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

Diff of /code/trunk/ChangeLog

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

revision 61 by nigel, Sat Feb 24 21:39:58 2007 UTC revision 1452 by zherczeg, Tue Jan 28 16:07:52 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 3.0 02-Jan-02  Version 8.35-RC1 xx-xxxx-201x
5    -----------------------------
6    
7    1.  A new flag is set, when property checks are present in an XCLASS.
8        When this flag is not set, PCRE can perform certain optimizations
9        such as studying these XCLASS-es.
10    
11    2.  The auto-possessification of character sets were improved: a normal
12        and an extended character set can be compared now. Furthermore
13        the JIT compiler optimizes more character set checks.
14    
15    3.  Got rid of some compiler warnings for potentially uninitialized variables
16        that show up only when compiled with -O2.
17    
18    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
19        of a match later then the end of the match. The pcretest program was not
20        handling the case sensibly - it was outputting from the start to the next
21        binary zero. It now reports this situation in a message, and outputs the
22        text from the end to the start.
23    
24    5.  Fast forward search is improved in JIT. Instead of the first three
25        characters, any three characters with fixed position can be searched.
26        Search order: first, last, middle.
27    
28    6.  Improve character range checks in JIT. Characters are read by an inprecise
29        function now, which returns with an unknown value if the character code is
30        above a certain treshold (e.g: 256). The only limitation is that the value
31        must be bigger than the treshold as well. This function is useful, when
32        the characters above the treshold are handled in the same way.
33    
34    7.  The macros whose names start with RAWUCHAR are placeholders for a future
35        mode in which only the bottom 21 bits of 32-bit data items are used. To
36        make this more memorable for those maintaining the code, the names have
37        been changed to start with UCHAR21, and an extensive comment has been added
38        to their definition.
39    
40    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
41        to the export list in Makefile.am (they were accidentally omitted from the
42        8.34 tarball).
43    
44    9.  The informational output from pcretest used the phrase "starting byte set"
45        which is inappropriate for the 16-bit and 32-bit libraries. As the output
46        for "first char" and "need char" really means "non-UTF-char", I've changed
47        "byte" to "char", and slightly reworded the output. The documentation about
48        these values has also been (I hope) clarified.
49    
50    10. Another JIT related optimization: use table jumps for selecting the correct
51        backtracking path, when more than four alternatives are present inside a
52        bracket.
53    
54    11. Empty match is not possible, when the minimum length is greater than zero,
55        and there is no \K in the pattern. JIT should avoid empty match checks in
56        such cases.
57    
58    12. In a caseless character class with UCP support, when a character with more
59        than one alternative case was not the first character of a range, not all
60        the alternative cases were added to the class. For example, s and \x{17f}
61        are both alternative cases for S: the class [RST] was handled correctly,
62        but [R-T] was not.
63    
64    13. The configure.ac file always checked for pthread support when JIT was
65        enabled. This is not used in Windows, so I have put this test inside a
66        check for the presence of windows.h (which was already tested for).
67    
68    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
69        The algorithm provides a way to skip certain starting offsets, and usually
70        faster than linear prefix searches.
71    
72    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
73        as for 'fr_FR' and 'french'. For some reason, however, it then used the
74        Windows-specific input and output files, which have 'french' screwed in.
75        So this could never have worked. One of the problems with locales is that
76        they aren't always the same. I have now updated RunTest so that it checks
77        the output of the locale test (test 3) against three different output
78        files, and it allows the test to pass if any one of them matches. With luck
79        this should make the test pass on some versions of Solaris where it was
80        failing. Because of the uncertainty, the script did not used to stop if
81        test 3 failed; it now does. If further versions of a French locale ever
82        come to light, they can now easily be added.
83    
84    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
85        there was a message during ./configure, but it did not stop. This now
86        provokes an error. The invalid example in README has been corrected.
87        If a value less than the minimum is given, the minimum value has always
88        been used, but now a warning is given.
89    
90    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
91        was a bug in the test system, which is now fixed. Also, the list of various
92        configurations that are tested for each release did not have one with both
93        16/32 bits and --enable-bar-anycrlf. It now does.
94    
95    18. pcretest was missing "-C bsr" for displaying the \R default setting.
96    
97    19. Little endian PowerPC systems are supported now by the JIT compiler.
98    
99    20. The fast forward newline mechanism could enter to an infinite loop on
100        certain invalid UTF-8 input. Although we don't support these cases
101        this issue can be fixed by a performance optimization.
102    
103    
104    Version 8.34 15-December-2013
105    -----------------------------
106    
107    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
108        executable memory. Patch inspired by Carsten Klein.
109    
110    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
111        this macro is never tested and has no effect, because the work to support
112        coverage involves only compiling and linking options and special targets in
113        the Makefile. The comment in config.h implied that defining the macro would
114        enable coverage support, which is totally false. There was also support for
115        setting this macro in the CMake files (my fault, I just copied it from
116        configure). SUPPORT_GCOV has now been removed.
117    
118    3.  Make a small performance improvement in strlen16() and strlen32() in
119        pcretest.
120    
121    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
122        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
123        these cases). There was also one in pcretest.c.
124    
125    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
126    
127    6.  In UTF mode, the code for checking whether a group could match an empty
128        string (which is used for indefinitely repeated groups to allow for
129        breaking an infinite loop) was broken when the group contained a repeated
130        negated single-character class with a character that occupied more than one
131        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
132        UTF-8 mode). The effect was undefined: the group might or might not be
133        deemed as matching an empty string, or the program might have crashed.
134    
135    7.  The code for checking whether a group could match an empty string was not
136        recognizing that \h, \H, \v, \V, and \R must match a character.
137    
138    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
139        an empty string. If it can, pcretest shows this in its information output.
140    
141    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
142        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
143        matched by pcre_exec() without using JIT:
144    
145        (a) If the rest of the pattern did not match after a maximal run of
146            grapheme clusters, the code for backing up to try with fewer of them
147            did not always back up over a full grapheme when characters that do not
148            have the modifier quality were involved, e.g. Hangul syllables.
149    
150        (b) If the match point in a subject started with modifier character, and
151            there was no match, the code could incorrectly back up beyond the match
152            point, and potentially beyond the first character in the subject,
153            leading to a segfault or an incorrect match result.
154    
155    10. A conditional group with an assertion condition could lead to PCRE
156        recording an incorrect first data item for a match if no other first data
157        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
158        first data item, and therefore matched "ca" after "c" instead of at the
159        start.
160    
161    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
162        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
163    
164    12. The source of pcregrep now includes z/OS-specific code so that it can be
165        compiled for z/OS as part of the special z/OS distribution.
166    
167    13. Added the -T and -TM options to pcretest.
168    
169    14. The code in pcre_compile.c for creating the table of named capturing groups
170        has been refactored. Instead of creating the table dynamically during the
171        actual compiling pass, the information is remembered during the pre-compile
172        pass (on the stack unless there are more than 20 named groups, in which
173        case malloc() is used) and the whole table is created before the actual
174        compile happens. This has simplified the code (it is now nearly 150 lines
175        shorter) and prepared the way for better handling of references to groups
176        with duplicate names.
177    
178    15. A back reference to a named subpattern when there is more than one of the
179        same name now checks them in the order in which they appear in the pattern.
180        The first one that is set is used for the reference. Previously only the
181        first one was inspected. This change makes PCRE more compatible with Perl.
182    
183    16. Unicode character properties were updated from Unicode 6.3.0.
184    
185    17. The compile-time code for auto-possessification has been refactored, based
186        on a patch by Zoltan Herczeg. It now happens after instead of during
187        compilation. The code is cleaner, and more cases are handled. The option
188        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
189        options in pcretest are provided to set it. It can also be set by
190        (*NO_AUTO_POSSESS) at the start of a pattern.
191    
192    18. The character VT has been added to the default ("C" locale) set of
193        characters that match \s and are generally treated as white space,
194        following this same change in Perl 5.18. There is now no difference between
195        "Perl space" and "POSIX space". Whether VT is treated as white space in
196        other locales depends on the locale.
197    
198    19. The code for checking named groups as conditions, either for being set or
199        for being recursed, has been refactored (this is related to 14 and 15
200        above). Processing unduplicated named groups should now be as fast at
201        numerical groups, and processing duplicated groups should be faster than
202        before.
203    
204    20. Two patches to the CMake build system, by Alexander Barkov:
205    
206          (1) Replace the "source" command by "." in CMakeLists.txt because
207              "source" is a bash-ism.
208    
209          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
210              without these the CMake build does not work on Solaris.
211    
212    21. Perl has changed its handling of \8 and \9. If there is no previously
213        encountered capturing group of those numbers, they are treated as the
214        literal characters 8 and 9 instead of a binary zero followed by the
215        literals. PCRE now does the same.
216    
217    22. Following Perl, added \o{} to specify codepoints in octal, making it
218        possible to specify values greater than 0777 and also making them
219        unambiguous.
220    
221    23. Perl now gives an error for missing closing braces after \x{... instead of
222        treating the string as literal. PCRE now does the same.
223    
224    24. RunTest used to grumble if an inappropriate test was selected explicitly,
225        but just skip it when running all tests. This make it awkward to run ranges
226        of tests when one of them was inappropriate. Now it just skips any
227        inappropriate tests, as it always did when running all tests.
228    
229    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
230        character types such as \d or \w, too many callouts were inserted, and the
231        data that they returned was rubbish.
232    
233    26. In UCP mode, \s was not matching two of the characters that Perl matches,
234        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
235        were matched by \h. The code has now been refactored so that the lists of
236        the horizontal and vertical whitespace characters used for \h and \v (which
237        are defined only in one place) are now also used for \s.
238    
239    27. Add JIT support for the 64 bit TileGX architecture.
240        Patch by Jiong Wang (Tilera Corporation).
241    
242    28. Possessive quantifiers for classes (both explicit and automatically
243        generated) now use special opcodes instead of wrapping in ONCE brackets.
244    
245    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
246        (because it's meaningless), this was not happening when PCRE_CASELESS was
247        set. Not wrong, but inefficient.
248    
249    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
250        properties for \w, \d, etc) is present in a test regex. Otherwise if the
251        test contains no characters greater than 255, Perl doesn't realise it
252        should be using Unicode semantics.
253    
254    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
255        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
256        does in Unicode mode.
257    
258    32. Added the "forbid" facility to pcretest so that putting tests into the
259        wrong test files can sometimes be quickly detected.
260    
261    33. There is now a limit (default 250) on the depth of nesting of parentheses.
262        This limit is imposed to control the amount of system stack used at compile
263        time. It can be changed at build time by --with-parens-nest-limit=xxx or
264        the equivalent in CMake.
265    
266    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
267        errors. Perl warns for these when in warning mode, but PCRE has no facility
268        for giving warnings.
269    
270    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
271        However, this was not working for (?!) because it is optimized to (*FAIL),
272        for which PCRE does not allow quantifiers. The optimization is now disabled
273        when a quantifier follows (?!). I can't see any use for this, but it makes
274        things uniform.
275    
276    36. Perl no longer allows group names to start with digits, so I have made this
277        change also in PCRE. It simplifies the code a bit.
278    
279    37. In extended mode, Perl ignores spaces before a + that indicates a
280        possessive quantifier. PCRE allowed a space before the quantifier, but not
281        before the possessive +. It now does.
282    
283    38. The use of \K (reset reported match start) within a repeated possessive
284        group such as (a\Kb)*+ was not working.
285    
286    40. Document that the same character tables must be used at compile time and
287        run time, and that the facility to pass tables to pcre_exec() and
288        pcre_dfa_exec() is for use only with saved/restored patterns.
289    
290    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
291        features for Builds with MSVC:
292    
293        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
294           builds). The .rc files can be used to set FileDescription and many other
295           attributes.
296    
297        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
298           This allows higher-level build scripts which want .pdb files to avoid
299           hard-coding the exact files needed."
300    
301    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
302        mean "start of word" and "end of word", respectively, as a transition aid.
303    
304    43. A minimizing repeat of a class containing codepoints greater than 255 in
305        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
306        compiled to use the heap for recursion.
307    
308    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
309        is configured.
310    
311    
312    Version 8.33 28-May-2013
313    ------------------------
314    
315    1.  Added 'U' to some constants that are compared to unsigned integers, to
316        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
317        variables that are added to signed variables, to ensure the result is
318        signed and can be negated.
319    
320    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
321        CMake config files.
322    
323    3.  Revise the creation of config.h.generic so that all boolean macros are
324        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
325        overriding via -D on the command line possible.
326    
327    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
328        to unsigned int is reported to make a quite noticeable speed difference in
329        a specific Windows environment. Testing on Linux did also appear to show
330        some benefit (and it is clearly not harmful). Also fixed the definition of
331        Xop which should be unsigned.
332    
333    5.  Related to (4), changing the definition of the intermediate variable cc
334        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
335        improvements.
336    
337    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
338        unnecessary spaces.
339    
340    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
341        and later.
342    
343    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
344    
345    9.  Optimizing fast_forward_start_bits in JIT.
346    
347    10. Adding support for callouts in JIT, and fixing some issues revealed
348        during this work. Namely:
349    
350        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
351    
352        (b) Minimum length was not checked before the matching is started.
353    
354    11. The value of capture_last that is passed to callouts was incorrect in some
355        cases when there was a capture on one path that was subsequently abandoned
356        after a backtrack. Also, the capture_last value is now reset after a
357        recursion, since all captures are also reset in this case.
358    
359    12. The interpreter no longer returns the "too many substrings" error in the
360        case when an overflowing capture is in a branch that is subsequently
361        abandoned after a backtrack.
362    
363    13. In the pathological case when an offset vector of size 2 is used, pcretest
364        now prints out the matched string after a yield of 0 or 1.
365    
366    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
367        Only supported by the JIT compiler at the moment.
368    
369    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
370    
371    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
372        offset of the starting point of the matching process, provided the offsets
373        vector is large enough.
374    
375    17. The \A escape now records a lookbehind value of 1, though its execution
376        does not actually inspect the previous character. This is to ensure that,
377        in partial multi-segment matching, at least one character from the old
378        segment is retained when a new segment is processed. Otherwise, if there
379        are no lookbehinds in the pattern, \A might match incorrectly at the start
380        of a new segment.
381    
382    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
383    
384    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
385        gives some modest performance improvement in 8-bit mode.
386    
387    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
388        be expressed in certain programming languages using Universal Character
389        Names.
390    
391    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
392        which points out that "non characters" are not "characters that may not
393        appear in Unicode strings" but rather "characters that are reserved for
394        internal use and have only local meaning".
395    
396    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
397        there was a conditional group that depended on an assertion, if the
398        assertion was false, the callout that immediately followed the alternation
399        in the condition was skipped when pcre_exec() was used for matching.
400    
401    23. Allow an explicit callout to be inserted before an assertion that is the
402        condition for a conditional group, for compatibility with automatic
403        callouts, which always insert a callout at this point.
404    
405    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
406        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
407    
408    25. (*PRUNE) is now supported by the JIT compiler.
409    
410    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
411    
412    27. Fix the case where there are two or more SKIPs with arguments that may be
413        ignored.
414    
415    28. (*SKIP) is now supported by the JIT compiler.
416    
417    29. (*THEN) is now supported by the JIT compiler.
418    
419    30. Update RunTest with additional test selector options.
420    
421    31. The way PCRE handles backtracking verbs has been changed in two ways.
