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

Diff of /code/trunk/ChangeLog

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

revision 127 by ph10, Mon Mar 19 11:44:45 2007 UTC revision 1490 by chpe, Thu Jun 19 07:51:39 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.1 12-Mar-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    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
51        to hold a pattern to be incorrectly computed (too small) when there were
52        named back references to duplicated names. This could cause "internal
53        error: code overflow" or "double free or corruption" or other memory
54        handling errors.
55    
56    11. When named subpatterns had the same prefixes, back references could be
57        confused. For example, in this pattern:
58    
59          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
60    
61        the reference to 'Name' was incorrectly treated as a reference to a
62        duplicate name.
63    
64    12. A pattern such as /^s?c/mi8 where the optional character has more than
65        one "other case" was incorrectly compiled such that it would only try to
66        match starting at "c".
67    
68    13. When a pattern starting with \s was studied, VT was not included in the
69        list of possible starting characters; this should have been part of the
70        8.34/18 patch.
71    
72    14. If a character class started [\Qx]... where x is any character, the class
73        was incorrectly terminated at the ].
74    
75    15. If a pattern that started with a caseless match for a character with more
76        than one "other case" was studied, PCRE did not set up the starting code
77        unit bit map for the list of possible characters. Now it does. This is an
78        optimization improvement, not a bug fix.
79    
80    16. The Unicode data tables have been updated to Unicode 7.0.0.
81    
82    
83    Version 8.35 04-April-2014
84    --------------------------
85    
86    1.  A new flag is set, when property checks are present in an XCLASS.
87        When this flag is not set, PCRE can perform certain optimizations
88        such as studying these XCLASS-es.
89    
90    2.  The auto-possessification of character sets were improved: a normal
91        and an extended character set can be compared now. Furthermore
92        the JIT compiler optimizes more character set checks.
93    
94    3.  Got rid of some compiler warnings for potentially uninitialized variables
95        that show up only when compiled with -O2.
96    
97    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
98        of a match later then the end of the match. The pcretest program was not
99        handling the case sensibly - it was outputting from the start to the next
100        binary zero. It now reports this situation in a message, and outputs the
101        text from the end to the start.
102    
103    5.  Fast forward search is improved in JIT. Instead of the first three
104        characters, any three characters with fixed position can be searched.
105        Search order: first, last, middle.
106    
107    6.  Improve character range checks in JIT. Characters are read by an inprecise
108        function now, which returns with an unknown value if the character code is
109        above a certain treshold (e.g: 256). The only limitation is that the value
110        must be bigger than the treshold as well. This function is useful, when
111        the characters above the treshold are handled in the same way.
112    
113    7.  The macros whose names start with RAWUCHAR are placeholders for a future
114        mode in which only the bottom 21 bits of 32-bit data items are used. To
115        make this more memorable for those maintaining the code, the names have
116        been changed to start with UCHAR21, and an extensive comment has been added
117        to their definition.
118    
119    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
120        to the export list in Makefile.am (they were accidentally omitted from the
121        8.34 tarball).
122    
123    9.  The informational output from pcretest used the phrase "starting byte set"
124        which is inappropriate for the 16-bit and 32-bit libraries. As the output
125        for "first char" and "need char" really means "non-UTF-char", I've changed
126        "byte" to "char", and slightly reworded the output. The documentation about
127        these values has also been (I hope) clarified.
128    
129    10. Another JIT related optimization: use table jumps for selecting the correct
130        backtracking path, when more than four alternatives are present inside a
131        bracket.
132    
133    11. Empty match is not possible, when the minimum length is greater than zero,
134        and there is no \K in the pattern. JIT should avoid empty match checks in
135        such cases.
136    
137    12. In a caseless character class with UCP support, when a character with more
138        than one alternative case was not the first character of a range, not all
139        the alternative cases were added to the class. For example, s and \x{17f}
140        are both alternative cases for S: the class [RST] was handled correctly,
141        but [R-T] was not.
142    
143    13. The configure.ac file always checked for pthread support when JIT was
144        enabled. This is not used in Windows, so I have put this test inside a
145        check for the presence of windows.h (which was already tested for).
146    
147    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
148        The algorithm provides a way to skip certain starting offsets, and usually
149        faster than linear prefix searches.
150    
151    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
152        as for 'fr_FR' and 'french'. For some reason, however, it then used the
153        Windows-specific input and output files, which have 'french' screwed in.
154        So this could never have worked. One of the problems with locales is that
155        they aren't always the same. I have now updated RunTest so that it checks
156        the output of the locale test (test 3) against three different output
157        files, and it allows the test to pass if any one of them matches. With luck
158        this should make the test pass on some versions of Solaris where it was
159        failing. Because of the uncertainty, the script did not used to stop if
160        test 3 failed; it now does. If further versions of a French locale ever
161        come to light, they can now easily be added.
162    
163    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
164        there was a message during ./configure, but it did not stop. This now
165        provokes an error. The invalid example in README has been corrected.
166        If a value less than the minimum is given, the minimum value has always
167        been used, but now a warning is given.
168    
169    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
170        was a bug in the test system, which is now fixed. Also, the list of various
171        configurations that are tested for each release did not have one with both
172        16/32 bits and --enable-bar-anycrlf. It now does.
173    
174    18. pcretest was missing "-C bsr" for displaying the \R default setting.
175    
176    19. Little endian PowerPC systems are supported now by the JIT compiler.
177    
178    20. The fast forward newline mechanism could enter to an infinite loop on
179        certain invalid UTF-8 input. Although we don't support these cases
180        this issue can be fixed by a performance optimization.
181    
182    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
183        not take account if existing stack usage. There is now a new global
184        variable called pcre_stack_guard that can be set to point to an external
185        function to check stack availability. It is called at the start of
186        processing every parenthesized group.
187    
188    22. A typo in the code meant that in ungreedy mode the max/min qualifier
189        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
190        not match "ab".
191    
192    23. When UTF was disabled, the JIT program reported some incorrect compile
193        errors. These messages are silenced now.
194    
195    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
196        compiler.
197    
198    25. Change all the temporary files used in RunGrepTest to be different to those
199        used by RunTest so that the tests can be run simultaneously, for example by
200        "make -j check".
201    
202    
203    Version 8.34 15-December-2013
204    -----------------------------
205    
206    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
207        executable memory. Patch inspired by Carsten Klein.
208    
209    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
210        this macro is never tested and has no effect, because the work to support
211        coverage involves only compiling and linking options and special targets in
212        the Makefile. The comment in config.h implied that defining the macro would
213        enable coverage support, which is totally false. There was also support for
214        setting this macro in the CMake files (my fault, I just copied it from
215        configure). SUPPORT_GCOV has now been removed.
216    
217    3.  Make a small performance improvement in strlen16() and strlen32() in
218        pcretest.
219    
220    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
221        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
222        these cases). There was also one in pcretest.c.
223    
224    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
225    
226    6.  In UTF mode, the code for checking whether a group could match an empty
227        string (which is used for indefinitely repeated groups to allow for
228        breaking an infinite loop) was broken when the group contained a repeated
229        negated single-character class with a character that occupied more than one
230        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
231        UTF-8 mode). The effect was undefined: the group might or might not be
232        deemed as matching an empty string, or the program might have crashed.
