ViewVC logotype

Diff of /code/trunk/ChangeLog

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

revision 45 by nigel, Sat Feb 24 21:39:25 2007 UTC revision 1482 by ph10, Wed May 28 11:10:58 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
4    Version 8.36 xx-xxx-2014
5    ------------------------
7    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
8        -Wmissing-field-initializers and -Wunused-parameter.
10    2.  The tests for quantifiers being too big (greater than 65535) were being
11        applied after reading the number, and stupidly assuming that integer
12        overflow would give a negative number. The tests are now applied as the
13        numbers are read.
15    3.  Tidy code in pcre_exec.c where two branches that used to be different are
16        now the same.
18    4.  The JIT compiler did not generate match limit checks for certain
19        bracketed expressions with quantifiers. This may lead to exponential
20        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
21        issue should be resolved now.
23    5.  Fixed an issue, which occures when nested alternatives are optimized
24        with table jumps.
26    6.  Inserted two casts and changed some ints to size_t in the light of some
27        reported 64-bit compiler warnings (Bugzilla 1477).
29    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
30        an empty string, which sometimes were behaving incorrectly in the
31        interpreter (though correctly in the JIT matcher). This pcretest input is
32        an example:
34          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
37        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
38        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
39        for an empty string was breaking the inner loop and carrying on at a lower
40        level, when possessive repeated groups should always return to a higher
41        level as they have no backtrack points in them. The empty string test now
42        occurs at the outer level.
44    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
45        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
47    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
48        doesn't).
50    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
51        to hold a pattern to be incorrectly computed (too small) when there were
52        named back references to duplicated names. This could cause "internal
53        error: code overflow" or "double free or corruption" or other memory
54        handling errors.
56    11. When named subpatterns had the same prefixes, back references could be
57        confused. For example, in this pattern:
59          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
61        the reference to 'Name' was incorrectly treated as a reference to a
62        duplicate name.
65    Version 8.35 04-April-2014
66    --------------------------
68    1.  A new flag is set, when property checks are present in an XCLASS.
69        When this flag is not set, PCRE can perform certain optimizations
70        such as studying these XCLASS-es.
72    2.  The auto-possessification of character sets were improved: a normal
73        and an extended character set can be compared now. Furthermore
74        the JIT compiler optimizes more character set checks.
76    3.  Got rid of some compiler warnings for potentially uninitialized variables
77        that show up only when compiled with -O2.
79    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
80        of a match later then the end of the match. The pcretest program was not
81        handling the case sensibly - it was outputting from the start to the next
82        binary zero. It now reports this situation in a message, and outputs the
83        text from the end to the start.
85    5.  Fast forward search is improved in JIT. Instead of the first three
86        characters, any three characters with fixed position can be searched.
87        Search order: first, last, middle.
89    6.  Improve character range checks in JIT. Characters are read by an inprecise
90        function now, which returns with an unknown value if the character code is
91        above a certain treshold (e.g: 256). The only limitation is that the value
92        must be bigger than the treshold as well. This function is useful, when
93        the characters above the treshold are handled in the same way.
95    7.  The macros whose names start with RAWUCHAR are placeholders for a future
96        mode in which only the bottom 21 bits of 32-bit data items are used. To
97        make this more memorable for those maintaining the code, the names have
98        been changed to start with UCHAR21, and an extensive comment has been added
99        to their definition.
101    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
102        to the export list in Makefile.am (they were accidentally omitted from the
103        8.34 tarball).
105    9.  The informational output from pcretest used the phrase "starting byte set"
106        which is inappropriate for the 16-bit and 32-bit libraries. As the output
107        for "first char" and "need char" really means "non-UTF-char", I've changed
108        "byte" to "char", and slightly reworded the output. The documentation about
109        these values has also been (I hope) clarified.
111    10. Another JIT related optimization: use table jumps for selecting the correct
112        backtracking path, when more than four alternatives are present inside a
113        bracket.
115    11. Empty match is not possible, when the minimum length is greater than zero,
116        and there is no \K in the pattern. JIT should avoid empty match checks in
117        such cases.
119    12. In a caseless character class with UCP support, when a character with more
120        than one alternative case was not the first character of a range, not all
121        the alternative cases were added to the class. For example, s and \x{17f}
122        are both alternative cases for S: the class [RST] was handled correctly,
123        but [R-T] was not.
125    13. The configure.ac file always checked for pthread support when JIT was
126        enabled. This is not used in Windows, so I have put this test inside a
127        check for the presence of windows.h (which was already tested for).
129    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
130        The algorithm provides a way to skip certain starting offsets, and usually
131        faster than linear prefix searches.
133    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
134        as for 'fr_FR' and 'french'. For some reason, however, it then used the
135        Windows-specific input and output files, which have 'french' screwed in.
136        So this could never have worked. One of the problems with locales is that
137        they aren't always the same. I have now updated RunTest so that it checks
138        the output of the locale test (test 3) against three different output
139        files, and it allows the test to pass if any one of them matches. With luck
140        this should make the test pass on some versions of Solaris where it was
141        failing. Because of the uncertainty, the script did not used to stop if
142        test 3 failed; it now does. If further versions of a French locale ever
143        come to light, they can now easily be added.
145    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
146        there was a message during ./configure, but it did not stop. This now
147        provokes an error. The invalid example in README has been corrected.
148        If a value less than the minimum is given, the minimum value has always
149        been used, but now a warning is given.
151    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
152        was a bug in the test system, which is now fixed. Also, the list of various
153        configurations that are tested for each release did not have one with both
154        16/32 bits and --enable-bar-anycrlf. It now does.
156    18. pcretest was missing "-C bsr" for displaying the \R default setting.
158    19. Little endian PowerPC systems are supported now by the JIT compiler.
160    20. The fast forward newline mechanism could enter to an infinite loop on
161        certain invalid UTF-8 input. Although we don't support these cases
162        this issue can be fixed by a performance optimization.
164    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
165        not take account if existing stack usage. There is now a new global
166        variable called pcre_stack_guard that can be set to point to an external
167        function to check stack availability. It is called at the start of
168        processing every parenthesized group.
170    22. A typo in the code meant that in ungreedy mode the max/min qualifier
171        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
172        not match "ab".
174    23. When UTF was disabled, the JIT program reported some incorrect compile
175        errors. These messages are silenced now.
177    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
178        compiler.
180    25. Change all the temporary files used in RunGrepTest to be different to those
181        used by RunTest so that the tests can be run simultaneously, for example by
182        "make -j check".
185    Version 8.34 15-December-2013
186    -----------------------------
188    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
189        executable memory. Patch inspired by Carsten Klein.
191    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
192        this macro is never tested and has no effect, because the work to support
193        coverage involves only compiling and linking options and special targets in
194        the Makefile. The comment in config.h implied that defining the macro would
195        enable coverage support, which is totally false. There was also support for
196        setting this macro in the CMake files (my fault, I just copied it from
197        configure). SUPPORT_GCOV has now been removed.
199    3.  Make a small performance improvement in strlen16() and strlen32() in
200        pcretest.
202    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
203        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
204        these cases). There was also one in pcretest.c.
206    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
208    6.  In UTF mode, the code for checking whether a group could match an empty
209        string (which is used for indefinitely repeated groups to allow for
210        breaking an infinite loop) was broken when the group contained a repeated
211        negated single-character class with a character that occupied more than one
212        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
213        UTF-8 mode). The effect was undefined: the group might or might not be
214        deemed as matching an empty string, or the program might have crashed.
216    7.  The code for checking whether a group could match an empty string was not
217        recognizing that \h, \H, \v, \V, and \R must match a character.
219    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
220        an empty string. If it can, pcretest shows this in its information output.
222    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
223        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
224        matched by pcre_exec() without using JIT:
226        (a) If the rest of the pattern did not match after a maximal run of
227            grapheme clusters, the code for backing up to try with fewer of them
228            did not always back up over a full grapheme when characters that do not
229            have the modifier quality were involved, e.g. Hangul syllables.
231        (b) If the match point in a subject started with modifier character, and
232            there was no match, the code could incorrectly back up beyond the match
233            point, and potentially beyond the first character in the subject,
234            leading to a segfault or an incorrect match result.
236    10. A conditional group with an assertion condition could lead to PCRE
237        recording an incorrect first data item for a match if no other first data
238        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
239        first data item, and therefore matched "ca" after "c" instead of at the
240        start.
242    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
243        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
245    12. The source of pcregrep now includes z/OS-specific code so that it can be
246        compiled for z/OS as part of the special z/OS distribution.
248    13. Added the -T and -TM options to pcretest.
250    14. The code in pcre_compile.c for creating the table of named capturing groups
251        has been refactored. Instead of creating the table dynamically during the
252        actual compiling pass, the information is remembered during the pre-compile
253        pass (on the stack unless there are more than 20 named groups, in which
254        case malloc() is used) and the whole table is created before the actual
255        compile happens. This has simplified the code (it is now nearly 150 lines
256        shorter) and prepared the way for better handling of references to groups
257        with duplicate names.
259    15. A back reference to a named subpattern when there is more than one of the
260        same name now checks them in the order in which they appear in the pattern.
261        The first one that is set is used for the reference. Previously only the
262        first one was inspected. This change makes PCRE more compatible with Perl.
264    16. Unicode character properties were updated from Unicode 6.3.0.
266    17. The compile-time code for auto-possessification has been refactored, based
267        on a patch by Zoltan Herczeg. It now happens after instead of during
268        compilation. The code is cleaner, and more cases are handled. The option
269        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
270        options in pcretest are provided to set it. It can also be set by
271        (*NO_AUTO_POSSESS) at the start of a pattern.
273    18. The character VT has been added to the default ("C" locale) set of
274        characters that match \s and are generally treated as white space,
275        following this same change in Perl 5.18. There is now no difference between
276        "Perl space" and "POSIX space". Whether VT is treated as white space in
277        other locales depends on the locale.
279    19. The code for checking named groups as conditions, either for being set or
280        for being recursed, has been refactored (this is related to 14 and 15
281        above). Processing unduplicated named groups should now be as fast at
282        numerical groups, and processing duplicated groups should be faster than
283        before.
285    20. Two patches to the CMake build system, by Alexander Barkov:
287          (1) Replace the "source" command by "." in CMakeLists.txt because
288              "source" is a bash-ism.
290          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
291              without these the CMake build does not work on Solaris.
293    21. Perl has changed its handling of \8 and \9. If there is no previously
294        encountered capturing group of those numbers, they are treated as the
295        literal characters 8 and 9 instead of a binary zero followed by the
296        literals. PCRE now does the same.
298    22. Following Perl, added \o{} to specify codepoints in octal, making it
299        possible to specify values greater than 0777 and also making them
300        unambiguous.
302    23. Perl now gives an error for missing closing braces after \x{... instead of
303        treating the string as literal. PCRE now does the same.
305    24. RunTest used to grumble if an inappropriate test was selected explicitly,
306        but just skip it when running all tests. This make it awkward to run ranges
307        of tests when one of them was inappropriate. Now it just skips any
308        inappropriate tests, as it always did when running all tests.
310    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
311        character types such as \d or \w, too many callouts were inserted, and the
312        data that they returned was rubbish.
314    26. In UCP mode, \s was not matching two of the characters that Perl matches,
315        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
316        were matched by \h. The code has now been refactored so that the lists of
317        the horizontal and vertical whitespace characters used for \h and \v (which
318        are defined only in one place) are now also used for \s.
320    27. Add JIT support for the 64 bit TileGX architecture.
321        Patch by Jiong Wang (Tilera Corporation).
323    28. Possessive quantifiers for classes (both explicit and automatically
324        generated) now use special opcodes instead of wrapping in ONCE brackets.
326    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
327        (because it's meaningless), this was not happening when PCRE_CASELESS was
328        set. Not wrong, but inefficient.
330    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
331        properties for \w, \d, etc) is present in a test regex. Otherwise if the
332        test contains no characters greater than 255, Perl doesn't realise it
333        should be using Unicode semantics.
335    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
336        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
337        does in Unicode mode.
339    32. Added the "forbid" facility to pcretest so that putting tests into the
340        wrong test files can sometimes be quickly detected.
342    33. There is now a limit (default 250) on the depth of nesting of parentheses.
343        This limit is imposed to control the amount of system stack used at compile
344        time. It can be changed at build time by --with-parens-nest-limit=xxx or
345        the equivalent in CMake.
347    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
348        errors. Perl warns for these when in warning mode, but PCRE has no facility
349        for giving warnings.
351    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
352        However, this was not working for (?!) because it is optimized to (*FAIL),
353        for which PCRE does not allow quantifiers. The optimization is now disabled
354        when a quantifier follows (?!). I can't see any use for this, but it makes
355        things uniform.
357    36. Perl no longer allows group names to start with digits, so I have made this
358        change also in PCRE. It simplifies the code a bit.
360    37. In extended mode, Perl ignores spaces before a + that indicates a
361        possessive quantifier. PCRE allowed a space before the quantifier, but not
362        before the possessive +. It now does.
364    38. The use of \K (reset reported match start) within a repeated possessive
365        group such as (a\Kb)*+ was not working.
367    40. Document that the same character tables must be used at compile time and
368        run time, and that the facility to pass tables to pcre_exec() and
369        pcre_dfa_exec() is for use only with saved/restored patterns.
371    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
372        features for Builds with MSVC:
374        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
375           builds). The .rc files can be used to set FileDescription and many other
376           attributes.
378        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
379           This allows higher-level build scripts which want .pdb files to avoid
380           hard-coding the exact files needed."
382    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
383        mean "start of word" and "end of word", respectively, as a transition aid.
385    43. A minimizing repeat of a class containing codepoints greater than 255 in
386        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
387        compiled to use the heap for recursion.
389    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
390        is configured.
393    Version 8.33 28-May-2013
394    ------------------------
396    1.  Added 'U' to some constants that are compared to unsigned integers, to
397        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
398        variables that are added to signed variables, to ensure the result is
399        signed and can be negated.
401    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
402        CMake config files.
404    3.  Revise the creation of config.h.generic so that all boolean macros are
405        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
406        overriding via -D on the command line possible.
408    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
409        to unsigned int is reported to make a quite noticeable speed difference in
410        a specific Windows environment. Testing on Linux did also appear to show
411        some benefit (and it is clearly not harmful). Also fixed the definition of
412        Xop which should be unsigned.
414    5.  Related to (4), changing the definition of the intermediate variable cc
415        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
416        improvements.
418    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
419        unnecessary spaces.
421    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
422        and later.
424    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
426    9.  Optimizing fast_forward_start_bits in JIT.
