ViewVC logotype

Diff of /code/trunk/ChangeLog

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

revision 61 by nigel, Sat Feb 24 21:39:58 2007 UTC revision 1459 by ph10, Tue Mar 4 10:45:15 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
4  Version 3.0 02-Jan-02  Version 8.35-RC1 04-March-2014
5    ------------------------------
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.
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.
15    3.  Got rid of some compiler warnings for potentially uninitialized variables
16        that show up only when compiled with -O2.
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.
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.
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.
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.
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).
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.
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.
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.
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.
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).
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.
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.
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.
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.
95    18. pcretest was missing "-C bsr" for displaying the \R default setting.
97    19. Little endian PowerPC systems are supported now by the JIT compiler.
99    20. The fast forward newline mechanism could enter to an infinite loop on
100        certain invalid UTF-8 input. Although we don't support these cases
101        this issue can be fixed by a performance optimization.
103    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
104        not take account if existing stack usage. There is now a new global
105        variable called pcre_stack_guard that can be set to point to an external
106        function to check stack availability. It is called at the start of
107        processing every parenthesized group.
110    Version 8.34 15-December-2013
111    -----------------------------
113    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
114        executable memory. Patch inspired by Carsten Klein.
116    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
117        this macro is never tested and has no effect, because the work to support
118        coverage involves only compiling and linking options and special targets in
119        the Makefile. The comment in config.h implied that defining the macro would
120        enable coverage support, which is totally false. There was also support for
121        setting this macro in the CMake files (my fault, I just copied it from
122        configure). SUPPORT_GCOV has now been removed.
124    3.  Make a small performance improvement in strlen16() and strlen32() in
125        pcretest.
127    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
128        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
129        these cases). There was also one in pcretest.c.
131    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
133    6.  In UTF mode, the code for checking whether a group could match an empty
134        string (which is used for indefinitely repeated groups to allow for
135        breaking an infinite loop) was broken when the group contained a repeated
136        negated single-character class with a character that occupied more than one
137        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
138        UTF-8 mode). The effect was undefined: the group might or might not be
139        deemed as matching an empty string, or the program might have crashed.
141    7.  The code for checking whether a group could match an empty string was not
142        recognizing that \h, \H, \v, \V, and \R must match a character.
144    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
145        an empty string. If it can, pcretest shows this in its information output.
147    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
148        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
149        matched by pcre_exec() without using JIT:
151        (a) If the rest of the pattern did not match after a maximal run of
152            grapheme clusters, the code for backing up to try with fewer of them
153            did not always back up over a full grapheme when characters that do not
154            have the modifier quality were involved, e.g. Hangul syllables.
156        (b) If the match point in a subject started with modifier character, and
157            there was no match, the code could incorrectly back up beyond the match
158            point, and potentially beyond the first character in the subject,
159            leading to a segfault or an incorrect match result.
161    10. A conditional group with an assertion condition could lead to PCRE
162        recording an incorrect first data item for a match if no other first data
163        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
164        first data item, and therefore matched "ca" after "c" instead of at the
165        start.
167    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
168        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
170    12. The source of pcregrep now includes z/OS-specific code so that it can be
171        compiled for z/OS as part of the special z/OS distribution.
173    13. Added the -T and -TM options to pcretest.
175    14. The code in pcre_compile.c for creating the table of named capturing groups
176        has been refactored. Instead of creating the table dynamically during the
177        actual compiling pass, the information is remembered during the pre-compile
178        pass (on the stack unless there are more than 20 named groups, in which
179        case malloc() is used) and the whole table is created before the actual
180        compile happens. This has simplified the code (it is now nearly 150 lines
181        shorter) and prepared the way for better handling of references to groups
182        with duplicate names.
184    15. A back reference to a named subpattern when there is more than one of the
185        same name now checks them in the order in which they appear in the pattern.
186        The first one that is set is used for the reference. Previously only the
187        first one was inspected. This change makes PCRE more compatible with Perl.
189    16. Unicode character properties were updated from Unicode 6.3.0.
191    17. The compile-time code for auto-possessification has been refactored, based
192        on a patch by Zoltan Herczeg. It now happens after instead of during
193        compilation. The code is cleaner, and more cases are handled. The option
194        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
195        options in pcretest are provided to set it. It can also be set by
196        (*NO_AUTO_POSSESS) at the start of a pattern.
198    18. The character VT has been added to the default ("C" locale) set of
199        characters that match \s and are generally treated as white space,
200        following this same change in Perl 5.18. There is now no difference between
201        "Perl space" and "POSIX space". Whether VT is treated as white space in
202        other locales depends on the locale.
204    19. The code for checking named groups as conditions, either for being set or
205        for being recursed, has been refactored (this is related to 14 and 15
206        above). Processing unduplicated named groups should now be as fast at
207        numerical groups, and processing duplicated groups should be faster than
208        before.
210    20. Two patches to the CMake build system, by Alexander Barkov:
212          (1) Replace the "source" command by "." in CMakeLists.txt because
213              "source" is a bash-ism.
215          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
216              without these the CMake build does not work on Solaris.
218    21. Perl has changed its handling of \8 and \9. If there is no previously
219        encountered capturing group of those numbers, they are treated as the
220        literal characters 8 and 9 instead of a binary zero followed by the
221        literals. PCRE now does the same.
223    22. Following Perl, added \o{} to specify codepoints in octal, making it
224        possible to specify values greater than 0777 and also making them
225        unambiguous.
227    23. Perl now gives an error for missing closing braces after \x{... instead of
228        treating the string as literal. PCRE now does the same.
230    24. RunTest used to grumble if an inappropriate test was selected explicitly,
231        but just skip it when running all tests. This make it awkward to run ranges
232        of tests when one of them was inappropriate. Now it just skips any
233        inappropriate tests, as it always did when running all tests.
235    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
236        character types such as \d or \w, too many callouts were inserted, and the
237        data that they returned was rubbish.
239    26. In UCP mode, \s was not matching two of the characters that Perl matches,
240        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
241        were matched by \h. The code has now been refactored so that the lists of
242        the horizontal and vertical whitespace characters used for \h and \v (which
243        are defined only in one place) are now also used for \s.
245    27. Add JIT support for the 64 bit TileGX architecture.
246        Patch by Jiong Wang (Tilera Corporation).
248    28. Possessive quantifiers for classes (both explicit and automatically
249        generated) now use special opcodes instead of wrapping in ONCE brackets.
251    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
252        (because it's meaningless), this was not happening when PCRE_CASELESS was
253        set. Not wrong, but inefficient.
255    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
256        properties for \w, \d, etc) is present in a test regex. Otherwise if the
257        test contains no characters greater than 255, Perl doesn't realise it
258        should be using Unicode semantics.
260    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
261        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
262        does in Unicode mode.
264    32. Added the "forbid" facility to pcretest so that putting tests into the
265        wrong test files can sometimes be quickly detected.
267    33. There is now a limit (default 250) on the depth of nesting of parentheses.
268        This limit is imposed to control the amount of system stack used at compile
269        time. It can be changed at build time by --with-parens-nest-limit=xxx or
270        the equivalent in CMake.
272    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
273        errors. Perl warns for these when in warning mode, but PCRE has no facility
274        for giving warnings.
276    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
277        However, this was not working for (?!) because it is optimized to (*FAIL),
278        for which PCRE does not allow quantifiers. The optimization is now disabled
279        when a quantifier follows (?!). I can't see any use for this, but it makes
280        things uniform.
282    36. Perl no longer allows group names to start with digits, so I have made this
283        change also in PCRE. It simplifies the code a bit.
285    37. In extended mode, Perl ignores spaces before a + that indicates a
286        possessive quantifier. PCRE allowed a space before the quantifier, but not
287        before the possessive +. It now does.
289    38. The use of \K (reset reported match start) within a repeated possessive
290        group such as (a\Kb)*+ was not working.
292    40. Document that the same character tables must be used at compile time and
293        run time, and that the facility to pass tables to pcre_exec() and
294        pcre_dfa_exec() is for use only with saved/restored patterns.
296    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
297        features for Builds with MSVC:
299        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
300           builds). The .rc files can be used to set FileDescription and many other
301           attributes.
303        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
304           This allows higher-level build scripts which want .pdb files to avoid
305           hard-coding the exact files needed."
307    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
308        mean "start of word" and "end of word", respectively, as a transition aid.
310    43. A minimizing repeat of a class containing codepoints greater than 255 in
311        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
312        compiled to use the heap for recursion.
314    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
315        is configured.
318    Version 8.33 28-May-2013
319    ------------------------
321    1.  Added 'U' to some constants that are compared to unsigned integers, to
322        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
323        variables that are added to signed variables, to ensure the result is
324        signed and can be negated.
326    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
327        CMake config files.
329    3.  Revise the creation of config.h.generic so that all boolean macros are
330        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
331        overriding via -D on the command line possible.
333    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
334        to unsigned int is reported to make a quite noticeable speed difference in
335        a specific Windows environment. Testing on Linux did also appear to show
336        some benefit (and it is clearly not harmful). Also fixed the definition of
337        Xop which should be unsigned.
339    5.  Related to (4), changing the definition of the intermediate variable cc
340        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
341        improvements.
343    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
344        unnecessary spaces.
346    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
347        and later.
349    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
351    9.  Optimizing fast_forward_start_bits in JIT.
353    10. Adding support for callouts in JIT, and fixing some issues revealed
354        during this work. Namely:
356        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
358        (b) Minimum length was not checked before the matching is started.
360    11. The value of capture_last that is passed to callouts was incorrect in some
361        cases when there was a capture on one path that was subsequently abandoned
362        after a backtrack. Also, the capture_last value is now reset after a
363        recursion, since all captures are also reset in this case.
365    12. The interpreter no longer returns the "too many substrings" error in the
366        case when an overflowing capture is in a branch that is subsequently
367        abandoned after a backtrack.
369    13. In the pathological case when an offset vector of size 2 is used, pcretest
370        now prints out the matched string after a yield of 0 or 1.
372    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
373        Only supported by the JIT compiler at the moment.
375    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
377    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
378        offset of the starting point of the matching process, provided the offsets
379        vector is large enough.
381    17. The \A escape now records a lookbehind value of 1, though its execution
382        does not actually inspect the previous character. This is to ensure that,
383        in partial multi-segment matching, at least one character from the old
384        segment is retained when a new segment is processed. Otherwise, if there
385        are no lookbehinds in the pattern, \A might match incorrectly at the start
386        of a new segment.
388    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
390    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
391        gives some modest performance improvement in 8-bit mode.
393    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
394        be expressed in certain programming languages using Universal Character
395        Names.
397    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
398        which points out that "non characters" are not "characters that may not
399        appear in Unicode strings" but rather "characters that are reserved for
400        internal use and have only local meaning".
402    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
403        there was a conditional group that depended on an assertion, if the
404        assertion was false, the callout that immediately followed the alternation
405        in the condition was skipped when pcre_exec() was used for matching.
407    23. Allow an explicit callout to be inserted before an assertion that is the
408        condition for a conditional group, for compatibility with automatic
409        callouts, which always insert a callout at this point.
411    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
412        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
414    25. (*PRUNE) is now supported by the JIT compiler.
416    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
418    27. Fix the case where there are two or more SKIPs with arguments that may be
419        ignored.
421    28. (*SKIP) is now supported by the JIT compiler.
423    29. (*THEN) is now supported by the JIT compiler.
425    30. Update RunTest with additional test selector options.
427    31. The way PCRE handles backtracking verbs has been changed in two ways.
429        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
430        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
431        backtracking. In some cases this makes it more Perl-compatible, but Perl's
432        rather obscure rules do not always do the same thing.
434        (2) Previously, backtracking verbs were confined within assertions. This is
435        no longer the case for positive assertions, except for (*ACCEPT). Again,
436        this sometimes improves Perl compatibility, and sometimes does not.
438    32. A number of tests that were in test 2 because Perl did things differently
439        have been moved to test 1, because either Perl or PCRE has changed, and
440        these tests are now compatible.
442    32. Backtracking control verbs are now handled in the same way in JIT and
443        interpreter.
445    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
446        contained a forward subroutine reference caused a compile error.
448    34. Auto-detect and optimize limited repetitions in JIT.
450    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
451        blocking (*UTF) etc.
453    36. In the interpreter, maximizing pattern repetitions for characters and
454        character types now use tail recursion, which reduces stack usage.
456    37. The value of the max lookbehind was not correctly preserved if a compiled
457        and saved regex was reloaded on a host of different endianness.
459    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
460        of the compiled pattern block, expand the flags field from 16 to 32 bits
461        because it was almost full.
463    39. Try madvise first before posix_madvise.
465    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
466        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
467        original change locked out all matches of empty strings. This has been
468        changed so that one match of an empty string per line is recognized.
469        Subsequent searches on the same line (for colouring or for --only-matching,
470        for example) do not recognize empty strings.
472    41. Applied a user patch to fix a number of spelling mistakes in comments.
474    42. Data lines longer than 65536 caused pcretest to crash.
476    43. Clarified the data type for length and startoffset arguments for pcre_exec
477        and pcre_dfa_exec in the function-specific man pages, where they were
478        explicitly stated to be in bytes, never having been updated. I also added
479        some clarification to the pcreapi man page.
481    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
482        a segmentation fault.
