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

Diff of /code/trunk/ChangeLog

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

revision 158 by ph10, Tue May 1 10:30:14 2007 UTC revision 1481 by ph10, Tue May 27 18:24:42 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-07  Version 8.36 xx-xxx-2014
5    ------------------------
6    
7    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
8        -Wmissing-field-initializers and -Wunused-parameter.
9    
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.
14    
15    3.  Tidy code in pcre_exec.c where two branches that used to be different are
16        now the same.
17    
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.
22    
23    5.  Fixed an issue, which occures when nested alternatives are optimized
24        with table jumps.
25    
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).
28    
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:
33    
34          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
35          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
36    
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.
43    
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".
46    
47    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
48        doesn't).
49    
50    
51    Version 8.35 04-April-2014
52    --------------------------
53    
54    1.  A new flag is set, when property checks are present in an XCLASS.
55        When this flag is not set, PCRE can perform certain optimizations
56        such as studying these XCLASS-es.
57    
58    2.  The auto-possessification of character sets were improved: a normal
59        and an extended character set can be compared now. Furthermore
60        the JIT compiler optimizes more character set checks.
61    
62    3.  Got rid of some compiler warnings for potentially uninitialized variables
63        that show up only when compiled with -O2.
64    
65    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
66        of a match later then the end of the match. The pcretest program was not
67        handling the case sensibly - it was outputting from the start to the next
68        binary zero. It now reports this situation in a message, and outputs the
69        text from the end to the start.
70    
71    5.  Fast forward search is improved in JIT. Instead of the first three
72        characters, any three characters with fixed position can be searched.
73        Search order: first, last, middle.
74    
75    6.  Improve character range checks in JIT. Characters are read by an inprecise
76        function now, which returns with an unknown value if the character code is
77        above a certain treshold (e.g: 256). The only limitation is that the value
78        must be bigger than the treshold as well. This function is useful, when
79        the characters above the treshold are handled in the same way.
80    
81    7.  The macros whose names start with RAWUCHAR are placeholders for a future
82        mode in which only the bottom 21 bits of 32-bit data items are used. To
83        make this more memorable for those maintaining the code, the names have
84        been changed to start with UCHAR21, and an extensive comment has been added
85        to their definition.
86    
87    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
88        to the export list in Makefile.am (they were accidentally omitted from the
89        8.34 tarball).
90    
91    9.  The informational output from pcretest used the phrase "starting byte set"
92        which is inappropriate for the 16-bit and 32-bit libraries. As the output
93        for "first char" and "need char" really means "non-UTF-char", I've changed
94        "byte" to "char", and slightly reworded the output. The documentation about
95        these values has also been (I hope) clarified.
96    
97    10. Another JIT related optimization: use table jumps for selecting the correct
98        backtracking path, when more than four alternatives are present inside a
99        bracket.
100    
101    11. Empty match is not possible, when the minimum length is greater than zero,
102        and there is no \K in the pattern. JIT should avoid empty match checks in
103        such cases.
104    
105    12. In a caseless character class with UCP support, when a character with more
106        than one alternative case was not the first character of a range, not all
107        the alternative cases were added to the class. For example, s and \x{17f}
108        are both alternative cases for S: the class [RST] was handled correctly,
109        but [R-T] was not.
110    
111    13. The configure.ac file always checked for pthread support when JIT was
112        enabled. This is not used in Windows, so I have put this test inside a
113        check for the presence of windows.h (which was already tested for).
114    
115    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
116        The algorithm provides a way to skip certain starting offsets, and usually
117        faster than linear prefix searches.
118    
119    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
120        as for 'fr_FR' and 'french'. For some reason, however, it then used the
121        Windows-specific input and output files, which have 'french' screwed in.
122        So this could never have worked. One of the problems with locales is that
123        they aren't always the same. I have now updated RunTest so that it checks
124        the output of the locale test (test 3) against three different output
125        files, and it allows the test to pass if any one of them matches. With luck
126        this should make the test pass on some versions of Solaris where it was
127        failing. Because of the uncertainty, the script did not used to stop if
128        test 3 failed; it now does. If further versions of a French locale ever
129        come to light, they can now easily be added.
130    
131    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
132        there was a message during ./configure, but it did not stop. This now
133        provokes an error. The invalid example in README has been corrected.
134        If a value less than the minimum is given, the minimum value has always
135        been used, but now a warning is given.
136    
137    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
138        was a bug in the test system, which is now fixed. Also, the list of various
139        configurations that are tested for each release did not have one with both
140        16/32 bits and --enable-bar-anycrlf. It now does.
141    
142    18. pcretest was missing "-C bsr" for displaying the \R default setting.
143    
144    19. Little endian PowerPC systems are supported now by the JIT compiler.
145    
146    20. The fast forward newline mechanism could enter to an infinite loop on
147        certain invalid UTF-8 input. Although we don't support these cases
148        this issue can be fixed by a performance optimization.
149    
150    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
151        not take account if existing stack usage. There is now a new global
152        variable called pcre_stack_guard that can be set to point to an external
153        function to check stack availability. It is called at the start of
154        processing every parenthesized group.
155    
156    22. A typo in the code meant that in ungreedy mode the max/min qualifier
157        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
158        not match "ab".
159    
160    23. When UTF was disabled, the JIT program reported some incorrect compile
161        errors. These messages are silenced now.
162    
163    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
164        compiler.
165    
166    25. Change all the temporary files used in RunGrepTest to be different to those
167        used by RunTest so that the tests can be run simultaneously, for example by
168        "make -j check".
169    
170    
171    Version 8.34 15-December-2013
172    -----------------------------
173    
174    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
175        executable memory. Patch inspired by Carsten Klein.
176    
177    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
178        this macro is never tested and has no effect, because the work to support
179        coverage involves only compiling and linking options and special targets in
180        the Makefile. The comment in config.h implied that defining the macro would
181        enable coverage support, which is totally false. There was also support for
182        setting this macro in the CMake files (my fault, I just copied it from
183        configure). SUPPORT_GCOV has now been removed.
184    
185    3.  Make a small performance improvement in strlen16() and strlen32() in
186        pcretest.
187    
188    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
189        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
190        these cases). There was also one in pcretest.c.
191    
192    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
193    
194    6.  In UTF mode, the code for checking whether a group could match an empty
195        string (which is used for indefinitely repeated groups to allow for
196        breaking an infinite loop) was broken when the group contained a repeated
197        negated single-character class with a character that occupied more than one
198        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
199        UTF-8 mode). The effect was undefined: the group might or might not be
200        deemed as matching an empty string, or the program might have crashed.
201    
202    7.  The code for checking whether a group could match an empty string was not
203        recognizing that \h, \H, \v, \V, and \R must match a character.
204    
205    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
206        an empty string. If it can, pcretest shows this in its information output.
207    
208    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
209        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
210        matched by pcre_exec() without using JIT:
211    
212        (a) If the rest of the pattern did not match after a maximal run of
213            grapheme clusters, the code for backing up to try with fewer of them
214            did not always back up over a full grapheme when characters that do not
215            have the modifier quality were involved, e.g. Hangul syllables.
216    
217        (b) If the match point in a subject started with modifier character, and
218            there was no match, the code could incorrectly back up beyond the match
219            point, and potentially beyond the first character in the subject,
220            leading to a segfault or an incorrect match result.
221    
222    10. A conditional group with an assertion condition could lead to PCRE
223        recording an incorrect first data item for a match if no other first data
224        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
225        first data item, and therefore matched "ca" after "c" instead of at the
226        start.
227    
228    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
229        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
230    
231    12. The source of pcregrep now includes z/OS-specific code so that it can be
232        compiled for z/OS as part of the special z/OS distribution.
233    
234    13. Added the -T and -TM options to pcretest.
235    
236    14. The code in pcre_compile.c for creating the table of named capturing groups
237        has been refactored. Instead of creating the table dynamically during the
238        actual compiling pass, the information is remembered during the pre-compile
239        pass (on the stack unless there are more than 20 named groups, in which
240        case malloc() is used) and the whole table is created before the actual
241        compile happens. This has simplified the code (it is now nearly 150 lines
242        shorter) and prepared the way for better handling of references to groups
243        with duplicate names.
244    
245    15. A back reference to a named subpattern when there is more than one of the
246        same name now checks them in the order in which they appear in the pattern.
247        The first one that is set is used for the reference. Previously only the
248        first one was inspected. This change makes PCRE more compatible with Perl.
249    
250    16. Unicode character properties were updated from Unicode 6.3.0.
251    
252    17. The compile-time code for auto-possessification has been refactored, based
253        on a patch by Zoltan Herczeg. It now happens after instead of during
254        compilation. The code is cleaner, and more cases are handled. The option
255        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
256        options in pcretest are provided to set it. It can also be set by
257        (*NO_AUTO_POSSESS) at the start of a pattern.
258    
259    18. The character VT has been added to the default ("C" locale) set of
260        characters that match \s and are generally treated as white space,
261        following this same change in Perl 5.18. There is now no difference between
262        "Perl space" and "POSIX space". Whether VT is treated as white space in
263        other locales depends on the locale.
264    
265    19. The code for checking named groups as conditions, either for being set or
266        for being recursed, has been refactored (this is related to 14 and 15
267        above). Processing unduplicated named groups should now be as fast at
268        numerical groups, and processing duplicated groups should be faster than
269        before.
270    
271    20. Two patches to the CMake build system, by Alexander Barkov:
272    
273          (1) Replace the "source" command by "." in CMakeLists.txt because
274              "source" is a bash-ism.
275    
276          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
277              without these the CMake build does not work on Solaris.
278    
279    21. Perl has changed its handling of \8 and \9. If there is no previously
280        encountered capturing group of those numbers, they are treated as the
281        literal characters 8 and 9 instead of a binary zero followed by the
282        literals. PCRE now does the same.
283    
284    22. Following Perl, added \o{} to specify codepoints in octal, making it
285        possible to specify values greater than 0777 and also making them
286        unambiguous.
287    
288    23. Perl now gives an error for missing closing braces after \x{... instead of
289        treating the string as literal. PCRE now does the same.
290    
291    24. RunTest used to grumble if an inappropriate test was selected explicitly,
292        but just skip it when running all tests. This make it awkward to run ranges
293        of tests when one of them was inappropriate. Now it just skips any
294        inappropriate tests, as it always did when running all tests.
295    
296    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
297        character types such as \d or \w, too many callouts were inserted, and the
298        data that they returned was rubbish.
299    
300    26. In UCP mode, \s was not matching two of the characters that Perl matches,
301        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
302        were matched by \h. The code has now been refactored so that the lists of
303        the horizontal and vertical whitespace characters used for \h and \v (which
304        are defined only in one place) are now also used for \s.
305    
306    27. Add JIT support for the 64 bit TileGX architecture.
307        Patch by Jiong Wang (Tilera Corporation).
308    
309    28. Possessive quantifiers for classes (both explicit and automatically
310        generated) now use special opcodes instead of wrapping in ONCE brackets.
311    
312    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
313        (because it's meaningless), this was not happening when PCRE_CASELESS was
314        set. Not wrong, but inefficient.
315    
316    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
317        properties for \w, \d, etc) is present in a test regex. Otherwise if the
318        test contains no characters greater than 255, Perl doesn't realise it
319        should be using Unicode semantics.
320    
321    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
322        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
323        does in Unicode mode.
324    
325    32. Added the "forbid" facility to pcretest so that putting tests into the
326        wrong test files can sometimes be quickly detected.
327    
328    33. There is now a limit (default 250) on the depth of nesting of parentheses.
329        This limit is imposed to control the amount of system stack used at compile
330        time. It can be changed at build time by --with-parens-nest-limit=xxx or
331        the equivalent in CMake.
332    
333    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
334        errors. Perl warns for these when in warning mode, but PCRE has no facility
335        for giving warnings.
336    
337    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
338        However, this was not working for (?!) because it is optimized to (*FAIL),
339        for which PCRE does not allow quantifiers. The optimization is now disabled
340        when a quantifier follows (?!). I can't see any use for this, but it makes
341        things uniform.