428    10. Adding support for callouts in JIT, and fixing some issues revealed
429        during this work. Namely:
431        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
433        (b) Minimum length was not checked before the matching is started.
435    11. The value of capture_last that is passed to callouts was incorrect in some
436        cases when there was a capture on one path that was subsequently abandoned
437        after a backtrack. Also, the capture_last value is now reset after a
438        recursion, since all captures are also reset in this case.
440    12. The interpreter no longer returns the "too many substrings" error in the
441        case when an overflowing capture is in a branch that is subsequently
442        abandoned after a backtrack.
444    13. In the pathological case when an offset vector of size 2 is used, pcretest
445        now prints out the matched string after a yield of 0 or 1.
447    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
448        Only supported by the JIT compiler at the moment.
450    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
452    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
453        offset of the starting point of the matching process, provided the offsets
454        vector is large enough.
456    17. The \A escape now records a lookbehind value of 1, though its execution
457        does not actually inspect the previous character. This is to ensure that,
458        in partial multi-segment matching, at least one character from the old
459        segment is retained when a new segment is processed. Otherwise, if there
460        are no lookbehinds in the pattern, \A might match incorrectly at the start
461        of a new segment.
463    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
465    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
466        gives some modest performance improvement in 8-bit mode.
468    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
469        be expressed in certain programming languages using Universal Character
470        Names.
472    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
473        which points out that "non characters" are not "characters that may not
474        appear in Unicode strings" but rather "characters that are reserved for
475        internal use and have only local meaning".
477    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
478        there was a conditional group that depended on an assertion, if the
479        assertion was false, the callout that immediately followed the alternation
480        in the condition was skipped when pcre_exec() was used for matching.
482    23. Allow an explicit callout to be inserted before an assertion that is the
483        condition for a conditional group, for compatibility with automatic
484        callouts, which always insert a callout at this point.
486    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
487        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
489    25. (*PRUNE) is now supported by the JIT compiler.
491    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
493    27. Fix the case where there are two or more SKIPs with arguments that may be
494        ignored.
496    28. (*SKIP) is now supported by the JIT compiler.
498    29. (*THEN) is now supported by the JIT compiler.
500    30. Update RunTest with additional test selector options.
502    31. The way PCRE handles backtracking verbs has been changed in two ways.
504        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
505        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
506        backtracking. In some cases this makes it more Perl-compatible, but Perl's
507        rather obscure rules do not always do the same thing.
509        (2) Previously, backtracking verbs were confined within assertions. This is
510        no longer the case for positive assertions, except for (*ACCEPT). Again,
511        this sometimes improves Perl compatibility, and sometimes does not.
513    32. A number of tests that were in test 2 because Perl did things differently
514        have been moved to test 1, because either Perl or PCRE has changed, and
515        these tests are now compatible.
517    32. Backtracking control verbs are now handled in the same way in JIT and
518        interpreter.
520    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
521        contained a forward subroutine reference caused a compile error.
523    34. Auto-detect and optimize limited repetitions in JIT.
525    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
526        blocking (*UTF) etc.
528    36. In the interpreter, maximizing pattern repetitions for characters and
529        character types now use tail recursion, which reduces stack usage.
531    37. The value of the max lookbehind was not correctly preserved if a compiled
532        and saved regex was reloaded on a host of different endianness.
534    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
535        of the compiled pattern block, expand the flags field from 16 to 32 bits
536        because it was almost full.
538    39. Try madvise first before posix_madvise.
540    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
541        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
542        original change locked out all matches of empty strings. This has been
543        changed so that one match of an empty string per line is recognized.
544        Subsequent searches on the same line (for colouring or for --only-matching,
545        for example) do not recognize empty strings.
547    41. Applied a user patch to fix a number of spelling mistakes in comments.
549    42. Data lines longer than 65536 caused pcretest to crash.
551    43. Clarified the data type for length and startoffset arguments for pcre_exec
552        and pcre_dfa_exec in the function-specific man pages, where they were
553        explicitly stated to be in bytes, never having been updated. I also added
554        some clarification to the pcreapi man page.
556    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
557        a segmentation fault.
560    Version 8.32 30-November-2012
561    -----------------------------
563    1.  Improved JIT compiler optimizations for first character search and single
564        character iterators.
566    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
567        Patch by Daniel Richard G.
569    3.  Single character iterator optimizations in the JIT compiler.
571    4.  Improved JIT compiler optimizations for character ranges.
573    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
574        Reported by Giuseppe D'Angelo.
576    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
577        of a line, was being set incorrectly in cases where .* appeared inside
578        atomic brackets at the start of a pattern, or where there was a subsequent
579        *PRUNE or *SKIP.
581    7.  Improved instruction cache flush for POWER/PowerPC.
582        Patch by Daniel Richard G.
584    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
585        grep:
587        (a) There is now no limit to the number of patterns to be matched.
589        (b) An error is given if a pattern is too long.
591        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
592            are now supported.
594        (d) --exclude-from and --include-from (multiple use) have been added.
596        (e) Exclusions and inclusions now apply to all files and directories, not
597            just to those obtained from scanning a directory recursively.
599        (f) Multiple uses of -f and --file-list are now supported.
601        (g) In a Windows environment, the default for -d has been changed from
602            "read" (the GNU grep default) to "skip", because otherwise the presence
603            of a directory in the file list provokes an error.
605        (h) The documentation has been revised and clarified in places.
607    9.  Improve the matching speed of capturing brackets.
609    10. Changed the meaning of \X so that it now matches a Unicode extended
610        grapheme cluster.
612    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
613        out POSIX threads when JIT support is configured.
615    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
617    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
618        structure could go wrong in environments where size_t is not the same size
619        as int.
621    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
622        set.
624    15. The EBCDIC support had decayed; later updates to the code had included
625        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
626        general tidy up of EBCDIC-related issues, and the documentation was also
627        not quite right. There is now a test that can be run on ASCII systems to
628        check some of the EBCDIC-related things (but is it not a full test).
630    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
631        in a small tidy to the code.
633    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
635    18. If the --only-matching (-o) option in pcregrep is specified multiple
636        times, each one causes appropriate output. For example, -o1 -o2 outputs the
637        substrings matched by the 1st and 2nd capturing parentheses. A separating
638        string can be specified by --om-separator (default empty).
640    19. Improving the first n character searches.
642    20. Turn case lists for horizontal and vertical white space into macros so that
643        they are defined only once.
645    21. This set of changes together give more compatible Unicode case-folding
646        behaviour for characters that have more than one other case when UCP
647        support is available.
649        (a) The Unicode property table now has offsets into a new table of sets of
650            three or more characters that are case-equivalent. The MultiStage2.py
651            script that generates these tables (the pcre_ucd.c file) now scans
652            CaseFolding.txt instead of UnicodeData.txt for character case
653            information.
655        (b) The code for adding characters or ranges of characters to a character
656            class has been abstracted into a generalized function that also handles
657            case-independence. In UTF-mode with UCP support, this uses the new data
658            to handle characters with more than one other case.
660        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
661            whose other case is greater than 256 are now correctly matched
662            caselessly. Previously, the high codepoint matched the low one, but not
663            vice versa.
665        (d) The processing of \h, \H, \v, and \ in character classes now makes use
666            of the new class addition function, using character lists defined as
667            macros alongside the case definitions of 20 above.
669        (e) Caseless back references now work with characters that have more than
670            one other case.
672        (f) General caseless matching of characters with more than one other case
673            is supported.
675    22. Unicode character properties were updated from Unicode 6.2.0
677    23. Improved CMake support under Windows. Patch by Daniel Richard G.
679    24. Add support for 32-bit character strings, and UTF-32
681    25. Major JIT compiler update (code refactoring and bugfixing).
682        Experimental Sparc 32 support is added.
684    26. Applied a modified version of Daniel Richard G's patch to create
685        pcre.h.generic and config.h.generic by "make" instead of in the
686        PrepareRelease script.
688    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
689        pcre_compile.c when checking for a zero character.
691    28. Introducing a native interface for JIT. Through this interface, the compiled
692        machine code can be directly executed. The purpose of this interface is to
693        provide fast pattern matching, so several sanity checks are not performed.
694        However, feature tests are still performed. The new interface provides
695        1.4x speedup compared to the old one.
697    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
698        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
699        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
701    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
702        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
703        the "old" RFC 2279). Instead, it ended up passing a negative length to
704        pcre_exec().
706    31. Add support for GCC's visibility feature to hide internal functions.
708    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
709        "unknown -C option" after outputting 0 or 1.
711    33. There is now support for generating a code coverage report for the test
712        suite in environments where gcc is the compiler and lcov is installed. This
713        is mainly for the benefit of the developers.
715    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
716        unaddressable using valgrind annotations, allowing valgrind to detect
717        invalid memory accesses. This is mainly for the benefit of the developers.
719    25. (*UTF) can now be used to start a pattern in any of the three libraries.
721    26. Give configure error if --enable-cpp but no C++ compiler found.
724    Version 8.31 06-July-2012
725    -------------------------
727    1.  Fixing a wrong JIT test case and some compiler warnings.
729    2.  Removed a bashism from the RunTest script.
731    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
732        to unsigned type, result still unsigned" that was given by an MS compiler
733        on encountering the code "-sizeof(xxx)".
735    4.  Partial matching support is added to the JIT compiler.
737    5.  Fixed several bugs concerned with partial matching of items that consist
738        of more than one character:
740        (a) /^(..)\1/ did not partially match "aba" because checking references was
741            done on an "all or nothing" basis. This also applied to repeated
742            references.
744        (b) \R did not give a hard partial match if \r was found at the end of the
745            subject.
747        (c) \X did not give a hard partial match after matching one or more
748            characters at the end of the subject.
750        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
751            a partial match for the string "\r".
753        (e) When newline was set to CRLF, the metacharacter "." did not recognize
754            a partial match for a CR character at the end of the subject string.
756    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
757        running pcretest, the text "(JIT)" added to the output whenever JIT is
758        actually used to run the match.
760    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
761        or /S+[+] with a digit between 1 and 7.
763    8.  OP_NOT now supports any UTF character not just single-byte ones.
765    9.  (*MARK) control verb is now supported by the JIT compiler.
767    10. The command "./RunTest list" lists the available tests without actually
768        running any of them. (Because I keep forgetting what they all are.)
772    12. Applied a (slightly modified) user-supplied patch that improves performance
773        when the heap is used for recursion (compiled with --disable-stack-for-
774        recursion). Instead of malloc and free for each heap frame each time a
775        logical recursion happens, frames are retained on a chain and re-used where
776        possible. This sometimes gives as much as 30% improvement.
778    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
779        call.
781    14. As documented, (*COMMIT) is now confined to within a positive assertion.
783    15. It is now possible to link pcretest with libedit as an alternative to
784        libreadline.
786    16. (*COMMIT) control verb is now supported by the JIT compiler.
788    17. The Unicode data tables have been updated to Unicode 6.1.0.
790    18. Added --file-list option to pcregrep.
792    19. Added binary file support to pcregrep, including the -a, --binary-files,
793        -I, and --text options.
795    20. The madvise function is renamed for posix_madvise for QNX compatibility
796        reasons. Fixed by Giuseppe D'Angelo.
798    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
800    22. Changed the option for creating symbolic links for 16-bit man pages from
801        -s to -sf so that re-installing does not cause issues.
803    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
805    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
806        matches in certain environments (the workspace was not being correctly
807        retained). Also added to pcre_dfa_exec() a simple plausibility check on
808        some of the workspace data at the beginning of a restart.
810    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
811        was not doing so when it should - probably a typo introduced by SVN 528
812        (change 8.10/14).
814    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
815        \w+ when the character tables indicated that \x{c4} was a word character.
816        There were several related cases, all because the tests for doing a table
817        lookup were testing for characters less than 127 instead of 255.
819    27. If a pattern contains capturing parentheses that are not used in a match,
820        their slots in the ovector are set to -1. For those that are higher than
821        any matched groups, this happens at the end of processing. In the case when
822        there were back references that the ovector was too small to contain
823        (causing temporary malloc'd memory to be used during matching), and the
824        highest capturing number was not used, memory off the end of the ovector
825        was incorrectly being set to -1. (It was using the size of the temporary
826        memory instead of the true size.)
828    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
829        ovector size, it uses memory at the end of the block that it has got.
831    29. Check for an overlong MARK name and give an error at compile time. The
832        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
834    30. JIT compiler update.
836    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
837        Rill for the patch.
839    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
841    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
843    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
844        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
845        was enabled.
847    35. Improve JIT code generation for greedy plus quantifier.
849    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
850        1 to "aa" instead of to an empty string. The bug affected repeated groups
851        that could potentially match an empty string.
853    37. Optimizing single character iterators in JIT.
855    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
856        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
857        codepoints that are too big for the mode are faulted, and in a UTF mode,
858        disallowed codepoints are also faulted.
860    39. If PCRE was compiled with UTF support, in three places in the DFA
861        matcher there was code that should only have been obeyed in UTF mode, but
862        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
863        processing when bytes with values greater than 127 were present. In 16-bit
864        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
865        both cases the values are those that cannot be the first data item in a UTF
866        character. The three items that might have provoked this were recursions,
867        possessively repeated groups, and atomic groups.
869    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
870        and pcregrep, because some OS require shared objects to be explicitly
871        passed to ld, causing the link step to fail if they are not.
873    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
874        mode, patterns that started with \h* or \R* might be incorrectly matched.
877    Version 8.30 04-February-2012
878    -----------------------------
880    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
881        name is defined in ctype.h.
883    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
884        only in quite long subpatterns.
886    3.  Removed the function pcre_info(), which has been obsolete and deprecated
887        since it was replaced by pcre_fullinfo() in February 2000.
889    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
890        match a (*MARK), and the match failed at the start of the subject, a
891        reference to memory before the start of the subject could occur. This bug
892        was introduced by fix 17 of release 8.21.
894    5.  A reference to an unset group with zero minimum repetition was giving
895        totally wrong answers (in non-JavaScript-compatibility mode). For example,
896        /(another)?(\1?)test/ matched against "hello world test". This bug was
897        introduced in release 8.13.
899    6.  Add support for 16-bit character strings (a large amount of work involving
900        many changes and refactorings).
902    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
903        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
904        from a file.
906    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
907        rounding is not applied in this particular case).
909    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
910        if they appear, or are escaped, in patterns.
912    10. Get rid of a number of -Wunused-but-set-variable warnings.
914    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
915        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
916        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
917        also returns the mark "x". This bug applied to capturing parentheses,
918        non-capturing parentheses, and atomic parentheses. It also applied to some
919        assertions.
