the following changes have been made:
1473        (a) The default value of the buffer size parameter has been increased from
1474            8K to 20K. (The actual buffer used is three times this size.)
1476        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1477            PCRE is built.
1479        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1480            to be set at run time.
1482        (d) Numerical values in pcregrep options can be followed by K or M, for
1483            example --buffer-size=50K.
1485        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1486            given and the return code is set to 2.
1488    45. Add a pointer to the latest mark to the callout data block.
1490    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1491        partial match of an empty string instead of no match. This was specific to
1492        the use of ".".
1494    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1495        complete match instead of a partial match. This bug was dependent on both
1496        the PCRE_UTF8 and PCRE_DOTALL options being set.
1498    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1499        starting byte set, because \b was not being ignored.
1502    Version 8.12 15-Jan-2011
1503    ------------------------
1505    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1506        checks for such things as part of the documentation building process.
1508    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1509        --match-limit and --recursion-limit options (added for 8.11). In
1510        particular, this made one of the standard tests fail. (The integer value
1511        went into the wrong half of a long int.)
1513    3.  If the --colour option was given to pcregrep with -v (invert match), it
1514        did strange things, either producing crazy output, or crashing. It should,
1515        of course, ignore a request for colour when reporting lines that do not
1516        match.
1518    4.  Another pcregrep bug caused similar problems if --colour was specified with
1519        -M (multiline) and the pattern match finished with a line ending.
1521    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1522        matched in multiline mode, the following line was shown as part of the
1523        match. This seems wrong, so I have changed it.
1525    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1526        the check for further matches in the same line (so they could be coloured)
1527        to overrun the end of the current line. If another match was found, it was
1528        incorrectly shown (and then shown again when found in the next line).
1530    7.  If pcregrep was compiled under Windows, there was a reference to the
1531        function pcregrep_exit() before it was defined. I am assuming this was
1532        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1533        reported by a user. I've moved the definition above the reference.
1536    Version 8.11 10-Dec-2010
1537    ------------------------
1539    1.  (*THEN) was not working properly if there were untried alternatives prior
1540        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1541        backtracked to try for "b" instead of moving to the next alternative branch
1542        at the same level (in this case, to look for "c"). The Perl documentation
1543        is clear that when (*THEN) is backtracked onto, it goes to the "next
1544        alternative in the innermost enclosing group".
1546    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1547        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1548        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1549        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1550        (*THEN).
1552    3.  If \s appeared in a character class, it removed the VT character from
1553        the class, even if it had been included by some previous item, for example
1554        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1555        of \s, but is part of the POSIX "space" class.)
1557    4.  A partial match never returns an empty string (because you can always
1558        match an empty string at the end of the subject); however the checking for
1559        an empty string was starting at the "start of match" point. This has been
1560        changed to the "earliest inspected character" point, because the returned
1561        data for a partial match starts at this character. This means that, for
1562        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1563        (previously it gave "no match").
1565    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1566        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1567        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1568        has an implication that the given string is incomplete (because a partial
1569        match is preferred over a full match). For this reason, these items now
1570        give a partial match in this situation. [Aside: previously, the one case
1571        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1572        match rather than a full match, which was wrong by the old rules, but is
1573        now correct.]
1575    6.  There was a bug in the handling of #-introduced comments, recognized when
1576        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1577        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1578        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1579        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1580        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1581        places in pcre_compile().
1583    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1584        comments when looking ahead for named forward references to subpatterns,
1585        the only newline sequence it recognized was NL. It now handles newlines
1586        according to the set newline convention.
1588    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1589        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1590        cater for a lack of strerror(). These oversights have been fixed.
1592    9.  Added --match-limit and --recursion-limit to pcregrep.
1594    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1596    11. When the -o option was used, pcregrep was setting a return code of 1, even
1597        when matches were found, and --line-buffered was not being honoured.
1599    12. Added an optional parentheses number to the -o and --only-matching options
1600        of pcregrep.
1602    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1603        can match an empty string. The code to do it in pcretest and pcredemo
1604        needed fixing:
1606        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1607            only one byte after an empty string match just before CRLF (this case
1608            just got forgotten; "any" and "anycrlf" were OK).
1610        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1611            mode when an empty string match preceded an ASCII character followed by
1612            a non-ASCII character. (The code for advancing by one character rather
1613            than one byte was nonsense.)
1615        (c) The pcredemo.c sample program did not have any code at all to handle
1616            the cases when CRLF is a valid newline sequence.
1618    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1619        as a starting offset was within the subject string. There is now a new
1620        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1621        negative or greater than the length of the string. In order to test this,
1622        pcretest is extended to allow the setting of negative starting offsets.
1624    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1625        starting offset points to the beginning of a UTF-8 character was
1626        unnecessarily clumsy. I tidied it up.
1628    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1629        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1631    17. Nobody had reported that the --include_dir option, which was added in
1632        release 7.7 should have been called --include-dir (hyphen, not underscore)
1633        for compatibility with GNU grep. I have changed it to --include-dir, but
1634        left --include_dir as an undocumented synonym, and the same for
1635        --exclude-dir, though that is not available in GNU grep, at least as of
1636        release 2.5.4.
1638    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1639        characters from a string of bytes) have been redefined so as not to use
1640        loops, in order to improve performance in some environments. At the same
1641        time, I abstracted some of the common code into auxiliary macros to save
1642        repetition (this should not affect the compiled code).
1644    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1645        compile-time error is now given if \c is not followed by an ASCII
1646        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1647        different, and any byte value is allowed.)
1649    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1650        START_OPTIMIZE option, which is now allowed at compile time - but just
1651        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1652        to pcregrep and other applications that have no direct access to PCRE
1653        options. The new /Y option in pcretest sets this option when calling
1654        pcre_compile().
1656    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1657        back references. Groups containing recursive back references were forced to
1658        be atomic by that change, but in the case of named groups, the amount of
1659        memory required was incorrectly computed, leading to "Failed: internal
1660        error: code overflow". This has been fixed.
1662    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1663        pcretest.c, to avoid build problems in some Borland environments.
1666    Version 8.10 25-Jun-2010
1667    ------------------------
1669    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1670        THEN.
1672    2.  (*ACCEPT) was not working when inside an atomic group.
1674    3.  Inside a character class, \B is treated as a literal by default, but
1675        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1676        causes the error). The code is unchanged, but I tidied the documentation.
1678    4.  Inside a character class, PCRE always treated \R and \X as literals,
1679        whereas Perl faults them if its -w option is set. I have changed PCRE so
1680        that it faults them when PCRE_EXTRA is set.
1682    5.  Added support for \N, which always matches any character other than
1683        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1685    6.  When compiling pcregrep with newer versions of gcc which may have
1686        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1687        declared with attribute warn_unused_result" were given. Just casting the
1688        result to (void) does not stop the warnings; a more elaborate fudge is
1689        needed. I've used a macro to implement this.
1691    7.  Minor change to pcretest.c to avoid a compiler warning.
1693    8.  Added four artifical Unicode properties to help with an option to make
1694        \s etc use properties (see next item). The new properties are: Xan
1695        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1697    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1698        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1699        this option. Modified pcretest to add /W to test this facility. Added
1700        REG_UCP to make it available via the POSIX interface.
1702    10. Added --line-buffered to pcregrep.
1704    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1705        studied, and the match started with a letter with a code point greater than
1706        127 whose first byte was different to the first byte of the other case of
1707        the letter, the other case of this starting letter was not recognized
1708        (#976).
1710    12. If a pattern that was studied started with a repeated Unicode property
1711        test, for example, \p{Nd}+, there was the theoretical possibility of
1712        setting up an incorrect bitmap of starting bytes, but fortunately it could
1713        not have actually happened in practice until change 8 above was made (it
1714        added property types that matched character-matching opcodes).
1716    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1717        possible starting bytes for non-anchored patterns.
1719    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1720        \R, and also a number of cases that involve Unicode properties, both
1721        explicit and implicit when PCRE_UCP is set.
1723    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1724        input, it could crash or give wrong results if characters with values
1725        greater than 0xc0 were present in the subject string. (Detail: it assumed
1726        UTF-8 input when processing these items.)
1728    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1729        size_t is 64-bit (#991).
1731    17. Added a check for running out of memory when PCRE is compiled with
1732        --disable-stack-for-recursion (#990).
1734    18. If the last data line in a file for pcretest does not have a newline on
1735        the end, a newline was missing in the output.
1737    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1738        less than 128) in its various bitmaps. However, there is a facility for
1739        generating tables according to the current locale when PCRE is compiled. It
1740        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1741        characters, are recognized by isspace() and therefore were getting set in
1742        these tables, and indeed these tables seem to approximate to ISO 8859. This
1743        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1744        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1745        which of course cannot start UTF-8 characters. I have changed the code so
1746        that only real ASCII characters (less than 128) and the correct starting
1747        bytes for UTF-8 encodings are set for characters greater than 127 when in
1748        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1749        altogether.)