233    
234    7.  The code for checking whether a group could match an empty string was not
235        recognizing that \h, \H, \v, \V, and \R must match a character.
236    
237    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
238        an empty string. If it can, pcretest shows this in its information output.
239    
240    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
241        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
242        matched by pcre_exec() without using JIT:
243    
244        (a) If the rest of the pattern did not match after a maximal run of
245            grapheme clusters, the code for backing up to try with fewer of them
246            did not always back up over a full grapheme when characters that do not
247            have the modifier quality were involved, e.g. Hangul syllables.
248    
249        (b) If the match point in a subject started with modifier character, and
250            there was no match, the code could incorrectly back up beyond the match
251            point, and potentially beyond the first character in the subject,
252            leading to a segfault or an incorrect match result.
253    
254    10. A conditional group with an assertion condition could lead to PCRE
255        recording an incorrect first data item for a match if no other first data
256        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
257        first data item, and therefore matched "ca" after "c" instead of at the
258        start.
259    
260    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
261        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
262    
263    12. The source of pcregrep now includes z/OS-specific code so that it can be
264        compiled for z/OS as part of the special z/OS distribution.
265    
266    13. Added the -T and -TM options to pcretest.
267    
268    14. The code in pcre_compile.c for creating the table of named capturing groups
269        has been refactored. Instead of creating the table dynamically during the
270        actual compiling pass, the information is remembered during the pre-compile
271        pass (on the stack unless there are more than 20 named groups, in which
272        case malloc() is used) and the whole table is created before the actual
273        compile happens. This has simplified the code (it is now nearly 150 lines
274        shorter) and prepared the way for better handling of references to groups
275        with duplicate names.
276    
277    15. A back reference to a named subpattern when there is more than one of the
278        same name now checks them in the order in which they appear in the pattern.
279        The first one that is set is used for the reference. Previously only the
280        first one was inspected. This change makes PCRE more compatible with Perl.
281    
282    16. Unicode character properties were updated from Unicode 6.3.0.
283    
284    17. The compile-time code for auto-possessification has been refactored, based
285        on a patch by Zoltan Herczeg. It now happens after instead of during
286        compilation. The code is cleaner, and more cases are handled. The option
287        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
288        options in pcretest are provided to set it. It can also be set by
289        (*NO_AUTO_POSSESS) at the start of a pattern.
290    
291    18. The character VT has been added to the default ("C" locale) set of
292        characters that match \s and are generally treated as white space,
293        following this same change in Perl 5.18. There is now no difference between
294        "Perl space" and "POSIX space". Whether VT is treated as white space in
295        other locales depends on the locale.
296    
297    19. The code for checking named groups as conditions, either for being set or
298        for being recursed, has been refactored (this is related to 14 and 15
299        above). Processing unduplicated named groups should now be as fast at
300        numerical groups, and processing duplicated groups should be faster than
301        before.
302    
303    20. Two patches to the CMake build system, by Alexander Barkov:
304    
305          (1) Replace the "source" command by "." in CMakeLists.txt because
306              "source" is a bash-ism.
307    
308          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
309              without these the CMake build does not work on Solaris.
310    
311    21. Perl has changed its handling of \8 and \9. If there is no previously
312        encountered capturing group of those numbers, they are treated as the
313        literal characters 8 and 9 instead of a binary zero followed by the
314        literals. PCRE now does the same.
315    
316    22. Following Perl, added \o{} to specify codepoints in octal, making it
317        possible to specify values greater than 0777 and also making them
318        unambiguous.
319    
320    23. Perl now gives an error for missing closing braces after \x{... instead of
321        treating the string as literal. PCRE now does the same.
322    
323    24. RunTest used to grumble if an inappropriate test was selected explicitly,
324        but just skip it when running all tests. This make it awkward to run ranges
325        of tests when one of them was inappropriate. Now it just skips any
326        inappropriate tests, as it always did when running all tests.
327    
328    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
329        character types such as \d or \w, too many callouts were inserted, and the
330        data that they returned was rubbish.
331    
332    26. In UCP mode, \s was not matching two of the characters that Perl matches,
333        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
334        were matched by \h. The code has now been refactored so that the lists of
335        the horizontal and vertical whitespace characters used for \h and \v (which
336        are defined only in one place) are now also used for \s.
337    
338    27. Add JIT support for the 64 bit TileGX architecture.
339        Patch by Jiong Wang (Tilera Corporation).
340    
341    28. Possessive quantifiers for classes (both explicit and automatically
342        generated) now use special opcodes instead of wrapping in ONCE brackets.
343    
344    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
345        (because it's meaningless), this was not happening when PCRE_CASELESS was
346        set. Not wrong, but inefficient.
347    
348    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
349        properties for \w, \d, etc) is present in a test regex. Otherwise if the
350        test contains no characters greater than 255, Perl doesn't realise it
351        should be using Unicode semantics.
352    
353    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
354        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
355        does in Unicode mode.
356    
357    32. Added the "forbid" facility to pcretest so that putting tests into the
358        wrong test files can sometimes be quickly detected.
359    
360    33. There is now a limit (default 250) on the depth of nesting of parentheses.
361        This limit is imposed to control the amount of system stack used at compile
362        time. It can be changed at build time by --with-parens-nest-limit=xxx or
363        the equivalent in CMake.
364    
365    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
366        errors. Perl warns for these when in warning mode, but PCRE has no facility
367        for giving warnings.
368    
369    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
370        However, this was not working for (?!) because it is optimized to (*FAIL),
371        for which PCRE does not allow quantifiers. The optimization is now disabled
372        when a quantifier follows (?!). I can't see any use for this, but it makes
373        things uniform.
374    
375    36. Perl no longer allows group names to start with digits, so I have made this
376        change also in PCRE. It simplifies the code a bit.
377    
378    37. In extended mode, Perl ignores spaces before a + that indicates a
379        possessive quantifier. PCRE allowed a space before the quantifier, but not
380        before the possessive +. It now does.
381    
382    38. The use of \K (reset reported match start) within a repeated possessive
383        group such as (a\Kb)*+ was not working.
384    
385    40. Document that the same character tables must be used at compile time and
386        run time, and that the facility to pass tables to pcre_exec() and
387        pcre_dfa_exec() is for use only with saved/restored patterns.
388    
389    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
390        features for Builds with MSVC:
391    
392        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
393           builds). The .rc files can be used to set FileDescription and many other
394           attributes.
395    
396        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
397           This allows higher-level build scripts which want .pdb files to avoid
398           hard-coding the exact files needed."
399    
400    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
401        mean "start of word" and "end of word", respectively, as a transition aid.
402    
403    43. A minimizing repeat of a class containing codepoints greater than 255 in
404        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
405        compiled to use the heap for recursion.
406    
407    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
408        is configured.
409    
410    
411    Version 8.33 28-May-2013
412    ------------------------
413    
414    1.  Added 'U' to some constants that are compared to unsigned integers, to
415        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
416        variables that are added to signed variables, to ensure the result is
417        signed and can be negated.
418    
419    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
420        CMake config files.
421    
422    3.  Revise the creation of config.h.generic so that all boolean macros are
423        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
424        overriding via -D on the command line possible.