921    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
922        information out of configure.ac instead of relying on pcre.h.generic, which
923        is not stored in the repository.
925    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
926        -lreadline.
928    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
930    15. Applied Graycode's patch to put the top-level frame on the stack rather
931        than the heap when not using the stack for recursion. This gives a
932        performance improvement in many cases when recursion is not deep.
934    16. Experimental code added to "pcretest -C" to output the stack frame size.
937    Version 8.21 12-Dec-2011
938    ------------------------
940    1.  Updating the JIT compiler.
942    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
943        are added as well.
945    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
946        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
947        calling _pcre_jit_exec. Some extra comments are added.
949    4.  (*MARK) settings inside atomic groups that do not contain any capturing
950        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
951        was introduced by change 18 for 8.20.
953    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
954        ECMA-262 standard.
956    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
957        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
958        This bug was probably introduced by change 9 of 8.13.
960    7.  While fixing 6 above, I noticed that a number of other items were being
961        incorrectly rejected as "not fixed length". This arose partly because newer
962        opcodes had not been added to the fixed-length checking code. I have (a)
963        corrected the bug and added tests for these items, and (b) arranged for an
964        error to occur if an unknown opcode is encountered while checking for fixed
965        length instead of just assuming "not fixed length". The items that were
966        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
967        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
968        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
970    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
971        being incorrectly compiled and would have given unpredicatble results.
973    9.  A possessively repeated subpattern with minimum repeat count greater than
974        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
975        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
976        the first (A) could occur when it should not.
978    10. Add a cast and remove a redundant test from the code.
980    11. JIT should use pcre_malloc/pcre_free for allocation.
982    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
983        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
984        is anything other than /usr, -L is still shown).
986    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
988    14. Perl does not support \N without a following name in a [] class; PCRE now
989        also gives an error.
991    15. If a forward reference was repeated with an upper limit of around 2000,
992        it caused the error "internal error: overran compiling workspace". The
993        maximum number of forward references (including repeats) was limited by the
994        internal workspace, and dependent on the LINK_SIZE. The code has been
995        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
996        the default depends on LINK_SIZE. There is a new upper limit (for safety)
997        of around 200,000 forward references. While doing this, I also speeded up
998        the filling in of repeated forward references.
1000    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1001        incorrectly expecting the subject to contain another "a" after the start.
1003    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1004        in the match, the SKIP should be ignored. This was not happening; instead
1005        the SKIP was being treated as NOMATCH. For patterns such as
1006        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1007        tested.
1009    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1010        now much more compatible with Perl, in particular in cases where the result
1011        is a non-match for a non-anchored pattern. For example, if
1012        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1013        "m", where previously it did not return a name. A side effect of this
1014        change is that for partial matches, the last encountered mark name is
1015        returned, as for non matches. A number of tests that were previously not
1016        Perl-compatible have been moved into the Perl-compatible test files. The
1017        refactoring has had the pleasing side effect of removing one argument from
1018        the match() function, thus reducing its stack requirements.
1020    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1021        subsequent uses of /S (without +) incorrectly behaved like /S+.
1023    21. Retrieve executable code size support for the JIT compiler and fixing
1024        some warnings.
1026    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1027        not work when the shorter character appeared right at the end of the
1028        subject string.
1030    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1031        systems.
1033    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1034        output it when the /M option is used in pcretest.
1036    25. The CheckMan script was not being included in the distribution. Also, added
1037        an explicit "perl" to run Perl scripts from the PrepareRelease script
1038        because this is reportedly needed in Windows.
1040    26. If study data was being save in a file and studying had not found a set of
1041        "starts with" bytes for the pattern, the data written to the file (though
1042        never used) was taken from uninitialized memory and so caused valgrind to
1043        complain.
1045    27. Updated RunTest.bat as provided by Sheri Pierce.
1047    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1049    29. Computation of memory usage for the table of capturing group names was
1050        giving an unnecessarily large value.
1053    Version 8.20 21-Oct-2011
1054    ------------------------
1056    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1057        a POSIX class. After further experiments with Perl, which convinced me that
1058        Perl has bugs and confusions, a closing square bracket is no longer allowed
1059        in a POSIX name. This bug also affected patterns with classes that started
1060        with full stops.
1062    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1063        captured substring, but while checking the failing first alternative,
1064        substring 1 is temporarily captured. If the output vector supplied to
1065        pcre_exec() was not big enough for this capture, the yield of the function
1066        was still zero ("insufficient space for captured substrings"). This cannot
1067        be totally fixed without adding another stack variable, which seems a lot
1068        of expense for a edge case. However, I have improved the situation in cases
1069        such as /(a)(b)x|abc/ matched against "abc", where the return code
1070        indicates that fewer than the maximum number of slots in the ovector have
1071        been set.
1073    3.  Related to (2) above: when there are more back references in a pattern than
1074        slots in the output vector, pcre_exec() uses temporary memory during
1075        matching, and copies in the captures as far as possible afterwards. It was
1076        using the entire output vector, but this conflicts with the specification
1077        that only 2/3 is used for passing back captured substrings. Now it uses
1078        only the first 2/3, for compatibility. This is, of course, another edge
1079        case.
1081    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1082        main code base, and can be used by building with --enable-jit. When this is
1083        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1084        runtime --no-jit option is given.
1086    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1087        ovector, the return from the function was zero, implying that there were
1088        other matches that did not fit. The correct "exactly full" value is now
1089        returned.
1091    6.  If a subpattern that was called recursively or as a subroutine contained
1092        (*PRUNE) or any other control that caused it to give a non-standard return,
1093        invalid errors such as "Error -26 (nested recursion at the same subject
1094        position)" or even infinite loops could occur.
1096    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1097        computing the minimum length on reaching *ACCEPT, and so ended up with the
1098        wrong value of 1 rather than 0. Further investigation indicates that
1099        computing a minimum subject length in the presence of *ACCEPT is difficult
1100        (think back references, subroutine calls), and so I have changed the code
1101        so that no minimum is registered for a pattern that contains *ACCEPT.
1103    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1104        it was not handled as intended. [But see 16 below.]
1106    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1107        Sheri Pierce.
1109    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1110        the first byte in a match must be "a".
1112    11. Change 17 for 8.13 increased the recursion depth for patterns like
1113        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1114        pattern contains any instances of (*THEN). If it does not, the old
1115        optimizations are restored. It would be nice to do this on a per-group
1116        basis, but at the moment that is not feasible.
1118    12. In some environments, the output of pcretest -C is CRLF terminated. This
1119        broke RunTest's code that checks for the link size. A single white space
1120        character after the value is now allowed for.
1122    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1123        For "fr", it uses the Windows-specific input and output files.
1125    14. If (*THEN) appeared in a group that was called recursively or as a
1126        subroutine, it did not work as intended. [But see next item.]
1128    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1129        pattern fragments (but not containing any | characters). If A and B are
1130        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1131        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1132        D. In other words, Perl considers parentheses that do not contain any |
1133        characters to be part of a surrounding alternative, whereas PCRE was
1134        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1135        differently. PCRE now behaves in the same way as Perl, except in the case
1136        of subroutine/recursion calls such as (?1) which have in any case always
1137        been different (but PCRE had them first :-).
1139    16. Related to 15 above: Perl does not treat the | in a conditional group as
1140        creating alternatives. Such a group is treated in the same way as an
1141        ordinary group without any | characters when processing (*THEN). PCRE has
1142        been changed to match Perl's behaviour.
1144    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1145        RunGrepTest script failed.
1147    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1148        inevitable for groups that contain captures, but it can lead to a lot of
1149        stack use in large patterns. The old behaviour has been restored for atomic
1150        groups that do not contain any capturing parentheses.
1152    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1153        suppress the check for a minimum subject length at run time. (If it was
1154        given to pcre_exec() or pcre_dfa_exec() it did work.)
1156    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1157        fail to work when decoding hex characters in data strings in EBCDIC
1158        environments.
1160    21. It appears that in at least one Mac OS environment, the isxdigit() function
1161        is implemented as a macro that evaluates to its argument more than once,
1162        contravening the C 90 Standard (I haven't checked a later standard). There
1163        was an instance in pcretest which caused it to go wrong when processing
1164        \x{...} escapes in subject strings. The has been rewritten to avoid using
1165        things like p++ in the argument of isxdigit().
1168    Version 8.13 16-Aug-2011
1169    ------------------------
1171    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1173    2.  Two minor typos in pcre_internal.h have been fixed.
1175    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1176        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1177        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1179    4.  There were a number of related bugs in the code for matching backrefences
1180        caselessly in UTF-8 mode when codes for the characters concerned were
1181        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1182        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1183        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1184        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1185        2-byte code at the end of the subject (it thought there wasn't enough data
1186        left).
1188    5.  Comprehensive information about what went wrong is now returned by
1189        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1190        as the output vector has at least 2 elements. The offset of the start of
1191        the failing character and a reason code are placed in the vector.
1193    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1194        now returned is for the first byte of the failing character, instead of the
1195        last byte inspected. This is an incompatible change, but I hope it is small
1196        enough not to be a problem. It makes the returned offset consistent with
1197        pcre_exec() and pcre_dfa_exec().
1199    7.  pcretest now gives a text phrase as well as the error number when
1200        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1201        failure, the offset and reason code are output.
1203    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1204        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1205        back over a single character (\n). This seems wrong (because it treated the
1206        two characters as a single entity when going forwards), conflicts with the
1207        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1208        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1209        has been changed.
1211    9.  Some internal refactoring has changed the processing so that the handling
1212        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1213        time (the PCRE_DOTALL option was changed this way some time ago: version
1214        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1215        which was always a bit of a fudge. It also means that there is one less
1216        argument for the match() function, which reduces its stack requirements
1217        slightly. This change also fixes an incompatibility with Perl: the pattern
1218        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1220    10. More internal refactoring has drastically reduced the number of recursive
1221        calls to match() for possessively repeated groups such as (abc)++ when
1222        using pcre_exec().
1224    11. While implementing 10, a number of bugs in the handling of groups were
1225        discovered and fixed:
1227        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1228        (a|)*(?1) gave a compile-time internal error.
1229        ((a|)+)+  did not notice that the outer group could match an empty string.
1230        (^a|^)+   was not marked as anchored.
1231        (.*a|.*)+ was not marked as matching at start or after a newline.
1233    12. Yet more internal refactoring has removed another argument from the match()
1234        function. Special calls to this function are now indicated by setting a
1235        value in a variable in the "match data" data block.
1237    13. Be more explicit in pcre_study() instead of relying on "default" for
1238        opcodes that mean there is no starting character; this means that when new
1239        ones are added and accidentally left out of pcre_study(), testing should
1240        pick them up.
1242    14. The -s option of pcretest has been documented for ages as being an old
1243        synonym of -m (show memory usage). I have changed it to mean "force study
1244        for every regex", that is, assume /S for every regex. This is similar to -i
1245        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1246        using it. It makes it easier to run collections of tests with and without
1247        study enabled, and thereby test pcre_study() more easily. All the standard
1248        tests are now run with and without -s (but some patterns can be marked as
1249        "never study" - see 20 below).
1251    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1252        restoration of the capturing data to the outer values was not happening
1253        correctly.
1255    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1256        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1257        pattern had matched an empty string, and so incorrectly returned a no
1258        match.
1260    17. There was optimizing code for the last branch of non-capturing parentheses,
1261        and also for the obeyed branch of a conditional subexpression, which used
1262        tail recursion to cut down on stack usage. Unfortunately, now that there is
1263        the possibility of (*THEN) occurring in these branches, tail recursion is
1264        no longer possible because the return has to be checked for (*THEN). These
1265        two optimizations have therefore been removed. [But see 8.20/11 above.]
1267    18. If a pattern containing \R was studied, it was assumed that \R always
1268        matched two bytes, thus causing the minimum subject length to be
1269        incorrectly computed because \R can also match just one byte.
1271    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1272        was incorrectly computed.
1274    20. If /S is present twice on a test pattern in pcretest input, it now
1275        *disables* studying, thereby overriding the use of -s on the command line
1276        (see 14 above). This is necessary for one or two tests to keep the output
1277        identical in both cases.
1279    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1280        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1282    22. When an atomic group that contained a capturing parenthesis was
1283        successfully matched, but the branch in which it appeared failed, the
1284        capturing was not being forgotten if a higher numbered group was later
1285        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1286        group 1 to "a", when in fact it should be unset. This applied to multi-
1287        branched capturing and non-capturing groups, repeated or not, and also to
1288        positive assertions (capturing in negative assertions does not happen
1289        in PCRE) and also to nested atomic groups.
1291    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1292        subject after a captured substring, to make it easier to tell which of a
1293        number of identical substrings has been captured.
1295    24. The way atomic groups are processed by pcre_exec() has been changed so that
1296        if they are repeated, backtracking one repetition now resets captured
1297        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1298        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1299        "aaa". Previously, it would have been "a". As part of this code
1300        refactoring, the way recursive calls are handled has also been changed.
1302    25. If an assertion condition captured any substrings, they were not passed
1303        back unless some other capturing happened later. For example, if
1304        (?(?=(a))a) was matched against "a", no capturing was returned.
1306    26. When studying a pattern that contained subroutine calls or assertions,
1307        the code for finding the minimum length of a possible match was handling
1308        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1309        group 1 called group 2 while simultaneously a separate group 2 called group
1310        1). A stack overflow occurred in this case. I have fixed this by limiting
1311        the recursion depth to 10.
1313    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1314        Fortmann. This supports explicit test numbers on the command line, and has
1315        argument validation and error reporting.
1317    28. An instance of \X with an unlimited repeat could fail if at any point the
1318        first character it looked at was a mark character.
1320    29. Some minor code refactoring concerning Unicode properties and scripts
1321        should reduce the stack requirement of match() slightly.
1323    30. Added the '=' option to pcretest to check the setting of unused capturing
1324        slots at the end of the pattern, which are documented as being -1, but are
1325        not included in the return count.
1327    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1328        compiled something random. Now it gives a compile-time error (as does
1329        Perl).
1331    32. A *MARK encountered during the processing of a positive assertion is now
1332        recorded and passed back (compatible with Perl).
1334    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1335        had alternative anchored branches, the search for a second match in a line
1336        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1337        matched the line "0102" twice. The same bug affected patterns that started
1338        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1339        twice.
1341    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1342        does, and because of capturing effects, quantifying parenthesized
1343        assertions may at times be useful. Quantifiers are now allowed for
1344        parenthesized assertions.
