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

Diff of /code/trunk/ChangeLog

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

revision 65 by nigel, Sat Feb 24 21:40:08 2007 UTC revision 1451 by zherczeg, Tue Jan 21 01:49:14 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    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    
100    Version 8.34 15-December-2013
101    -----------------------------
102    
103    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
104        executable memory. Patch inspired by Carsten Klein.
105    
106    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
107        this macro is never tested and has no effect, because the work to support
108        coverage involves only compiling and linking options and special targets in
109        the Makefile. The comment in config.h implied that defining the macro would
110        enable coverage support, which is totally false. There was also support for
111        setting this macro in the CMake files (my fault, I just copied it from
112        configure). SUPPORT_GCOV has now been removed.
113    
114    3.  Make a small performance improvement in strlen16() and strlen32() in
115        pcretest.
116    
117    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
118        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
119        these cases). There was also one in pcretest.c.
120    
121    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
122    
123    6.  In UTF mode, the code for checking whether a group could match an empty
124        string (which is used for indefinitely repeated groups to allow for
125        breaking an infinite loop) was broken when the group contained a repeated
126        negated single-character class with a character that occupied more than one
127        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
128        UTF-8 mode). The effect was undefined: the group might or might not be
129        deemed as matching an empty string, or the program might have crashed.
130    
131    7.  The code for checking whether a group could match an empty string was not
132        recognizing that \h, \H, \v, \V, and \R must match a character.
133    
134    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
135        an empty string. If it can, pcretest shows this in its information output.
136    
137    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
138        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
139        matched by pcre_exec() without using JIT:
140    
141        (a) If the rest of the pattern did not match after a maximal run of
142            grapheme clusters, the code for backing up to try with fewer of them
143            did not always back up over a full grapheme when characters that do not
144            have the modifier quality were involved, e.g. Hangul syllables.
145    
146        (b) If the match point in a subject started with modifier character, and
147            there was no match, the code could incorrectly back up beyond the match
148            point, and potentially beyond the first character in the subject,
149            leading to a segfault or an incorrect match result.
150    
151    10. A conditional group with an assertion condition could lead to PCRE
152        recording an incorrect first data item for a match if no other first data
153        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
154        first data item, and therefore matched "ca" after "c" instead of at the
155        start.
156    
157    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
158        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
159    
160    12. The source of pcregrep now includes z/OS-specific code so that it can be
161        compiled for z/OS as part of the special z/OS distribution.
162    
163    13. Added the -T and -TM options to pcretest.
164    
165    14. The code in pcre_compile.c for creating the table of named capturing groups
166        has been refactored. Instead of creating the table dynamically during the
167        actual compiling pass, the information is remembered during the pre-compile
168        pass (on the stack unless there are more than 20 named groups, in which
169        case malloc() is used) and the whole table is created before the actual
170        compile happens. This has simplified the code (it is now nearly 150 lines
171        shorter) and prepared the way for better handling of references to groups
172        with duplicate names.
173    
174    15. A back reference to a named subpattern when there is more than one of the
175        same name now checks them in the order in which they appear in the pattern.
176        The first one that is set is used for the reference. Previously only the
177        first one was inspected. This change makes PCRE more compatible with Perl.
178    
179    16. Unicode character properties were updated from Unicode 6.3.0.
180    
181    17. The compile-time code for auto-possessification has been refactored, based
182        on a patch by Zoltan Herczeg. It now happens after instead of during
183        compilation. The code is cleaner, and more cases are handled. The option
184        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
185        options in pcretest are provided to set it. It can also be set by
186        (*NO_AUTO_POSSESS) at the start of a pattern.
187    
188    18. The character VT has been added to the default ("C" locale) set of
189        characters that match \s and are generally treated as white space,
190        following this same change in Perl 5.18. There is now no difference between
191        "Perl space" and "POSIX space". Whether VT is treated as white space in
192        other locales depends on the locale.
193    
194    19. The code for checking named groups as conditions, either for being set or
195        for being recursed, has been refactored (this is related to 14 and 15
196        above). Processing unduplicated named groups should now be as fast at
197        numerical groups, and processing duplicated groups should be faster than
198        before.
199    
200    20. Two patches to the CMake build system, by Alexander Barkov:
201    
202          (1) Replace the "source" command by "." in CMakeLists.txt because
203              "source" is a bash-ism.
204    
205          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
206              without these the CMake build does not work on Solaris.
207    
208    21. Perl has changed its handling of \8 and \9. If there is no previously
209        encountered capturing group of those numbers, they are treated as the
210        literal characters 8 and 9 instead of a binary zero followed by the
211        literals. PCRE now does the same.
212    
213    22. Following Perl, added \o{} to specify codepoints in octal, making it
214        possible to specify values greater than 0777 and also making them
215        unambiguous.
216    
217    23. Perl now gives an error for missing closing braces after \x{... instead of
218        treating the string as literal. PCRE now does the same.
219    
220    24. RunTest used to grumble if an inappropriate test was selected explicitly,
221        but just skip it when running all tests. This make it awkward to run ranges
222        of tests when one of them was inappropriate. Now it just skips any
223        inappropriate tests, as it always did when running all tests.
224    
225    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
226        character types such as \d or \w, too many callouts were inserted, and the
227        data that they returned was rubbish.
228    
229    26. In UCP mode, \s was not matching two of the characters that Perl matches,
230        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
231        were matched by \h. The code has now been refactored so that the lists of
232        the horizontal and vertical whitespace characters used for \h and \v (which
233        are defined only in one place) are now also used for \s.
234    
235    27. Add JIT support for the 64 bit TileGX architecture.
236        Patch by Jiong Wang (Tilera Corporation).
237    
238    28. Possessive quantifiers for classes (both explicit and automatically
239        generated) now use special opcodes instead of wrapping in ONCE brackets.
240    
241    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
242        (because it's meaningless), this was not happening when PCRE_CASELESS was
243        set. Not wrong, but inefficient.
244    
245    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
246        properties for \w, \d, etc) is present in a test regex. Otherwise if the
247        test contains no characters greater than 255, Perl doesn't realise it
248        should be using Unicode semantics.
249    
250    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
251        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
252        does in Unicode mode.
253    
254    32. Added the "forbid" facility to pcretest so that putting tests into the
255        wrong test files can sometimes be quickly detected.
256    
257    33. There is now a limit (default 250) on the depth of nesting of parentheses.
258        This limit is imposed to control the amount of system stack used at compile
259        time. It can be changed at build time by --with-parens-nest-limit=xxx or
260        the equivalent in CMake.
261    
262    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
263        errors. Perl warns for these when in warning mode, but PCRE has no facility
264        for giving warnings.
265    
266    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
267        However, this was not working for (?!) because it is optimized to (*FAIL),
268        for which PCRE does not allow quantifiers. The optimization is now disabled
269        when a quantifier follows (?!). I can't see any use for this, but it makes
270        things uniform.
271    
272    36. Perl no longer allows group names to start with digits, so I have made this
273        change also in PCRE. It simplifies the code a bit.
274    
275    37. In extended mode, Perl ignores spaces before a + that indicates a
276        possessive quantifier. PCRE allowed a space before the quantifier, but not
277        before the possessive +. It now does.
278    
279    38. The use of \K (reset reported match start) within a repeated possessive
280        group such as (a\Kb)*+ was not working.
281    
282    40. Document that the same character tables must be used at compile time and
283        run time, and that the facility to pass tables to pcre_exec() and
284        pcre_dfa_exec() is for use only with saved/restored patterns.
285    
286    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
287        features for Builds with MSVC:
288    
289        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
290           builds). The .rc files can be used to set FileDescription and many other
291           attributes.
292    
293        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
294           This allows higher-level build scripts which want .pdb files to avoid
295           hard-coding the exact files needed."
296    
297    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
298        mean "start of word" and "end of word", respectively, as a transition aid.
299    
300    43. A minimizing repeat of a class containing codepoints greater than 255 in
301        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
302        compiled to use the heap for recursion.
303    
304    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
305        is configured.
306    
307    
308    Version 8.33 28-May-2013
309    ------------------------
310    
311    1.  Added 'U' to some constants that are compared to unsigned integers, to
312        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
313        variables that are added to signed variables, to ensure the result is
314        signed and can be negated.
315    
316    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
317        CMake config files.
318    
319    3.  Revise the creation of config.h.generic so that all boolean macros are
320        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
321        overriding via -D on the command line possible.
322    
323    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
324        to unsigned int is reported to make a quite noticeable speed difference in
325        a specific Windows environment. Testing on Linux did also appear to show
326        some benefit (and it is clearly not harmful). Also fixed the definition of
327        Xop which should be unsigned.
328    
329    5.  Related to (4), changing the definition of the intermediate variable cc
330        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
331        improvements.
332    
333    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
334        unnecessary spaces.
335    
336    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
337        and later.
338    
339    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
340    
341    9.  Optimizing fast_forward_start_bits in JIT.
342    
343    10. Adding support for callouts in JIT, and fixing some issues revealed
344        during this work. Namely:
345    
346        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
347    
348        (b) Minimum length was not checked before the matching is started.
349    
350    11. The value of capture_last that is passed to callouts was incorrect in some
351        cases when there was a capture on one path that was subsequently abandoned
352        after a backtrack. Also, the capture_last value is now reset after a
353        recursion, since all captures are also reset in this case.
354    
355    12. The interpreter no longer returns the "too many substrings" error in the
356        case when an overflowing capture is in a branch that is subsequently
357        abandoned after a backtrack.
358    
359    13. In the pathological case when an offset vector of size 2 is used, pcretest
360        now prints out the matched string after a yield of 0 or 1.
361    
362    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
363        Only supported by the JIT compiler at the moment.
364    
365    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
366    
367    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
368        offset of the starting point of the matching process, provided the offsets
369        vector is large enough.
370    
371    17. The \A escape now records a lookbehind value of 1, though its execution
372        does not actually inspect the previous character. This is to ensure that,
373        in partial multi-segment matching, at least one character from the old
374        segment is retained when a new segment is processed. Otherwise, if there
375        are no lookbehinds in the pattern, \A might match incorrectly at the start
376        of a new segment.
377    
378    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
379    
380    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
381        gives some modest performance improvement in 8-bit mode.
382    
383    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
384        be expressed in certain programming languages using Universal Character
385        Names.
386    
387    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
388        which points out that "non characters" are not "characters that may not
389        appear in Unicode strings" but rather "characters that are reserved for
390        internal use and have only local meaning".
391    
392    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
393        there was a conditional group that depended on an assertion, if the
394        assertion was false, the callout that immediately followed the alternation
395        in the condition was skipped when pcre_exec() was used for matching.
396    
397    23. Allow an explicit callout to be inserted before an assertion that is the
398        condition for a conditional group, for compatibility with automatic
399        callouts, which always insert a callout at this point.
400    
401    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
402        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
403    
404    25. (*PRUNE) is now supported by the JIT compiler.
405    
406    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
407    
408    27. Fix the case where there are two or more SKIPs with arguments that may be
409        ignored.
410    
411    28. (*SKIP) is now supported by the JIT compiler.
412    
413    29. (*THEN) is now supported by the JIT compiler.
414    
415    30. Update RunTest with additional test selector options.
416    
417    31. The way PCRE handles backtracking verbs has been changed in two ways.
418    
419        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
420        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
421        backtracking. In some cases this makes it more Perl-compatible, but Perl's
422        rather obscure rules do not always do the same thing.
423    
424        (2) Previously, backtracking verbs were confined within assertions. This is
425        no longer the case for positive assertions, except for (*ACCEPT). Again,
426        this sometimes improves Perl compatibility, and sometimes does not.
427    
428    32. A number of tests that were in test 2 because Perl did things differently
429        have been moved to test 1, because either Perl or PCRE has changed, and
430        these tests are now compatible.
431    
432    32. Backtracking control verbs are now handled in the same way in JIT and
433        interpreter.
434    
435    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
436        contained a forward subroutine reference caused a compile error.
437    
438    34. Auto-detect and optimize limited repetitions in JIT.
439    
440    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
441        blocking (*UTF) etc.
442    
443    36. In the interpreter, maximizing pattern repetitions for characters and
444        character types now use tail recursion, which reduces stack usage.
445    
446    37. The value of the max lookbehind was not correctly preserved if a compiled
447        and saved regex was reloaded on a host of different endianness.
448    
449    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
450        of the compiled pattern block, expand the flags field from 16 to 32 bits
451        because it was almost full.
452    
453    39. Try madvise first before posix_madvise.
454    
455    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
456        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
457        original change locked out all matches of empty strings. This has been
458        changed so that one match of an empty string per line is recognized.
459        Subsequent searches on the same line (for colouring or for --only-matching,
460        for example) do not recognize empty strings.
461    
462    41. Applied a user patch to fix a number of spelling mistakes in comments.
463    
464    42. Data lines longer than 65536 caused pcretest to crash.
465    
466    43. Clarified the data type for length and startoffset arguments for pcre_exec
467        and pcre_dfa_exec in the function-specific man pages, where they were
468        explicitly stated to be in bytes, never having been updated. I also added
469        some clarification to the pcreapi man page.
470    
471    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
472        a segmentation fault.
473    
474    
475    Version 8.32 30-November-2012
476    -----------------------------
477    
478    1.  Improved JIT compiler optimizations for first character search and single
479        character iterators.
480    
481    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
482        Patch by Daniel Richard G.
483    
484    3.  Single character iterator optimizations in the JIT compiler.
485    
486    4.  Improved JIT compiler optimizations for character ranges.
487    
488    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
489        Reported by Giuseppe D'Angelo.
490    
491    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
492        of a line, was being set incorrectly in cases where .* appeared inside
493        atomic brackets at the start of a pattern, or where there was a subsequent
494        *PRUNE or *SKIP.
495    
496    7.  Improved instruction cache flush for POWER/PowerPC.
497        Patch by Daniel Richard G.
498    
499    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
500        grep:
501    
502        (a) There is now no limit to the number of patterns to be matched.
503    
504        (b) An error is given if a pattern is too long.
505    
506        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
507            are now supported.
508    
509        (d) --exclude-from and --include-from (multiple use) have been added.
510    
511        (e) Exclusions and inclusions now apply to all files and directories, not
512            just to those obtained from scanning a directory recursively.
513    
514        (f) Multiple uses of -f and --file-list are now supported.
515    
516        (g) In a Windows environment, the default for -d has been changed from
517            "read" (the GNU grep default) to "skip", because otherwise the presence
518            of a directory in the file list provokes an error.
519    
520        (h) The documentation has been revised and clarified in places.
521    
522    9.  Improve the matching speed of capturing brackets.
523    
524    10. Changed the meaning of \X so that it now matches a Unicode extended
525        grapheme cluster.
526    
527    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
528        out POSIX threads when JIT support is configured.
529    
530    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
531    
532    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
533        structure could go wrong in environments where size_t is not the same size
534        as int.
535    
536    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
537        set.
538    
539    15. The EBCDIC support had decayed; later updates to the code had included
540        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
541        general tidy up of EBCDIC-related issues, and the documentation was also
542        not quite right. There is now a test that can be run on ASCII systems to
543        check some of the EBCDIC-related things (but is it not a full test).
544    
545    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
546        in a small tidy to the code.
547    
548    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
549    
550    18. If the --only-matching (-o) option in pcregrep is specified multiple
551        times, each one causes appropriate output. For example, -o1 -o2 outputs the
552        substrings matched by the 1st and 2nd capturing parentheses. A separating
553        string can be specified by --om-separator (default empty).
554    
555    19. Improving the first n character searches.
556    
557    20. Turn case lists for horizontal and vertical white space into macros so that
558        they are defined only once.
559    
560    21. This set of changes together give more compatible Unicode case-folding
561        behaviour for characters that have more than one other case when UCP
562        support is available.
563    
564        (a) The Unicode property table now has offsets into a new table of sets of
565            three or more characters that are case-equivalent. The MultiStage2.py
566            script that generates these tables (the pcre_ucd.c file) now scans
567            CaseFolding.txt instead of UnicodeData.txt for character case
568            information.