425    
426    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
427        to unsigned int is reported to make a quite noticeable speed difference in
428        a specific Windows environment. Testing on Linux did also appear to show
429        some benefit (and it is clearly not harmful). Also fixed the definition of
430        Xop which should be unsigned.
431    
432    5.  Related to (4), changing the definition of the intermediate variable cc
433        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
434        improvements.
435    
436    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
437        unnecessary spaces.
438    
439    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
440        and later.
441    
442    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
443    
444    9.  Optimizing fast_forward_start_bits in JIT.
445    
446    10. Adding support for callouts in JIT, and fixing some issues revealed
447        during this work. Namely:
448    
449        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
450    
451        (b) Minimum length was not checked before the matching is started.
452    
453    11. The value of capture_last that is passed to callouts was incorrect in some
454        cases when there was a capture on one path that was subsequently abandoned
455        after a backtrack. Also, the capture_last value is now reset after a
456        recursion, since all captures are also reset in this case.
457    
458    12. The interpreter no longer returns the "too many substrings" error in the
459        case when an overflowing capture is in a branch that is subsequently
460        abandoned after a backtrack.
461    
462    13. In the pathological case when an offset vector of size 2 is used, pcretest
463        now prints out the matched string after a yield of 0 or 1.
464    
465    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
466        Only supported by the JIT compiler at the moment.
467    
468    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
469    
470    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
471        offset of the starting point of the matching process, provided the offsets
472        vector is large enough.
473    
474    17. The \A escape now records a lookbehind value of 1, though its execution
475        does not actually inspect the previous character. This is to ensure that,
476        in partial multi-segment matching, at least one character from the old
477        segment is retained when a new segment is processed. Otherwise, if there
478        are no lookbehinds in the pattern, \A might match incorrectly at the start
479        of a new segment.
480    
481    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
482    
483    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
484        gives some modest performance improvement in 8-bit mode.
485    
486    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
487        be expressed in certain programming languages using Universal Character
488        Names.
489    
490    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
491        which points out that "non characters" are not "characters that may not
492        appear in Unicode strings" but rather "characters that are reserved for
493        internal use and have only local meaning".
494    
495    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
496        there was a conditional group that depended on an assertion, if the
497        assertion was false, the callout that immediately followed the alternation
498        in the condition was skipped when pcre_exec() was used for matching.
499    
500    23. Allow an explicit callout to be inserted before an assertion that is the
501        condition for a conditional group, for compatibility with automatic
502        callouts, which always insert a callout at this point.
503    
504    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
505        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
506    
507    25. (*PRUNE) is now supported by the JIT compiler.
508    
509    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
510    
511    27. Fix the case where there are two or more SKIPs with arguments that may be
512        ignored.
513    
514    28. (*SKIP) is now supported by the JIT compiler.
515    
516    29. (*THEN) is now supported by the JIT compiler.
517    
518    30. Update RunTest with additional test selector options.
519    
520    31. The way PCRE handles backtracking verbs has been changed in two ways.
521    
522        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
523        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
524        backtracking. In some cases this makes it more Perl-compatible, but Perl's
525        rather obscure rules do not always do the same thing.
526    
527        (2) Previously, backtracking verbs were confined within assertions. This is
528        no longer the case for positive assertions, except for (*ACCEPT). Again,
529        this sometimes improves Perl compatibility, and sometimes does not.
530    
531    32. A number of tests that were in test 2 because Perl did things differently
532        have been moved to test 1, because either Perl or PCRE has changed, and
533        these tests are now compatible.
534    
535    32. Backtracking control verbs are now handled in the same way in JIT and
536        interpreter.
537    
538    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
539        contained a forward subroutine reference caused a compile error.
540    
541    34. Auto-detect and optimize limited repetitions in JIT.
542    
543    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
544        blocking (*UTF) etc.
545    
546    36. In the interpreter, maximizing pattern repetitions for characters and
547        character types now use tail recursion, which reduces stack usage.
548    
549    37. The value of the max lookbehind was not correctly preserved if a compiled
550        and saved regex was reloaded on a host of different endianness.
551    
552    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
553        of the compiled pattern block, expand the flags field from 16 to 32 bits
554        because it was almost full.
555    
556    39. Try madvise first before posix_madvise.
557    
558    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
559        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
560        original change locked out all matches of empty strings. This has been
561        changed so that one match of an empty string per line is recognized.
562        Subsequent searches on the same line (for colouring or for --only-matching,
563        for example) do not recognize empty strings.
564    
565    41. Applied a user patch to fix a number of spelling mistakes in comments.
566    
567    42. Data lines longer than 65536 caused pcretest to crash.
568    
569    43. Clarified the data type for length and startoffset arguments for pcre_exec
570        and pcre_dfa_exec in the function-specific man pages, where they were
571        explicitly stated to be in bytes, never having been updated. I also added
572        some clarification to the pcreapi man page.
573    
574    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
575        a segmentation fault.
576    
577    
578    Version 8.32 30-November-2012
579    -----------------------------
580    
581    1.  Improved JIT compiler optimizations for first character search and single
582        character iterators.
583    
584    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
585        Patch by Daniel Richard G.
586    
587    3.  Single character iterator optimizations in the JIT compiler.
588    
589    4.  Improved JIT compiler optimizations for character ranges.
590    
591    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
592        Reported by Giuseppe D'Angelo.
593    
594    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
595        of a line, was being set incorrectly in cases where .* appeared inside
596        atomic brackets at the start of a pattern, or where there was a subsequent
597        *PRUNE or *SKIP.
598    
599    7.  Improved instruction cache flush for POWER/PowerPC.
600        Patch by Daniel Richard G.
601    
602    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
603        grep:
604    
605        (a) There is now no limit to the number of patterns to be matched.
606    
607        (b) An error is given if a pattern is too long.
608    
609        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
610            are now supported.
611    
612        (d) --exclude-from and --include-from (multiple use) have been added.
613    
614        (e) Exclusions and inclusions now apply to all files and directories, not
615            just to those obtained from scanning a directory recursively.
616    
617        (f) Multiple uses of -f and --file-list are now supported.
618    
619        (g) In a Windows environment, the default for -d has been changed from
620            "read" (the GNU grep default) to "skip", because otherwise the presence
621            of a directory in the file list provokes an error.
622    
623        (h) The documentation has been revised and clarified in places.
624    
625    9.  Improve the matching speed of capturing brackets.
626    
627    10. Changed the meaning of \X so that it now matches a Unicode extended
628        grapheme cluster.
629    
630    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
631        out POSIX threads when JIT support is configured.
632    
633    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
634    
635    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
636        structure could go wrong in environments where size_t is not the same size
637        as int.
638    
639    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
640        set.
641    
642    15. The EBCDIC support had decayed; later updates to the code had included
643        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
644        general tidy up of EBCDIC-related issues, and the documentation was also
645        not quite right. There is now a test that can be run on ASCII systems to
646        check some of the EBCDIC-related things (but is it not a full test).
647    
648    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
649        in a small tidy to the code.
650    
651    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
652    
653    18. If the --only-matching (-o) option in pcregrep is specified multiple
654        times, each one causes appropriate output. For example, -o1 -o2 outputs the
655        substrings matched by the 1st and 2nd capturing parentheses. A separating
656        string can be specified by --om-separator (default empty).