485    Version 8.32 30-November-2012
486    -----------------------------
488    1.  Improved JIT compiler optimizations for first character search and single
489        character iterators.
491    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
492        Patch by Daniel Richard G.
494    3.  Single character iterator optimizations in the JIT compiler.
496    4.  Improved JIT compiler optimizations for character ranges.
498    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
499        Reported by Giuseppe D'Angelo.
501    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
502        of a line, was being set incorrectly in cases where .* appeared inside
503        atomic brackets at the start of a pattern, or where there was a subsequent
504        *PRUNE or *SKIP.
506    7.  Improved instruction cache flush for POWER/PowerPC.
507        Patch by Daniel Richard G.
509    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
510        grep:
512        (a) There is now no limit to the number of patterns to be matched.
514        (b) An error is given if a pattern is too long.
516        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
517            are now supported.
519        (d) --exclude-from and --include-from (multiple use) have been added.
521        (e) Exclusions and inclusions now apply to all files and directories, not
522            just to those obtained from scanning a directory recursively.
524        (f) Multiple uses of -f and --file-list are now supported.
526        (g) In a Windows environment, the default for -d has been changed from
527            "read" (the GNU grep default) to "skip", because otherwise the presence
528            of a directory in the file list provokes an error.
530        (h) The documentation has been revised and clarified in places.
532    9.  Improve the matching speed of capturing brackets.
534    10. Changed the meaning of \X so that it now matches a Unicode extended
535        grapheme cluster.
537    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
538        out POSIX threads when JIT support is configured.
540    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
542    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
543        structure could go wrong in environments where size_t is not the same size
544        as int.
546    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
547        set.
549    15. The EBCDIC support had decayed; later updates to the code had included
550        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
551        general tidy up of EBCDIC-related issues, and the documentation was also
552        not quite right. There is now a test that can be run on ASCII systems to
553        check some of the EBCDIC-related things (but is it not a full test).
555    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
556        in a small tidy to the code.
558    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
560    18. If the --only-matching (-o) option in pcregrep is specified multiple
561        times, each one causes appropriate output. For example, -o1 -o2 outputs the
562        substrings matched by the 1st and 2nd capturing parentheses. A separating
563        string can be specified by --om-separator (default empty).
565    19. Improving the first n character searches.
567    20. Turn case lists for horizontal and vertical white space into macros so that
568        they are defined only once.
570    21. This set of changes together give more compatible Unicode case-folding
571        behaviour for characters that have more than one other case when UCP
572        support is available.
574        (a) The Unicode property table now has offsets into a new table of sets of
575            three or more characters that are case-equivalent. The MultiStage2.py
576            script that generates these tables (the pcre_ucd.c file) now scans
577            CaseFolding.txt instead of UnicodeData.txt for character case
578            information.
580        (b) The code for adding characters or ranges of characters to a character
581            class has been abstracted into a generalized function that also handles
582            case-independence. In UTF-mode with UCP support, this uses the new data
583            to handle characters with more than one other case.
585        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
586            whose other case is greater than 256 are now correctly matched
587            caselessly. Previously, the high codepoint matched the low one, but not
588            vice versa.
590        (d) The processing of \h, \H, \v, and \ in character classes now makes use
591            of the new class addition function, using character lists defined as
592            macros alongside the case definitions of 20 above.
594        (e) Caseless back references now work with characters that have more than
595            one other case.
597        (f) General caseless matching of characters with more than one other case
598            is supported.
600    22. Unicode character properties were updated from Unicode 6.2.0
602    23. Improved CMake support under Windows. Patch by Daniel Richard G.
604    24. Add support for 32-bit character strings, and UTF-32
606    25. Major JIT compiler update (code refactoring and bugfixing).
607        Experimental Sparc 32 support is added.
609    26. Applied a modified version of Daniel Richard G's patch to create
610        pcre.h.generic and config.h.generic by "make" instead of in the
611        PrepareRelease script.
613    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
614        pcre_compile.c when checking for a zero character.
616    28. Introducing a native interface for JIT. Through this interface, the compiled
617        machine code can be directly executed. The purpose of this interface is to
618        provide fast pattern matching, so several sanity checks are not performed.
619        However, feature tests are still performed. The new interface provides
620        1.4x speedup compared to the old one.
622    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
623        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
624        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
626    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
627        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
628        the "old" RFC 2279). Instead, it ended up passing a negative length to
629        pcre_exec().
631    31. Add support for GCC's visibility feature to hide internal functions.
633    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
634        "unknown -C option" after outputting 0 or 1.
636    33. There is now support for generating a code coverage report for the test
637        suite in environments where gcc is the compiler and lcov is installed. This
638        is mainly for the benefit of the developers.
640    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
641        unaddressable using valgrind annotations, allowing valgrind to detect
642        invalid memory accesses. This is mainly for the benefit of the developers.
644    25. (*UTF) can now be used to start a pattern in any of the three libraries.
646    26. Give configure error if --enable-cpp but no C++ compiler found.
649    Version 8.31 06-July-2012
650    -------------------------
652    1.  Fixing a wrong JIT test case and some compiler warnings.
654    2.  Removed a bashism from the RunTest script.
656    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
657        to unsigned type, result still unsigned" that was given by an MS compiler
658        on encountering the code "-sizeof(xxx)".
660    4.  Partial matching support is added to the JIT compiler.
662    5.  Fixed several bugs concerned with partial matching of items that consist
663        of more than one character:
665        (a) /^(..)\1/ did not partially match "aba" because checking references was
666            done on an "all or nothing" basis. This also applied to repeated
667            references.
669        (b) \R did not give a hard partial match if \r was found at the end of the
670            subject.
672        (c) \X did not give a hard partial match after matching one or more
673            characters at the end of the subject.
675        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
676            a partial match for the string "\r".
678        (e) When newline was set to CRLF, the metacharacter "." did not recognize
679            a partial match for a CR character at the end of the subject string.
681    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
682        running pcretest, the text "(JIT)" added to the output whenever JIT is
683        actually used to run the match.
685    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
686        or /S+[+] with a digit between 1 and 7.
688    8.  OP_NOT now supports any UTF character not just single-byte ones.
690    9.  (*MARK) control verb is now supported by the JIT compiler.
692    10. The command "./RunTest list" lists the available tests without actually
693        running any of them. (Because I keep forgetting what they all are.)
697    12. Applied a (slightly modified) user-supplied patch that improves performance
698        when the heap is used for recursion (compiled with --disable-stack-for-
699        recursion). Instead of malloc and free for each heap frame each time a
700        logical recursion happens, frames are retained on a chain and re-used where
701        possible. This sometimes gives as much as 30% improvement.
703    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
704        call.
706    14. As documented, (*COMMIT) is now confined to within a positive assertion.
708    15. It is now possible to link pcretest with libedit as an alternative to
709        libreadline.
711    16. (*COMMIT) control verb is now supported by the JIT compiler.
713    17. The Unicode data tables have been updated to Unicode 6.1.0.
715    18. Added --file-list option to pcregrep.
717    19. Added binary file support to pcregrep, including the -a, --binary-files,
718        -I, and --text options.
720    20. The madvise function is renamed for posix_madvise for QNX compatibility
721        reasons. Fixed by Giuseppe D'Angelo.
723    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
725    22. Changed the option for creating symbolic links for 16-bit man pages from
726        -s to -sf so that re-installing does not cause issues.
728    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
730    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
731        matches in certain environments (the workspace was not being correctly
732        retained). Also added to pcre_dfa_exec() a simple plausibility check on
733        some of the workspace data at the beginning of a restart.
735    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
736        was not doing so when it should - probably a typo introduced by SVN 528
737        (change 8.10/14).
739    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
740        \w+ when the character tables indicated that \x{c4} was a word character.
741        There were several related cases, all because the tests for doing a table
742        lookup were testing for characters less than 127 instead of 255.
744    27. If a pattern contains capturing parentheses that are not used in a match,
745        their slots in the ovector are set to -1. For those that are higher than
746        any matched groups, this happens at the end of processing. In the case when
747        there were back references that the ovector was too small to contain
748        (causing temporary malloc'd memory to be used during matching), and the
749        highest capturing number was not used, memory off the end of the ovector
750        was incorrectly being set to -1. (It was using the size of the temporary
751        memory instead of the true size.)
753    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
754        ovector size, it uses memory at the end of the block that it has got.
756    29. Check for an overlong MARK name and give an error at compile time. The
757        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
759    30. JIT compiler update.
761    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
762        Rill for the patch.
764    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
766    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
768    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
769        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
770        was enabled.
772    35. Improve JIT code generation for greedy plus quantifier.
774    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
775        1 to "aa" instead of to an empty string. The bug affected repeated groups
776        that could potentially match an empty string.
778    37. Optimizing single character iterators in JIT.
780    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
781        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
782        codepoints that are too big for the mode are faulted, and in a UTF mode,
783        disallowed codepoints are also faulted.
785    39. If PCRE was compiled with UTF support, in three places in the DFA
786        matcher there was code that should only have been obeyed in UTF mode, but
787        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
788        processing when bytes with values greater than 127 were present. In 16-bit
789        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
790        both cases the values are those that cannot be the first data item in a UTF
791        character. The three items that might have provoked this were recursions,
792        possessively repeated groups, and atomic groups.
794    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
795        and pcregrep, because some OS require shared objects to be explicitly
796        passed to ld, causing the link step to fail if they are not.
798    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
799        mode, patterns that started with \h* or \R* might be incorrectly matched.
802    Version 8.30 04-February-2012
803    -----------------------------
805    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
806        name is defined in ctype.h.
808    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
809        only in quite long subpatterns.
811    3.  Removed the function pcre_info(), which has been obsolete and deprecated
812        since it was replaced by pcre_fullinfo() in February 2000.
814    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
815        match a (*MARK), and the match failed at the start of the subject, a
816        reference to memory before the start of the subject could occur. This bug
817        was introduced by fix 17 of release 8.21.
819    5.  A reference to an unset group with zero minimum repetition was giving
820        totally wrong answers (in non-JavaScript-compatibility mode). For example,
821        /(another)?(\1?)test/ matched against "hello world test". This bug was
822        introduced in release 8.13.
824    6.  Add support for 16-bit character strings (a large amount of work involving
825        many changes and refactorings).
827    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
828        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
829        from a file.
831    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
832        rounding is not applied in this particular case).
834    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
835        if they appear, or are escaped, in patterns.
837    10. Get rid of a number of -Wunused-but-set-variable warnings.
839    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
840        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
841        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
842        also returns the mark "x". This bug applied to capturing parentheses,
843        non-capturing parentheses, and atomic parentheses. It also applied to some
844        assertions.
846    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
847        information out of configure.ac instead of relying on pcre.h.generic, which
848        is not stored in the repository.
850    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
851        -lreadline.
853    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
855    15. Applied Graycode's patch to put the top-level frame on the stack rather
856        than the heap when not using the stack for recursion. This gives a
857        performance improvement in many cases when recursion is not deep.
859    16. Experimental code added to "pcretest -C" to output the stack frame size.
862    Version 8.21 12-Dec-2011
863    ------------------------
865    1.  Updating the JIT compiler.
867    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
868        are added as well.
870    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
871        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
872        calling _pcre_jit_exec. Some extra comments are added.
874    4.  (*MARK) settings inside atomic groups that do not contain any capturing
875        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
876        was introduced by change 18 for 8.20.
878    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
879        ECMA-262 standard.
881    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
882        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
883        This bug was probably introduced by change 9 of 8.13.
885    7.  While fixing 6 above, I noticed that a number of other items were being
886        incorrectly rejected as "not fixed length". This arose partly because newer
887        opcodes had not been added to the fixed-length checking code. I have (a)
888        corrected the bug and added tests for these items, and (b) arranged for an
889        error to occur if an unknown opcode is encountered while checking for fixed
890        length instead of just assuming "not fixed length". The items that were
891        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
892        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
893        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
895    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
896        being incorrectly compiled and would have given unpredicatble results.
898    9.  A possessively repeated subpattern with minimum repeat count greater than
899        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
900        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
901        the first (A) could occur when it should not.
903    10. Add a cast and remove a redundant test from the code.
905    11. JIT should use pcre_malloc/pcre_free for allocation.
907    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
908        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
909        is anything other than /usr, -L is still shown).
911    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
913    14. Perl does not support \N without a following name in a [] class; PCRE now
914        also gives an error.
916    15. If a forward reference was repeated with an upper limit of around 2000,
917        it caused the error "internal error: overran compiling workspace". The
918        maximum number of forward references (including repeats) was limited by the
919        internal workspace, and dependent on the LINK_SIZE. The code has been
920        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
921        the default depends on LINK_SIZE. There is a new upper limit (for safety)
922        of around 200,000 forward references. While doing this, I also speeded up
923        the filling in of repeated forward references.
925    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
926        incorrectly expecting the subject to contain another "a" after the start.
928    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
929        in the match, the SKIP should be ignored. This was not happening; instead
930        the SKIP was being treated as NOMATCH. For patterns such as
931        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
932        tested.
934    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
935        now much more compatible with Perl, in particular in cases where the result
936        is a non-match for a non-anchored pattern. For example, if
937        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
938        "m", where previously it did not return a name. A side effect of this
939        change is that for partial matches, the last encountered mark name is
940        returned, as for non matches. A number of tests that were previously not
941        Perl-compatible have been moved into the Perl-compatible test files. The
942        refactoring has had the pleasing side effect of removing one argument from
943        the match() function, thus reducing its stack requirements.