569    
570        (b) The code for adding characters or ranges of characters to a character
571            class has been abstracted into a generalized function that also handles
572            case-independence. In UTF-mode with UCP support, this uses the new data
573            to handle characters with more than one other case.
574    
575        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
576            whose other case is greater than 256 are now correctly matched
577            caselessly. Previously, the high codepoint matched the low one, but not
578            vice versa.
579    
580        (d) The processing of \h, \H, \v, and \ in character classes now makes use
581            of the new class addition function, using character lists defined as
582            macros alongside the case definitions of 20 above.
583    
584        (e) Caseless back references now work with characters that have more than
585            one other case.
586    
587        (f) General caseless matching of characters with more than one other case
588            is supported.
589    
590    22. Unicode character properties were updated from Unicode 6.2.0
591    
592    23. Improved CMake support under Windows. Patch by Daniel Richard G.
593    
594    24. Add support for 32-bit character strings, and UTF-32
595    
596    25. Major JIT compiler update (code refactoring and bugfixing).
597        Experimental Sparc 32 support is added.
598    
599    26. Applied a modified version of Daniel Richard G's patch to create
600        pcre.h.generic and config.h.generic by "make" instead of in the
601        PrepareRelease script.
602    
603    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
604        pcre_compile.c when checking for a zero character.
605    
606    28. Introducing a native interface for JIT. Through this interface, the compiled
607        machine code can be directly executed. The purpose of this interface is to
608        provide fast pattern matching, so several sanity checks are not performed.
609        However, feature tests are still performed. The new interface provides
610        1.4x speedup compared to the old one.
611    
612    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
613        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
614        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
615    
616    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
617        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
618        the "old" RFC 2279). Instead, it ended up passing a negative length to
619        pcre_exec().
620    
621    31. Add support for GCC's visibility feature to hide internal functions.
622    
623    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
624        "unknown -C option" after outputting 0 or 1.
625    
626    33. There is now support for generating a code coverage report for the test
627        suite in environments where gcc is the compiler and lcov is installed. This
628        is mainly for the benefit of the developers.
629    
630    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
631        unaddressable using valgrind annotations, allowing valgrind to detect
632        invalid memory accesses. This is mainly for the benefit of the developers.
633    
634    25. (*UTF) can now be used to start a pattern in any of the three libraries.
635    
636    26. Give configure error if --enable-cpp but no C++ compiler found.
637    
638    
639    Version 8.31 06-July-2012
640    -------------------------
641    
642    1.  Fixing a wrong JIT test case and some compiler warnings.
643    
644    2.  Removed a bashism from the RunTest script.
645    
646    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
647        to unsigned type, result still unsigned" that was given by an MS compiler
648        on encountering the code "-sizeof(xxx)".
649    
650    4.  Partial matching support is added to the JIT compiler.
651    
652    5.  Fixed several bugs concerned with partial matching of items that consist
653        of more than one character:
654    
655        (a) /^(..)\1/ did not partially match "aba" because checking references was
656            done on an "all or nothing" basis. This also applied to repeated
657            references.
658    
659        (b) \R did not give a hard partial match if \r was found at the end of the
660            subject.
661    
662        (c) \X did not give a hard partial match after matching one or more
663            characters at the end of the subject.
664    
665        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
666            a partial match for the string "\r".
667    
668        (e) When newline was set to CRLF, the metacharacter "." did not recognize
669            a partial match for a CR character at the end of the subject string.
670    
671    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
672        running pcretest, the text "(JIT)" added to the output whenever JIT is
673        actually used to run the match.
674    
675    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
676        or /S+[+] with a digit between 1 and 7.
677    
678    8.  OP_NOT now supports any UTF character not just single-byte ones.
679    
680    9.  (*MARK) control verb is now supported by the JIT compiler.
681    
682    10. The command "./RunTest list" lists the available tests without actually
683        running any of them. (Because I keep forgetting what they all are.)
684    
685    11. Add PCRE_INFO_MAXLOOKBEHIND.
686    
687    12. Applied a (slightly modified) user-supplied patch that improves performance
688        when the heap is used for recursion (compiled with --disable-stack-for-
689        recursion). Instead of malloc and free for each heap frame each time a
690        logical recursion happens, frames are retained on a chain and re-used where
691        possible. This sometimes gives as much as 30% improvement.
692    
693    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
694        call.
695    
696    14. As documented, (*COMMIT) is now confined to within a positive assertion.
697    
698    15. It is now possible to link pcretest with libedit as an alternative to
699        libreadline.
700    
701    16. (*COMMIT) control verb is now supported by the JIT compiler.
702    
703    17. The Unicode data tables have been updated to Unicode 6.1.0.
704    
705    18. Added --file-list option to pcregrep.
706    
707    19. Added binary file support to pcregrep, including the -a, --binary-files,
708        -I, and --text options.
709    
710    20. The madvise function is renamed for posix_madvise for QNX compatibility
711        reasons. Fixed by Giuseppe D'Angelo.
712    
713    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
714    
715    22. Changed the option for creating symbolic links for 16-bit man pages from
716        -s to -sf so that re-installing does not cause issues.
717    
718    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
719    
720    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
721        matches in certain environments (the workspace was not being correctly
722        retained). Also added to pcre_dfa_exec() a simple plausibility check on
723        some of the workspace data at the beginning of a restart.
724    
725    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
726        was not doing so when it should - probably a typo introduced by SVN 528
727        (change 8.10/14).
728    
729    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
730        \w+ when the character tables indicated that \x{c4} was a word character.
731        There were several related cases, all because the tests for doing a table
732        lookup were testing for characters less than 127 instead of 255.
733    
734    27. If a pattern contains capturing parentheses that are not used in a match,
735        their slots in the ovector are set to -1. For those that are higher than
736        any matched groups, this happens at the end of processing. In the case when
737        there were back references that the ovector was too small to contain
738        (causing temporary malloc'd memory to be used during matching), and the
739        highest capturing number was not used, memory off the end of the ovector
740        was incorrectly being set to -1. (It was using the size of the temporary
741        memory instead of the true size.)
742    
743    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
744        ovector size, it uses memory at the end of the block that it has got.
745    
746    29. Check for an overlong MARK name and give an error at compile time. The
747        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
748    
749    30. JIT compiler update.
750    
751    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
752        Rill for the patch.
753    
754    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
755    
756    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
757    
758    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
759        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
760        was enabled.
761    
762    35. Improve JIT code generation for greedy plus quantifier.
763    
764    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
765        1 to "aa" instead of to an empty string. The bug affected repeated groups
766        that could potentially match an empty string.
767    
768    37. Optimizing single character iterators in JIT.
769    
770    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
771        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
772        codepoints that are too big for the mode are faulted, and in a UTF mode,
773        disallowed codepoints are also faulted.
774    
775    39. If PCRE was compiled with UTF support, in three places in the DFA
776        matcher there was code that should only have been obeyed in UTF mode, but
777        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
778        processing when bytes with values greater than 127 were present. In 16-bit
779        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
780        both cases the values are those that cannot be the first data item in a UTF
781        character. The three items that might have provoked this were recursions,
782        possessively repeated groups, and atomic groups.
783    
784    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
785        and pcregrep, because some OS require shared objects to be explicitly
786        passed to ld, causing the link step to fail if they are not.
787    
788    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
789        mode, patterns that started with \h* or \R* might be incorrectly matched.
790    
791    
792    Version 8.30 04-February-2012
793    -----------------------------
794    
795    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
796        name is defined in ctype.h.
797    
798    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
799        only in quite long subpatterns.
800    
801    3.  Removed the function pcre_info(), which has been obsolete and deprecated
802        since it was replaced by pcre_fullinfo() in February 2000.
803    
804    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
805        match a (*MARK), and the match failed at the start of the subject, a
806        reference to memory before the start of the subject could occur. This bug
807        was introduced by fix 17 of release 8.21.
808    
809    5.  A reference to an unset group with zero minimum repetition was giving
810        totally wrong answers (in non-JavaScript-compatibility mode). For example,
811        /(another)?(\1?)test/ matched against "hello world test". This bug was
812        introduced in release 8.13.
813    
814    6.  Add support for 16-bit character strings (a large amount of work involving
815        many changes and refactorings).
816    
817    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
818        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
819        from a file.
820    
821    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
822        rounding is not applied in this particular case).
823    
824    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
825        if they appear, or are escaped, in patterns.
826    
827    10. Get rid of a number of -Wunused-but-set-variable warnings.
828    
829    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
830        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
831        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
832        also returns the mark "x". This bug applied to capturing parentheses,
833        non-capturing parentheses, and atomic parentheses. It also applied to some
834        assertions.
835    
836    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
837        information out of configure.ac instead of relying on pcre.h.generic, which
838        is not stored in the repository.
839    
840    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
841        -lreadline.
842    
843    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
844    
845    15. Applied Graycode's patch to put the top-level frame on the stack rather
846        than the heap when not using the stack for recursion. This gives a
847        performance improvement in many cases when recursion is not deep.
848    
849    16. Experimental code added to "pcretest -C" to output the stack frame size.
850    
851    
852    Version 8.21 12-Dec-2011
853    ------------------------
854    
855    1.  Updating the JIT compiler.
856    
857    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
858        are added as well.
859    
860    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
861        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
862        calling _pcre_jit_exec. Some extra comments are added.
863    
864    4.  (*MARK) settings inside atomic groups that do not contain any capturing
865        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
866        was introduced by change 18 for 8.20.
867    
868    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
869        ECMA-262 standard.
870    
871    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
872        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
873        This bug was probably introduced by change 9 of 8.13.
874    
875    7.  While fixing 6 above, I noticed that a number of other items were being
876        incorrectly rejected as "not fixed length". This arose partly because newer
877        opcodes had not been added to the fixed-length checking code. I have (a)
878        corrected the bug and added tests for these items, and (b) arranged for an
879        error to occur if an unknown opcode is encountered while checking for fixed
880        length instead of just assuming "not fixed length". The items that were
881        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
882        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
883        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
884    
885    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
886        being incorrectly compiled and would have given unpredicatble results.
887    
888    9.  A possessively repeated subpattern with minimum repeat count greater than
889        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
890        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
891        the first (A) could occur when it should not.
892    
893    10. Add a cast and remove a redundant test from the code.
894    
895    11. JIT should use pcre_malloc/pcre_free for allocation.
896    
897    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
898        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
899        is anything other than /usr, -L is still shown).
900    
901    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
902    
903    14. Perl does not support \N without a following name in a [] class; PCRE now
904        also gives an error.
905    
906    15. If a forward reference was repeated with an upper limit of around 2000,
907        it caused the error "internal error: overran compiling workspace". The
908        maximum number of forward references (including repeats) was limited by the
909        internal workspace, and dependent on the LINK_SIZE. The code has been
910        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
911        the default depends on LINK_SIZE. There is a new upper limit (for safety)
912        of around 200,000 forward references. While doing this, I also speeded up
913        the filling in of repeated forward references.
914    
915    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
916        incorrectly expecting the subject to contain another "a" after the start.
917    
918    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
919        in the match, the SKIP should be ignored. This was not happening; instead
920        the SKIP was being treated as NOMATCH. For patterns such as
921        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
922        tested.
923    
924    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
925        now much more compatible with Perl, in particular in cases where the result
926        is a non-match for a non-anchored pattern. For example, if
927        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
928        "m", where previously it did not return a name. A side effect of this
929        change is that for partial matches, the last encountered mark name is
930        returned, as for non matches. A number of tests that were previously not
931        Perl-compatible have been moved into the Perl-compatible test files. The
932        refactoring has had the pleasing side effect of removing one argument from
933        the match() function, thus reducing its stack requirements.
934    
935    19. If the /S+ option was used in pcretest to study a pattern using JIT,
936        subsequent uses of /S (without +) incorrectly behaved like /S+.
937    
938    21. Retrieve executable code size support for the JIT compiler and fixing
939        some warnings.
940    
941    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
942        not work when the shorter character appeared right at the end of the
943        subject string.
944    
945    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
946        systems.
947    
948    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
949        output it when the /M option is used in pcretest.
950    
951    25. The CheckMan script was not being included in the distribution. Also, added
952        an explicit "perl" to run Perl scripts from the PrepareRelease script
953        because this is reportedly needed in Windows.
954    
955    26. If study data was being save in a file and studying had not found a set of
956        "starts with" bytes for the pattern, the data written to the file (though
957        never used) was taken from uninitialized memory and so caused valgrind to
958        complain.
959    
960    27. Updated RunTest.bat as provided by Sheri Pierce.
961    
962    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
963    
964    29. Computation of memory usage for the table of capturing group names was
965        giving an unnecessarily large value.
966    
967    
968    Version 8.20 21-Oct-2011
969    ------------------------
970    
971    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
972        a POSIX class. After further experiments with Perl, which convinced me that
973        Perl has bugs and confusions, a closing square bracket is no longer allowed
974        in a POSIX name. This bug also affected patterns with classes that started
975        with full stops.
976    
977    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
978        captured substring, but while checking the failing first alternative,
979        substring 1 is temporarily captured. If the output vector supplied to
980        pcre_exec() was not big enough for this capture, the yield of the function
981        was still zero ("insufficient space for captured substrings"). This cannot
982        be totally fixed without adding another stack variable, which seems a lot
983        of expense for a edge case. However, I have improved the situation in cases
984        such as /(a)(b)x|abc/ matched against "abc", where the return code
985        indicates that fewer than the maximum number of slots in the ovector have
986        been set.
987    
988    3.  Related to (2) above: when there are more back references in a pattern than
989        slots in the output vector, pcre_exec() uses temporary memory during
990        matching, and copies in the captures as far as possible afterwards. It was
991        using the entire output vector, but this conflicts with the specification
992        that only 2/3 is used for passing back captured substrings. Now it uses
993        only the first 2/3, for compatibility. This is, of course, another edge
994        case.
995    
996    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
997        main code base, and can be used by building with --enable-jit. When this is
998        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
999        runtime --no-jit option is given.
1000    
1001    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1002        ovector, the return from the function was zero, implying that there were
1003        other matches that did not fit. The correct "exactly full" value is now
1004        returned.
1005    
1006    6.  If a subpattern that was called recursively or as a subroutine contained
1007        (*PRUNE) or any other control that caused it to give a non-standard return,
1008        invalid errors such as "Error -26 (nested recursion at the same subject
1009        position)" or even infinite loops could occur.
1010    
1011    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1012        computing the minimum length on reaching *ACCEPT, and so ended up with the
1013        wrong value of 1 rather than 0. Further investigation indicates that
1014        computing a minimum subject length in the presence of *ACCEPT is difficult
1015        (think back references, subroutine calls), and so I have changed the code
1016        so that no minimum is registered for a pattern that contains *ACCEPT.
1017    
1018    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1019        it was not handled as intended. [But see 16 below.]
1020    
1021    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1022        Sheri Pierce.
1023    
1024    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1025        the first byte in a match must be "a".
1026    
1027    11. Change 17 for 8.13 increased the recursion depth for patterns like
1028        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1029        pattern contains any instances of (*THEN). If it does not, the old
1030        optimizations are restored. It would be nice to do this on a per-group
1031        basis, but at the moment that is not feasible.
1032    
1033    12. In some environments, the output of pcretest -C is CRLF terminated. This
1034        broke RunTest's code that checks for the link size. A single white space
1035        character after the value is now allowed for.
1036    
1037    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1038        For "fr", it uses the Windows-specific input and output files.
1039    
1040    14. If (*THEN) appeared in a group that was called recursively or as a
1041        subroutine, it did not work as intended. [But see next item.]
1042    
1043    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1044        pattern fragments (but not containing any | characters). If A and B are
1045        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1046        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1047        D. In other words, Perl considers parentheses that do not contain any |
1048        characters to be part of a surrounding alternative, whereas PCRE was
1049        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1050        differently. PCRE now behaves in the same way as Perl, except in the case
1051        of subroutine/recursion calls such as (?1) which have in any case always
1052        been different (but PCRE had them first :-).
1053    
1054    16. Related to 15 above: Perl does not treat the | in a conditional group as
1055        creating alternatives. Such a group is treated in the same way as an
1056        ordinary group without any | characters when processing (*THEN). PCRE has
1057        been changed to match Perl's behaviour.