1751    20. Added the /T option to pcretest so as to be able to run tests with non-
1752        standard character tables, thus making it possible to include the tests
1753        used for 19 above in the standard set of tests.
1755    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1756        reference to a subpattern the other side of a comment that contains an
1757        opening parenthesis caused either an internal compiling error, or a
1758        reference to the wrong subpattern.
1761    Version 8.02 19-Mar-2010
1762    ------------------------
1764    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1766    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1767        configured.
1769    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1770        original author of that file, following a query about its status.
1772    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1773        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1775    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1776        quantifier applied to a forward-referencing subroutine call, could compile
1777        incorrect code or give the error "internal error: previously-checked
1778        referenced subpattern not found".
1780    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1781        variables to point to external functions. For these systems, therefore,
1782        pcre_malloc etc. are now initialized to local functions that call the
1783        relevant global functions.
1785    7.  There were two entries missing in the vectors called coptable and poptable
1786        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1787        I've fixed the data, and added a kludgy way of testing at compile time that
1788        the lengths are correct (equal to the number of opcodes).
1790    8.  Following on from 7, I added a similar kludge to check the length of the
1791        eint vector in pcreposix.c.
1793    9.  Error texts for pcre_compile() are held as one long string to avoid too
1794        much relocation at load time. To find a text, the string is searched,
1795        counting zeros. There was no check for running off the end of the string,
1796        which could happen if a new error number was added without updating the
1797        string.
1799    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1801    11. \K was not working if it appeared in an atomic group or in a group that
1802        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1803        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1804        the assertion is positive, but not if it is negative.
1806    12. Change 11 fortuitously reduced the size of the stack frame used in the
1807        "match()" function of pcre_exec.c by one pointer. Forthcoming
1808        implementation of support for (*MARK) will need an extra pointer on the
1809        stack; I have reserved it now, so that the stack frame size does not
1810        decrease.
1812    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1813        item in branch that calls a recursion is a subroutine call - as in the
1814        second branch in the above example - was incorrectly given the compile-
1815        time error "recursive call could loop indefinitely" because pcre_compile()
1816        was not correctly checking the subroutine for matching a non-empty string.
1818    14. The checks for overrunning compiling workspace could trigger after an
1819        overrun had occurred. This is a "should never occur" error, but it can be
1820        triggered by pathological patterns such as hundreds of nested parentheses.
1821        The checks now trigger 100 bytes before the end of the workspace.
1823    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1826    Version 8.01 19-Jan-2010
1827    ------------------------
1829    1.  If a pattern contained a conditional subpattern with only one branch (in
1830        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1831        computed the wrong minimum data length (which is of course zero for such
1832        subpatterns). This could cause incorrect "no match" results.
1834    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1835        the pattern is reset in the first branch, pcre_compile() failed with
1836        "internal error: code overflow at offset...". This happened only when
1837        the reset was to the original external option setting. (An optimization
1838        abstracts leading options settings into an external setting, which was the
1839        cause of this.)
1841    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1842        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1843        assertion pattern did not match (meaning that the assertion was true), it
1844        was incorrectly treated as false if the SKIP had been reached during the
1845        matching. This also applied to assertions used as conditions.
1847    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1848        assertion subpattern, including such a pattern used as a condition,
1849        unpredictable results occurred, instead of the error return
1852    5.  The C++ GlobalReplace function was not working like Perl for the special
1853        situation when an empty string is matched. It now does the fancy magic
1854        stuff that is necessary.
1856    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1857        removed. (These were left over from very, very early versions of PCRE.)
1859    7.  Some cosmetic changes to the code to make life easier when compiling it
1860        as part of something else:
1862        (a) Change DEBUG to PCRE_DEBUG.
1864        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1865            called "current" as "current_branch", to prevent a collision with the
1866            Linux macro when compiled as a kernel module.
1868        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1869            prevent a collision with the Linux macro when compiled as a kernel
1870            module.
1872    8.  In pcre_compile() there are some checks for integer overflows that used to
1873        cast potentially large values to (double). This has been changed to that
1874        when building, a check for int64_t is made, and if it is found, it is used
1875        instead, thus avoiding the use of floating point arithmetic. (There is no
1876        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1877        double.
1879    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1880        2005 (difference between two addresses compared to an unsigned value).
1882    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1883        custom one, because of the following reported problem in Windows:
1885          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1886              under Win32.
1887          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1888              therefore missing the function definition.
1889          - The compiler thus generates a "C" signature for the test function.
1890          - The linker fails to find the "C" function.
1891          - PCRE fails to configure if asked to do so against libbz2.
1893    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1894        messages were output:
1896          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1897          rerunning libtoolize, to keep the correct libtool macros in-tree.
1898          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1900        I have done both of these things.
1902    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1903        most of the time, it *can* run out if it is given a pattern that contains a
1904        runaway infinite recursion. I updated the discussion in the pcrestack man
1905        page.
1907    13. Now that we have gone to the x.xx style of version numbers, the minor
1908        version may start with zero. Using 08 or 09 is a bad idea because users
1909        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1910        interpreted as invalid octal numbers. I've updated the previous comment in
1911        configure.ac, and also added a check that gives an error if 08 or 09 are
1912        used.
1914    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1915        causing partial matching to fail when the end of the subject matched \W
1916        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1918    15. There were some discrepancies between the declarations in pcre_internal.h
1919        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1920        their definitions. The declarations used "const uschar *" and the
1921        definitions used USPTR. Even though USPTR is normally defined as "const
1922        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1923        reported that: "This difference in casting confuses some C++ compilers, for
1924        example, SunCC recognizes above declarations as different functions and
1925        generates broken code for hbpcre." I have changed the declarations to use
1926        USPTR.
1928    16. GNU libtool is named differently on some systems. The autogen.sh script now
1929        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1930        (FreeBSD).
1932    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1933        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1934        comment: "Figure out how to create a longlong from a string: strtoll and
1935        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1936        instance, but it only takes 2 args instead of 3!"
1938    18. A subtle bug concerned with back references has been fixed by a change of
1939        specification, with a corresponding code fix. A pattern such as
1940        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1941        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1942        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1943        same bug. Such groups have to be quantified to be useful, or contained
1944        inside another quantified group. (If there's no repetition, the reference
1945        can never match.) The problem arises because, having left the group and
1946        moved on to the rest of the pattern, a later failure that backtracks into
1947        the group uses the captured value from the final iteration of the group
1948        rather than the correct earlier one. I have fixed this in PCRE by forcing
1949        any group that contains a reference to itself to be an atomic group; that
1950        is, there cannot be any backtracking into it once it has completed. This is
1951        similar to recursive and subroutine calls.
1954    Version 8.00 19-Oct-09
1955    ----------------------
1957    1.  The table for translating pcre_compile() error codes into POSIX error codes
1958        was out-of-date, and there was no check on the pcre_compile() error code
1959        being within the table. This could lead to an OK return being given in
1960        error.
1962    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1963        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1964        in a Windows environment.
1966    3.  The pcregrep --count option prints the count for each file even when it is
1967        zero, as does GNU grep. However, pcregrep was also printing all files when
1968        --files-with-matches was added. Now, when both options are given, it prints
1969        counts only for those files that have at least one match. (GNU grep just
1970        prints the file name in this circumstance, but including the count seems
1971        more useful - otherwise, why use --count?) Also ensured that the
1972        combination -clh just lists non-zero counts, with no names.
1974    4.  The long form of the pcregrep -F option was incorrectly implemented as
1975        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1976        but it seems right to fix it, and I didn't think it was worth preserving
1977        the old behaviour.
1979    5.  The command line items --regex=pattern and --regexp=pattern were not
1980        recognized by pcregrep, which required --regex pattern or --regexp pattern
1981        (with a space rather than an '='). The man page documented the '=' forms,
1982        which are compatible with GNU grep; these now work.
1984    6.  No libpcreposix.pc file was created for pkg-config; there was just
1985        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1987    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1988        when UCP support is not needed, by modifying the Python script that
1989        generates it from Unicode data files. This should not matter if the module
1990        is correctly used as a library, but I received one complaint about 50K of
1991        unwanted data. My guess is that the person linked everything into his
1992        program rather than using a library. Anyway, it does no harm.
1994    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1995        was a minimum greater than 1 for a wide character in a possessive
1996        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1997        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1998        character. Chaos in the form of incorrect output or a compiling loop could
1999        result.