422    
423        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
424        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
425        backtracking. In some cases this makes it more Perl-compatible, but Perl's
426        rather obscure rules do not always do the same thing.
427    
428        (2) Previously, backtracking verbs were confined within assertions. This is
429        no longer the case for positive assertions, except for (*ACCEPT). Again,
430        this sometimes improves Perl compatibility, and sometimes does not.
431    
432    32. A number of tests that were in test 2 because Perl did things differently
433        have been moved to test 1, because either Perl or PCRE has changed, and
434        these tests are now compatible.
435    
436    32. Backtracking control verbs are now handled in the same way in JIT and
437        interpreter.
438    
439    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
440        contained a forward subroutine reference caused a compile error.
441    
442    34. Auto-detect and optimize limited repetitions in JIT.
443    
444    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
445        blocking (*UTF) etc.
446    
447    36. In the interpreter, maximizing pattern repetitions for characters and
448        character types now use tail recursion, which reduces stack usage.
449    
450    37. The value of the max lookbehind was not correctly preserved if a compiled
451        and saved regex was reloaded on a host of different endianness.
452    
453    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
454        of the compiled pattern block, expand the flags field from 16 to 32 bits
455        because it was almost full.
456    
457    39. Try madvise first before posix_madvise.
458    
459    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
460        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
461        original change locked out all matches of empty strings. This has been
462        changed so that one match of an empty string per line is recognized.
463        Subsequent searches on the same line (for colouring or for --only-matching,
464        for example) do not recognize empty strings.
465    
466    41. Applied a user patch to fix a number of spelling mistakes in comments.
467    
468    42. Data lines longer than 65536 caused pcretest to crash.
469    
470    43. Clarified the data type for length and startoffset arguments for pcre_exec
471        and pcre_dfa_exec in the function-specific man pages, where they were
472        explicitly stated to be in bytes, never having been updated. I also added
473        some clarification to the pcreapi man page.
474    
475    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
476        a segmentation fault.
477    
478    
479    Version 8.32 30-November-2012
480    -----------------------------
481    
482    1.  Improved JIT compiler optimizations for first character search and single
483        character iterators.
484    
485    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
486        Patch by Daniel Richard G.
487    
488    3.  Single character iterator optimizations in the JIT compiler.
489    
490    4.  Improved JIT compiler optimizations for character ranges.
491    
492    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
493        Reported by Giuseppe D'Angelo.
494    
495    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
496        of a line, was being set incorrectly in cases where .* appeared inside
497        atomic brackets at the start of a pattern, or where there was a subsequent
498        *PRUNE or *SKIP.
499    
500    7.  Improved instruction cache flush for POWER/PowerPC.
501        Patch by Daniel Richard G.
502    
503    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
504        grep:
505    
506        (a) There is now no limit to the number of patterns to be matched.
507    
508        (b) An error is given if a pattern is too long.
509    
510        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
511            are now supported.
512    
513        (d) --exclude-from and --include-from (multiple use) have been added.
514    
515        (e) Exclusions and inclusions now apply to all files and directories, not
516            just to those obtained from scanning a directory recursively.
517    
518        (f) Multiple uses of -f and --file-list are now supported.
519    
520        (g) In a Windows environment, the default for -d has been changed from
521            "read" (the GNU grep default) to "skip", because otherwise the presence
522            of a directory in the file list provokes an error.
523    
524        (h) The documentation has been revised and clarified in places.
525    
526    9.  Improve the matching speed of capturing brackets.
527    
528    10. Changed the meaning of \X so that it now matches a Unicode extended
529        grapheme cluster.
530    
531    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
532        out POSIX threads when JIT support is configured.
533    
534    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
535    
536    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
537        structure could go wrong in environments where size_t is not the same size
538        as int.
539    
540    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
541        set.
542    
543    15. The EBCDIC support had decayed; later updates to the code had included
544        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
545        general tidy up of EBCDIC-related issues, and the documentation was also
546        not quite right. There is now a test that can be run on ASCII systems to
547        check some of the EBCDIC-related things (but is it not a full test).
548    
549    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
550        in a small tidy to the code.
551    
552    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
553    
554    18. If the --only-matching (-o) option in pcregrep is specified multiple
555        times, each one causes appropriate output. For example, -o1 -o2 outputs the
556        substrings matched by the 1st and 2nd capturing parentheses. A separating
557        string can be specified by --om-separator (default empty).
558    
559    19. Improving the first n character searches.
560    
561    20. Turn case lists for horizontal and vertical white space into macros so that
562        they are defined only once.
563    
564    21. This set of changes together give more compatible Unicode case-folding
565        behaviour for characters that have more than one other case when UCP
566        support is available.
567    
568        (a) The Unicode property table now has offsets into a new table of sets of
569            three or more characters that are case-equivalent. The MultiStage2.py
570            script that generates these tables (the pcre_ucd.c file) now scans
571            CaseFolding.txt instead of UnicodeData.txt for character case
572            information.
573    
574        (b) The code for adding characters or ranges of characters to a character
575            class has been abstracted into a generalized function that also handles
576            case-independence. In UTF-mode with UCP support, this uses the new data
577            to handle characters with more than one other case.
578    
579        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
580            whose other case is greater than 256 are now correctly matched
581            caselessly. Previously, the high codepoint matched the low one, but not
582            vice versa.
583    
584        (d) The processing of \h, \H, \v, and \ in character classes now makes use
585            of the new class addition function, using character lists defined as
586            macros alongside the case definitions of 20 above.
587    
588        (e) Caseless back references now work with characters that have more than
589            one other case.
590    
591        (f) General caseless matching of characters with more than one other case
592            is supported.
593    
594    22. Unicode character properties were updated from Unicode 6.2.0
595    
596    23. Improved CMake support under Windows. Patch by Daniel Richard G.
597    
598    24. Add support for 32-bit character strings, and UTF-32
599    
600    25. Major JIT compiler update (code refactoring and bugfixing).
601        Experimental Sparc 32 support is added.
602    
603    26. Applied a modified version of Daniel Richard G's patch to create
604        pcre.h.generic and config.h.generic by "make" instead of in the
605        PrepareRelease script.
606    
607    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
608        pcre_compile.c when checking for a zero character.
609    
610    28. Introducing a native interface for JIT. Through this interface, the compiled
611        machine code can be directly executed. The purpose of this interface is to
612        provide fast pattern matching, so several sanity checks are not performed.
613        However, feature tests are still performed. The new interface provides
614        1.4x speedup compared to the old one.
615    
616    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
617        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
618        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
619    
620    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
621        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
622        the "old" RFC 2279). Instead, it ended up passing a negative length to
623        pcre_exec().
624    
625    31. Add support for GCC's visibility feature to hide internal functions.
626    
627    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
628        "unknown -C option" after outputting 0 or 1.
629    
630    33. There is now support for generating a code coverage report for the test
631        suite in environments where gcc is the compiler and lcov is installed. This
632        is mainly for the benefit of the developers.
633    
634    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
635        unaddressable using valgrind annotations, allowing valgrind to detect
636        invalid memory accesses. This is mainly for the benefit of the developers.
637    
638    25. (*UTF) can now be used to start a pattern in any of the three libraries.
639    
640    26. Give configure error if --enable-cpp but no C++ compiler found.
641    
642    
643    Version 8.31 06-July-2012
644    -------------------------
645    
646    1.  Fixing a wrong JIT test case and some compiler warnings.
647    
648    2.  Removed a bashism from the RunTest script.
649    
650    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
651        to unsigned type, result still unsigned" that was given by an MS compiler
652        on encountering the code "-sizeof(xxx)".
653    
654    4.  Partial matching support is added to the JIT compiler.
655    
656    5.  Fixed several bugs concerned with partial matching of items that consist
657        of more than one character:
658    
659        (a) /^(..)\1/ did not partially match "aba" because checking references was
660            done on an "all or nothing" basis. This also applied to repeated
661            references.
662    
663        (b) \R did not give a hard partial match if \r was found at the end of the
664            subject.
665    
666        (c) \X did not give a hard partial match after matching one or more
667            characters at the end of the subject.
668    
669        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
670            a partial match for the string "\r".
671    
672        (e) When newline was set to CRLF, the metacharacter "." did not recognize
673            a partial match for a CR character at the end of the subject string.
674    
675    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
676        running pcretest, the text "(JIT)" added to the output whenever JIT is
677        actually used to run the match.
678    
679    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
680        or /S+[+] with a digit between 1 and 7.
681    
682    8.  OP_NOT now supports any UTF character not just single-byte ones.
683    
684    9.  (*MARK) control verb is now supported by the JIT compiler.
685    
686    10. The command "./RunTest list" lists the available tests without actually
687        running any of them. (Because I keep forgetting what they all are.)
688    
689    11. Add PCRE_INFO_MAXLOOKBEHIND.
690    
691    12. Applied a (slightly modified) user-supplied patch that improves performance
692        when the heap is used for recursion (compiled with --disable-stack-for-
693        recursion). Instead of malloc and free for each heap frame each time a
694        logical recursion happens, frames are retained on a chain and re-used where
695        possible. This sometimes gives as much as 30% improvement.
696    
697    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
698        call.
699    
700    14. As documented, (*COMMIT) is now confined to within a positive assertion.
701    
702    15. It is now possible to link pcretest with libedit as an alternative to
703        libreadline.
704    
705    16. (*COMMIT) control verb is now supported by the JIT compiler.
706    
707    17. The Unicode data tables have been updated to Unicode 6.1.0.
708    
709    18. Added --file-list option to pcregrep.
710    
711    19. Added binary file support to pcregrep, including the -a, --binary-files,
712        -I, and --text options.
713    
714    20. The madvise function is renamed for posix_madvise for QNX compatibility
715        reasons. Fixed by Giuseppe D'Angelo.
716    
717    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
718    
719    22. Changed the option for creating symbolic links for 16-bit man pages from
720        -s to -sf so that re-installing does not cause issues.
721    
722    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
723    
724    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
725        matches in certain environments (the workspace was not being correctly
726        retained). Also added to pcre_dfa_exec() a simple plausibility check on
727        some of the workspace data at the beginning of a restart.
728    
729    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
730        was not doing so when it should - probably a typo introduced by SVN 528
731        (change 8.10/14).
732    
733    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
734        \w+ when the character tables indicated that \x{c4} was a word character.
735        There were several related cases, all because the tests for doing a table
736        lookup were testing for characters less than 127 instead of 255.
737    
738    27. If a pattern contains capturing parentheses that are not used in a match,
739        their slots in the ovector are set to -1. For those that are higher than
740        any matched groups, this happens at the end of processing. In the case when
741        there were back references that the ovector was too small to contain
742        (causing temporary malloc'd memory to be used during matching), and the
743        highest capturing number was not used, memory off the end of the ovector
744        was incorrectly being set to -1. (It was using the size of the temporary
745        memory instead of the true size.)
746    
747    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
748        ovector size, it uses memory at the end of the block that it has got.
749    
750    29. Check for an overlong MARK name and give an error at compile time. The
751        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
752    
753    30. JIT compiler update.
754    
755    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
756        Rill for the patch.
757    
758    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
759    
760    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
761    
762    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
763        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
764        was enabled.
765    
766    35. Improve JIT code generation for greedy plus quantifier.
767    
768    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
769        1 to "aa" instead of to an empty string. The bug affected repeated groups
770        that could potentially match an empty string.
771    
772    37. Optimizing single character iterators in JIT.
773    
774    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
775        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
776        codepoints that are too big for the mode are faulted, and in a UTF mode,
777        disallowed codepoints are also faulted.
778    
779    39. If PCRE was compiled with UTF support, in three places in the DFA
780        matcher there was code that should only have been obeyed in UTF mode, but
781        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
782        processing when bytes with values greater than 127 were present. In 16-bit
783        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
784        both cases the values are those that cannot be the first data item in a UTF
785        character. The three items that might have provoked this were recursions,
786        possessively repeated groups, and atomic groups.
787    
788    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
789        and pcregrep, because some OS require shared objects to be explicitly
790        passed to ld, causing the link step to fail if they are not.
791    
792    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
793        mode, patterns that started with \h* or \R* might be incorrectly matched.
794    
795    
796    Version 8.30 04-February-2012
797    -----------------------------
798    
799    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
800        name is defined in ctype.h.
801    
802    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
803        only in quite long subpatterns.
804    
805    3.  Removed the function pcre_info(), which has been obsolete and deprecated
806        since it was replaced by pcre_fullinfo() in February 2000.
807    
808    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
809        match a (*MARK), and the match failed at the start of the subject, a
810        reference to memory before the start of the subject could occur. This bug
811        was introduced by fix 17 of release 8.21.
812    
813    5.  A reference to an unset group with zero minimum repetition was giving
814        totally wrong answers (in non-JavaScript-compatibility mode). For example,
815        /(another)?(\1?)test/ matched against "hello world test". This bug was
816        introduced in release 8.13.
817    
818    6.  Add support for 16-bit character strings (a large amount of work involving
819        many changes and refactorings).
820    
821    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
822        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
823        from a file.
824    
825    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
826        rounding is not applied in this particular case).
827    
828    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
829        if they appear, or are escaped, in patterns.
830    
831    10. Get rid of a number of -Wunused-but-set-variable warnings.
832    
833    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
834        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
835        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
836        also returns the mark "x". This bug applied to capturing parentheses,
837        non-capturing parentheses, and atomic parentheses. It also applied to some
838        assertions.
839    
840    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
841        information out of configure.ac instead of relying on pcre.h.generic, which
842        is not stored in the repository.
843    
844    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
845        -lreadline.
846    
847    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
848    
849    15. Applied Graycode's patch to put the top-level frame on the stack rather
850        than the heap when not using the stack for recursion. This gives a
851        performance improvement in many cases when recursion is not deep.
852    
853    16. Experimental code added to "pcretest -C" to output the stack frame size.
854    
855    
856    Version 8.21 12-Dec-2011
857    ------------------------
858    
859    1.  Updating the JIT compiler.
860    
861    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
862        are added as well.
863    
864    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
865        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
866        calling _pcre_jit_exec. Some extra comments are added.
867    
868    4.  (*MARK) settings inside atomic groups that do not contain any capturing
869        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
870        was introduced by change 18 for 8.20.
871    
872    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
873        ECMA-262 standard.
874    
875    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
876        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
877        This bug was probably introduced by change 9 of 8.13.
878    
879    7.  While fixing 6 above, I noticed that a number of other items were being
880        incorrectly rejected as "not fixed length". This arose partly because newer
881        opcodes had not been added to the fixed-length checking code. I have (a)
882        corrected the bug and added tests for these items, and (b) arranged for an
883        error to occur if an unknown opcode is encountered while checking for fixed
884        length instead of just assuming "not fixed length". The items that were
885        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
886        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
887        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
888    
889    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
890        being incorrectly compiled and would have given unpredicatble results.
891    
892    9.  A possessively repeated subpattern with minimum repeat count greater than
893        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
894        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
895        the first (A) could occur when it should not.
896    
897    10. Add a cast and remove a redundant test from the code.
898    
899    11. JIT should use pcre_malloc/pcre_free for allocation.
900    
901    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
902        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
903        is anything other than /usr, -L is still shown).
904    
905    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
906    
907    14. Perl does not support \N without a following name in a [] class; PCRE now
908        also gives an error.