1058    
1059    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1060        RunGrepTest script failed.
1061    
1062    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1063        inevitable for groups that contain captures, but it can lead to a lot of
1064        stack use in large patterns. The old behaviour has been restored for atomic
1065        groups that do not contain any capturing parentheses.
1066    
1067    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1068        suppress the check for a minimum subject length at run time. (If it was
1069        given to pcre_exec() or pcre_dfa_exec() it did work.)
1070    
1071    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1072        fail to work when decoding hex characters in data strings in EBCDIC
1073        environments.
1074    
1075    21. It appears that in at least one Mac OS environment, the isxdigit() function
1076        is implemented as a macro that evaluates to its argument more than once,
1077        contravening the C 90 Standard (I haven't checked a later standard). There
1078        was an instance in pcretest which caused it to go wrong when processing
1079        \x{...} escapes in subject strings. The has been rewritten to avoid using
1080        things like p++ in the argument of isxdigit().
1081    
1082    
1083    Version 8.13 16-Aug-2011
1084    ------------------------
1085    
1086    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1087    
1088    2.  Two minor typos in pcre_internal.h have been fixed.
1089    
1090    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1091        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1092        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1093    
1094    4.  There were a number of related bugs in the code for matching backrefences
1095        caselessly in UTF-8 mode when codes for the characters concerned were
1096        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1097        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1098        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1099        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1100        2-byte code at the end of the subject (it thought there wasn't enough data
1101        left).
1102    
1103    5.  Comprehensive information about what went wrong is now returned by
1104        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1105        as the output vector has at least 2 elements. The offset of the start of
1106        the failing character and a reason code are placed in the vector.
1107    
1108    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1109        now returned is for the first byte of the failing character, instead of the
1110        last byte inspected. This is an incompatible change, but I hope it is small
1111        enough not to be a problem. It makes the returned offset consistent with
1112        pcre_exec() and pcre_dfa_exec().
1113    
1114    7.  pcretest now gives a text phrase as well as the error number when
1115        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1116        failure, the offset and reason code are output.
1117    
1118    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1119        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1120        back over a single character (\n). This seems wrong (because it treated the
1121        two characters as a single entity when going forwards), conflicts with the
1122        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1123        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1124        has been changed.
1125    
1126    9.  Some internal refactoring has changed the processing so that the handling
1127        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1128        time (the PCRE_DOTALL option was changed this way some time ago: version
1129        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1130        which was always a bit of a fudge. It also means that there is one less
1131        argument for the match() function, which reduces its stack requirements
1132        slightly. This change also fixes an incompatibility with Perl: the pattern
1133        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1134    
1135    10. More internal refactoring has drastically reduced the number of recursive
1136        calls to match() for possessively repeated groups such as (abc)++ when
1137        using pcre_exec().
1138    
1139    11. While implementing 10, a number of bugs in the handling of groups were
1140        discovered and fixed:
1141    
1142        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1143        (a|)*(?1) gave a compile-time internal error.
1144        ((a|)+)+  did not notice that the outer group could match an empty string.
1145        (^a|^)+   was not marked as anchored.
1146        (.*a|.*)+ was not marked as matching at start or after a newline.
1147    
1148    12. Yet more internal refactoring has removed another argument from the match()
1149        function. Special calls to this function are now indicated by setting a
1150        value in a variable in the "match data" data block.
1151    
1152    13. Be more explicit in pcre_study() instead of relying on "default" for
1153        opcodes that mean there is no starting character; this means that when new
1154        ones are added and accidentally left out of pcre_study(), testing should
1155        pick them up.
1156    
1157    14. The -s option of pcretest has been documented for ages as being an old
1158        synonym of -m (show memory usage). I have changed it to mean "force study
1159        for every regex", that is, assume /S for every regex. This is similar to -i
1160        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1161        using it. It makes it easier to run collections of tests with and without
1162        study enabled, and thereby test pcre_study() more easily. All the standard
1163        tests are now run with and without -s (but some patterns can be marked as
1164        "never study" - see 20 below).
1165    
1166    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1167        restoration of the capturing data to the outer values was not happening
1168        correctly.
1169    
1170    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1171        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1172        pattern had matched an empty string, and so incorrectly returned a no
1173        match.
1174    
1175    17. There was optimizing code for the last branch of non-capturing parentheses,
1176        and also for the obeyed branch of a conditional subexpression, which used
1177        tail recursion to cut down on stack usage. Unfortunately, now that there is
1178        the possibility of (*THEN) occurring in these branches, tail recursion is
1179        no longer possible because the return has to be checked for (*THEN). These
1180        two optimizations have therefore been removed. [But see 8.20/11 above.]
1181    
1182    18. If a pattern containing \R was studied, it was assumed that \R always
1183        matched two bytes, thus causing the minimum subject length to be
1184        incorrectly computed because \R can also match just one byte.
1185    
1186    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1187        was incorrectly computed.
1188    
1189    20. If /S is present twice on a test pattern in pcretest input, it now
1190        *disables* studying, thereby overriding the use of -s on the command line
1191        (see 14 above). This is necessary for one or two tests to keep the output
1192        identical in both cases.
1193    
1194    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1195        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1196    
1197    22. When an atomic group that contained a capturing parenthesis was
1198        successfully matched, but the branch in which it appeared failed, the
1199        capturing was not being forgotten if a higher numbered group was later
1200        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1201        group 1 to "a", when in fact it should be unset. This applied to multi-
1202        branched capturing and non-capturing groups, repeated or not, and also to
1203        positive assertions (capturing in negative assertions does not happen
1204        in PCRE) and also to nested atomic groups.
1205    
1206    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1207        subject after a captured substring, to make it easier to tell which of a
1208        number of identical substrings has been captured.
1209    
1210    24. The way atomic groups are processed by pcre_exec() has been changed so that
1211        if they are repeated, backtracking one repetition now resets captured
1212        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1213        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1214        "aaa". Previously, it would have been "a". As part of this code
1215        refactoring, the way recursive calls are handled has also been changed.
1216    
1217    25. If an assertion condition captured any substrings, they were not passed
1218        back unless some other capturing happened later. For example, if
1219        (?(?=(a))a) was matched against "a", no capturing was returned.
1220    
1221    26. When studying a pattern that contained subroutine calls or assertions,
1222        the code for finding the minimum length of a possible match was handling
1223        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1224        group 1 called group 2 while simultaneously a separate group 2 called group
1225        1). A stack overflow occurred in this case. I have fixed this by limiting
1226        the recursion depth to 10.
1227    
1228    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1229        Fortmann. This supports explicit test numbers on the command line, and has
1230        argument validation and error reporting.
1231    
1232    28. An instance of \X with an unlimited repeat could fail if at any point the
1233        first character it looked at was a mark character.
1234    
1235    29. Some minor code refactoring concerning Unicode properties and scripts
1236        should reduce the stack requirement of match() slightly.
1237    
1238    30. Added the '=' option to pcretest to check the setting of unused capturing
1239        slots at the end of the pattern, which are documented as being -1, but are
1240        not included in the return count.
1241    
1242    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1243        compiled something random. Now it gives a compile-time error (as does
1244        Perl).
1245    
1246    32. A *MARK encountered during the processing of a positive assertion is now
1247        recorded and passed back (compatible with Perl).
1248    
1249    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1250        had alternative anchored branches, the search for a second match in a line
1251        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1252        matched the line "0102" twice. The same bug affected patterns that started
1253        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1254        twice.
1255    
1256    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1257        does, and because of capturing effects, quantifying parenthesized
1258        assertions may at times be useful. Quantifiers are now allowed for
1259        parenthesized assertions.
1260    
1261    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1262    
1263    36. \g was being checked for fancy things in a character class, when it should
1264        just be a literal "g".
1265    
1266    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1267        appearance of a nested POSIX class supersedes an apparent external class.
1268        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1269        unescaped square brackets may also appear as part of class names. For
1270        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1271        more like Perl. (But see 8.20/1 above.)
1272    
1273    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1274        was because it thought it was \N{name}, which is not supported).
1275    
1276    39. Add minix to OS list not supporting the -S option in pcretest.
1277    
1278    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1279        cannot analyze patterns in sufficient detail to catch mutual recursions
1280        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1281        subgroup is called recursively as a subpattern for a second time at the
1282        same position in the subject string. In previous releases this might have
1283        been caught by the recursion limit, or it might have run out of stack.
1284    
1285    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1286        happen only once. PCRE was, however incorrectly giving a compile time error
1287        "recursive call could loop indefinitely" because it cannot analyze the
1288        pattern in sufficient detail. The compile time test no longer happens when
1289        PCRE is compiling a conditional subpattern, but actual runaway loops are
1290        now caught at runtime (see 40 above).
1291    
1292    42. It seems that Perl allows any characters other than a closing parenthesis
1293        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1294        has been changed to be the same.
1295    
1296    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1297        as not to get warnings when autogen.sh is called. Also changed
1298        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1299    
1300    44. To help people who use pcregrep to scan files containing exceedingly long
1301        lines, the following changes have been made:
1302    
1303        (a) The default value of the buffer size parameter has been increased from
1304            8K to 20K. (The actual buffer used is three times this size.)
1305    
1306        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1307            PCRE is built.
1308    
1309        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1310            to be set at run time.
1311    
1312        (d) Numerical values in pcregrep options can be followed by K or M, for
1313            example --buffer-size=50K.
1314    
1315        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1316            given and the return code is set to 2.
1317    
1318    45. Add a pointer to the latest mark to the callout data block.
1319    
1320    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1321        partial match of an empty string instead of no match. This was specific to
1322        the use of ".".
1323    
1324    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1325        complete match instead of a partial match. This bug was dependent on both
1326        the PCRE_UTF8 and PCRE_DOTALL options being set.
1327    
1328    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1329        starting byte set, because \b was not being ignored.
1330    
1331    
1332    Version 8.12 15-Jan-2011
1333    ------------------------
1334    
1335    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1336        checks for such things as part of the documentation building process.
1337    
1338    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1339        --match-limit and --recursion-limit options (added for 8.11). In
1340        particular, this made one of the standard tests fail. (The integer value
1341        went into the wrong half of a long int.)
1342    
1343    3.  If the --colour option was given to pcregrep with -v (invert match), it
1344        did strange things, either producing crazy output, or crashing. It should,
1345        of course, ignore a request for colour when reporting lines that do not
1346        match.
1347    
1348    4.  Another pcregrep bug caused similar problems if --colour was specified with
1349        -M (multiline) and the pattern match finished with a line ending.
1350    
1351    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1352        matched in multiline mode, the following line was shown as part of the
1353        match. This seems wrong, so I have changed it.
1354    
1355    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1356        the check for further matches in the same line (so they could be coloured)
1357        to overrun the end of the current line. If another match was found, it was
1358        incorrectly shown (and then shown again when found in the next line).
1359    
1360    7.  If pcregrep was compiled under Windows, there was a reference to the
1361        function pcregrep_exit() before it was defined. I am assuming this was
1362        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1363        reported by a user. I've moved the definition above the reference.
1364    
1365    
1366    Version 8.11 10-Dec-2010
1367    ------------------------
1368    
1369    1.  (*THEN) was not working properly if there were untried alternatives prior
1370        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1371        backtracked to try for "b" instead of moving to the next alternative branch
1372        at the same level (in this case, to look for "c"). The Perl documentation
1373        is clear that when (*THEN) is backtracked onto, it goes to the "next
1374        alternative in the innermost enclosing group".
1375    
1376    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1377        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1378        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1379        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1380        (*THEN).
1381    
1382    3.  If \s appeared in a character class, it removed the VT character from
1383        the class, even if it had been included by some previous item, for example
1384        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1385        of \s, but is part of the POSIX "space" class.)
1386    
1387    4.  A partial match never returns an empty string (because you can always
1388        match an empty string at the end of the subject); however the checking for
1389        an empty string was starting at the "start of match" point. This has been
1390        changed to the "earliest inspected character" point, because the returned
1391        data for a partial match starts at this character. This means that, for
1392        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1393        (previously it gave "no match").
1394    
1395    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1396        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1397        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1398        has an implication that the given string is incomplete (because a partial
1399        match is preferred over a full match). For this reason, these items now
1400        give a partial match in this situation. [Aside: previously, the one case
1401        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1402        match rather than a full match, which was wrong by the old rules, but is
1403        now correct.]
1404    
1405    6.  There was a bug in the handling of #-introduced comments, recognized when
1406        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1407        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1408        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1409        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1410        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1411        places in pcre_compile().
1412    
1413    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1414        comments when looking ahead for named forward references to subpatterns,
1415        the only newline sequence it recognized was NL. It now handles newlines
1416        according to the set newline convention.
1417    
1418    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1419        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1420        cater for a lack of strerror(). These oversights have been fixed.
1421    
1422    9.  Added --match-limit and --recursion-limit to pcregrep.
1423    
1424    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1425    
1426    11. When the -o option was used, pcregrep was setting a return code of 1, even
1427        when matches were found, and --line-buffered was not being honoured.
1428    
1429    12. Added an optional parentheses number to the -o and --only-matching options
1430        of pcregrep.
1431    
1432    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1433        can match an empty string. The code to do it in pcretest and pcredemo
1434        needed fixing:
1435    
1436        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1437            only one byte after an empty string match just before CRLF (this case
1438            just got forgotten; "any" and "anycrlf" were OK).
1439    
1440        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1441            mode when an empty string match preceded an ASCII character followed by
1442            a non-ASCII character. (The code for advancing by one character rather
1443            than one byte was nonsense.)
1444    
1445        (c) The pcredemo.c sample program did not have any code at all to handle
1446            the cases when CRLF is a valid newline sequence.
1447    
1448    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1449        as a starting offset was within the subject string. There is now a new
1450        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1451        negative or greater than the length of the string. In order to test this,
1452        pcretest is extended to allow the setting of negative starting offsets.
1453    
1454    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1455        starting offset points to the beginning of a UTF-8 character was
1456        unnecessarily clumsy. I tidied it up.
1457    
1458    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1459        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1460    
1461    17. Nobody had reported that the --include_dir option, which was added in
1462        release 7.7 should have been called --include-dir (hyphen, not underscore)
1463        for compatibility with GNU grep. I have changed it to --include-dir, but
1464        left --include_dir as an undocumented synonym, and the same for
1465        --exclude-dir, though that is not available in GNU grep, at least as of
1466        release 2.5.4.
1467    
1468    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1469        characters from a string of bytes) have been redefined so as not to use
1470        loops, in order to improve performance in some environments. At the same
1471        time, I abstracted some of the common code into auxiliary macros to save
1472        repetition (this should not affect the compiled code).
1473    
1474    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1475        compile-time error is now given if \c is not followed by an ASCII
1476        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1477        different, and any byte value is allowed.)
1478    
1479    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1480        START_OPTIMIZE option, which is now allowed at compile time - but just
1481        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1482        to pcregrep and other applications that have no direct access to PCRE
1483        options. The new /Y option in pcretest sets this option when calling
1484        pcre_compile().
1485    
1486    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1487        back references. Groups containing recursive back references were forced to
1488        be atomic by that change, but in the case of named groups, the amount of
1489        memory required was incorrectly computed, leading to "Failed: internal
1490        error: code overflow". This has been fixed.
1491    
1492    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1493        pcretest.c, to avoid build problems in some Borland environments.
1494    
1495    
1496    Version 8.10 25-Jun-2010
1497    ------------------------
1498    
1499    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1500        THEN.
1501    
1502    2.  (*ACCEPT) was not working when inside an atomic group.
1503    
1504    3.  Inside a character class, \B is treated as a literal by default, but
1505        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1506        causes the error). The code is unchanged, but I tidied the documentation.
1507    
1508    4.  Inside a character class, PCRE always treated \R and \X as literals,
1509        whereas Perl faults them if its -w option is set. I have changed PCRE so
1510        that it faults them when PCRE_EXTRA is set.