945    19. If the /S+ option was used in pcretest to study a pattern using JIT,
946        subsequent uses of /S (without +) incorrectly behaved like /S+.
948    21. Retrieve executable code size support for the JIT compiler and fixing
949        some warnings.
951    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
952        not work when the shorter character appeared right at the end of the
953        subject string.
955    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
956        systems.
958    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
959        output it when the /M option is used in pcretest.
961    25. The CheckMan script was not being included in the distribution. Also, added
962        an explicit "perl" to run Perl scripts from the PrepareRelease script
963        because this is reportedly needed in Windows.
965    26. If study data was being save in a file and studying had not found a set of
966        "starts with" bytes for the pattern, the data written to the file (though
967        never used) was taken from uninitialized memory and so caused valgrind to
968        complain.
970    27. Updated RunTest.bat as provided by Sheri Pierce.
972    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
974    29. Computation of memory usage for the table of capturing group names was
975        giving an unnecessarily large value.
978    Version 8.20 21-Oct-2011
979    ------------------------
981    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
982        a POSIX class. After further experiments with Perl, which convinced me that
983        Perl has bugs and confusions, a closing square bracket is no longer allowed
984        in a POSIX name. This bug also affected patterns with classes that started
985        with full stops.
987    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
988        captured substring, but while checking the failing first alternative,
989        substring 1 is temporarily captured. If the output vector supplied to
990        pcre_exec() was not big enough for this capture, the yield of the function
991        was still zero ("insufficient space for captured substrings"). This cannot
992        be totally fixed without adding another stack variable, which seems a lot
993        of expense for a edge case. However, I have improved the situation in cases
994        such as /(a)(b)x|abc/ matched against "abc", where the return code
995        indicates that fewer than the maximum number of slots in the ovector have
996        been set.
998    3.  Related to (2) above: when there are more back references in a pattern than
999        slots in the output vector, pcre_exec() uses temporary memory during
1000        matching, and copies in the captures as far as possible afterwards. It was
1001        using the entire output vector, but this conflicts with the specification
1002        that only 2/3 is used for passing back captured substrings. Now it uses
1003        only the first 2/3, for compatibility. This is, of course, another edge
1004        case.
1006    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1007        main code base, and can be used by building with --enable-jit. When this is
1008        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1009        runtime --no-jit option is given.
1011    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1012        ovector, the return from the function was zero, implying that there were
1013        other matches that did not fit. The correct "exactly full" value is now
1014        returned.
1016    6.  If a subpattern that was called recursively or as a subroutine contained
1017        (*PRUNE) or any other control that caused it to give a non-standard return,
1018        invalid errors such as "Error -26 (nested recursion at the same subject
1019        position)" or even infinite loops could occur.
1021    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1022        computing the minimum length on reaching *ACCEPT, and so ended up with the
1023        wrong value of 1 rather than 0. Further investigation indicates that
1024        computing a minimum subject length in the presence of *ACCEPT is difficult
1025        (think back references, subroutine calls), and so I have changed the code
1026        so that no minimum is registered for a pattern that contains *ACCEPT.
1028    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1029        it was not handled as intended. [But see 16 below.]
1031    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1032        Sheri Pierce.
1034    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1035        the first byte in a match must be "a".
1037    11. Change 17 for 8.13 increased the recursion depth for patterns like
1038        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1039        pattern contains any instances of (*THEN). If it does not, the old
1040        optimizations are restored. It would be nice to do this on a per-group
1041        basis, but at the moment that is not feasible.
1043    12. In some environments, the output of pcretest -C is CRLF terminated. This
1044        broke RunTest's code that checks for the link size. A single white space
1045        character after the value is now allowed for.
1047    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1048        For "fr", it uses the Windows-specific input and output files.
1050    14. If (*THEN) appeared in a group that was called recursively or as a
1051        subroutine, it did not work as intended. [But see next item.]
1053    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1054        pattern fragments (but not containing any | characters). If A and B are
1055        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1056        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1057        D. In other words, Perl considers parentheses that do not contain any |
1058        characters to be part of a surrounding alternative, whereas PCRE was
1059        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1060        differently. PCRE now behaves in the same way as Perl, except in the case
1061        of subroutine/recursion calls such as (?1) which have in any case always
1062        been different (but PCRE had them first :-).
1064    16. Related to 15 above: Perl does not treat the | in a conditional group as
1065        creating alternatives. Such a group is treated in the same way as an
1066        ordinary group without any | characters when processing (*THEN). PCRE has
1067        been changed to match Perl's behaviour.
1069    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1070        RunGrepTest script failed.
1072    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1073        inevitable for groups that contain captures, but it can lead to a lot of
1074        stack use in large patterns. The old behaviour has been restored for atomic
1075        groups that do not contain any capturing parentheses.
1077    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1078        suppress the check for a minimum subject length at run time. (If it was
1079        given to pcre_exec() or pcre_dfa_exec() it did work.)
1081    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1082        fail to work when decoding hex characters in data strings in EBCDIC
1083        environments.
1085    21. It appears that in at least one Mac OS environment, the isxdigit() function
1086        is implemented as a macro that evaluates to its argument more than once,
1087        contravening the C 90 Standard (I haven't checked a later standard). There
1088        was an instance in pcretest which caused it to go wrong when processing
1089        \x{...} escapes in subject strings. The has been rewritten to avoid using
1090        things like p++ in the argument of isxdigit().
1093    Version 8.13 16-Aug-2011
1094    ------------------------
1096    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1098    2.  Two minor typos in pcre_internal.h have been fixed.
1100    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1101        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1102        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1104    4.  There were a number of related bugs in the code for matching backrefences
1105        caselessly in UTF-8 mode when codes for the characters concerned were
1106        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1107        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1108        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1109        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1110        2-byte code at the end of the subject (it thought there wasn't enough data
1111        left).
1113    5.  Comprehensive information about what went wrong is now returned by
1114        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1115        as the output vector has at least 2 elements. The offset of the start of
1116        the failing character and a reason code are placed in the vector.
1118    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1119        now returned is for the first byte of the failing character, instead of the
1120        last byte inspected. This is an incompatible change, but I hope it is small
1121        enough not to be a problem. It makes the returned offset consistent with
1122        pcre_exec() and pcre_dfa_exec().
1124    7.  pcretest now gives a text phrase as well as the error number when
1125        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1126        failure, the offset and reason code are output.
1128    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1129        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1130        back over a single character (\n). This seems wrong (because it treated the
1131        two characters as a single entity when going forwards), conflicts with the
1132        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1133        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1134        has been changed.
1136    9.  Some internal refactoring has changed the processing so that the handling
1137        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1138        time (the PCRE_DOTALL option was changed this way some time ago: version
1139        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1140        which was always a bit of a fudge. It also means that there is one less
1141        argument for the match() function, which reduces its stack requirements
1142        slightly. This change also fixes an incompatibility with Perl: the pattern
1143        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1145    10. More internal refactoring has drastically reduced the number of recursive
1146        calls to match() for possessively repeated groups such as (abc)++ when
1147        using pcre_exec().
1149    11. While implementing 10, a number of bugs in the handling of groups were
1150        discovered and fixed:
1152        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1153        (a|)*(?1) gave a compile-time internal error.
1154        ((a|)+)+  did not notice that the outer group could match an empty string.
1155        (^a|^)+   was not marked as anchored.
1156        (.*a|.*)+ was not marked as matching at start or after a newline.
1158    12. Yet more internal refactoring has removed another argument from the match()
1159        function. Special calls to this function are now indicated by setting a
1160        value in a variable in the "match data" data block.
1162    13. Be more explicit in pcre_study() instead of relying on "default" for
1163        opcodes that mean there is no starting character; this means that when new
1164        ones are added and accidentally left out of pcre_study(), testing should
1165        pick them up.
1167    14. The -s option of pcretest has been documented for ages as being an old
1168        synonym of -m (show memory usage). I have changed it to mean "force study
1169        for every regex", that is, assume /S for every regex. This is similar to -i
1170        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1171        using it. It makes it easier to run collections of tests with and without
1172        study enabled, and thereby test pcre_study() more easily. All the standard
1173        tests are now run with and without -s (but some patterns can be marked as
1174        "never study" - see 20 below).
1176    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1177        restoration of the capturing data to the outer values was not happening
1178        correctly.
1180    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1181        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1182        pattern had matched an empty string, and so incorrectly returned a no
1183        match.
1185    17. There was optimizing code for the last branch of non-capturing parentheses,
1186        and also for the obeyed branch of a conditional subexpression, which used
1187        tail recursion to cut down on stack usage. Unfortunately, now that there is
1188        the possibility of (*THEN) occurring in these branches, tail recursion is
1189        no longer possible because the return has to be checked for (*THEN). These
1190        two optimizations have therefore been removed. [But see 8.20/11 above.]
1192    18. If a pattern containing \R was studied, it was assumed that \R always
1193        matched two bytes, thus causing the minimum subject length to be
1194        incorrectly computed because \R can also match just one byte.
1196    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1197        was incorrectly computed.
1199    20. If /S is present twice on a test pattern in pcretest input, it now
1200        *disables* studying, thereby overriding the use of -s on the command line
1201        (see 14 above). This is necessary for one or two tests to keep the output
1202        identical in both cases.
1204    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1205        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1207    22. When an atomic group that contained a capturing parenthesis was
1208        successfully matched, but the branch in which it appeared failed, the
1209        capturing was not being forgotten if a higher numbered group was later
1210        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1211        group 1 to "a", when in fact it should be unset. This applied to multi-
1212        branched capturing and non-capturing groups, repeated or not, and also to
1213        positive assertions (capturing in negative assertions does not happen
1214        in PCRE) and also to nested atomic groups.
1216    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1217        subject after a captured substring, to make it easier to tell which of a
1218        number of identical substrings has been captured.
1220    24. The way atomic groups are processed by pcre_exec() has been changed so that
1221        if they are repeated, backtracking one repetition now resets captured
1222        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1223        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1224        "aaa". Previously, it would have been "a". As part of this code
1225        refactoring, the way recursive calls are handled has also been changed.
1227    25. If an assertion condition captured any substrings, they were not passed
1228        back unless some other capturing happened later. For example, if
1229        (?(?=(a))a) was matched against "a", no capturing was returned.
1231    26. When studying a pattern that contained subroutine calls or assertions,
1232        the code for finding the minimum length of a possible match was handling
1233        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1234        group 1 called group 2 while simultaneously a separate group 2 called group
1235        1). A stack overflow occurred in this case. I have fixed this by limiting
1236        the recursion depth to 10.
1238    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1239        Fortmann. This supports explicit test numbers on the command line, and has
1240        argument validation and error reporting.
1242    28. An instance of \X with an unlimited repeat could fail if at any point the
1243        first character it looked at was a mark character.
1245    29. Some minor code refactoring concerning Unicode properties and scripts
1246        should reduce the stack requirement of match() slightly.
1248    30. Added the '=' option to pcretest to check the setting of unused capturing
1249        slots at the end of the pattern, which are documented as being -1, but are
1250        not included in the return count.
1252    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1253        compiled something random. Now it gives a compile-time error (as does
1254        Perl).
1256    32. A *MARK encountered during the processing of a positive assertion is now
1257        recorded and passed back (compatible with Perl).
1259    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1260        had alternative anchored branches, the search for a second match in a line
1261        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1262        matched the line "0102" twice. The same bug affected patterns that started
1263        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1264        twice.
1266    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1267        does, and because of capturing effects, quantifying parenthesized
1268        assertions may at times be useful. Quantifiers are now allowed for
1269        parenthesized assertions.
1271    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1273    36. \g was being checked for fancy things in a character class, when it should
1274        just be a literal "g".
1276    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1277        appearance of a nested POSIX class supersedes an apparent external class.
1278        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1279        unescaped square brackets may also appear as part of class names. For
1280        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1281        more like Perl. (But see 8.20/1 above.)
1283    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1284        was because it thought it was \N{name}, which is not supported).
1286    39. Add minix to OS list not supporting the -S option in pcretest.
1288    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1289        cannot analyze patterns in sufficient detail to catch mutual recursions
1290        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1291        subgroup is called recursively as a subpattern for a second time at the
1292        same position in the subject string. In previous releases this might have
1293        been caught by the recursion limit, or it might have run out of stack.
1295    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1296        happen only once. PCRE was, however incorrectly giving a compile time error
1297        "recursive call could loop indefinitely" because it cannot analyze the
1298        pattern in sufficient detail. The compile time test no longer happens when
1299        PCRE is compiling a conditional subpattern, but actual runaway loops are
1300        now caught at runtime (see 40 above).
1302    42. It seems that Perl allows any characters other than a closing parenthesis
1303        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1304        has been changed to be the same.
1306    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1307        as not to get warnings when autogen.sh is called. Also changed
1308        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1310    44. To help people who use pcregrep to scan files containing exceedingly long
1311        lines, the following changes have been made:
1313        (a) The default value of the buffer size parameter has been increased from
1314            8K to 20K. (The actual buffer used is three times this size.)
1316        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1317            PCRE is built.
1319        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1320            to be set at run time.
1322        (d) Numerical values in pcregrep options can be followed by K or M, for
1323            example --buffer-size=50K.
1325        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1326            given and the return code is set to 2.