1346    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1348    36. \g was being checked for fancy things in a character class, when it should
1349        just be a literal "g".
1351    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1352        appearance of a nested POSIX class supersedes an apparent external class.
1353        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1354        unescaped square brackets may also appear as part of class names. For
1355        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1356        more like Perl. (But see 8.20/1 above.)
1358    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1359        was because it thought it was \N{name}, which is not supported).
1361    39. Add minix to OS list not supporting the -S option in pcretest.
1363    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1364        cannot analyze patterns in sufficient detail to catch mutual recursions
1365        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1366        subgroup is called recursively as a subpattern for a second time at the
1367        same position in the subject string. In previous releases this might have
1368        been caught by the recursion limit, or it might have run out of stack.
1370    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1371        happen only once. PCRE was, however incorrectly giving a compile time error
1372        "recursive call could loop indefinitely" because it cannot analyze the
1373        pattern in sufficient detail. The compile time test no longer happens when
1374        PCRE is compiling a conditional subpattern, but actual runaway loops are
1375        now caught at runtime (see 40 above).
1377    42. It seems that Perl allows any characters other than a closing parenthesis
1378        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1379        has been changed to be the same.
1381    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1382        as not to get warnings when autogen.sh is called. Also changed
1383        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1385    44. To help people who use pcregrep to scan files containing exceedingly long
1386        lines, the following changes have been made:
1388        (a) The default value of the buffer size parameter has been increased from
1389            8K to 20K. (The actual buffer used is three times this size.)
1391        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1392            PCRE is built.
1394        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1395            to be set at run time.
1397        (d) Numerical values in pcregrep options can be followed by K or M, for
1398            example --buffer-size=50K.
1400        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1401            given and the return code is set to 2.
1403    45. Add a pointer to the latest mark to the callout data block.
1405    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1406        partial match of an empty string instead of no match. This was specific to
1407        the use of ".".
1409    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1410        complete match instead of a partial match. This bug was dependent on both
1411        the PCRE_UTF8 and PCRE_DOTALL options being set.
1413    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1414        starting byte set, because \b was not being ignored.
1417    Version 8.12 15-Jan-2011
1418    ------------------------
1420    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1421        checks for such things as part of the documentation building process.
1423    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1424        --match-limit and --recursion-limit options (added for 8.11). In
1425        particular, this made one of the standard tests fail. (The integer value
1426        went into the wrong half of a long int.)
1428    3.  If the --colour option was given to pcregrep with -v (invert match), it
1429        did strange things, either producing crazy output, or crashing. It should,
1430        of course, ignore a request for colour when reporting lines that do not
1431        match.
1433    4.  Another pcregrep bug caused similar problems if --colour was specified with
1434        -M (multiline) and the pattern match finished with a line ending.
1436    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1437        matched in multiline mode, the following line was shown as part of the
1438        match. This seems wrong, so I have changed it.
1440    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1441        the check for further matches in the same line (so they could be coloured)
1442        to overrun the end of the current line. If another match was found, it was
1443        incorrectly shown (and then shown again when found in the next line).
1445    7.  If pcregrep was compiled under Windows, there was a reference to the
1446        function pcregrep_exit() before it was defined. I am assuming this was
1447        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1448        reported by a user. I've moved the definition above the reference.
1451    Version 8.11 10-Dec-2010
1452    ------------------------
1454    1.  (*THEN) was not working properly if there were untried alternatives prior
1455        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1456        backtracked to try for "b" instead of moving to the next alternative branch
1457        at the same level (in this case, to look for "c"). The Perl documentation
1458        is clear that when (*THEN) is backtracked onto, it goes to the "next
1459        alternative in the innermost enclosing group".
1461    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1462        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1463        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1464        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1465        (*THEN).
1467    3.  If \s appeared in a character class, it removed the VT character from
1468        the class, even if it had been included by some previous item, for example
1469        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1470        of \s, but is part of the POSIX "space" class.)
1472    4.  A partial match never returns an empty string (because you can always
1473        match an empty string at the end of the subject); however the checking for
1474        an empty string was starting at the "start of match" point. This has been
1475        changed to the "earliest inspected character" point, because the returned
1476        data for a partial match starts at this character. This means that, for
1477        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1478        (previously it gave "no match").
1480    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1481        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1482        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1483        has an implication that the given string is incomplete (because a partial
1484        match is preferred over a full match). For this reason, these items now
1485        give a partial match in this situation. [Aside: previously, the one case
1486        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1487        match rather than a full match, which was wrong by the old rules, but is
1488        now correct.]
1490    6.  There was a bug in the handling of #-introduced comments, recognized when
1491        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1492        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1493        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1494        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1495        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1496        places in pcre_compile().
1498    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1499        comments when looking ahead for named forward references to subpatterns,
1500        the only newline sequence it recognized was NL. It now handles newlines
1501        according to the set newline convention.
1503    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1504        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1505        cater for a lack of strerror(). These oversights have been fixed.
1507    9.  Added --match-limit and --recursion-limit to pcregrep.
1509    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1511    11. When the -o option was used, pcregrep was setting a return code of 1, even
1512        when matches were found, and --line-buffered was not being honoured.
1514    12. Added an optional parentheses number to the -o and --only-matching options
1515        of pcregrep.
1517    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1518        can match an empty string. The code to do it in pcretest and pcredemo
1519        needed fixing:
1521        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1522            only one byte after an empty string match just before CRLF (this case
1523            just got forgotten; "any" and "anycrlf" were OK).
1525        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1526            mode when an empty string match preceded an ASCII character followed by
1527            a non-ASCII character. (The code for advancing by one character rather
1528            than one byte was nonsense.)
1530        (c) The pcredemo.c sample program did not have any code at all to handle
1531            the cases when CRLF is a valid newline sequence.
1533    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1534        as a starting offset was within the subject string. There is now a new
1535        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1536        negative or greater than the length of the string. In order to test this,
1537        pcretest is extended to allow the setting of negative starting offsets.
1539    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1540        starting offset points to the beginning of a UTF-8 character was
1541        unnecessarily clumsy. I tidied it up.
1543    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1544        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1546    17. Nobody had reported that the --include_dir option, which was added in
1547        release 7.7 should have been called --include-dir (hyphen, not underscore)
1548        for compatibility with GNU grep. I have changed it to --include-dir, but
1549        left --include_dir as an undocumented synonym, and the same for
1550        --exclude-dir, though that is not available in GNU grep, at least as of
1551        release 2.5.4.
1553    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1554        characters from a string of bytes) have been redefined so as not to use
1555        loops, in order to improve performance in some environments. At the same
1556        time, I abstracted some of the common code into auxiliary macros to save
1557        repetition (this should not affect the compiled code).
1559    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1560        compile-time error is now given if \c is not followed by an ASCII
1561        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1562        different, and any byte value is allowed.)
1564    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1565        START_OPTIMIZE option, which is now allowed at compile time - but just
1566        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1567        to pcregrep and other applications that have no direct access to PCRE
1568        options. The new /Y option in pcretest sets this option when calling
1569        pcre_compile().
1571    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1572        back references. Groups containing recursive back references were forced to
1573        be atomic by that change, but in the case of named groups, the amount of
1574        memory required was incorrectly computed, leading to "Failed: internal
1575        error: code overflow". This has been fixed.
1577    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1578        pcretest.c, to avoid build problems in some Borland environments.
1581    Version 8.10 25-Jun-2010
1582    ------------------------
1584    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1585        THEN.
1587    2.  (*ACCEPT) was not working when inside an atomic group.
1589    3.  Inside a character class, \B is treated as a literal by default, but
1590        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1591        causes the error). The code is unchanged, but I tidied the documentation.
1593    4.  Inside a character class, PCRE always treated \R and \X as literals,
1594        whereas Perl faults them if its -w option is set. I have changed PCRE so
1595        that it faults them when PCRE_EXTRA is set.
1597    5.  Added support for \N, which always matches any character other than
1598        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1600    6.  When compiling pcregrep with newer versions of gcc which may have
1601        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1602        declared with attribute warn_unused_result" were given. Just casting the
1603        result to (void) does not stop the warnings; a more elaborate fudge is
1604        needed. I've used a macro to implement this.
1606    7.  Minor change to pcretest.c to avoid a compiler warning.
1608    8.  Added four artifical Unicode properties to help with an option to make
1609        \s etc use properties (see next item). The new properties are: Xan
1610        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1612    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1613        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1614        this option. Modified pcretest to add /W to test this facility. Added
1615        REG_UCP to make it available via the POSIX interface.
1617    10. Added --line-buffered to pcregrep.
1619    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1620        studied, and the match started with a letter with a code point greater than
1621        127 whose first byte was different to the first byte of the other case of
1622        the letter, the other case of this starting letter was not recognized
1623        (#976).
1625    12. If a pattern that was studied started with a repeated Unicode property
1626        test, for example, \p{Nd}+, there was the theoretical possibility of
1627        setting up an incorrect bitmap of starting bytes, but fortunately it could
1628        not have actually happened in practice until change 8 above was made (it
1629        added property types that matched character-matching opcodes).
1631    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1632        possible starting bytes for non-anchored patterns.
1634    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1635        \R, and also a number of cases that involve Unicode properties, both
1636        explicit and implicit when PCRE_UCP is set.
1638    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1639        input, it could crash or give wrong results if characters with values
1640        greater than 0xc0 were present in the subject string. (Detail: it assumed
1641        UTF-8 input when processing these items.)
1643    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1644        size_t is 64-bit (#991).
1646    17. Added a check for running out of memory when PCRE is compiled with
1647        --disable-stack-for-recursion (#990).
1649    18. If the last data line in a file for pcretest does not have a newline on
1650        the end, a newline was missing in the output.
1652    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1653        less than 128) in its various bitmaps. However, there is a facility for
1654        generating tables according to the current locale when PCRE is compiled. It
1655        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1656        characters, are recognized by isspace() and therefore were getting set in
1657        these tables, and indeed these tables seem to approximate to ISO 8859. This
1658        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1659        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1660        which of course cannot start UTF-8 characters. I have changed the code so
1661        that only real ASCII characters (less than 128) and the correct starting
1662        bytes for UTF-8 encodings are set for characters greater than 127 when in
1663        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1664        altogether.)
1666    20. Added the /T option to pcretest so as to be able to run tests with non-
1667        standard character tables, thus making it possible to include the tests
1668        used for 19 above in the standard set of tests.
1670    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1671        reference to a subpattern the other side of a comment that contains an
1672        opening parenthesis caused either an internal compiling error, or a
1673        reference to the wrong subpattern.
1676    Version 8.02 19-Mar-2010
1677    ------------------------
1679    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1681    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1682        configured.
1684    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1685        original author of that file, following a query about its status.
1687    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1688        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1690    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1691        quantifier applied to a forward-referencing subroutine call, could compile
1692        incorrect code or give the error "internal error: previously-checked
1693        referenced subpattern not found".
1695    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1696        variables to point to external functions. For these systems, therefore,
1697        pcre_malloc etc. are now initialized to local functions that call the
1698        relevant global functions.
1700    7.  There were two entries missing in the vectors called coptable and poptable
1701        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1702        I've fixed the data, and added a kludgy way of testing at compile time that
1703        the lengths are correct (equal to the number of opcodes).
1705    8.  Following on from 7, I added a similar kludge to check the length of the
1706        eint vector in pcreposix.c.
1708    9.  Error texts for pcre_compile() are held as one long string to avoid too
1709        much relocation at load time. To find a text, the string is searched,
1710        counting zeros. There was no check for running off the end of the string,
1711        which could happen if a new error number was added without updating the
1712        string.
1714    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1716    11. \K was not working if it appeared in an atomic group or in a group that
1717        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1718        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1719        the assertion is positive, but not if it is negative.
1721    12. Change 11 fortuitously reduced the size of the stack frame used in the
1722        "match()" function of pcre_exec.c by one pointer. Forthcoming
1723        implementation of support for (*MARK) will need an extra pointer on the
1724        stack; I have reserved it now, so that the stack frame size does not
1725        decrease.
1727    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1728        item in branch that calls a recursion is a subroutine call - as in the
1729        second branch in the above example - was incorrectly given the compile-
1730        time error "recursive call could loop indefinitely" because pcre_compile()
1731        was not correctly checking the subroutine for matching a non-empty string.
1733    14. The checks for overrunning compiling workspace could trigger after an
1734        overrun had occurred. This is a "should never occur" error, but it can be
1735        triggered by pathological patterns such as hundreds of nested parentheses.
1736        The checks now trigger 100 bytes before the end of the workspace.
1738    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1741    Version 8.01 19-Jan-2010
1742    ------------------------
1744    1.  If a pattern contained a conditional subpattern with only one branch (in
1745        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1746        computed the wrong minimum data length (which is of course zero for such
1747        subpatterns). This could cause incorrect "no match" results.
1749    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1750        the pattern is reset in the first branch, pcre_compile() failed with
1751        "internal error: code overflow at offset...". This happened only when
1752        the reset was to the original external option setting. (An optimization
1753        abstracts leading options settings into an external setting, which was the
1754        cause of this.)
1756    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1757        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1758        assertion pattern did not match (meaning that the assertion was true), it
1759        was incorrectly treated as false if the SKIP had been reached during the
1760        matching. This also applied to assertions used as conditions.
1762    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1763        assertion subpattern, including such a pattern used as a condition,
1764        unpredictable results occurred, instead of the error return
1767    5.  The C++ GlobalReplace function was not working like Perl for the special
1768        situation when an empty string is matched. It now does the fancy magic
1769        stuff that is necessary.
1771    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1772        removed. (These were left over from very, very early versions of PCRE.)
1774    7.  Some cosmetic changes to the code to make life easier when compiling it
1775        as part of something else:
1777        (a) Change DEBUG to PCRE_DEBUG.
1779        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1780            called "current" as "current_branch", to prevent a collision with the
1781            Linux macro when compiled as a kernel module.
1783        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1784            prevent a collision with the Linux macro when compiled as a kernel
1785            module.
1787    8.  In pcre_compile() there are some checks for integer overflows that used to
1788        cast potentially large values to (double). This has been changed to that
1789        when building, a check for int64_t is made, and if it is found, it is used
1790        instead, thus avoiding the use of floating point arithmetic. (There is no
1791        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1792        double.
1794    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1795        2005 (difference between two addresses compared to an unsigned value).
1797    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1798        custom one, because of the following reported problem in Windows:
1800          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1801              under Win32.
1802          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1803              therefore missing the function definition.
1804          - The compiler thus generates a "C" signature for the test function.
1805          - The linker fails to find the "C" function.