1511    
1512    5.  Added support for \N, which always matches any character other than
1513        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1514    
1515    6.  When compiling pcregrep with newer versions of gcc which may have
1516        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1517        declared with attribute warn_unused_result" were given. Just casting the
1518        result to (void) does not stop the warnings; a more elaborate fudge is
1519        needed. I've used a macro to implement this.
1520    
1521    7.  Minor change to pcretest.c to avoid a compiler warning.
1522    
1523    8.  Added four artifical Unicode properties to help with an option to make
1524        \s etc use properties (see next item). The new properties are: Xan
1525        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1526    
1527    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1528        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1529        this option. Modified pcretest to add /W to test this facility. Added
1530        REG_UCP to make it available via the POSIX interface.
1531    
1532    10. Added --line-buffered to pcregrep.
1533    
1534    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1535        studied, and the match started with a letter with a code point greater than
1536        127 whose first byte was different to the first byte of the other case of
1537        the letter, the other case of this starting letter was not recognized
1538        (#976).
1539    
1540    12. If a pattern that was studied started with a repeated Unicode property
1541        test, for example, \p{Nd}+, there was the theoretical possibility of
1542        setting up an incorrect bitmap of starting bytes, but fortunately it could
1543        not have actually happened in practice until change 8 above was made (it
1544        added property types that matched character-matching opcodes).
1545    
1546    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1547        possible starting bytes for non-anchored patterns.
1548    
1549    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1550        \R, and also a number of cases that involve Unicode properties, both
1551        explicit and implicit when PCRE_UCP is set.
1552    
1553    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1554        input, it could crash or give wrong results if characters with values
1555        greater than 0xc0 were present in the subject string. (Detail: it assumed
1556        UTF-8 input when processing these items.)
1557    
1558    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1559        size_t is 64-bit (#991).
1560    
1561    17. Added a check for running out of memory when PCRE is compiled with
1562        --disable-stack-for-recursion (#990).
1563    
1564    18. If the last data line in a file for pcretest does not have a newline on
1565        the end, a newline was missing in the output.
1566    
1567    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1568        less than 128) in its various bitmaps. However, there is a facility for
1569        generating tables according to the current locale when PCRE is compiled. It
1570        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1571        characters, are recognized by isspace() and therefore were getting set in
1572        these tables, and indeed these tables seem to approximate to ISO 8859. This
1573        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1574        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1575        which of course cannot start UTF-8 characters. I have changed the code so
1576        that only real ASCII characters (less than 128) and the correct starting
1577        bytes for UTF-8 encodings are set for characters greater than 127 when in
1578        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1579        altogether.)
1580    
1581    20. Added the /T option to pcretest so as to be able to run tests with non-
1582        standard character tables, thus making it possible to include the tests
1583        used for 19 above in the standard set of tests.
1584    
1585    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1586        reference to a subpattern the other side of a comment that contains an
1587        opening parenthesis caused either an internal compiling error, or a
1588        reference to the wrong subpattern.
1589    
1590    
1591    Version 8.02 19-Mar-2010
1592    ------------------------
1593    
1594    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1595    
1596    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1597        configured.
1598    
1599    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1600        original author of that file, following a query about its status.
1601    
1602    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1603        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1604    
1605    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1606        quantifier applied to a forward-referencing subroutine call, could compile
1607        incorrect code or give the error "internal error: previously-checked
1608        referenced subpattern not found".
1609    
1610    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1611        variables to point to external functions. For these systems, therefore,
1612        pcre_malloc etc. are now initialized to local functions that call the
1613        relevant global functions.
1614    
1615    7.  There were two entries missing in the vectors called coptable and poptable
1616        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1617        I've fixed the data, and added a kludgy way of testing at compile time that
1618        the lengths are correct (equal to the number of opcodes).
1619    
1620    8.  Following on from 7, I added a similar kludge to check the length of the
1621        eint vector in pcreposix.c.
1622    
1623    9.  Error texts for pcre_compile() are held as one long string to avoid too
1624        much relocation at load time. To find a text, the string is searched,
1625        counting zeros. There was no check for running off the end of the string,
1626        which could happen if a new error number was added without updating the
1627        string.
1628    
1629    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1630    
1631    11. \K was not working if it appeared in an atomic group or in a group that
1632        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1633        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1634        the assertion is positive, but not if it is negative.
1635    
1636    12. Change 11 fortuitously reduced the size of the stack frame used in the
1637        "match()" function of pcre_exec.c by one pointer. Forthcoming
1638        implementation of support for (*MARK) will need an extra pointer on the
1639        stack; I have reserved it now, so that the stack frame size does not
1640        decrease.
1641    
1642    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1643        item in branch that calls a recursion is a subroutine call - as in the
1644        second branch in the above example - was incorrectly given the compile-
1645        time error "recursive call could loop indefinitely" because pcre_compile()
1646        was not correctly checking the subroutine for matching a non-empty string.
1647    
1648    14. The checks for overrunning compiling workspace could trigger after an
1649        overrun had occurred. This is a "should never occur" error, but it can be
1650        triggered by pathological patterns such as hundreds of nested parentheses.
1651        The checks now trigger 100 bytes before the end of the workspace.
1652    
1653    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1654    
1655    
1656    Version 8.01 19-Jan-2010
1657    ------------------------
1658    
1659    1.  If a pattern contained a conditional subpattern with only one branch (in
1660        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1661        computed the wrong minimum data length (which is of course zero for such
1662        subpatterns). This could cause incorrect "no match" results.
1663    
1664    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1665        the pattern is reset in the first branch, pcre_compile() failed with
1666        "internal error: code overflow at offset...". This happened only when
1667        the reset was to the original external option setting. (An optimization
1668        abstracts leading options settings into an external setting, which was the
1669        cause of this.)
1670    
1671    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1672        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1673        assertion pattern did not match (meaning that the assertion was true), it
1674        was incorrectly treated as false if the SKIP had been reached during the
1675        matching. This also applied to assertions used as conditions.
1676    
1677    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1678        assertion subpattern, including such a pattern used as a condition,
1679        unpredictable results occurred, instead of the error return
1680        PCRE_ERROR_DFA_UITEM.
1681    
1682    5.  The C++ GlobalReplace function was not working like Perl for the special
1683        situation when an empty string is matched. It now does the fancy magic
1684        stuff that is necessary.
1685    
1686    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1687        removed. (These were left over from very, very early versions of PCRE.)
1688    
1689    7.  Some cosmetic changes to the code to make life easier when compiling it
1690        as part of something else:
1691    
1692        (a) Change DEBUG to PCRE_DEBUG.
1693    
1694        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1695            called "current" as "current_branch", to prevent a collision with the
1696            Linux macro when compiled as a kernel module.
1697    
1698        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1699            prevent a collision with the Linux macro when compiled as a kernel
1700            module.
1701    
1702    8.  In pcre_compile() there are some checks for integer overflows that used to
1703        cast potentially large values to (double). This has been changed to that
1704        when building, a check for int64_t is made, and if it is found, it is used
1705        instead, thus avoiding the use of floating point arithmetic. (There is no
1706        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1707        double.
1708    
1709    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1710        2005 (difference between two addresses compared to an unsigned value).
1711    
1712    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1713        custom one, because of the following reported problem in Windows:
1714    
1715          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1716              under Win32.
1717          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1718              therefore missing the function definition.
1719          - The compiler thus generates a "C" signature for the test function.
1720          - The linker fails to find the "C" function.
1721          - PCRE fails to configure if asked to do so against libbz2.
1722    
1723    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1724        messages were output:
1725    
1726          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1727          rerunning libtoolize, to keep the correct libtool macros in-tree.
1728          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1729    
1730        I have done both of these things.
1731    
1732    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1733        most of the time, it *can* run out if it is given a pattern that contains a
1734        runaway infinite recursion. I updated the discussion in the pcrestack man
1735        page.
1736    
1737    13. Now that we have gone to the x.xx style of version numbers, the minor
1738        version may start with zero. Using 08 or 09 is a bad idea because users
1739        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1740        interpreted as invalid octal numbers. I've updated the previous comment in
1741        configure.ac, and also added a check that gives an error if 08 or 09 are
1742        used.
1743    
1744    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1745        causing partial matching to fail when the end of the subject matched \W
1746        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1747    
1748    15. There were some discrepancies between the declarations in pcre_internal.h
1749        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1750        their definitions. The declarations used "const uschar *" and the
1751        definitions used USPTR. Even though USPTR is normally defined as "const
1752        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1753        reported that: "This difference in casting confuses some C++ compilers, for
1754        example, SunCC recognizes above declarations as different functions and
1755        generates broken code for hbpcre." I have changed the declarations to use
1756        USPTR.
1757    
1758    16. GNU libtool is named differently on some systems. The autogen.sh script now
1759        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1760        (FreeBSD).
1761    
1762    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1763        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1764        comment: "Figure out how to create a longlong from a string: strtoll and
1765        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1766        instance, but it only takes 2 args instead of 3!"
1767    
1768    18. A subtle bug concerned with back references has been fixed by a change of
1769        specification, with a corresponding code fix. A pattern such as
1770        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1771        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1772        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1773        same bug. Such groups have to be quantified to be useful, or contained
1774        inside another quantified group. (If there's no repetition, the reference
1775        can never match.) The problem arises because, having left the group and
1776        moved on to the rest of the pattern, a later failure that backtracks into
1777        the group uses the captured value from the final iteration of the group
1778        rather than the correct earlier one. I have fixed this in PCRE by forcing
1779        any group that contains a reference to itself to be an atomic group; that
1780        is, there cannot be any backtracking into it once it has completed. This is
1781        similar to recursive and subroutine calls.
1782    
1783    
1784    Version 8.00 19-Oct-09
1785    ----------------------
1786    
1787    1.  The table for translating pcre_compile() error codes into POSIX error codes
1788        was out-of-date, and there was no check on the pcre_compile() error code
1789        being within the table. This could lead to an OK return being given in
1790        error.
1791    
1792    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1793        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1794        in a Windows environment.
1795    
1796    3.  The pcregrep --count option prints the count for each file even when it is
1797        zero, as does GNU grep. However, pcregrep was also printing all files when
1798        --files-with-matches was added. Now, when both options are given, it prints
1799        counts only for those files that have at least one match. (GNU grep just
1800        prints the file name in this circumstance, but including the count seems
1801        more useful - otherwise, why use --count?) Also ensured that the
1802        combination -clh just lists non-zero counts, with no names.
1803    
1804    4.  The long form of the pcregrep -F option was incorrectly implemented as
1805        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1806        but it seems right to fix it, and I didn't think it was worth preserving
1807        the old behaviour.
1808    
1809    5.  The command line items --regex=pattern and --regexp=pattern were not
1810        recognized by pcregrep, which required --regex pattern or --regexp pattern
1811        (with a space rather than an '='). The man page documented the '=' forms,
1812        which are compatible with GNU grep; these now work.
1813    
1814    6.  No libpcreposix.pc file was created for pkg-config; there was just
1815        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1816    
1817    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1818        when UCP support is not needed, by modifying the Python script that
1819        generates it from Unicode data files. This should not matter if the module
1820        is correctly used as a library, but I received one complaint about 50K of
1821        unwanted data. My guess is that the person linked everything into his
1822        program rather than using a library. Anyway, it does no harm.
1823    
1824    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1825        was a minimum greater than 1 for a wide character in a possessive
1826        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1827        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1828        character. Chaos in the form of incorrect output or a compiling loop could
1829        result.
1830    
1831    9.  The restrictions on what a pattern can contain when partial matching is
1832        requested for pcre_exec() have been removed. All patterns can now be
1833        partially matched by this function. In addition, if there are at least two
1834        slots in the offset vector, the offset of the earliest inspected character
1835        for the match and the offset of the end of the subject are set in them when
1836        PCRE_ERROR_PARTIAL is returned.
1837    
1838    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1839        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1840        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1841        and may be more useful for multi-segment matching.
1842    
1843    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1844        used to be given if ever the end of the subject was reached; now it is
1845        given only if matching could not proceed because another character was
1846        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1847        string "Z", which now yields "no match" instead of "partial match". In the
1848        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1849        final character ended with (*FAIL).
1850    
1851    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1852        if the pattern had a "must contain" character that was already found in the
1853        earlier partial match, unless partial matching was again requested. For
1854        example, with the pattern /dog.(body)?/, the "must contain" character is
1855        "g". If the first part-match was for the string "dog", restarting with
1856        "sbody" failed. This bug has been fixed.
1857    
1858    13. The string returned by pcre_dfa_exec() after a partial match has been
1859        changed so that it starts at the first inspected character rather than the
1860        first character of the match. This makes a difference only if the pattern
1861        starts with a lookbehind assertion or \b or \B (\K is not supported by
1862        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1863        matching functions compatible, and I think it's the right thing to do.
1864    
1865    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1866        so that the demonstration program is easily available in environments where
1867        PCRE has not been installed from source.
1868    
1869    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1870        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1871        library.
1872    
1873    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1874        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1875        is not the first non-POSIX option to be added. Clearly some people find
1876        these options useful.
1877    
1878    17. If a caller to the POSIX matching function regexec() passes a non-zero
1879        value for nmatch with a NULL value for pmatch, the value of
1880        nmatch is forced to zero.
1881    
1882    18. RunGrepTest did not have a test for the availability of the -u option of
1883        the diff command, as RunTest does. It now checks in the same way as
1884        RunTest, and also checks for the -b option.
1885    
1886    19. If an odd number of negated classes containing just a single character
1887        interposed, within parentheses, between a forward reference to a named
1888        subpattern and the definition of the subpattern, compilation crashed with
1889        an internal error, complaining that it could not find the referenced
1890        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1891        [The bug was that it was starting one character too far in when skipping
1892        over the character class, thus treating the ] as data rather than
1893        terminating the class. This meant it could skip too much.]
1894    
1895    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1896        /g option in pcretest when the pattern contains \K, which makes it possible
1897        to have an empty string match not at the start, even when the pattern is
1898        anchored. Updated pcretest and pcredemo to use this option.
1899    
1900    21. If the maximum number of capturing subpatterns in a recursion was greater
1901        than the maximum at the outer level, the higher number was returned, but
1902        with unset values at the outer level. The correct (outer level) value is
1903        now given.
1904    
1905    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1906        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1907        make it do so. The string so far is captured, making this feature
1908        compatible with Perl.
1909    
1910    23. The tests have been re-organized, adding tests 11 and 12, to make it
1911        possible to check the Perl 5.10 features against Perl 5.10.
1912    
1913    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1914        pattern matches a fixed length string. PCRE did not allow this; now it
1915        does. Neither allows recursion.
1916    
1917    25. I finally figured out how to implement a request to provide the minimum
1918        length of subject string that was needed in order to match a given pattern.
1919        (It was back references and recursion that I had previously got hung up
1920        on.) This code has now been added to pcre_study(); it finds a lower bound
1921        to the length of subject needed. It is not necessarily the greatest lower
1922        bound, but using it to avoid searching strings that are too short does give
1923        some useful speed-ups. The value is available to calling programs via
1924        pcre_fullinfo().
1925    
1926    26. While implementing 25, I discovered to my embarrassment that pcretest had
1927        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1928        study optimizations had never been tested with that matching function.
1929        Oops. What is worse, even when it was passed study data, there was a bug in
1930        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1931        were also very few tests of studied patterns with pcre_dfa_exec().
1932    
1933    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1934        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1935        on the other side of the coin, they are no longer allowed to have different
1936        names, because these cannot be distinguished in PCRE, and this has caused
1937        confusion. (This is a difference from Perl.)
1938    
1939    28. When duplicate subpattern names are present (necessarily with different
1940        numbers, as required by 27 above), and a test is made by name in a
1941        conditional pattern, either for a subpattern having been matched, or for
1942        recursion in such a pattern, all the associated numbered subpatterns are
1943        tested, and the overall condition is true if the condition is true for any
1944        one of them. This is the way Perl works, and is also more like the way
1945        testing by number works.
1946    
1947    
1948    Version 7.9 11-Apr-09
1949    ---------------------
1950    
1951    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1952        (pcretest), all targets were linked against these libraries. This included
1953        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1954        libraries. This caused unwanted dependencies to be created. This problem
1955        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1956        pcretest is linked with readline.