1328    45. Add a pointer to the latest mark to the callout data block.
1330    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1331        partial match of an empty string instead of no match. This was specific to
1332        the use of ".".
1334    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1335        complete match instead of a partial match. This bug was dependent on both
1336        the PCRE_UTF8 and PCRE_DOTALL options being set.
1338    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1339        starting byte set, because \b was not being ignored.
1342    Version 8.12 15-Jan-2011
1343    ------------------------
1345    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1346        checks for such things as part of the documentation building process.
1348    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1349        --match-limit and --recursion-limit options (added for 8.11). In
1350        particular, this made one of the standard tests fail. (The integer value
1351        went into the wrong half of a long int.)
1353    3.  If the --colour option was given to pcregrep with -v (invert match), it
1354        did strange things, either producing crazy output, or crashing. It should,
1355        of course, ignore a request for colour when reporting lines that do not
1356        match.
1358    4.  Another pcregrep bug caused similar problems if --colour was specified with
1359        -M (multiline) and the pattern match finished with a line ending.
1361    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1362        matched in multiline mode, the following line was shown as part of the
1363        match. This seems wrong, so I have changed it.
1365    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1366        the check for further matches in the same line (so they could be coloured)
1367        to overrun the end of the current line. If another match was found, it was
1368        incorrectly shown (and then shown again when found in the next line).
1370    7.  If pcregrep was compiled under Windows, there was a reference to the
1371        function pcregrep_exit() before it was defined. I am assuming this was
1372        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1373        reported by a user. I've moved the definition above the reference.
1376    Version 8.11 10-Dec-2010
1377    ------------------------
1379    1.  (*THEN) was not working properly if there were untried alternatives prior
1380        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1381        backtracked to try for "b" instead of moving to the next alternative branch
1382        at the same level (in this case, to look for "c"). The Perl documentation
1383        is clear that when (*THEN) is backtracked onto, it goes to the "next
1384        alternative in the innermost enclosing group".
1386    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1387        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1388        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1389        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1390        (*THEN).
1392    3.  If \s appeared in a character class, it removed the VT character from
1393        the class, even if it had been included by some previous item, for example
1394        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1395        of \s, but is part of the POSIX "space" class.)
1397    4.  A partial match never returns an empty string (because you can always
1398        match an empty string at the end of the subject); however the checking for
1399        an empty string was starting at the "start of match" point. This has been
1400        changed to the "earliest inspected character" point, because the returned
1401        data for a partial match starts at this character. This means that, for
1402        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1403        (previously it gave "no match").
1405    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1406        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1407        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1408        has an implication that the given string is incomplete (because a partial
1409        match is preferred over a full match). For this reason, these items now
1410        give a partial match in this situation. [Aside: previously, the one case
1411        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1412        match rather than a full match, which was wrong by the old rules, but is
1413        now correct.]
1415    6.  There was a bug in the handling of #-introduced comments, recognized when
1416        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1417        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1418        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1419        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1420        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1421        places in pcre_compile().
1423    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1424        comments when looking ahead for named forward references to subpatterns,
1425        the only newline sequence it recognized was NL. It now handles newlines
1426        according to the set newline convention.
1428    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1429        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1430        cater for a lack of strerror(). These oversights have been fixed.
1432    9.  Added --match-limit and --recursion-limit to pcregrep.
1434    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1436    11. When the -o option was used, pcregrep was setting a return code of 1, even
1437        when matches were found, and --line-buffered was not being honoured.
1439    12. Added an optional parentheses number to the -o and --only-matching options
1440        of pcregrep.
1442    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1443        can match an empty string. The code to do it in pcretest and pcredemo
1444        needed fixing:
1446        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1447            only one byte after an empty string match just before CRLF (this case
1448            just got forgotten; "any" and "anycrlf" were OK).
1450        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1451            mode when an empty string match preceded an ASCII character followed by
1452            a non-ASCII character. (The code for advancing by one character rather
1453            than one byte was nonsense.)
1455        (c) The pcredemo.c sample program did not have any code at all to handle
1456            the cases when CRLF is a valid newline sequence.
1458    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1459        as a starting offset was within the subject string. There is now a new
1460        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1461        negative or greater than the length of the string. In order to test this,
1462        pcretest is extended to allow the setting of negative starting offsets.
1464    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1465        starting offset points to the beginning of a UTF-8 character was
1466        unnecessarily clumsy. I tidied it up.
1468    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1469        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1471    17. Nobody had reported that the --include_dir option, which was added in
1472        release 7.7 should have been called --include-dir (hyphen, not underscore)
1473        for compatibility with GNU grep. I have changed it to --include-dir, but
1474        left --include_dir as an undocumented synonym, and the same for
1475        --exclude-dir, though that is not available in GNU grep, at least as of
1476        release 2.5.4.
1478    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1479        characters from a string of bytes) have been redefined so as not to use
1480        loops, in order to improve performance in some environments. At the same
1481        time, I abstracted some of the common code into auxiliary macros to save
1482        repetition (this should not affect the compiled code).
1484    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1485        compile-time error is now given if \c is not followed by an ASCII
1486        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1487        different, and any byte value is allowed.)
1489    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1490        START_OPTIMIZE option, which is now allowed at compile time - but just
1491        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1492        to pcregrep and other applications that have no direct access to PCRE
1493        options. The new /Y option in pcretest sets this option when calling
1494        pcre_compile().
1496    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1497        back references. Groups containing recursive back references were forced to
1498        be atomic by that change, but in the case of named groups, the amount of
1499        memory required was incorrectly computed, leading to "Failed: internal
1500        error: code overflow". This has been fixed.
1502    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1503        pcretest.c, to avoid build problems in some Borland environments.
1506    Version 8.10 25-Jun-2010
1507    ------------------------
1509    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1510        THEN.
1512    2.  (*ACCEPT) was not working when inside an atomic group.
1514    3.  Inside a character class, \B is treated as a literal by default, but
1515        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1516        causes the error). The code is unchanged, but I tidied the documentation.
1518    4.  Inside a character class, PCRE always treated \R and \X as literals,
1519        whereas Perl faults them if its -w option is set. I have changed PCRE so
1520        that it faults them when PCRE_EXTRA is set.
1522    5.  Added support for \N, which always matches any character other than
1523        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1525    6.  When compiling pcregrep with newer versions of gcc which may have
1526        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1527        declared with attribute warn_unused_result" were given. Just casting the
1528        result to (void) does not stop the warnings; a more elaborate fudge is
1529        needed. I've used a macro to implement this.
1531    7.  Minor change to pcretest.c to avoid a compiler warning.
1533    8.  Added four artifical Unicode properties to help with an option to make
1534        \s etc use properties (see next item). The new properties are: Xan
1535        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1537    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1538        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1539        this option. Modified pcretest to add /W to test this facility. Added
1540        REG_UCP to make it available via the POSIX interface.
1542    10. Added --line-buffered to pcregrep.
1544    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1545        studied, and the match started with a letter with a code point greater than
1546        127 whose first byte was different to the first byte of the other case of
1547        the letter, the other case of this starting letter was not recognized
1548        (#976).
1550    12. If a pattern that was studied started with a repeated Unicode property
1551        test, for example, \p{Nd}+, there was the theoretical possibility of
1552        setting up an incorrect bitmap of starting bytes, but fortunately it could
1553        not have actually happened in practice until change 8 above was made (it
1554        added property types that matched character-matching opcodes).
1556    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1557        possible starting bytes for non-anchored patterns.
1559    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1560        \R, and also a number of cases that involve Unicode properties, both
1561        explicit and implicit when PCRE_UCP is set.
1563    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1564        input, it could crash or give wrong results if characters with values
1565        greater than 0xc0 were present in the subject string. (Detail: it assumed
1566        UTF-8 input when processing these items.)
1568    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1569        size_t is 64-bit (#991).
1571    17. Added a check for running out of memory when PCRE is compiled with
1572        --disable-stack-for-recursion (#990).
1574    18. If the last data line in a file for pcretest does not have a newline on
1575        the end, a newline was missing in the output.
1577    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1578        less than 128) in its various bitmaps. However, there is a facility for
1579        generating tables according to the current locale when PCRE is compiled. It
1580        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1581        characters, are recognized by isspace() and therefore were getting set in
1582        these tables, and indeed these tables seem to approximate to ISO 8859. This
1583        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1584        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1585        which of course cannot start UTF-8 characters. I have changed the code so
1586        that only real ASCII characters (less than 128) and the correct starting
1587        bytes for UTF-8 encodings are set for characters greater than 127 when in
1588        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1589        altogether.)
1591    20. Added the /T option to pcretest so as to be able to run tests with non-
1592        standard character tables, thus making it possible to include the tests
1593        used for 19 above in the standard set of tests.
1595    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1596        reference to a subpattern the other side of a comment that contains an
1597        opening parenthesis caused either an internal compiling error, or a
1598        reference to the wrong subpattern.
1601    Version 8.02 19-Mar-2010
1602    ------------------------
1604    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1606    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1607        configured.
1609    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1610        original author of that file, following a query about its status.
1612    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1613        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1615    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1616        quantifier applied to a forward-referencing subroutine call, could compile
1617        incorrect code or give the error "internal error: previously-checked
1618        referenced subpattern not found".
1620    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1621        variables to point to external functions. For these systems, therefore,
1622        pcre_malloc etc. are now initialized to local functions that call the
1623        relevant global functions.
1625    7.  There were two entries missing in the vectors called coptable and poptable
1626        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1627        I've fixed the data, and added a kludgy way of testing at compile time that
1628        the lengths are correct (equal to the number of opcodes).
1630    8.  Following on from 7, I added a similar kludge to check the length of the
1631        eint vector in pcreposix.c.
1633    9.  Error texts for pcre_compile() are held as one long string to avoid too
1634        much relocation at load time. To find a text, the string is searched,
1635        counting zeros. There was no check for running off the end of the string,
1636        which could happen if a new error number was added without updating the
1637        string.
1639    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1641    11. \K was not working if it appeared in an atomic group or in a group that
1642        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1643        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1644        the assertion is positive, but not if it is negative.
1646    12. Change 11 fortuitously reduced the size of the stack frame used in the
1647        "match()" function of pcre_exec.c by one pointer. Forthcoming
1648        implementation of support for (*MARK) will need an extra pointer on the
1649        stack; I have reserved it now, so that the stack frame size does not
1650        decrease.
1652    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1653        item in branch that calls a recursion is a subroutine call - as in the
1654        second branch in the above example - was incorrectly given the compile-
1655        time error "recursive call could loop indefinitely" because pcre_compile()
1656        was not correctly checking the subroutine for matching a non-empty string.
1658    14. The checks for overrunning compiling workspace could trigger after an
1659        overrun had occurred. This is a "should never occur" error, but it can be
1660        triggered by pathological patterns such as hundreds of nested parentheses.
1661        The checks now trigger 100 bytes before the end of the workspace.
1663    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1666    Version 8.01 19-Jan-2010
1667    ------------------------
1669    1.  If a pattern contained a conditional subpattern with only one branch (in
1670        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1671        computed the wrong minimum data length (which is of course zero for such
1672        subpatterns). This could cause incorrect "no match" results.
1674    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1675        the pattern is reset in the first branch, pcre_compile() failed with
1676        "internal error: code overflow at offset...". This happened only when
1677        the reset was to the original external option setting. (An optimization
1678        abstracts leading options settings into an external setting, which was the
1679        cause of this.)
1681    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1682        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1683        assertion pattern did not match (meaning that the assertion was true), it
1684        was incorrectly treated as false if the SKIP had been reached during the
1685        matching. This also applied to assertions used as conditions.
1687    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1688        assertion subpattern, including such a pattern used as a condition,
1689        unpredictable results occurred, instead of the error return
1692    5.  The C++ GlobalReplace function was not working like Perl for the special
1693        situation when an empty string is matched. It now does the fancy magic
1694        stuff that is necessary.
1696    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1697        removed. (These were left over from very, very early versions of PCRE.)
1699    7.  Some cosmetic changes to the code to make life easier when compiling it
1700        as part of something else:
1702        (a) Change DEBUG to PCRE_DEBUG.
1704        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1705            called "current" as "current_branch", to prevent a collision with the
1706            Linux macro when compiled as a kernel module.
1708        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1709            prevent a collision with the Linux macro when compiled as a kernel
1710            module.
1712    8.  In pcre_compile() there are some checks for integer overflows that used to
1713        cast potentially large values to (double). This has been changed to that
1714        when building, a check for int64_t is made, and if it is found, it is used
1715        instead, thus avoiding the use of floating point arithmetic. (There is no
1716        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1717        double.
1719    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1720        2005 (difference between two addresses compared to an unsigned value).
1722    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1723        custom one, because of the following reported problem in Windows:
1725          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1726              under Win32.
1727          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1728              therefore missing the function definition.
1729          - The compiler thus generates a "C" signature for the test function.
1730          - The linker fails to find the "C" function.
1731          - PCRE fails to configure if asked to do so against libbz2.
1733    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1734        messages were output:
1736          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1737          rerunning libtoolize, to keep the correct libtool macros in-tree.
1738          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1740        I have done both of these things.
1742    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1743        most of the time, it *can* run out if it is given a pattern that contains a
1744        runaway infinite recursion. I updated the discussion in the pcrestack man
1745        page.