909    
910    15. If a forward reference was repeated with an upper limit of around 2000,
911        it caused the error "internal error: overran compiling workspace". The
912        maximum number of forward references (including repeats) was limited by the
913        internal workspace, and dependent on the LINK_SIZE. The code has been
914        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
915        the default depends on LINK_SIZE. There is a new upper limit (for safety)
916        of around 200,000 forward references. While doing this, I also speeded up
917        the filling in of repeated forward references.
918    
919    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
920        incorrectly expecting the subject to contain another "a" after the start.
921    
922    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
923        in the match, the SKIP should be ignored. This was not happening; instead
924        the SKIP was being treated as NOMATCH. For patterns such as
925        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
926        tested.
927    
928    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
929        now much more compatible with Perl, in particular in cases where the result
930        is a non-match for a non-anchored pattern. For example, if
931        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
932        "m", where previously it did not return a name. A side effect of this
933        change is that for partial matches, the last encountered mark name is
934        returned, as for non matches. A number of tests that were previously not
935        Perl-compatible have been moved into the Perl-compatible test files. The
936        refactoring has had the pleasing side effect of removing one argument from
937        the match() function, thus reducing its stack requirements.
938    
939    19. If the /S+ option was used in pcretest to study a pattern using JIT,
940        subsequent uses of /S (without +) incorrectly behaved like /S+.
941    
942    21. Retrieve executable code size support for the JIT compiler and fixing
943        some warnings.
944    
945    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
946        not work when the shorter character appeared right at the end of the
947        subject string.
948    
949    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
950        systems.
951    
952    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
953        output it when the /M option is used in pcretest.
954    
955    25. The CheckMan script was not being included in the distribution. Also, added
956        an explicit "perl" to run Perl scripts from the PrepareRelease script
957        because this is reportedly needed in Windows.
958    
959    26. If study data was being save in a file and studying had not found a set of
960        "starts with" bytes for the pattern, the data written to the file (though
961        never used) was taken from uninitialized memory and so caused valgrind to
962        complain.
963    
964    27. Updated RunTest.bat as provided by Sheri Pierce.
965    
966    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
967    
968    29. Computation of memory usage for the table of capturing group names was
969        giving an unnecessarily large value.
970    
971    
972    Version 8.20 21-Oct-2011
973    ------------------------
974    
975    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
976        a POSIX class. After further experiments with Perl, which convinced me that
977        Perl has bugs and confusions, a closing square bracket is no longer allowed
978        in a POSIX name. This bug also affected patterns with classes that started
979        with full stops.
980    
981    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
982        captured substring, but while checking the failing first alternative,
983        substring 1 is temporarily captured. If the output vector supplied to
984        pcre_exec() was not big enough for this capture, the yield of the function
985        was still zero ("insufficient space for captured substrings"). This cannot
986        be totally fixed without adding another stack variable, which seems a lot
987        of expense for a edge case. However, I have improved the situation in cases
988        such as /(a)(b)x|abc/ matched against "abc", where the return code
989        indicates that fewer than the maximum number of slots in the ovector have
990        been set.
991    
992    3.  Related to (2) above: when there are more back references in a pattern than
993        slots in the output vector, pcre_exec() uses temporary memory during
994        matching, and copies in the captures as far as possible afterwards. It was
995        using the entire output vector, but this conflicts with the specification
996        that only 2/3 is used for passing back captured substrings. Now it uses
997        only the first 2/3, for compatibility. This is, of course, another edge
998        case.
999    
1000    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1001        main code base, and can be used by building with --enable-jit. When this is
1002        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1003        runtime --no-jit option is given.
1004    
1005    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1006        ovector, the return from the function was zero, implying that there were
1007        other matches that did not fit. The correct "exactly full" value is now
1008        returned.
1009    
1010    6.  If a subpattern that was called recursively or as a subroutine contained
1011        (*PRUNE) or any other control that caused it to give a non-standard return,
1012        invalid errors such as "Error -26 (nested recursion at the same subject
1013        position)" or even infinite loops could occur.
1014    
1015    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1016        computing the minimum length on reaching *ACCEPT, and so ended up with the
1017        wrong value of 1 rather than 0. Further investigation indicates that
1018        computing a minimum subject length in the presence of *ACCEPT is difficult
1019        (think back references, subroutine calls), and so I have changed the code
1020        so that no minimum is registered for a pattern that contains *ACCEPT.
1021    
1022    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1023        it was not handled as intended. [But see 16 below.]
1024    
1025    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1026        Sheri Pierce.
1027    
1028    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1029        the first byte in a match must be "a".
1030    
1031    11. Change 17 for 8.13 increased the recursion depth for patterns like
1032        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1033        pattern contains any instances of (*THEN). If it does not, the old
1034        optimizations are restored. It would be nice to do this on a per-group
1035        basis, but at the moment that is not feasible.
1036    
1037    12. In some environments, the output of pcretest -C is CRLF terminated. This
1038        broke RunTest's code that checks for the link size. A single white space
1039        character after the value is now allowed for.
1040    
1041    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1042        For "fr", it uses the Windows-specific input and output files.
1043    
1044    14. If (*THEN) appeared in a group that was called recursively or as a
1045        subroutine, it did not work as intended. [But see next item.]
1046    
1047    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1048        pattern fragments (but not containing any | characters). If A and B are
1049        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1050        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1051        D. In other words, Perl considers parentheses that do not contain any |
1052        characters to be part of a surrounding alternative, whereas PCRE was
1053        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1054        differently. PCRE now behaves in the same way as Perl, except in the case
1055        of subroutine/recursion calls such as (?1) which have in any case always
1056        been different (but PCRE had them first :-).
1057    
1058    16. Related to 15 above: Perl does not treat the | in a conditional group as
1059        creating alternatives. Such a group is treated in the same way as an
1060        ordinary group without any | characters when processing (*THEN). PCRE has
1061        been changed to match Perl's behaviour.
1062    
1063    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1064        RunGrepTest script failed.
1065    
1066    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1067        inevitable for groups that contain captures, but it can lead to a lot of
1068        stack use in large patterns. The old behaviour has been restored for atomic
1069        groups that do not contain any capturing parentheses.
1070    
1071    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1072        suppress the check for a minimum subject length at run time. (If it was
1073        given to pcre_exec() or pcre_dfa_exec() it did work.)
1074    
1075    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1076        fail to work when decoding hex characters in data strings in EBCDIC
1077        environments.
1078    
1079    21. It appears that in at least one Mac OS environment, the isxdigit() function
1080        is implemented as a macro that evaluates to its argument more than once,
1081        contravening the C 90 Standard (I haven't checked a later standard). There
1082        was an instance in pcretest which caused it to go wrong when processing
1083        \x{...} escapes in subject strings. The has been rewritten to avoid using
1084        things like p++ in the argument of isxdigit().
1085    
1086    
1087    Version 8.13 16-Aug-2011
1088    ------------------------
1089    
1090    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1091    
1092    2.  Two minor typos in pcre_internal.h have been fixed.
1093    
1094    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1095        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1096        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1097    
1098    4.  There were a number of related bugs in the code for matching backrefences
1099        caselessly in UTF-8 mode when codes for the characters concerned were
1100        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1101        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1102        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1103        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1104        2-byte code at the end of the subject (it thought there wasn't enough data
1105        left).
1106    
1107    5.  Comprehensive information about what went wrong is now returned by
1108        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1109        as the output vector has at least 2 elements. The offset of the start of
1110        the failing character and a reason code are placed in the vector.
1111    
1112    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1113        now returned is for the first byte of the failing character, instead of the
1114        last byte inspected. This is an incompatible change, but I hope it is small
1115        enough not to be a problem. It makes the returned offset consistent with
1116        pcre_exec() and pcre_dfa_exec().
1117    
1118    7.  pcretest now gives a text phrase as well as the error number when
1119        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1120        failure, the offset and reason code are output.
1121    
1122    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1123        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1124        back over a single character (\n). This seems wrong (because it treated the
1125        two characters as a single entity when going forwards), conflicts with the
1126        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1127        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1128        has been changed.
1129    
1130    9.  Some internal refactoring has changed the processing so that the handling
1131        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1132        time (the PCRE_DOTALL option was changed this way some time ago: version
1133        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1134        which was always a bit of a fudge. It also means that there is one less
1135        argument for the match() function, which reduces its stack requirements
1136        slightly. This change also fixes an incompatibility with Perl: the pattern
1137        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1138    
1139    10. More internal refactoring has drastically reduced the number of recursive
1140        calls to match() for possessively repeated groups such as (abc)++ when
1141        using pcre_exec().
1142    
1143    11. While implementing 10, a number of bugs in the handling of groups were
1144        discovered and fixed:
1145    
1146        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1147        (a|)*(?1) gave a compile-time internal error.
1148        ((a|)+)+  did not notice that the outer group could match an empty string.
1149        (^a|^)+   was not marked as anchored.
1150        (.*a|.*)+ was not marked as matching at start or after a newline.
1151    
1152    12. Yet more internal refactoring has removed another argument from the match()
1153        function. Special calls to this function are now indicated by setting a
1154        value in a variable in the "match data" data block.
1155    
1156    13. Be more explicit in pcre_study() instead of relying on "default" for
1157        opcodes that mean there is no starting character; this means that when new
1158        ones are added and accidentally left out of pcre_study(), testing should
1159        pick them up.
1160    
1161    14. The -s option of pcretest has been documented for ages as being an old
1162        synonym of -m (show memory usage). I have changed it to mean "force study
1163        for every regex", that is, assume /S for every regex. This is similar to -i
1164        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1165        using it. It makes it easier to run collections of tests with and without
1166        study enabled, and thereby test pcre_study() more easily. All the standard
1167        tests are now run with and without -s (but some patterns can be marked as
1168        "never study" - see 20 below).
1169    
1170    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1171        restoration of the capturing data to the outer values was not happening
1172        correctly.
1173    
1174    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1175        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1176        pattern had matched an empty string, and so incorrectly returned a no
1177        match.
1178    
1179    17. There was optimizing code for the last branch of non-capturing parentheses,
1180        and also for the obeyed branch of a conditional subexpression, which used
1181        tail recursion to cut down on stack usage. Unfortunately, now that there is
1182        the possibility of (*THEN) occurring in these branches, tail recursion is
1183        no longer possible because the return has to be checked for (*THEN). These
1184        two optimizations have therefore been removed. [But see 8.20/11 above.]
1185    
1186    18. If a pattern containing \R was studied, it was assumed that \R always
1187        matched two bytes, thus causing the minimum subject length to be
1188        incorrectly computed because \R can also match just one byte.
1189    
1190    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1191        was incorrectly computed.
1192    
1193    20. If /S is present twice on a test pattern in pcretest input, it now
1194        *disables* studying, thereby overriding the use of -s on the command line
1195        (see 14 above). This is necessary for one or two tests to keep the output
1196        identical in both cases.
1197    
1198    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1199        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1200    
1201    22. When an atomic group that contained a capturing parenthesis was
1202        successfully matched, but the branch in which it appeared failed, the
1203        capturing was not being forgotten if a higher numbered group was later
1204        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1205        group 1 to "a", when in fact it should be unset. This applied to multi-
1206        branched capturing and non-capturing groups, repeated or not, and also to
1207        positive assertions (capturing in negative assertions does not happen
1208        in PCRE) and also to nested atomic groups.
1209    
1210    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1211        subject after a captured substring, to make it easier to tell which of a
1212        number of identical substrings has been captured.
1213    
1214    24. The way atomic groups are processed by pcre_exec() has been changed so that
1215        if they are repeated, backtracking one repetition now resets captured
1216        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1217        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1218        "aaa". Previously, it would have been "a". As part of this code
1219        refactoring, the way recursive calls are handled has also been changed.
1220    
1221    25. If an assertion condition captured any substrings, they were not passed
1222        back unless some other capturing happened later. For example, if
1223        (?(?=(a))a) was matched against "a", no capturing was returned.
1224    
1225    26. When studying a pattern that contained subroutine calls or assertions,
1226        the code for finding the minimum length of a possible match was handling
1227        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1228        group 1 called group 2 while simultaneously a separate group 2 called group
1229        1). A stack overflow occurred in this case. I have fixed this by limiting
1230        the recursion depth to 10.
1231    
1232    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1233        Fortmann. This supports explicit test numbers on the command line, and has
1234        argument validation and error reporting.
1235    
1236    28. An instance of \X with an unlimited repeat could fail if at any point the
1237        first character it looked at was a mark character.
1238    
1239    29. Some minor code refactoring concerning Unicode properties and scripts
1240        should reduce the stack requirement of match() slightly.
1241    
1242    30. Added the '=' option to pcretest to check the setting of unused capturing
1243        slots at the end of the pattern, which are documented as being -1, but are
1244        not included in the return count.
1245    
1246    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1247        compiled something random. Now it gives a compile-time error (as does
1248        Perl).
1249    
1250    32. A *MARK encountered during the processing of a positive assertion is now
1251        recorded and passed back (compatible with Perl).
1252    
1253    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1254        had alternative anchored branches, the search for a second match in a line
1255        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1256        matched the line "0102" twice. The same bug affected patterns that started
1257        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1258        twice.
1259    
1260    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1261        does, and because of capturing effects, quantifying parenthesized
1262        assertions may at times be useful. Quantifiers are now allowed for
1263        parenthesized assertions.
1264    
1265    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1266    
1267    36. \g was being checked for fancy things in a character class, when it should
1268        just be a literal "g".
1269    
1270    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1271        appearance of a nested POSIX class supersedes an apparent external class.
1272        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1273        unescaped square brackets may also appear as part of class names. For
1274        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1275        more like Perl. (But see 8.20/1 above.)
1276    
1277    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1278        was because it thought it was \N{name}, which is not supported).
1279    
1280    39. Add minix to OS list not supporting the -S option in pcretest.
1281    
1282    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1283        cannot analyze patterns in sufficient detail to catch mutual recursions
1284        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1285        subgroup is called recursively as a subpattern for a second time at the
1286        same position in the subject string. In previous releases this might have
1287        been caught by the recursion limit, or it might have run out of stack.
1288    
1289    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1290        happen only once. PCRE was, however incorrectly giving a compile time error
1291        "recursive call could loop indefinitely" because it cannot analyze the
1292        pattern in sufficient detail. The compile time test no longer happens when
1293        PCRE is compiling a conditional subpattern, but actual runaway loops are
1294        now caught at runtime (see 40 above).
1295    
1296    42. It seems that Perl allows any characters other than a closing parenthesis
1297        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1298        has been changed to be the same.
1299    
1300    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1301        as not to get warnings when autogen.sh is called. Also changed
1302        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1303    
1304    44. To help people who use pcregrep to scan files containing exceedingly long
1305        lines, the following changes have been made:
1306    
1307        (a) The default value of the buffer size parameter has been increased from
1308            8K to 20K. (The actual buffer used is three times this size.)
1309    
1310        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1311            PCRE is built.
1312    
1313        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1314            to be set at run time.
1315    
1316        (d) Numerical values in pcregrep options can be followed by K or M, for
1317            example --buffer-size=50K.
1318    
1319        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1320            given and the return code is set to 2.
1321    
1322    45. Add a pointer to the latest mark to the callout data block.
1323    
1324    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1325        partial match of an empty string instead of no match. This was specific to
1326        the use of ".".
1327    
1328    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1329        complete match instead of a partial match. This bug was dependent on both
1330        the PCRE_UTF8 and PCRE_DOTALL options being set.
1331    
1332    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1333        starting byte set, because \b was not being ignored.