657    
658    19. Improving the first n character searches.
659    
660    20. Turn case lists for horizontal and vertical white space into macros so that
661        they are defined only once.
662    
663    21. This set of changes together give more compatible Unicode case-folding
664        behaviour for characters that have more than one other case when UCP
665        support is available.
666    
667        (a) The Unicode property table now has offsets into a new table of sets of
668            three or more characters that are case-equivalent. The MultiStage2.py
669            script that generates these tables (the pcre_ucd.c file) now scans
670            CaseFolding.txt instead of UnicodeData.txt for character case
671            information.
672    
673        (b) The code for adding characters or ranges of characters to a character
674            class has been abstracted into a generalized function that also handles
675            case-independence. In UTF-mode with UCP support, this uses the new data
676            to handle characters with more than one other case.
677    
678        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
679            whose other case is greater than 256 are now correctly matched
680            caselessly. Previously, the high codepoint matched the low one, but not
681            vice versa.
682    
683        (d) The processing of \h, \H, \v, and \ in character classes now makes use
684            of the new class addition function, using character lists defined as
685            macros alongside the case definitions of 20 above.
686    
687        (e) Caseless back references now work with characters that have more than
688            one other case.
689    
690        (f) General caseless matching of characters with more than one other case
691            is supported.
692    
693    22. Unicode character properties were updated from Unicode 6.2.0
694    
695    23. Improved CMake support under Windows. Patch by Daniel Richard G.
696    
697    24. Add support for 32-bit character strings, and UTF-32
698    
699    25. Major JIT compiler update (code refactoring and bugfixing).
700        Experimental Sparc 32 support is added.
701    
702    26. Applied a modified version of Daniel Richard G's patch to create
703        pcre.h.generic and config.h.generic by "make" instead of in the
704        PrepareRelease script.
705    
706    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
707        pcre_compile.c when checking for a zero character.
708    
709    28. Introducing a native interface for JIT. Through this interface, the compiled
710        machine code can be directly executed. The purpose of this interface is to
711        provide fast pattern matching, so several sanity checks are not performed.
712        However, feature tests are still performed. The new interface provides
713        1.4x speedup compared to the old one.
714    
715    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
716        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
717        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
718    
719    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
720        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
721        the "old" RFC 2279). Instead, it ended up passing a negative length to
722        pcre_exec().
723    
724    31. Add support for GCC's visibility feature to hide internal functions.
725    
726    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
727        "unknown -C option" after outputting 0 or 1.
728    
729    33. There is now support for generating a code coverage report for the test
730        suite in environments where gcc is the compiler and lcov is installed. This
731        is mainly for the benefit of the developers.
732    
733    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
734        unaddressable using valgrind annotations, allowing valgrind to detect
735        invalid memory accesses. This is mainly for the benefit of the developers.
736    
737    25. (*UTF) can now be used to start a pattern in any of the three libraries.
738    
739    26. Give configure error if --enable-cpp but no C++ compiler found.
740    
741    
742    Version 8.31 06-July-2012
743    -------------------------
744    
745    1.  Fixing a wrong JIT test case and some compiler warnings.
746    
747    2.  Removed a bashism from the RunTest script.
748    
749    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
750        to unsigned type, result still unsigned" that was given by an MS compiler
751        on encountering the code "-sizeof(xxx)".
752    
753    4.  Partial matching support is added to the JIT compiler.
754    
755    5.  Fixed several bugs concerned with partial matching of items that consist
756        of more than one character:
757    
758        (a) /^(..)\1/ did not partially match "aba" because checking references was
759            done on an "all or nothing" basis. This also applied to repeated
760            references.
761    
762        (b) \R did not give a hard partial match if \r was found at the end of the
763            subject.
764    
765        (c) \X did not give a hard partial match after matching one or more
766            characters at the end of the subject.
767    
768        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
769            a partial match for the string "\r".
770    
771        (e) When newline was set to CRLF, the metacharacter "." did not recognize
772            a partial match for a CR character at the end of the subject string.
773    
774    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
775        running pcretest, the text "(JIT)" added to the output whenever JIT is
776        actually used to run the match.
777    
778    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
779        or /S+[+] with a digit between 1 and 7.
780    
781    8.  OP_NOT now supports any UTF character not just single-byte ones.
782    
783    9.  (*MARK) control verb is now supported by the JIT compiler.
784    
785    10. The command "./RunTest list" lists the available tests without actually
786        running any of them. (Because I keep forgetting what they all are.)
787    
788    11. Add PCRE_INFO_MAXLOOKBEHIND.
789    
790    12. Applied a (slightly modified) user-supplied patch that improves performance
791        when the heap is used for recursion (compiled with --disable-stack-for-
792        recursion). Instead of malloc and free for each heap frame each time a
793        logical recursion happens, frames are retained on a chain and re-used where
794        possible. This sometimes gives as much as 30% improvement.
795    
796    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
797        call.
798    
799    14. As documented, (*COMMIT) is now confined to within a positive assertion.
800    
801    15. It is now possible to link pcretest with libedit as an alternative to
802        libreadline.
803    
804    16. (*COMMIT) control verb is now supported by the JIT compiler.
805    
806    17. The Unicode data tables have been updated to Unicode 6.1.0.
807    
808    18. Added --file-list option to pcregrep.
809    
810    19. Added binary file support to pcregrep, including the -a, --binary-files,
811        -I, and --text options.
812    
813    20. The madvise function is renamed for posix_madvise for QNX compatibility
814        reasons. Fixed by Giuseppe D'Angelo.
815    
816    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
817    
818    22. Changed the option for creating symbolic links for 16-bit man pages from
819        -s to -sf so that re-installing does not cause issues.
820    
821    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
822    
823    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
824        matches in certain environments (the workspace was not being correctly
825        retained). Also added to pcre_dfa_exec() a simple plausibility check on
826        some of the workspace data at the beginning of a restart.
827    
828    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
829        was not doing so when it should - probably a typo introduced by SVN 528
830        (change 8.10/14).
831    
832    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
833        \w+ when the character tables indicated that \x{c4} was a word character.
834        There were several related cases, all because the tests for doing a table
835        lookup were testing for characters less than 127 instead of 255.
836    
837    27. If a pattern contains capturing parentheses that are not used in a match,
838        their slots in the ovector are set to -1. For those that are higher than
839        any matched groups, this happens at the end of processing. In the case when
840        there were back references that the ovector was too small to contain
841        (causing temporary malloc'd memory to be used during matching), and the
842        highest capturing number was not used, memory off the end of the ovector
843        was incorrectly being set to -1. (It was using the size of the temporary
844        memory instead of the true size.)
845    
846    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
847        ovector size, it uses memory at the end of the block that it has got.
848    
849    29. Check for an overlong MARK name and give an error at compile time. The
850        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
851    
852    30. JIT compiler update.
853    
854    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
855        Rill for the patch.
856    
857    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
858    
859    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
860    
861    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
862        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
863        was enabled.
864    
865    35. Improve JIT code generation for greedy plus quantifier.
866    
867    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
868        1 to "aa" instead of to an empty string. The bug affected repeated groups
869        that could potentially match an empty string.
870    
871    37. Optimizing single character iterators in JIT.