1806          - PCRE fails to configure if asked to do so against libbz2.
1808    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1809        messages were output:
1811          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1812          rerunning libtoolize, to keep the correct libtool macros in-tree.
1813          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1815        I have done both of these things.
1817    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1818        most of the time, it *can* run out if it is given a pattern that contains a
1819        runaway infinite recursion. I updated the discussion in the pcrestack man
1820        page.
1822    13. Now that we have gone to the x.xx style of version numbers, the minor
1823        version may start with zero. Using 08 or 09 is a bad idea because users
1824        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1825        interpreted as invalid octal numbers. I've updated the previous comment in
1826        configure.ac, and also added a check that gives an error if 08 or 09 are
1827        used.
1829    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1830        causing partial matching to fail when the end of the subject matched \W
1831        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1833    15. There were some discrepancies between the declarations in pcre_internal.h
1834        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1835        their definitions. The declarations used "const uschar *" and the
1836        definitions used USPTR. Even though USPTR is normally defined as "const
1837        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1838        reported that: "This difference in casting confuses some C++ compilers, for
1839        example, SunCC recognizes above declarations as different functions and
1840        generates broken code for hbpcre." I have changed the declarations to use
1841        USPTR.
1843    16. GNU libtool is named differently on some systems. The autogen.sh script now
1844        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1845        (FreeBSD).
1847    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1848        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1849        comment: "Figure out how to create a longlong from a string: strtoll and
1850        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1851        instance, but it only takes 2 args instead of 3!"
1853    18. A subtle bug concerned with back references has been fixed by a change of
1854        specification, with a corresponding code fix. A pattern such as
1855        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1856        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1857        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1858        same bug. Such groups have to be quantified to be useful, or contained
1859        inside another quantified group. (If there's no repetition, the reference
1860        can never match.) The problem arises because, having left the group and
1861        moved on to the rest of the pattern, a later failure that backtracks into
1862        the group uses the captured value from the final iteration of the group
1863        rather than the correct earlier one. I have fixed this in PCRE by forcing
1864        any group that contains a reference to itself to be an atomic group; that
1865        is, there cannot be any backtracking into it once it has completed. This is
1866        similar to recursive and subroutine calls.
1869    Version 8.00 19-Oct-09
1870    ----------------------
1872    1.  The table for translating pcre_compile() error codes into POSIX error codes
1873        was out-of-date, and there was no check on the pcre_compile() error code
1874        being within the table. This could lead to an OK return being given in
1875        error.
1877    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1878        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1879        in a Windows environment.
1881    3.  The pcregrep --count option prints the count for each file even when it is
1882        zero, as does GNU grep. However, pcregrep was also printing all files when
1883        --files-with-matches was added. Now, when both options are given, it prints
1884        counts only for those files that have at least one match. (GNU grep just
1885        prints the file name in this circumstance, but including the count seems
1886        more useful - otherwise, why use --count?) Also ensured that the
1887        combination -clh just lists non-zero counts, with no names.
1889    4.  The long form of the pcregrep -F option was incorrectly implemented as
1890        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1891        but it seems right to fix it, and I didn't think it was worth preserving
1892        the old behaviour.
1894    5.  The command line items --regex=pattern and --regexp=pattern were not
1895        recognized by pcregrep, which required --regex pattern or --regexp pattern
1896        (with a space rather than an '='). The man page documented the '=' forms,
1897        which are compatible with GNU grep; these now work.
1899    6.  No libpcreposix.pc file was created for pkg-config; there was just
1900        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1902    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1903        when UCP support is not needed, by modifying the Python script that
1904        generates it from Unicode data files. This should not matter if the module
1905        is correctly used as a library, but I received one complaint about 50K of
1906        unwanted data. My guess is that the person linked everything into his
1907        program rather than using a library. Anyway, it does no harm.
1909    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1910        was a minimum greater than 1 for a wide character in a possessive
1911        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1912        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1913        character. Chaos in the form of incorrect output or a compiling loop could
1914        result.
1916    9.  The restrictions on what a pattern can contain when partial matching is
1917        requested for pcre_exec() have been removed. All patterns can now be
1918        partially matched by this function. In addition, if there are at least two
1919        slots in the offset vector, the offset of the earliest inspected character
1920        for the match and the offset of the end of the subject are set in them when
1921        PCRE_ERROR_PARTIAL is returned.
1923    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1924        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1925        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1926        and may be more useful for multi-segment matching.
1928    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1929        used to be given if ever the end of the subject was reached; now it is
1930        given only if matching could not proceed because another character was
1931        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1932        string "Z", which now yields "no match" instead of "partial match". In the
1933        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1934        final character ended with (*FAIL).
1936    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1937        if the pattern had a "must contain" character that was already found in the
1938        earlier partial match, unless partial matching was again requested. For
1939        example, with the pattern /dog.(body)?/, the "must contain" character is
1940        "g". If the first part-match was for the string "dog", restarting with
1941        "sbody" failed. This bug has been fixed.
1943    13. The string returned by pcre_dfa_exec() after a partial match has been
1944        changed so that it starts at the first inspected character rather than the
1945        first character of the match. This makes a difference only if the pattern
1946        starts with a lookbehind assertion or \b or \B (\K is not supported by
1947        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1948        matching functions compatible, and I think it's the right thing to do.
1950    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1951        so that the demonstration program is easily available in environments where
1952        PCRE has not been installed from source.
1954    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1955        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1956        library.
1958    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1959        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1960        is not the first non-POSIX option to be added. Clearly some people find
1961        these options useful.
1963    17. If a caller to the POSIX matching function regexec() passes a non-zero
1964        value for nmatch with a NULL value for pmatch, the value of
1965        nmatch is forced to zero.
1967    18. RunGrepTest did not have a test for the availability of the -u option of
1968        the diff command, as RunTest does. It now checks in the same way as
1969        RunTest, and also checks for the -b option.
1971    19. If an odd number of negated classes containing just a single character
1972        interposed, within parentheses, between a forward reference to a named
1973        subpattern and the definition of the subpattern, compilation crashed with
1974        an internal error, complaining that it could not find the referenced
1975        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1976        [The bug was that it was starting one character too far in when skipping
1977        over the character class, thus treating the ] as data rather than
1978        terminating the class. This meant it could skip too much.]
1980    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1981        /g option in pcretest when the pattern contains \K, which makes it possible
1982        to have an empty string match not at the start, even when the pattern is
1983        anchored. Updated pcretest and pcredemo to use this option.
1985    21. If the maximum number of capturing subpatterns in a recursion was greater
1986        than the maximum at the outer level, the higher number was returned, but
1987        with unset values at the outer level. The correct (outer level) value is
1988        now given.
1990    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1991        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1992        make it do so. The string so far is captured, making this feature
1993        compatible with Perl.
1995    23. The tests have been re-organized, adding tests 11 and 12, to make it
1996        possible to check the Perl 5.10 features against Perl 5.10.
1998    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1999        pattern matches a fixed length string. PCRE did not allow this; now it
2000        does. Neither allows recursion.
2002    25. I finally figured out how to implement a request to provide the minimum
2003        length of subject string that was needed in order to match a given pattern.
2004        (It was back references and recursion that I had previously got hung up
2005        on.) This code has now been added to pcre_study(); it finds a lower bound
2006        to the length of subject needed. It is not necessarily the greatest lower
2007        bound, but using it to avoid searching strings that are too short does give
2008        some useful speed-ups. The value is available to calling programs via
2009        pcre_fullinfo().
2011    26. While implementing 25, I discovered to my embarrassment that pcretest had
2012        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2013        study optimizations had never been tested with that matching function.
2014        Oops. What is worse, even when it was passed study data, there was a bug in
2015        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2016        were also very few tests of studied patterns with pcre_dfa_exec().
2018    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2019        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2020        on the other side of the coin, they are no longer allowed to have different
2021        names, because these cannot be distinguished in PCRE, and this has caused
2022        confusion. (This is a difference from Perl.)
2024    28. When duplicate subpattern names are present (necessarily with different
2025        numbers, as required by 27 above), and a test is made by name in a
2026        conditional pattern, either for a subpattern having been matched, or for
2027        recursion in such a pattern, all the associated numbered subpatterns are
2028        tested, and the overall condition is true if the condition is true for any
2029        one of them. This is the way Perl works, and is also more like the way
2030        testing by number works.
2033    Version 7.9 11-Apr-09
2034    ---------------------
2036    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2037        (pcretest), all targets were linked against these libraries. This included
2038        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2039        libraries. This caused unwanted dependencies to be created. This problem
2040        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2041        pcretest is linked with readline.
2043    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2044        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2045        moved outside it again, because FALSE and TRUE are already defined in AIX,
2046        but BOOL is not.
2048    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2049        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2051    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2052        hyphens) following file names and line numbers when outputting matching
2053        lines. This is not true; no spaces are inserted. I have also clarified the
2054        wording for the --colour (or --color) option.
2056    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2057        was not coloured; this is different to GNU grep, so I have changed it to be
2058        the same.
2060    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2061        each matching line was coloured. Now it goes on to look for further matches
2062        of any of the test patterns, which is the same behaviour as GNU grep.
2064    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2065        doesn't make sense to accept an empty string match in pcregrep, so I have
2066        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2067        seems to be how GNU grep behaves. [But see later change 40 for release
2068        8.33.]
2070    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2071        start or after a newline", because the conditional assertion was not being
2072        correctly handled. The rule now is that both the assertion and what follows
2073        in the first alternative must satisfy the test.
2075    9.  If auto-callout was enabled in a pattern with a conditional group whose
2076        condition was an assertion, PCRE could crash during matching, both with
2077        pcre_exec() and pcre_dfa_exec().
2079    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2080        used for matching.
2082    11. Unicode property support in character classes was not working for
2083        characters (bytes) greater than 127 when not in UTF-8 mode.
2085    12. Added the -M command line option to pcretest.
2087    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2089    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2091    16. Added comments and documentation about mis-use of no_arg in the C++
2092        wrapper.
2094    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2095        from Martin Jerabek that uses macro names for all relevant character and
2096        string constants.
2098    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2099        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2100        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2101        these, but not everybody uses configure.
2103    19. A conditional group that had only one branch was not being correctly
2104        recognized as an item that could match an empty string. This meant that an
2105        enclosing group might also not be so recognized, causing infinite looping
2106        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2107        with the subject "ab", where knowledge that the repeated group can match
2108        nothing is needed in order to break the loop.
2110    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2111        exec(), but without supplying a callout function, matching went wrong.
2113    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2114        leak if the size of the offset vector was greater than 30. When the vector
2115        is smaller, the saved offsets during recursion go onto a local stack
2116        vector, but for larger vectors malloc() is used. It was failing to free
2117        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2118        error, in fact).
2120    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2121        heapframe that is used only when UTF-8 support is enabled. This caused no
2122        problem, but was untidy.
2124    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2125        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2126        included within another project.
2128    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2129        slightly modified by me:
2131          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2132              not building pcregrep.
2134          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2135              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2137    25. Forward references, both numeric and by name, in patterns that made use of
2138        duplicate group numbers, could behave incorrectly or give incorrect errors,
2139        because when scanning forward to find the reference group, PCRE was not
2140        taking into account the duplicate group numbers. A pattern such as
2141        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2143    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2144        the feature of a custom pointer more persuasive (as requested by a user).
2146    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2147        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2148        pre-defined.
2150    28. Added support for (*UTF8) at the start of a pattern.
2152    29. Arrange for flags added by the "release type" setting in CMake to be shown
2153        in the configuration summary.
2156    Version 7.8 05-Sep-08
2157    ---------------------
2159    1.  Replaced UCP searching code with optimized version as implemented for Ad
2160        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2161        stage table and inline lookup instead of a function, giving speed ups of 2
2162        to 5 times on some simple patterns that I tested. Permission was given to
2163        distribute the MultiStage2.py script that generates the tables (it's not in
2164        the tarball, but is in the Subversion repository).
2166    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2167        scripts.
2169    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2170        a group with a zero qualifier. The result of the study could be incorrect,
2171        or the function might crash, depending on the pattern.
2173    4.  Caseless matching was not working for non-ASCII characters in back
2174        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2175        It now works when Unicode Property Support is available.
2177    5.  In pcretest, an escape such as \x{de} in the data was always generating
2178        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2179        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2180        truncation.
2182    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2184    7.  Added two (int) casts to pcregrep when printing the difference of two
2185        pointers, in case they are 64-bit values.
2187    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2188        test 2 if it fails.
2190    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2191        and a #define of that name to empty if it is not externally set. This is to
2192        allow users of MSVC to set it if necessary.
2194    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2195        the convenience functions in the pcre_get.c source file.
2197    11. An option change at the start of a pattern that had top-level alternatives
2198        could cause overwriting and/or a crash. This command provoked a crash in
2199        some environments:
2201          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2203        This potential security problem was recorded as CVE-2008-2371.
2205    12. For a pattern where the match had to start at the beginning or immediately
2206        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2207        pcre_dfa_exec() could read past the end of the passed subject if there was
2208        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2209        pcretest so that it places the subject at the end of its malloc-ed buffer.
2211    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2212        exec() might read past the end of the data buffer in UTF-8 mode.
2214    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2215        the data contained the byte 0x85 as part of a UTF-8 character within its
2216        first line. This applied both to normal and DFA matching.
2218    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2219        /^[^d]*?$/8 failed to match "abc".
2221    16. Added a missing copyright notice to pcrecpp_internal.h.
2223    17. Make it more clear in the documentation that values returned from
2224        pcre_exec() in ovector are byte offsets, not character counts.
2226    18. Tidied a few places to stop certain compilers from issuing warnings.
2228    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2229        supplied by Stefan Weber. I made a further small update for 7.8 because
2230        there is a change of source arrangements: the pcre_searchfuncs.c module is
2231        replaced by pcre_ucd.c.
2234    Version 7.7 07-May-08
2235    ---------------------
2237    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2238        a string to a long long, pretend we don't even have a long long." This is
2239        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2241    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2242        pre-7.6 versions, which defined a global no_arg variable instead of putting
2243        it in the RE class. (See also #8 below.)
2245    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2246        Lopes.
2248    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2250        (1) The include/exclude patterns were being applied to the whole pathnames
2251            of files, instead of just to the final components.
2253        (2) If there was more than one level of directory, the subdirectories were
2254            skipped unless they satisfied the include/exclude conditions. This is
2255            inconsistent with GNU grep (and could even be seen as contrary to the
2256            pcregrep specification - which I improved to make it absolutely clear).
2257            The action now is always to scan all levels of directory, and just
2258            apply the include/exclude patterns to regular files.