2001    9.  The restrictions on what a pattern can contain when partial matching is
2002        requested for pcre_exec() have been removed. All patterns can now be
2003        partially matched by this function. In addition, if there are at least two
2004        slots in the offset vector, the offset of the earliest inspected character
2005        for the match and the offset of the end of the subject are set in them when
2006        PCRE_ERROR_PARTIAL is returned.
2008    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2009        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2010        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2011        and may be more useful for multi-segment matching.
2013    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2014        used to be given if ever the end of the subject was reached; now it is
2015        given only if matching could not proceed because another character was
2016        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2017        string "Z", which now yields "no match" instead of "partial match". In the
2018        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2019        final character ended with (*FAIL).
2021    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2022        if the pattern had a "must contain" character that was already found in the
2023        earlier partial match, unless partial matching was again requested. For
2024        example, with the pattern /dog.(body)?/, the "must contain" character is
2025        "g". If the first part-match was for the string "dog", restarting with
2026        "sbody" failed. This bug has been fixed.
2028    13. The string returned by pcre_dfa_exec() after a partial match has been
2029        changed so that it starts at the first inspected character rather than the
2030        first character of the match. This makes a difference only if the pattern
2031        starts with a lookbehind assertion or \b or \B (\K is not supported by
2032        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2033        matching functions compatible, and I think it's the right thing to do.
2035    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2036        so that the demonstration program is easily available in environments where
2037        PCRE has not been installed from source.
2039    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2040        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2041        library.
2043    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2044        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2045        is not the first non-POSIX option to be added. Clearly some people find
2046        these options useful.
2048    17. If a caller to the POSIX matching function regexec() passes a non-zero
2049        value for nmatch with a NULL value for pmatch, the value of
2050        nmatch is forced to zero.
2052    18. RunGrepTest did not have a test for the availability of the -u option of
2053        the diff command, as RunTest does. It now checks in the same way as
2054        RunTest, and also checks for the -b option.
2056    19. If an odd number of negated classes containing just a single character
2057        interposed, within parentheses, between a forward reference to a named
2058        subpattern and the definition of the subpattern, compilation crashed with
2059        an internal error, complaining that it could not find the referenced
2060        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2061        [The bug was that it was starting one character too far in when skipping
2062        over the character class, thus treating the ] as data rather than
2063        terminating the class. This meant it could skip too much.]
2065    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2066        /g option in pcretest when the pattern contains \K, which makes it possible
2067        to have an empty string match not at the start, even when the pattern is
2068        anchored. Updated pcretest and pcredemo to use this option.
2070    21. If the maximum number of capturing subpatterns in a recursion was greater
2071        than the maximum at the outer level, the higher number was returned, but
2072        with unset values at the outer level. The correct (outer level) value is
2073        now given.
2075    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2076        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2077        make it do so. The string so far is captured, making this feature
2078        compatible with Perl.
2080    23. The tests have been re-organized, adding tests 11 and 12, to make it
2081        possible to check the Perl 5.10 features against Perl 5.10.
2083    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2084        pattern matches a fixed length string. PCRE did not allow this; now it
2085        does. Neither allows recursion.
2087    25. I finally figured out how to implement a request to provide the minimum
2088        length of subject string that was needed in order to match a given pattern.
2089        (It was back references and recursion that I had previously got hung up
2090        on.) This code has now been added to pcre_study(); it finds a lower bound
2091        to the length of subject needed. It is not necessarily the greatest lower
2092        bound, but using it to avoid searching strings that are too short does give
2093        some useful speed-ups. The value is available to calling programs via
2094        pcre_fullinfo().
2096    26. While implementing 25, I discovered to my embarrassment that pcretest had
2097        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2098        study optimizations had never been tested with that matching function.
2099        Oops. What is worse, even when it was passed study data, there was a bug in
2100        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2101        were also very few tests of studied patterns with pcre_dfa_exec().
2103    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2104        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2105        on the other side of the coin, they are no longer allowed to have different
2106        names, because these cannot be distinguished in PCRE, and this has caused
2107        confusion. (This is a difference from Perl.)
2109    28. When duplicate subpattern names are present (necessarily with different
2110        numbers, as required by 27 above), and a test is made by name in a
2111        conditional pattern, either for a subpattern having been matched, or for
2112        recursion in such a pattern, all the associated numbered subpatterns are
2113        tested, and the overall condition is true if the condition is true for any
2114        one of them. This is the way Perl works, and is also more like the way
2115        testing by number works.
2118    Version 7.9 11-Apr-09
2119    ---------------------
2121    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2122        (pcretest), all targets were linked against these libraries. This included
2123        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2124        libraries. This caused unwanted dependencies to be created. This problem
2125        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2126        pcretest is linked with readline.
2128    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2129        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2130        moved outside it again, because FALSE and TRUE are already defined in AIX,
2131        but BOOL is not.
2133    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2134        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2136    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2137        hyphens) following file names and line numbers when outputting matching
2138        lines. This is not true; no spaces are inserted. I have also clarified the
2139        wording for the --colour (or --color) option.
2141    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2142        was not coloured; this is different to GNU grep, so I have changed it to be
2143        the same.
2145    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2146        each matching line was coloured. Now it goes on to look for further matches
2147        of any of the test patterns, which is the same behaviour as GNU grep.
2149    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2150        doesn't make sense to accept an empty string match in pcregrep, so I have
2151        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2152        seems to be how GNU grep behaves. [But see later change 40 for release
2153        8.33.]
2155    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2156        start or after a newline", because the conditional assertion was not being
2157        correctly handled. The rule now is that both the assertion and what follows
2158        in the first alternative must satisfy the test.
2160    9.  If auto-callout was enabled in a pattern with a conditional group whose
2161        condition was an assertion, PCRE could crash during matching, both with
2162        pcre_exec() and pcre_dfa_exec().
2164    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2165        used for matching.
2167    11. Unicode property support in character classes was not working for
2168        characters (bytes) greater than 127 when not in UTF-8 mode.
2170    12. Added the -M command line option to pcretest.
2172    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2174    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2176    16. Added comments and documentation about mis-use of no_arg in the C++
2177        wrapper.
2179    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2180        from Martin Jerabek that uses macro names for all relevant character and
2181        string constants.
2183    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2184        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2185        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2186        these, but not everybody uses configure.
2188    19. A conditional group that had only one branch was not being correctly
2189        recognized as an item that could match an empty string. This meant that an
2190        enclosing group might also not be so recognized, causing infinite looping
2191        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2192        with the subject "ab", where knowledge that the repeated group can match
2193        nothing is needed in order to break the loop.
2195    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2196        exec(), but without supplying a callout function, matching went wrong.
2198    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2199        leak if the size of the offset vector was greater than 30. When the vector
2200        is smaller, the saved offsets during recursion go onto a local stack
2201        vector, but for larger vectors malloc() is used. It was failing to free
2202        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2203        error, in fact).
2205    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2206        heapframe that is used only when UTF-8 support is enabled. This caused no
2207        problem, but was untidy.
2209    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2210        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2211        included within another project.
2213    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2214        slightly modified by me:
2216          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2217              not building pcregrep.
2219          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2220              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2222    25. Forward references, both numeric and by name, in patterns that made use of
2223        duplicate group numbers, could behave incorrectly or give incorrect errors,
2224        because when scanning forward to find the reference group, PCRE was not
2225        taking into account the duplicate group numbers. A pattern such as
2226        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2228    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2229        the feature of a custom pointer more persuasive (as requested by a user).
2231    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2232        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2233        pre-defined.
2235    28. Added support for (*UTF8) at the start of a pattern.
2237    29. Arrange for flags added by the "release type" setting in CMake to be shown
2238        in the configuration summary.
2241    Version 7.8 05-Sep-08
2242    ---------------------
2244    1.  Replaced UCP searching code with optimized version as implemented for Ad
2245        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2246        stage table and inline lookup instead of a function, giving speed ups of 2
2247        to 5 times on some simple patterns that I tested. Permission was given to
2248        distribute the MultiStage2.py script that generates the tables (it's not in
2249        the tarball, but is in the Subversion repository).
2251    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2252        scripts.
2254    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2255        a group with a zero qualifier. The result of the study could be incorrect,
2256        or the function might crash, depending on the pattern.
2258    4.  Caseless matching was not working for non-ASCII characters in back
2259        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2260        It now works when Unicode Property Support is available.
2262    5.  In pcretest, an escape such as \x{de} in the data was always generating
2263        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2264        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2265        truncation.
2267    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2269    7.  Added two (int) casts to pcregrep when printing the difference of two
2270        pointers, in case they are 64-bit values.
2272    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2273        test 2 if it fails.
2275    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2276        and a #define of that name to empty if it is not externally set. This is to
2277        allow users of MSVC to set it if necessary.