1747    13. Now that we have gone to the x.xx style of version numbers, the minor
1748        version may start with zero. Using 08 or 09 is a bad idea because users
1749        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1750        interpreted as invalid octal numbers. I've updated the previous comment in
1751        configure.ac, and also added a check that gives an error if 08 or 09 are
1752        used.
1754    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1755        causing partial matching to fail when the end of the subject matched \W
1756        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1758    15. There were some discrepancies between the declarations in pcre_internal.h
1759        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1760        their definitions. The declarations used "const uschar *" and the
1761        definitions used USPTR. Even though USPTR is normally defined as "const
1762        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1763        reported that: "This difference in casting confuses some C++ compilers, for
1764        example, SunCC recognizes above declarations as different functions and
1765        generates broken code for hbpcre." I have changed the declarations to use
1766        USPTR.
1768    16. GNU libtool is named differently on some systems. The autogen.sh script now
1769        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1770        (FreeBSD).
1772    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1773        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1774        comment: "Figure out how to create a longlong from a string: strtoll and
1775        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1776        instance, but it only takes 2 args instead of 3!"
1778    18. A subtle bug concerned with back references has been fixed by a change of
1779        specification, with a corresponding code fix. A pattern such as
1780        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1781        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1782        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1783        same bug. Such groups have to be quantified to be useful, or contained
1784        inside another quantified group. (If there's no repetition, the reference
1785        can never match.) The problem arises because, having left the group and
1786        moved on to the rest of the pattern, a later failure that backtracks into
1787        the group uses the captured value from the final iteration of the group
1788        rather than the correct earlier one. I have fixed this in PCRE by forcing
1789        any group that contains a reference to itself to be an atomic group; that
1790        is, there cannot be any backtracking into it once it has completed. This is
1791        similar to recursive and subroutine calls.
1794    Version 8.00 19-Oct-09
1795    ----------------------
1797    1.  The table for translating pcre_compile() error codes into POSIX error codes
1798        was out-of-date, and there was no check on the pcre_compile() error code
1799        being within the table. This could lead to an OK return being given in
1800        error.
1802    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1803        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1804        in a Windows environment.
1806    3.  The pcregrep --count option prints the count for each file even when it is
1807        zero, as does GNU grep. However, pcregrep was also printing all files when
1808        --files-with-matches was added. Now, when both options are given, it prints
1809        counts only for those files that have at least one match. (GNU grep just
1810        prints the file name in this circumstance, but including the count seems
1811        more useful - otherwise, why use --count?) Also ensured that the
1812        combination -clh just lists non-zero counts, with no names.
1814    4.  The long form of the pcregrep -F option was incorrectly implemented as
1815        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1816        but it seems right to fix it, and I didn't think it was worth preserving
1817        the old behaviour.
1819    5.  The command line items --regex=pattern and --regexp=pattern were not
1820        recognized by pcregrep, which required --regex pattern or --regexp pattern
1821        (with a space rather than an '='). The man page documented the '=' forms,
1822        which are compatible with GNU grep; these now work.
1824    6.  No libpcreposix.pc file was created for pkg-config; there was just
1825        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1827    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1828        when UCP support is not needed, by modifying the Python script that
1829        generates it from Unicode data files. This should not matter if the module
1830        is correctly used as a library, but I received one complaint about 50K of
1831        unwanted data. My guess is that the person linked everything into his
1832        program rather than using a library. Anyway, it does no harm.
1834    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1835        was a minimum greater than 1 for a wide character in a possessive
1836        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1837        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1838        character. Chaos in the form of incorrect output or a compiling loop could
1839        result.
1841    9.  The restrictions on what a pattern can contain when partial matching is
1842        requested for pcre_exec() have been removed. All patterns can now be
1843        partially matched by this function. In addition, if there are at least two
1844        slots in the offset vector, the offset of the earliest inspected character
1845        for the match and the offset of the end of the subject are set in them when
1846        PCRE_ERROR_PARTIAL is returned.
1848    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1849        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1850        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1851        and may be more useful for multi-segment matching.
1853    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1854        used to be given if ever the end of the subject was reached; now it is
1855        given only if matching could not proceed because another character was
1856        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1857        string "Z", which now yields "no match" instead of "partial match". In the
1858        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1859        final character ended with (*FAIL).
1861    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1862        if the pattern had a "must contain" character that was already found in the
1863        earlier partial match, unless partial matching was again requested. For
1864        example, with the pattern /dog.(body)?/, the "must contain" character is
1865        "g". If the first part-match was for the string "dog", restarting with
1866        "sbody" failed. This bug has been fixed.
1868    13. The string returned by pcre_dfa_exec() after a partial match has been
1869        changed so that it starts at the first inspected character rather than the
1870        first character of the match. This makes a difference only if the pattern
1871        starts with a lookbehind assertion or \b or \B (\K is not supported by
1872        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1873        matching functions compatible, and I think it's the right thing to do.
1875    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1876        so that the demonstration program is easily available in environments where
1877        PCRE has not been installed from source.
1879    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1880        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1881        library.
1883    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1884        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1885        is not the first non-POSIX option to be added. Clearly some people find
1886        these options useful.
1888    17. If a caller to the POSIX matching function regexec() passes a non-zero
1889        value for nmatch with a NULL value for pmatch, the value of
1890        nmatch is forced to zero.
1892    18. RunGrepTest did not have a test for the availability of the -u option of
1893        the diff command, as RunTest does. It now checks in the same way as
1894        RunTest, and also checks for the -b option.
1896    19. If an odd number of negated classes containing just a single character
1897        interposed, within parentheses, between a forward reference to a named
1898        subpattern and the definition of the subpattern, compilation crashed with
1899        an internal error, complaining that it could not find the referenced
1900        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1901        [The bug was that it was starting one character too far in when skipping
1902        over the character class, thus treating the ] as data rather than
1903        terminating the class. This meant it could skip too much.]
1905    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1906        /g option in pcretest when the pattern contains \K, which makes it possible
1907        to have an empty string match not at the start, even when the pattern is
1908        anchored. Updated pcretest and pcredemo to use this option.
1910    21. If the maximum number of capturing subpatterns in a recursion was greater
1911        than the maximum at the outer level, the higher number was returned, but
1912        with unset values at the outer level. The correct (outer level) value is
1913        now given.
1915    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1916        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1917        make it do so. The string so far is captured, making this feature
1918        compatible with Perl.
1920    23. The tests have been re-organized, adding tests 11 and 12, to make it
1921        possible to check the Perl 5.10 features against Perl 5.10.
1923    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1924        pattern matches a fixed length string. PCRE did not allow this; now it
1925        does. Neither allows recursion.
1927    25. I finally figured out how to implement a request to provide the minimum
1928        length of subject string that was needed in order to match a given pattern.
1929        (It was back references and recursion that I had previously got hung up
1930        on.) This code has now been added to pcre_study(); it finds a lower bound
1931        to the length of subject needed. It is not necessarily the greatest lower
1932        bound, but using it to avoid searching strings that are too short does give
1933        some useful speed-ups. The value is available to calling programs via
1934        pcre_fullinfo().
1936    26. While implementing 25, I discovered to my embarrassment that pcretest had
1937        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1938        study optimizations had never been tested with that matching function.
1939        Oops. What is worse, even when it was passed study data, there was a bug in
1940        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1941        were also very few tests of studied patterns with pcre_dfa_exec().
1943    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1944        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1945        on the other side of the coin, they are no longer allowed to have different
1946        names, because these cannot be distinguished in PCRE, and this has caused
1947        confusion. (This is a difference from Perl.)
1949    28. When duplicate subpattern names are present (necessarily with different
1950        numbers, as required by 27 above), and a test is made by name in a
1951        conditional pattern, either for a subpattern having been matched, or for
1952        recursion in such a pattern, all the associated numbered subpatterns are
1953        tested, and the overall condition is true if the condition is true for any
1954        one of them. This is the way Perl works, and is also more like the way
1955        testing by number works.
1958    Version 7.9 11-Apr-09
1959    ---------------------
1961    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1962        (pcretest), all targets were linked against these libraries. This included
1963        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1964        libraries. This caused unwanted dependencies to be created. This problem
1965        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1966        pcretest is linked with readline.
1968    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1969        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1970        moved outside it again, because FALSE and TRUE are already defined in AIX,
1971        but BOOL is not.
1973    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1974        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1976    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1977        hyphens) following file names and line numbers when outputting matching
1978        lines. This is not true; no spaces are inserted. I have also clarified the
1979        wording for the --colour (or --color) option.
1981    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1982        was not coloured; this is different to GNU grep, so I have changed it to be
1983        the same.
1985    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1986        each matching line was coloured. Now it goes on to look for further matches
1987        of any of the test patterns, which is the same behaviour as GNU grep.
1989    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1990        doesn't make sense to accept an empty string match in pcregrep, so I have
1991        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1992        seems to be how GNU grep behaves. [But see later change 40 for release
1993        8.33.]
1995    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1996        start or after a newline", because the conditional assertion was not being
1997        correctly handled. The rule now is that both the assertion and what follows
1998        in the first alternative must satisfy the test.
2000    9.  If auto-callout was enabled in a pattern with a conditional group whose
2001        condition was an assertion, PCRE could crash during matching, both with
2002        pcre_exec() and pcre_dfa_exec().
2004    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2005        used for matching.
2007    11. Unicode property support in character classes was not working for
2008        characters (bytes) greater than 127 when not in UTF-8 mode.
2010    12. Added the -M command line option to pcretest.
2012    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2014    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2016    16. Added comments and documentation about mis-use of no_arg in the C++
2017        wrapper.
2019    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2020        from Martin Jerabek that uses macro names for all relevant character and
2021        string constants.
2023    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2024        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2025        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2026        these, but not everybody uses configure.
2028    19. A conditional group that had only one branch was not being correctly
2029        recognized as an item that could match an empty string. This meant that an
2030        enclosing group might also not be so recognized, causing infinite looping
2031        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2032        with the subject "ab", where knowledge that the repeated group can match
2033        nothing is needed in order to break the loop.
2035    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2036        exec(), but without supplying a callout function, matching went wrong.
2038    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2039        leak if the size of the offset vector was greater than 30. When the vector
2040        is smaller, the saved offsets during recursion go onto a local stack
2041        vector, but for larger vectors malloc() is used. It was failing to free
2042        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2043        error, in fact).
2045    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2046        heapframe that is used only when UTF-8 support is enabled. This caused no
2047        problem, but was untidy.
2049    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2050        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2051        included within another project.
2053    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2054        slightly modified by me:
2056          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2057              not building pcregrep.
2059          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2060              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2062    25. Forward references, both numeric and by name, in patterns that made use of
2063        duplicate group numbers, could behave incorrectly or give incorrect errors,
2064        because when scanning forward to find the reference group, PCRE was not
2065        taking into account the duplicate group numbers. A pattern such as
2066        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2068    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2069        the feature of a custom pointer more persuasive (as requested by a user).
2071    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2072        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2073        pre-defined.
2075    28. Added support for (*UTF8) at the start of a pattern.
2077    29. Arrange for flags added by the "release type" setting in CMake to be shown
2078        in the configuration summary.
2081    Version 7.8 05-Sep-08
2082    ---------------------
2084    1.  Replaced UCP searching code with optimized version as implemented for Ad
2085        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2086        stage table and inline lookup instead of a function, giving speed ups of 2
2087        to 5 times on some simple patterns that I tested. Permission was given to
2088        distribute the MultiStage2.py script that generates the tables (it's not in
2089        the tarball, but is in the Subversion repository).
2091    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2092        scripts.
2094    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2095        a group with a zero qualifier. The result of the study could be incorrect,
2096        or the function might crash, depending on the pattern.
2098    4.  Caseless matching was not working for non-ASCII characters in back
2099        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2100        It now works when Unicode Property Support is available.
2102    5.  In pcretest, an escape such as \x{de} in the data was always generating
2103        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2104        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2105        truncation.
2107    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2109    7.  Added two (int) casts to pcregrep when printing the difference of two
2110        pointers, in case they are 64-bit values.
2112    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2113        test 2 if it fails.
2115    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2116        and a #define of that name to empty if it is not externally set. This is to
2117        allow users of MSVC to set it if necessary.
2119    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2120        the convenience functions in the pcre_get.c source file.
2122    11. An option change at the start of a pattern that had top-level alternatives
2123        could cause overwriting and/or a crash. This command provoked a crash in
2124        some environments:
2126          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2128        This potential security problem was recorded as CVE-2008-2371.
2130    12. For a pattern where the match had to start at the beginning or immediately
2131        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2132        pcre_dfa_exec() could read past the end of the passed subject if there was
2133        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2134        pcretest so that it places the subject at the end of its malloc-ed buffer.
2136    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2137        exec() might read past the end of the data buffer in UTF-8 mode.
2139    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2140        the data contained the byte 0x85 as part of a UTF-8 character within its
2141        first line. This applied both to normal and DFA matching.
2143    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2144        /^[^d]*?$/8 failed to match "abc".
2146    16. Added a missing copyright notice to pcrecpp_internal.h.
2148    17. Make it more clear in the documentation that values returned from
2149        pcre_exec() in ovector are byte offsets, not character counts.
2151    18. Tidied a few places to stop certain compilers from issuing warnings.
2153    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2154        supplied by Stefan Weber. I made a further small update for 7.8 because
2155        there is a change of source arrangements: the pcre_searchfuncs.c module is
2156        replaced by pcre_ucd.c.