2260    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2261        --exclude_dir in the tests to avoid scanning .svn directories.
2263    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2264        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2265        doesn't support NULs in patterns.
2267    7.  Added some missing "const"s to declarations of static tables in
2268        pcre_compile.c and pcre_dfa_exec.c.
2270    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2271        caused by fix #2  above. (Subsequently also a second patch to fix the
2272        first patch. And a third patch - this was a messy problem.)
2274    9.  Applied Craig's patch to remove the use of push_back().
2276    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2277        matching function regexec().
2279    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2280        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2281        references. PCRE supports relative numbers with this syntax (I don't think
2282        Oniguruma does).
2284    12. Previously, a group with a zero repeat such as (...){0} was completely
2285        omitted from the compiled regex. However, this means that if the group
2286        was called as a subroutine from elsewhere in the pattern, things went wrong
2287        (an internal error was given). Such groups are now left in the compiled
2288        pattern, with a new opcode that causes them to be skipped at execution
2289        time.
2291    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2292        to the way PCRE behaves:
2294        (a) A lone ] character is dis-allowed (Perl treats it as data).
2296        (b) A back reference to an unmatched subpattern matches an empty string
2297            (Perl fails the current match path).
2299        (c) A data ] in a character class must be notated as \] because if the
2300            first data character in a class is ], it defines an empty class. (In
2301            Perl it is not possible to have an empty class.) The empty class []
2302            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2303            The negative empty class [^] matches any one character, independently
2304            of the DOTALL setting.
2306    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2307        non-existent subpattern following a character class starting with ']' and
2308        containing () gave an internal compiling error instead of "reference to
2309        non-existent subpattern". Fortunately, when the pattern did exist, the
2310        compiled code was correct. (When scanning forwards to check for the
2311        existence of the subpattern, it was treating the data ']' as terminating
2312        the class, so got the count wrong. When actually compiling, the reference
2313        was subsequently set up correctly.)
2315    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2316        it was being rejected as not supported by pcre_dfa_exec(), even though
2317        other assertions are supported. I have made pcre_dfa_exec() support
2318        (*FAIL).
2320    16. The implementation of 13c above involved the invention of a new opcode,
2321        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2322        cannot be changed at match time, I realized I could make a small
2323        improvement to matching performance by compiling OP_ALLANY instead of
2324        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2325        on the OP_ANY path.
2327    17. Compiling pcretest on Windows with readline support failed without the
2328        following two fixes: (1) Make the unistd.h include conditional on
2329        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2331    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2332        ncurses library to be included for pcretest when ReadLine support is
2333        requested, but also to allow for it to be overridden. This patch came from
2334        Daniel Bergström.
2336    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2337        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2338        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2339        spotting this.
2342    Version 7.6 28-Jan-08
2343    ---------------------
2345    1.  A character class containing a very large number of characters with
2346        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2347        overflow.
2349    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2350        HAVE_LONG_LONG is not defined.
2352    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2353        bring them up to date and include new features. This patch includes:
2355        - Fixed PH's badly added libz and libbz2 support.
2356        - Fixed a problem with static linking.
2357        - Added pcredemo. [But later removed - see 7 below.]
2358        - Fixed dftables problem and added an option.
2359        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2360            HAVE_LONG_LONG.
2361        - Added readline support for pcretest.
2362        - Added an listing of the option settings after cmake has run.
2364    4.  A user submitted a patch to Makefile that makes it easy to create
2365        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2366        Makefile.am that cause it to include this special target, without
2367        affecting anything else. Note that the same mingw target plus all
2368        the other distribution libraries and programs are now supported
2369        when configuring with CMake (see 6 below) instead of with
2370        Configure/Make.
2372    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2373        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2374        exported in the Windows port". It has not yet been confirmed that the patch
2375        solves the problem, but it does no harm.
2377    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2378        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2379        with CMake, and also correct the comment about stack recursion.
2381    7.  Remove the automatic building of pcredemo from the ./configure system and
2382        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2383        of a program that users should build themselves after PCRE is installed, so
2384        building it automatically is not really right. What is more, it gave
2385        trouble in some build environments.
2387    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2390    Version 7.5 10-Jan-08
2391    ---------------------
2393    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2394        values in parens when parsing an RE using the C++ wrapper."
2396    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2397        Characters greater than 255 were excluded from the class instead of being
2398        included.
2400    3.  The same bug as (2) above applied to negated POSIX classes such as
2401        [:^space:].
2403    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2404        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2405        I have changed it.
2407    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2408        first named subpattern) and a construct such as (?&a) would reference the
2409        first named subpattern whose name started with "a" (in other words, the
2410        length check was missing). Both these problems are fixed. "Subpattern name
2411        expected" is now given for (?&) (a zero-length name), and this patch also
2412        makes it give the same error for \k'' (previously it complained that that
2413        was a reference to a non-existent subpattern).
2415    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2416        this is right because (?- can be followed by option settings as well as by
2417        digits. I have, however, made the messages clearer.
2419    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2420        than the number used in the conditional) now cause a compile-time error.
2421        This is actually not compatible with Perl, which accepts such patterns, but
2422        treats the conditional as always being FALSE (as PCRE used to), but it
2423        seems to me that giving a diagnostic is better.
2425    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2426        and messages.
2428    9.  Fix two occurrences of "backslash" in comments that should have been
2429        "backspace".
2431    10. Remove two redundant lines of code that can never be obeyed (their function
2432        was moved elsewhere).
2434    11. The program that makes PCRE's Unicode character property table had a bug
2435        which caused it to generate incorrect table entries for sequences of
2436        characters that have the same character type, but are in different scripts.
2437        It amalgamated them into a single range, with the script of the first of
2438        them. In other words, some characters were in the wrong script. There were
2439        thirteen such cases, affecting characters in the following ranges:
2441          U+002b0 - U+002c1
2442          U+0060c - U+0060d
2443          U+0061e - U+00612
2444          U+0064b - U+0065e
2445          U+0074d - U+0076d
2446          U+01800 - U+01805
2447          U+01d00 - U+01d77
2448          U+01d9b - U+01dbf
2449          U+0200b - U+0200f
2450          U+030fc - U+030fe
2451          U+03260 - U+0327f
2452          U+0fb46 - U+0fbb1
2453          U+10450 - U+1049d
2455    12. The -o option (show only the matching part of a line) for pcregrep was not
2456        compatible with GNU grep in that, if there was more than one match in a
2457        line, it showed only the first of them. It now behaves in the same way as
2458        GNU grep.
2460    13. If the -o and -v options were combined for pcregrep, it printed a blank
2461        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2462        does the same. The return code can be used to tell if there were any
2463        non-matching lines.
2465    14. Added --file-offsets and --line-offsets to pcregrep.
2467    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2468        infinitely looping recursion. The bug was that positive lookaheads were not
2469        being skipped when checking for a possible empty match (negative lookaheads
2470        and both kinds of lookbehind were skipped).
2472    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2473        inclusion of <windows.h> to before rather than after the definition of
2474        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2476    17. Specifying a possessive quantifier with a specific limit for a Unicode
2477        character property caused pcre_compile() to compile bad code, which led at
2478        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2479        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2480        caused the error; without that there was no problem.
2482    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2484    19. Added --enable-pcretest-libreadline.
2486    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2487        RE::GlobalReplace(). As a result, the number of replacements returned was
2488        double what it should be. I removed one of the increments, but Craig sent a
2489        later patch that removed the other one (the right fix) and added unit tests
2490        that check the return values (which was not done before).
2492    21. Several CMake things:
2494        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2495            the names libpcre and libpcreposix, not just pcre and pcreposix.
2497        (2) The above change means that pcretest and pcregrep are now correctly
2498            linked with the newly-built libraries, not previously installed ones.
2502    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2503        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2504        UTF-8 newline character). The key issue is that the pattern starts .*;
2505        this means that the match must be either at the beginning, or after a
2506        newline. The bug was in the code for advancing after a failed match and
2507        checking that the new position followed a newline. It was not taking
2508        account of UTF-8 characters correctly.
2510    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2511        character classes. PCRE was not treating the sequence [:...:] as a
2512        character class unless the ... were all letters. Perl, however, seems to
2513        allow any characters between [: and :], though of course it rejects as
2514        unknown any "names" that contain non-letters, because all the known class
2515        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2516        for example, whereas PCRE did not - it did not recognize a POSIX character
2517        class. This seemed a bit dangerous, so the code has been changed to be
2518        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2519        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2520        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2521        Perl does, and where it didn't before.
2523    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2524        Windows environments %n is disabled by default.
2527    Version 7.4 21-Sep-07
2528    ---------------------
2530    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2531        means that a class such as [\s] counted as "explicit reference to CR or
2532        LF". That isn't really right - the whole point of the change was to try to
2533        help when there was an actual mention of one of the two characters. So now
2534        the change happens only if \r or \n (or a literal CR or LF) character is
2535        encountered.
2537    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2538        of both had grown to the point where there were only 3 bits left.
2539        Fortunately, there was spare space in the data structure, and so I have
2540        moved the internal flags into a new 16-bit field to free up more option
2541        bits.
2543    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2544        but did not set the internal JCHANGED flag - either of these is enough to
2545        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2546        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2547        start sets both bits.
2549    4.  Added options (at build time, compile time, exec time) to change \R from
2550        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2552    5.  doc/pcresyntax.html was missing from the distribution.
2554    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2555        compatibility, even though it is no longer used.
2557    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2558        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2559        windows.h file is present (where different names are used). [This was
2560        reversed later after testing - see 16 below.]
2562    8.  Changed all #include <config.h> to #include "config.h". There were also
2563        some further <pcre.h> cases that I changed to "pcre.h".
2565    9.  When pcregrep was used with the --colour option, it missed the line ending
2566        sequence off the lines that it output.
2568    10. It was pointed out to me that arrays of string pointers cause lots of
2569        relocations when a shared library is dynamically loaded. A technique of
2570        using a single long string with a table of offsets can drastically reduce
2571        these. I have refactored PCRE in four places to do this. The result is
2572        dramatic:
2574          Originally:                          290
2575          After changing UCP table:            187
2576          After changing error message table:   43
2577          After changing table of "verbs"       36
2578          After changing table of Posix names   22
2580        Thanks to the folks working on Gregex for glib for this insight.
2582    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2583        unicode-properties was also set.
2585    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2587    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2588        checked only for CRLF.
2590    14. Added casts to pcretest.c to avoid compiler warnings.
2592    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2594    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2595        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2596        entirely. This removes changes made in 7 above.
2598    17. The CMake files have been updated, and there is now more information about
2599        building with CMake in the NON-UNIX-USE document.
2602    Version 7.3 28-Aug-07
2603    ---------------------
2605     1. In the rejigging of the build system that eventually resulted in 7.1, the
2606        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2607        brackets there is not right, since it causes compilers to look for an
2608        installed pcre.h, not the version that is in the source that is being
2609        compiled (which of course may be different). I have changed it back to:
2611          #include "pcre.h"
2613        I have a vague recollection that the change was concerned with compiling in
2614        different directories, but in the new build system, that is taken care of
2615        by the VPATH setting the Makefile.
2617     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2618        when the subject happened to end in the byte 0x85 (e.g. if the last
2619        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2620        characters but of course it shouldn't be taken as a newline when it is part
2621        of another character. The bug was that, for an unlimited repeat of . in
2622        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2623        characters when looking for a newline.
2625     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2627     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2628        in debug output.
2630     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2631        long printing in the pcrecpp unittest when running under MinGW.
2633     6. ESC_K was left out of the EBCDIC table.
2635     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2636        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2637        limit also applies to "virtual nesting" when a pattern is recursive, and in
2638        this case 1000 isn't so big. I have been able to remove this limit at the
2639        expense of backing off one optimization in certain circumstances. Normally,
2640        when pcre_exec() would call its internal match() function recursively and
2641        immediately return the result unconditionally, it uses a "tail recursion"
2642        feature to save stack. However, when a subpattern that can match an empty
2643        string has an unlimited repetition quantifier, it no longer makes this
2644        optimization. That gives it a stack frame in which to save the data for
2645        checking that an empty string has been matched. Previously this was taken
2646        from the 1000-entry workspace that had been reserved. So now there is no
2647        explicit limit, but more stack is used.
2649     8. Applied Daniel's patches to solve problems with the import/export magic
2650        syntax that is required for Windows, and which was going wrong for the
2651        pcreposix and pcrecpp parts of the library. These were overlooked when this
2652        problem was solved for the main library.
2654     9. There were some crude static tests to avoid integer overflow when computing
2655        the size of patterns that contain repeated groups with explicit upper
2656        limits. As the maximum quantifier is 65535, the maximum group length was
2657        set at 30,000 so that the product of these two numbers did not overflow a
2658        32-bit integer. However, it turns out that people want to use groups that
2659        are longer than 30,000 bytes (though not repeat them that many times).
2660        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2661        made it possible to implement the integer overflow checks in a much more
2662        dynamic way, which I have now done. The artificial limitation on group
2663        length has been removed - we now have only the limit on the total length of
2664        the compiled pattern, which depends on the LINK_SIZE setting.
2666    10. Fixed a bug in the documentation for get/copy named substring when
2667        duplicate names are permitted. If none of the named substrings are set, the
2668        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2669        empty string.
2671    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2672        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2673        because the ] is interpreted as the first data character and the
2674        terminating ] is not found. PCRE has been made compatible with Perl in this
2675        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2676        cause memory overwriting.
2678    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2679        string has been matched (to stop an infinite loop). It was not recognizing
2680        a conditional subpattern that could match an empty string if that
2681        subpattern was within another subpattern. For example, it looped when
2682        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2683        condition was not nested. This bug has been fixed.
2685    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2686        past the start of the subject in the presence of bytes with the top bit
2687        set, for example "\x8aBCD".
2689    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2690        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2692    14. Optimized (?!) to (*FAIL).
2694    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2695        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2696        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2697        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2698        does: it's just the validity check that is more restrictive.
2700    16. Inserted checks for integer overflows during escape sequence (backslash)
2701        processing, and also fixed erroneous offset values for syntax errors during
2702        backslash processing.
2704    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2705        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2707    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2708        caused an overrun.
2710    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2711        something other than just ASCII characters) inside a group that had an
2712        unlimited repeat caused a loop at compile time (while checking to see
2713        whether the group could match an empty string).
2715    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2716        [\P{Any}] did so. (Error in the code for printing property names.)
2718    21. An orphan \E inside a character class could cause a crash.
2720    22. A repeated capturing bracket such as (A)? could cause a wild memory
2721        reference during compilation.