2279    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2280        the convenience functions in the pcre_get.c source file.
2282    11. An option change at the start of a pattern that had top-level alternatives
2283        could cause overwriting and/or a crash. This command provoked a crash in
2284        some environments:
2286          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2288        This potential security problem was recorded as CVE-2008-2371.
2290    12. For a pattern where the match had to start at the beginning or immediately
2291        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2292        pcre_dfa_exec() could read past the end of the passed subject if there was
2293        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2294        pcretest so that it places the subject at the end of its malloc-ed buffer.
2296    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2297        exec() might read past the end of the data buffer in UTF-8 mode.
2299    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2300        the data contained the byte 0x85 as part of a UTF-8 character within its
2301        first line. This applied both to normal and DFA matching.
2303    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2304        /^[^d]*?$/8 failed to match "abc".
2306    16. Added a missing copyright notice to pcrecpp_internal.h.
2308    17. Make it more clear in the documentation that values returned from
2309        pcre_exec() in ovector are byte offsets, not character counts.
2311    18. Tidied a few places to stop certain compilers from issuing warnings.
2313    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2314        supplied by Stefan Weber. I made a further small update for 7.8 because
2315        there is a change of source arrangements: the pcre_searchfuncs.c module is
2316        replaced by pcre_ucd.c.
2319    Version 7.7 07-May-08
2320    ---------------------
2322    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2323        a string to a long long, pretend we don't even have a long long." This is
2324        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2326    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2327        pre-7.6 versions, which defined a global no_arg variable instead of putting
2328        it in the RE class. (See also #8 below.)
2330    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2331        Lopes.
2333    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2335        (1) The include/exclude patterns were being applied to the whole pathnames
2336            of files, instead of just to the final components.
2338        (2) If there was more than one level of directory, the subdirectories were
2339            skipped unless they satisfied the include/exclude conditions. This is
2340            inconsistent with GNU grep (and could even be seen as contrary to the
2341            pcregrep specification - which I improved to make it absolutely clear).
2342            The action now is always to scan all levels of directory, and just
2343            apply the include/exclude patterns to regular files.
2345    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2346        --exclude_dir in the tests to avoid scanning .svn directories.
2348    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2349        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2350        doesn't support NULs in patterns.
2352    7.  Added some missing "const"s to declarations of static tables in
2353        pcre_compile.c and pcre_dfa_exec.c.
2355    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2356        caused by fix #2  above. (Subsequently also a second patch to fix the
2357        first patch. And a third patch - this was a messy problem.)
2359    9.  Applied Craig's patch to remove the use of push_back().
2361    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2362        matching function regexec().
2364    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2365        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2366        references. PCRE supports relative numbers with this syntax (I don't think
2367        Oniguruma does).
2369    12. Previously, a group with a zero repeat such as (...){0} was completely
2370        omitted from the compiled regex. However, this means that if the group
2371        was called as a subroutine from elsewhere in the pattern, things went wrong
2372        (an internal error was given). Such groups are now left in the compiled
2373        pattern, with a new opcode that causes them to be skipped at execution
2374        time.
2376    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2377        to the way PCRE behaves:
2379        (a) A lone ] character is dis-allowed (Perl treats it as data).
2381        (b) A back reference to an unmatched subpattern matches an empty string
2382            (Perl fails the current match path).
2384        (c) A data ] in a character class must be notated as \] because if the
2385            first data character in a class is ], it defines an empty class. (In
2386            Perl it is not possible to have an empty class.) The empty class []
2387            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2388            The negative empty class [^] matches any one character, independently
2389            of the DOTALL setting.
2391    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2392        non-existent subpattern following a character class starting with ']' and
2393        containing () gave an internal compiling error instead of "reference to
2394        non-existent subpattern". Fortunately, when the pattern did exist, the
2395        compiled code was correct. (When scanning forwards to check for the
2396        existence of the subpattern, it was treating the data ']' as terminating
2397        the class, so got the count wrong. When actually compiling, the reference
2398        was subsequently set up correctly.)
2400    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2401        it was being rejected as not supported by pcre_dfa_exec(), even though
2402        other assertions are supported. I have made pcre_dfa_exec() support
2403        (*FAIL).
2405    16. The implementation of 13c above involved the invention of a new opcode,
2406        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2407        cannot be changed at match time, I realized I could make a small
2408        improvement to matching performance by compiling OP_ALLANY instead of
2409        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2410        on the OP_ANY path.
2412    17. Compiling pcretest on Windows with readline support failed without the
2413        following two fixes: (1) Make the unistd.h include conditional on
2414        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2416    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2417        ncurses library to be included for pcretest when ReadLine support is
2418        requested, but also to allow for it to be overridden. This patch came from
2419        Daniel Bergström.
2421    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2422        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2423        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2424        spotting this.
2427    Version 7.6 28-Jan-08
2428    ---------------------
2430    1.  A character class containing a very large number of characters with
2431        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2432        overflow.
2434    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2435        HAVE_LONG_LONG is not defined.
2437    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2438        bring them up to date and include new features. This patch includes:
2440        - Fixed PH's badly added libz and libbz2 support.
2441        - Fixed a problem with static linking.
2442        - Added pcredemo. [But later removed - see 7 below.]
2443        - Fixed dftables problem and added an option.
2444        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2445            HAVE_LONG_LONG.
2446        - Added readline support for pcretest.
2447        - Added an listing of the option settings after cmake has run.
2449    4.  A user submitted a patch to Makefile that makes it easy to create
2450        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2451        Makefile.am that cause it to include this special target, without
2452        affecting anything else. Note that the same mingw target plus all
2453        the other distribution libraries and programs are now supported
2454        when configuring with CMake (see 6 below) instead of with
2455        Configure/Make.
2457    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2458        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2459        exported in the Windows port". It has not yet been confirmed that the patch
2460        solves the problem, but it does no harm.
2462    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2463        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2464        with CMake, and also correct the comment about stack recursion.
2466    7.  Remove the automatic building of pcredemo from the ./configure system and
2467        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2468        of a program that users should build themselves after PCRE is installed, so
2469        building it automatically is not really right. What is more, it gave
2470        trouble in some build environments.
2472    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2475    Version 7.5 10-Jan-08
2476    ---------------------
2478    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2479        values in parens when parsing an RE using the C++ wrapper."
2481    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2482        Characters greater than 255 were excluded from the class instead of being
2483        included.
2485    3.  The same bug as (2) above applied to negated POSIX classes such as
2486        [:^space:].
2488    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2489        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2490        I have changed it.
2492    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2493        first named subpattern) and a construct such as (?&a) would reference the
2494        first named subpattern whose name started with "a" (in other words, the
2495        length check was missing). Both these problems are fixed. "Subpattern name
2496        expected" is now given for (?&) (a zero-length name), and this patch also
2497        makes it give the same error for \k'' (previously it complained that that
2498        was a reference to a non-existent subpattern).
2500    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2501        this is right because (?- can be followed by option settings as well as by
2502        digits. I have, however, made the messages clearer.
2504    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2505        than the number used in the conditional) now cause a compile-time error.
2506        This is actually not compatible with Perl, which accepts such patterns, but
2507        treats the conditional as always being FALSE (as PCRE used to), but it
2508        seems to me that giving a diagnostic is better.
2510    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2511        and messages.
2513    9.  Fix two occurrences of "backslash" in comments that should have been
2514        "backspace".
2516    10. Remove two redundant lines of code that can never be obeyed (their function
2517        was moved elsewhere).
2519    11. The program that makes PCRE's Unicode character property table had a bug
2520        which caused it to generate incorrect table entries for sequences of
2521        characters that have the same character type, but are in different scripts.
2522        It amalgamated them into a single range, with the script of the first of
2523        them. In other words, some characters were in the wrong script. There were
2524        thirteen such cases, affecting characters in the following ranges:
2526          U+002b0 - U+002c1
2527          U+0060c - U+0060d
2528          U+0061e - U+00612
2529          U+0064b - U+0065e
2530          U+0074d - U+0076d
2531          U+01800 - U+01805
2532          U+01d00 - U+01d77
2533          U+01d9b - U+01dbf
2534          U+0200b - U+0200f
2535          U+030fc - U+030fe
2536          U+03260 - U+0327f
2537          U+0fb46 - U+0fbb1
2538          U+10450 - U+1049d
2540    12. The -o option (show only the matching part of a line) for pcregrep was not
2541        compatible with GNU grep in that, if there was more than one match in a
2542        line, it showed only the first of them. It now behaves in the same way as
2543        GNU grep.
2545    13. If the -o and -v options were combined for pcregrep, it printed a blank
2546        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2547        does the same. The return code can be used to tell if there were any
2548        non-matching lines.