342    
343    36. Perl no longer allows group names to start with digits, so I have made this
344        change also in PCRE. It simplifies the code a bit.
345    
346    37. In extended mode, Perl ignores spaces before a + that indicates a
347        possessive quantifier. PCRE allowed a space before the quantifier, but not
348        before the possessive +. It now does.
349    
350    38. The use of \K (reset reported match start) within a repeated possessive
351        group such as (a\Kb)*+ was not working.
352    
353    40. Document that the same character tables must be used at compile time and
354        run time, and that the facility to pass tables to pcre_exec() and
355        pcre_dfa_exec() is for use only with saved/restored patterns.
356    
357    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
358        features for Builds with MSVC:
359    
360        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
361           builds). The .rc files can be used to set FileDescription and many other
362           attributes.
363    
364        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
365           This allows higher-level build scripts which want .pdb files to avoid
366           hard-coding the exact files needed."
367    
368    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
369        mean "start of word" and "end of word", respectively, as a transition aid.
370    
371    43. A minimizing repeat of a class containing codepoints greater than 255 in
372        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
373        compiled to use the heap for recursion.
374    
375    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
376        is configured.
377    
378    
379    Version 8.33 28-May-2013
380    ------------------------
381    
382    1.  Added 'U' to some constants that are compared to unsigned integers, to
383        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
384        variables that are added to signed variables, to ensure the result is
385        signed and can be negated.
386    
387    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
388        CMake config files.
389    
390    3.  Revise the creation of config.h.generic so that all boolean macros are
391        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
392        overriding via -D on the command line possible.
393    
394    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
395        to unsigned int is reported to make a quite noticeable speed difference in
396        a specific Windows environment. Testing on Linux did also appear to show
397        some benefit (and it is clearly not harmful). Also fixed the definition of
398        Xop which should be unsigned.
399    
400    5.  Related to (4), changing the definition of the intermediate variable cc
401        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
402        improvements.
403    
404    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
405        unnecessary spaces.
406    
407    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
408        and later.
409    
410    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
411    
412    9.  Optimizing fast_forward_start_bits in JIT.
413    
414    10. Adding support for callouts in JIT, and fixing some issues revealed
415        during this work. Namely:
416    
417        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
418    
419        (b) Minimum length was not checked before the matching is started.
420    
421    11. The value of capture_last that is passed to callouts was incorrect in some
422        cases when there was a capture on one path that was subsequently abandoned
423        after a backtrack. Also, the capture_last value is now reset after a
424        recursion, since all captures are also reset in this case.
425    
426    12. The interpreter no longer returns the "too many substrings" error in the
427        case when an overflowing capture is in a branch that is subsequently
428        abandoned after a backtrack.
429    
430    13. In the pathological case when an offset vector of size 2 is used, pcretest
431        now prints out the matched string after a yield of 0 or 1.
432    
433    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
434        Only supported by the JIT compiler at the moment.
435    
436    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
437    
438    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
439        offset of the starting point of the matching process, provided the offsets
440        vector is large enough.
441    
442    17. The \A escape now records a lookbehind value of 1, though its execution
443        does not actually inspect the previous character. This is to ensure that,
444        in partial multi-segment matching, at least one character from the old
445        segment is retained when a new segment is processed. Otherwise, if there
446        are no lookbehinds in the pattern, \A might match incorrectly at the start
447        of a new segment.
448    
449    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
450    
451    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
452        gives some modest performance improvement in 8-bit mode.
453    
454    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
455        be expressed in certain programming languages using Universal Character
456        Names.
457    
458    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
459        which points out that "non characters" are not "characters that may not
460        appear in Unicode strings" but rather "characters that are reserved for
461        internal use and have only local meaning".
462    
463    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
464        there was a conditional group that depended on an assertion, if the
465        assertion was false, the callout that immediately followed the alternation
466        in the condition was skipped when pcre_exec() was used for matching.
467    
468    23. Allow an explicit callout to be inserted before an assertion that is the
469        condition for a conditional group, for compatibility with automatic
470        callouts, which always insert a callout at this point.
471    
472    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
473        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
474    
475    25. (*PRUNE) is now supported by the JIT compiler.
476    
477    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
478    
479    27. Fix the case where there are two or more SKIPs with arguments that may be
480        ignored.
481    
482    28. (*SKIP) is now supported by the JIT compiler.
483    
484    29. (*THEN) is now supported by the JIT compiler.
485    
486    30. Update RunTest with additional test selector options.
487    
488    31. The way PCRE handles backtracking verbs has been changed in two ways.
489    
490        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
491        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
492        backtracking. In some cases this makes it more Perl-compatible, but Perl's
493        rather obscure rules do not always do the same thing.
494    
495        (2) Previously, backtracking verbs were confined within assertions. This is
496        no longer the case for positive assertions, except for (*ACCEPT). Again,
497        this sometimes improves Perl compatibility, and sometimes does not.
498    
499    32. A number of tests that were in test 2 because Perl did things differently
500        have been moved to test 1, because either Perl or PCRE has changed, and
501        these tests are now compatible.
502    
503    32. Backtracking control verbs are now handled in the same way in JIT and
504        interpreter.
505    
506    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
507        contained a forward subroutine reference caused a compile error.
508    
509    34. Auto-detect and optimize limited repetitions in JIT.
510    
511    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
512        blocking (*UTF) etc.
513    
514    36. In the interpreter, maximizing pattern repetitions for characters and
515        character types now use tail recursion, which reduces stack usage.
516    
517    37. The value of the max lookbehind was not correctly preserved if a compiled
518        and saved regex was reloaded on a host of different endianness.
519    
520    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
521        of the compiled pattern block, expand the flags field from 16 to 32 bits
522        because it was almost full.
523    
524    39. Try madvise first before posix_madvise.
525    
526    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
527        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
528        original change locked out all matches of empty strings. This has been
529        changed so that one match of an empty string per line is recognized.
530        Subsequent searches on the same line (for colouring or for --only-matching,
531        for example) do not recognize empty strings.
532    
533    41. Applied a user patch to fix a number of spelling mistakes in comments.
534    
535    42. Data lines longer than 65536 caused pcretest to crash.
536    
537    43. Clarified the data type for length and startoffset arguments for pcre_exec
538        and pcre_dfa_exec in the function-specific man pages, where they were
539        explicitly stated to be in bytes, never having been updated. I also added
540        some clarification to the pcreapi man page.
541    
542    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
543        a segmentation fault.
544    
545    
546    Version 8.32 30-November-2012
547    -----------------------------
548    
549    1.  Improved JIT compiler optimizations for first character search and single
550        character iterators.
551    
552    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
553        Patch by Daniel Richard G.
554    
555    3.  Single character iterator optimizations in the JIT compiler.
556    
557    4.  Improved JIT compiler optimizations for character ranges.
558    
559    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
560        Reported by Giuseppe D'Angelo.
561    
562    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
563        of a line, was being set incorrectly in cases where .* appeared inside
564        atomic brackets at the start of a pattern, or where there was a subsequent
565        *PRUNE or *SKIP.
566    
567    7.  Improved instruction cache flush for POWER/PowerPC.
568        Patch by Daniel Richard G.
569    
570    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
571        grep:
572    
573        (a) There is now no limit to the number of patterns to be matched.
574    
575        (b) An error is given if a pattern is too long.
576    
577        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
578            are now supported.
579    
580        (d) --exclude-from and --include-from (multiple use) have been added.
581    
582        (e) Exclusions and inclusions now apply to all files and directories, not
583            just to those obtained from scanning a directory recursively.
584    
585        (f) Multiple uses of -f and --file-list are now supported.
586    
587        (g) In a Windows environment, the default for -d has been changed from
588            "read" (the GNU grep default) to "skip", because otherwise the presence
589            of a directory in the file list provokes an error.
590    
591        (h) The documentation has been revised and clarified in places.
592    
593    9.  Improve the matching speed of capturing brackets.
594    
595    10. Changed the meaning of \X so that it now matches a Unicode extended
596        grapheme cluster.
597    
598    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
599        out POSIX threads when JIT support is configured.
600    
601    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
602    
603    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
604        structure could go wrong in environments where size_t is not the same size
605        as int.
606    
607    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
608        set.
609    
610    15. The EBCDIC support had decayed; later updates to the code had included
611        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
612        general tidy up of EBCDIC-related issues, and the documentation was also
613        not quite right. There is now a test that can be run on ASCII systems to
614        check some of the EBCDIC-related things (but is it not a full test).
615    
616    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
617        in a small tidy to the code.
618    
619    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
620    
621    18. If the --only-matching (-o) option in pcregrep is specified multiple
622        times, each one causes appropriate output. For example, -o1 -o2 outputs the
623        substrings matched by the 1st and 2nd capturing parentheses. A separating
624        string can be specified by --om-separator (default empty).
625    
626    19. Improving the first n character searches.
627    
628    20. Turn case lists for horizontal and vertical white space into macros so that
629        they are defined only once.
630    
631    21. This set of changes together give more compatible Unicode case-folding
632        behaviour for characters that have more than one other case when UCP
633        support is available.
634    
635        (a) The Unicode property table now has offsets into a new table of sets of
636            three or more characters that are case-equivalent. The MultiStage2.py
637            script that generates these tables (the pcre_ucd.c file) now scans
638            CaseFolding.txt instead of UnicodeData.txt for character case
639            information.
640    
641        (b) The code for adding characters or ranges of characters to a character
642            class has been abstracted into a generalized function that also handles
643            case-independence. In UTF-mode with UCP support, this uses the new data
644            to handle characters with more than one other case.
645    
646        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
647            whose other case is greater than 256 are now correctly matched
648            caselessly. Previously, the high codepoint matched the low one, but not
649            vice versa.
650    
651        (d) The processing of \h, \H, \v, and \ in character classes now makes use
652            of the new class addition function, using character lists defined as
653            macros alongside the case definitions of 20 above.
654    
655        (e) Caseless back references now work with characters that have more than
656            one other case.
657    
658        (f) General caseless matching of characters with more than one other case
659            is supported.
660    
661    22. Unicode character properties were updated from Unicode 6.2.0
662    
663    23. Improved CMake support under Windows. Patch by Daniel Richard G.
664    
665    24. Add support for 32-bit character strings, and UTF-32
666    
667    25. Major JIT compiler update (code refactoring and bugfixing).
668        Experimental Sparc 32 support is added.
669    
670    26. Applied a modified version of Daniel Richard G's patch to create
671        pcre.h.generic and config.h.generic by "make" instead of in the
672        PrepareRelease script.
673    
674    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
675        pcre_compile.c when checking for a zero character.
676    
677    28. Introducing a native interface for JIT. Through this interface, the compiled
678        machine code can be directly executed. The purpose of this interface is to
679        provide fast pattern matching, so several sanity checks are not performed.
680        However, feature tests are still performed. The new interface provides
681        1.4x speedup compared to the old one.
682    
683    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
684        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
685        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
686    
687    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
688        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
689        the "old" RFC 2279). Instead, it ended up passing a negative length to
690        pcre_exec().
691    
692    31. Add support for GCC's visibility feature to hide internal functions.
693    
694    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
695        "unknown -C option" after outputting 0 or 1.
696    
697    33. There is now support for generating a code coverage report for the test
698        suite in environments where gcc is the compiler and lcov is installed. This
699        is mainly for the benefit of the developers.
700    
701    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
702        unaddressable using valgrind annotations, allowing valgrind to detect
703        invalid memory accesses. This is mainly for the benefit of the developers.
704    
705    25. (*UTF) can now be used to start a pattern in any of the three libraries.
706    
707    26. Give configure error if --enable-cpp but no C++ compiler found.
708    
709    
710    Version 8.31 06-July-2012
711    -------------------------
712    
713    1.  Fixing a wrong JIT test case and some compiler warnings.
714    
715    2.  Removed a bashism from the RunTest script.
716    
717    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
718        to unsigned type, result still unsigned" that was given by an MS compiler
719        on encountering the code "-sizeof(xxx)".