2159    Version 7.7 07-May-08
2160    ---------------------
2162    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2163        a string to a long long, pretend we don't even have a long long." This is
2164        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2166    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2167        pre-7.6 versions, which defined a global no_arg variable instead of putting
2168        it in the RE class. (See also #8 below.)
2170    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2171        Lopes.
2173    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2175        (1) The include/exclude patterns were being applied to the whole pathnames
2176            of files, instead of just to the final components.
2178        (2) If there was more than one level of directory, the subdirectories were
2179            skipped unless they satisfied the include/exclude conditions. This is
2180            inconsistent with GNU grep (and could even be seen as contrary to the
2181            pcregrep specification - which I improved to make it absolutely clear).
2182            The action now is always to scan all levels of directory, and just
2183            apply the include/exclude patterns to regular files.
2185    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2186        --exclude_dir in the tests to avoid scanning .svn directories.
2188    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2189        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2190        doesn't support NULs in patterns.
2192    7.  Added some missing "const"s to declarations of static tables in
2193        pcre_compile.c and pcre_dfa_exec.c.
2195    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2196        caused by fix #2  above. (Subsequently also a second patch to fix the
2197        first patch. And a third patch - this was a messy problem.)
2199    9.  Applied Craig's patch to remove the use of push_back().
2201    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2202        matching function regexec().
2204    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2205        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2206        references. PCRE supports relative numbers with this syntax (I don't think
2207        Oniguruma does).
2209    12. Previously, a group with a zero repeat such as (...){0} was completely
2210        omitted from the compiled regex. However, this means that if the group
2211        was called as a subroutine from elsewhere in the pattern, things went wrong
2212        (an internal error was given). Such groups are now left in the compiled
2213        pattern, with a new opcode that causes them to be skipped at execution
2214        time.
2216    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2217        to the way PCRE behaves:
2219        (a) A lone ] character is dis-allowed (Perl treats it as data).
2221        (b) A back reference to an unmatched subpattern matches an empty string
2222            (Perl fails the current match path).
2224        (c) A data ] in a character class must be notated as \] because if the
2225            first data character in a class is ], it defines an empty class. (In
2226            Perl it is not possible to have an empty class.) The empty class []
2227            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2228            The negative empty class [^] matches any one character, independently
2229            of the DOTALL setting.
2231    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2232        non-existent subpattern following a character class starting with ']' and
2233        containing () gave an internal compiling error instead of "reference to
2234        non-existent subpattern". Fortunately, when the pattern did exist, the
2235        compiled code was correct. (When scanning forwards to check for the
2236        existence of the subpattern, it was treating the data ']' as terminating
2237        the class, so got the count wrong. When actually compiling, the reference
2238        was subsequently set up correctly.)
2240    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2241        it was being rejected as not supported by pcre_dfa_exec(), even though
2242        other assertions are supported. I have made pcre_dfa_exec() support
2243        (*FAIL).
2245    16. The implementation of 13c above involved the invention of a new opcode,
2246        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2247        cannot be changed at match time, I realized I could make a small
2248        improvement to matching performance by compiling OP_ALLANY instead of
2249        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2250        on the OP_ANY path.
2252    17. Compiling pcretest on Windows with readline support failed without the
2253        following two fixes: (1) Make the unistd.h include conditional on
2254        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2256    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2257        ncurses library to be included for pcretest when ReadLine support is
2258        requested, but also to allow for it to be overridden. This patch came from
2259        Daniel Bergström.
2261    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2262        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2263        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2264        spotting this.
2267    Version 7.6 28-Jan-08
2268    ---------------------
2270    1.  A character class containing a very large number of characters with
2271        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2272        overflow.
2274    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2275        HAVE_LONG_LONG is not defined.
2277    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2278        bring them up to date and include new features. This patch includes:
2280        - Fixed PH's badly added libz and libbz2 support.
2281        - Fixed a problem with static linking.
2282        - Added pcredemo. [But later removed - see 7 below.]
2283        - Fixed dftables problem and added an option.
2284        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2285            HAVE_LONG_LONG.
2286        - Added readline support for pcretest.
2287        - Added an listing of the option settings after cmake has run.
2289    4.  A user submitted a patch to Makefile that makes it easy to create
2290        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2291        Makefile.am that cause it to include this special target, without
2292        affecting anything else. Note that the same mingw target plus all
2293        the other distribution libraries and programs are now supported
2294        when configuring with CMake (see 6 below) instead of with
2295        Configure/Make.
2297    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2298        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2299        exported in the Windows port". It has not yet been confirmed that the patch
2300        solves the problem, but it does no harm.
2302    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2303        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2304        with CMake, and also correct the comment about stack recursion.
2306    7.  Remove the automatic building of pcredemo from the ./configure system and
2307        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2308        of a program that users should build themselves after PCRE is installed, so
2309        building it automatically is not really right. What is more, it gave
2310        trouble in some build environments.
2312    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2315    Version 7.5 10-Jan-08
2316    ---------------------
2318    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2319        values in parens when parsing an RE using the C++ wrapper."
2321    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2322        Characters greater than 255 were excluded from the class instead of being
2323        included.
2325    3.  The same bug as (2) above applied to negated POSIX classes such as
2326        [:^space:].
2328    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2329        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2330        I have changed it.
2332    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2333        first named subpattern) and a construct such as (?&a) would reference the
2334        first named subpattern whose name started with "a" (in other words, the
2335        length check was missing). Both these problems are fixed. "Subpattern name
2336        expected" is now given for (?&) (a zero-length name), and this patch also
2337        makes it give the same error for \k'' (previously it complained that that
2338        was a reference to a non-existent subpattern).
2340    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2341        this is right because (?- can be followed by option settings as well as by
2342        digits. I have, however, made the messages clearer.
2344    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2345        than the number used in the conditional) now cause a compile-time error.
2346        This is actually not compatible with Perl, which accepts such patterns, but
2347        treats the conditional as always being FALSE (as PCRE used to), but it
2348        seems to me that giving a diagnostic is better.
2350    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2351        and messages.
2353    9.  Fix two occurrences of "backslash" in comments that should have been
2354        "backspace".
2356    10. Remove two redundant lines of code that can never be obeyed (their function
2357        was moved elsewhere).
2359    11. The program that makes PCRE's Unicode character property table had a bug
2360        which caused it to generate incorrect table entries for sequences of
2361        characters that have the same character type, but are in different scripts.
2362        It amalgamated them into a single range, with the script of the first of
2363        them. In other words, some characters were in the wrong script. There were
2364        thirteen such cases, affecting characters in the following ranges:
2366          U+002b0 - U+002c1
2367          U+0060c - U+0060d
2368          U+0061e - U+00612
2369          U+0064b - U+0065e
2370          U+0074d - U+0076d
2371          U+01800 - U+01805
2372          U+01d00 - U+01d77
2373          U+01d9b - U+01dbf
2374          U+0200b - U+0200f
2375          U+030fc - U+030fe
2376          U+03260 - U+0327f
2377          U+0fb46 - U+0fbb1
2378          U+10450 - U+1049d
2380    12. The -o option (show only the matching part of a line) for pcregrep was not
2381        compatible with GNU grep in that, if there was more than one match in a
2382        line, it showed only the first of them. It now behaves in the same way as
2383        GNU grep.
2385    13. If the -o and -v options were combined for pcregrep, it printed a blank
2386        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2387        does the same. The return code can be used to tell if there were any
2388        non-matching lines.
2390    14. Added --file-offsets and --line-offsets to pcregrep.
2392    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2393        infinitely looping recursion. The bug was that positive lookaheads were not
2394        being skipped when checking for a possible empty match (negative lookaheads
2395        and both kinds of lookbehind were skipped).
2397    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2398        inclusion of <windows.h> to before rather than after the definition of
2399        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2401    17. Specifying a possessive quantifier with a specific limit for a Unicode
2402        character property caused pcre_compile() to compile bad code, which led at
2403        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2404        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2405        caused the error; without that there was no problem.
2407    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2409    19. Added --enable-pcretest-libreadline.
2411    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2412        RE::GlobalReplace(). As a result, the number of replacements returned was
2413        double what it should be. I removed one of the increments, but Craig sent a
2414        later patch that removed the other one (the right fix) and added unit tests
2415        that check the return values (which was not done before).
2417    21. Several CMake things:
2419        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2420            the names libpcre and libpcreposix, not just pcre and pcreposix.
2422        (2) The above change means that pcretest and pcregrep are now correctly
2423            linked with the newly-built libraries, not previously installed ones.
2427    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2428        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2429        UTF-8 newline character). The key issue is that the pattern starts .*;
2430        this means that the match must be either at the beginning, or after a
2431        newline. The bug was in the code for advancing after a failed match and
2432        checking that the new position followed a newline. It was not taking
2433        account of UTF-8 characters correctly.
2435    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2436        character classes. PCRE was not treating the sequence [:...:] as a
2437        character class unless the ... were all letters. Perl, however, seems to
2438        allow any characters between [: and :], though of course it rejects as
2439        unknown any "names" that contain non-letters, because all the known class
2440        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2441        for example, whereas PCRE did not - it did not recognize a POSIX character
2442        class. This seemed a bit dangerous, so the code has been changed to be
2443        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2444        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2445        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2446        Perl does, and where it didn't before.
2448    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2449        Windows environments %n is disabled by default.
2452    Version 7.4 21-Sep-07
2453    ---------------------
2455    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2456        means that a class such as [\s] counted as "explicit reference to CR or
2457        LF". That isn't really right - the whole point of the change was to try to
2458        help when there was an actual mention of one of the two characters. So now
2459        the change happens only if \r or \n (or a literal CR or LF) character is
2460        encountered.
2462    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2463        of both had grown to the point where there were only 3 bits left.
2464        Fortunately, there was spare space in the data structure, and so I have
2465        moved the internal flags into a new 16-bit field to free up more option
2466        bits.
2468    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2469        but did not set the internal JCHANGED flag - either of these is enough to
2470        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2471        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2472        start sets both bits.
2474    4.  Added options (at build time, compile time, exec time) to change \R from
2475        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2477    5.  doc/pcresyntax.html was missing from the distribution.
2479    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2480        compatibility, even though it is no longer used.
2482    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2483        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2484        windows.h file is present (where different names are used). [This was
2485        reversed later after testing - see 16 below.]
2487    8.  Changed all #include <config.h> to #include "config.h". There were also
2488        some further <pcre.h> cases that I changed to "pcre.h".
2490    9.  When pcregrep was used with the --colour option, it missed the line ending
2491        sequence off the lines that it output.
2493    10. It was pointed out to me that arrays of string pointers cause lots of
2494        relocations when a shared library is dynamically loaded. A technique of
2495        using a single long string with a table of offsets can drastically reduce
2496        these. I have refactored PCRE in four places to do this. The result is
2497        dramatic:
2499          Originally:                          290
2500          After changing UCP table:            187
2501          After changing error message table:   43
2502          After changing table of "verbs"       36
2503          After changing table of Posix names   22
2505        Thanks to the folks working on Gregex for glib for this insight.
2507    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2508        unicode-properties was also set.
2510    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2512    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2513        checked only for CRLF.
2515    14. Added casts to pcretest.c to avoid compiler warnings.
2517    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2519    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2520        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2521        entirely. This removes changes made in 7 above.
2523    17. The CMake files have been updated, and there is now more information about
2524        building with CMake in the NON-UNIX-USE document.
2527    Version 7.3 28-Aug-07
2528    ---------------------
2530     1. In the rejigging of the build system that eventually resulted in 7.1, the
2531        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2532        brackets there is not right, since it causes compilers to look for an
2533        installed pcre.h, not the version that is in the source that is being
2534        compiled (which of course may be different). I have changed it back to:
2536          #include "pcre.h"
2538        I have a vague recollection that the change was concerned with compiling in
2539        different directories, but in the new build system, that is taken care of
2540        by the VPATH setting the Makefile.
2542     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2543        when the subject happened to end in the byte 0x85 (e.g. if the last
2544        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2545        characters but of course it shouldn't be taken as a newline when it is part
2546        of another character. The bug was that, for an unlimited repeat of . in
2547        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2548        characters when looking for a newline.
2550     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2552     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2553        in debug output.
2555     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2556        long printing in the pcrecpp unittest when running under MinGW.
2558     6. ESC_K was left out of the EBCDIC table.
2560     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2561        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2562        limit also applies to "virtual nesting" when a pattern is recursive, and in
2563        this case 1000 isn't so big. I have been able to remove this limit at the
2564        expense of backing off one optimization in certain circumstances. Normally,
2565        when pcre_exec() would call its internal match() function recursively and
2566        immediately return the result unconditionally, it uses a "tail recursion"
2567        feature to save stack. However, when a subpattern that can match an empty
2568        string has an unlimited repetition quantifier, it no longer makes this
2569        optimization. That gives it a stack frame in which to save the data for
2570        checking that an empty string has been matched. Previously this was taken
2571        from the 1000-entry workspace that had been reserved. So now there is no
2572        explicit limit, but more stack is used.
2574     8. Applied Daniel's patches to solve problems with the import/export magic
2575        syntax that is required for Windows, and which was going wrong for the
2576        pcreposix and pcrecpp parts of the library. These were overlooked when this
2577        problem was solved for the main library.