1334    
1335    
1336    Version 8.12 15-Jan-2011
1337    ------------------------
1338    
1339    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1340        checks for such things as part of the documentation building process.
1341    
1342    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1343        --match-limit and --recursion-limit options (added for 8.11). In
1344        particular, this made one of the standard tests fail. (The integer value
1345        went into the wrong half of a long int.)
1346    
1347    3.  If the --colour option was given to pcregrep with -v (invert match), it
1348        did strange things, either producing crazy output, or crashing. It should,
1349        of course, ignore a request for colour when reporting lines that do not
1350        match.
1351    
1352    4.  Another pcregrep bug caused similar problems if --colour was specified with
1353        -M (multiline) and the pattern match finished with a line ending.
1354    
1355    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1356        matched in multiline mode, the following line was shown as part of the
1357        match. This seems wrong, so I have changed it.
1358    
1359    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1360        the check for further matches in the same line (so they could be coloured)
1361        to overrun the end of the current line. If another match was found, it was
1362        incorrectly shown (and then shown again when found in the next line).
1363    
1364    7.  If pcregrep was compiled under Windows, there was a reference to the
1365        function pcregrep_exit() before it was defined. I am assuming this was
1366        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1367        reported by a user. I've moved the definition above the reference.
1368    
1369    
1370    Version 8.11 10-Dec-2010
1371    ------------------------
1372    
1373    1.  (*THEN) was not working properly if there were untried alternatives prior
1374        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1375        backtracked to try for "b" instead of moving to the next alternative branch
1376        at the same level (in this case, to look for "c"). The Perl documentation
1377        is clear that when (*THEN) is backtracked onto, it goes to the "next
1378        alternative in the innermost enclosing group".
1379    
1380    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1381        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1382        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1383        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1384        (*THEN).
1385    
1386    3.  If \s appeared in a character class, it removed the VT character from
1387        the class, even if it had been included by some previous item, for example
1388        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1389        of \s, but is part of the POSIX "space" class.)
1390    
1391    4.  A partial match never returns an empty string (because you can always
1392        match an empty string at the end of the subject); however the checking for
1393        an empty string was starting at the "start of match" point. This has been
1394        changed to the "earliest inspected character" point, because the returned
1395        data for a partial match starts at this character. This means that, for
1396        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1397        (previously it gave "no match").
1398    
1399    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1400        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1401        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1402        has an implication that the given string is incomplete (because a partial
1403        match is preferred over a full match). For this reason, these items now
1404        give a partial match in this situation. [Aside: previously, the one case
1405        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1406        match rather than a full match, which was wrong by the old rules, but is
1407        now correct.]
1408    
1409    6.  There was a bug in the handling of #-introduced comments, recognized when
1410        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1411        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1412        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1413        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1414        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1415        places in pcre_compile().
1416    
1417    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1418        comments when looking ahead for named forward references to subpatterns,
1419        the only newline sequence it recognized was NL. It now handles newlines
1420        according to the set newline convention.
1421    
1422    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1423        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1424        cater for a lack of strerror(). These oversights have been fixed.
1425    
1426    9.  Added --match-limit and --recursion-limit to pcregrep.
1427    
1428    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1429    
1430    11. When the -o option was used, pcregrep was setting a return code of 1, even
1431        when matches were found, and --line-buffered was not being honoured.
1432    
1433    12. Added an optional parentheses number to the -o and --only-matching options
1434        of pcregrep.
1435    
1436    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1437        can match an empty string. The code to do it in pcretest and pcredemo
1438        needed fixing:
1439    
1440        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1441            only one byte after an empty string match just before CRLF (this case
1442            just got forgotten; "any" and "anycrlf" were OK).
1443    
1444        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1445            mode when an empty string match preceded an ASCII character followed by
1446            a non-ASCII character. (The code for advancing by one character rather
1447            than one byte was nonsense.)
1448    
1449        (c) The pcredemo.c sample program did not have any code at all to handle
1450            the cases when CRLF is a valid newline sequence.
1451    
1452    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1453        as a starting offset was within the subject string. There is now a new
1454        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1455        negative or greater than the length of the string. In order to test this,
1456        pcretest is extended to allow the setting of negative starting offsets.
1457    
1458    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1459        starting offset points to the beginning of a UTF-8 character was
1460        unnecessarily clumsy. I tidied it up.
1461    
1462    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1463        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1464    
1465    17. Nobody had reported that the --include_dir option, which was added in
1466        release 7.7 should have been called --include-dir (hyphen, not underscore)
1467        for compatibility with GNU grep. I have changed it to --include-dir, but
1468        left --include_dir as an undocumented synonym, and the same for
1469        --exclude-dir, though that is not available in GNU grep, at least as of
1470        release 2.5.4.
1471    
1472    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1473        characters from a string of bytes) have been redefined so as not to use
1474        loops, in order to improve performance in some environments. At the same
1475        time, I abstracted some of the common code into auxiliary macros to save
1476        repetition (this should not affect the compiled code).
1477    
1478    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1479        compile-time error is now given if \c is not followed by an ASCII
1480        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1481        different, and any byte value is allowed.)
1482    
1483    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1484        START_OPTIMIZE option, which is now allowed at compile time - but just
1485        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1486        to pcregrep and other applications that have no direct access to PCRE
1487        options. The new /Y option in pcretest sets this option when calling
1488        pcre_compile().
1489    
1490    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1491        back references. Groups containing recursive back references were forced to
1492        be atomic by that change, but in the case of named groups, the amount of
1493        memory required was incorrectly computed, leading to "Failed: internal
1494        error: code overflow". This has been fixed.
1495    
1496    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1497        pcretest.c, to avoid build problems in some Borland environments.
1498    
1499    
1500    Version 8.10 25-Jun-2010
1501    ------------------------
1502    
1503    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1504        THEN.
1505    
1506    2.  (*ACCEPT) was not working when inside an atomic group.
1507    
1508    3.  Inside a character class, \B is treated as a literal by default, but
1509        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1510        causes the error). The code is unchanged, but I tidied the documentation.
1511    
1512    4.  Inside a character class, PCRE always treated \R and \X as literals,
1513        whereas Perl faults them if its -w option is set. I have changed PCRE so
1514        that it faults them when PCRE_EXTRA is set.
1515    
1516    5.  Added support for \N, which always matches any character other than
1517        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1518    
1519    6.  When compiling pcregrep with newer versions of gcc which may have
1520        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1521        declared with attribute warn_unused_result" were given. Just casting the
1522        result to (void) does not stop the warnings; a more elaborate fudge is
1523        needed. I've used a macro to implement this.
1524    
1525    7.  Minor change to pcretest.c to avoid a compiler warning.
1526    
1527    8.  Added four artifical Unicode properties to help with an option to make
1528        \s etc use properties (see next item). The new properties are: Xan
1529        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1530    
1531    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1532        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1533        this option. Modified pcretest to add /W to test this facility. Added
1534        REG_UCP to make it available via the POSIX interface.
1535    
1536    10. Added --line-buffered to pcregrep.
1537    
1538    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1539        studied, and the match started with a letter with a code point greater than
1540        127 whose first byte was different to the first byte of the other case of
1541        the letter, the other case of this starting letter was not recognized
1542        (#976).
1543    
1544    12. If a pattern that was studied started with a repeated Unicode property
1545        test, for example, \p{Nd}+, there was the theoretical possibility of
1546        setting up an incorrect bitmap of starting bytes, but fortunately it could
1547        not have actually happened in practice until change 8 above was made (it
1548        added property types that matched character-matching opcodes).
1549    
1550    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1551        possible starting bytes for non-anchored patterns.
1552    
1553    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1554        \R, and also a number of cases that involve Unicode properties, both
1555        explicit and implicit when PCRE_UCP is set.
1556    
1557    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1558        input, it could crash or give wrong results if characters with values
1559        greater than 0xc0 were present in the subject string. (Detail: it assumed
1560        UTF-8 input when processing these items.)
1561    
1562    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1563        size_t is 64-bit (#991).
1564    
1565    17. Added a check for running out of memory when PCRE is compiled with
1566        --disable-stack-for-recursion (#990).
1567    
1568    18. If the last data line in a file for pcretest does not have a newline on
1569        the end, a newline was missing in the output.
1570    
1571    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1572        less than 128) in its various bitmaps. However, there is a facility for
1573        generating tables according to the current locale when PCRE is compiled. It
1574        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1575        characters, are recognized by isspace() and therefore were getting set in
1576        these tables, and indeed these tables seem to approximate to ISO 8859. This
1577        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1578        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1579        which of course cannot start UTF-8 characters. I have changed the code so
1580        that only real ASCII characters (less than 128) and the correct starting
1581        bytes for UTF-8 encodings are set for characters greater than 127 when in
1582        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1583        altogether.)
1584    
1585    20. Added the /T option to pcretest so as to be able to run tests with non-
1586        standard character tables, thus making it possible to include the tests
1587        used for 19 above in the standard set of tests.
1588    
1589    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1590        reference to a subpattern the other side of a comment that contains an
1591        opening parenthesis caused either an internal compiling error, or a
1592        reference to the wrong subpattern.
1593    
1594    
1595    Version 8.02 19-Mar-2010
1596    ------------------------
1597    
1598    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1599    
1600    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1601        configured.
1602    
1603    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1604        original author of that file, following a query about its status.
1605    
1606    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1607        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1608    
1609    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1610        quantifier applied to a forward-referencing subroutine call, could compile
1611        incorrect code or give the error "internal error: previously-checked
1612        referenced subpattern not found".
1613    
1614    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1615        variables to point to external functions. For these systems, therefore,
1616        pcre_malloc etc. are now initialized to local functions that call the
1617        relevant global functions.
1618    
1619    7.  There were two entries missing in the vectors called coptable and poptable
1620        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1621        I've fixed the data, and added a kludgy way of testing at compile time that
1622        the lengths are correct (equal to the number of opcodes).
1623    
1624    8.  Following on from 7, I added a similar kludge to check the length of the
1625        eint vector in pcreposix.c.
1626    
1627    9.  Error texts for pcre_compile() are held as one long string to avoid too
1628        much relocation at load time. To find a text, the string is searched,
1629        counting zeros. There was no check for running off the end of the string,
1630        which could happen if a new error number was added without updating the
1631        string.
1632    
1633    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1634    
1635    11. \K was not working if it appeared in an atomic group or in a group that
1636        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1637        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1638        the assertion is positive, but not if it is negative.
1639    
1640    12. Change 11 fortuitously reduced the size of the stack frame used in the
1641        "match()" function of pcre_exec.c by one pointer. Forthcoming
1642        implementation of support for (*MARK) will need an extra pointer on the
1643        stack; I have reserved it now, so that the stack frame size does not
1644        decrease.
1645    
1646    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1647        item in branch that calls a recursion is a subroutine call - as in the
1648        second branch in the above example - was incorrectly given the compile-
1649        time error "recursive call could loop indefinitely" because pcre_compile()
1650        was not correctly checking the subroutine for matching a non-empty string.
1651    
1652    14. The checks for overrunning compiling workspace could trigger after an
1653        overrun had occurred. This is a "should never occur" error, but it can be
1654        triggered by pathological patterns such as hundreds of nested parentheses.
1655        The checks now trigger 100 bytes before the end of the workspace.
1656    
1657    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1658    
1659    
1660    Version 8.01 19-Jan-2010
1661    ------------------------
1662    
1663    1.  If a pattern contained a conditional subpattern with only one branch (in
1664        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1665        computed the wrong minimum data length (which is of course zero for such
1666        subpatterns). This could cause incorrect "no match" results.
1667    
1668    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1669        the pattern is reset in the first branch, pcre_compile() failed with
1670        "internal error: code overflow at offset...". This happened only when
1671        the reset was to the original external option setting. (An optimization
1672        abstracts leading options settings into an external setting, which was the
1673        cause of this.)
1674    
1675    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1676        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1677        assertion pattern did not match (meaning that the assertion was true), it
1678        was incorrectly treated as false if the SKIP had been reached during the
1679        matching. This also applied to assertions used as conditions.
1680    
1681    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1682        assertion subpattern, including such a pattern used as a condition,
1683        unpredictable results occurred, instead of the error return
1684        PCRE_ERROR_DFA_UITEM.
1685    
1686    5.  The C++ GlobalReplace function was not working like Perl for the special
1687        situation when an empty string is matched. It now does the fancy magic
1688        stuff that is necessary.
1689    
1690    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1691        removed. (These were left over from very, very early versions of PCRE.)
1692    
1693    7.  Some cosmetic changes to the code to make life easier when compiling it
1694        as part of something else:
1695    
1696        (a) Change DEBUG to PCRE_DEBUG.
1697    
1698        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1699            called "current" as "current_branch", to prevent a collision with the
1700            Linux macro when compiled as a kernel module.
1701    
1702        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1703            prevent a collision with the Linux macro when compiled as a kernel
1704            module.
1705    
1706    8.  In pcre_compile() there are some checks for integer overflows that used to
1707        cast potentially large values to (double). This has been changed to that
1708        when building, a check for int64_t is made, and if it is found, it is used
1709        instead, thus avoiding the use of floating point arithmetic. (There is no
1710        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1711        double.
1712    
1713    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1714        2005 (difference between two addresses compared to an unsigned value).
1715    
1716    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1717        custom one, because of the following reported problem in Windows:
1718    
1719          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1720              under Win32.
1721          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1722              therefore missing the function definition.
1723          - The compiler thus generates a "C" signature for the test function.
1724          - The linker fails to find the "C" function.
1725          - PCRE fails to configure if asked to do so against libbz2.
1726    
1727    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1728        messages were output:
1729    
1730          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1731          rerunning libtoolize, to keep the correct libtool macros in-tree.
1732          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1733    
1734        I have done both of these things.
1735    
1736    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1737        most of the time, it *can* run out if it is given a pattern that contains a
1738        runaway infinite recursion. I updated the discussion in the pcrestack man
1739        page.
1740    
1741    13. Now that we have gone to the x.xx style of version numbers, the minor
1742        version may start with zero. Using 08 or 09 is a bad idea because users
1743        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1744        interpreted as invalid octal numbers. I've updated the previous comment in
1745        configure.ac, and also added a check that gives an error if 08 or 09 are
1746        used.
1747    
1748    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1749        causing partial matching to fail when the end of the subject matched \W
1750        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1751    
1752    15. There were some discrepancies between the declarations in pcre_internal.h
1753        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1754        their definitions. The declarations used "const uschar *" and the
1755        definitions used USPTR. Even though USPTR is normally defined as "const
1756        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1757        reported that: "This difference in casting confuses some C++ compilers, for
1758        example, SunCC recognizes above declarations as different functions and
1759        generates broken code for hbpcre." I have changed the declarations to use
1760        USPTR.
1761    
1762    16. GNU libtool is named differently on some systems. The autogen.sh script now
1763        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1764        (FreeBSD).
1765    
1766    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1767        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1768        comment: "Figure out how to create a longlong from a string: strtoll and
1769        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1770        instance, but it only takes 2 args instead of 3!"
1771    
1772    18. A subtle bug concerned with back references has been fixed by a change of
1773        specification, with a corresponding code fix. A pattern such as
1774        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1775        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1776        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1777        same bug. Such groups have to be quantified to be useful, or contained
1778        inside another quantified group. (If there's no repetition, the reference
1779        can never match.) The problem arises because, having left the group and
1780        moved on to the rest of the pattern, a later failure that backtracks into
1781        the group uses the captured value from the final iteration of the group
1782        rather than the correct earlier one. I have fixed this in PCRE by forcing
1783        any group that contains a reference to itself to be an atomic group; that
1784        is, there cannot be any backtracking into it once it has completed. This is
1785        similar to recursive and subroutine calls.