720    
721    4.  Partial matching support is added to the JIT compiler.
722    
723    5.  Fixed several bugs concerned with partial matching of items that consist
724        of more than one character:
725    
726        (a) /^(..)\1/ did not partially match "aba" because checking references was
727            done on an "all or nothing" basis. This also applied to repeated
728            references.
729    
730        (b) \R did not give a hard partial match if \r was found at the end of the
731            subject.
732    
733        (c) \X did not give a hard partial match after matching one or more
734            characters at the end of the subject.
735    
736        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
737            a partial match for the string "\r".
738    
739        (e) When newline was set to CRLF, the metacharacter "." did not recognize
740            a partial match for a CR character at the end of the subject string.
741    
742    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
743        running pcretest, the text "(JIT)" added to the output whenever JIT is
744        actually used to run the match.
745    
746    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
747        or /S+[+] with a digit between 1 and 7.
748    
749    8.  OP_NOT now supports any UTF character not just single-byte ones.
750    
751    9.  (*MARK) control verb is now supported by the JIT compiler.
752    
753    10. The command "./RunTest list" lists the available tests without actually
754        running any of them. (Because I keep forgetting what they all are.)
755    
756    11. Add PCRE_INFO_MAXLOOKBEHIND.
757    
758    12. Applied a (slightly modified) user-supplied patch that improves performance
759        when the heap is used for recursion (compiled with --disable-stack-for-
760        recursion). Instead of malloc and free for each heap frame each time a
761        logical recursion happens, frames are retained on a chain and re-used where
762        possible. This sometimes gives as much as 30% improvement.
763    
764    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
765        call.
766    
767    14. As documented, (*COMMIT) is now confined to within a positive assertion.
768    
769    15. It is now possible to link pcretest with libedit as an alternative to
770        libreadline.
771    
772    16. (*COMMIT) control verb is now supported by the JIT compiler.
773    
774    17. The Unicode data tables have been updated to Unicode 6.1.0.
775    
776    18. Added --file-list option to pcregrep.
777    
778    19. Added binary file support to pcregrep, including the -a, --binary-files,
779        -I, and --text options.
780    
781    20. The madvise function is renamed for posix_madvise for QNX compatibility
782        reasons. Fixed by Giuseppe D'Angelo.
783    
784    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
785    
786    22. Changed the option for creating symbolic links for 16-bit man pages from
787        -s to -sf so that re-installing does not cause issues.
788    
789    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
790    
791    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
792        matches in certain environments (the workspace was not being correctly
793        retained). Also added to pcre_dfa_exec() a simple plausibility check on
794        some of the workspace data at the beginning of a restart.
795    
796    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
797        was not doing so when it should - probably a typo introduced by SVN 528
798        (change 8.10/14).
799    
800    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
801        \w+ when the character tables indicated that \x{c4} was a word character.
802        There were several related cases, all because the tests for doing a table
803        lookup were testing for characters less than 127 instead of 255.
804    
805    27. If a pattern contains capturing parentheses that are not used in a match,
806        their slots in the ovector are set to -1. For those that are higher than
807        any matched groups, this happens at the end of processing. In the case when
808        there were back references that the ovector was too small to contain
809        (causing temporary malloc'd memory to be used during matching), and the
810        highest capturing number was not used, memory off the end of the ovector
811        was incorrectly being set to -1. (It was using the size of the temporary
812        memory instead of the true size.)
813    
814    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
815        ovector size, it uses memory at the end of the block that it has got.
816    
817    29. Check for an overlong MARK name and give an error at compile time. The
818        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
819    
820    30. JIT compiler update.
821    
822    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
823        Rill for the patch.
824    
825    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
826    
827    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
828    
829    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
830        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
831        was enabled.
832    
833    35. Improve JIT code generation for greedy plus quantifier.
834    
835    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
836        1 to "aa" instead of to an empty string. The bug affected repeated groups
837        that could potentially match an empty string.
838    
839    37. Optimizing single character iterators in JIT.
840    
841    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
842        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
843        codepoints that are too big for the mode are faulted, and in a UTF mode,
844        disallowed codepoints are also faulted.
845    
846    39. If PCRE was compiled with UTF support, in three places in the DFA
847        matcher there was code that should only have been obeyed in UTF mode, but
848        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
849        processing when bytes with values greater than 127 were present. In 16-bit
850        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
851        both cases the values are those that cannot be the first data item in a UTF
852        character. The three items that might have provoked this were recursions,
853        possessively repeated groups, and atomic groups.
854    
855    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
856        and pcregrep, because some OS require shared objects to be explicitly
857        passed to ld, causing the link step to fail if they are not.
858    
859    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
860        mode, patterns that started with \h* or \R* might be incorrectly matched.
861    
862    
863    Version 8.30 04-February-2012
864    -----------------------------
865    
866    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
867        name is defined in ctype.h.
868    
869    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
870        only in quite long subpatterns.
871    
872    3.  Removed the function pcre_info(), which has been obsolete and deprecated
873        since it was replaced by pcre_fullinfo() in February 2000.
874    
875    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
876        match a (*MARK), and the match failed at the start of the subject, a
877        reference to memory before the start of the subject could occur. This bug
878        was introduced by fix 17 of release 8.21.
879    
880    5.  A reference to an unset group with zero minimum repetition was giving
881        totally wrong answers (in non-JavaScript-compatibility mode). For example,
882        /(another)?(\1?)test/ matched against "hello world test". This bug was
883        introduced in release 8.13.
884    
885    6.  Add support for 16-bit character strings (a large amount of work involving
886        many changes and refactorings).
887    
888    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
889        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
890        from a file.
891    
892    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
893        rounding is not applied in this particular case).
894    
895    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
896        if they appear, or are escaped, in patterns.
897    
898    10. Get rid of a number of -Wunused-but-set-variable warnings.
899    
900    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
901        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
902        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
903        also returns the mark "x". This bug applied to capturing parentheses,
904        non-capturing parentheses, and atomic parentheses. It also applied to some
905        assertions.
906    
907    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
908        information out of configure.ac instead of relying on pcre.h.generic, which
909        is not stored in the repository.
910    
911    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
912        -lreadline.
913    
914    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
915    
916    15. Applied Graycode's patch to put the top-level frame on the stack rather
917        than the heap when not using the stack for recursion. This gives a
918        performance improvement in many cases when recursion is not deep.
919    
920    16. Experimental code added to "pcretest -C" to output the stack frame size.
921    
922    
923    Version 8.21 12-Dec-2011
924    ------------------------
925    
926    1.  Updating the JIT compiler.
927    
928    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
929        are added as well.
930    
931    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
932        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
933        calling _pcre_jit_exec. Some extra comments are added.
934    
935    4.  (*MARK) settings inside atomic groups that do not contain any capturing
936        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
937        was introduced by change 18 for 8.20.
938    
939    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
940        ECMA-262 standard.
941    
942    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
943        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
944        This bug was probably introduced by change 9 of 8.13.
945    
946    7.  While fixing 6 above, I noticed that a number of other items were being
947        incorrectly rejected as "not fixed length". This arose partly because newer
948        opcodes had not been added to the fixed-length checking code. I have (a)
949        corrected the bug and added tests for these items, and (b) arranged for an
950        error to occur if an unknown opcode is encountered while checking for fixed
951        length instead of just assuming "not fixed length". The items that were
952        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
953        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
954        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
955    
956    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
957        being incorrectly compiled and would have given unpredicatble results.
958    
959    9.  A possessively repeated subpattern with minimum repeat count greater than
960        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
961        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
962        the first (A) could occur when it should not.
963    
964    10. Add a cast and remove a redundant test from the code.
965    
966    11. JIT should use pcre_malloc/pcre_free for allocation.
967    
968    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
969        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
970        is anything other than /usr, -L is still shown).
971    
972    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
973    
974    14. Perl does not support \N without a following name in a [] class; PCRE now
975        also gives an error.
976    
977    15. If a forward reference was repeated with an upper limit of around 2000,
978        it caused the error "internal error: overran compiling workspace". The
979        maximum number of forward references (including repeats) was limited by the
980        internal workspace, and dependent on the LINK_SIZE. The code has been
981        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
982        the default depends on LINK_SIZE. There is a new upper limit (for safety)
983        of around 200,000 forward references. While doing this, I also speeded up
984        the filling in of repeated forward references.
985    
986    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
987        incorrectly expecting the subject to contain another "a" after the start.
988    
989    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
990        in the match, the SKIP should be ignored. This was not happening; instead
991        the SKIP was being treated as NOMATCH. For patterns such as
992        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
993        tested.
994    
995    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
996        now much more compatible with Perl, in particular in cases where the result
997        is a non-match for a non-anchored pattern. For example, if
998        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
999        "m", where previously it did not return a name. A side effect of this
1000        change is that for partial matches, the last encountered mark name is
1001        returned, as for non matches. A number of tests that were previously not
1002        Perl-compatible have been moved into the Perl-compatible test files. The
1003        refactoring has had the pleasing side effect of removing one argument from
1004        the match() function, thus reducing its stack requirements.
1005    
1006    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1007        subsequent uses of /S (without +) incorrectly behaved like /S+.
1008    
1009    21. Retrieve executable code size support for the JIT compiler and fixing
1010        some warnings.
1011    
1012    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1013        not work when the shorter character appeared right at the end of the
1014        subject string.
1015    
1016    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1017        systems.
1018    
1019    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1020        output it when the /M option is used in pcretest.
1021    
1022    25. The CheckMan script was not being included in the distribution. Also, added
1023        an explicit "perl" to run Perl scripts from the PrepareRelease script
1024        because this is reportedly needed in Windows.
1025    
1026    26. If study data was being save in a file and studying had not found a set of
1027        "starts with" bytes for the pattern, the data written to the file (though
1028        never used) was taken from uninitialized memory and so caused valgrind to
1029        complain.
1030    
1031    27. Updated RunTest.bat as provided by Sheri Pierce.
1032    
1033    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1034    
1035    29. Computation of memory usage for the table of capturing group names was
1036        giving an unnecessarily large value.
1037    
1038    
1039    Version 8.20 21-Oct-2011
1040    ------------------------
1041    
1042    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1043        a POSIX class. After further experiments with Perl, which convinced me that
1044        Perl has bugs and confusions, a closing square bracket is no longer allowed
1045        in a POSIX name. This bug also affected patterns with classes that started
1046        with full stops.
1047    
1048    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1049        captured substring, but while checking the failing first alternative,
1050        substring 1 is temporarily captured. If the output vector supplied to
1051        pcre_exec() was not big enough for this capture, the yield of the function
1052        was still zero ("insufficient space for captured substrings"). This cannot
1053        be totally fixed without adding another stack variable, which seems a lot
1054        of expense for a edge case. However, I have improved the situation in cases
1055        such as /(a)(b)x|abc/ matched against "abc", where the return code
1056        indicates that fewer than the maximum number of slots in the ovector have
1057        been set.
1058    
1059    3.  Related to (2) above: when there are more back references in a pattern than
1060        slots in the output vector, pcre_exec() uses temporary memory during
1061        matching, and copies in the captures as far as possible afterwards. It was
1062        using the entire output vector, but this conflicts with the specification
1063        that only 2/3 is used for passing back captured substrings. Now it uses
1064        only the first 2/3, for compatibility. This is, of course, another edge
1065        case.
1066    
1067    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1068        main code base, and can be used by building with --enable-jit. When this is
1069        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1070        runtime --no-jit option is given.
1071    
1072    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1073        ovector, the return from the function was zero, implying that there were
1074        other matches that did not fit. The correct "exactly full" value is now
1075        returned.
1076    
1077    6.  If a subpattern that was called recursively or as a subroutine contained
1078        (*PRUNE) or any other control that caused it to give a non-standard return,
1079        invalid errors such as "Error -26 (nested recursion at the same subject
1080        position)" or even infinite loops could occur.