2579     9. There were some crude static tests to avoid integer overflow when computing
2580        the size of patterns that contain repeated groups with explicit upper
2581        limits. As the maximum quantifier is 65535, the maximum group length was
2582        set at 30,000 so that the product of these two numbers did not overflow a
2583        32-bit integer. However, it turns out that people want to use groups that
2584        are longer than 30,000 bytes (though not repeat them that many times).
2585        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2586        made it possible to implement the integer overflow checks in a much more
2587        dynamic way, which I have now done. The artificial limitation on group
2588        length has been removed - we now have only the limit on the total length of
2589        the compiled pattern, which depends on the LINK_SIZE setting.
2591    10. Fixed a bug in the documentation for get/copy named substring when
2592        duplicate names are permitted. If none of the named substrings are set, the
2593        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2594        empty string.
2596    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2597        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2598        because the ] is interpreted as the first data character and the
2599        terminating ] is not found. PCRE has been made compatible with Perl in this
2600        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2601        cause memory overwriting.
2603    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2604        string has been matched (to stop an infinite loop). It was not recognizing
2605        a conditional subpattern that could match an empty string if that
2606        subpattern was within another subpattern. For example, it looped when
2607        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2608        condition was not nested. This bug has been fixed.
2610    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2611        past the start of the subject in the presence of bytes with the top bit
2612        set, for example "\x8aBCD".
2614    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2615        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2617    14. Optimized (?!) to (*FAIL).
2619    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2620        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2621        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2622        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2623        does: it's just the validity check that is more restrictive.
2625    16. Inserted checks for integer overflows during escape sequence (backslash)
2626        processing, and also fixed erroneous offset values for syntax errors during
2627        backslash processing.
2629    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2630        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2632    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2633        caused an overrun.
2635    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2636        something other than just ASCII characters) inside a group that had an
2637        unlimited repeat caused a loop at compile time (while checking to see
2638        whether the group could match an empty string).
2640    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2641        [\P{Any}] did so. (Error in the code for printing property names.)
2643    21. An orphan \E inside a character class could cause a crash.
2645    22. A repeated capturing bracket such as (A)? could cause a wild memory
2646        reference during compilation.
2648    23. There are several functions in pcre_compile() that scan along a compiled
2649        expression for various reasons (e.g. to see if it's fixed length for look
2650        behind). There were bugs in these functions when a repeated \p or \P was
2651        present in the pattern. These operators have additional parameters compared
2652        with \d, etc, and these were not being taken into account when moving along
2653        the compiled data. Specifically:
2655        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2656            length.
2658        (b) An item such as \pL+ within a repeated group could cause crashes or
2659            loops.
2661        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2662            "reference to non-existent subpattern" error.
2664        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2666    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2667        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2669    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2671    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2672        character were causing crashes (broken optimization).
2674    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2675        \p or \P) caused a compile-time loop.
2677    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2678        break. For example, the unstudied pattern [\r\n]A does not match the string
2679        "\r\nA" because change 7.0/46 below moves the current point on by two
2680        characters after failing to match at the start. However, the pattern \nA
2681        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2682        the same is true. There doesn't seem any very clean way out of this, but
2683        what I have chosen to do makes the common cases work: PCRE now takes note
2684        of whether there can be an explicit match for \r or \n anywhere in the
2685        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2686        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2687        pattern has explicit CR or LF references.
2689    29. Added (*CR) etc for changing newline setting at start of pattern.
2692    Version 7.2 19-Jun-07
2693    ---------------------
2695     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2696        which is apparently normally available under Windows.
2698     2. Re-jig the pcregrep tests with different newline settings in an attempt
2699        to make them independent of the local environment's newline setting.
2701     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2703     4. Some of the "internals" tests were previously cut out when the link size
2704        was not 2, because the output contained actual offsets. The recent new
2705        "Z" feature of pcretest means that these can be cut out, making the tests
2706        usable with all link sizes.
2708     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2709        stack recursion. This gives a massive performance boost under BSD, but just
2710        a small improvement under Linux. However, it saves one field in the frame
2711        in all cases.
2713     6. Added more features from the forthcoming Perl 5.10:
2715        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2716            recursion call. It refers to the nth most recently opened parentheses.
2718        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2719            to be opened parentheses.
2721        (c) Conditions that refer to capturing parentheses can be specified
2722            relatively, for example, (?(-2)... or (?(+3)...
2724        (d) \K resets the start of the current match so that everything before
2725            is not part of it.
2727        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2729        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2730            reference syntax.
2732        (g) (?| introduces a group in which the numbering of parentheses in each
2733            alternative starts with the same number.
2735        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2737     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2740     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2741        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2742        for detecting groups that can match an empty string.
2744     9. A pattern with a very large number of alternatives (more than several
2745        hundred) was running out of internal workspace during the pre-compile
2746        phase, where pcre_compile() figures out how much memory will be needed. A
2747        bit of new cunning has reduced the workspace needed for groups with
2748        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2749        workspace instead of running out of the 4096 that are available.
2751    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2753    11. Applied patch from Google to remove an optimization that didn't quite work.
2754        The report of the bug said:
2756          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2757          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2758          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2760    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2761        it matched the wrong number of bytes.
2764    Version 7.1 24-Apr-07
2765    ---------------------
2767     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2768        that is more "standard", making use of automake and other Autotools. There
2769        is some re-arrangement of the files and adjustment of comments consequent
2770        on this.
2772     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2773        for recursive directory scanning broke on some systems because the files
2774        are not scanned in any specific order and on different systems the order
2775        was different. A call to "sort" has been inserted into RunGrepTest for the
2776        approprate test as a short-term fix. In the longer term there may be an
2777        alternative.
2779     3. I had an email from Eric Raymond about problems translating some of PCRE's
2780        man pages to HTML (despite the fact that I distribute HTML pages, some
2781        people do their own conversions for various reasons). The problems
2782        concerned the use of low-level troff macros .br and .in. I have therefore
2783        removed all such uses from the man pages (some were redundant, some could
2784        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2785        HTML has been updated to handle .nf/.fi and to complain if it encounters
2786        .br or .in.
2788     4. Updated comments in configure.ac that get placed in config.h.in and also
2789        arranged for config.h to be included in the distribution, with the name
2790        config.h.generic, for the benefit of those who have to compile without
2791        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2793     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2794        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2795        makevp.bat for the current PCRE, using the additional files
2796        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2798     6. A Windows user reported a minor discrepancy with test 2, which turned out
2799        to be caused by a trailing space on an input line that had got lost in his
2800        copy. The trailing space was an accident, so I've just removed it.
2802     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2803        that is needed.
2805     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2806        as "const" (a) because they are and (b) because it helps the PHP
2807        maintainers who have recently made a script to detect big data structures
2808        in the php code that should be moved to the .rodata section. I remembered
2809        to update Builducptable as well, so it won't revert if ucptable.h is ever
2810        re-created.
2812     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2813        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2814        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2815        support is not required. This saves 1.5-2K of code, which is important in
2816        some applications.
2818        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2819        so as not to refer to the tables, even though these functions will never be
2820        called when UTF-8 support is disabled. Otherwise there are problems with a
2821        shared library.
2823    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2825        (a) It was defining its arguments as char * instead of void *.
2827        (b) It was assuming that all moves were upwards in memory; this was true
2828            a long time ago when I wrote it, but is no longer the case.
2830        The emulated memove() is provided for those environments that have neither
2831        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2832        is clearly not the case, as these two bugs were recently reported.
2834    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2835        and Detrail to create the HTML documentation, the .txt form of the man
2836        pages, and it removes trailing spaces from listed files. It also creates
2837        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2838        case, it wraps all the #defines with #ifndefs. This script should be run
2839        before "make dist".
2841    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2842        with Unicode property support.
2844        (a) For a maximizing quantifier, if the two different cases of the
2845            character were of different lengths in their UTF-8 codings (there are
2846            some cases like this - I found 11), and the matching function had to
2847            back up over a mixture of the two cases, it incorrectly assumed they
2848            were both the same length.
2850        (b) When PCRE was configured to use the heap rather than the stack for
2851            recursion during matching, it was not correctly preserving the data for
2852            the other case of a UTF-8 character when checking ahead for a match
2853            while processing a minimizing repeat. If the check also involved
2854            matching a wide character, but failed, corruption could cause an
2855            erroneous result when trying to check for a repeat of the original
2856            character.
2858    13. Some tidying changes to the testing mechanism:
2860        (a) The RunTest script now detects the internal link size and whether there
2861            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2862            values substituted by "configure". (The RunGrepTest script already did
2863            this for UTF-8.) The configure.ac script no longer substitutes the
2864            relevant variables.
2866        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2867            with length and offset values. This means that the output is different
2868            for different internal link sizes. Test 2 is skipped for link sizes
2869            other than 2 because of this, bypassing the problem. Unfortunately,
2870            there was also a test in test 3 (the locale tests) that used /B and
2871            failed for link sizes other than 2. Rather than cut the whole test out,
2872            I have added a new /Z option to pcretest that replaces the length and
2873            offset values with spaces. This is now used to make test 3 independent
2874            of link size. (Test 2 will be tidied up later.)
2876    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2877        segmentation fault instead of returning the appropriate error message.
2879    15. In multiline mode when the newline sequence was set to "any", the pattern
2880        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2881        This doesn't seem right; it now treats the CRLF combination as the line
2882        ending, and so does not match in that case. It's only a pattern such as ^$
2883        that would hit this one: something like ^ABC$ would have failed after \r
2884        and then tried again after \r\n.
2886    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2887        in an attempt to make files that differ only in their line terminators
2888        compare equal. This works on Linux.
2890    17. Under certain error circumstances pcregrep might try to free random memory
2891        as it exited. This is now fixed, thanks to valgrind.
2893    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2894        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2895        was because its rules for how to advance for /g after matching an empty
2896        string at the end of a line did not allow for this case. They now check for
2897        it specially.
2899    20. pcretest is supposed to handle patterns and data of any length, by
2900        extending its buffers when necessary. It was getting this wrong when the
2901        buffer for a data line had to be extended.
2903    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2904        CRLF as a newline sequence.
2906    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2907        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2908        I have nevertheless tidied it up.
2910    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2912    24. Added a man page for pcre-config.
2915    Version 7.0 19-Dec-06
2916    ---------------------
2918     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2919        moving to gcc 4.1.1.
2921     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2922        sys/time.h, which is documented as needed for this function. It doesn't
2923        seem to matter on Linux, but it showed up on some releases of OS X.
2925     3. It seems that there are systems where bytes whose values are greater than
2926        127 match isprint() in the "C" locale. The "C" locale should be the
2927        default when a C program starts up. In most systems, only ASCII printing
2928        characters match isprint(). This difference caused the output from pcretest
2929        to vary, making some of the tests fail. I have changed pcretest so that:
2931        (a) When it is outputting text in the compiled version of a pattern, bytes
2932            other than 32-126 are always shown as hex escapes.
2934        (b) When it is outputting text that is a matched part of a subject string,
2935            it does the same, unless a different locale has been set for the match
2936            (using the /L modifier). In this case, it uses isprint() to decide.
2938     4. Fixed a major bug that caused incorrect computation of the amount of memory
2939        required for a compiled pattern when options that changed within the
2940        pattern affected the logic of the preliminary scan that determines the
2941        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2942        that the computed length was too small. The symptoms of this bug were
2943        either the PCRE error "internal error: code overflow" from pcre_compile(),
2944        or a glibc crash with a message such as "pcretest: free(): invalid next
2945        size (fast)". Examples of patterns that provoked this bug (shown in
2946        pcretest format) are:
2948          /(?-x: )/x
2949          /(?x)(?-x: \s*#\s*)/
2950          /((?i)[\x{c0}])/8
2951          /(?i:[\x{c0}])/8
2953        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2954        is now done differently.
2956     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2957        wrapper classes; (b) implement a new function in the C++ scanner that is
2958        more efficient than the old way of doing things because it avoids levels of
2959        recursion in the regex matching; (c) add a paragraph to the documentation
2960        for the FullMatch() function.
2962     6. The escape sequence \n was being treated as whatever was defined as
2963        "newline". Not only was this contrary to the documentation, which states
2964        that \n is character 10 (hex 0A), but it also went horribly wrong when
2965        "newline" was defined as CRLF. This has been fixed.
2967     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2968        was being set to -1 for the "end of line" case (supposedly a value that no
2969        character can have). Though this value is never used (the check for end of
2970        line is "zero bytes in current character"), it caused compiler complaints.
2971        I've changed it to 0xffffffff.
2973     8. In pcre_version.c, the version string was being built by a sequence of
2974        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2975        string (as it is for production releases) called a macro with an empty
2976        argument. The C standard says the result of this is undefined. The gcc
2977        compiler treats it as an empty string (which was what was wanted) but it is
2978        reported that Visual C gives an error. The source has been hacked around to
2979        avoid this problem.
2981     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2982        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2983        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2984        of them did).
2986    10. Originally, pcretest opened its input and output without "b"; then I was
2987        told that "b" was needed in some environments, so it was added for release
2988        5.0 to both the input and output. (It makes no difference on Unix-like
2989        systems.) Later I was told that it is wrong for the input on Windows. I've
2990        now abstracted the modes into two macros, to make it easier to fiddle with
2991        them, and removed "b" from the input mode under Windows.
2993    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2995    12. Added -help and --help to pcretest as an official way of being reminded
2996        of the options.