2550    14. Added --file-offsets and --line-offsets to pcregrep.
2552    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2553        infinitely looping recursion. The bug was that positive lookaheads were not
2554        being skipped when checking for a possible empty match (negative lookaheads
2555        and both kinds of lookbehind were skipped).
2557    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2558        inclusion of <windows.h> to before rather than after the definition of
2559        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2561    17. Specifying a possessive quantifier with a specific limit for a Unicode
2562        character property caused pcre_compile() to compile bad code, which led at
2563        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2564        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2565        caused the error; without that there was no problem.
2567    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2569    19. Added --enable-pcretest-libreadline.
2571    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2572        RE::GlobalReplace(). As a result, the number of replacements returned was
2573        double what it should be. I removed one of the increments, but Craig sent a
2574        later patch that removed the other one (the right fix) and added unit tests
2575        that check the return values (which was not done before).
2577    21. Several CMake things:
2579        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2580            the names libpcre and libpcreposix, not just pcre and pcreposix.
2582        (2) The above change means that pcretest and pcregrep are now correctly
2583            linked with the newly-built libraries, not previously installed ones.
2587    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2588        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2589        UTF-8 newline character). The key issue is that the pattern starts .*;
2590        this means that the match must be either at the beginning, or after a
2591        newline. The bug was in the code for advancing after a failed match and
2592        checking that the new position followed a newline. It was not taking
2593        account of UTF-8 characters correctly.
2595    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2596        character classes. PCRE was not treating the sequence [:...:] as a
2597        character class unless the ... were all letters. Perl, however, seems to
2598        allow any characters between [: and :], though of course it rejects as
2599        unknown any "names" that contain non-letters, because all the known class
2600        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2601        for example, whereas PCRE did not - it did not recognize a POSIX character
2602        class. This seemed a bit dangerous, so the code has been changed to be
2603        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2604        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2605        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2606        Perl does, and where it didn't before.
2608    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2609        Windows environments %n is disabled by default.
2612    Version 7.4 21-Sep-07
2613    ---------------------
2615    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2616        means that a class such as [\s] counted as "explicit reference to CR or
2617        LF". That isn't really right - the whole point of the change was to try to
2618        help when there was an actual mention of one of the two characters. So now
2619        the change happens only if \r or \n (or a literal CR or LF) character is
2620        encountered.
2622    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2623        of both had grown to the point where there were only 3 bits left.
2624        Fortunately, there was spare space in the data structure, and so I have
2625        moved the internal flags into a new 16-bit field to free up more option
2626        bits.
2628    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2629        but did not set the internal JCHANGED flag - either of these is enough to
2630        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2631        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2632        start sets both bits.
2634    4.  Added options (at build time, compile time, exec time) to change \R from
2635        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2637    5.  doc/pcresyntax.html was missing from the distribution.
2639    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2640        compatibility, even though it is no longer used.
2642    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2643        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2644        windows.h file is present (where different names are used). [This was
2645        reversed later after testing - see 16 below.]
2647    8.  Changed all #include <config.h> to #include "config.h". There were also
2648        some further <pcre.h> cases that I changed to "pcre.h".
2650    9.  When pcregrep was used with the --colour option, it missed the line ending
2651        sequence off the lines that it output.
2653    10. It was pointed out to me that arrays of string pointers cause lots of
2654        relocations when a shared library is dynamically loaded. A technique of
2655        using a single long string with a table of offsets can drastically reduce
2656        these. I have refactored PCRE in four places to do this. The result is
2657        dramatic:
2659          Originally:                          290
2660          After changing UCP table:            187
2661          After changing error message table:   43
2662          After changing table of "verbs"       36
2663          After changing table of Posix names   22
2665        Thanks to the folks working on Gregex for glib for this insight.
2667    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2668        unicode-properties was also set.
2670    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2672    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2673        checked only for CRLF.
2675    14. Added casts to pcretest.c to avoid compiler warnings.
2677    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2679    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2680        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2681        entirely. This removes changes made in 7 above.
2683    17. The CMake files have been updated, and there is now more information about
2684        building with CMake in the NON-UNIX-USE document.
2687    Version 7.3 28-Aug-07
2688    ---------------------
2690     1. In the rejigging of the build system that eventually resulted in 7.1, the
2691        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2692        brackets there is not right, since it causes compilers to look for an
2693        installed pcre.h, not the version that is in the source that is being
2694        compiled (which of course may be different). I have changed it back to:
2696          #include "pcre.h"
2698        I have a vague recollection that the change was concerned with compiling in
2699        different directories, but in the new build system, that is taken care of
2700        by the VPATH setting the Makefile.
2702     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2703        when the subject happened to end in the byte 0x85 (e.g. if the last
2704        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2705        characters but of course it shouldn't be taken as a newline when it is part
2706        of another character. The bug was that, for an unlimited repeat of . in
2707        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2708        characters when looking for a newline.
2710     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2712     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2713        in debug output.
2715     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2716        long printing in the pcrecpp unittest when running under MinGW.
2718     6. ESC_K was left out of the EBCDIC table.
2720     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2721        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2722        limit also applies to "virtual nesting" when a pattern is recursive, and in
2723        this case 1000 isn't so big. I have been able to remove this limit at the
2724        expense of backing off one optimization in certain circumstances. Normally,
2725        when pcre_exec() would call its internal match() function recursively and
2726        immediately return the result unconditionally, it uses a "tail recursion"
2727        feature to save stack. However, when a subpattern that can match an empty
2728        string has an unlimited repetition quantifier, it no longer makes this
2729        optimization. That gives it a stack frame in which to save the data for
2730        checking that an empty string has been matched. Previously this was taken
2731        from the 1000-entry workspace that had been reserved. So now there is no
2732        explicit limit, but more stack is used.
2734     8. Applied Daniel's patches to solve problems with the import/export magic
2735        syntax that is required for Windows, and which was going wrong for the
2736        pcreposix and pcrecpp parts of the library. These were overlooked when this
2737        problem was solved for the main library.
2739     9. There were some crude static tests to avoid integer overflow when computing
2740        the size of patterns that contain repeated groups with explicit upper
2741        limits. As the maximum quantifier is 65535, the maximum group length was
2742        set at 30,000 so that the product of these two numbers did not overflow a
2743        32-bit integer. However, it turns out that people want to use groups that
2744        are longer than 30,000 bytes (though not repeat them that many times).
2745        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2746        made it possible to implement the integer overflow checks in a much more
2747        dynamic way, which I have now done. The artificial limitation on group
2748        length has been removed - we now have only the limit on the total length of
2749        the compiled pattern, which depends on the LINK_SIZE setting.
2751    10. Fixed a bug in the documentation for get/copy named substring when
2752        duplicate names are permitted. If none of the named substrings are set, the
2753        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2754        empty string.
2756    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2757        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2758        because the ] is interpreted as the first data character and the
2759        terminating ] is not found. PCRE has been made compatible with Perl in this
2760        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2761        cause memory overwriting.
2763    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2764        string has been matched (to stop an infinite loop). It was not recognizing
2765        a conditional subpattern that could match an empty string if that
2766        subpattern was within another subpattern. For example, it looped when
2767        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2768        condition was not nested. This bug has been fixed.
2770    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2771        past the start of the subject in the presence of bytes with the top bit
2772        set, for example "\x8aBCD".
2774    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2775        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2777    14. Optimized (?!) to (*FAIL).
2779    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2780        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2781        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2782        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2783        does: it's just the validity check that is more restrictive.
2785    16. Inserted checks for integer overflows during escape sequence (backslash)
2786        processing, and also fixed erroneous offset values for syntax errors during
2787        backslash processing.
2789    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2790        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2792    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2793        caused an overrun.
2795    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2796        something other than just ASCII characters) inside a group that had an
2797        unlimited repeat caused a loop at compile time (while checking to see
2798        whether the group could match an empty string).
2800    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2801        [\P{Any}] did so. (Error in the code for printing property names.)
2803    21. An orphan \E inside a character class could cause a crash.
2805    22. A repeated capturing bracket such as (A)? could cause a wild memory
2806        reference during compilation.
2808    23. There are several functions in pcre_compile() that scan along a compiled
2809        expression for various reasons (e.g. to see if it's fixed length for look
2810        behind). There were bugs in these functions when a repeated \p or \P was
2811        present in the pattern. These operators have additional parameters compared
2812        with \d, etc, and these were not being taken into account when moving along
2813        the compiled data. Specifically:
2815        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2816            length.
2818        (b) An item such as \pL+ within a repeated group could cause crashes or
2819            loops.
2821        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2822            "reference to non-existent subpattern" error.
2824        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2826    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2827        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2829    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2831    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2832        character were causing crashes (broken optimization).