1081    
1082    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1083        computing the minimum length on reaching *ACCEPT, and so ended up with the
1084        wrong value of 1 rather than 0. Further investigation indicates that
1085        computing a minimum subject length in the presence of *ACCEPT is difficult
1086        (think back references, subroutine calls), and so I have changed the code
1087        so that no minimum is registered for a pattern that contains *ACCEPT.
1088    
1089    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1090        it was not handled as intended. [But see 16 below.]
1091    
1092    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1093        Sheri Pierce.
1094    
1095    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1096        the first byte in a match must be "a".
1097    
1098    11. Change 17 for 8.13 increased the recursion depth for patterns like
1099        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1100        pattern contains any instances of (*THEN). If it does not, the old
1101        optimizations are restored. It would be nice to do this on a per-group
1102        basis, but at the moment that is not feasible.
1103    
1104    12. In some environments, the output of pcretest -C is CRLF terminated. This
1105        broke RunTest's code that checks for the link size. A single white space
1106        character after the value is now allowed for.
1107    
1108    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1109        For "fr", it uses the Windows-specific input and output files.
1110    
1111    14. If (*THEN) appeared in a group that was called recursively or as a
1112        subroutine, it did not work as intended. [But see next item.]
1113    
1114    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1115        pattern fragments (but not containing any | characters). If A and B are
1116        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1117        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1118        D. In other words, Perl considers parentheses that do not contain any |
1119        characters to be part of a surrounding alternative, whereas PCRE was
1120        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1121        differently. PCRE now behaves in the same way as Perl, except in the case
1122        of subroutine/recursion calls such as (?1) which have in any case always
1123        been different (but PCRE had them first :-).
1124    
1125    16. Related to 15 above: Perl does not treat the | in a conditional group as
1126        creating alternatives. Such a group is treated in the same way as an
1127        ordinary group without any | characters when processing (*THEN). PCRE has
1128        been changed to match Perl's behaviour.
1129    
1130    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1131        RunGrepTest script failed.
1132    
1133    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1134        inevitable for groups that contain captures, but it can lead to a lot of
1135        stack use in large patterns. The old behaviour has been restored for atomic
1136        groups that do not contain any capturing parentheses.
1137    
1138    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1139        suppress the check for a minimum subject length at run time. (If it was
1140        given to pcre_exec() or pcre_dfa_exec() it did work.)
1141    
1142    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1143        fail to work when decoding hex characters in data strings in EBCDIC
1144        environments.
1145    
1146    21. It appears that in at least one Mac OS environment, the isxdigit() function
1147        is implemented as a macro that evaluates to its argument more than once,
1148        contravening the C 90 Standard (I haven't checked a later standard). There
1149        was an instance in pcretest which caused it to go wrong when processing
1150        \x{...} escapes in subject strings. The has been rewritten to avoid using
1151        things like p++ in the argument of isxdigit().
1152    
1153    
1154    Version 8.13 16-Aug-2011
1155    ------------------------
1156    
1157    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1158    
1159    2.  Two minor typos in pcre_internal.h have been fixed.
1160    
1161    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1162        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1163        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1164    
1165    4.  There were a number of related bugs in the code for matching backrefences
1166        caselessly in UTF-8 mode when codes for the characters concerned were
1167        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1168        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1169        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1170        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1171        2-byte code at the end of the subject (it thought there wasn't enough data
1172        left).
1173    
1174    5.  Comprehensive information about what went wrong is now returned by
1175        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1176        as the output vector has at least 2 elements. The offset of the start of
1177        the failing character and a reason code are placed in the vector.
1178    
1179    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1180        now returned is for the first byte of the failing character, instead of the
1181        last byte inspected. This is an incompatible change, but I hope it is small
1182        enough not to be a problem. It makes the returned offset consistent with
1183        pcre_exec() and pcre_dfa_exec().
1184    
1185    7.  pcretest now gives a text phrase as well as the error number when
1186        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1187        failure, the offset and reason code are output.
1188    
1189    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1190        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1191        back over a single character (\n). This seems wrong (because it treated the
1192        two characters as a single entity when going forwards), conflicts with the
1193        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1194        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1195        has been changed.
1196    
1197    9.  Some internal refactoring has changed the processing so that the handling
1198        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1199        time (the PCRE_DOTALL option was changed this way some time ago: version
1200        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1201        which was always a bit of a fudge. It also means that there is one less
1202        argument for the match() function, which reduces its stack requirements
1203        slightly. This change also fixes an incompatibility with Perl: the pattern
1204        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1205    
1206    10. More internal refactoring has drastically reduced the number of recursive
1207        calls to match() for possessively repeated groups such as (abc)++ when
1208        using pcre_exec().
1209    
1210    11. While implementing 10, a number of bugs in the handling of groups were
1211        discovered and fixed:
1212    
1213        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1214        (a|)*(?1) gave a compile-time internal error.
1215        ((a|)+)+  did not notice that the outer group could match an empty string.
1216        (^a|^)+   was not marked as anchored.
1217        (.*a|.*)+ was not marked as matching at start or after a newline.
1218    
1219    12. Yet more internal refactoring has removed another argument from the match()
1220        function. Special calls to this function are now indicated by setting a
1221        value in a variable in the "match data" data block.
1222    
1223    13. Be more explicit in pcre_study() instead of relying on "default" for
1224        opcodes that mean there is no starting character; this means that when new
1225        ones are added and accidentally left out of pcre_study(), testing should
1226        pick them up.
1227    
1228    14. The -s option of pcretest has been documented for ages as being an old
1229        synonym of -m (show memory usage). I have changed it to mean "force study
1230        for every regex", that is, assume /S for every regex. This is similar to -i
1231        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1232        using it. It makes it easier to run collections of tests with and without
1233        study enabled, and thereby test pcre_study() more easily. All the standard
1234        tests are now run with and without -s (but some patterns can be marked as
1235        "never study" - see 20 below).
1236    
1237    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1238        restoration of the capturing data to the outer values was not happening
1239        correctly.
1240    
1241    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1242        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1243        pattern had matched an empty string, and so incorrectly returned a no
1244        match.
1245    
1246    17. There was optimizing code for the last branch of non-capturing parentheses,
1247        and also for the obeyed branch of a conditional subexpression, which used
1248        tail recursion to cut down on stack usage. Unfortunately, now that there is
1249        the possibility of (*THEN) occurring in these branches, tail recursion is
1250        no longer possible because the return has to be checked for (*THEN). These
1251        two optimizations have therefore been removed. [But see 8.20/11 above.]
1252    
1253    18. If a pattern containing \R was studied, it was assumed that \R always
1254        matched two bytes, thus causing the minimum subject length to be
1255        incorrectly computed because \R can also match just one byte.
1256    
1257    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1258        was incorrectly computed.
1259    
1260    20. If /S is present twice on a test pattern in pcretest input, it now
1261        *disables* studying, thereby overriding the use of -s on the command line
1262        (see 14 above). This is necessary for one or two tests to keep the output
1263        identical in both cases.
1264    
1265    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1266        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1267    
1268    22. When an atomic group that contained a capturing parenthesis was
1269        successfully matched, but the branch in which it appeared failed, the
1270        capturing was not being forgotten if a higher numbered group was later
1271        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1272        group 1 to "a", when in fact it should be unset. This applied to multi-
1273        branched capturing and non-capturing groups, repeated or not, and also to
1274        positive assertions (capturing in negative assertions does not happen
1275        in PCRE) and also to nested atomic groups.
1276    
1277    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1278        subject after a captured substring, to make it easier to tell which of a
1279        number of identical substrings has been captured.
1280    
1281    24. The way atomic groups are processed by pcre_exec() has been changed so that
1282        if they are repeated, backtracking one repetition now resets captured
1283        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1284        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1285        "aaa". Previously, it would have been "a". As part of this code
1286        refactoring, the way recursive calls are handled has also been changed.
1287    
1288    25. If an assertion condition captured any substrings, they were not passed
1289        back unless some other capturing happened later. For example, if
1290        (?(?=(a))a) was matched against "a", no capturing was returned.
1291    
1292    26. When studying a pattern that contained subroutine calls or assertions,
1293        the code for finding the minimum length of a possible match was handling
1294        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1295        group 1 called group 2 while simultaneously a separate group 2 called group
1296        1). A stack overflow occurred in this case. I have fixed this by limiting
1297        the recursion depth to 10.
1298    
1299    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1300        Fortmann. This supports explicit test numbers on the command line, and has
1301        argument validation and error reporting.
1302    
1303    28. An instance of \X with an unlimited repeat could fail if at any point the
1304        first character it looked at was a mark character.
1305    
1306    29. Some minor code refactoring concerning Unicode properties and scripts
1307        should reduce the stack requirement of match() slightly.
1308    
1309    30. Added the '=' option to pcretest to check the setting of unused capturing
1310        slots at the end of the pattern, which are documented as being -1, but are
1311        not included in the return count.
1312    
1313    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1314        compiled something random. Now it gives a compile-time error (as does
1315        Perl).
1316    
1317    32. A *MARK encountered during the processing of a positive assertion is now
1318        recorded and passed back (compatible with Perl).
1319    
1320    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1321        had alternative anchored branches, the search for a second match in a line
1322        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1323        matched the line "0102" twice. The same bug affected patterns that started
1324        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1325        twice.
1326    
1327    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1328        does, and because of capturing effects, quantifying parenthesized
1329        assertions may at times be useful. Quantifiers are now allowed for
1330        parenthesized assertions.
1331    
1332    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1333    
1334    36. \g was being checked for fancy things in a character class, when it should
1335        just be a literal "g".
1336    
1337    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1338        appearance of a nested POSIX class supersedes an apparent external class.
1339        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1340        unescaped square brackets may also appear as part of class names. For
1341        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1342        more like Perl. (But see 8.20/1 above.)
1343    
1344    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1345        was because it thought it was \N{name}, which is not supported).
1346    
1347    39. Add minix to OS list not supporting the -S option in pcretest.
1348    
1349    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1350        cannot analyze patterns in sufficient detail to catch mutual recursions
1351        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1352        subgroup is called recursively as a subpattern for a second time at the
1353        same position in the subject string. In previous releases this might have
1354        been caught by the recursion limit, or it might have run out of stack.
1355    
1356    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1357        happen only once. PCRE was, however incorrectly giving a compile time error
1358        "recursive call could loop indefinitely" because it cannot analyze the
1359        pattern in sufficient detail. The compile time test no longer happens when
1360        PCRE is compiling a conditional subpattern, but actual runaway loops are
1361        now caught at runtime (see 40 above).
1362    
1363    42. It seems that Perl allows any characters other than a closing parenthesis
1364        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1365        has been changed to be the same.
1366    
1367    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1368        as not to get warnings when autogen.sh is called. Also changed
1369        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1370    
1371    44. To help people who use pcregrep to scan files containing exceedingly long
1372        lines, the following changes have been made:
1373    
1374        (a) The default value of the buffer size parameter has been increased from
1375            8K to 20K. (The actual buffer used is three times this size.)
1376    
1377        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1378            PCRE is built.
1379    
1380        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1381            to be set at run time.
1382    
1383        (d) Numerical values in pcregrep options can be followed by K or M, for
1384            example --buffer-size=50K.
1385    
1386        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1387            given and the return code is set to 2.
1388    
1389    45. Add a pointer to the latest mark to the callout data block.
1390    
1391    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1392        partial match of an empty string instead of no match. This was specific to
1393        the use of ".".
1394    
1395    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1396        complete match instead of a partial match. This bug was dependent on both
1397        the PCRE_UTF8 and PCRE_DOTALL options being set.
1398    
1399    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1400        starting byte set, because \b was not being ignored.
1401    
1402    
1403    Version 8.12 15-Jan-2011
1404    ------------------------
1405    
1406    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1407        checks for such things as part of the documentation building process.
1408    
1409    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1410        --match-limit and --recursion-limit options (added for 8.11). In
1411        particular, this made one of the standard tests fail. (The integer value
1412        went into the wrong half of a long int.)
1413    
1414    3.  If the --colour option was given to pcregrep with -v (invert match), it
1415        did strange things, either producing crazy output, or crashing. It should,
1416        of course, ignore a request for colour when reporting lines that do not
1417        match.