1786    
1787    
1788    Version 8.00 19-Oct-09
1789    ----------------------
1790    
1791    1.  The table for translating pcre_compile() error codes into POSIX error codes
1792        was out-of-date, and there was no check on the pcre_compile() error code
1793        being within the table. This could lead to an OK return being given in
1794        error.
1795    
1796    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1797        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1798        in a Windows environment.
1799    
1800    3.  The pcregrep --count option prints the count for each file even when it is
1801        zero, as does GNU grep. However, pcregrep was also printing all files when
1802        --files-with-matches was added. Now, when both options are given, it prints
1803        counts only for those files that have at least one match. (GNU grep just
1804        prints the file name in this circumstance, but including the count seems
1805        more useful - otherwise, why use --count?) Also ensured that the
1806        combination -clh just lists non-zero counts, with no names.
1807    
1808    4.  The long form of the pcregrep -F option was incorrectly implemented as
1809        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1810        but it seems right to fix it, and I didn't think it was worth preserving
1811        the old behaviour.
1812    
1813    5.  The command line items --regex=pattern and --regexp=pattern were not
1814        recognized by pcregrep, which required --regex pattern or --regexp pattern
1815        (with a space rather than an '='). The man page documented the '=' forms,
1816        which are compatible with GNU grep; these now work.
1817    
1818    6.  No libpcreposix.pc file was created for pkg-config; there was just
1819        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1820    
1821    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1822        when UCP support is not needed, by modifying the Python script that
1823        generates it from Unicode data files. This should not matter if the module
1824        is correctly used as a library, but I received one complaint about 50K of
1825        unwanted data. My guess is that the person linked everything into his
1826        program rather than using a library. Anyway, it does no harm.
1827    
1828    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1829        was a minimum greater than 1 for a wide character in a possessive
1830        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1831        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1832        character. Chaos in the form of incorrect output or a compiling loop could
1833        result.
1834    
1835    9.  The restrictions on what a pattern can contain when partial matching is
1836        requested for pcre_exec() have been removed. All patterns can now be
1837        partially matched by this function. In addition, if there are at least two
1838        slots in the offset vector, the offset of the earliest inspected character
1839        for the match and the offset of the end of the subject are set in them when
1840        PCRE_ERROR_PARTIAL is returned.
1841    
1842    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1843        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1844        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1845        and may be more useful for multi-segment matching.
1846    
1847    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1848        used to be given if ever the end of the subject was reached; now it is
1849        given only if matching could not proceed because another character was
1850        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1851        string "Z", which now yields "no match" instead of "partial match". In the
1852        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1853        final character ended with (*FAIL).
1854    
1855    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1856        if the pattern had a "must contain" character that was already found in the
1857        earlier partial match, unless partial matching was again requested. For
1858        example, with the pattern /dog.(body)?/, the "must contain" character is
1859        "g". If the first part-match was for the string "dog", restarting with
1860        "sbody" failed. This bug has been fixed.
1861    
1862    13. The string returned by pcre_dfa_exec() after a partial match has been
1863        changed so that it starts at the first inspected character rather than the
1864        first character of the match. This makes a difference only if the pattern
1865        starts with a lookbehind assertion or \b or \B (\K is not supported by
1866        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1867        matching functions compatible, and I think it's the right thing to do.
1868    
1869    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1870        so that the demonstration program is easily available in environments where
1871        PCRE has not been installed from source.
1872    
1873    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1874        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1875        library.
1876    
1877    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1878        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1879        is not the first non-POSIX option to be added. Clearly some people find
1880        these options useful.
1881    
1882    17. If a caller to the POSIX matching function regexec() passes a non-zero
1883        value for nmatch with a NULL value for pmatch, the value of
1884        nmatch is forced to zero.
1885    
1886    18. RunGrepTest did not have a test for the availability of the -u option of
1887        the diff command, as RunTest does. It now checks in the same way as
1888        RunTest, and also checks for the -b option.
1889    
1890    19. If an odd number of negated classes containing just a single character
1891        interposed, within parentheses, between a forward reference to a named
1892        subpattern and the definition of the subpattern, compilation crashed with
1893        an internal error, complaining that it could not find the referenced
1894        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1895        [The bug was that it was starting one character too far in when skipping
1896        over the character class, thus treating the ] as data rather than
1897        terminating the class. This meant it could skip too much.]
1898    
1899    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1900        /g option in pcretest when the pattern contains \K, which makes it possible
1901        to have an empty string match not at the start, even when the pattern is
1902        anchored. Updated pcretest and pcredemo to use this option.
1903    
1904    21. If the maximum number of capturing subpatterns in a recursion was greater
1905        than the maximum at the outer level, the higher number was returned, but
1906        with unset values at the outer level. The correct (outer level) value is
1907        now given.
1908    
1909    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1910        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1911        make it do so. The string so far is captured, making this feature
1912        compatible with Perl.
1913    
1914    23. The tests have been re-organized, adding tests 11 and 12, to make it
1915        possible to check the Perl 5.10 features against Perl 5.10.
1916    
1917    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1918        pattern matches a fixed length string. PCRE did not allow this; now it
1919        does. Neither allows recursion.
1920    
1921    25. I finally figured out how to implement a request to provide the minimum
1922        length of subject string that was needed in order to match a given pattern.
1923        (It was back references and recursion that I had previously got hung up
1924        on.) This code has now been added to pcre_study(); it finds a lower bound
1925        to the length of subject needed. It is not necessarily the greatest lower
1926        bound, but using it to avoid searching strings that are too short does give
1927        some useful speed-ups. The value is available to calling programs via
1928        pcre_fullinfo().
1929    
1930    26. While implementing 25, I discovered to my embarrassment that pcretest had
1931        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1932        study optimizations had never been tested with that matching function.
1933        Oops. What is worse, even when it was passed study data, there was a bug in
1934        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1935        were also very few tests of studied patterns with pcre_dfa_exec().
1936    
1937    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1938        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1939        on the other side of the coin, they are no longer allowed to have different
1940        names, because these cannot be distinguished in PCRE, and this has caused
1941        confusion. (This is a difference from Perl.)
1942    
1943    28. When duplicate subpattern names are present (necessarily with different
1944        numbers, as required by 27 above), and a test is made by name in a
1945        conditional pattern, either for a subpattern having been matched, or for
1946        recursion in such a pattern, all the associated numbered subpatterns are
1947        tested, and the overall condition is true if the condition is true for any
1948        one of them. This is the way Perl works, and is also more like the way
1949        testing by number works.
1950    
1951    
1952    Version 7.9 11-Apr-09
1953    ---------------------
1954    
1955    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1956        (pcretest), all targets were linked against these libraries. This included
1957        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1958        libraries. This caused unwanted dependencies to be created. This problem
1959        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1960        pcretest is linked with readline.
1961    
1962    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1963        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1964        moved outside it again, because FALSE and TRUE are already defined in AIX,
1965        but BOOL is not.
1966    
1967    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1968        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1969    
1970    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1971        hyphens) following file names and line numbers when outputting matching
1972        lines. This is not true; no spaces are inserted. I have also clarified the
1973        wording for the --colour (or --color) option.
1974    
1975    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1976        was not coloured; this is different to GNU grep, so I have changed it to be
1977        the same.
1978    
1979    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1980        each matching line was coloured. Now it goes on to look for further matches
1981        of any of the test patterns, which is the same behaviour as GNU grep.
1982    
1983    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1984        doesn't make sense to accept an empty string match in pcregrep, so I have
1985        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1986        seems to be how GNU grep behaves. [But see later change 40 for release
1987        8.33.]
1988    
1989    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1990        start or after a newline", because the conditional assertion was not being
1991        correctly handled. The rule now is that both the assertion and what follows
1992        in the first alternative must satisfy the test.
1993    
1994    9.  If auto-callout was enabled in a pattern with a conditional group whose
1995        condition was an assertion, PCRE could crash during matching, both with
1996        pcre_exec() and pcre_dfa_exec().
1997    
1998    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1999        used for matching.
2000    
2001    11. Unicode property support in character classes was not working for
2002        characters (bytes) greater than 127 when not in UTF-8 mode.
2003    
2004    12. Added the -M command line option to pcretest.
2005    
2006    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2007    
2008    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2009    
2010    16. Added comments and documentation about mis-use of no_arg in the C++
2011        wrapper.
2012    
2013    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2014        from Martin Jerabek that uses macro names for all relevant character and
2015        string constants.
2016    
2017    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2018        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2019        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2020        these, but not everybody uses configure.
2021    
2022    19. A conditional group that had only one branch was not being correctly
2023        recognized as an item that could match an empty string. This meant that an
2024        enclosing group might also not be so recognized, causing infinite looping
2025        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2026        with the subject "ab", where knowledge that the repeated group can match
2027        nothing is needed in order to break the loop.
2028    
2029    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2030        exec(), but without supplying a callout function, matching went wrong.
2031    
2032    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2033        leak if the size of the offset vector was greater than 30. When the vector
2034        is smaller, the saved offsets during recursion go onto a local stack
2035        vector, but for larger vectors malloc() is used. It was failing to free
2036        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2037        error, in fact).
2038    
2039    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2040        heapframe that is used only when UTF-8 support is enabled. This caused no
2041        problem, but was untidy.
2042    
2043    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2044        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2045        included within another project.
2046    
2047    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2048        slightly modified by me:
2049    
2050          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2051              not building pcregrep.
2052    
2053          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2054              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2055    
2056    25. Forward references, both numeric and by name, in patterns that made use of
2057        duplicate group numbers, could behave incorrectly or give incorrect errors,
2058        because when scanning forward to find the reference group, PCRE was not
2059        taking into account the duplicate group numbers. A pattern such as
2060        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2061    
2062    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2063        the feature of a custom pointer more persuasive (as requested by a user).
2064    
2065    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2066        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2067        pre-defined.
2068    
2069    28. Added support for (*UTF8) at the start of a pattern.
2070    
2071    29. Arrange for flags added by the "release type" setting in CMake to be shown
2072        in the configuration summary.
2073    
2074    
2075    Version 7.8 05-Sep-08
2076    ---------------------
2077    
2078    1.  Replaced UCP searching code with optimized version as implemented for Ad
2079        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2080        stage table and inline lookup instead of a function, giving speed ups of 2
2081        to 5 times on some simple patterns that I tested. Permission was given to
2082        distribute the MultiStage2.py script that generates the tables (it's not in
2083        the tarball, but is in the Subversion repository).
2084    
2085    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2086        scripts.
2087    
2088    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2089        a group with a zero qualifier. The result of the study could be incorrect,
2090        or the function might crash, depending on the pattern.
2091    
2092    4.  Caseless matching was not working for non-ASCII characters in back
2093        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2094        It now works when Unicode Property Support is available.
2095    
2096    5.  In pcretest, an escape such as \x{de} in the data was always generating
2097        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2098        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2099        truncation.
2100    
2101    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2102    
2103    7.  Added two (int) casts to pcregrep when printing the difference of two
2104        pointers, in case they are 64-bit values.
2105    
2106    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2107        test 2 if it fails.
2108    
2109    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2110        and a #define of that name to empty if it is not externally set. This is to
2111        allow users of MSVC to set it if necessary.
2112    
2113    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2114        the convenience functions in the pcre_get.c source file.
2115    
2116    11. An option change at the start of a pattern that had top-level alternatives
2117        could cause overwriting and/or a crash. This command provoked a crash in
2118        some environments:
2119    
2120          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2121    
2122        This potential security problem was recorded as CVE-2008-2371.
2123    
2124    12. For a pattern where the match had to start at the beginning or immediately
2125        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2126        pcre_dfa_exec() could read past the end of the passed subject if there was
2127        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2128        pcretest so that it places the subject at the end of its malloc-ed buffer.
2129    
2130    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2131        exec() might read past the end of the data buffer in UTF-8 mode.
2132    
2133    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2134        the data contained the byte 0x85 as part of a UTF-8 character within its
2135        first line. This applied both to normal and DFA matching.
2136    
2137    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2138        /^[^d]*?$/8 failed to match "abc".
2139    
2140    16. Added a missing copyright notice to pcrecpp_internal.h.
2141    
2142    17. Make it more clear in the documentation that values returned from
2143        pcre_exec() in ovector are byte offsets, not character counts.
2144    
2145    18. Tidied a few places to stop certain compilers from issuing warnings.
2146    
2147    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2148        supplied by Stefan Weber. I made a further small update for 7.8 because
2149        there is a change of source arrangements: the pcre_searchfuncs.c module is
2150        replaced by pcre_ucd.c.
2151    
2152    
2153    Version 7.7 07-May-08
2154    ---------------------
2155    
2156    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2157        a string to a long long, pretend we don't even have a long long." This is
2158        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2159    
2160    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2161        pre-7.6 versions, which defined a global no_arg variable instead of putting
2162        it in the RE class. (See also #8 below.)
2163    
2164    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2165        Lopes.
2166    
2167    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2168    
2169        (1) The include/exclude patterns were being applied to the whole pathnames
2170            of files, instead of just to the final components.
2171    
2172        (2) If there was more than one level of directory, the subdirectories were
2173            skipped unless they satisfied the include/exclude conditions. This is
2174            inconsistent with GNU grep (and could even be seen as contrary to the
2175            pcregrep specification - which I improved to make it absolutely clear).
2176            The action now is always to scan all levels of directory, and just
2177            apply the include/exclude patterns to regular files.
2178    
2179    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2180        --exclude_dir in the tests to avoid scanning .svn directories.
2181    
2182    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2183        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2184        doesn't support NULs in patterns.
2185    
2186    7.  Added some missing "const"s to declarations of static tables in
2187        pcre_compile.c and pcre_dfa_exec.c.
2188    
2189    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2190        caused by fix #2  above. (Subsequently also a second patch to fix the
2191        first patch. And a third patch - this was a messy problem.)
2192    
2193    9.  Applied Craig's patch to remove the use of push_back().
2194    
2195    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2196        matching function regexec().
2197    
2198    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2199        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2200        references. PCRE supports relative numbers with this syntax (I don't think
2201        Oniguruma does).
2202    
2203    12. Previously, a group with a zero repeat such as (...){0} was completely
2204        omitted from the compiled regex. However, this means that if the group
2205        was called as a subroutine from elsewhere in the pattern, things went wrong
2206        (an internal error was given). Such groups are now left in the compiled
2207        pattern, with a new opcode that causes them to be skipped at execution
2208        time.
2209    
2210    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2211        to the way PCRE behaves:
2212    
2213        (a) A lone ] character is dis-allowed (Perl treats it as data).
2214    
2215        (b) A back reference to an unmatched subpattern matches an empty string
2216            (Perl fails the current match path).
2217    
2218        (c) A data ] in a character class must be notated as \] because if the
2219            first data character in a class is ], it defines an empty class. (In
2220            Perl it is not possible to have an empty class.) The empty class []
2221            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2222            The negative empty class [^] matches any one character, independently
2223            of the DOTALL setting.
2224    
2225    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2226        non-existent subpattern following a character class starting with ']' and
2227        containing () gave an internal compiling error instead of "reference to
2228        non-existent subpattern". Fortunately, when the pattern did exist, the
2229        compiled code was correct. (When scanning forwards to check for the
2230        existence of the subpattern, it was treating the data ']' as terminating
2231        the class, so got the count wrong. When actually compiling, the reference
2232        was subsequently set up correctly.)