872    
873    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
874        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
875        codepoints that are too big for the mode are faulted, and in a UTF mode,
876        disallowed codepoints are also faulted.
877    
878    39. If PCRE was compiled with UTF support, in three places in the DFA
879        matcher there was code that should only have been obeyed in UTF mode, but
880        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
881        processing when bytes with values greater than 127 were present. In 16-bit
882        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
883        both cases the values are those that cannot be the first data item in a UTF
884        character. The three items that might have provoked this were recursions,
885        possessively repeated groups, and atomic groups.
886    
887    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
888        and pcregrep, because some OS require shared objects to be explicitly
889        passed to ld, causing the link step to fail if they are not.
890    
891    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
892        mode, patterns that started with \h* or \R* might be incorrectly matched.
893    
894    
895    Version 8.30 04-February-2012
896    -----------------------------
897    
898    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
899        name is defined in ctype.h.
900    
901    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
902        only in quite long subpatterns.
903    
904    3.  Removed the function pcre_info(), which has been obsolete and deprecated
905        since it was replaced by pcre_fullinfo() in February 2000.
906    
907    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
908        match a (*MARK), and the match failed at the start of the subject, a
909        reference to memory before the start of the subject could occur. This bug
910        was introduced by fix 17 of release 8.21.
911    
912    5.  A reference to an unset group with zero minimum repetition was giving
913        totally wrong answers (in non-JavaScript-compatibility mode). For example,
914        /(another)?(\1?)test/ matched against "hello world test". This bug was
915        introduced in release 8.13.
916    
917    6.  Add support for 16-bit character strings (a large amount of work involving
918        many changes and refactorings).
919    
920    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
921        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
922        from a file.
923    
924    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
925        rounding is not applied in this particular case).
926    
927    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
928        if they appear, or are escaped, in patterns.
929    
930    10. Get rid of a number of -Wunused-but-set-variable warnings.
931    
932    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
933        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
934        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
935        also returns the mark "x". This bug applied to capturing parentheses,
936        non-capturing parentheses, and atomic parentheses. It also applied to some
937        assertions.
938    
939    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
940        information out of configure.ac instead of relying on pcre.h.generic, which
941        is not stored in the repository.
942    
943    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
944        -lreadline.
945    
946    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
947    
948    15. Applied Graycode's patch to put the top-level frame on the stack rather
949        than the heap when not using the stack for recursion. This gives a
950        performance improvement in many cases when recursion is not deep.
951    
952    16. Experimental code added to "pcretest -C" to output the stack frame size.
953    
954    
955    Version 8.21 12-Dec-2011
956    ------------------------
957    
958    1.  Updating the JIT compiler.
959    
960    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
961        are added as well.
962    
963    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
964        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
965        calling _pcre_jit_exec. Some extra comments are added.
966    
967    4.  (*MARK) settings inside atomic groups that do not contain any capturing
968        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
969        was introduced by change 18 for 8.20.
970    
971    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
972        ECMA-262 standard.
973    
974    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
975        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
976        This bug was probably introduced by change 9 of 8.13.
977    
978    7.  While fixing 6 above, I noticed that a number of other items were being
979        incorrectly rejected as "not fixed length". This arose partly because newer
980        opcodes had not been added to the fixed-length checking code. I have (a)
981        corrected the bug and added tests for these items, and (b) arranged for an
982        error to occur if an unknown opcode is encountered while checking for fixed
983        length instead of just assuming "not fixed length". The items that were
984        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
985        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
986        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
987    
988    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
989        being incorrectly compiled and would have given unpredicatble results.
990    
991    9.  A possessively repeated subpattern with minimum repeat count greater than
992        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
993        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
994        the first (A) could occur when it should not.
995    
996    10. Add a cast and remove a redundant test from the code.
997    
998    11. JIT should use pcre_malloc/pcre_free for allocation.
999    
1000    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1001        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1002        is anything other than /usr, -L is still shown).
1003    
1004    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1005    
1006    14. Perl does not support \N without a following name in a [] class; PCRE now
1007        also gives an error.
1008    
1009    15. If a forward reference was repeated with an upper limit of around 2000,
1010        it caused the error "internal error: overran compiling workspace". The
1011        maximum number of forward references (including repeats) was limited by the
1012        internal workspace, and dependent on the LINK_SIZE. The code has been
1013        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1014        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1015        of around 200,000 forward references. While doing this, I also speeded up
1016        the filling in of repeated forward references.
1017    
1018    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1019        incorrectly expecting the subject to contain another "a" after the start.
1020    
1021    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1022        in the match, the SKIP should be ignored. This was not happening; instead
1023        the SKIP was being treated as NOMATCH. For patterns such as
1024        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1025        tested.
1026    
1027    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1028        now much more compatible with Perl, in particular in cases where the result
1029        is a non-match for a non-anchored pattern. For example, if
1030        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1031        "m", where previously it did not return a name. A side effect of this
1032        change is that for partial matches, the last encountered mark name is
1033        returned, as for non matches. A number of tests that were previously not
1034        Perl-compatible have been moved into the Perl-compatible test files. The
1035        refactoring has had the pleasing side effect of removing one argument from
1036        the match() function, thus reducing its stack requirements.
1037    
1038    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1039        subsequent uses of /S (without +) incorrectly behaved like /S+.
1040    
1041    21. Retrieve executable code size support for the JIT compiler and fixing
1042        some warnings.
1043    
1044    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1045        not work when the shorter character appeared right at the end of the
1046        subject string.
1047    
1048    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1049        systems.
1050    
1051    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1052        output it when the /M option is used in pcretest.
1053    
1054    25. The CheckMan script was not being included in the distribution. Also, added
1055        an explicit "perl" to run Perl scripts from the PrepareRelease script
1056        because this is reportedly needed in Windows.
1057    
1058    26. If study data was being save in a file and studying had not found a set of
1059        "starts with" bytes for the pattern, the data written to the file (though
1060        never used) was taken from uninitialized memory and so caused valgrind to
1061        complain.
1062    
1063    27. Updated RunTest.bat as provided by Sheri Pierce.
1064    
1065    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1066    
1067    29. Computation of memory usage for the table of capturing group names was
1068        giving an unnecessarily large value.
1069    
1070    
1071    Version 8.20 21-Oct-2011
1072    ------------------------
1073    
1074    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1075        a POSIX class. After further experiments with Perl, which convinced me that
1076        Perl has bugs and confusions, a closing square bracket is no longer allowed
1077        in a POSIX name. This bug also affected patterns with classes that started
1078        with full stops.
1079    
1080    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1081        captured substring, but while checking the failing first alternative,
1082        substring 1 is temporarily captured. If the output vector supplied to
1083        pcre_exec() was not big enough for this capture, the yield of the function
1084        was still zero ("insufficient space for captured substrings"). This cannot
1085        be totally fixed without adding another stack variable, which seems a lot
1086        of expense for a edge case. However, I have improved the situation in cases
1087        such as /(a)(b)x|abc/ matched against "abc", where the return code
1088        indicates that fewer than the maximum number of slots in the ovector have
1089        been set.