1418    
1419    4.  Another pcregrep bug caused similar problems if --colour was specified with
1420        -M (multiline) and the pattern match finished with a line ending.
1421    
1422    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1423        matched in multiline mode, the following line was shown as part of the
1424        match. This seems wrong, so I have changed it.
1425    
1426    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1427        the check for further matches in the same line (so they could be coloured)
1428        to overrun the end of the current line. If another match was found, it was
1429        incorrectly shown (and then shown again when found in the next line).
1430    
1431    7.  If pcregrep was compiled under Windows, there was a reference to the
1432        function pcregrep_exit() before it was defined. I am assuming this was
1433        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1434        reported by a user. I've moved the definition above the reference.
1435    
1436    
1437    Version 8.11 10-Dec-2010
1438    ------------------------
1439    
1440    1.  (*THEN) was not working properly if there were untried alternatives prior
1441        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1442        backtracked to try for "b" instead of moving to the next alternative branch
1443        at the same level (in this case, to look for "c"). The Perl documentation
1444        is clear that when (*THEN) is backtracked onto, it goes to the "next
1445        alternative in the innermost enclosing group".
1446    
1447    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1448        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1449        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1450        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1451        (*THEN).
1452    
1453    3.  If \s appeared in a character class, it removed the VT character from
1454        the class, even if it had been included by some previous item, for example
1455        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1456        of \s, but is part of the POSIX "space" class.)
1457    
1458    4.  A partial match never returns an empty string (because you can always
1459        match an empty string at the end of the subject); however the checking for
1460        an empty string was starting at the "start of match" point. This has been
1461        changed to the "earliest inspected character" point, because the returned
1462        data for a partial match starts at this character. This means that, for
1463        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1464        (previously it gave "no match").
1465    
1466    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1467        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1468        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1469        has an implication that the given string is incomplete (because a partial
1470        match is preferred over a full match). For this reason, these items now
1471        give a partial match in this situation. [Aside: previously, the one case
1472        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1473        match rather than a full match, which was wrong by the old rules, but is
1474        now correct.]
1475    
1476    6.  There was a bug in the handling of #-introduced comments, recognized when
1477        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1478        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1479        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1480        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1481        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1482        places in pcre_compile().
1483    
1484    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1485        comments when looking ahead for named forward references to subpatterns,
1486        the only newline sequence it recognized was NL. It now handles newlines
1487        according to the set newline convention.
1488    
1489    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1490        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1491        cater for a lack of strerror(). These oversights have been fixed.
1492    
1493    9.  Added --match-limit and --recursion-limit to pcregrep.
1494    
1495    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1496    
1497    11. When the -o option was used, pcregrep was setting a return code of 1, even
1498        when matches were found, and --line-buffered was not being honoured.
1499    
1500    12. Added an optional parentheses number to the -o and --only-matching options
1501        of pcregrep.
1502    
1503    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1504        can match an empty string. The code to do it in pcretest and pcredemo
1505        needed fixing:
1506    
1507        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1508            only one byte after an empty string match just before CRLF (this case
1509            just got forgotten; "any" and "anycrlf" were OK).
1510    
1511        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1512            mode when an empty string match preceded an ASCII character followed by
1513            a non-ASCII character. (The code for advancing by one character rather
1514            than one byte was nonsense.)
1515    
1516        (c) The pcredemo.c sample program did not have any code at all to handle
1517            the cases when CRLF is a valid newline sequence.
1518    
1519    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1520        as a starting offset was within the subject string. There is now a new
1521        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1522        negative or greater than the length of the string. In order to test this,
1523        pcretest is extended to allow the setting of negative starting offsets.
1524    
1525    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1526        starting offset points to the beginning of a UTF-8 character was
1527        unnecessarily clumsy. I tidied it up.
1528    
1529    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1530        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1531    
1532    17. Nobody had reported that the --include_dir option, which was added in
1533        release 7.7 should have been called --include-dir (hyphen, not underscore)
1534        for compatibility with GNU grep. I have changed it to --include-dir, but
1535        left --include_dir as an undocumented synonym, and the same for
1536        --exclude-dir, though that is not available in GNU grep, at least as of
1537        release 2.5.4.
1538    
1539    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1540        characters from a string of bytes) have been redefined so as not to use
1541        loops, in order to improve performance in some environments. At the same
1542        time, I abstracted some of the common code into auxiliary macros to save
1543        repetition (this should not affect the compiled code).
1544    
1545    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1546        compile-time error is now given if \c is not followed by an ASCII
1547        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1548        different, and any byte value is allowed.)
1549    
1550    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1551        START_OPTIMIZE option, which is now allowed at compile time - but just
1552        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1553        to pcregrep and other applications that have no direct access to PCRE
1554        options. The new /Y option in pcretest sets this option when calling
1555        pcre_compile().
1556    
1557    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1558        back references. Groups containing recursive back references were forced to
1559        be atomic by that change, but in the case of named groups, the amount of
1560        memory required was incorrectly computed, leading to "Failed: internal
1561        error: code overflow". This has been fixed.
1562    
1563    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1564        pcretest.c, to avoid build problems in some Borland environments.
1565    
1566    
1567    Version 8.10 25-Jun-2010
1568    ------------------------
1569    
1570    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1571        THEN.
1572    
1573    2.  (*ACCEPT) was not working when inside an atomic group.
1574    
1575    3.  Inside a character class, \B is treated as a literal by default, but
1576        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1577        causes the error). The code is unchanged, but I tidied the documentation.
1578    
1579    4.  Inside a character class, PCRE always treated \R and \X as literals,
1580        whereas Perl faults them if its -w option is set. I have changed PCRE so
1581        that it faults them when PCRE_EXTRA is set.
1582    
1583    5.  Added support for \N, which always matches any character other than
1584        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1585    
1586    6.  When compiling pcregrep with newer versions of gcc which may have
1587        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1588        declared with attribute warn_unused_result" were given. Just casting the
1589        result to (void) does not stop the warnings; a more elaborate fudge is
1590        needed. I've used a macro to implement this.
1591    
1592    7.  Minor change to pcretest.c to avoid a compiler warning.
1593    
1594    8.  Added four artifical Unicode properties to help with an option to make
1595        \s etc use properties (see next item). The new properties are: Xan
1596        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1597    
1598    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1599        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1600        this option. Modified pcretest to add /W to test this facility. Added
1601        REG_UCP to make it available via the POSIX interface.
1602    
1603    10. Added --line-buffered to pcregrep.
1604    
1605    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1606        studied, and the match started with a letter with a code point greater than
1607        127 whose first byte was different to the first byte of the other case of
1608        the letter, the other case of this starting letter was not recognized
1609        (#976).
1610    
1611    12. If a pattern that was studied started with a repeated Unicode property
1612        test, for example, \p{Nd}+, there was the theoretical possibility of
1613        setting up an incorrect bitmap of starting bytes, but fortunately it could
1614        not have actually happened in practice until change 8 above was made (it
1615        added property types that matched character-matching opcodes).
1616    
1617    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1618        possible starting bytes for non-anchored patterns.
1619    
1620    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1621        \R, and also a number of cases that involve Unicode properties, both
1622        explicit and implicit when PCRE_UCP is set.
1623    
1624    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1625        input, it could crash or give wrong results if characters with values
1626        greater than 0xc0 were present in the subject string. (Detail: it assumed
1627        UTF-8 input when processing these items.)
1628    
1629    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1630        size_t is 64-bit (#991).
1631    
1632    17. Added a check for running out of memory when PCRE is compiled with
1633        --disable-stack-for-recursion (#990).
1634    
1635    18. If the last data line in a file for pcretest does not have a newline on
1636        the end, a newline was missing in the output.
1637    
1638    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1639        less than 128) in its various bitmaps. However, there is a facility for
1640        generating tables according to the current locale when PCRE is compiled. It
1641        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1642        characters, are recognized by isspace() and therefore were getting set in
1643        these tables, and indeed these tables seem to approximate to ISO 8859. This
1644        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1645        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1646        which of course cannot start UTF-8 characters. I have changed the code so
1647        that only real ASCII characters (less than 128) and the correct starting
1648        bytes for UTF-8 encodings are set for characters greater than 127 when in
1649        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1650        altogether.)
1651    
1652    20. Added the /T option to pcretest so as to be able to run tests with non-
1653        standard character tables, thus making it possible to include the tests
1654        used for 19 above in the standard set of tests.
1655    
1656    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1657        reference to a subpattern the other side of a comment that contains an
1658        opening parenthesis caused either an internal compiling error, or a
1659        reference to the wrong subpattern.
1660    
1661    
1662    Version 8.02 19-Mar-2010
1663    ------------------------
1664    
1665    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1666    
1667    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1668        configured.
1669    
1670    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1671        original author of that file, following a query about its status.
1672    
1673    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1674        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1675    
1676    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1677        quantifier applied to a forward-referencing subroutine call, could compile
1678        incorrect code or give the error "internal error: previously-checked
1679        referenced subpattern not found".
1680    
1681    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1682        variables to point to external functions. For these systems, therefore,
1683        pcre_malloc etc. are now initialized to local functions that call the
1684        relevant global functions.
1685    
1686    7.  There were two entries missing in the vectors called coptable and poptable
1687        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1688        I've fixed the data, and added a kludgy way of testing at compile time that
1689        the lengths are correct (equal to the number of opcodes).
1690    
1691    8.  Following on from 7, I added a similar kludge to check the length of the
1692        eint vector in pcreposix.c.
1693    
1694    9.  Error texts for pcre_compile() are held as one long string to avoid too
1695        much relocation at load time. To find a text, the string is searched,
1696        counting zeros. There was no check for running off the end of the string,
1697        which could happen if a new error number was added without updating the
1698        string.
1699    
1700    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1701    
1702    11. \K was not working if it appeared in an atomic group or in a group that
1703        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1704        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1705        the assertion is positive, but not if it is negative.
1706    
1707    12. Change 11 fortuitously reduced the size of the stack frame used in the
1708        "match()" function of pcre_exec.c by one pointer. Forthcoming
1709        implementation of support for (*MARK) will need an extra pointer on the
1710        stack; I have reserved it now, so that the stack frame size does not
1711        decrease.
1712    
1713    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1714        item in branch that calls a recursion is a subroutine call - as in the
1715        second branch in the above example - was incorrectly given the compile-
1716        time error "recursive call could loop indefinitely" because pcre_compile()
1717        was not correctly checking the subroutine for matching a non-empty string.
1718    
1719    14. The checks for overrunning compiling workspace could trigger after an
1720        overrun had occurred. This is a "should never occur" error, but it can be
1721        triggered by pathological patterns such as hundreds of nested parentheses.
1722        The checks now trigger 100 bytes before the end of the workspace.
1723    
1724    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1725    
1726    
1727    Version 8.01 19-Jan-2010
1728    ------------------------
1729    
1730    1.  If a pattern contained a conditional subpattern with only one branch (in
1731        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1732        computed the wrong minimum data length (which is of course zero for such
1733        subpatterns). This could cause incorrect "no match" results.
1734    
1735    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1736        the pattern is reset in the first branch, pcre_compile() failed with
1737        "internal error: code overflow at offset...". This happened only when
1738        the reset was to the original external option setting. (An optimization
1739        abstracts leading options settings into an external setting, which was the
1740        cause of this.)
1741    
1742    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1743        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1744        assertion pattern did not match (meaning that the assertion was true), it
1745        was incorrectly treated as false if the SKIP had been reached during the
1746        matching. This also applied to assertions used as conditions.
1747    
1748    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1749        assertion subpattern, including such a pattern used as a condition,
1750        unpredictable results occurred, instead of the error return
1751        PCRE_ERROR_DFA_UITEM.
1752    
1753    5.  The C++ GlobalReplace function was not working like Perl for the special
1754        situation when an empty string is matched. It now does the fancy magic
1755        stuff that is necessary.
1756    
1757    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1758        removed. (These were left over from very, very early versions of PCRE.)