2834    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2835        \p or \P) caused a compile-time loop.
2837    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2838        break. For example, the unstudied pattern [\r\n]A does not match the string
2839        "\r\nA" because change 7.0/46 below moves the current point on by two
2840        characters after failing to match at the start. However, the pattern \nA
2841        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2842        the same is true. There doesn't seem any very clean way out of this, but
2843        what I have chosen to do makes the common cases work: PCRE now takes note
2844        of whether there can be an explicit match for \r or \n anywhere in the
2845        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2846        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2847        pattern has explicit CR or LF references.
2849    29. Added (*CR) etc for changing newline setting at start of pattern.
2852    Version 7.2 19-Jun-07
2853    ---------------------
2855     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2856        which is apparently normally available under Windows.
2858     2. Re-jig the pcregrep tests with different newline settings in an attempt
2859        to make them independent of the local environment's newline setting.
2861     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2863     4. Some of the "internals" tests were previously cut out when the link size
2864        was not 2, because the output contained actual offsets. The recent new
2865        "Z" feature of pcretest means that these can be cut out, making the tests
2866        usable with all link sizes.
2868     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2869        stack recursion. This gives a massive performance boost under BSD, but just
2870        a small improvement under Linux. However, it saves one field in the frame
2871        in all cases.
2873     6. Added more features from the forthcoming Perl 5.10:
2875        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2876            recursion call. It refers to the nth most recently opened parentheses.
2878        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2879            to be opened parentheses.
2881        (c) Conditions that refer to capturing parentheses can be specified
2882            relatively, for example, (?(-2)... or (?(+3)...
2884        (d) \K resets the start of the current match so that everything before
2885            is not part of it.
2887        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2889        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2890            reference syntax.
2892        (g) (?| introduces a group in which the numbering of parentheses in each
2893            alternative starts with the same number.
2895        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2897     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2900     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2901        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2902        for detecting groups that can match an empty string.
2904     9. A pattern with a very large number of alternatives (more than several
2905        hundred) was running out of internal workspace during the pre-compile
2906        phase, where pcre_compile() figures out how much memory will be needed. A
2907        bit of new cunning has reduced the workspace needed for groups with
2908        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2909        workspace instead of running out of the 4096 that are available.
2911    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2913    11. Applied patch from Google to remove an optimization that didn't quite work.
2914        The report of the bug said:
2916          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2917          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2918          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2920    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2921        it matched the wrong number of bytes.
2924    Version 7.1 24-Apr-07
2925    ---------------------
2927     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2928        that is more "standard", making use of automake and other Autotools. There
2929        is some re-arrangement of the files and adjustment of comments consequent
2930        on this.
2932     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2933        for recursive directory scanning broke on some systems because the files
2934        are not scanned in any specific order and on different systems the order
2935        was different. A call to "sort" has been inserted into RunGrepTest for the
2936        approprate test as a short-term fix. In the longer term there may be an
2937        alternative.
2939     3. I had an email from Eric Raymond about problems translating some of PCRE's
2940        man pages to HTML (despite the fact that I distribute HTML pages, some
2941        people do their own conversions for various reasons). The problems
2942        concerned the use of low-level troff macros .br and .in. I have therefore
2943        removed all such uses from the man pages (some were redundant, some could
2944        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2945        HTML has been updated to handle .nf/.fi and to complain if it encounters
2946        .br or .in.
2948     4. Updated comments in configure.ac that get placed in config.h.in and also
2949        arranged for config.h to be included in the distribution, with the name
2950        config.h.generic, for the benefit of those who have to compile without
2951        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2953     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2954        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2955        makevp.bat for the current PCRE, using the additional files
2956        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2958     6. A Windows user reported a minor discrepancy with test 2, which turned out
2959        to be caused by a trailing space on an input line that had got lost in his
2960        copy. The trailing space was an accident, so I've just removed it.
2962     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2963        that is needed.
2965     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2966        as "const" (a) because they are and (b) because it helps the PHP
2967        maintainers who have recently made a script to detect big data structures
2968        in the php code that should be moved to the .rodata section. I remembered
2969        to update Builducptable as well, so it won't revert if ucptable.h is ever
2970        re-created.
2972     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2973        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2974        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2975        support is not required. This saves 1.5-2K of code, which is important in
2976        some applications.
2978        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2979        so as not to refer to the tables, even though these functions will never be
2980        called when UTF-8 support is disabled. Otherwise there are problems with a
2981        shared library.
2983    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2985        (a) It was defining its arguments as char * instead of void *.
2987        (b) It was assuming that all moves were upwards in memory; this was true
2988            a long time ago when I wrote it, but is no longer the case.
2990        The emulated memove() is provided for those environments that have neither
2991        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2992        is clearly not the case, as these two bugs were recently reported.
2994    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2995        and Detrail to create the HTML documentation, the .txt form of the man
2996        pages, and it removes trailing spaces from listed files. It also creates
2997        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2998        case, it wraps all the #defines with #ifndefs. This script should be run
2999        before "make dist".
3001    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3002        with Unicode property support.
3004        (a) For a maximizing quantifier, if the two different cases of the
3005            character were of different lengths in their UTF-8 codings (there are
3006            some cases like this - I found 11), and the matching function had to
3007            back up over a mixture of the two cases, it incorrectly assumed they
3008            were both the same length.
3010        (b) When PCRE was configured to use the heap rather than the stack for
3011            recursion during matching, it was not correctly preserving the data for
3012            the other case of a UTF-8 character when checking ahead for a match
3013            while processing a minimizing repeat. If the check also involved
3014            matching a wide character, but failed, corruption could cause an
3015            erroneous result when trying to check for a repeat of the original
3016            character.
3018    13. Some tidying changes to the testing mechanism:
3020        (a) The RunTest script now detects the internal link size and whether there
3021            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3022            values substituted by "configure". (The RunGrepTest script already did
3023            this for UTF-8.) The configure.ac script no longer substitutes the
3024            relevant variables.
3026        (b) The debugging options /B and /D in pcretest show the compiled bytecode
3027            with length and offset values. This means that the output is different
3028            for different internal link sizes. Test 2 is skipped for link sizes
3029            other than 2 because of this, bypassing the problem. Unfortunately,
3030            there was also a test in test 3 (the locale tests) that used /B and
3031            failed for link sizes other than 2. Rather than cut the whole test out,
3032            I have added a new /Z option to pcretest that replaces the length and
3033            offset values with spaces. This is now used to make test 3 independent
3034            of link size. (Test 2 will be tidied up later.)
3036    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3037        segmentation fault instead of returning the appropriate error message.
3039    15. In multiline mode when the newline sequence was set to "any", the pattern
3040        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3041        This doesn't seem right; it now treats the CRLF combination as the line
3042        ending, and so does not match in that case. It's only a pattern such as ^$
3043        that would hit this one: something like ^ABC$ would have failed after \r
3044        and then tried again after \r\n.
3046    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3047        in an attempt to make files that differ only in their line terminators
3048        compare equal. This works on Linux.
3050    17. Under certain error circumstances pcregrep might try to free random memory
3051        as it exited. This is now fixed, thanks to valgrind.
3053    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3054        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3055        was because its rules for how to advance for /g after matching an empty
3056        string at the end of a line did not allow for this case. They now check for
3057        it specially.
3059    20. pcretest is supposed to handle patterns and data of any length, by
3060        extending its buffers when necessary. It was getting this wrong when the
3061        buffer for a data line had to be extended.
3063    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3064        CRLF as a newline sequence.
3066    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3067        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3068        I have nevertheless tidied it up.
3070    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3072    24. Added a man page for pcre-config.
3075    Version 7.0 19-Dec-06
3076    ---------------------
3078     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3079        moving to gcc 4.1.1.
3081     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3082        sys/time.h, which is documented as needed for this function. It doesn't
3083        seem to matter on Linux, but it showed up on some releases of OS X.
3085     3. It seems that there are systems where bytes whose values are greater than
3086        127 match isprint() in the "C" locale. The "C" locale should be the
3087        default when a C program starts up. In most systems, only ASCII printing
3088        characters match isprint(). This difference caused the output from pcretest
3089        to vary, making some of the tests fail. I have changed pcretest so that:
3091        (a) When it is outputting text in the compiled version of a pattern, bytes
3092            other than 32-126 are always shown as hex escapes.
3094        (b) When it is outputting text that is a matched part of a subject string,
3095            it does the same, unless a different locale has been set for the match
3096            (using the /L modifier). In this case, it uses isprint() to decide.