2998    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2999        and pcrecpp.cc because they annoy compilers at high warning levels.
3001    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3003    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3004        scripts are not C programs :-) and which was not noticed because it works
3005        on Linux.
3007    16. pcretest is supposed to handle any length of pattern and data line (as one
3008        line or as a continued sequence of lines) by extending its input buffer if
3009        necessary. This feature was broken for very long pattern lines, leading to
3010        a string of junk being passed to pcre_compile() if the pattern was longer
3011        than about 50K.
3013    17. I have done a major re-factoring of the way pcre_compile() computes the
3014        amount of memory needed for a compiled pattern. Previously, there was code
3015        that made a preliminary scan of the pattern in order to do this. That was
3016        OK when PCRE was new, but as the facilities have expanded, it has become
3017        harder and harder to keep it in step with the real compile phase, and there
3018        have been a number of bugs (see for example, 4 above). I have now found a
3019        cunning way of running the real compile function in a "fake" mode that
3020        enables it to compute how much memory it would need, while actually only
3021        ever using a few hundred bytes of working memory and without too many
3022        tests of the mode. This should make future maintenance and development
3023        easier. A side effect of this work is that the limit of 200 on the nesting
3024        depth of parentheses has been removed (though this was never a serious
3025        limitation, I suspect). However, there is a downside: pcre_compile() now
3026        runs more slowly than before (30% or more, depending on the pattern). I
3027        hope this isn't a big issue. There is no effect on runtime performance.
3029    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3030        newline (only possible for the last line of a file) and it was a
3031        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3033    19. Added additional timing features to pcretest. (1) The -tm option now times
3034        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3035        separate command line item, by a number that specifies the number of
3036        repeats to use when timing. The default is 50000; this gives better
3037        precision, but takes uncomfortably long for very large patterns.
3039    20. Extended pcre_study() to be more clever in cases where a branch of a
3040        subpattern has no definite first character. For example, (a*|b*)[cd] would
3041        previously give no result from pcre_study(). Now it recognizes that the
3042        first character must be a, b, c, or d.
3044    21. There was an incorrect error "recursive call could loop indefinitely" if
3045        a subpattern (or the entire pattern) that was being tested for matching an
3046        empty string contained only one non-empty item after a nested subpattern.
3047        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3048        incorrectly, because the \d was being skipped in the check.
3050    22. The pcretest program now has a new pattern option /B and a command line
3051        option -b, which is equivalent to adding /B to every pattern. This causes
3052        it to show the compiled bytecode, without the additional information that
3053        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3054        is the same as /B/I).
3056    23. A new optimization is now able automatically to treat some sequences such
3057        as a*b as a*+b. More specifically, if something simple (such as a character
3058        or a simple class like \d) has an unlimited quantifier, and is followed by
3059        something that cannot possibly match the quantified thing, the quantifier
3060        is automatically "possessified".
3062    24. A recursive reference to a subpattern whose number was greater than 39
3063        went wrong under certain circumstances in UTF-8 mode. This bug could also
3064        have affected the operation of pcre_study().
3066    25. Realized that a little bit of performance could be had by replacing
3067        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3069    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3071    27. Possessive quantifiers such as a++ were previously implemented by turning
3072        them into atomic groups such as ($>a+). Now they have their own opcodes,
3073        which improves performance. This includes the automatically created ones
3074        from 23 above.
3076    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3077        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3078        the first matched character to be a colon. This applied both to named and
3079        numbered groups.
3081    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3083    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3084        building PCRE on Windows easier, so I have included it in the distribution.
3086    31. There is now a check in pcretest against a ridiculously large number being
3087        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3088        loop, the loop is abandoned.
3090    32. Forward references to subpatterns in conditions such as (?(2)...) where
3091        subpattern 2 is defined later cause pcre_compile() to search forwards in
3092        the pattern for the relevant set of parentheses. This search went wrong
3093        when there were unescaped parentheses in a character class, parentheses
3094        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3096    33. "Subroutine" calls and backreferences were previously restricted to
3097        referencing subpatterns earlier in the regex. This restriction has now
3098        been removed.
3100    34. Added a number of extra features that are going to be in Perl 5.10. On the
3101        whole, these are just syntactic alternatives for features that PCRE had
3102        previously implemented using the Python syntax or my own invention. The
3103        other formats are all retained for compatibility.
3105        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3106            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3107            also .NET compatible.
3109        (b) A recursion or subroutine call to a named group can now be defined as
3110            (?&name) as well as (?P>name).
3112        (c) A backreference to a named group can now be defined as \k<name> or
3113            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3114            5.10, are also .NET compatible.
3116        (d) A conditional reference to a named group can now use the syntax
3117            (?(<name>) or (?('name') as well as (?(name).
3119        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3120            groups (named and numbered) that are never evaluated inline, but can be
3121            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3122            is always false. There may be only one alternative in such a group.
3124        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3125            as the simple (?(R). The condition is true only if the most recent
3126            recursion is that of the given number or name. It does not search out
3127            through the entire recursion stack.
3129        (g) The escape \gN or \g{N} has been added, where N is a positive or
3130            negative number, specifying an absolute or relative reference.
3132    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3133        some "unreachable code" warnings.
3135    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3136        things, this adds five new scripts.
3138    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3139        There were also incompatibilities regarding the handling of \Q..\E inside
3140        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3141        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3143    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3144        matches an empty string, and forcibly breaks the loop. There were bugs in
3145        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3146        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3147        separate and independent bugs (that affected different cases) have been
3148        fixed.
3150    39. Refactored the code to abolish the use of different opcodes for small
3151        capturing bracket numbers. This is a tidy that I avoided doing when I
3152        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3153        The new approach is not only tidier, it makes it possible to reduce the
3154        memory needed to fix the previous bug (38).
3156    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3157        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3158        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3159        mode.
3161    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3162        report.
3164    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3165        copy construction and assignment in the C++ wrapper.
3167    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3168        couple of bugs that could have given wrong results in the "--newline=crlf"
3169        case.
3171    44. Added a number of casts and did some reorganization of signed/unsigned int
3172        variables following suggestions from Dair Grant. Also renamed the variable
3173        "this" as "item" because it is a C++ keyword.
3175    45. Arranged for dftables to add
3177          #include "pcre_internal.h"
3179        to pcre_chartables.c because without it, gcc 4.x may remove the array
3180        definition from the final binary if PCRE is built into a static library and
3181        dead code stripping is activated.
3183    46. For an unanchored pattern, if a match attempt fails at the start of a
3184        newline sequence, and the newline setting is CRLF or ANY, and the next two
3185        characters are CRLF, advance by two characters instead of one.
3188    Version 6.7 04-Jul-06
3189    ---------------------
3191     1. In order to handle tests when input lines are enormously long, pcretest has
3192        been re-factored so that it automatically extends its buffers when
3193        necessary. The code is crude, but this _is_ just a test program. The
3194        default size has been increased from 32K to 50K.
3196     2. The code in pcre_study() was using the value of the re argument before
3197        testing it for NULL. (Of course, in any sensible call of the function, it
3198        won't be NULL.)
3200     3. The memmove() emulation function in pcre_internal.h, which is used on
3201        systems that lack both memmove() and bcopy() - that is, hardly ever -
3202        was missing a "static" storage class specifier.
3204     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3205        containing an extended class (one that cannot be represented by a bitmap
3206        because it contains high-valued characters or Unicode property items, e.g.
3207        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3208        pattern, but PCRE should not loop if you do not (it no longer does).
3209        [Detail: two cases were found: (a) a repeated subpattern containing an
3210        extended class; (b) a recursive reference to a subpattern that followed a
3211        previous extended class. It wasn't skipping over the extended class
3212        correctly when UTF-8 mode was not set.]
3214     5. A negated single-character class was not being recognized as fixed-length
3215        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3216        compile error "lookbehind assertion is not fixed length".
3218     6. The RunPerlTest auxiliary script was showing an unexpected difference
3219        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3220        write a Perl script that can interpret lines of an input file either as
3221        byte characters or as UTF-8, which is what "perltest" was being required to
3222        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3223        can't do is switch easily at run time between having the "use utf8;" pragma
3224        or not. In the end, I fudged it by using the RunPerlTest script to insert
3225        "use utf8;" explicitly for the UTF-8 tests.
3227     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3228        the end of the subject string, contrary to the documentation and to what
3229        Perl does. This was true of both matching functions. Now it matches only at
3230        the start of the subject and immediately after *internal* newlines.
3232     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3233        a pointer to an int instead of a pointer to an unsigned long int. This
3234        caused problems on 64-bit systems.
3236     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3237        instance of the 'standard' template library not being so standard".
3239    10. There was no check on the number of named subpatterns nor the maximum
3240        length of a subpattern name. The product of these values is used to compute
3241        the size of the memory block for a compiled pattern. By supplying a very
3242        long subpattern name and a large number of named subpatterns, the size
3243        computation could be caused to overflow. This is now prevented by limiting
3244        the length of names to 32 characters, and the number of named subpatterns
3245        to 10,000.
3247    11. Subpatterns that are repeated with specific counts have to be replicated in
3248        the compiled pattern. The size of memory for this was computed from the
3249        length of the subpattern and the repeat count. The latter is limited to
3250        65535, but there was no limit on the former, meaning that integer overflow
3251        could in principle occur. The compiled length of a repeated subpattern is
3252        now limited to 30,000 bytes in order to prevent this.
3254    12. Added the optional facility to have named substrings with the same name.
3256    13. Added the ability to use a named substring as a condition, using the
3257        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3258        are numbers (not recommended). Forward references are permitted.
3260    14. Added forward references in named backreferences (if you see what I mean).
3262    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3263        pattern could run off the end of the subject. For example, the pattern
3264        "(?s)(.{1,5})"8 did this with the subject "ab".
3266    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3267        PCRE_CASELESS was set when matching characters that were quantified with ?
3268        or *.
3270    17. A character class other than a single negated character that had a minimum
3271        but no maximum quantifier - for example [ab]{6,} - was not handled
3272        correctly by pce_dfa_exec(). It would match only one character.
3274    18. A valid (though odd) pattern that looked like a POSIX character
3275        class but used an invalid character after [ (for example [[,abc,]]) caused
3276        pcre_compile() to give the error "Failed: internal error: code overflow" or
3277        in some cases to crash with a glibc free() error. This could even happen if
3278        the pattern terminated after [[ but there just happened to be a sequence of
3279        letters, a binary zero, and a closing ] in the memory that followed.
3281    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3282        over the years. Originally (before any Unicode support), just the bottom 8
3283        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3284        output from "man perlunicode" includes this:
3286          The regular expression compiler produces polymorphic opcodes.  That
3287          is, the pattern adapts to the data and automatically switches to
3288          the Unicode character scheme when presented with Unicode data--or
3289          instead uses a traditional byte scheme when presented with byte
3290          data.
3292        Sadly, a wide octal escape does not cause a switch, and in a string with
3293        no other multibyte characters, these octal escapes are treated as before.
3294        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3295        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3296        Unicode string.
3298        I have not perpetrated such confusion in PCRE. Up till now, it took just
3299        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3300        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3301        translate to the appropriate multibyte character.
3303    29. Applied some refactoring to reduce the number of warnings from Microsoft
3304        and Borland compilers. This has included removing the fudge introduced
3305        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3306        a warning about an unused variable.
3308    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3309        characters since release 4.0, because Perl (from release 5.004) does not.
3310        [Or at least, is documented not to: some releases seem to be in conflict
3311        with the documentation.] However, when a pattern was studied with
3312        pcre_study() and all its branches started with \s, PCRE still included VT
3313        as a possible starting character. Of course, this did no harm; it just
3314        caused an unnecessary match attempt.
3316    22. Removed a now-redundant internal flag bit that recorded the fact that case
3317        dependency changed within the pattern. This was once needed for "required
3318        byte" processing, but is no longer used. This recovers a now-scarce options
3319        bit. Also moved the least significant internal flag bit to the most-
3320        significant bit of the word, which was not previously used (hangover from
3321        the days when it was an int rather than a uint) to free up another bit for
3322        the future.
3324    23. Added support for CRLF line endings as well as CR and LF. As well as the
3325        default being selectable at build time, it can now be changed at runtime
3326        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3327        specify that it is scanning data with non-default line endings.
3329    24. Changed the definition of CXXLINK to make it agree with the definition of
3330        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3332    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3333        recursions. This makes a big different to stack usage for some patterns.
3335    26. If a subpattern containing a named recursion or subroutine reference such
3336        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3337        the space required for the compiled pattern went wrong and gave too small a
3338        value. Depending on the environment, this could lead to "Failed: internal
3339        error: code overflow at offset 49" or "glibc detected double free or
3340        corruption" errors.
3342    27. Applied patches from Google (a) to support the new newline modes and (b) to
3343        advance over multibyte UTF-8 characters in GlobalReplace.
3345    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3346        difference for some implementation of PCRE in some Windows version.
3348    29. Added some extra testing facilities to pcretest:
3350        \q<number>   in a data line sets the "match limit" value
3351        \Q<number>   in a data line sets the "match recursion limt" value
3352        -S <number>  sets the stack size, where <number> is in megabytes
3354        The -S option isn't available for Windows.
3357    Version 6.6 06-Feb-06
3358    ---------------------
3360     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3361        in pcreposix.h. I have copied the definition from pcre.h.