1957    
1958    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1959        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1960        moved outside it again, because FALSE and TRUE are already defined in AIX,
1961        but BOOL is not.
1962    
1963    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1964        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1965    
1966    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1967        hyphens) following file names and line numbers when outputting matching
1968        lines. This is not true; no spaces are inserted. I have also clarified the
1969        wording for the --colour (or --color) option.
1970    
1971    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1972        was not coloured; this is different to GNU grep, so I have changed it to be
1973        the same.
1974    
1975    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1976        each matching line was coloured. Now it goes on to look for further matches
1977        of any of the test patterns, which is the same behaviour as GNU grep.
1978    
1979    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1980        doesn't make sense to accept an empty string match in pcregrep, so I have
1981        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1982        seems to be how GNU grep behaves. [But see later change 40 for release
1983        8.33.]
1984    
1985    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1986        start or after a newline", because the conditional assertion was not being
1987        correctly handled. The rule now is that both the assertion and what follows
1988        in the first alternative must satisfy the test.
1989    
1990    9.  If auto-callout was enabled in a pattern with a conditional group whose
1991        condition was an assertion, PCRE could crash during matching, both with
1992        pcre_exec() and pcre_dfa_exec().
1993    
1994    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1995        used for matching.
1996    
1997    11. Unicode property support in character classes was not working for
1998        characters (bytes) greater than 127 when not in UTF-8 mode.
1999    
2000    12. Added the -M command line option to pcretest.
2001    
2002    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2003    
2004    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2005    
2006    16. Added comments and documentation about mis-use of no_arg in the C++
2007        wrapper.
2008    
2009    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2010        from Martin Jerabek that uses macro names for all relevant character and
2011        string constants.
2012    
2013    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2014        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2015        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2016        these, but not everybody uses configure.
2017    
2018    19. A conditional group that had only one branch was not being correctly
2019        recognized as an item that could match an empty string. This meant that an
2020        enclosing group might also not be so recognized, causing infinite looping
2021        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2022        with the subject "ab", where knowledge that the repeated group can match
2023        nothing is needed in order to break the loop.
2024    
2025    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2026        exec(), but without supplying a callout function, matching went wrong.
2027    
2028    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2029        leak if the size of the offset vector was greater than 30. When the vector
2030        is smaller, the saved offsets during recursion go onto a local stack
2031        vector, but for larger vectors malloc() is used. It was failing to free
2032        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2033        error, in fact).
2034    
2035    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2036        heapframe that is used only when UTF-8 support is enabled. This caused no
2037        problem, but was untidy.
2038    
2039    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2040        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2041        included within another project.
2042    
2043    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2044        slightly modified by me:
2045    
2046          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2047              not building pcregrep.
2048    
2049          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2050              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2051    
2052    25. Forward references, both numeric and by name, in patterns that made use of
2053        duplicate group numbers, could behave incorrectly or give incorrect errors,
2054        because when scanning forward to find the reference group, PCRE was not
2055        taking into account the duplicate group numbers. A pattern such as
2056        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2057    
2058    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2059        the feature of a custom pointer more persuasive (as requested by a user).
2060    
2061    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2062        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2063        pre-defined.
2064    
2065    28. Added support for (*UTF8) at the start of a pattern.
2066    
2067    29. Arrange for flags added by the "release type" setting in CMake to be shown
2068        in the configuration summary.
2069    
2070    
2071    Version 7.8 05-Sep-08
2072    ---------------------
2073    
2074    1.  Replaced UCP searching code with optimized version as implemented for Ad
2075        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2076        stage table and inline lookup instead of a function, giving speed ups of 2
2077        to 5 times on some simple patterns that I tested. Permission was given to
2078        distribute the MultiStage2.py script that generates the tables (it's not in
2079        the tarball, but is in the Subversion repository).
2080    
2081    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2082        scripts.
2083    
2084    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2085        a group with a zero qualifier. The result of the study could be incorrect,
2086        or the function might crash, depending on the pattern.
2087    
2088    4.  Caseless matching was not working for non-ASCII characters in back
2089        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2090        It now works when Unicode Property Support is available.
2091    
2092    5.  In pcretest, an escape such as \x{de} in the data was always generating
2093        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2094        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2095        truncation.
2096    
2097    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2098    
2099    7.  Added two (int) casts to pcregrep when printing the difference of two
2100        pointers, in case they are 64-bit values.
2101    
2102    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2103        test 2 if it fails.
2104    
2105    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2106        and a #define of that name to empty if it is not externally set. This is to
2107        allow users of MSVC to set it if necessary.
2108    
2109    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2110        the convenience functions in the pcre_get.c source file.
2111    
2112    11. An option change at the start of a pattern that had top-level alternatives
2113        could cause overwriting and/or a crash. This command provoked a crash in
2114        some environments:
2115    
2116          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2117    
2118        This potential security problem was recorded as CVE-2008-2371.
2119    
2120    12. For a pattern where the match had to start at the beginning or immediately
2121        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2122        pcre_dfa_exec() could read past the end of the passed subject if there was
2123        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2124        pcretest so that it places the subject at the end of its malloc-ed buffer.
2125    
2126    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2127        exec() might read past the end of the data buffer in UTF-8 mode.
2128    
2129    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2130        the data contained the byte 0x85 as part of a UTF-8 character within its
2131        first line. This applied both to normal and DFA matching.
2132    
2133    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2134        /^[^d]*?$/8 failed to match "abc".
2135    
2136    16. Added a missing copyright notice to pcrecpp_internal.h.
2137    
2138    17. Make it more clear in the documentation that values returned from
2139        pcre_exec() in ovector are byte offsets, not character counts.
2140    
2141    18. Tidied a few places to stop certain compilers from issuing warnings.
2142    
2143    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2144        supplied by Stefan Weber. I made a further small update for 7.8 because
2145        there is a change of source arrangements: the pcre_searchfuncs.c module is
2146        replaced by pcre_ucd.c.
2147    
2148    
2149    Version 7.7 07-May-08
2150    ---------------------
2151    
2152    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2153        a string to a long long, pretend we don't even have a long long." This is
2154        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2155    
2156    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2157        pre-7.6 versions, which defined a global no_arg variable instead of putting
2158        it in the RE class. (See also #8 below.)
2159    
2160    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2161        Lopes.
2162    
2163    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2164    
2165        (1) The include/exclude patterns were being applied to the whole pathnames
2166            of files, instead of just to the final components.
2167    
2168        (2) If there was more than one level of directory, the subdirectories were
2169            skipped unless they satisfied the include/exclude conditions. This is
2170            inconsistent with GNU grep (and could even be seen as contrary to the
2171            pcregrep specification - which I improved to make it absolutely clear).
2172            The action now is always to scan all levels of directory, and just
2173            apply the include/exclude patterns to regular files.
2174    
2175    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2176        --exclude_dir in the tests to avoid scanning .svn directories.
2177    
2178    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2179        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2180        doesn't support NULs in patterns.
2181    
2182    7.  Added some missing "const"s to declarations of static tables in
2183        pcre_compile.c and pcre_dfa_exec.c.
2184    
2185    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2186        caused by fix #2  above. (Subsequently also a second patch to fix the
2187        first patch. And a third patch - this was a messy problem.)
2188    
2189    9.  Applied Craig's patch to remove the use of push_back().
2190    
2191    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2192        matching function regexec().
2193    
2194    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2195        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2196        references. PCRE supports relative numbers with this syntax (I don't think
2197        Oniguruma does).
2198    
2199    12. Previously, a group with a zero repeat such as (...){0} was completely
2200        omitted from the compiled regex. However, this means that if the group
2201        was called as a subroutine from elsewhere in the pattern, things went wrong
2202        (an internal error was given). Such groups are now left in the compiled
2203        pattern, with a new opcode that causes them to be skipped at execution
2204        time.
2205    
2206    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2207        to the way PCRE behaves:
2208    
2209        (a) A lone ] character is dis-allowed (Perl treats it as data).
2210    
2211        (b) A back reference to an unmatched subpattern matches an empty string
2212            (Perl fails the current match path).
2213    
2214        (c) A data ] in a character class must be notated as \] because if the
2215            first data character in a class is ], it defines an empty class. (In
2216            Perl it is not possible to have an empty class.) The empty class []
2217            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2218            The negative empty class [^] matches any one character, independently
2219            of the DOTALL setting.
2220    
2221    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2222        non-existent subpattern following a character class starting with ']' and
2223        containing () gave an internal compiling error instead of "reference to
2224        non-existent subpattern". Fortunately, when the pattern did exist, the
2225        compiled code was correct. (When scanning forwards to check for the
2226        existence of the subpattern, it was treating the data ']' as terminating
2227        the class, so got the count wrong. When actually compiling, the reference
2228        was subsequently set up correctly.)
2229    
2230    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2231        it was being rejected as not supported by pcre_dfa_exec(), even though
2232        other assertions are supported. I have made pcre_dfa_exec() support
2233        (*FAIL).
2234    
2235    16. The implementation of 13c above involved the invention of a new opcode,
2236        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2237        cannot be changed at match time, I realized I could make a small
2238        improvement to matching performance by compiling OP_ALLANY instead of
2239        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2240        on the OP_ANY path.
2241    
2242    17. Compiling pcretest on Windows with readline support failed without the
2243        following two fixes: (1) Make the unistd.h include conditional on
2244        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2245    
2246    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2247        ncurses library to be included for pcretest when ReadLine support is
2248        requested, but also to allow for it to be overridden. This patch came from
2249        Daniel Bergström.
2250    
2251    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2252        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2253        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2254        spotting this.
2255    
2256    
2257    Version 7.6 28-Jan-08
2258    ---------------------
2259    
2260    1.  A character class containing a very large number of characters with
2261        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2262        overflow.
2263    
2264    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2265        HAVE_LONG_LONG is not defined.
2266    
2267    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2268        bring them up to date and include new features. This patch includes:
2269    
2270        - Fixed PH's badly added libz and libbz2 support.
2271        - Fixed a problem with static linking.
2272        - Added pcredemo. [But later removed - see 7 below.]
2273        - Fixed dftables problem and added an option.
2274        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2275            HAVE_LONG_LONG.
2276        - Added readline support for pcretest.
2277        - Added an listing of the option settings after cmake has run.
2278    
2279    4.  A user submitted a patch to Makefile that makes it easy to create
2280        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2281        Makefile.am that cause it to include this special target, without
2282        affecting anything else. Note that the same mingw target plus all
2283        the other distribution libraries and programs are now supported
2284        when configuring with CMake (see 6 below) instead of with
2285        Configure/Make.
2286    
2287    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2288        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2289        exported in the Windows port". It has not yet been confirmed that the patch
2290        solves the problem, but it does no harm.
2291    
2292    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2293        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2294        with CMake, and also correct the comment about stack recursion.
2295    
2296    7.  Remove the automatic building of pcredemo from the ./configure system and
2297        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2298        of a program that users should build themselves after PCRE is installed, so
2299        building it automatically is not really right. What is more, it gave
2300        trouble in some build environments.
2301    
2302    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2303    
2304    
2305    Version 7.5 10-Jan-08
2306    ---------------------
2307    
2308    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2309        values in parens when parsing an RE using the C++ wrapper."
2310    
2311    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2312        Characters greater than 255 were excluded from the class instead of being
2313        included.
2314    
2315    3.  The same bug as (2) above applied to negated POSIX classes such as
2316        [:^space:].
2317    
2318    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2319        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2320        I have changed it.
2321    
2322    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2323        first named subpattern) and a construct such as (?&a) would reference the
2324        first named subpattern whose name started with "a" (in other words, the
2325        length check was missing). Both these problems are fixed. "Subpattern name
2326        expected" is now given for (?&) (a zero-length name), and this patch also
2327        makes it give the same error for \k'' (previously it complained that that
2328        was a reference to a non-existent subpattern).
2329    
2330    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2331        this is right because (?- can be followed by option settings as well as by
2332        digits. I have, however, made the messages clearer.
2333    
2334    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2335        than the number used in the conditional) now cause a compile-time error.
2336        This is actually not compatible with Perl, which accepts such patterns, but
2337        treats the conditional as always being FALSE (as PCRE used to), but it
2338        seems to me that giving a diagnostic is better.
2339    
2340    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2341        and messages.
2342    
2343    9.  Fix two occurrences of "backslash" in comments that should have been
2344        "backspace".
2345    
2346    10. Remove two redundant lines of code that can never be obeyed (their function
2347        was moved elsewhere).
2348    
2349    11. The program that makes PCRE's Unicode character property table had a bug
2350        which caused it to generate incorrect table entries for sequences of
2351        characters that have the same character type, but are in different scripts.
2352        It amalgamated them into a single range, with the script of the first of
2353        them. In other words, some characters were in the wrong script. There were
2354        thirteen such cases, affecting characters in the following ranges:
2355    
2356          U+002b0 - U+002c1
2357          U+0060c - U+0060d
2358          U+0061e - U+00612
2359          U+0064b - U+0065e
2360          U+0074d - U+0076d
2361          U+01800 - U+01805
2362          U+01d00 - U+01d77
2363          U+01d9b - U+01dbf
2364          U+0200b - U+0200f
2365          U+030fc - U+030fe
2366          U+03260 - U+0327f
2367          U+0fb46 - U+0fbb1
2368          U+10450 - U+1049d
2369    
2370    12. The -o option (show only the matching part of a line) for pcregrep was not
2371        compatible with GNU grep in that, if there was more than one match in a
2372        line, it showed only the first of them. It now behaves in the same way as
2373        GNU grep.
2374    
2375    13. If the -o and -v options were combined for pcregrep, it printed a blank
2376        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2377        does the same. The return code can be used to tell if there were any
2378        non-matching lines.
2379    
2380    14. Added --file-offsets and --line-offsets to pcregrep.
2381    
2382    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2383        infinitely looping recursion. The bug was that positive lookaheads were not
2384        being skipped when checking for a possible empty match (negative lookaheads
2385        and both kinds of lookbehind were skipped).
2386    
2387    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2388        inclusion of <windows.h> to before rather than after the definition of
2389        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2390    
2391    17. Specifying a possessive quantifier with a specific limit for a Unicode
2392        character property caused pcre_compile() to compile bad code, which led at
2393        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2394        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2395        caused the error; without that there was no problem.
2396    
2397    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2398    
2399    19. Added --enable-pcretest-libreadline.
2400    
2401    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2402        RE::GlobalReplace(). As a result, the number of replacements returned was
2403        double what it should be. I removed one of the increments, but Craig sent a
2404        later patch that removed the other one (the right fix) and added unit tests
2405        that check the return values (which was not done before).
2406    
2407    21. Several CMake things:
2408    
2409        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2410            the names libpcre and libpcreposix, not just pcre and pcreposix.
2411    
2412        (2) The above change means that pcretest and pcregrep are now correctly
2413            linked with the newly-built libraries, not previously installed ones.
2414    
2415        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2416    
2417    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2418        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2419        UTF-8 newline character). The key issue is that the pattern starts .*;
2420        this means that the match must be either at the beginning, or after a
2421        newline. The bug was in the code for advancing after a failed match and
2422        checking that the new position followed a newline. It was not taking
2423        account of UTF-8 characters correctly.
2424    
2425    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2426        character classes. PCRE was not treating the sequence [:...:] as a
2427        character class unless the ... were all letters. Perl, however, seems to
2428        allow any characters between [: and :], though of course it rejects as
2429        unknown any "names" that contain non-letters, because all the known class
2430        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2431        for example, whereas PCRE did not - it did not recognize a POSIX character
2432        class. This seemed a bit dangerous, so the code has been changed to be
2433        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2434        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2435        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2436        Perl does, and where it didn't before.
2437    
2438    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2439        Windows environments %n is disabled by default.
2440    
2441    
2442    Version 7.4 21-Sep-07
2443    ---------------------
2444    
2445    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2446        means that a class such as [\s] counted as "explicit reference to CR or
2447        LF". That isn't really right - the whole point of the change was to try to
2448        help when there was an actual mention of one of the two characters. So now
2449        the change happens only if \r or \n (or a literal CR or LF) character is
2450        encountered.