1759    
1760    7.  Some cosmetic changes to the code to make life easier when compiling it
1761        as part of something else:
1762    
1763        (a) Change DEBUG to PCRE_DEBUG.
1764    
1765        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1766            called "current" as "current_branch", to prevent a collision with the
1767            Linux macro when compiled as a kernel module.
1768    
1769        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1770            prevent a collision with the Linux macro when compiled as a kernel
1771            module.
1772    
1773    8.  In pcre_compile() there are some checks for integer overflows that used to
1774        cast potentially large values to (double). This has been changed to that
1775        when building, a check for int64_t is made, and if it is found, it is used
1776        instead, thus avoiding the use of floating point arithmetic. (There is no
1777        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1778        double.
1779    
1780    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1781        2005 (difference between two addresses compared to an unsigned value).
1782    
1783    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1784        custom one, because of the following reported problem in Windows:
1785    
1786          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1787              under Win32.
1788          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1789              therefore missing the function definition.
1790          - The compiler thus generates a "C" signature for the test function.
1791          - The linker fails to find the "C" function.
1792          - PCRE fails to configure if asked to do so against libbz2.
1793    
1794    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1795        messages were output:
1796    
1797          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1798          rerunning libtoolize, to keep the correct libtool macros in-tree.
1799          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1800    
1801        I have done both of these things.
1802    
1803    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1804        most of the time, it *can* run out if it is given a pattern that contains a
1805        runaway infinite recursion. I updated the discussion in the pcrestack man
1806        page.
1807    
1808    13. Now that we have gone to the x.xx style of version numbers, the minor
1809        version may start with zero. Using 08 or 09 is a bad idea because users
1810        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1811        interpreted as invalid octal numbers. I've updated the previous comment in
1812        configure.ac, and also added a check that gives an error if 08 or 09 are
1813        used.
1814    
1815    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1816        causing partial matching to fail when the end of the subject matched \W
1817        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1818    
1819    15. There were some discrepancies between the declarations in pcre_internal.h
1820        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1821        their definitions. The declarations used "const uschar *" and the
1822        definitions used USPTR. Even though USPTR is normally defined as "const
1823        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1824        reported that: "This difference in casting confuses some C++ compilers, for
1825        example, SunCC recognizes above declarations as different functions and
1826        generates broken code for hbpcre." I have changed the declarations to use
1827        USPTR.
1828    
1829    16. GNU libtool is named differently on some systems. The autogen.sh script now
1830        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1831        (FreeBSD).
1832    
1833    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1834        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1835        comment: "Figure out how to create a longlong from a string: strtoll and
1836        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1837        instance, but it only takes 2 args instead of 3!"
1838    
1839    18. A subtle bug concerned with back references has been fixed by a change of
1840        specification, with a corresponding code fix. A pattern such as
1841        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1842        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1843        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1844        same bug. Such groups have to be quantified to be useful, or contained
1845        inside another quantified group. (If there's no repetition, the reference
1846        can never match.) The problem arises because, having left the group and
1847        moved on to the rest of the pattern, a later failure that backtracks into
1848        the group uses the captured value from the final iteration of the group
1849        rather than the correct earlier one. I have fixed this in PCRE by forcing
1850        any group that contains a reference to itself to be an atomic group; that
1851        is, there cannot be any backtracking into it once it has completed. This is
1852        similar to recursive and subroutine calls.
1853    
1854    
1855    Version 8.00 19-Oct-09
1856    ----------------------
1857    
1858    1.  The table for translating pcre_compile() error codes into POSIX error codes
1859        was out-of-date, and there was no check on the pcre_compile() error code
1860        being within the table. This could lead to an OK return being given in
1861        error.
1862    
1863    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1864        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1865        in a Windows environment.
1866    
1867    3.  The pcregrep --count option prints the count for each file even when it is
1868        zero, as does GNU grep. However, pcregrep was also printing all files when
1869        --files-with-matches was added. Now, when both options are given, it prints
1870        counts only for those files that have at least one match. (GNU grep just
1871        prints the file name in this circumstance, but including the count seems
1872        more useful - otherwise, why use --count?) Also ensured that the
1873        combination -clh just lists non-zero counts, with no names.
1874    
1875    4.  The long form of the pcregrep -F option was incorrectly implemented as
1876        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1877        but it seems right to fix it, and I didn't think it was worth preserving
1878        the old behaviour.
1879    
1880    5.  The command line items --regex=pattern and --regexp=pattern were not
1881        recognized by pcregrep, which required --regex pattern or --regexp pattern
1882        (with a space rather than an '='). The man page documented the '=' forms,
1883        which are compatible with GNU grep; these now work.
1884    
1885    6.  No libpcreposix.pc file was created for pkg-config; there was just
1886        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1887    
1888    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1889        when UCP support is not needed, by modifying the Python script that
1890        generates it from Unicode data files. This should not matter if the module
1891        is correctly used as a library, but I received one complaint about 50K of
1892        unwanted data. My guess is that the person linked everything into his
1893        program rather than using a library. Anyway, it does no harm.
1894    
1895    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1896        was a minimum greater than 1 for a wide character in a possessive
1897        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1898        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1899        character. Chaos in the form of incorrect output or a compiling loop could
1900        result.
1901    
1902    9.  The restrictions on what a pattern can contain when partial matching is
1903        requested for pcre_exec() have been removed. All patterns can now be
1904        partially matched by this function. In addition, if there are at least two
1905        slots in the offset vector, the offset of the earliest inspected character
1906        for the match and the offset of the end of the subject are set in them when
1907        PCRE_ERROR_PARTIAL is returned.
1908    
1909    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1910        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1911        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1912        and may be more useful for multi-segment matching.
1913    
1914    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1915        used to be given if ever the end of the subject was reached; now it is
1916        given only if matching could not proceed because another character was
1917        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1918        string "Z", which now yields "no match" instead of "partial match". In the
1919        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1920        final character ended with (*FAIL).
1921    
1922    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1923        if the pattern had a "must contain" character that was already found in the
1924        earlier partial match, unless partial matching was again requested. For
1925        example, with the pattern /dog.(body)?/, the "must contain" character is
1926        "g". If the first part-match was for the string "dog", restarting with
1927        "sbody" failed. This bug has been fixed.
1928    
1929    13. The string returned by pcre_dfa_exec() after a partial match has been
1930        changed so that it starts at the first inspected character rather than the
1931        first character of the match. This makes a difference only if the pattern
1932        starts with a lookbehind assertion or \b or \B (\K is not supported by
1933        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1934        matching functions compatible, and I think it's the right thing to do.
1935    
1936    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1937        so that the demonstration program is easily available in environments where
1938        PCRE has not been installed from source.
1939    
1940    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1941        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1942        library.
1943    
1944    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1945        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1946        is not the first non-POSIX option to be added. Clearly some people find
1947        these options useful.
1948    
1949    17. If a caller to the POSIX matching function regexec() passes a non-zero
1950        value for nmatch with a NULL value for pmatch, the value of
1951        nmatch is forced to zero.
1952    
1953    18. RunGrepTest did not have a test for the availability of the -u option of
1954        the diff command, as RunTest does. It now checks in the same way as
1955        RunTest, and also checks for the -b option.
1956    
1957    19. If an odd number of negated classes containing just a single character
1958        interposed, within parentheses, between a forward reference to a named
1959        subpattern and the definition of the subpattern, compilation crashed with
1960        an internal error, complaining that it could not find the referenced
1961        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1962        [The bug was that it was starting one character too far in when skipping
1963        over the character class, thus treating the ] as data rather than
1964        terminating the class. This meant it could skip too much.]
1965    
1966    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1967        /g option in pcretest when the pattern contains \K, which makes it possible
1968        to have an empty string match not at the start, even when the pattern is
1969        anchored. Updated pcretest and pcredemo to use this option.
1970    
1971    21. If the maximum number of capturing subpatterns in a recursion was greater
1972        than the maximum at the outer level, the higher number was returned, but
1973        with unset values at the outer level. The correct (outer level) value is
1974        now given.
1975    
1976    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1977        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1978        make it do so. The string so far is captured, making this feature
1979        compatible with Perl.
1980    
1981    23. The tests have been re-organized, adding tests 11 and 12, to make it
1982        possible to check the Perl 5.10 features against Perl 5.10.
1983    
1984    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1985        pattern matches a fixed length string. PCRE did not allow this; now it
1986        does. Neither allows recursion.
1987    
1988    25. I finally figured out how to implement a request to provide the minimum
1989        length of subject string that was needed in order to match a given pattern.
1990        (It was back references and recursion that I had previously got hung up
1991        on.) This code has now been added to pcre_study(); it finds a lower bound
1992        to the length of subject needed. It is not necessarily the greatest lower
1993        bound, but using it to avoid searching strings that are too short does give
1994        some useful speed-ups. The value is available to calling programs via
1995        pcre_fullinfo().
1996    
1997    26. While implementing 25, I discovered to my embarrassment that pcretest had
1998        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1999        study optimizations had never been tested with that matching function.
2000        Oops. What is worse, even when it was passed study data, there was a bug in
2001        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2002        were also very few tests of studied patterns with pcre_dfa_exec().
2003    
2004    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2005        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2006        on the other side of the coin, they are no longer allowed to have different
2007        names, because these cannot be distinguished in PCRE, and this has caused
2008        confusion. (This is a difference from Perl.)
2009    
2010    28. When duplicate subpattern names are present (necessarily with different
2011        numbers, as required by 27 above), and a test is made by name in a
2012        conditional pattern, either for a subpattern having been matched, or for
2013        recursion in such a pattern, all the associated numbered subpatterns are
2014        tested, and the overall condition is true if the condition is true for any
2015        one of them. This is the way Perl works, and is also more like the way
2016        testing by number works.
2017    
2018    
2019    Version 7.9 11-Apr-09
2020    ---------------------
2021    
2022    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2023        (pcretest), all targets were linked against these libraries. This included
2024        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2025        libraries. This caused unwanted dependencies to be created. This problem
2026        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2027        pcretest is linked with readline.
2028    
2029    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2030        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2031        moved outside it again, because FALSE and TRUE are already defined in AIX,
2032        but BOOL is not.
2033    
2034    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2035        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2036    
2037    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2038        hyphens) following file names and line numbers when outputting matching
2039        lines. This is not true; no spaces are inserted. I have also clarified the
2040        wording for the --colour (or --color) option.
2041    
2042    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2043        was not coloured; this is different to GNU grep, so I have changed it to be
2044        the same.
2045    
2046    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2047        each matching line was coloured. Now it goes on to look for further matches
2048        of any of the test patterns, which is the same behaviour as GNU grep.
2049    
2050    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2051        doesn't make sense to accept an empty string match in pcregrep, so I have
2052        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2053        seems to be how GNU grep behaves. [But see later change 40 for release
2054        8.33.]
2055    
2056    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2057        start or after a newline", because the conditional assertion was not being
2058        correctly handled. The rule now is that both the assertion and what follows
2059        in the first alternative must satisfy the test.
2060    
2061    9.  If auto-callout was enabled in a pattern with a conditional group whose
2062        condition was an assertion, PCRE could crash during matching, both with
2063        pcre_exec() and pcre_dfa_exec().
2064    
2065    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2066        used for matching.
2067    
2068    11. Unicode property support in character classes was not working for
2069        characters (bytes) greater than 127 when not in UTF-8 mode.
2070    
2071    12. Added the -M command line option to pcretest.
2072    
2073    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2074    
2075    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2076    
2077    16. Added comments and documentation about mis-use of no_arg in the C++
2078        wrapper.
2079    
2080    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2081        from Martin Jerabek that uses macro names for all relevant character and
2082        string constants.
2083    
2084    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2085        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2086        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2087        these, but not everybody uses configure.
2088    
2089    19. A conditional group that had only one branch was not being correctly
2090        recognized as an item that could match an empty string. This meant that an
2091        enclosing group might also not be so recognized, causing infinite looping
2092        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2093        with the subject "ab", where knowledge that the repeated group can match
2094        nothing is needed in order to break the loop.