1090    
1091    3.  Related to (2) above: when there are more back references in a pattern than
1092        slots in the output vector, pcre_exec() uses temporary memory during
1093        matching, and copies in the captures as far as possible afterwards. It was
1094        using the entire output vector, but this conflicts with the specification
1095        that only 2/3 is used for passing back captured substrings. Now it uses
1096        only the first 2/3, for compatibility. This is, of course, another edge
1097        case.
1098    
1099    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1100        main code base, and can be used by building with --enable-jit. When this is
1101        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1102        runtime --no-jit option is given.
1103    
1104    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1105        ovector, the return from the function was zero, implying that there were
1106        other matches that did not fit. The correct "exactly full" value is now
1107        returned.
1108    
1109    6.  If a subpattern that was called recursively or as a subroutine contained
1110        (*PRUNE) or any other control that caused it to give a non-standard return,
1111        invalid errors such as "Error -26 (nested recursion at the same subject
1112        position)" or even infinite loops could occur.
1113    
1114    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1115        computing the minimum length on reaching *ACCEPT, and so ended up with the
1116        wrong value of 1 rather than 0. Further investigation indicates that
1117        computing a minimum subject length in the presence of *ACCEPT is difficult
1118        (think back references, subroutine calls), and so I have changed the code
1119        so that no minimum is registered for a pattern that contains *ACCEPT.
1120    
1121    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1122        it was not handled as intended. [But see 16 below.]
1123    
1124    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1125        Sheri Pierce.
1126    
1127    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1128        the first byte in a match must be "a".
1129    
1130    11. Change 17 for 8.13 increased the recursion depth for patterns like
1131        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1132        pattern contains any instances of (*THEN). If it does not, the old
1133        optimizations are restored. It would be nice to do this on a per-group
1134        basis, but at the moment that is not feasible.
1135    
1136    12. In some environments, the output of pcretest -C is CRLF terminated. This
1137        broke RunTest's code that checks for the link size. A single white space
1138        character after the value is now allowed for.
1139    
1140    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1141        For "fr", it uses the Windows-specific input and output files.
1142    
1143    14. If (*THEN) appeared in a group that was called recursively or as a
1144        subroutine, it did not work as intended. [But see next item.]
1145    
1146    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1147        pattern fragments (but not containing any | characters). If A and B are
1148        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1149        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1150        D. In other words, Perl considers parentheses that do not contain any |
1151        characters to be part of a surrounding alternative, whereas PCRE was
1152        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1153        differently. PCRE now behaves in the same way as Perl, except in the case
1154        of subroutine/recursion calls such as (?1) which have in any case always
1155        been different (but PCRE had them first :-).
1156    
1157    16. Related to 15 above: Perl does not treat the | in a conditional group as
1158        creating alternatives. Such a group is treated in the same way as an
1159        ordinary group without any | characters when processing (*THEN). PCRE has
1160        been changed to match Perl's behaviour.
1161    
1162    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1163        RunGrepTest script failed.
1164    
1165    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1166        inevitable for groups that contain captures, but it can lead to a lot of
1167        stack use in large patterns. The old behaviour has been restored for atomic
1168        groups that do not contain any capturing parentheses.
1169    
1170    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1171        suppress the check for a minimum subject length at run time. (If it was
1172        given to pcre_exec() or pcre_dfa_exec() it did work.)
1173    
1174    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1175        fail to work when decoding hex characters in data strings in EBCDIC
1176        environments.
1177    
1178    21. It appears that in at least one Mac OS environment, the isxdigit() function
1179        is implemented as a macro that evaluates to its argument more than once,
1180        contravening the C 90 Standard (I haven't checked a later standard). There
1181        was an instance in pcretest which caused it to go wrong when processing
1182        \x{...} escapes in subject strings. The has been rewritten to avoid using
1183        things like p++ in the argument of isxdigit().
1184    
1185    
1186    Version 8.13 16-Aug-2011
1187    ------------------------
1188    
1189    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1190    
1191    2.  Two minor typos in pcre_internal.h have been fixed.
1192    
1193    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1194        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1195        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1196    
1197    4.  There were a number of related bugs in the code for matching backrefences
1198        caselessly in UTF-8 mode when codes for the characters concerned were
1199        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1200        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1201        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1202        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1203        2-byte code at the end of the subject (it thought there wasn't enough data
1204        left).
1205    
1206    5.  Comprehensive information about what went wrong is now returned by
1207        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1208        as the output vector has at least 2 elements. The offset of the start of
1209        the failing character and a reason code are placed in the vector.
1210    
1211    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1212        now returned is for the first byte of the failing character, instead of the
1213        last byte inspected. This is an incompatible change, but I hope it is small
1214        enough not to be a problem. It makes the returned offset consistent with
1215        pcre_exec() and pcre_dfa_exec().
1216    
1217    7.  pcretest now gives a text phrase as well as the error number when
1218        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1219        failure, the offset and reason code are output.
1220    
1221    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1222        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1223        back over a single character (\n). This seems wrong (because it treated the
1224        two characters as a single entity when going forwards), conflicts with the
1225        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1226        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1227        has been changed.
1228    
1229    9.  Some internal refactoring has changed the processing so that the handling
1230        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1231        time (the PCRE_DOTALL option was changed this way some time ago: version
1232        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1233        which was always a bit of a fudge. It also means that there is one less
1234        argument for the match() function, which reduces its stack requirements
1235        slightly. This change also fixes an incompatibility with Perl: the pattern
1236        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1237    
1238    10. More internal refactoring has drastically reduced the number of recursive
1239        calls to match() for possessively repeated groups such as (abc)++ when
1240        using pcre_exec().
1241    
1242    11. While implementing 10, a number of bugs in the handling of groups were
1243        discovered and fixed:
1244    
1245        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1246        (a|)*(?1) gave a compile-time internal error.
1247        ((a|)+)+  did not notice that the outer group could match an empty string.
1248        (^a|^)+   was not marked as anchored.
1249        (.*a|.*)+ was not marked as matching at start or after a newline.
1250    
1251    12. Yet more internal refactoring has removed another argument from the match()
1252        function. Special calls to this function are now indicated by setting a
1253        value in a variable in the "match data" data block.
1254    
1255    13. Be more explicit in pcre_study() instead of relying on "default" for
1256        opcodes that mean there is no starting character; this means that when new
1257        ones are added and accidentally left out of pcre_study(), testing should
1258        pick them up.
1259    
1260    14. The -s option of pcretest has been documented for ages as being an old
1261        synonym of -m (show memory usage). I have changed it to mean "force study
1262        for every regex", that is, assume /S for every regex. This is similar to -i
1263        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1264        using it. It makes it easier to run collections of tests with and without
1265        study enabled, and thereby test pcre_study() more easily. All the standard
1266        tests are now run with and without -s (but some patterns can be marked as
1267        "never study" - see 20 below).
1268    
1269    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1270        restoration of the capturing data to the outer values was not happening
1271        correctly.
1272    
1273    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1274        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1275        pattern had matched an empty string, and so incorrectly returned a no
1276        match.
1277    
1278    17. There was optimizing code for the last branch of non-capturing parentheses,
1279        and also for the obeyed branch of a conditional subexpression, which used
1280        tail recursion to cut down on stack usage. Unfortunately, now that there is
1281        the possibility of (*THEN) occurring in these branches, tail recursion is
1282        no longer possible because the return has to be checked for (*THEN). These
1283        two optimizations have therefore been removed. [But see 8.20/11 above.]