2451    
2452    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2453        of both had grown to the point where there were only 3 bits left.
2454        Fortunately, there was spare space in the data structure, and so I have
2455        moved the internal flags into a new 16-bit field to free up more option
2456        bits.
2457    
2458    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2459        but did not set the internal JCHANGED flag - either of these is enough to
2460        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2461        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2462        start sets both bits.
2463    
2464    4.  Added options (at build time, compile time, exec time) to change \R from
2465        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2466    
2467    5.  doc/pcresyntax.html was missing from the distribution.
2468    
2469    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2470        compatibility, even though it is no longer used.
2471    
2472    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2473        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2474        windows.h file is present (where different names are used). [This was
2475        reversed later after testing - see 16 below.]
2476    
2477    8.  Changed all #include <config.h> to #include "config.h". There were also
2478        some further <pcre.h> cases that I changed to "pcre.h".
2479    
2480    9.  When pcregrep was used with the --colour option, it missed the line ending
2481        sequence off the lines that it output.
2482    
2483    10. It was pointed out to me that arrays of string pointers cause lots of
2484        relocations when a shared library is dynamically loaded. A technique of
2485        using a single long string with a table of offsets can drastically reduce
2486        these. I have refactored PCRE in four places to do this. The result is
2487        dramatic:
2488    
2489          Originally:                          290
2490          After changing UCP table:            187
2491          After changing error message table:   43
2492          After changing table of "verbs"       36
2493          After changing table of Posix names   22
2494    
2495        Thanks to the folks working on Gregex for glib for this insight.
2496    
2497    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2498        unicode-properties was also set.
2499    
2500    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2501    
2502    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2503        checked only for CRLF.
2504    
2505    14. Added casts to pcretest.c to avoid compiler warnings.
2506    
2507    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2508    
2509    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2510        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2511        entirely. This removes changes made in 7 above.
2512    
2513    17. The CMake files have been updated, and there is now more information about
2514        building with CMake in the NON-UNIX-USE document.
2515    
2516    
2517    Version 7.3 28-Aug-07
2518    ---------------------
2519    
2520     1. In the rejigging of the build system that eventually resulted in 7.1, the
2521        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2522        brackets there is not right, since it causes compilers to look for an
2523        installed pcre.h, not the version that is in the source that is being
2524        compiled (which of course may be different). I have changed it back to:
2525    
2526          #include "pcre.h"
2527    
2528        I have a vague recollection that the change was concerned with compiling in
2529        different directories, but in the new build system, that is taken care of
2530        by the VPATH setting the Makefile.
2531    
2532     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2533        when the subject happened to end in the byte 0x85 (e.g. if the last
2534        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2535        characters but of course it shouldn't be taken as a newline when it is part
2536        of another character. The bug was that, for an unlimited repeat of . in
2537        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2538        characters when looking for a newline.
2539    
2540     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2541    
2542     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2543        in debug output.
2544    
2545     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2546        long printing in the pcrecpp unittest when running under MinGW.
2547    
2548     6. ESC_K was left out of the EBCDIC table.
2549    
2550     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2551        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2552        limit also applies to "virtual nesting" when a pattern is recursive, and in
2553        this case 1000 isn't so big. I have been able to remove this limit at the
2554        expense of backing off one optimization in certain circumstances. Normally,
2555        when pcre_exec() would call its internal match() function recursively and
2556        immediately return the result unconditionally, it uses a "tail recursion"
2557        feature to save stack. However, when a subpattern that can match an empty
2558        string has an unlimited repetition quantifier, it no longer makes this
2559        optimization. That gives it a stack frame in which to save the data for
2560        checking that an empty string has been matched. Previously this was taken
2561        from the 1000-entry workspace that had been reserved. So now there is no
2562        explicit limit, but more stack is used.
2563    
2564     8. Applied Daniel's patches to solve problems with the import/export magic
2565        syntax that is required for Windows, and which was going wrong for the
2566        pcreposix and pcrecpp parts of the library. These were overlooked when this
2567        problem was solved for the main library.
2568    
2569     9. There were some crude static tests to avoid integer overflow when computing
2570        the size of patterns that contain repeated groups with explicit upper
2571        limits. As the maximum quantifier is 65535, the maximum group length was
2572        set at 30,000 so that the product of these two numbers did not overflow a
2573        32-bit integer. However, it turns out that people want to use groups that
2574        are longer than 30,000 bytes (though not repeat them that many times).
2575        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2576        made it possible to implement the integer overflow checks in a much more
2577        dynamic way, which I have now done. The artificial limitation on group
2578        length has been removed - we now have only the limit on the total length of
2579        the compiled pattern, which depends on the LINK_SIZE setting.
2580    
2581    10. Fixed a bug in the documentation for get/copy named substring when
2582        duplicate names are permitted. If none of the named substrings are set, the
2583        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2584        empty string.
2585    
2586    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2587        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2588        because the ] is interpreted as the first data character and the
2589        terminating ] is not found. PCRE has been made compatible with Perl in this
2590        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2591        cause memory overwriting.
2592    
2593    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2594        string has been matched (to stop an infinite loop). It was not recognizing
2595        a conditional subpattern that could match an empty string if that
2596        subpattern was within another subpattern. For example, it looped when
2597        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2598        condition was not nested. This bug has been fixed.
2599    
2600    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2601        past the start of the subject in the presence of bytes with the top bit
2602        set, for example "\x8aBCD".
2603    
2604    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2605        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2606    
2607    14. Optimized (?!) to (*FAIL).
2608    
2609    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2610        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2611        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2612        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2613        does: it's just the validity check that is more restrictive.
2614    
2615    16. Inserted checks for integer overflows during escape sequence (backslash)
2616        processing, and also fixed erroneous offset values for syntax errors during
2617        backslash processing.
2618    
2619    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2620        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2621    
2622    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2623        caused an overrun.
2624    
2625    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2626        something other than just ASCII characters) inside a group that had an
2627        unlimited repeat caused a loop at compile time (while checking to see
2628        whether the group could match an empty string).
2629    
2630    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2631        [\P{Any}] did so. (Error in the code for printing property names.)
2632    
2633    21. An orphan \E inside a character class could cause a crash.
2634    
2635    22. A repeated capturing bracket such as (A)? could cause a wild memory
2636        reference during compilation.
2637    
2638    23. There are several functions in pcre_compile() that scan along a compiled
2639        expression for various reasons (e.g. to see if it's fixed length for look
2640        behind). There were bugs in these functions when a repeated \p or \P was
2641        present in the pattern. These operators have additional parameters compared
2642        with \d, etc, and these were not being taken into account when moving along
2643        the compiled data. Specifically:
2644    
2645        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2646            length.
2647    
2648        (b) An item such as \pL+ within a repeated group could cause crashes or
2649            loops.
2650    
2651        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2652            "reference to non-existent subpattern" error.
2653    
2654        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2655    
2656    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2657        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2658    
2659    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2660    
2661    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2662        character were causing crashes (broken optimization).
2663    
2664    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2665        \p or \P) caused a compile-time loop.
2666    
2667    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2668        break. For example, the unstudied pattern [\r\n]A does not match the string
2669        "\r\nA" because change 7.0/46 below moves the current point on by two
2670        characters after failing to match at the start. However, the pattern \nA
2671        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2672        the same is true. There doesn't seem any very clean way out of this, but
2673        what I have chosen to do makes the common cases work: PCRE now takes note
2674        of whether there can be an explicit match for \r or \n anywhere in the
2675        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2676        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2677        pattern has explicit CR or LF references.
2678    
2679    29. Added (*CR) etc for changing newline setting at start of pattern.
2680    
2681    
2682    Version 7.2 19-Jun-07
2683    ---------------------
2684    
2685     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2686        which is apparently normally available under Windows.
2687    
2688     2. Re-jig the pcregrep tests with different newline settings in an attempt
2689        to make them independent of the local environment's newline setting.
2690    
2691     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2692    
2693     4. Some of the "internals" tests were previously cut out when the link size
2694        was not 2, because the output contained actual offsets. The recent new
2695        "Z" feature of pcretest means that these can be cut out, making the tests
2696        usable with all link sizes.
2697    
2698     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2699        stack recursion. This gives a massive performance boost under BSD, but just
2700        a small improvement under Linux. However, it saves one field in the frame
2701        in all cases.
2702    
2703     6. Added more features from the forthcoming Perl 5.10:
2704    
2705        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2706            recursion call. It refers to the nth most recently opened parentheses.
2707    
2708        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2709            to be opened parentheses.
2710    
2711        (c) Conditions that refer to capturing parentheses can be specified
2712            relatively, for example, (?(-2)... or (?(+3)...
2713    
2714        (d) \K resets the start of the current match so that everything before
2715            is not part of it.
2716    
2717        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2718    
2719        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2720            reference syntax.
2721    
2722        (g) (?| introduces a group in which the numbering of parentheses in each
2723            alternative starts with the same number.
2724    
2725        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2726    
2727     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2728        PCRE_INFO_JCHANGED.
2729    
2730     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2731        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2732        for detecting groups that can match an empty string.
2733    
2734     9. A pattern with a very large number of alternatives (more than several
2735        hundred) was running out of internal workspace during the pre-compile
2736        phase, where pcre_compile() figures out how much memory will be needed. A
2737        bit of new cunning has reduced the workspace needed for groups with
2738        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2739        workspace instead of running out of the 4096 that are available.
2740    
2741    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2742    
2743    11. Applied patch from Google to remove an optimization that didn't quite work.
2744        The report of the bug said:
2745    
2746          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2747          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2748          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2749    
2750    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2751        it matched the wrong number of bytes.
2752    
2753    
2754    Version 7.1 24-Apr-07
2755    ---------------------
2756    
2757     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2758        that is more "standard", making use of automake and other Autotools. There
2759        is some re-arrangement of the files and adjustment of comments consequent
2760        on this.
2761    
2762     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2763        for recursive directory scanning broke on some systems because the files
2764        are not scanned in any specific order and on different systems the order
2765        was different. A call to "sort" has been inserted into RunGrepTest for the
2766        approprate test as a short-term fix. In the longer term there may be an
2767        alternative.
2768    
2769     3. I had an email from Eric Raymond about problems translating some of PCRE's
2770        man pages to HTML (despite the fact that I distribute HTML pages, some
2771        people do their own conversions for various reasons). The problems
2772        concerned the use of low-level troff macros .br and .in. I have therefore
2773        removed all such uses from the man pages (some were redundant, some could
2774        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2775        HTML has been updated to handle .nf/.fi and to complain if it encounters
2776        .br or .in.
2777    
2778     4. Updated comments in configure.ac that get placed in config.h.in and also
2779        arranged for config.h to be included in the distribution, with the name
2780        config.h.generic, for the benefit of those who have to compile without
2781        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2782    
2783     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2784        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2785        makevp.bat for the current PCRE, using the additional files
2786        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2787    
2788     6. A Windows user reported a minor discrepancy with test 2, which turned out
2789        to be caused by a trailing space on an input line that had got lost in his
2790        copy. The trailing space was an accident, so I've just removed it.
2791    
2792     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2793        that is needed.
2794    
2795     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2796        as "const" (a) because they are and (b) because it helps the PHP
2797        maintainers who have recently made a script to detect big data structures
2798        in the php code that should be moved to the .rodata section. I remembered
2799        to update Builducptable as well, so it won't revert if ucptable.h is ever
2800        re-created.
2801    
2802     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2803        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2804        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2805        support is not required. This saves 1.5-2K of code, which is important in
2806        some applications.
2807    
2808        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2809        so as not to refer to the tables, even though these functions will never be
2810        called when UTF-8 support is disabled. Otherwise there are problems with a
2811        shared library.
2812    
2813    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2814    
2815        (a) It was defining its arguments as char * instead of void *.
2816    
2817        (b) It was assuming that all moves were upwards in memory; this was true
2818            a long time ago when I wrote it, but is no longer the case.
2819    
2820        The emulated memove() is provided for those environments that have neither
2821        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2822        is clearly not the case, as these two bugs were recently reported.
2823    
2824    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2825        and Detrail to create the HTML documentation, the .txt form of the man
2826        pages, and it removes trailing spaces from listed files. It also creates
2827        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2828        case, it wraps all the #defines with #ifndefs. This script should be run
2829        before "make dist".
2830    
2831    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2832        with Unicode property support.
2833    
2834        (a) For a maximizing quantifier, if the two different cases of the
2835            character were of different lengths in their UTF-8 codings (there are
2836            some cases like this - I found 11), and the matching function had to
2837            back up over a mixture of the two cases, it incorrectly assumed they
2838            were both the same length.
2839    
2840        (b) When PCRE was configured to use the heap rather than the stack for
2841            recursion during matching, it was not correctly preserving the data for
2842            the other case of a UTF-8 character when checking ahead for a match
2843            while processing a minimizing repeat. If the check also involved
2844            matching a wide character, but failed, corruption could cause an
2845            erroneous result when trying to check for a repeat of the original
2846            character.
2847    
2848    13. Some tidying changes to the testing mechanism:
2849    
2850        (a) The RunTest script now detects the internal link size and whether there
2851            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2852            values substituted by "configure". (The RunGrepTest script already did
2853            this for UTF-8.) The configure.ac script no longer substitutes the
2854            relevant variables.
2855    
2856        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2857            with length and offset values. This means that the output is different
2858            for different internal link sizes. Test 2 is skipped for link sizes
2859            other than 2 because of this, bypassing the problem. Unfortunately,
2860            there was also a test in test 3 (the locale tests) that used /B and
2861            failed for link sizes other than 2. Rather than cut the whole test out,
2862            I have added a new /Z option to pcretest that replaces the length and
2863            offset values with spaces. This is now used to make test 3 independent
2864            of link size. (Test 2 will be tidied up later.)
2865    
2866    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2867        segmentation fault instead of returning the appropriate error message.
2868    
2869    15. In multiline mode when the newline sequence was set to "any", the pattern
2870        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2871        This doesn't seem right; it now treats the CRLF combination as the line
2872        ending, and so does not match in that case. It's only a pattern such as ^$
2873        that would hit this one: something like ^ABC$ would have failed after \r
2874        and then tried again after \r\n.
2875    
2876    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2877        in an attempt to make files that differ only in their line terminators
2878        compare equal. This works on Linux.
2879    
2880    17. Under certain error circumstances pcregrep might try to free random memory
2881        as it exited. This is now fixed, thanks to valgrind.
2882    
2883    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2884        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2885        was because its rules for how to advance for /g after matching an empty
2886        string at the end of a line did not allow for this case. They now check for
2887        it specially.
2888    
2889    20. pcretest is supposed to handle patterns and data of any length, by
2890        extending its buffers when necessary. It was getting this wrong when the
2891        buffer for a data line had to be extended.
2892    
2893    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2894        CRLF as a newline sequence.
2895    
2896    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2897        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2898        I have nevertheless tidied it up.
2899    
2900    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2901    
2902    24. Added a man page for pcre-config.
2903    
2904    
2905    Version 7.0 19-Dec-06
2906    ---------------------
2907    
2908     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2909        moving to gcc 4.1.1.
2910    
2911     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2912        sys/time.h, which is documented as needed for this function. It doesn't
2913        seem to matter on Linux, but it showed up on some releases of OS X.
2914    
2915     3. It seems that there are systems where bytes whose values are greater than
2916        127 match isprint() in the "C" locale. The "C" locale should be the
2917        default when a C program starts up. In most systems, only ASCII printing
2918        characters match isprint(). This difference caused the output from pcretest
2919        to vary, making some of the tests fail. I have changed pcretest so that:
2920    
2921        (a) When it is outputting text in the compiled version of a pattern, bytes
2922            other than 32-126 are always shown as hex escapes.
2923    
2924        (b) When it is outputting text that is a matched part of a subject string,
2925            it does the same, unless a different locale has been set for the match
2926            (using the /L modifier). In this case, it uses isprint() to decide.