2233    
2234    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2235        it was being rejected as not supported by pcre_dfa_exec(), even though
2236        other assertions are supported. I have made pcre_dfa_exec() support
2237        (*FAIL).
2238    
2239    16. The implementation of 13c above involved the invention of a new opcode,
2240        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2241        cannot be changed at match time, I realized I could make a small
2242        improvement to matching performance by compiling OP_ALLANY instead of
2243        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2244        on the OP_ANY path.
2245    
2246    17. Compiling pcretest on Windows with readline support failed without the
2247        following two fixes: (1) Make the unistd.h include conditional on
2248        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2249    
2250    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2251        ncurses library to be included for pcretest when ReadLine support is
2252        requested, but also to allow for it to be overridden. This patch came from
2253        Daniel Bergström.
2254    
2255    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2256        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2257        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2258        spotting this.
2259    
2260    
2261    Version 7.6 28-Jan-08
2262    ---------------------
2263    
2264    1.  A character class containing a very large number of characters with
2265        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2266        overflow.
2267    
2268    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2269        HAVE_LONG_LONG is not defined.
2270    
2271    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2272        bring them up to date and include new features. This patch includes:
2273    
2274        - Fixed PH's badly added libz and libbz2 support.
2275        - Fixed a problem with static linking.
2276        - Added pcredemo. [But later removed - see 7 below.]
2277        - Fixed dftables problem and added an option.
2278        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2279            HAVE_LONG_LONG.
2280        - Added readline support for pcretest.
2281        - Added an listing of the option settings after cmake has run.
2282    
2283    4.  A user submitted a patch to Makefile that makes it easy to create
2284        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2285        Makefile.am that cause it to include this special target, without
2286        affecting anything else. Note that the same mingw target plus all
2287        the other distribution libraries and programs are now supported
2288        when configuring with CMake (see 6 below) instead of with
2289        Configure/Make.
2290    
2291    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2292        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2293        exported in the Windows port". It has not yet been confirmed that the patch
2294        solves the problem, but it does no harm.
2295    
2296    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2297        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2298        with CMake, and also correct the comment about stack recursion.
2299    
2300    7.  Remove the automatic building of pcredemo from the ./configure system and
2301        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2302        of a program that users should build themselves after PCRE is installed, so
2303        building it automatically is not really right. What is more, it gave
2304        trouble in some build environments.
2305    
2306    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2307    
2308    
2309    Version 7.5 10-Jan-08
2310    ---------------------
2311    
2312    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2313        values in parens when parsing an RE using the C++ wrapper."
2314    
2315    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2316        Characters greater than 255 were excluded from the class instead of being
2317        included.
2318    
2319    3.  The same bug as (2) above applied to negated POSIX classes such as
2320        [:^space:].
2321    
2322    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2323        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2324        I have changed it.
2325    
2326    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2327        first named subpattern) and a construct such as (?&a) would reference the
2328        first named subpattern whose name started with "a" (in other words, the
2329        length check was missing). Both these problems are fixed. "Subpattern name
2330        expected" is now given for (?&) (a zero-length name), and this patch also
2331        makes it give the same error for \k'' (previously it complained that that
2332        was a reference to a non-existent subpattern).
2333    
2334    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2335        this is right because (?- can be followed by option settings as well as by
2336        digits. I have, however, made the messages clearer.
2337    
2338    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2339        than the number used in the conditional) now cause a compile-time error.
2340        This is actually not compatible with Perl, which accepts such patterns, but
2341        treats the conditional as always being FALSE (as PCRE used to), but it
2342        seems to me that giving a diagnostic is better.
2343    
2344    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2345        and messages.
2346    
2347    9.  Fix two occurrences of "backslash" in comments that should have been
2348        "backspace".
2349    
2350    10. Remove two redundant lines of code that can never be obeyed (their function
2351        was moved elsewhere).
2352    
2353    11. The program that makes PCRE's Unicode character property table had a bug
2354        which caused it to generate incorrect table entries for sequences of
2355        characters that have the same character type, but are in different scripts.
2356        It amalgamated them into a single range, with the script of the first of
2357        them. In other words, some characters were in the wrong script. There were
2358        thirteen such cases, affecting characters in the following ranges:
2359    
2360          U+002b0 - U+002c1
2361          U+0060c - U+0060d
2362          U+0061e - U+00612
2363          U+0064b - U+0065e
2364          U+0074d - U+0076d
2365          U+01800 - U+01805
2366          U+01d00 - U+01d77
2367          U+01d9b - U+01dbf
2368          U+0200b - U+0200f
2369          U+030fc - U+030fe
2370          U+03260 - U+0327f
2371          U+0fb46 - U+0fbb1
2372          U+10450 - U+1049d
2373    
2374    12. The -o option (show only the matching part of a line) for pcregrep was not
2375        compatible with GNU grep in that, if there was more than one match in a
2376        line, it showed only the first of them. It now behaves in the same way as
2377        GNU grep.
2378    
2379    13. If the -o and -v options were combined for pcregrep, it printed a blank
2380        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2381        does the same. The return code can be used to tell if there were any
2382        non-matching lines.
2383    
2384    14. Added --file-offsets and --line-offsets to pcregrep.
2385    
2386    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2387        infinitely looping recursion. The bug was that positive lookaheads were not
2388        being skipped when checking for a possible empty match (negative lookaheads
2389        and both kinds of lookbehind were skipped).
2390    
2391    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2392        inclusion of <windows.h> to before rather than after the definition of
2393        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2394    
2395    17. Specifying a possessive quantifier with a specific limit for a Unicode
2396        character property caused pcre_compile() to compile bad code, which led at
2397        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2398        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2399        caused the error; without that there was no problem.
2400    
2401    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2402    
2403    19. Added --enable-pcretest-libreadline.
2404    
2405    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2406        RE::GlobalReplace(). As a result, the number of replacements returned was
2407        double what it should be. I removed one of the increments, but Craig sent a
2408        later patch that removed the other one (the right fix) and added unit tests
2409        that check the return values (which was not done before).
2410    
2411    21. Several CMake things:
2412    
2413        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2414            the names libpcre and libpcreposix, not just pcre and pcreposix.
2415    
2416        (2) The above change means that pcretest and pcregrep are now correctly
2417            linked with the newly-built libraries, not previously installed ones.
2418    
2419        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2420    
2421    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2422        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2423        UTF-8 newline character). The key issue is that the pattern starts .*;
2424        this means that the match must be either at the beginning, or after a
2425        newline. The bug was in the code for advancing after a failed match and
2426        checking that the new position followed a newline. It was not taking
2427        account of UTF-8 characters correctly.
2428    
2429    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2430        character classes. PCRE was not treating the sequence [:...:] as a
2431        character class unless the ... were all letters. Perl, however, seems to
2432        allow any characters between [: and :], though of course it rejects as
2433        unknown any "names" that contain non-letters, because all the known class
2434        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2435        for example, whereas PCRE did not - it did not recognize a POSIX character
2436        class. This seemed a bit dangerous, so the code has been changed to be
2437        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2438        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2439        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2440        Perl does, and where it didn't before.
2441    
2442    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2443        Windows environments %n is disabled by default.
2444    
2445    
2446    Version 7.4 21-Sep-07
2447    ---------------------
2448    
2449    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2450        means that a class such as [\s] counted as "explicit reference to CR or
2451        LF". That isn't really right - the whole point of the change was to try to
2452        help when there was an actual mention of one of the two characters. So now
2453        the change happens only if \r or \n (or a literal CR or LF) character is
2454        encountered.
2455    
2456    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2457        of both had grown to the point where there were only 3 bits left.
2458        Fortunately, there was spare space in the data structure, and so I have
2459        moved the internal flags into a new 16-bit field to free up more option
2460        bits.
2461    
2462    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2463        but did not set the internal JCHANGED flag - either of these is enough to
2464        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2465        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2466        start sets both bits.
2467    
2468    4.  Added options (at build time, compile time, exec time) to change \R from
2469        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2470    
2471    5.  doc/pcresyntax.html was missing from the distribution.
2472    
2473    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2474        compatibility, even though it is no longer used.
2475    
2476    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2477        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2478        windows.h file is present (where different names are used). [This was
2479        reversed later after testing - see 16 below.]
2480    
2481    8.  Changed all #include <config.h> to #include "config.h". There were also
2482        some further <pcre.h> cases that I changed to "pcre.h".
2483    
2484    9.  When pcregrep was used with the --colour option, it missed the line ending
2485        sequence off the lines that it output.
2486    
2487    10. It was pointed out to me that arrays of string pointers cause lots of
2488        relocations when a shared library is dynamically loaded. A technique of
2489        using a single long string with a table of offsets can drastically reduce
2490        these. I have refactored PCRE in four places to do this. The result is
2491        dramatic:
2492    
2493          Originally:                          290
2494          After changing UCP table:            187
2495          After changing error message table:   43
2496          After changing table of "verbs"       36
2497          After changing table of Posix names   22
2498    
2499        Thanks to the folks working on Gregex for glib for this insight.
2500    
2501    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2502        unicode-properties was also set.
2503    
2504    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2505    
2506    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2507        checked only for CRLF.
2508    
2509    14. Added casts to pcretest.c to avoid compiler warnings.
2510    
2511    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2512    
2513    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2514        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2515        entirely. This removes changes made in 7 above.
2516    
2517    17. The CMake files have been updated, and there is now more information about
2518        building with CMake in the NON-UNIX-USE document.
2519    
2520    
2521    Version 7.3 28-Aug-07
2522    ---------------------
2523    
2524     1. In the rejigging of the build system that eventually resulted in 7.1, the
2525        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2526        brackets there is not right, since it causes compilers to look for an
2527        installed pcre.h, not the version that is in the source that is being
2528        compiled (which of course may be different). I have changed it back to:
2529    
2530          #include "pcre.h"
2531    
2532        I have a vague recollection that the change was concerned with compiling in
2533        different directories, but in the new build system, that is taken care of
2534        by the VPATH setting the Makefile.
2535    
2536     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2537        when the subject happened to end in the byte 0x85 (e.g. if the last
2538        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2539        characters but of course it shouldn't be taken as a newline when it is part
2540        of another character. The bug was that, for an unlimited repeat of . in
2541        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2542        characters when looking for a newline.
2543    
2544     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2545    
2546     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2547        in debug output.
2548    
2549     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2550        long printing in the pcrecpp unittest when running under MinGW.
2551    
2552     6. ESC_K was left out of the EBCDIC table.
2553    
2554     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2555        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2556        limit also applies to "virtual nesting" when a pattern is recursive, and in
2557        this case 1000 isn't so big. I have been able to remove this limit at the
2558        expense of backing off one optimization in certain circumstances. Normally,
2559        when pcre_exec() would call its internal match() function recursively and
2560        immediately return the result unconditionally, it uses a "tail recursion"
2561        feature to save stack. However, when a subpattern that can match an empty
2562        string has an unlimited repetition quantifier, it no longer makes this
2563        optimization. That gives it a stack frame in which to save the data for
2564        checking that an empty string has been matched. Previously this was taken
2565        from the 1000-entry workspace that had been reserved. So now there is no
2566        explicit limit, but more stack is used.
2567    
2568     8. Applied Daniel's patches to solve problems with the import/export magic
2569        syntax that is required for Windows, and which was going wrong for the
2570        pcreposix and pcrecpp parts of the library. These were overlooked when this
2571        problem was solved for the main library.
2572    
2573     9. There were some crude static tests to avoid integer overflow when computing
2574        the size of patterns that contain repeated groups with explicit upper
2575        limits. As the maximum quantifier is 65535, the maximum group length was
2576        set at 30,000 so that the product of these two numbers did not overflow a
2577        32-bit integer. However, it turns out that people want to use groups that
2578        are longer than 30,000 bytes (though not repeat them that many times).
2579        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2580        made it possible to implement the integer overflow checks in a much more
2581        dynamic way, which I have now done. The artificial limitation on group
2582        length has been removed - we now have only the limit on the total length of
2583        the compiled pattern, which depends on the LINK_SIZE setting.
2584    
2585    10. Fixed a bug in the documentation for get/copy named substring when
2586        duplicate names are permitted. If none of the named substrings are set, the
2587        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2588        empty string.
2589    
2590    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2591        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2592        because the ] is interpreted as the first data character and the
2593        terminating ] is not found. PCRE has been made compatible with Perl in this
2594        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2595        cause memory overwriting.
2596    
2597    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2598        string has been matched (to stop an infinite loop). It was not recognizing
2599        a conditional subpattern that could match an empty string if that
2600        subpattern was within another subpattern. For example, it looped when
2601        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2602        condition was not nested. This bug has been fixed.
2603    
2604    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2605        past the start of the subject in the presence of bytes with the top bit
2606        set, for example "\x8aBCD".
2607    
2608    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2609        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2610    
2611    14. Optimized (?!) to (*FAIL).
2612    
2613    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2614        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2615        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2616        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2617        does: it's just the validity check that is more restrictive.
2618    
2619    16. Inserted checks for integer overflows during escape sequence (backslash)
2620        processing, and also fixed erroneous offset values for syntax errors during
2621        backslash processing.
2622    
2623    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2624        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2625    
2626    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2627        caused an overrun.
2628    
2629    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2630        something other than just ASCII characters) inside a group that had an
2631        unlimited repeat caused a loop at compile time (while checking to see
2632        whether the group could match an empty string).
2633    
2634    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2635        [\P{Any}] did so. (Error in the code for printing property names.)
2636    
2637    21. An orphan \E inside a character class could cause a crash.
2638    
2639    22. A repeated capturing bracket such as (A)? could cause a wild memory
2640        reference during compilation.
2641    
2642    23. There are several functions in pcre_compile() that scan along a compiled
2643        expression for various reasons (e.g. to see if it's fixed length for look
2644        behind). There were bugs in these functions when a repeated \p or \P was
2645        present in the pattern. These operators have additional parameters compared
2646        with \d, etc, and these were not being taken into account when moving along
2647        the compiled data. Specifically:
2648    
2649        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2650            length.
2651    
2652        (b) An item such as \pL+ within a repeated group could cause crashes or
2653            loops.
2654    
2655        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2656            "reference to non-existent subpattern" error.
2657    
2658        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2659    
2660    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2661        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2662    
2663    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2664    
2665    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2666        character were causing crashes (broken optimization).
2667    
2668    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2669        \p or \P) caused a compile-time loop.
2670    
2671    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2672        break. For example, the unstudied pattern [\r\n]A does not match the string
2673        "\r\nA" because change 7.0/46 below moves the current point on by two
2674        characters after failing to match at the start. However, the pattern \nA
2675        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2676        the same is true. There doesn't seem any very clean way out of this, but
2677        what I have chosen to do makes the common cases work: PCRE now takes note
2678        of whether there can be an explicit match for \r or \n anywhere in the
2679        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2680        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2681        pattern has explicit CR or LF references.
2682    
2683    29. Added (*CR) etc for changing newline setting at start of pattern.
2684    
2685    
2686    Version 7.2 19-Jun-07
2687    ---------------------
2688    
2689     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2690        which is apparently normally available under Windows.
2691    
2692     2. Re-jig the pcregrep tests with different newline settings in an attempt
2693        to make them independent of the local environment's newline setting.