2723    23. There are several functions in pcre_compile() that scan along a compiled
2724        expression for various reasons (e.g. to see if it's fixed length for look
2725        behind). There were bugs in these functions when a repeated \p or \P was
2726        present in the pattern. These operators have additional parameters compared
2727        with \d, etc, and these were not being taken into account when moving along
2728        the compiled data. Specifically:
2730        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2731            length.
2733        (b) An item such as \pL+ within a repeated group could cause crashes or
2734            loops.
2736        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2737            "reference to non-existent subpattern" error.
2739        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2741    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2742        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2744    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2746    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2747        character were causing crashes (broken optimization).
2749    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2750        \p or \P) caused a compile-time loop.
2752    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2753        break. For example, the unstudied pattern [\r\n]A does not match the string
2754        "\r\nA" because change 7.0/46 below moves the current point on by two
2755        characters after failing to match at the start. However, the pattern \nA
2756        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2757        the same is true. There doesn't seem any very clean way out of this, but
2758        what I have chosen to do makes the common cases work: PCRE now takes note
2759        of whether there can be an explicit match for \r or \n anywhere in the
2760        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2761        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2762        pattern has explicit CR or LF references.
2764    29. Added (*CR) etc for changing newline setting at start of pattern.
2767    Version 7.2 19-Jun-07
2768    ---------------------
2770     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2771        which is apparently normally available under Windows.
2773     2. Re-jig the pcregrep tests with different newline settings in an attempt
2774        to make them independent of the local environment's newline setting.
2776     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2778     4. Some of the "internals" tests were previously cut out when the link size
2779        was not 2, because the output contained actual offsets. The recent new
2780        "Z" feature of pcretest means that these can be cut out, making the tests
2781        usable with all link sizes.
2783     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2784        stack recursion. This gives a massive performance boost under BSD, but just
2785        a small improvement under Linux. However, it saves one field in the frame
2786        in all cases.
2788     6. Added more features from the forthcoming Perl 5.10:
2790        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2791            recursion call. It refers to the nth most recently opened parentheses.
2793        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2794            to be opened parentheses.
2796        (c) Conditions that refer to capturing parentheses can be specified
2797            relatively, for example, (?(-2)... or (?(+3)...
2799        (d) \K resets the start of the current match so that everything before
2800            is not part of it.
2802        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2804        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2805            reference syntax.
2807        (g) (?| introduces a group in which the numbering of parentheses in each
2808            alternative starts with the same number.
2810        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2812     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2815     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2816        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2817        for detecting groups that can match an empty string.
2819     9. A pattern with a very large number of alternatives (more than several
2820        hundred) was running out of internal workspace during the pre-compile
2821        phase, where pcre_compile() figures out how much memory will be needed. A
2822        bit of new cunning has reduced the workspace needed for groups with
2823        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2824        workspace instead of running out of the 4096 that are available.
2826    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2828    11. Applied patch from Google to remove an optimization that didn't quite work.
2829        The report of the bug said:
2831          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2832          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2833          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2835    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2836        it matched the wrong number of bytes.
2839    Version 7.1 24-Apr-07
2840    ---------------------
2842     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2843        that is more "standard", making use of automake and other Autotools. There
2844        is some re-arrangement of the files and adjustment of comments consequent
2845        on this.
2847     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2848        for recursive directory scanning broke on some systems because the files
2849        are not scanned in any specific order and on different systems the order
2850        was different. A call to "sort" has been inserted into RunGrepTest for the
2851        approprate test as a short-term fix. In the longer term there may be an
2852        alternative.
2854     3. I had an email from Eric Raymond about problems translating some of PCRE's
2855        man pages to HTML (despite the fact that I distribute HTML pages, some
2856        people do their own conversions for various reasons). The problems
2857        concerned the use of low-level troff macros .br and .in. I have therefore
2858        removed all such uses from the man pages (some were redundant, some could
2859        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2860        HTML has been updated to handle .nf/.fi and to complain if it encounters
2861        .br or .in.
2863     4. Updated comments in configure.ac that get placed in config.h.in and also
2864        arranged for config.h to be included in the distribution, with the name
2865        config.h.generic, for the benefit of those who have to compile without
2866        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2868     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2869        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2870        makevp.bat for the current PCRE, using the additional files
2871        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2873     6. A Windows user reported a minor discrepancy with test 2, which turned out
2874        to be caused by a trailing space on an input line that had got lost in his
2875        copy. The trailing space was an accident, so I've just removed it.
2877     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2878        that is needed.
2880     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2881        as "const" (a) because they are and (b) because it helps the PHP
2882        maintainers who have recently made a script to detect big data structures
2883        in the php code that should be moved to the .rodata section. I remembered
2884        to update Builducptable as well, so it won't revert if ucptable.h is ever
2885        re-created.
2887     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2888        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2889        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2890        support is not required. This saves 1.5-2K of code, which is important in
2891        some applications.
2893        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2894        so as not to refer to the tables, even though these functions will never be
2895        called when UTF-8 support is disabled. Otherwise there are problems with a
2896        shared library.
2898    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2900        (a) It was defining its arguments as char * instead of void *.
2902        (b) It was assuming that all moves were upwards in memory; this was true
2903            a long time ago when I wrote it, but is no longer the case.
2905        The emulated memove() is provided for those environments that have neither
2906        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2907        is clearly not the case, as these two bugs were recently reported.
2909    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2910        and Detrail to create the HTML documentation, the .txt form of the man
2911        pages, and it removes trailing spaces from listed files. It also creates
2912        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2913        case, it wraps all the #defines with #ifndefs. This script should be run
2914        before "make dist".
2916    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2917        with Unicode property support.
2919        (a) For a maximizing quantifier, if the two different cases of the
2920            character were of different lengths in their UTF-8 codings (there are
2921            some cases like this - I found 11), and the matching function had to
2922            back up over a mixture of the two cases, it incorrectly assumed they
2923            were both the same length.
2925        (b) When PCRE was configured to use the heap rather than the stack for
2926            recursion during matching, it was not correctly preserving the data for
2927            the other case of a UTF-8 character when checking ahead for a match
2928            while processing a minimizing repeat. If the check also involved
2929            matching a wide character, but failed, corruption could cause an
2930            erroneous result when trying to check for a repeat of the original
2931            character.
2933    13. Some tidying changes to the testing mechanism:
2935        (a) The RunTest script now detects the internal link size and whether there
2936            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2937            values substituted by "configure". (The RunGrepTest script already did
2938            this for UTF-8.) The configure.ac script no longer substitutes the
2939            relevant variables.
2941        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2942            with length and offset values. This means that the output is different
2943            for different internal link sizes. Test 2 is skipped for link sizes
2944            other than 2 because of this, bypassing the problem. Unfortunately,
2945            there was also a test in test 3 (the locale tests) that used /B and
2946            failed for link sizes other than 2. Rather than cut the whole test out,
2947            I have added a new /Z option to pcretest that replaces the length and
2948            offset values with spaces. This is now used to make test 3 independent
2949            of link size. (Test 2 will be tidied up later.)
2951    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2952        segmentation fault instead of returning the appropriate error message.
2954    15. In multiline mode when the newline sequence was set to "any", the pattern
2955        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2956        This doesn't seem right; it now treats the CRLF combination as the line
2957        ending, and so does not match in that case. It's only a pattern such as ^$
2958        that would hit this one: something like ^ABC$ would have failed after \r
2959        and then tried again after \r\n.
2961    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2962        in an attempt to make files that differ only in their line terminators
2963        compare equal. This works on Linux.
2965    17. Under certain error circumstances pcregrep might try to free random memory
2966        as it exited. This is now fixed, thanks to valgrind.
2968    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2969        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2970        was because its rules for how to advance for /g after matching an empty
2971        string at the end of a line did not allow for this case. They now check for
2972        it specially.
2974    20. pcretest is supposed to handle patterns and data of any length, by
2975        extending its buffers when necessary. It was getting this wrong when the
2976        buffer for a data line had to be extended.
2978    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2979        CRLF as a newline sequence.
2981    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2982        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2983        I have nevertheless tidied it up.
2985    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2987    24. Added a man page for pcre-config.
2990    Version 7.0 19-Dec-06
2991    ---------------------
2993     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2994        moving to gcc 4.1.1.
2996     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2997        sys/time.h, which is documented as needed for this function. It doesn't
2998        seem to matter on Linux, but it showed up on some releases of OS X.
3000     3. It seems that there are systems where bytes whose values are greater than
3001        127 match isprint() in the "C" locale. The "C" locale should be the
3002        default when a C program starts up. In most systems, only ASCII printing
3003        characters match isprint(). This difference caused the output from pcretest
3004        to vary, making some of the tests fail. I have changed pcretest so that:
3006        (a) When it is outputting text in the compiled version of a pattern, bytes
3007            other than 32-126 are always shown as hex escapes.
3009        (b) When it is outputting text that is a matched part of a subject string,
3010            it does the same, unless a different locale has been set for the match
3011            (using the /L modifier). In this case, it uses isprint() to decide.
3013     4. Fixed a major bug that caused incorrect computation of the amount of memory
3014        required for a compiled pattern when options that changed within the
3015        pattern affected the logic of the preliminary scan that determines the
3016        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3017        that the computed length was too small. The symptoms of this bug were
3018        either the PCRE error "internal error: code overflow" from pcre_compile(),
3019        or a glibc crash with a message such as "pcretest: free(): invalid next
3020        size (fast)". Examples of patterns that provoked this bug (shown in
3021        pcretest format) are:
3023          /(?-x: )/x
3024          /(?x)(?-x: \s*#\s*)/
3025          /((?i)[\x{c0}])/8
3026          /(?i:[\x{c0}])/8
3028        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3029        is now done differently.
3031     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3032        wrapper classes; (b) implement a new function in the C++ scanner that is
3033        more efficient than the old way of doing things because it avoids levels of
3034        recursion in the regex matching; (c) add a paragraph to the documentation
3035        for the FullMatch() function.
3037     6. The escape sequence \n was being treated as whatever was defined as
3038        "newline". Not only was this contrary to the documentation, which states
3039        that \n is character 10 (hex 0A), but it also went horribly wrong when
3040        "newline" was defined as CRLF. This has been fixed.
3042     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3043        was being set to -1 for the "end of line" case (supposedly a value that no
3044        character can have). Though this value is never used (the check for end of
3045        line is "zero bytes in current character"), it caused compiler complaints.
3046        I've changed it to 0xffffffff.
3048     8. In pcre_version.c, the version string was being built by a sequence of
3049        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3050        string (as it is for production releases) called a macro with an empty
3051        argument. The C standard says the result of this is undefined. The gcc
3052        compiler treats it as an empty string (which was what was wanted) but it is
3053        reported that Visual C gives an error. The source has been hacked around to
3054        avoid this problem.
3056     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3057        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3058        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3059        of them did).
3061    10. Originally, pcretest opened its input and output without "b"; then I was
3062        told that "b" was needed in some environments, so it was added for release
3063        5.0 to both the input and output. (It makes no difference on Unix-like
3064        systems.) Later I was told that it is wrong for the input on Windows. I've
3065        now abstracted the modes into two macros, to make it easier to fiddle with
3066        them, and removed "b" from the input mode under Windows.
3068    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3070    12. Added -help and --help to pcretest as an official way of being reminded
3071        of the options.
3073    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3074        and pcrecpp.cc because they annoy compilers at high warning levels.
3076    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3078    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3079        scripts are not C programs :-) and which was not noticed because it works
3080        on Linux.
3082    16. pcretest is supposed to handle any length of pattern and data line (as one
3083        line or as a continued sequence of lines) by extending its input buffer if
3084        necessary. This feature was broken for very long pattern lines, leading to
3085        a string of junk being passed to pcre_compile() if the pattern was longer
3086        than about 50K.
3088    17. I have done a major re-factoring of the way pcre_compile() computes the
3089        amount of memory needed for a compiled pattern. Previously, there was code
3090        that made a preliminary scan of the pattern in order to do this. That was
3091        OK when PCRE was new, but as the facilities have expanded, it has become
3092        harder and harder to keep it in step with the real compile phase, and there
3093        have been a number of bugs (see for example, 4 above). I have now found a
3094        cunning way of running the real compile function in a "fake" mode that
3095        enables it to compute how much memory it would need, while actually only
3096        ever using a few hundred bytes of working memory and without too many
3097        tests of the mode. This should make future maintenance and development
3098        easier. A side effect of this work is that the limit of 200 on the nesting
3099        depth of parentheses has been removed (though this was never a serious
3100        limitation, I suspect). However, there is a downside: pcre_compile() now
3101        runs more slowly than before (30% or more, depending on the pattern). I
3102        hope this isn't a big issue. There is no effect on runtime performance.
3104    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3105        newline (only possible for the last line of a file) and it was a
3106        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3108    19. Added additional timing features to pcretest. (1) The -tm option now times
3109        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3110        separate command line item, by a number that specifies the number of
3111        repeats to use when timing. The default is 50000; this gives better
3112        precision, but takes uncomfortably long for very large patterns.
3114    20. Extended pcre_study() to be more clever in cases where a branch of a
3115        subpattern has no definite first character. For example, (a*|b*)[cd] would
3116        previously give no result from pcre_study(). Now it recognizes that the
3117        first character must be a, b, c, or d.
3119    21. There was an incorrect error "recursive call could loop indefinitely" if
3120        a subpattern (or the entire pattern) that was being tested for matching an
3121        empty string contained only one non-empty item after a nested subpattern.
3122        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3123        incorrectly, because the \d was being skipped in the check.
3125    22. The pcretest program now has a new pattern option /B and a command line
3126        option -b, which is equivalent to adding /B to every pattern. This causes
3127        it to show the compiled bytecode, without the additional information that
3128        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3129        is the same as /B/I).
3131    23. A new optimization is now able automatically to treat some sequences such
3132        as a*b as a*+b. More specifically, if something simple (such as a character
3133        or a simple class like \d) has an unlimited quantifier, and is followed by
3134        something that cannot possibly match the quantified thing, the quantifier
3135        is automatically "possessified".
3137    24. A recursive reference to a subpattern whose number was greater than 39
3138        went wrong under certain circumstances in UTF-8 mode. This bug could also
3139        have affected the operation of pcre_study().
3141    25. Realized that a little bit of performance could be had by replacing
3142        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3144    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3146    27. Possessive quantifiers such as a++ were previously implemented by turning
3147        them into atomic groups such as ($>a+). Now they have their own opcodes,
3148        which improves performance. This includes the automatically created ones
3149        from 23 above.