2927    
2928     4. Fixed a major bug that caused incorrect computation of the amount of memory
2929        required for a compiled pattern when options that changed within the
2930        pattern affected the logic of the preliminary scan that determines the
2931        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2932        that the computed length was too small. The symptoms of this bug were
2933        either the PCRE error "internal error: code overflow" from pcre_compile(),
2934        or a glibc crash with a message such as "pcretest: free(): invalid next
2935        size (fast)". Examples of patterns that provoked this bug (shown in
2936        pcretest format) are:
2937    
2938          /(?-x: )/x
2939          /(?x)(?-x: \s*#\s*)/
2940          /((?i)[\x{c0}])/8
2941          /(?i:[\x{c0}])/8
2942    
2943        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2944        is now done differently.
2945    
2946     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2947        wrapper classes; (b) implement a new function in the C++ scanner that is
2948        more efficient than the old way of doing things because it avoids levels of
2949        recursion in the regex matching; (c) add a paragraph to the documentation
2950        for the FullMatch() function.
2951    
2952     6. The escape sequence \n was being treated as whatever was defined as
2953        "newline". Not only was this contrary to the documentation, which states
2954        that \n is character 10 (hex 0A), but it also went horribly wrong when
2955        "newline" was defined as CRLF. This has been fixed.
2956    
2957     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2958        was being set to -1 for the "end of line" case (supposedly a value that no
2959        character can have). Though this value is never used (the check for end of
2960        line is "zero bytes in current character"), it caused compiler complaints.
2961        I've changed it to 0xffffffff.
2962    
2963     8. In pcre_version.c, the version string was being built by a sequence of
2964        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2965        string (as it is for production releases) called a macro with an empty
2966        argument. The C standard says the result of this is undefined. The gcc
2967        compiler treats it as an empty string (which was what was wanted) but it is
2968        reported that Visual C gives an error. The source has been hacked around to
2969        avoid this problem.
2970    
2971     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2972        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2973        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2974        of them did).
2975    
2976    10. Originally, pcretest opened its input and output without "b"; then I was
2977        told that "b" was needed in some environments, so it was added for release
2978        5.0 to both the input and output. (It makes no difference on Unix-like
2979        systems.) Later I was told that it is wrong for the input on Windows. I've
2980        now abstracted the modes into two macros, to make it easier to fiddle with
2981        them, and removed "b" from the input mode under Windows.
2982    
2983    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2984    
2985    12. Added -help and --help to pcretest as an official way of being reminded
2986        of the options.
2987    
2988    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2989        and pcrecpp.cc because they annoy compilers at high warning levels.
2990    
2991    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2992    
2993    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2994        scripts are not C programs :-) and which was not noticed because it works
2995        on Linux.
2996    
2997    16. pcretest is supposed to handle any length of pattern and data line (as one
2998        line or as a continued sequence of lines) by extending its input buffer if
2999        necessary. This feature was broken for very long pattern lines, leading to
3000        a string of junk being passed to pcre_compile() if the pattern was longer
3001        than about 50K.
3002    
3003    17. I have done a major re-factoring of the way pcre_compile() computes the
3004        amount of memory needed for a compiled pattern. Previously, there was code
3005        that made a preliminary scan of the pattern in order to do this. That was
3006        OK when PCRE was new, but as the facilities have expanded, it has become
3007        harder and harder to keep it in step with the real compile phase, and there
3008        have been a number of bugs (see for example, 4 above). I have now found a
3009        cunning way of running the real compile function in a "fake" mode that
3010        enables it to compute how much memory it would need, while actually only
3011        ever using a few hundred bytes of working memory and without too many
3012        tests of the mode. This should make future maintenance and development
3013        easier. A side effect of this work is that the limit of 200 on the nesting
3014        depth of parentheses has been removed (though this was never a serious
3015        limitation, I suspect). However, there is a downside: pcre_compile() now
3016        runs more slowly than before (30% or more, depending on the pattern). I
3017        hope this isn't a big issue. There is no effect on runtime performance.
3018    
3019    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3020        newline (only possible for the last line of a file) and it was a
3021        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3022    
3023    19. Added additional timing features to pcretest. (1) The -tm option now times
3024        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3025        separate command line item, by a number that specifies the number of
3026        repeats to use when timing. The default is 50000; this gives better
3027        precision, but takes uncomfortably long for very large patterns.
3028    
3029    20. Extended pcre_study() to be more clever in cases where a branch of a
3030        subpattern has no definite first character. For example, (a*|b*)[cd] would
3031        previously give no result from pcre_study(). Now it recognizes that the
3032        first character must be a, b, c, or d.
3033    
3034    21. There was an incorrect error "recursive call could loop indefinitely" if
3035        a subpattern (or the entire pattern) that was being tested for matching an
3036        empty string contained only one non-empty item after a nested subpattern.
3037        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3038        incorrectly, because the \d was being skipped in the check.
3039    
3040    22. The pcretest program now has a new pattern option /B and a command line
3041        option -b, which is equivalent to adding /B to every pattern. This causes
3042        it to show the compiled bytecode, without the additional information that
3043        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3044        is the same as /B/I).
3045    
3046    23. A new optimization is now able automatically to treat some sequences such
3047        as a*b as a*+b. More specifically, if something simple (such as a character
3048        or a simple class like \d) has an unlimited quantifier, and is followed by
3049        something that cannot possibly match the quantified thing, the quantifier
3050        is automatically "possessified".
3051    
3052    24. A recursive reference to a subpattern whose number was greater than 39
3053        went wrong under certain circumstances in UTF-8 mode. This bug could also
3054        have affected the operation of pcre_study().
3055    
3056    25. Realized that a little bit of performance could be had by replacing
3057        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3058    
3059    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3060    
3061    27. Possessive quantifiers such as a++ were previously implemented by turning
3062        them into atomic groups such as ($>a+). Now they have their own opcodes,
3063        which improves performance. This includes the automatically created ones
3064        from 23 above.
3065    
3066    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3067        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3068        the first matched character to be a colon. This applied both to named and
3069        numbered groups.
3070    
3071    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3072    
3073    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3074        building PCRE on Windows easier, so I have included it in the distribution.
3075    
3076    31. There is now a check in pcretest against a ridiculously large number being
3077        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3078        loop, the loop is abandoned.
3079    
3080    32. Forward references to subpatterns in conditions such as (?(2)...) where
3081        subpattern 2 is defined later cause pcre_compile() to search forwards in
3082        the pattern for the relevant set of parentheses. This search went wrong
3083        when there were unescaped parentheses in a character class, parentheses
3084        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3085    
3086    33. "Subroutine" calls and backreferences were previously restricted to
3087        referencing subpatterns earlier in the regex. This restriction has now
3088        been removed.
3089    
3090    34. Added a number of extra features that are going to be in Perl 5.10. On the
3091        whole, these are just syntactic alternatives for features that PCRE had
3092        previously implemented using the Python syntax or my own invention. The
3093        other formats are all retained for compatibility.
3094    
3095        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3096            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3097            also .NET compatible.
3098    
3099        (b) A recursion or subroutine call to a named group can now be defined as
3100            (?&name) as well as (?P>name).
3101    
3102        (c) A backreference to a named group can now be defined as \k<name> or
3103            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3104            5.10, are also .NET compatible.
3105    
3106        (d) A conditional reference to a named group can now use the syntax
3107            (?(<name>) or (?('name') as well as (?(name).
3108    
3109        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3110            groups (named and numbered) that are never evaluated inline, but can be
3111            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3112            is always false. There may be only one alternative in such a group.
3113    
3114        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3115            as the simple (?(R). The condition is true only if the most recent
3116            recursion is that of the given number or name. It does not search out
3117            through the entire recursion stack.
3118    
3119        (g) The escape \gN or \g{N} has been added, where N is a positive or
3120            negative number, specifying an absolute or relative reference.
3121    
3122    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3123        some "unreachable code" warnings.
3124    
3125    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3126        things, this adds five new scripts.
3127    
3128    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3129        There were also incompatibilities regarding the handling of \Q..\E inside
3130        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3131        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3132    
3133    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3134        matches an empty string, and forcibly breaks the loop. There were bugs in
3135        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3136        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3137        separate and independent bugs (that affected different cases) have been
3138        fixed.
3139    
3140    39. Refactored the code to abolish the use of different opcodes for small
3141        capturing bracket numbers. This is a tidy that I avoided doing when I
3142        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3143        The new approach is not only tidier, it makes it possible to reduce the
3144        memory needed to fix the previous bug (38).
3145    
3146    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3147        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3148        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3149        mode.
3150    
3151    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3152        report.
3153    
3154    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3155        copy construction and assignment in the C++ wrapper.
3156    
3157    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3158        couple of bugs that could have given wrong results in the "--newline=crlf"
3159        case.
3160    
3161    44. Added a number of casts and did some reorganization of signed/unsigned int
3162        variables following suggestions from Dair Grant. Also renamed the variable
3163        "this" as "item" because it is a C++ keyword.
3164    
3165    45. Arranged for dftables to add
3166    
3167          #include "pcre_internal.h"
3168    
3169        to pcre_chartables.c because without it, gcc 4.x may remove the array
3170        definition from the final binary if PCRE is built into a static library and
3171        dead code stripping is activated.
3172    
3173    46. For an unanchored pattern, if a match attempt fails at the start of a
3174        newline sequence, and the newline setting is CRLF or ANY, and the next two
3175        characters are CRLF, advance by two characters instead of one.
3176    
3177    
3178    Version 6.7 04-Jul-06
3179    ---------------------
3180    
3181     1. In order to handle tests when input lines are enormously long, pcretest has
3182        been re-factored so that it automatically extends its buffers when
3183        necessary. The code is crude, but this _is_ just a test program. The
3184        default size has been increased from 32K to 50K.
3185    
3186     2. The code in pcre_study() was using the value of the re argument before
3187        testing it for NULL. (Of course, in any sensible call of the function, it
3188        won't be NULL.)
3189    
3190     3. The memmove() emulation function in pcre_internal.h, which is used on
3191        systems that lack both memmove() and bcopy() - that is, hardly ever -
3192        was missing a "static" storage class specifier.
3193    
3194     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3195        containing an extended class (one that cannot be represented by a bitmap
3196        because it contains high-valued characters or Unicode property items, e.g.
3197        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3198        pattern, but PCRE should not loop if you do not (it no longer does).
3199        [Detail: two cases were found: (a) a repeated subpattern containing an
3200        extended class; (b) a recursive reference to a subpattern that followed a
3201        previous extended class. It wasn't skipping over the extended class
3202        correctly when UTF-8 mode was not set.]
3203    
3204     5. A negated single-character class was not being recognized as fixed-length
3205        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3206        compile error "lookbehind assertion is not fixed length".
3207    
3208     6. The RunPerlTest auxiliary script was showing an unexpected difference
3209        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3210        write a Perl script that can interpret lines of an input file either as
3211        byte characters or as UTF-8, which is what "perltest" was being required to
3212        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3213        can't do is switch easily at run time between having the "use utf8;" pragma
3214        or not. In the end, I fudged it by using the RunPerlTest script to insert
3215        "use utf8;" explicitly for the UTF-8 tests.
3216    
3217     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3218        the end of the subject string, contrary to the documentation and to what
3219        Perl does. This was true of both matching functions. Now it matches only at
3220        the start of the subject and immediately after *internal* newlines.
3221    
3222     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3223        a pointer to an int instead of a pointer to an unsigned long int. This
3224        caused problems on 64-bit systems.
3225    
3226     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3227        instance of the 'standard' template library not being so standard".
3228    
3229    10. There was no check on the number of named subpatterns nor the maximum
3230        length of a subpattern name. The product of these values is used to compute
3231        the size of the memory block for a compiled pattern. By supplying a very
3232        long subpattern name and a large number of named subpatterns, the size
3233        computation could be caused to overflow. This is now prevented by limiting
3234        the length of names to 32 characters, and the number of named subpatterns
3235        to 10,000.
3236    
3237    11. Subpatterns that are repeated with specific counts have to be replicated in
3238        the compiled pattern. The size of memory for this was computed from the
3239        length of the subpattern and the repeat count. The latter is limited to
3240        65535, but there was no limit on the former, meaning that integer overflow
3241        could in principle occur. The compiled length of a repeated subpattern is
3242        now limited to 30,000 bytes in order to prevent this.
3243    
3244    12. Added the optional facility to have named substrings with the same name.
3245    
3246    13. Added the ability to use a named substring as a condition, using the
3247        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3248        are numbers (not recommended). Forward references are permitted.
3249    
3250    14. Added forward references in named backreferences (if you see what I mean).
3251    
3252    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3253        pattern could run off the end of the subject. For example, the pattern
3254        "(?s)(.{1,5})"8 did this with the subject "ab".
3255    
3256    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3257        PCRE_CASELESS was set when matching characters that were quantified with ?
3258        or *.
3259    
3260    17. A character class other than a single negated character that had a minimum
3261        but no maximum quantifier - for example [ab]{6,} - was not handled
3262        correctly by pce_dfa_exec(). It would match only one character.
3263    
3264    18. A valid (though odd) pattern that looked like a POSIX character
3265        class but used an invalid character after [ (for example [[,abc,]]) caused
3266        pcre_compile() to give the error "Failed: internal error: code overflow" or
3267        in some cases to crash with a glibc free() error. This could even happen if
3268        the pattern terminated after [[ but there just happened to be a sequence of
3269        letters, a binary zero, and a closing ] in the memory that followed.
3270    
3271    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3272        over the years. Originally (before any Unicode support), just the bottom 8
3273        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3274        output from "man perlunicode" includes this:
3275    
3276          The regular expression compiler produces polymorphic opcodes.  That
3277          is, the pattern adapts to the data and automatically switches to
3278          the Unicode character scheme when presented with Unicode data--or
3279          instead uses a traditional byte scheme when presented with byte
3280          data.
3281    
3282        Sadly, a wide octal escape does not cause a switch, and in a string with
3283        no other multibyte characters, these octal escapes are treated as before.
3284        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3285        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3286        Unicode string.
3287    
3288        I have not perpetrated such confusion in PCRE. Up till now, it took just
3289        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3290        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3291        translate to the appropriate multibyte character.
3292    
3293    29. Applied some refactoring to reduce the number of warnings from Microsoft
3294        and Borland compilers. This has included removing the fudge introduced
3295        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3296        a warning about an unused variable.
3297    
3298    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3299        characters since release 4.0, because Perl (from release 5.004) does not.
3300        [Or at least, is documented not to: some releases seem to be in conflict
3301        with the documentation.] However, when a pattern was studied with
3302        pcre_study() and all its branches started with \s, PCRE still included VT
3303        as a possible starting character. Of course, this did no harm; it just
3304        caused an unnecessary match attempt.
3305    
3306    22. Removed a now-redundant internal flag bit that recorded the fact that case
3307        dependency changed within the pattern. This was once needed for "required
3308        byte" processing, but is no longer used. This recovers a now-scarce options
3309        bit. Also moved the least significant internal flag bit to the most-
3310        significant bit of the word, which was not previously used (hangover from
3311        the days when it was an int rather than a uint) to free up another bit for
3312        the future.
3313    
3314    23. Added support for CRLF line endings as well as CR and LF. As well as the
3315        default being selectable at build time, it can now be changed at runtime
3316        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3317        specify that it is scanning data with non-default line endings.
3318    
3319    24. Changed the definition of CXXLINK to make it agree with the definition of
3320        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3321    
3322    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3323        recursions. This makes a big different to stack usage for some patterns.
3324    
3325    26. If a subpattern containing a named recursion or subroutine reference such
3326        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3327        the space required for the compiled pattern went wrong and gave too small a
3328        value. Depending on the environment, this could lead to "Failed: internal
3329        error: code overflow at offset 49" or "glibc detected double free or
3330        corruption" errors.
3331    
3332    27. Applied patches from Google (a) to support the new newline modes and (b) to
3333        advance over multibyte UTF-8 characters in GlobalReplace.
3334    
3335    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3336        difference for some implementation of PCRE in some Windows version.