1284    
1285    18. If a pattern containing \R was studied, it was assumed that \R always
1286        matched two bytes, thus causing the minimum subject length to be
1287        incorrectly computed because \R can also match just one byte.
1288    
1289    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1290        was incorrectly computed.
1291    
1292    20. If /S is present twice on a test pattern in pcretest input, it now
1293        *disables* studying, thereby overriding the use of -s on the command line
1294        (see 14 above). This is necessary for one or two tests to keep the output
1295        identical in both cases.
1296    
1297    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1298        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1299    
1300    22. When an atomic group that contained a capturing parenthesis was
1301        successfully matched, but the branch in which it appeared failed, the
1302        capturing was not being forgotten if a higher numbered group was later
1303        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1304        group 1 to "a", when in fact it should be unset. This applied to multi-
1305        branched capturing and non-capturing groups, repeated or not, and also to
1306        positive assertions (capturing in negative assertions does not happen
1307        in PCRE) and also to nested atomic groups.
1308    
1309    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1310        subject after a captured substring, to make it easier to tell which of a
1311        number of identical substrings has been captured.
1312    
1313    24. The way atomic groups are processed by pcre_exec() has been changed so that
1314        if they are repeated, backtracking one repetition now resets captured
1315        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1316        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1317        "aaa". Previously, it would have been "a". As part of this code
1318        refactoring, the way recursive calls are handled has also been changed.
1319    
1320    25. If an assertion condition captured any substrings, they were not passed
1321        back unless some other capturing happened later. For example, if
1322        (?(?=(a))a) was matched against "a", no capturing was returned.
1323    
1324    26. When studying a pattern that contained subroutine calls or assertions,
1325        the code for finding the minimum length of a possible match was handling
1326        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1327        group 1 called group 2 while simultaneously a separate group 2 called group
1328        1). A stack overflow occurred in this case. I have fixed this by limiting
1329        the recursion depth to 10.
1330    
1331    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1332        Fortmann. This supports explicit test numbers on the command line, and has
1333        argument validation and error reporting.
1334    
1335    28. An instance of \X with an unlimited repeat could fail if at any point the
1336        first character it looked at was a mark character.
1337    
1338    29. Some minor code refactoring concerning Unicode properties and scripts
1339        should reduce the stack requirement of match() slightly.
1340    
1341    30. Added the '=' option to pcretest to check the setting of unused capturing
1342        slots at the end of the pattern, which are documented as being -1, but are
1343        not included in the return count.
1344    
1345    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1346        compiled something random. Now it gives a compile-time error (as does
1347        Perl).
1348    
1349    32. A *MARK encountered during the processing of a positive assertion is now
1350        recorded and passed back (compatible with Perl).
1351    
1352    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1353        had alternative anchored branches, the search for a second match in a line
1354        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1355        matched the line "0102" twice. The same bug affected patterns that started
1356        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1357        twice.
1358    
1359    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1360        does, and because of capturing effects, quantifying parenthesized
1361        assertions may at times be useful. Quantifiers are now allowed for
1362        parenthesized assertions.
1363    
1364    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1365    
1366    36. \g was being checked for fancy things in a character class, when it should
1367        just be a literal "g".
1368    
1369    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1370        appearance of a nested POSIX class supersedes an apparent external class.
1371        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1372        unescaped square brackets may also appear as part of class names. For
1373        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1374        more like Perl. (But see 8.20/1 above.)
1375    
1376    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1377        was because it thought it was \N{name}, which is not supported).
1378    
1379    39. Add minix to OS list not supporting the -S option in pcretest.
1380    
1381    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1382        cannot analyze patterns in sufficient detail to catch mutual recursions
1383        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1384        subgroup is called recursively as a subpattern for a second time at the
1385        same position in the subject string. In previous releases this might have
1386        been caught by the recursion limit, or it might have run out of stack.
1387    
1388    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1389        happen only once. PCRE was, however incorrectly giving a compile time error
1390        "recursive call could loop indefinitely" because it cannot analyze the
1391        pattern in sufficient detail. The compile time test no longer happens when
1392        PCRE is compiling a conditional subpattern, but actual runaway loops are
1393        now caught at runtime (see 40 above).
1394    
1395    42. It seems that Perl allows any characters other than a closing parenthesis
1396        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1397        has been changed to be the same.
1398    
1399    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1400        as not to get warnings when autogen.sh is called. Also changed
1401        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1402    
1403    44. To help people who use pcregrep to scan files containing exceedingly long
1404        lines, the following changes have been made:
1405    
1406        (a) The default value of the buffer size parameter has been increased from
1407            8K to 20K. (The actual buffer used is three times this size.)
1408    
1409        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1410            PCRE is built.
1411    
1412        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1413            to be set at run time.
1414    
1415        (d) Numerical values in pcregrep options can be followed by K or M, for
1416            example --buffer-size=50K.
1417    
1418        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1419            given and the return code is set to 2.
1420    
1421    45. Add a pointer to the latest mark to the callout data block.
1422    
1423    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1424        partial match of an empty string instead of no match. This was specific to
1425        the use of ".".
1426    
1427    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1428        complete match instead of a partial match. This bug was dependent on both
1429        the PCRE_UTF8 and PCRE_DOTALL options being set.
1430    
1431    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1432        starting byte set, because \b was not being ignored.
1433    
1434    
1435    Version 8.12 15-Jan-2011
1436    ------------------------
1437    
1438    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1439        checks for such things as part of the documentation building process.
1440    
1441    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1442        --match-limit and --recursion-limit options (added for 8.11). In
1443        particular, this made one of the standard tests fail. (The integer value
1444        went into the wrong half of a long int.)
1445    
1446    3.  If the --colour option was given to pcregrep with -v (invert match), it
1447        did strange things, either producing crazy output, or crashing. It should,
1448        of course, ignore a request for colour when reporting lines that do not
1449        match.
1450    
1451    4.  Another pcregrep bug caused similar problems if --colour was specified with
1452        -M (multiline) and the pattern match finished with a line ending.
1453    
1454    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1455        matched in multiline mode, the following line was shown as part of the
1456        match. This seems wrong, so I have changed it.
1457    
1458    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1459        the check for further matches in the same line (so they could be coloured)
1460        to overrun the end of the current line. If another match was found, it was
1461        incorrectly shown (and then shown again when found in the next line).
1462    
1463    7.  If pcregrep was compiled under Windows, there was a reference to the
1464        function pcregrep_exit() before it was defined. I am assuming this was
1465        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1466        reported by a user. I've moved the definition above the reference.
1467    
1468    
1469    Version 8.11 10-Dec-2010
1470    ------------------------
1471    
1472    1.  (*THEN) was not working properly if there were untried alternatives prior
1473        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1474        backtracked to try for "b" instead of moving to the next alternative branch
1475        at the same level (in this case, to look for "c"). The Perl documentation
1476        is clear that when (*THEN) is backtracked onto, it goes to the "next
1477        alternative in the innermost enclosing group".
1478    
1479    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1480        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1481        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1482        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1483        (*THEN).