3151    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3152        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3153        the first matched character to be a colon. This applied both to named and
3154        numbered groups.
3156    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3158    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3159        building PCRE on Windows easier, so I have included it in the distribution.
3161    31. There is now a check in pcretest against a ridiculously large number being
3162        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3163        loop, the loop is abandoned.
3165    32. Forward references to subpatterns in conditions such as (?(2)...) where
3166        subpattern 2 is defined later cause pcre_compile() to search forwards in
3167        the pattern for the relevant set of parentheses. This search went wrong
3168        when there were unescaped parentheses in a character class, parentheses
3169        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3171    33. "Subroutine" calls and backreferences were previously restricted to
3172        referencing subpatterns earlier in the regex. This restriction has now
3173        been removed.
3175    34. Added a number of extra features that are going to be in Perl 5.10. On the
3176        whole, these are just syntactic alternatives for features that PCRE had
3177        previously implemented using the Python syntax or my own invention. The
3178        other formats are all retained for compatibility.
3180        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3181            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3182            also .NET compatible.
3184        (b) A recursion or subroutine call to a named group can now be defined as
3185            (?&name) as well as (?P>name).
3187        (c) A backreference to a named group can now be defined as \k<name> or
3188            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3189            5.10, are also .NET compatible.
3191        (d) A conditional reference to a named group can now use the syntax
3192            (?(<name>) or (?('name') as well as (?(name).
3194        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3195            groups (named and numbered) that are never evaluated inline, but can be
3196            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3197            is always false. There may be only one alternative in such a group.
3199        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3200            as the simple (?(R). The condition is true only if the most recent
3201            recursion is that of the given number or name. It does not search out
3202            through the entire recursion stack.
3204        (g) The escape \gN or \g{N} has been added, where N is a positive or
3205            negative number, specifying an absolute or relative reference.
3207    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3208        some "unreachable code" warnings.
3210    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3211        things, this adds five new scripts.
3213    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3214        There were also incompatibilities regarding the handling of \Q..\E inside
3215        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3216        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3218    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3219        matches an empty string, and forcibly breaks the loop. There were bugs in
3220        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3221        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3222        separate and independent bugs (that affected different cases) have been
3223        fixed.
3225    39. Refactored the code to abolish the use of different opcodes for small
3226        capturing bracket numbers. This is a tidy that I avoided doing when I
3227        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3228        The new approach is not only tidier, it makes it possible to reduce the
3229        memory needed to fix the previous bug (38).
3231    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3232        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3233        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3234        mode.
3236    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3237        report.
3239    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3240        copy construction and assignment in the C++ wrapper.
3242    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3243        couple of bugs that could have given wrong results in the "--newline=crlf"
3244        case.
3246    44. Added a number of casts and did some reorganization of signed/unsigned int
3247        variables following suggestions from Dair Grant. Also renamed the variable
3248        "this" as "item" because it is a C++ keyword.
3250    45. Arranged for dftables to add
3252          #include "pcre_internal.h"
3254        to pcre_chartables.c because without it, gcc 4.x may remove the array
3255        definition from the final binary if PCRE is built into a static library and
3256        dead code stripping is activated.
3258    46. For an unanchored pattern, if a match attempt fails at the start of a
3259        newline sequence, and the newline setting is CRLF or ANY, and the next two
3260        characters are CRLF, advance by two characters instead of one.
3263    Version 6.7 04-Jul-06
3264    ---------------------
3266     1. In order to handle tests when input lines are enormously long, pcretest has
3267        been re-factored so that it automatically extends its buffers when
3268        necessary. The code is crude, but this _is_ just a test program. The
3269        default size has been increased from 32K to 50K.
3271     2. The code in pcre_study() was using the value of the re argument before
3272        testing it for NULL. (Of course, in any sensible call of the function, it
3273        won't be NULL.)
3275     3. The memmove() emulation function in pcre_internal.h, which is used on
3276        systems that lack both memmove() and bcopy() - that is, hardly ever -
3277        was missing a "static" storage class specifier.
3279     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3280        containing an extended class (one that cannot be represented by a bitmap
3281        because it contains high-valued characters or Unicode property items, e.g.
3282        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3283        pattern, but PCRE should not loop if you do not (it no longer does).
3284        [Detail: two cases were found: (a) a repeated subpattern containing an
3285        extended class; (b) a recursive reference to a subpattern that followed a
3286        previous extended class. It wasn't skipping over the extended class
3287        correctly when UTF-8 mode was not set.]
3289     5. A negated single-character class was not being recognized as fixed-length
3290        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3291        compile error "lookbehind assertion is not fixed length".
3293     6. The RunPerlTest auxiliary script was showing an unexpected difference
3294        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3295        write a Perl script that can interpret lines of an input file either as
3296        byte characters or as UTF-8, which is what "perltest" was being required to
3297        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3298        can't do is switch easily at run time between having the "use utf8;" pragma
3299        or not. In the end, I fudged it by using the RunPerlTest script to insert
3300        "use utf8;" explicitly for the UTF-8 tests.
3302     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3303        the end of the subject string, contrary to the documentation and to what
3304        Perl does. This was true of both matching functions. Now it matches only at
3305        the start of the subject and immediately after *internal* newlines.
3307     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3308        a pointer to an int instead of a pointer to an unsigned long int. This
3309        caused problems on 64-bit systems.
3311     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3312        instance of the 'standard' template library not being so standard".
3314    10. There was no check on the number of named subpatterns nor the maximum
3315        length of a subpattern name. The product of these values is used to compute
3316        the size of the memory block for a compiled pattern. By supplying a very
3317        long subpattern name and a large number of named subpatterns, the size
3318        computation could be caused to overflow. This is now prevented by limiting
3319        the length of names to 32 characters, and the number of named subpatterns
3320        to 10,000.
3322    11. Subpatterns that are repeated with specific counts have to be replicated in
3323        the compiled pattern. The size of memory for this was computed from the
3324        length of the subpattern and the repeat count. The latter is limited to
3325        65535, but there was no limit on the former, meaning that integer overflow
3326        could in principle occur. The compiled length of a repeated subpattern is
3327        now limited to 30,000 bytes in order to prevent this.
3329    12. Added the optional facility to have named substrings with the same name.
3331    13. Added the ability to use a named substring as a condition, using the
3332        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3333        are numbers (not recommended). Forward references are permitted.
3335    14. Added forward references in named backreferences (if you see what I mean).
3337    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3338        pattern could run off the end of the subject. For example, the pattern
3339        "(?s)(.{1,5})"8 did this with the subject "ab".
3341    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3342        PCRE_CASELESS was set when matching characters that were quantified with ?
3343        or *.
3345    17. A character class other than a single negated character that had a minimum
3346        but no maximum quantifier - for example [ab]{6,} - was not handled
3347        correctly by pce_dfa_exec(). It would match only one character.
3349    18. A valid (though odd) pattern that looked like a POSIX character
3350        class but used an invalid character after [ (for example [[,abc,]]) caused
3351        pcre_compile() to give the error "Failed: internal error: code overflow" or
3352        in some cases to crash with a glibc free() error. This could even happen if
3353        the pattern terminated after [[ but there just happened to be a sequence of
3354        letters, a binary zero, and a closing ] in the memory that followed.
3356    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3357        over the years. Originally (before any Unicode support), just the bottom 8
3358        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3359        output from "man perlunicode" includes this:
3361          The regular expression compiler produces polymorphic opcodes.  That
3362          is, the pattern adapts to the data and automatically switches to
3363          the Unicode character scheme when presented with Unicode data--or
3364          instead uses a traditional byte scheme when presented with byte
3365          data.
3367        Sadly, a wide octal escape does not cause a switch, and in a string with
3368        no other multibyte characters, these octal escapes are treated as before.
3369        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3370        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3371        Unicode string.
3373        I have not perpetrated such confusion in PCRE. Up till now, it took just
3374        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3375        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3376        translate to the appropriate multibyte character.
3378    29. Applied some refactoring to reduce the number of warnings from Microsoft
3379        and Borland compilers. This has included removing the fudge introduced
3380        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3381        a warning about an unused variable.
3383    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3384        characters since release 4.0, because Perl (from release 5.004) does not.
3385        [Or at least, is documented not to: some releases seem to be in conflict
3386        with the documentation.] However, when a pattern was studied with
3387        pcre_study() and all its branches started with \s, PCRE still included VT
3388        as a possible starting character. Of course, this did no harm; it just
3389        caused an unnecessary match attempt.
3391    22. Removed a now-redundant internal flag bit that recorded the fact that case
3392        dependency changed within the pattern. This was once needed for "required
3393        byte" processing, but is no longer used. This recovers a now-scarce options
3394        bit. Also moved the least significant internal flag bit to the most-
3395        significant bit of the word, which was not previously used (hangover from
3396        the days when it was an int rather than a uint) to free up another bit for
3397        the future.
3399    23. Added support for CRLF line endings as well as CR and LF. As well as the
3400        default being selectable at build time, it can now be changed at runtime
3401        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3402        specify that it is scanning data with non-default line endings.
3404    24. Changed the definition of CXXLINK to make it agree with the definition of
3405        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3407    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3408        recursions. This makes a big different to stack usage for some patterns.
3410    26. If a subpattern containing a named recursion or subroutine reference such
3411        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3412        the space required for the compiled pattern went wrong and gave too small a
3413        value. Depending on the environment, this could lead to "Failed: internal
3414        error: code overflow at offset 49" or "glibc detected double free or
3415        corruption" errors.
3417    27. Applied patches from Google (a) to support the new newline modes and (b) to
3418        advance over multibyte UTF-8 characters in GlobalReplace.
3420    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3421        difference for some implementation of PCRE in some Windows version.
3423    29. Added some extra testing facilities to pcretest:
3425        \q<number>   in a data line sets the "match limit" value
3426        \Q<number>   in a data line sets the "match recursion limt" value
3427        -S <number>  sets the stack size, where <number> is in megabytes
3429        The -S option isn't available for Windows.
3432    Version 6.6 06-Feb-06
3433    ---------------------
3435     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3436        in pcreposix.h. I have copied the definition from pcre.h.
3438     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3439        because pcre.h is no longer a built file.
3441     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3442        not normally included in the compiled code.
3445    Version 6.5 01-Feb-06
3446    ---------------------
3448     1. When using the partial match feature with pcre_dfa_exec(), it was not
3449        anchoring the second and subsequent partial matches at the new starting
3450        point. This could lead to incorrect results. For example, with the pattern
3451        /1234/, partially matching against "123" and then "a4" gave a match.
3453     2. Changes to pcregrep:
3455        (a) All non-match returns from pcre_exec() were being treated as failures
3456            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3457            error message is output. Some extra information is given for the
3458            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3459            probably the only errors that are likely to be caused by users (by
3460            specifying a regex that has nested indefinite repeats, for instance).
3461            If there are more than 20 of these errors, pcregrep is abandoned.
3463        (b) A binary zero was treated as data while matching, but terminated the
3464            output line if it was written out. This has been fixed: binary zeroes
3465            are now no different to any other data bytes.
3467        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3468            used to set a locale for matching. The --locale=xxxx long option has
3469            been added (no short equivalent) to specify a locale explicitly on the
3470            pcregrep command, overriding the environment variables.
3472        (d) When -B was used with -n, some line numbers in the output were one less
3473            than they should have been.
3475        (e) Added the -o (--only-matching) option.
3477        (f) If -A or -C was used with -c (count only), some lines of context were
3478            accidentally printed for the final match.
3480        (g) Added the -H (--with-filename) option.
3482        (h) The combination of options -rh failed to suppress file names for files
3483            that were found from directory arguments.
3485        (i) Added the -D (--devices) and -d (--directories) options.
3487        (j) Added the -F (--fixed-strings) option.
3489        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3491        (l) Added the --colo(u)r option.
3493        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3494            is not present by default.
3496     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3497        items such as (?R) or (?1), when the recursion could match a number of
3498        alternatives. If it matched one of the alternatives, but subsequently,
3499        outside the recursion, there was a failure, the code tried to back up into
3500        the recursion. However, because of the way PCRE is implemented, this is not
3501        possible, and the result was an incorrect result from the match.
3503        In order to prevent this happening, the specification of recursion has
3504        been changed so that all such subpatterns are automatically treated as
3505        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3507     4. I had overlooked the fact that, in some locales, there are characters for
3508        which isalpha() is true but neither isupper() nor islower() are true. In
3509        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3510        and ordfeminine) are like this. This affected the treatment of \w and \W
3511        when they appeared in character classes, but not when they appeared outside
3512        a character class. The bit map for "word" characters is now created
3513        separately from the results of isalnum() instead of just taking it from the
3514        upper, lower, and digit maps. (Plus the underscore character, of course.)
3516     5. The above bug also affected the handling of POSIX character classes such as
3517        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3518        permanent tables. Instead, the bit maps for such a class were previously
3519        created as the appropriate unions of the upper, lower, and digit bitmaps.
3520        Now they are created by subtraction from the [[:word:]] class, which has
3521        its own bitmap.
3523     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3524        It is created by subtracting the vertical space characters (\x09, \x0a,
3525        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3526        subtraction was done in the overall bitmap for a character class, meaning
3527        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3528        be recognized. This bug has been fixed.
3530     7. Patches from the folks at Google:
3532          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3533          real life, but is still worth protecting against".
3535          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3536          regular expressions".
3538          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3539          have it.
3541          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3542          "configure" and the latter not, in order to fix a problem somebody had
3543          with compiling the Arg class on HP-UX.
3545          (e) Improve the error-handling of the C++ wrapper a little bit.
3547          (f) New tests for checking recursion limiting.
3549     8. The pcre_memmove() function, which is used only if the environment does not
3550        have a standard memmove() function (and is therefore rarely compiled),
3551        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3552        returning a result (though PCRE never actually uses the result).
3554     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3555        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3556        returned instead of calling malloc() with an overflowing number that would
3557        most likely cause subsequent chaos.
3559    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3561    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3562        with this option is matched, the nmatch and pmatch options of regexec() are
3563        ignored.
3565    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3566        provided in case anyone wants to the the POSIX interface with UTF-8
3567        strings.
3569    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3570        C++ linking (needed for some HP-UX environments).
3572    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3573        (unused parameter) and in the pcre_printint() function (omitted "default"
3574        switch label when the default is to do nothing).
3576    15. Added some code to make it possible, when PCRE is compiled as a C++
3577        library, to replace subject pointers for pcre_exec() with a smart pointer
3578        class, thus making it possible to process discontinuous strings.
3580    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3581        much the same function. They were added by different people who were trying