3337    
3338    29. Added some extra testing facilities to pcretest:
3339    
3340        \q<number>   in a data line sets the "match limit" value
3341        \Q<number>   in a data line sets the "match recursion limt" value
3342        -S <number>  sets the stack size, where <number> is in megabytes
3343    
3344        The -S option isn't available for Windows.
3345    
3346    
3347    Version 6.6 06-Feb-06
3348    ---------------------
3349    
3350     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3351        in pcreposix.h. I have copied the definition from pcre.h.
3352    
3353     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3354        because pcre.h is no longer a built file.
3355    
3356     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3357        not normally included in the compiled code.
3358    
3359    
3360    Version 6.5 01-Feb-06
3361    ---------------------
3362    
3363     1. When using the partial match feature with pcre_dfa_exec(), it was not
3364        anchoring the second and subsequent partial matches at the new starting
3365        point. This could lead to incorrect results. For example, with the pattern
3366        /1234/, partially matching against "123" and then "a4" gave a match.
3367    
3368     2. Changes to pcregrep:
3369    
3370        (a) All non-match returns from pcre_exec() were being treated as failures
3371            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3372            error message is output. Some extra information is given for the
3373            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3374            probably the only errors that are likely to be caused by users (by
3375            specifying a regex that has nested indefinite repeats, for instance).
3376            If there are more than 20 of these errors, pcregrep is abandoned.
3377    
3378        (b) A binary zero was treated as data while matching, but terminated the
3379            output line if it was written out. This has been fixed: binary zeroes
3380            are now no different to any other data bytes.
3381    
3382        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3383            used to set a locale for matching. The --locale=xxxx long option has
3384            been added (no short equivalent) to specify a locale explicitly on the
3385            pcregrep command, overriding the environment variables.
3386    
3387        (d) When -B was used with -n, some line numbers in the output were one less
3388            than they should have been.
3389    
3390        (e) Added the -o (--only-matching) option.
3391    
3392        (f) If -A or -C was used with -c (count only), some lines of context were
3393            accidentally printed for the final match.
3394    
3395        (g) Added the -H (--with-filename) option.
3396    
3397        (h) The combination of options -rh failed to suppress file names for files
3398            that were found from directory arguments.
3399    
3400        (i) Added the -D (--devices) and -d (--directories) options.
3401    
3402        (j) Added the -F (--fixed-strings) option.
3403    
3404        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3405    
3406        (l) Added the --colo(u)r option.
3407    
3408        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3409            is not present by default.
3410    
3411     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3412        items such as (?R) or (?1), when the recursion could match a number of
3413        alternatives. If it matched one of the alternatives, but subsequently,
3414        outside the recursion, there was a failure, the code tried to back up into
3415        the recursion. However, because of the way PCRE is implemented, this is not
3416        possible, and the result was an incorrect result from the match.
3417    
3418        In order to prevent this happening, the specification of recursion has
3419        been changed so that all such subpatterns are automatically treated as
3420        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3421    
3422     4. I had overlooked the fact that, in some locales, there are characters for
3423        which isalpha() is true but neither isupper() nor islower() are true. In
3424        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3425        and ordfeminine) are like this. This affected the treatment of \w and \W
3426        when they appeared in character classes, but not when they appeared outside
3427        a character class. The bit map for "word" characters is now created
3428        separately from the results of isalnum() instead of just taking it from the
3429        upper, lower, and digit maps. (Plus the underscore character, of course.)
3430    
3431     5. The above bug also affected the handling of POSIX character classes such as
3432        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3433        permanent tables. Instead, the bit maps for such a class were previously
3434        created as the appropriate unions of the upper, lower, and digit bitmaps.
3435        Now they are created by subtraction from the [[:word:]] class, which has
3436        its own bitmap.
3437    
3438     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3439        It is created by subtracting the vertical space characters (\x09, \x0a,
3440        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3441        subtraction was done in the overall bitmap for a character class, meaning
3442        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3443        be recognized. This bug has been fixed.
3444    
3445     7. Patches from the folks at Google:
3446    
3447          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3448          real life, but is still worth protecting against".
3449    
3450          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3451          regular expressions".
3452    
3453          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3454          have it.
3455    
3456          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3457          "configure" and the latter not, in order to fix a problem somebody had
3458          with compiling the Arg class on HP-UX.
3459    
3460          (e) Improve the error-handling of the C++ wrapper a little bit.
3461    
3462          (f) New tests for checking recursion limiting.
3463    
3464     8. The pcre_memmove() function, which is used only if the environment does not
3465        have a standard memmove() function (and is therefore rarely compiled),
3466        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3467        returning a result (though PCRE never actually uses the result).
3468    
3469     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3470        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3471        returned instead of calling malloc() with an overflowing number that would
3472        most likely cause subsequent chaos.
3473    
3474    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3475    
3476    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3477        with this option is matched, the nmatch and pmatch options of regexec() are
3478        ignored.
3479    
3480    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3481        provided in case anyone wants to the the POSIX interface with UTF-8
3482        strings.
3483    
3484    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3485        C++ linking (needed for some HP-UX environments).
3486    
3487    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3488        (unused parameter) and in the pcre_printint() function (omitted "default"
3489        switch label when the default is to do nothing).
3490    
3491    15. Added some code to make it possible, when PCRE is compiled as a C++
3492        library, to replace subject pointers for pcre_exec() with a smart pointer
3493        class, thus making it possible to process discontinuous strings.
3494    
3495    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3496        much the same function. They were added by different people who were trying
3497        to make PCRE easy to compile on non-Unix systems. It has been suggested
3498        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3499        for compiling on Windows systems. I have therefore replaced it with
3500        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3501        defaults to "extern" for C or "extern C" for C++, which works fine on
3502        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3503        SCOPE with something explicit in config.h. In addition:
3504    
3505        (a) pcreposix.h still had just "extern" instead of either of these macros;
3506            I have replaced it with PCRE_DATA_SCOPE.
3507    
3508        (b) Functions such as _pcre_xclass(), which are internal to the library,
3509            but external in the C sense, all had PCRE_EXPORT in their definitions.
3510            This is apparently wrong for the Windows case, so I have removed it.
3511            (It makes no difference on Unix-like systems.)
3512    
3513    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3514        of recursive calls to match(). This is different to MATCH_LIMIT because
3515        that limits the total number of calls to match(), not all of which increase
3516        the depth of recursion. Limiting the recursion depth limits the amount of
3517        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3518        when PCRE is compiled, and changed at run time. A patch from Google adds
3519        this functionality to the C++ interface.
3520    
3521    18. Changes to the handling of Unicode character properties:
3522    
3523        (a) Updated the table to Unicode 4.1.0.
3524    
3525        (b) Recognize characters that are not in the table as "Cn" (undefined).
3526    
3527        (c) I revised the way the table is implemented to a much improved format
3528            which includes recognition of ranges. It now supports the ranges that
3529            are defined in UnicodeData.txt, and it also amalgamates other
3530            characters into ranges. This has reduced the number of entries in the
3531            table from around 16,000 to around 3,000, thus reducing its size
3532            considerably. I realized I did not need to use a tree structure after
3533            all - a binary chop search is just as efficient. Having reduced the
3534            number of entries, I extended their size from 6 bytes to 8 bytes to
3535            allow for more data.
3536    
3537        (d) Added support for Unicode script names via properties such as \p{Han}.
3538    
3539    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3540        matching that character.
3541    
3542    20. When matching a repeated Unicode property with a minimum greater than zero,
3543        (for example \pL{2,}), PCRE could look past the end of the subject if it
3544        reached it while seeking the minimum number of characters. This could
3545        happen only if some of the characters were more than one byte long, because
3546        there is a check for at least the minimum number of bytes.
3547    
3548    21. Refactored the implementation of \p and \P so as to be more general, to
3549        allow for more different types of property in future. This has changed the
3550        compiled form incompatibly. Anybody with saved compiled patterns that use
3551        \p or \P will have to recompile them.
3552    
3553    22. Added "Any" and "L&" to the supported property types.
3554    
3555    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3556        but give a compile time error if the value is greater than 0xff.
3557    
3558    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3559        accidentally not being installed or uninstalled.
3560    
3561    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3562        made were to insert the current release number. This seemed silly, because
3563        it made things harder for people building PCRE on systems that don't run
3564        "configure". I have turned pcre.h into a distributed file, no longer built
3565        by "configure", with the version identification directly included. There is
3566        no longer a pcre.h.in file.
3567    
3568        However, this change necessitated a change to the pcre-config script as
3569        well. It is built from pcre-config.in, and one of the substitutions was the
3570        release number. I have updated configure.ac so that ./configure now finds
3571        the release number by grepping pcre.h.
3572    
3573    26. Added the ability to run the tests under valgrind.
3574    
3575    
3576    Version 6.4 05-Sep-05
3577    ---------------------
3578    
3579     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3580        "--" to be printed when multiple files were scanned, even when none of the
3581        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3582        consider it to be a bug, and have restored the previous behaviour.
3583    
3584     2. A couple of code tidies to get rid of compiler warnings.
3585    
3586     3. The pcretest program used to cheat by referring to symbols in the library
3587        whose names begin with _pcre_. These are internal symbols that are not
3588        really supposed to be visible externally, and in some environments it is
3589        possible to suppress them. The cheating is now confined to including
3590        certain files from the library's source, which is a bit cleaner.
3591    
3592     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3593        file's purpose clearer.
3594    
3595     5. Reorganized pcre_ucp_findchar().
3596    
3597    
3598    Version 6.3 15-Aug-05
3599    ---------------------
3600    
3601     1. The file libpcre.pc.in did not have general read permission in the tarball.
3602    
3603     2. There were some problems when building without C++ support:
3604    
3605        (a) If C++ support was not built, "make install" and "make test" still
3606            tried to test it.
3607    
3608        (b) There were problems when the value of CXX was explicitly set. Some
3609            changes have been made to try to fix these, and ...
3610    
3611        (c) --disable-cpp can now be used to explicitly disable C++ support.
3612    
3613        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3614            backslash in a target when C++ was disabled. This confuses some
3615            versions of "make", apparently. Using an intermediate variable solves
3616            this. (Same for CPP_LOBJ.)
3617    
3618     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3619        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3620        necessary on certain architectures.
3621    
3622     4. Added a setting of -export-symbols-regex to the link command to remove
3623        those symbols that are exported in the C sense, but actually are local
3624        within the library, and not documented. Their names all begin with
3625        "_pcre_". This is not a perfect job, because (a) we have to except some
3626        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3627        available (and never for static libraries). I have made a note to try to
3628        find a way round (a) in the future.
3629    
3630    
3631    Version 6.2 01-Aug-05
3632    ---------------------
3633    
3634     1. There was no test for integer overflow of quantifier values. A construction
3635        such as {1111111111111111} would give undefined results. What is worse, if
3636        a minimum quantifier for a parenthesized subpattern overflowed and became
3637        negative, the calculation of the memory size went wrong. This could have
3638        led to memory overwriting.
3639    
3640     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3641    
3642     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3643        operating environments where this matters.
3644    
3645     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3646        PCRE options from within the C++ wrapper.
3647    
3648     5. Named capturing subpatterns were not being correctly counted when a pattern
3649        was compiled. This caused two problems: (a) If there were more than 100
3650        such subpatterns, the calculation of the memory needed for the whole
3651        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3652        back references of the form \12, where the number was greater than 9, were
3653        not recognized as back references, even though there were sufficient
3654        previous subpatterns.
3655    
3656     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3657        versions of gcc, e.g. 2.95.4.
3658    
3659    
3660    Version 6.1 21-Jun-05
3661    ---------------------
3662    
3663     1. There was one reference to the variable "posix" in pcretest.c that was not
3664        surrounded by "#if !defined NOPOSIX".
3665    
3666     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3667        the cross-check on the old pcre_info() function, for the benefit of the
3668        cut-down version of PCRE that is currently imported into Exim.
3669    
3670     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3671        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3672        patterns that start (?i) but I don't think that matters. The use of (?i) is
3673        just an example; this all applies to the other options as well.
3674    
3675     4. Since libtool seems to echo the compile commands it is issuing, the output
3676        from "make" can be reduced a bit by putting "@" in front of each libtool
3677        compile command.
3678    
3679     5. Patch from the folks at Google for configure.in to be a bit more thorough
3680        in checking for a suitable C++ installation before trying to compile the
3681        C++ stuff. This should fix a reported problem when a compiler was present,
3682        but no suitable headers.
3683    
3684     6. The man pages all had just "PCRE" as their title. I have changed them to
3685        be the relevant file name. I have also arranged that these names are
3686        retained in the file doc/pcre.txt, which is a concatenation in text format
3687        of all the man pages except the little individual ones for each function.
3688    
3689     7. The NON-UNIX-USE file had not been updated for the different set of source
3690        files that come with release 6. I also added a few comments about the C++
3691        wrapper.
3692    
3693    
3694    Version 6.0 07-Jun-05
3695    ---------------------
3696    
3697     1. Some minor internal re-organization to help with my DFA experiments.
3698    
3699     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3700        didn't matter for the library itself when fully configured, but did matter
3701        when compiling without UCP support, or within Exim, where the ucp files are
3702        not imported.
3703    
3704     3. Refactoring of the library code to split up the various functions into
3705        different source modules. The addition of the new DFA matching code (see
3706        below) to a single monolithic source would have made it really too
3707        unwieldy, quite apart from causing all the code to be include in a
3708        statically linked application, when only some functions are used. This is
3709        relevant even without the DFA addition now that patterns can be compiled in
3710        one application and matched in another.
3711    
3712        The downside of splitting up is that there have to be some external
3713        functions and data tables that are used internally in different modules of
3714        the library but which are not part of the API. These have all had their
3715        names changed to start with "_pcre_" so that they are unlikely to clash
3716        with other external names.
3717    
3718     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3719        a different (DFA) algorithm. Although it is slower than the original
3720        function, it does have some advantages for certain types of matching
3721        problem.
3722    
3723     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3724        including restarting after a partial match.
3725    
3726     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3727        defined when compiling for Windows was sent to me. I have put it into the
3728        code, though I have no means of testing or verifying it.
3729    
3730     7. Added the pcre_refcount() auxiliary function.
3731    
3732     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3733        match before or at the first newline in the subject string. In pcretest,
3734        the /f option on a pattern can be used to set this.
3735    
3736     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3737        would behave wrongly. This has been present in PCRE since release 4.0.
3738    
3739    10. A number of changes to the pcregrep command:
3740    
3741        (a) Refactored how -x works; insert ^(...)$ instead of setting
3742            PCRE_ANCHORED and checking the length, in preparation for adding
3743            something similar for -w.
3744    
3745        (b) Added the -w (match as a word) option.
3746    
3747        (c) Refactored the way lines are read and buffered so as to have more
3748            than one at a time available.
3749    
3750        (d) Implemented a pcregrep test script.
3751    
3752        (e) Added the -M (multiline match) option. This allows patterns to match
3753            over several lines of the subject. The buffering ensures that at least
3754            8K, or the rest of the document (whichever is the shorter) is available
3755            for matching (and similarly the previous 8K for lookbehind assertions).
3756    
3757        (f) Changed the --help output so that it now says
3758    
3759              -w, --word-regex(p)
3760    
3761            instead of two lines, one with "regex" and the other with "regexp"
3762            because that confused at least one person since the short forms are the
3763            same. (This required a bit of code, as the output is generated
3764            automatically from a table. It wasn't just a text change.)
3765    
3766        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3767            option but starts with a hyphen. Could be a pattern or a path name
3768            starting with a hyphen, for instance.
3769    
3770        (h) "-" can be given as a file name to represent stdin.
3771    
3772        (i) When file names are being printed, "(standard input)" is used for
3773            the standard input, for compatibility with GNU grep. Previously
3774            "<stdin>" was used.
3775    
3776        (j) The option --label=xxx can be used to supply a name to be used for
3777            stdin when file names are being printed. There is no short form.
3778    
3779        (k) Re-factored the options decoding logic because we are going to add
3780            two more options that take data. Such options can now be given in four
3781            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3782    
3783        (l) Added the -A, -B, and -C options for requesting that lines of context
3784            around matches be printed.
3785    
3786        (m) Added the -L option to print the names of files that do not contain
3787            any matching lines, that is, the complement of -l.
3788    
3789