3098     4. Fixed a major bug that caused incorrect computation of the amount of memory
3099        required for a compiled pattern when options that changed within the
3100        pattern affected the logic of the preliminary scan that determines the
3101        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3102        that the computed length was too small. The symptoms of this bug were
3103        either the PCRE error "internal error: code overflow" from pcre_compile(),
3104        or a glibc crash with a message such as "pcretest: free(): invalid next
3105        size (fast)". Examples of patterns that provoked this bug (shown in
3106        pcretest format) are:
3108          /(?-x: )/x
3109          /(?x)(?-x: \s*#\s*)/
3110          /((?i)[\x{c0}])/8
3111          /(?i:[\x{c0}])/8
3113        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3114        is now done differently.
3116     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3117        wrapper classes; (b) implement a new function in the C++ scanner that is
3118        more efficient than the old way of doing things because it avoids levels of
3119        recursion in the regex matching; (c) add a paragraph to the documentation
3120        for the FullMatch() function.
3122     6. The escape sequence \n was being treated as whatever was defined as
3123        "newline". Not only was this contrary to the documentation, which states
3124        that \n is character 10 (hex 0A), but it also went horribly wrong when
3125        "newline" was defined as CRLF. This has been fixed.
3127     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3128        was being set to -1 for the "end of line" case (supposedly a value that no
3129        character can have). Though this value is never used (the check for end of
3130        line is "zero bytes in current character"), it caused compiler complaints.
3131        I've changed it to 0xffffffff.
3133     8. In pcre_version.c, the version string was being built by a sequence of
3134        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3135        string (as it is for production releases) called a macro with an empty
3136        argument. The C standard says the result of this is undefined. The gcc
3137        compiler treats it as an empty string (which was what was wanted) but it is
3138        reported that Visual C gives an error. The source has been hacked around to
3139        avoid this problem.
3141     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3142        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3143        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3144        of them did).
3146    10. Originally, pcretest opened its input and output without "b"; then I was
3147        told that "b" was needed in some environments, so it was added for release
3148        5.0 to both the input and output. (It makes no difference on Unix-like
3149        systems.) Later I was told that it is wrong for the input on Windows. I've
3150        now abstracted the modes into two macros, to make it easier to fiddle with
3151        them, and removed "b" from the input mode under Windows.
3153    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3155    12. Added -help and --help to pcretest as an official way of being reminded
3156        of the options.
3158    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3159        and pcrecpp.cc because they annoy compilers at high warning levels.
3161    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3163    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3164        scripts are not C programs :-) and which was not noticed because it works
3165        on Linux.
3167    16. pcretest is supposed to handle any length of pattern and data line (as one
3168        line or as a continued sequence of lines) by extending its input buffer if
3169        necessary. This feature was broken for very long pattern lines, leading to
3170        a string of junk being passed to pcre_compile() if the pattern was longer
3171        than about 50K.
3173    17. I have done a major re-factoring of the way pcre_compile() computes the
3174        amount of memory needed for a compiled pattern. Previously, there was code
3175        that made a preliminary scan of the pattern in order to do this. That was
3176        OK when PCRE was new, but as the facilities have expanded, it has become
3177        harder and harder to keep it in step with the real compile phase, and there
3178        have been a number of bugs (see for example, 4 above). I have now found a
3179        cunning way of running the real compile function in a "fake" mode that
3180        enables it to compute how much memory it would need, while actually only
3181        ever using a few hundred bytes of working memory and without too many
3182        tests of the mode. This should make future maintenance and development
3183        easier. A side effect of this work is that the limit of 200 on the nesting
3184        depth of parentheses has been removed (though this was never a serious
3185        limitation, I suspect). However, there is a downside: pcre_compile() now
3186        runs more slowly than before (30% or more, depending on the pattern). I
3187        hope this isn't a big issue. There is no effect on runtime performance.
3189    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3190        newline (only possible for the last line of a file) and it was a
3191        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3193    19. Added additional timing features to pcretest. (1) The -tm option now times
3194        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3195        separate command line item, by a number that specifies the number of
3196        repeats to use when timing. The default is 50000; this gives better
3197        precision, but takes uncomfortably long for very large patterns.
3199    20. Extended pcre_study() to be more clever in cases where a branch of a
3200        subpattern has no definite first character. For example, (a*|b*)[cd] would
3201        previously give no result from pcre_study(). Now it recognizes that the
3202        first character must be a, b, c, or d.
3204    21. There was an incorrect error "recursive call could loop indefinitely" if
3205        a subpattern (or the entire pattern) that was being tested for matching an
3206        empty string contained only one non-empty item after a nested subpattern.
3207        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3208        incorrectly, because the \d was being skipped in the check.
3210    22. The pcretest program now has a new pattern option /B and a command line
3211        option -b, which is equivalent to adding /B to every pattern. This causes
3212        it to show the compiled bytecode, without the additional information that
3213        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3214        is the same as /B/I).
3216    23. A new optimization is now able automatically to treat some sequences such
3217        as a*b as a*+b. More specifically, if something simple (such as a character
3218        or a simple class like \d) has an unlimited quantifier, and is followed by
3219        something that cannot possibly match the quantified thing, the quantifier
3220        is automatically "possessified".
3222    24. A recursive reference to a subpattern whose number was greater than 39
3223        went wrong under certain circumstances in UTF-8 mode. This bug could also
3224        have affected the operation of pcre_study().
3226    25. Realized that a little bit of performance could be had by replacing
3227        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3229    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3231    27. Possessive quantifiers such as a++ were previously implemented by turning
3232        them into atomic groups such as ($>a+). Now they have their own opcodes,
3233        which improves performance. This includes the automatically created ones
3234        from 23 above.
3236    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3237        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3238        the first matched character to be a colon. This applied both to named and
3239        numbered groups.
3241    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3243    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3244        building PCRE on Windows easier, so I have included it in the distribution.
3246    31. There is now a check in pcretest against a ridiculously large number being
3247        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3248        loop, the loop is abandoned.
3250    32. Forward references to subpatterns in conditions such as (?(2)...) where
3251        subpattern 2 is defined later cause pcre_compile() to search forwards in
3252        the pattern for the relevant set of parentheses. This search went wrong
3253        when there were unescaped parentheses in a character class, parentheses
3254        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3256    33. "Subroutine" calls and backreferences were previously restricted to
3257        referencing subpatterns earlier in the regex. This restriction has now
3258        been removed.
3260    34. Added a number of extra features that are going to be in Perl 5.10. On the
3261        whole, these are just syntactic alternatives for features that PCRE had
3262        previously implemented using the Python syntax or my own invention. The
3263        other formats are all retained for compatibility.
3265        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3266            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3267            also .NET compatible.
3269        (b) A recursion or subroutine call to a named group can now be defined as
3270            (?&name) as well as (?P>name).
3272        (c) A backreference to a named group can now be defined as \k<name> or
3273            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3274            5.10, are also .NET compatible.
3276        (d) A conditional reference to a named group can now use the syntax
3277            (?(<name>) or (?('name') as well as (?(name).
3279        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3280            groups (named and numbered) that are never evaluated inline, but can be
3281            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3282            is always false. There may be only one alternative in such a group.
3284        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3285            as the simple (?(R). The condition is true only if the most recent
3286            recursion is that of the given number or name. It does not search out
3287            through the entire recursion stack.
3289        (g) The escape \gN or \g{N} has been added, where N is a positive or
3290            negative number, specifying an absolute or relative reference.
3292    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3293        some "unreachable code" warnings.
3295    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3296        things, this adds five new scripts.
3298    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3299        There were also incompatibilities regarding the handling of \Q..\E inside
3300        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3301        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3303    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3304        matches an empty string, and forcibly breaks the loop. There were bugs in
3305        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3306        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3307        separate and independent bugs (that affected different cases) have been
3308        fixed.
3310    39. Refactored the code to abolish the use of different opcodes for small
3311        capturing bracket numbers. This is a tidy that I avoided doing when I
3312        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3313        The new approach is not only tidier, it makes it possible to reduce the
3314        memory needed to fix the previous bug (38).
3316    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3317        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3318        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3319        mode.
3321    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3322        report.
3324    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3325        copy construction and assignment in the C++ wrapper.
3327    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3328        couple of bugs that could have given wrong results in the "--newline=crlf"
3329        case.
3331    44. Added a number of casts and did some reorganization of signed/unsigned int
3332        variables following suggestions from Dair Grant. Also renamed the variable
3333        "this" as "item" because it is a C++ keyword.
3335    45. Arranged for dftables to add
3337          #include "pcre_internal.h"
3339        to pcre_chartables.c because without it, gcc 4.x may remove the array
3340        definition from the final binary if PCRE is built into a static library and
3341        dead code stripping is activated.
3343    46. For an unanchored pattern, if a match attempt fails at the start of a
3344        newline sequence, and the newline setting is CRLF or ANY, and the next two
3345        characters are CRLF, advance by two characters instead of one.