2095    
2096    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2097        exec(), but without supplying a callout function, matching went wrong.
2098    
2099    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2100        leak if the size of the offset vector was greater than 30. When the vector
2101        is smaller, the saved offsets during recursion go onto a local stack
2102        vector, but for larger vectors malloc() is used. It was failing to free
2103        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2104        error, in fact).
2105    
2106    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2107        heapframe that is used only when UTF-8 support is enabled. This caused no
2108        problem, but was untidy.
2109    
2110    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2111        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2112        included within another project.
2113    
2114    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2115        slightly modified by me:
2116    
2117          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2118              not building pcregrep.
2119    
2120          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2121              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2122    
2123    25. Forward references, both numeric and by name, in patterns that made use of
2124        duplicate group numbers, could behave incorrectly or give incorrect errors,
2125        because when scanning forward to find the reference group, PCRE was not
2126        taking into account the duplicate group numbers. A pattern such as
2127        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2128    
2129    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2130        the feature of a custom pointer more persuasive (as requested by a user).
2131    
2132    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2133        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2134        pre-defined.
2135    
2136    28. Added support for (*UTF8) at the start of a pattern.
2137    
2138    29. Arrange for flags added by the "release type" setting in CMake to be shown
2139        in the configuration summary.
2140    
2141    
2142    Version 7.8 05-Sep-08
2143    ---------------------
2144    
2145    1.  Replaced UCP searching code with optimized version as implemented for Ad
2146        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2147        stage table and inline lookup instead of a function, giving speed ups of 2
2148        to 5 times on some simple patterns that I tested. Permission was given to
2149        distribute the MultiStage2.py script that generates the tables (it's not in
2150        the tarball, but is in the Subversion repository).
2151    
2152    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2153        scripts.
2154    
2155    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2156        a group with a zero qualifier. The result of the study could be incorrect,
2157        or the function might crash, depending on the pattern.
2158    
2159    4.  Caseless matching was not working for non-ASCII characters in back
2160        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2161        It now works when Unicode Property Support is available.
2162    
2163    5.  In pcretest, an escape such as \x{de} in the data was always generating
2164        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2165        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2166        truncation.
2167    
2168    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2169    
2170    7.  Added two (int) casts to pcregrep when printing the difference of two
2171        pointers, in case they are 64-bit values.
2172    
2173    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2174        test 2 if it fails.
2175    
2176    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2177        and a #define of that name to empty if it is not externally set. This is to
2178        allow users of MSVC to set it if necessary.
2179    
2180    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2181        the convenience functions in the pcre_get.c source file.
2182    
2183    11. An option change at the start of a pattern that had top-level alternatives
2184        could cause overwriting and/or a crash. This command provoked a crash in
2185        some environments:
2186    
2187          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2188    
2189        This potential security problem was recorded as CVE-2008-2371.
2190    
2191    12. For a pattern where the match had to start at the beginning or immediately
2192        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2193        pcre_dfa_exec() could read past the end of the passed subject if there was
2194        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2195        pcretest so that it places the subject at the end of its malloc-ed buffer.
2196    
2197    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2198        exec() might read past the end of the data buffer in UTF-8 mode.
2199    
2200    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2201        the data contained the byte 0x85 as part of a UTF-8 character within its
2202        first line. This applied both to normal and DFA matching.
2203    
2204    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2205        /^[^d]*?$/8 failed to match "abc".
2206    
2207    16. Added a missing copyright notice to pcrecpp_internal.h.
2208    
2209    17. Make it more clear in the documentation that values returned from
2210        pcre_exec() in ovector are byte offsets, not character counts.
2211    
2212    18. Tidied a few places to stop certain compilers from issuing warnings.
2213    
2214    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2215        supplied by Stefan Weber. I made a further small update for 7.8 because
2216        there is a change of source arrangements: the pcre_searchfuncs.c module is
2217        replaced by pcre_ucd.c.
2218    
2219    
2220    Version 7.7 07-May-08
2221    ---------------------
2222    
2223    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2224        a string to a long long, pretend we don't even have a long long." This is
2225        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2226    
2227    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2228        pre-7.6 versions, which defined a global no_arg variable instead of putting
2229        it in the RE class. (See also #8 below.)
2230    
2231    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2232        Lopes.
2233    
2234    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2235    
2236        (1) The include/exclude patterns were being applied to the whole pathnames
2237            of files, instead of just to the final components.
2238    
2239        (2) If there was more than one level of directory, the subdirectories were
2240            skipped unless they satisfied the include/exclude conditions. This is
2241            inconsistent with GNU grep (and could even be seen as contrary to the
2242            pcregrep specification - which I improved to make it absolutely clear).
2243            The action now is always to scan all levels of directory, and just
2244            apply the include/exclude patterns to regular files.
2245    
2246    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2247        --exclude_dir in the tests to avoid scanning .svn directories.
2248    
2249    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2250        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2251        doesn't support NULs in patterns.
2252    
2253    7.  Added some missing "const"s to declarations of static tables in
2254        pcre_compile.c and pcre_dfa_exec.c.
2255    
2256    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2257        caused by fix #2  above. (Subsequently also a second patch to fix the
2258        first patch. And a third patch - this was a messy problem.)
2259    
2260    9.  Applied Craig's patch to remove the use of push_back().
2261    
2262    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2263        matching function regexec().
2264    
2265    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2266        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2267        references. PCRE supports relative numbers with this syntax (I don't think
2268        Oniguruma does).
2269    
2270    12. Previously, a group with a zero repeat such as (...){0} was completely
2271        omitted from the compiled regex. However, this means that if the group
2272        was called as a subroutine from elsewhere in the pattern, things went wrong
2273        (an internal error was given). Such groups are now left in the compiled
2274        pattern, with a new opcode that causes them to be skipped at execution
2275        time.
2276    
2277    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2278        to the way PCRE behaves:
2279    
2280        (a) A lone ] character is dis-allowed (Perl treats it as data).
2281    
2282        (b) A back reference to an unmatched subpattern matches an empty string
2283            (Perl fails the current match path).
2284    
2285        (c) A data ] in a character class must be notated as \] because if the
2286            first data character in a class is ], it defines an empty class. (In
2287            Perl it is not possible to have an empty class.) The empty class []
2288            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2289            The negative empty class [^] matches any one character, independently
2290            of the DOTALL setting.
2291    
2292    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2293        non-existent subpattern following a character class starting with ']' and
2294        containing () gave an internal compiling error instead of "reference to
2295        non-existent subpattern". Fortunately, when the pattern did exist, the
2296        compiled code was correct. (When scanning forwards to check for the
2297        existence of the subpattern, it was treating the data ']' as terminating
2298        the class, so got the count wrong. When actually compiling, the reference
2299        was subsequently set up correctly.)
2300    
2301    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2302        it was being rejected as not supported by pcre_dfa_exec(), even though
2303        other assertions are supported. I have made pcre_dfa_exec() support
2304        (*FAIL).
2305    
2306    16. The implementation of 13c above involved the invention of a new opcode,
2307        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2308        cannot be changed at match time, I realized I could make a small
2309        improvement to matching performance by compiling OP_ALLANY instead of
2310        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2311        on the OP_ANY path.
2312    
2313    17. Compiling pcretest on Windows with readline support failed without the
2314        following two fixes: (1) Make the unistd.h include conditional on
2315        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2316    
2317    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2318        ncurses library to be included for pcretest when ReadLine support is
2319        requested, but also to allow for it to be overridden. This patch came from
2320        Daniel Bergström.
2321    
2322    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2323        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2324        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2325        spotting this.
2326    
2327    
2328    Version 7.6 28-Jan-08
2329    ---------------------
2330    
2331    1.  A character class containing a very large number of characters with
2332        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2333        overflow.
2334    
2335    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2336        HAVE_LONG_LONG is not defined.
2337    
2338    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2339        bring them up to date and include new features. This patch includes:
2340    
2341        - Fixed PH's badly added libz and libbz2 support.
2342        - Fixed a problem with static linking.
2343        - Added pcredemo. [But later removed - see 7 below.]
2344        - Fixed dftables problem and added an option.
2345        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2346            HAVE_LONG_LONG.
2347        - Added readline support for pcretest.
2348        - Added an listing of the option settings after cmake has run.
2349    
2350    4.  A user submitted a patch to Makefile that makes it easy to create
2351        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2352        Makefile.am that cause it to include this special target, without
2353        affecting anything else. Note that the same mingw target plus all
2354        the other distribution libraries and programs are now supported
2355        when configuring with CMake (see 6 below) instead of with
2356        Configure/Make.
2357    
2358    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2359        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2360        exported in the Windows port". It has not yet been confirmed that the patch
2361        solves the problem, but it does no harm.
2362    
2363    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2364        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2365        with CMake, and also correct the comment about stack recursion.
2366    
2367    7.  Remove the automatic building of pcredemo from the ./configure system and
2368        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2369        of a program that users should build themselves after PCRE is installed, so
2370        building it automatically is not really right. What is more, it gave
2371        trouble in some build environments.
2372    
2373    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2374    
2375    
2376    Version 7.5 10-Jan-08
2377    ---------------------
2378    
2379    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2380        values in parens when parsing an RE using the C++ wrapper."
2381    
2382    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2383        Characters greater than 255 were excluded from the class instead of being
2384        included.
2385    
2386    3.  The same bug as (2) above applied to negated POSIX classes such as
2387        [:^space:].
2388    
2389    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2390        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2391        I have changed it.
2392    
2393    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2394        first named subpattern) and a construct such as (?&a) would reference the
2395        first named subpattern whose name started with "a" (in other words, the
2396        length check was missing). Both these problems are fixed. "Subpattern name
2397        expected" is now given for (?&) (a zero-length name), and this patch also
2398        makes it give the same error for \k'' (previously it complained that that
2399        was a reference to a non-existent subpattern).
2400    
2401    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2402        this is right because (?- can be followed by option settings as well as by
2403        digits. I have, however, made the messages clearer.
2404    
2405    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2406        than the number used in the conditional) now cause a compile-time error.
2407        This is actually not compatible with Perl, which accepts such patterns, but
2408        treats the conditional as always being FALSE (as PCRE used to), but it
2409        seems to me that giving a diagnostic is better.
2410    
2411    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2412        and messages.
2413    
2414    9.  Fix two occurrences of "backslash" in comments that should have been
2415        "backspace".
2416    
2417    10. Remove two redundant lines of code that can never be obeyed (their function
2418        was moved elsewhere).
2419    
2420    11. The program that makes PCRE's Unicode character property table had a bug
2421        which caused it to generate incorrect table entries for sequences of
2422        characters that have the same character type, but are in different scripts.
2423        It amalgamated them into a single range, with the script of the first of
2424        them. In other words, some characters were in the wrong script. There were
2425        thirteen such cases, affecting characters in the following ranges:
2426    
2427          U+002b0 - U+002c1
2428          U+0060c - U+0060d
2429          U+0061e - U+00612
2430          U+0064b - U+0065e
2431          U+0074d - U+0076d
2432          U+01800 - U+01805
2433          U+01d00 - U+01d77
2434          U+01d9b - U+01dbf
2435          U+0200b - U+0200f
2436          U+030fc - U+030fe
2437          U+03260 - U+0327f
2438          U+0fb46 - U+0fbb1
2439          U+10450 - U+1049d
2440    
2441    12. The -o option (show only the matching part of a line) for pcregrep was not
2442        compatible with GNU grep in that, if there was more than one match in a
2443        line, it showed only the first of them. It now behaves in the same way as
2444        GNU grep.
2445    
2446    13. If the -o and -v options were combined for pcregrep, it printed a blank
2447        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2448        does the same. The return code can be used to tell if there were any
2449        non-matching lines.
2450    
2451    14. Added --file-offsets and --line-offsets to pcregrep.
2452    
2453    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2454        infinitely looping recursion. The bug was that positive lookaheads were not
2455        being skipped when checking for a possible empty match (negative lookaheads
2456        and both kinds of lookbehind were skipped).