2694    
2695     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2696    
2697     4. Some of the "internals" tests were previously cut out when the link size
2698        was not 2, because the output contained actual offsets. The recent new
2699        "Z" feature of pcretest means that these can be cut out, making the tests
2700        usable with all link sizes.
2701    
2702     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2703        stack recursion. This gives a massive performance boost under BSD, but just
2704        a small improvement under Linux. However, it saves one field in the frame
2705        in all cases.
2706    
2707     6. Added more features from the forthcoming Perl 5.10:
2708    
2709        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2710            recursion call. It refers to the nth most recently opened parentheses.
2711    
2712        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2713            to be opened parentheses.
2714    
2715        (c) Conditions that refer to capturing parentheses can be specified
2716            relatively, for example, (?(-2)... or (?(+3)...
2717    
2718        (d) \K resets the start of the current match so that everything before
2719            is not part of it.
2720    
2721        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2722    
2723        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2724            reference syntax.
2725    
2726        (g) (?| introduces a group in which the numbering of parentheses in each
2727            alternative starts with the same number.
2728    
2729        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2730    
2731     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2732        PCRE_INFO_JCHANGED.
2733    
2734     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2735        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2736        for detecting groups that can match an empty string.
2737    
2738     9. A pattern with a very large number of alternatives (more than several
2739        hundred) was running out of internal workspace during the pre-compile
2740        phase, where pcre_compile() figures out how much memory will be needed. A
2741        bit of new cunning has reduced the workspace needed for groups with
2742        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2743        workspace instead of running out of the 4096 that are available.
2744    
2745    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2746    
2747    11. Applied patch from Google to remove an optimization that didn't quite work.
2748        The report of the bug said:
2749    
2750          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2751          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2752          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2753    
2754    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2755        it matched the wrong number of bytes.
2756    
2757    
2758    Version 7.1 24-Apr-07
2759    ---------------------
2760    
2761     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2762        that is more "standard", making use of automake and other Autotools. There
2763        is some re-arrangement of the files and adjustment of comments consequent
2764        on this.
2765    
2766     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2767        for recursive directory scanning broke on some systems because the files
2768        are not scanned in any specific order and on different systems the order
2769        was different. A call to "sort" has been inserted into RunGrepTest for the
2770        approprate test as a short-term fix. In the longer term there may be an
2771        alternative.
2772    
2773     3. I had an email from Eric Raymond about problems translating some of PCRE's
2774        man pages to HTML (despite the fact that I distribute HTML pages, some
2775        people do their own conversions for various reasons). The problems
2776        concerned the use of low-level troff macros .br and .in. I have therefore
2777        removed all such uses from the man pages (some were redundant, some could
2778        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2779        HTML has been updated to handle .nf/.fi and to complain if it encounters
2780        .br or .in.
2781    
2782     4. Updated comments in configure.ac that get placed in config.h.in and also
2783        arranged for config.h to be included in the distribution, with the name
2784        config.h.generic, for the benefit of those who have to compile without
2785        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2786    
2787     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2788        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2789        makevp.bat for the current PCRE, using the additional files
2790        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2791    
2792     6. A Windows user reported a minor discrepancy with test 2, which turned out
2793        to be caused by a trailing space on an input line that had got lost in his
2794        copy. The trailing space was an accident, so I've just removed it.
2795    
2796     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2797        that is needed.
2798    
2799     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2800        as "const" (a) because they are and (b) because it helps the PHP
2801        maintainers who have recently made a script to detect big data structures
2802        in the php code that should be moved to the .rodata section. I remembered
2803        to update Builducptable as well, so it won't revert if ucptable.h is ever
2804        re-created.
2805    
2806     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2807        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2808        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2809        support is not required. This saves 1.5-2K of code, which is important in
2810        some applications.
2811    
2812        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2813        so as not to refer to the tables, even though these functions will never be
2814        called when UTF-8 support is disabled. Otherwise there are problems with a
2815        shared library.
2816    
2817    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2818    
2819        (a) It was defining its arguments as char * instead of void *.
2820    
2821        (b) It was assuming that all moves were upwards in memory; this was true
2822            a long time ago when I wrote it, but is no longer the case.
2823    
2824        The emulated memove() is provided for those environments that have neither
2825        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2826        is clearly not the case, as these two bugs were recently reported.
2827    
2828    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2829        and Detrail to create the HTML documentation, the .txt form of the man
2830        pages, and it removes trailing spaces from listed files. It also creates
2831        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2832        case, it wraps all the #defines with #ifndefs. This script should be run
2833        before "make dist".
2834    
2835    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2836        with Unicode property support.
2837    
2838        (a) For a maximizing quantifier, if the two different cases of the
2839            character were of different lengths in their UTF-8 codings (there are
2840            some cases like this - I found 11), and the matching function had to
2841            back up over a mixture of the two cases, it incorrectly assumed they
2842            were both the same length.
2843    
2844        (b) When PCRE was configured to use the heap rather than the stack for
2845            recursion during matching, it was not correctly preserving the data for
2846            the other case of a UTF-8 character when checking ahead for a match
2847            while processing a minimizing repeat. If the check also involved
2848            matching a wide character, but failed, corruption could cause an
2849            erroneous result when trying to check for a repeat of the original
2850            character.
2851    
2852    13. Some tidying changes to the testing mechanism:
2853    
2854        (a) The RunTest script now detects the internal link size and whether there
2855            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2856            values substituted by "configure". (The RunGrepTest script already did
2857            this for UTF-8.) The configure.ac script no longer substitutes the
2858            relevant variables.
2859    
2860        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2861            with length and offset values. This means that the output is different
2862            for different internal link sizes. Test 2 is skipped for link sizes
2863            other than 2 because of this, bypassing the problem. Unfortunately,
2864            there was also a test in test 3 (the locale tests) that used /B and
2865            failed for link sizes other than 2. Rather than cut the whole test out,
2866            I have added a new /Z option to pcretest that replaces the length and
2867            offset values with spaces. This is now used to make test 3 independent
2868            of link size. (Test 2 will be tidied up later.)
2869    
2870    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2871        segmentation fault instead of returning the appropriate error message.
2872    
2873    15. In multiline mode when the newline sequence was set to "any", the pattern
2874        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2875        This doesn't seem right; it now treats the CRLF combination as the line
2876        ending, and so does not match in that case. It's only a pattern such as ^$
2877        that would hit this one: something like ^ABC$ would have failed after \r
2878        and then tried again after \r\n.
2879    
2880    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2881        in an attempt to make files that differ only in their line terminators
2882        compare equal. This works on Linux.
2883    
2884    17. Under certain error circumstances pcregrep might try to free random memory
2885        as it exited. This is now fixed, thanks to valgrind.
2886    
2887    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2888        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2889        was because its rules for how to advance for /g after matching an empty
2890        string at the end of a line did not allow for this case. They now check for
2891        it specially.
2892    
2893    20. pcretest is supposed to handle patterns and data of any length, by
2894        extending its buffers when necessary. It was getting this wrong when the
2895        buffer for a data line had to be extended.
2896    
2897    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2898        CRLF as a newline sequence.
2899    
2900    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2901        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2902        I have nevertheless tidied it up.
2903    
2904    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2905    
2906    24. Added a man page for pcre-config.
2907    
2908    
2909    Version 7.0 19-Dec-06
2910    ---------------------
2911    
2912     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2913        moving to gcc 4.1.1.
2914    
2915     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2916        sys/time.h, which is documented as needed for this function. It doesn't
2917        seem to matter on Linux, but it showed up on some releases of OS X.
2918    
2919     3. It seems that there are systems where bytes whose values are greater than
2920        127 match isprint() in the "C" locale. The "C" locale should be the
2921        default when a C program starts up. In most systems, only ASCII printing
2922        characters match isprint(). This difference caused the output from pcretest
2923        to vary, making some of the tests fail. I have changed pcretest so that:
2924    
2925        (a) When it is outputting text in the compiled version of a pattern, bytes
2926            other than 32-126 are always shown as hex escapes.
2927    
2928        (b) When it is outputting text that is a matched part of a subject string,
2929            it does the same, unless a different locale has been set for the match
2930            (using the /L modifier). In this case, it uses isprint() to decide.
2931    
2932     4. Fixed a major bug that caused incorrect computation of the amount of memory
2933        required for a compiled pattern when options that changed within the
2934        pattern affected the logic of the preliminary scan that determines the
2935        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2936        that the computed length was too small. The symptoms of this bug were
2937        either the PCRE error "internal error: code overflow" from pcre_compile(),
2938        or a glibc crash with a message such as "pcretest: free(): invalid next
2939        size (fast)". Examples of patterns that provoked this bug (shown in
2940        pcretest format) are:
2941    
2942          /(?-x: )/x
2943          /(?x)(?-x: \s*#\s*)/
2944          /((?i)[\x{c0}])/8
2945          /(?i:[\x{c0}])/8
2946    
2947        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2948        is now done differently.
2949    
2950     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2951        wrapper classes; (b) implement a new function in the C++ scanner that is
2952        more efficient than the old way of doing things because it avoids levels of
2953        recursion in the regex matching; (c) add a paragraph to the documentation
2954        for the FullMatch() function.
2955    
2956     6. The escape sequence \n was being treated as whatever was defined as
2957        "newline". Not only was this contrary to the documentation, which states
2958        that \n is character 10 (hex 0A), but it also went horribly wrong when
2959        "newline" was defined as CRLF. This has been fixed.
2960    
2961     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2962        was being set to -1 for the "end of line" case (supposedly a value that no
2963        character can have). Though this value is never used (the check for end of
2964        line is "zero bytes in current character"), it caused compiler complaints.
2965        I've changed it to 0xffffffff.
2966    
2967     8. In pcre_version.c, the version string was being built by a sequence of
2968        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2969        string (as it is for production releases) called a macro with an empty
2970        argument. The C standard says the result of this is undefined. The gcc
2971        compiler treats it as an empty string (which was what was wanted) but it is
2972        reported that Visual C gives an error. The source has been hacked around to
2973        avoid this problem.
2974    
2975     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2976        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2977        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2978        of them did).
2979    
2980    10. Originally, pcretest opened its input and output without "b"; then I was
2981        told that "b" was needed in some environments, so it was added for release
2982        5.0 to both the input and output. (It makes no difference on Unix-like
2983        systems.) Later I was told that it is wrong for the input on Windows. I've
2984        now abstracted the modes into two macros, to make it easier to fiddle with
2985        them, and removed "b" from the input mode under Windows.
2986    
2987    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2988    
2989    12. Added -help and --help to pcretest as an official way of being reminded
2990        of the options.
2991    
2992    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2993        and pcrecpp.cc because they annoy compilers at high warning levels.
2994    
2995    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2996    
2997    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2998        scripts are not C programs :-) and which was not noticed because it works
2999        on Linux.
3000    
3001    16. pcretest is supposed to handle any length of pattern and data line (as one
3002        line or as a continued sequence of lines) by extending its input buffer if
3003        necessary. This feature was broken for very long pattern lines, leading to
3004        a string of junk being passed to pcre_compile() if the pattern was longer
3005        than about 50K.
3006    
3007    17. I have done a major re-factoring of the way pcre_compile() computes the
3008        amount of memory needed for a compiled pattern. Previously, there was code
3009        that made a preliminary scan of the pattern in order to do this. That was
3010        OK when PCRE was new, but as the facilities have expanded, it has become
3011        harder and harder to keep it in step with the real compile phase, and there
3012        have been a number of bugs (see for example, 4 above). I have now found a
3013        cunning way of running the real compile function in a "fake" mode that
3014        enables it to compute how much memory it would need, while actually only
3015        ever using a few hundred bytes of working memory and without too many
3016        tests of the mode. This should make future maintenance and development
3017        easier. A side effect of this work is that the limit of 200 on the nesting
3018        depth of parentheses has been removed (though this was never a serious
3019        limitation, I suspect). However, there is a downside: pcre_compile() now
3020        runs more slowly than before (30% or more, depending on the pattern). I
3021        hope this isn't a big issue. There is no effect on runtime performance.
3022    
3023    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3024        newline (only possible for the last line of a file) and it was a
3025        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3026    
3027    19. Added additional timing features to pcretest. (1) The -tm option now times
3028        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3029        separate command line item, by a number that specifies the number of
3030        repeats to use when timing. The default is 50000; this gives better
3031        precision, but takes uncomfortably long for very large patterns.
3032    
3033    20. Extended pcre_study() to be more clever in cases where a branch of a
3034        subpattern has no definite first character. For example, (a*|b*)[cd] would
3035        previously give no result from pcre_study(). Now it recognizes that the
3036        first character must be a, b, c, or d.
3037    
3038    21. There was an incorrect error "recursive call could loop indefinitely" if
3039        a subpattern (or the entire pattern) that was being tested for matching an
3040        empty string contained only one non-empty item after a nested subpattern.
3041        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3042        incorrectly, because the \d was being skipped in the check.
3043    
3044    22. The pcretest program now has a new pattern option /B and a command line
3045        option -b, which is equivalent to adding /B to every pattern. This causes
3046        it to show the compiled bytecode, without the additional information that
3047        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3048        is the same as /B/I).
3049    
3050    23. A new optimization is now able automatically to treat some sequences such
3051        as a*b as a*+b. More specifically, if something simple (such as a character
3052        or a simple class like \d) has an unlimited quantifier, and is followed by
3053        something that cannot possibly match the quantified thing, the quantifier
3054        is automatically "possessified".
3055    
3056    24. A recursive reference to a subpattern whose number was greater than 39
3057        went wrong under certain circumstances in UTF-8 mode. This bug could also
3058        have affected the operation of pcre_study().
3059    
3060    25. Realized that a little bit of performance could be had by replacing
3061        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3062    
3063    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3064    
3065    27. Possessive quantifiers such as a++ were previously implemented by turning
3066        them into atomic groups such as ($>a+). Now they have their own opcodes,
3067        which improves performance. This includes the automatically created ones
3068        from 23 above.
3069    
3070    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3071        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3072        the first matched character to be a colon. This applied both to named and
3073        numbered groups.
3074    
3075    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3076    
3077    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3078        building PCRE on Windows easier, so I have included it in the distribution.
3079    
3080    31. There is now a check in pcretest against a ridiculously large number being
3081        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3082        loop, the loop is abandoned.
3083    
3084    32. Forward references to subpatterns in conditions such as (?(2)...) where
3085        subpattern 2 is defined later cause pcre_compile() to search forwards in
3086        the pattern for the relevant set of parentheses. This search went wrong
3087        when there were unescaped parentheses in a character class, parentheses
3088        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3089    
3090    33. "Subroutine" calls and backreferences were previously restricted to
3091        referencing subpatterns earlier in the regex. This restriction has now
3092        been removed.
3093    
3094    34. Added a number of extra features that are going to be in Perl 5.10. On the
3095        whole, these are just syntactic alternatives for features that PCRE had
3096        previously implemented using the Python syntax or my own invention. The
3097        other formats are all retained for compatibility.
3098    
3099        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3100            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3101            also .NET compatible.
3102    
3103        (b) A recursion or subroutine call to a named group can now be defined as
3104            (?&name) as well as (?P>name).
3105    
3106        (c) A backreference to a named group can now be defined as \k<name> or
3107            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3108            5.10, are also .NET compatible.
3109    
3110        (d) A conditional reference to a named group can now use the syntax
3111            (?(<name>) or (?('name') as well as (?(name).