3348    Version 6.7 04-Jul-06
3349    ---------------------
3351     1. In order to handle tests when input lines are enormously long, pcretest has
3352        been re-factored so that it automatically extends its buffers when
3353        necessary. The code is crude, but this _is_ just a test program. The
3354        default size has been increased from 32K to 50K.
3356     2. The code in pcre_study() was using the value of the re argument before
3357        testing it for NULL. (Of course, in any sensible call of the function, it
3358        won't be NULL.)
3360     3. The memmove() emulation function in pcre_internal.h, which is used on
3361        systems that lack both memmove() and bcopy() - that is, hardly ever -
3362        was missing a "static" storage class specifier.
3364     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3365        containing an extended class (one that cannot be represented by a bitmap
3366        because it contains high-valued characters or Unicode property items, e.g.
3367        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3368        pattern, but PCRE should not loop if you do not (it no longer does).
3369        [Detail: two cases were found: (a) a repeated subpattern containing an
3370        extended class; (b) a recursive reference to a subpattern that followed a
3371        previous extended class. It wasn't skipping over the extended class
3372        correctly when UTF-8 mode was not set.]
3374     5. A negated single-character class was not being recognized as fixed-length
3375        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3376        compile error "lookbehind assertion is not fixed length".
3378     6. The RunPerlTest auxiliary script was showing an unexpected difference
3379        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3380        write a Perl script that can interpret lines of an input file either as
3381        byte characters or as UTF-8, which is what "perltest" was being required to
3382        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3383        can't do is switch easily at run time between having the "use utf8;" pragma
3384        or not. In the end, I fudged it by using the RunPerlTest script to insert
3385        "use utf8;" explicitly for the UTF-8 tests.
3387     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3388        the end of the subject string, contrary to the documentation and to what
3389        Perl does. This was true of both matching functions. Now it matches only at
3390        the start of the subject and immediately after *internal* newlines.
3392     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3393        a pointer to an int instead of a pointer to an unsigned long int. This
3394        caused problems on 64-bit systems.
3396     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3397        instance of the 'standard' template library not being so standard".
3399    10. There was no check on the number of named subpatterns nor the maximum
3400        length of a subpattern name. The product of these values is used to compute
3401        the size of the memory block for a compiled pattern. By supplying a very
3402        long subpattern name and a large number of named subpatterns, the size
3403        computation could be caused to overflow. This is now prevented by limiting
3404        the length of names to 32 characters, and the number of named subpatterns
3405        to 10,000.
3407    11. Subpatterns that are repeated with specific counts have to be replicated in
3408        the compiled pattern. The size of memory for this was computed from the
3409        length of the subpattern and the repeat count. The latter is limited to
3410        65535, but there was no limit on the former, meaning that integer overflow
3411        could in principle occur. The compiled length of a repeated subpattern is
3412        now limited to 30,000 bytes in order to prevent this.
3414    12. Added the optional facility to have named substrings with the same name.
3416    13. Added the ability to use a named substring as a condition, using the
3417        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3418        are numbers (not recommended). Forward references are permitted.
3420    14. Added forward references in named backreferences (if you see what I mean).
3422    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3423        pattern could run off the end of the subject. For example, the pattern
3424        "(?s)(.{1,5})"8 did this with the subject "ab".
3426    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3427        PCRE_CASELESS was set when matching characters that were quantified with ?
3428        or *.
3430    17. A character class other than a single negated character that had a minimum
3431        but no maximum quantifier - for example [ab]{6,} - was not handled
3432        correctly by pce_dfa_exec(). It would match only one character.
3434    18. A valid (though odd) pattern that looked like a POSIX character
3435        class but used an invalid character after [ (for example [[,abc,]]) caused
3436        pcre_compile() to give the error "Failed: internal error: code overflow" or
3437        in some cases to crash with a glibc free() error. This could even happen if
3438        the pattern terminated after [[ but there just happened to be a sequence of
3439        letters, a binary zero, and a closing ] in the memory that followed.
3441    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3442        over the years. Originally (before any Unicode support), just the bottom 8
3443        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3444        output from "man perlunicode" includes this:
3446          The regular expression compiler produces polymorphic opcodes.  That
3447          is, the pattern adapts to the data and automatically switches to
3448          the Unicode character scheme when presented with Unicode data--or
3449          instead uses a traditional byte scheme when presented with byte
3450          data.
3452        Sadly, a wide octal escape does not cause a switch, and in a string with
3453        no other multibyte characters, these octal escapes are treated as before.
3454        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3455        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3456        Unicode string.
3458        I have not perpetrated such confusion in PCRE. Up till now, it took just
3459        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3460        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3461        translate to the appropriate multibyte character.
3463    29. Applied some refactoring to reduce the number of warnings from Microsoft
3464        and Borland compilers. This has included removing the fudge introduced
3465        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3466        a warning about an unused variable.
3468    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3469        characters since release 4.0, because Perl (from release 5.004) does not.
3470        [Or at least, is documented not to: some releases seem to be in conflict
3471        with the documentation.] However, when a pattern was studied with
3472        pcre_study() and all its branches started with \s, PCRE still included VT
3473        as a possible starting character. Of course, this did no harm; it just
3474        caused an unnecessary match attempt.
3476    22. Removed a now-redundant internal flag bit that recorded the fact that case
3477        dependency changed within the pattern. This was once needed for "required
3478        byte" processing, but is no longer used. This recovers a now-scarce options
3479        bit. Also moved the least significant internal flag bit to the most-
3480        significant bit of the word, which was not previously used (hangover from
3481        the days when it was an int rather than a uint) to free up another bit for
3482        the future.
3484    23. Added support for CRLF line endings as well as CR and LF. As well as the
3485        default being selectable at build time, it can now be changed at runtime
3486        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3487        specify that it is scanning data with non-default line endings.
3489    24. Changed the definition of CXXLINK to make it agree with the definition of
3490        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3492    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3493        recursions. This makes a big different to stack usage for some patterns.
3495    26. If a subpattern containing a named recursion or subroutine reference such
3496        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3497        the space required for the compiled pattern went wrong and gave too small a
3498        value. Depending on the environment, this could lead to "Failed: internal
3499        error: code overflow at offset 49" or "glibc detected double free or
3500        corruption" errors.
3502    27. Applied patches from Google (a) to support the new newline modes and (b) to
3503        advance over multibyte UTF-8 characters in GlobalReplace.
3505    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3506        difference for some implementation of PCRE in some Windows version.
3508    29. Added some extra testing facilities to pcretest:
3510        \q<number>   in a data line sets the "match limit" value
3511        \Q<number>   in a data line sets the "match recursion limt" value
3512        -S <number>  sets the stack size, where <number> is in megabytes
3514        The -S option isn't available for Windows.
3517    Version 6.6 06-Feb-06
3518    ---------------------
3520     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3521        in pcreposix.h. I have copied the definition from pcre.h.
3523     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3524        because pcre.h is no longer a built file.
3526     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3527        not normally included in the compiled code.
3530    Version 6.5 01-Feb-06
3531    ---------------------
3533     1. When using the partial match feature with pcre_dfa_exec(), it was not
3534        anchoring the second and subsequent partial matches at the new starting
3535        point. This could lead to incorrect results. For example, with the pattern
3536        /1234/, partially matching against "123" and then "a4" gave a match.
3538     2. Changes to pcregrep:
3540        (a) All non-match returns from pcre_exec() were being treated as failures
3541            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3542            error message is output. Some extra information is given for the
3543            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3544            probably the only errors that are likely to be caused by users (by
3545            specifying a regex that has nested indefinite repeats, for instance).
3546            If there are more than 20 of these errors, pcregrep is abandoned.
3548        (b) A binary zero was treated as data while matching, but terminated the
3549            output line if it was written out. This has been fixed: binary zeroes
3550            are now no different to any other data bytes.
3552        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3553            used to set a locale for matching. The --locale=xxxx long option has
3554            been added (no short equivalent) to specify a locale explicitly on the
3555            pcregrep command, overriding the environment variables.
3557        (d) When -B was used with -n, some line numbers in the output were one less
3558            than they should have been.
3560        (e) Added the -o (--only-matching) option.
3562        (f) If -A or -C was used with -c (count only), some lines of context were
3563            accidentally printed for the final match.
3565        (g) Added the -H (--with-filename) option.
3567        (h) The combination of options -rh failed to suppress file names for files
3568            that were found from directory arguments.
3570        (i) Added the -D (--devices) and -d (--directories) options.
3572        (j) Added the -F (--fixed-strings) option.
3574        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3576        (l) Added the --colo(u)r option.
3578        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3579            is not present by default.
3581     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3582        items such as (?R) or (?1), when the recursion could match a number of