2457    
2458    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2459        inclusion of <windows.h> to before rather than after the definition of
2460        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2461    
2462    17. Specifying a possessive quantifier with a specific limit for a Unicode
2463        character property caused pcre_compile() to compile bad code, which led at
2464        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2465        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2466        caused the error; without that there was no problem.
2467    
2468    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2469    
2470    19. Added --enable-pcretest-libreadline.
2471    
2472    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2473        RE::GlobalReplace(). As a result, the number of replacements returned was
2474        double what it should be. I removed one of the increments, but Craig sent a
2475        later patch that removed the other one (the right fix) and added unit tests
2476        that check the return values (which was not done before).
2477    
2478    21. Several CMake things:
2479    
2480        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2481            the names libpcre and libpcreposix, not just pcre and pcreposix.
2482    
2483        (2) The above change means that pcretest and pcregrep are now correctly
2484            linked with the newly-built libraries, not previously installed ones.
2485    
2486        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2487    
2488    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2489        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2490        UTF-8 newline character). The key issue is that the pattern starts .*;
2491        this means that the match must be either at the beginning, or after a
2492        newline. The bug was in the code for advancing after a failed match and
2493        checking that the new position followed a newline. It was not taking
2494        account of UTF-8 characters correctly.
2495    
2496    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2497        character classes. PCRE was not treating the sequence [:...:] as a
2498        character class unless the ... were all letters. Perl, however, seems to
2499        allow any characters between [: and :], though of course it rejects as
2500        unknown any "names" that contain non-letters, because all the known class
2501        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2502        for example, whereas PCRE did not - it did not recognize a POSIX character
2503        class. This seemed a bit dangerous, so the code has been changed to be
2504        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2505        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2506        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2507        Perl does, and where it didn't before.
2508    
2509    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2510        Windows environments %n is disabled by default.
2511    
2512    
2513    Version 7.4 21-Sep-07
2514    ---------------------
2515    
2516    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2517        means that a class such as [\s] counted as "explicit reference to CR or
2518        LF". That isn't really right - the whole point of the change was to try to
2519        help when there was an actual mention of one of the two characters. So now
2520        the change happens only if \r or \n (or a literal CR or LF) character is
2521        encountered.
2522    
2523    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2524        of both had grown to the point where there were only 3 bits left.
2525        Fortunately, there was spare space in the data structure, and so I have
2526        moved the internal flags into a new 16-bit field to free up more option
2527        bits.
2528    
2529    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2530        but did not set the internal JCHANGED flag - either of these is enough to
2531        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2532        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2533        start sets both bits.
2534    
2535    4.  Added options (at build time, compile time, exec time) to change \R from
2536        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2537    
2538    5.  doc/pcresyntax.html was missing from the distribution.
2539    
2540    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2541        compatibility, even though it is no longer used.
2542    
2543    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2544        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2545        windows.h file is present (where different names are used). [This was
2546        reversed later after testing - see 16 below.]
2547    
2548    8.  Changed all #include <config.h> to #include "config.h". There were also
2549        some further <pcre.h> cases that I changed to "pcre.h".
2550    
2551    9.  When pcregrep was used with the --colour option, it missed the line ending
2552        sequence off the lines that it output.
2553    
2554    10. It was pointed out to me that arrays of string pointers cause lots of
2555        relocations when a shared library is dynamically loaded. A technique of
2556        using a single long string with a table of offsets can drastically reduce
2557        these. I have refactored PCRE in four places to do this. The result is
2558        dramatic:
2559    
2560          Originally:                          290
2561          After changing UCP table:            187
2562          After changing error message table:   43
2563          After changing table of "verbs"       36
2564          After changing table of Posix names   22
2565    
2566        Thanks to the folks working on Gregex for glib for this insight.
2567    
2568    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2569        unicode-properties was also set.
2570    
2571    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2572    
2573    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2574        checked only for CRLF.
2575    
2576    14. Added casts to pcretest.c to avoid compiler warnings.
2577    
2578    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2579    
2580    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2581        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2582        entirely. This removes changes made in 7 above.
2583    
2584    17. The CMake files have been updated, and there is now more information about
2585        building with CMake in the NON-UNIX-USE document.
2586    
2587    
2588    Version 7.3 28-Aug-07
2589    ---------------------
2590    
2591     1. In the rejigging of the build system that eventually resulted in 7.1, the
2592        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2593        brackets there is not right, since it causes compilers to look for an
2594        installed pcre.h, not the version that is in the source that is being
2595        compiled (which of course may be different). I have changed it back to:
2596    
2597          #include "pcre.h"
2598    
2599        I have a vague recollection that the change was concerned with compiling in
2600        different directories, but in the new build system, that is taken care of
2601        by the VPATH setting the Makefile.
2602    
2603     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2604        when the subject happened to end in the byte 0x85 (e.g. if the last
2605        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2606        characters but of course it shouldn't be taken as a newline when it is part
2607        of another character. The bug was that, for an unlimited repeat of . in
2608        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2609        characters when looking for a newline.
2610    
2611     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2612    
2613     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2614        in debug output.
2615    
2616     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2617        long printing in the pcrecpp unittest when running under MinGW.
2618    
2619     6. ESC_K was left out of the EBCDIC table.
2620    
2621     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2622        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2623        limit also applies to "virtual nesting" when a pattern is recursive, and in
2624        this case 1000 isn't so big. I have been able to remove this limit at the
2625        expense of backing off one optimization in certain circumstances. Normally,
2626        when pcre_exec() would call its internal match() function recursively and
2627        immediately return the result unconditionally, it uses a "tail recursion"
2628        feature to save stack. However, when a subpattern that can match an empty
2629        string has an unlimited repetition quantifier, it no longer makes this
2630        optimization. That gives it a stack frame in which to save the data for
2631        checking that an empty string has been matched. Previously this was taken
2632        from the 1000-entry workspace that had been reserved. So now there is no
2633        explicit limit, but more stack is used.
2634    
2635     8. Applied Daniel's patches to solve problems with the import/export magic
2636        syntax that is required for Windows, and which was going wrong for the
2637        pcreposix and pcrecpp parts of the library. These were overlooked when this
2638        problem was solved for the main library.
2639    
2640     9. There were some crude static tests to avoid integer overflow when computing
2641        the size of patterns that contain repeated groups with explicit upper
2642        limits. As the maximum quantifier is 65535, the maximum group length was
2643        set at 30,000 so that the product of these two numbers did not overflow a
2644        32-bit integer. However, it turns out that people want to use groups that
2645        are longer than 30,000 bytes (though not repeat them that many times).
2646        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2647        made it possible to implement the integer overflow checks in a much more
2648        dynamic way, which I have now done. The artificial limitation on group
2649        length has been removed - we now have only the limit on the total length of
2650        the compiled pattern, which depends on the LINK_SIZE setting.
2651    
2652    10. Fixed a bug in the documentation for get/copy named substring when
2653        duplicate names are permitted. If none of the named substrings are set, the
2654        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2655        empty string.
2656    
2657    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2658        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2659        because the ] is interpreted as the first data character and the
2660        terminating ] is not found. PCRE has been made compatible with Perl in this
2661        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2662        cause memory overwriting.
2663    
2664    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2665        string has been matched (to stop an infinite loop). It was not recognizing
2666        a conditional subpattern that could match an empty string if that
2667        subpattern was within another subpattern. For example, it looped when
2668        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2669        condition was not nested. This bug has been fixed.
2670    
2671    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2672        past the start of the subject in the presence of bytes with the top bit
2673        set, for example "\x8aBCD".
2674    
2675    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2676        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2677    
2678    14. Optimized (?!) to (*FAIL).
2679    
2680    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2681        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2682        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2683        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2684        does: it's just the validity check that is more restrictive.
2685    
2686    16. Inserted checks for integer overflows during escape sequence (backslash)
2687        processing, and also fixed erroneous offset values for syntax errors during
2688        backslash processing.
2689    
2690    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2691        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2692    
2693    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2694        caused an overrun.
2695    
2696    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2697        something other than just ASCII characters) inside a group that had an
2698        unlimited repeat caused a loop at compile time (while checking to see
2699        whether the group could match an empty string).
2700    
2701    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2702        [\P{Any}] did so. (Error in the code for printing property names.)
2703    
2704    21. An orphan \E inside a character class could cause a crash.
2705    
2706    22. A repeated capturing bracket such as (A)? could cause a wild memory
2707        reference during compilation.
2708    
2709    23. There are several functions in pcre_compile() that scan along a compiled
2710        expression for various reasons (e.g. to see if it's fixed length for look
2711        behind). There were bugs in these functions when a repeated \p or \P was
2712        present in the pattern. These operators have additional parameters compared
2713        with \d, etc, and these were not being taken into account when moving along
2714        the compiled data. Specifically:
2715    
2716        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2717            length.
2718    
2719        (b) An item such as \pL+ within a repeated group could cause crashes or
2720            loops.
2721    
2722        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2723            "reference to non-existent subpattern" error.
2724    
2725        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2726    
2727    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2728        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2729    
2730    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2731    
2732    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2733        character were causing crashes (broken optimization).
2734    
2735    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2736        \p or \P) caused a compile-time loop.
2737    
2738    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2739        break. For example, the unstudied pattern [\r\n]A does not match the string
2740        "\r\nA" because change 7.0/46 below moves the current point on by two
2741        characters after failing to match at the start. However, the pattern \nA
2742        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2743        the same is true. There doesn't seem any very clean way out of this, but
2744        what I have chosen to do makes the common cases work: PCRE now takes note
2745        of whether there can be an explicit match for \r or \n anywhere in the
2746        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2747        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2748        pattern has explicit CR or LF references.
2749    
2750    29. Added (*CR) etc for changing newline setting at start of pattern.
2751    
2752    
2753    Version 7.2 19-Jun-07
2754  ---------------------  ---------------------
2755    
2756   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2757      which is apparently normally available under Windows.      which is apparently normally available under Windows.
2758    
2759   2. Re-jig the pcregrep tests with different newline settings in an attempt   2. Re-jig the pcregrep tests with different newline settings in an attempt
2760      to make them independent of the local environment's newline setting.      to make them independent of the local environment's newline setting.
2761    
2762     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2763    
2764     4. Some of the "internals" tests were previously cut out when the link size
2765        was not 2, because the output contained actual offsets. The recent new
2766        "Z" feature of pcretest means that these can be cut out, making the tests
2767        usable with all link sizes.
2768    
2769     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2770        stack recursion. This gives a massive performance boost under BSD, but just
2771        a small improvement under Linux. However, it saves one field in the frame
2772        in all cases.
2773    
2774     6. Added more features from the forthcoming Perl 5.10:
2775    
2776        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2777            recursion call. It refers to the nth most recently opened parentheses.
2778    
2779        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2780            to be opened parentheses.
2781    
2782        (c) Conditions that refer to capturing parentheses can be specified
2783            relatively, for example, (?(-2)... or (?(+3)...
2784    
2785        (d) \K resets the start of the current match so that everything before
2786            is not part of it.
2787    
2788        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2789    
2790        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2791            reference syntax.
2792    
2793        (g) (?| introduces a group in which the numbering of parentheses in each
2794            alternative starts with the same number.
2795    
2796        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2797    
2798     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2799        PCRE_INFO_JCHANGED.
2800    
2801     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2802        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2803        for detecting groups that can match an empty string.
2804    
2805     9. A pattern with a very large number of alternatives (more than several
2806        hundred) was running out of internal workspace during the pre-compile
2807        phase, where pcre_compile() figures out how much memory will be needed. A
2808        bit of new cunning has reduced the workspace needed for groups with
2809        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2810        workspace instead of running out of the 4096 that are available.
2811    
2812    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2813    
2814    11. Applied patch from Google to remove an optimization that didn't quite work.
2815        The report of the bug said:
2816    
2817          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2818          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2819          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2820    
2821    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2822        it matched the wrong number of bytes.
2823    
2824    
2825  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
2826  ---------------------  ---------------------

Legend:
Removed from v.158  
changed lines
  Added in v.1481

  ViewVC Help
Powered by ViewVC 1.1.5