3112    
3113        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3114            groups (named and numbered) that are never evaluated inline, but can be
3115            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3116            is always false. There may be only one alternative in such a group.
3117    
3118        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3119            as the simple (?(R). The condition is true only if the most recent
3120            recursion is that of the given number or name. It does not search out
3121            through the entire recursion stack.
3122    
3123        (g) The escape \gN or \g{N} has been added, where N is a positive or
3124            negative number, specifying an absolute or relative reference.
3125    
3126    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3127        some "unreachable code" warnings.
3128    
3129    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3130        things, this adds five new scripts.
3131    
3132    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3133        There were also incompatibilities regarding the handling of \Q..\E inside
3134        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3135        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3136    
3137    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3138        matches an empty string, and forcibly breaks the loop. There were bugs in
3139        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3140        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3141        separate and independent bugs (that affected different cases) have been
3142        fixed.
3143    
3144    39. Refactored the code to abolish the use of different opcodes for small
3145        capturing bracket numbers. This is a tidy that I avoided doing when I
3146        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3147        The new approach is not only tidier, it makes it possible to reduce the
3148        memory needed to fix the previous bug (38).
3149    
3150    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3151        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3152        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3153        mode.
3154    
3155    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3156        report.
3157    
3158    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3159        copy construction and assignment in the C++ wrapper.
3160    
3161    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3162        couple of bugs that could have given wrong results in the "--newline=crlf"
3163        case.
3164    
3165    44. Added a number of casts and did some reorganization of signed/unsigned int
3166        variables following suggestions from Dair Grant. Also renamed the variable
3167        "this" as "item" because it is a C++ keyword.
3168    
3169    45. Arranged for dftables to add
3170    
3171          #include "pcre_internal.h"
3172    
3173        to pcre_chartables.c because without it, gcc 4.x may remove the array
3174        definition from the final binary if PCRE is built into a static library and
3175        dead code stripping is activated.
3176    
3177    46. For an unanchored pattern, if a match attempt fails at the start of a
3178        newline sequence, and the newline setting is CRLF or ANY, and the next two
3179        characters are CRLF, advance by two characters instead of one.
3180    
3181    
3182    Version 6.7 04-Jul-06
3183    ---------------------
3184    
3185     1. In order to handle tests when input lines are enormously long, pcretest has
3186        been re-factored so that it automatically extends its buffers when
3187        necessary. The code is crude, but this _is_ just a test program. The
3188        default size has been increased from 32K to 50K.
3189    
3190     2. The code in pcre_study() was using the value of the re argument before
3191        testing it for NULL. (Of course, in any sensible call of the function, it
3192        won't be NULL.)
3193    
3194     3. The memmove() emulation function in pcre_internal.h, which is used on
3195        systems that lack both memmove() and bcopy() - that is, hardly ever -
3196        was missing a "static" storage class specifier.
3197    
3198     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3199        containing an extended class (one that cannot be represented by a bitmap
3200        because it contains high-valued characters or Unicode property items, e.g.
3201        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3202        pattern, but PCRE should not loop if you do not (it no longer does).
3203        [Detail: two cases were found: (a) a repeated subpattern containing an
3204        extended class; (b) a recursive reference to a subpattern that followed a
3205        previous extended class. It wasn't skipping over the extended class
3206        correctly when UTF-8 mode was not set.]
3207    
3208     5. A negated single-character class was not being recognized as fixed-length
3209        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3210        compile error "lookbehind assertion is not fixed length".
3211    
3212     6. The RunPerlTest auxiliary script was showing an unexpected difference
3213        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3214        write a Perl script that can interpret lines of an input file either as
3215        byte characters or as UTF-8, which is what "perltest" was being required to
3216        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3217        can't do is switch easily at run time between having the "use utf8;" pragma
3218        or not. In the end, I fudged it by using the RunPerlTest script to insert
3219        "use utf8;" explicitly for the UTF-8 tests.
3220    
3221     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3222        the end of the subject string, contrary to the documentation and to what
3223        Perl does. This was true of both matching functions. Now it matches only at
3224        the start of the subject and immediately after *internal* newlines.
3225    
3226     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3227        a pointer to an int instead of a pointer to an unsigned long int. This
3228        caused problems on 64-bit systems.
3229    
3230     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3231        instance of the 'standard' template library not being so standard".
3232    
3233    10. There was no check on the number of named subpatterns nor the maximum
3234        length of a subpattern name. The product of these values is used to compute
3235        the size of the memory block for a compiled pattern. By supplying a very
3236        long subpattern name and a large number of named subpatterns, the size
3237        computation could be caused to overflow. This is now prevented by limiting
3238        the length of names to 32 characters, and the number of named subpatterns
3239        to 10,000.
3240    
3241    11. Subpatterns that are repeated with specific counts have to be replicated in
3242        the compiled pattern. The size of memory for this was computed from the
3243        length of the subpattern and the repeat count. The latter is limited to
3244        65535, but there was no limit on the former, meaning that integer overflow
3245        could in principle occur. The compiled length of a repeated subpattern is
3246        now limited to 30,000 bytes in order to prevent this.
3247    
3248    12. Added the optional facility to have named substrings with the same name.
3249    
3250    13. Added the ability to use a named substring as a condition, using the
3251        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3252        are numbers (not recommended). Forward references are permitted.
3253    
3254    14. Added forward references in named backreferences (if you see what I mean).
3255    
3256    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3257        pattern could run off the end of the subject. For example, the pattern
3258        "(?s)(.{1,5})"8 did this with the subject "ab".
3259    
3260    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3261        PCRE_CASELESS was set when matching characters that were quantified with ?
3262        or *.
3263    
3264    17. A character class other than a single negated character that had a minimum
3265        but no maximum quantifier - for example [ab]{6,} - was not handled
3266        correctly by pce_dfa_exec(). It would match only one character.
3267    
3268    18. A valid (though odd) pattern that looked like a POSIX character
3269        class but used an invalid character after [ (for example [[,abc,]]) caused
3270        pcre_compile() to give the error "Failed: internal error: code overflow" or
3271        in some cases to crash with a glibc free() error. This could even happen if
3272        the pattern terminated after [[ but there just happened to be a sequence of
3273        letters, a binary zero, and a closing ] in the memory that followed.
3274    
3275    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3276        over the years. Originally (before any Unicode support), just the bottom 8
3277        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3278        output from "man perlunicode" includes this:
3279    
3280          The regular expression compiler produces polymorphic opcodes.  That
3281          is, the pattern adapts to the data and automatically switches to
3282          the Unicode character scheme when presented with Unicode data--or
3283          instead uses a traditional byte scheme when presented with byte
3284          data.
3285    
3286        Sadly, a wide octal escape does not cause a switch, and in a string with
3287        no other multibyte characters, these octal escapes are treated as before.
3288        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3289        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3290        Unicode string.
3291    
3292        I have not perpetrated such confusion in PCRE. Up till now, it took just
3293        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3294        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3295        translate to the appropriate multibyte character.
3296    
3297    29. Applied some refactoring to reduce the number of warnings from Microsoft
3298        and Borland compilers. This has included removing the fudge introduced
3299        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3300        a warning about an unused variable.
3301    
3302    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3303        characters since release 4.0, because Perl (from release 5.004) does not.
3304        [Or at least, is documented not to: some releases seem to be in conflict
3305        with the documentation.] However, when a pattern was studied with
3306        pcre_study() and all its branches started with \s, PCRE still included VT
3307        as a possible starting character. Of course, this did no harm; it just
3308        caused an unnecessary match attempt.
3309    
3310    22. Removed a now-redundant internal flag bit that recorded the fact that case
3311        dependency changed within the pattern. This was once needed for "required
3312        byte" processing, but is no longer used. This recovers a now-scarce options
3313        bit. Also moved the least significant internal flag bit to the most-
3314        significant bit of the word, which was not previously used (hangover from
3315        the days when it was an int rather than a uint) to free up another bit for
3316        the future.
3317    
3318    23. Added support for CRLF line endings as well as CR and LF. As well as the
3319        default being selectable at build time, it can now be changed at runtime
3320        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3321        specify that it is scanning data with non-default line endings.
3322    
3323    24. Changed the definition of CXXLINK to make it agree with the definition of
3324        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3325    
3326    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3327        recursions. This makes a big different to stack usage for some patterns.
3328    
3329    26. If a subpattern containing a named recursion or subroutine reference such
3330        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3331        the space required for the compiled pattern went wrong and gave too small a
3332        value. Depending on the environment, this could lead to "Failed: internal
3333        error: code overflow at offset 49" or "glibc detected double free or
3334        corruption" errors.
3335    
3336    27. Applied patches from Google (a) to support the new newline modes and (b) to
3337        advance over multibyte UTF-8 characters in GlobalReplace.
3338    
3339    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3340        difference for some implementation of PCRE in some Windows version.
3341    
3342    29. Added some extra testing facilities to pcretest:
3343    
3344        \q<number>   in a data line sets the "match limit" value
3345        \Q<number>   in a data line sets the "match recursion limt" value
3346        -S <number>  sets the stack size, where <number> is in megabytes
3347    
3348        The -S option isn't available for Windows.
3349    
3350    
3351    Version 6.6 06-Feb-06
3352    ---------------------
3353    
3354     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3355        in pcreposix.h. I have copied the definition from pcre.h.
3356    
3357     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3358        because pcre.h is no longer a built file.
3359    
3360     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3361        not normally included in the compiled code.
3362    
3363    
3364    Version 6.5 01-Feb-06
3365    ---------------------
3366    
3367     1. When using the partial match feature with pcre_dfa_exec(), it was not
3368        anchoring the second and subsequent partial matches at the new starting
3369        point. This could lead to incorrect results. For example, with the pattern
3370        /1234/, partially matching against "123" and then "a4" gave a match.
3371    
3372     2. Changes to pcregrep:
3373    
3374        (a) All non-match returns from pcre_exec() were being treated as failures
3375            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3376            error message is output. Some extra information is given for the
3377            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3378            probably the only errors that are likely to be caused by users (by
3379            specifying a regex that has nested indefinite repeats, for instance).
3380            If there are more than 20 of these errors, pcregrep is abandoned.
3381    
3382        (b) A binary zero was treated as data while matching, but terminated the
3383            output line if it was written out. This has been fixed: binary zeroes
3384            are now no different to any other data bytes.
3385    
3386        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3387            used to set a locale for matching. The --locale=xxxx long option has
3388            been added (no short equivalent) to specify a locale explicitly on the
3389            pcregrep command, overriding the environment variables.
3390    
3391        (d) When -B was used with -n, some line numbers in the output were one less
3392            than they should have been.
3393    
3394        (e) Added the -o (--only-matching) option.
3395    
3396        (f) If -A or -C was used with -c (count only), some lines of context were
3397            accidentally printed for the final match.
3398    
3399        (g) Added the -H (--with-filename) option.
3400    
3401        (h) The combination of options -rh failed to suppress file names for files
3402            that were found from directory arguments.
3403    
3404        (i) Added the -D (--devices) and -d (--directories) options.
3405    
3406        (j) Added the -F (--fixed-strings) option.
3407    
3408        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3409    
3410        (l) Added the --colo(u)r option.
3411    
3412        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3413            is not present by default.
3414    
3415     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3416        items such as (?R) or (?1), when the recursion could match a number of
3417        alternatives. If it matched one of the alternatives, but subsequently,
3418        outside the recursion, there was a failure, the code tried to back up into
3419        the recursion. However, because of the way PCRE is implemented, this is not
3420        possible, and the result was an incorrect result from the match.
3421    
3422        In order to prevent this happening, the specification of recursion has
3423        been changed so that all such subpatterns are automatically treated as
3424        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3425    
3426     4. I had overlooked the fact that, in some locales, there are characters for
3427        which isalpha() is true but neither isupper() nor islower() are true. In
3428        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3429        and ordfeminine) are like this. This affected the treatment of \w and \W
3430        when they appeared in character classes, but not when they appeared outside
3431        a character class. The bit map for "word" characters is now created
3432        separately from the results of isalnum() instead of just taking it from the
3433        upper, lower, and digit maps. (Plus the underscore character, of course.)
3434    
3435     5. The above bug also affected the handling of POSIX character classes such as
3436        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3437        permanent tables. Instead, the bit maps for such a class were previously
3438        created as the appropriate unions of the upper, lower, and digit bitmaps.
3439        Now they are created by subtraction from the [[:word:]] class, which has
3440        its own bitmap.
3441    
3442     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3443        It is created by subtracting the vertical space characters (\x09, \x0a,
3444        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3445        subtraction was done in the overall bitmap for a character class, meaning
3446        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3447        be recognized. This bug has been fixed.
3448    
3449     7. Patches from the folks at Google:
3450    
3451          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3452          real life, but is still worth protecting against".
3453    
3454          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3455          regular expressions".
3456    
3457          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3458          have it.
3459    
3460          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3461          "configure" and the latter not, in order to fix a problem somebody had
3462          with compiling the Arg class on HP-UX.
3463    
3464          (e) Improve the error-handling of the C++ wrapper a little bit.
3465    
3466          (f) New tests for checking recursion limiting.
3467    
3468     8. The pcre_memmove() function, which is used only if the environment does not
3469        have a standard memmove() function (and is therefore rarely compiled),
3470        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3471        returning a result (though PCRE never actually uses the result).
3472    
3473     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3474        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3475        returned instead of calling malloc() with an overflowing number that would
3476        most likely cause subsequent chaos.
3477    
3478    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3479    
3480    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3481        with this option is matched, the nmatch and pmatch options of regexec() are
3482        ignored.
3483    
3484    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3485        provided in case anyone wants to the the POSIX interface with UTF-8
3486        strings.
3487    
3488    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3489        C++ linking (needed for some HP-UX environments).
3490    
3491    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3492        (unused parameter) and in the pcre_printint() function (omitted "default"
3493        switch label when the default is to do nothing).
3494    
3495    15. Added some code to make it possible, when PCRE is compiled as a C++
3496        library, to replace subject pointers for pcre_exec() with a smart pointer
3497        class, thus making it possible to process discontinuous strings.
3498    
3499    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3500        much the same function. They were added by different people who were trying
3501        to make PCRE easy to compile on non-Unix systems. It has been suggested
3502        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3503        for compiling on Windows systems. I have therefore replaced it with
3504        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3505        defaults to "extern" for C or "extern C" for C++, which works fine on
3506        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3507        SCOPE with something explicit in config.h. In addition:
3508    
3509        (a) pcreposix.h still had just "extern" instead of either of these macros;
3510            I have replaced it with PCRE_DATA_SCOPE.
3511    
3512        (b) Functions such as _pcre_xclass(), which are internal to the library,
3513            but external in the C sense, all had PCRE_EXPORT in their definitions.
3514            This is apparently wrong for the Windows case, so I have removed it.
3515            (It makes no difference on Unix-like systems.)
3516    
3517    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3518        of recursive calls to match(). This is different to MATCH_LIMIT because
3519        that limits the total number of calls to match(), not all of which increase
3520        the depth of recursion. Limiting the recursion depth limits the amount of
3521        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3522        when PCRE is compiled, and changed at run time. A patch from Google adds
3523        this functionality to the C++ interface.
3524    
3525    18. Changes to the handling of Unicode character properties:
3526    
3527        (a) Updated the table to Unicode 4.1.0.
3528    
3529        (b) Recognize characters that are not in the table as "Cn" (undefined).
3530    
3531        (c) I revised the way the table is implemented to a much improved format