1484    
1485    3.  If \s appeared in a character class, it removed the VT character from
1486        the class, even if it had been included by some previous item, for example
1487        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1488        of \s, but is part of the POSIX "space" class.)
1489    
1490    4.  A partial match never returns an empty string (because you can always
1491        match an empty string at the end of the subject); however the checking for
1492        an empty string was starting at the "start of match" point. This has been
1493        changed to the "earliest inspected character" point, because the returned
1494        data for a partial match starts at this character. This means that, for
1495        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1496        (previously it gave "no match").
1497    
1498    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1499        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1500        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1501        has an implication that the given string is incomplete (because a partial
1502        match is preferred over a full match). For this reason, these items now
1503        give a partial match in this situation. [Aside: previously, the one case
1504        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1505        match rather than a full match, which was wrong by the old rules, but is
1506        now correct.]
1507    
1508    6.  There was a bug in the handling of #-introduced comments, recognized when
1509        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1510        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1511        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1512        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1513        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1514        places in pcre_compile().
1515    
1516    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1517        comments when looking ahead for named forward references to subpatterns,
1518        the only newline sequence it recognized was NL. It now handles newlines
1519        according to the set newline convention.
1520    
1521    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1522        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1523        cater for a lack of strerror(). These oversights have been fixed.
1524    
1525    9.  Added --match-limit and --recursion-limit to pcregrep.
1526    
1527    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1528    
1529    11. When the -o option was used, pcregrep was setting a return code of 1, even
1530        when matches were found, and --line-buffered was not being honoured.
1531    
1532    12. Added an optional parentheses number to the -o and --only-matching options
1533        of pcregrep.
1534    
1535    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1536        can match an empty string. The code to do it in pcretest and pcredemo
1537        needed fixing:
1538    
1539        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1540            only one byte after an empty string match just before CRLF (this case
1541            just got forgotten; "any" and "anycrlf" were OK).
1542    
1543        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1544            mode when an empty string match preceded an ASCII character followed by
1545            a non-ASCII character. (The code for advancing by one character rather
1546            than one byte was nonsense.)
1547    
1548        (c) The pcredemo.c sample program did not have any code at all to handle
1549            the cases when CRLF is a valid newline sequence.
1550    
1551    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1552        as a starting offset was within the subject string. There is now a new
1553        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1554        negative or greater than the length of the string. In order to test this,
1555        pcretest is extended to allow the setting of negative starting offsets.
1556    
1557    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1558        starting offset points to the beginning of a UTF-8 character was
1559        unnecessarily clumsy. I tidied it up.
1560    
1561    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1562        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1563    
1564    17. Nobody had reported that the --include_dir option, which was added in
1565        release 7.7 should have been called --include-dir (hyphen, not underscore)
1566        for compatibility with GNU grep. I have changed it to --include-dir, but
1567        left --include_dir as an undocumented synonym, and the same for
1568        --exclude-dir, though that is not available in GNU grep, at least as of
1569        release 2.5.4.
1570    
1571    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1572        characters from a string of bytes) have been redefined so as not to use
1573        loops, in order to improve performance in some environments. At the same
1574        time, I abstracted some of the common code into auxiliary macros to save
1575        repetition (this should not affect the compiled code).
1576    
1577    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1578        compile-time error is now given if \c is not followed by an ASCII
1579        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1580        different, and any byte value is allowed.)
1581    
1582    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1583        START_OPTIMIZE option, which is now allowed at compile time - but just
1584        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1585        to pcregrep and other applications that have no direct access to PCRE
1586        options. The new /Y option in pcretest sets this option when calling
1587        pcre_compile().
1588    
1589    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1590        back references. Groups containing recursive back references were forced to
1591        be atomic by that change, but in the case of named groups, the amount of
1592        memory required was incorrectly computed, leading to "Failed: internal
1593        error: code overflow". This has been fixed.
1594    
1595    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1596        pcretest.c, to avoid build problems in some Borland environments.
1597    
1598    
1599    Version 8.10 25-Jun-2010
1600    ------------------------
1601    
1602    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1603        THEN.
1604    
1605    2.  (*ACCEPT) was not working when inside an atomic group.
1606    
1607    3.  Inside a character class, \B is treated as a literal by default, but
1608        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1609        causes the error). The code is unchanged, but I tidied the documentation.
1610    
1611    4.  Inside a character class, PCRE always treated \R and \X as literals,
1612        whereas Perl faults them if its -w option is set. I have changed PCRE so
1613        that it faults them when PCRE_EXTRA is set.
1614    
1615    5.  Added support for \N, which always matches any character other than
1616        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1617    
1618    6.  When compiling pcregrep with newer versions of gcc which may have
1619        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1620        declared with attribute warn_unused_result" were given. Just casting the
1621        result to (void) does not stop the warnings; a more elaborate fudge is
1622        needed. I've used a macro to implement this.
1623    
1624    7.  Minor change to pcretest.c to avoid a compiler warning.
1625    
1626    8.  Added four artifical Unicode properties to help with an option to make
1627        \s etc use properties (see next item). The new properties are: Xan
1628        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1629    
1630    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1631        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1632        this option. Modified pcretest to add /W to test this facility. Added
1633        REG_UCP to make it available via the POSIX interface.
1634    
1635    10. Added --line-buffered to pcregrep.
1636    
1637    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1638        studied, and the match started with a letter with a code point greater than
1639        127 whose first byte was different to the first byte of the other case of
1640        the letter, the other case of this starting letter was not recognized
1641        (#976).
1642    
1643    12. If a pattern that was studied started with a repeated Unicode property
1644        test, for example, \p{Nd}+, there was the theoretical possibility of
1645        setting up an incorrect bitmap of starting bytes, but fortunately it could
1646        not have actually happened in practice until change 8 above was made (it
1647        added property types that matched character-matching opcodes).
1648    
1649    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1650        possible starting bytes for non-anchored patterns.
1651    
1652    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1653        \R, and also a number of cases that involve Unicode properties, both
1654        explicit and implicit when PCRE_UCP is set.
1655    
1656    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1657        input, it could crash or give wrong results if characters with values
1658        greater than 0xc0 were present in the subject string. (Detail: it assumed
1659        UTF-8 input when processing these items.)
1660    
1661    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1662        size_t is 64-bit (#991).
1663    
1664    17. Added a check for running out of memory when PCRE is compiled with
1665        --disable-stack-for-recursion (#990).
1666    
1667    18. If the last data line in a file for pcretest does not have a newline on
1668        the end, a newline was missing in the output.
1669    
1670    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1671        less than 128) in its various bitmaps. However, there is a facility for
1672        generating tables according to the current locale when PCRE is compiled. It
1673        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1674        characters, are recognized by isspace() and therefore were getting set in
1675        these tables, and indeed these tables seem to approximate to ISO 8859. This
1676        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1677        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1678        which of course cannot start UTF-8 characters. I have changed the code so
1679        that only real ASCII characters (less than 128) and the correct starting
1680        bytes for UTF-8 encodings are set for characters greater than 127 when in
1681        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1682        altogether.)
1683    
1684    20. Added the /T option to pcretest so as to be able to run tests with non-
1685        standard character tables, thus making it possible to include the tests
1686        used for 19 above in the standard set of tests.
1687    
1688    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1689        reference to a subpattern the other side of a comment that contains an
1690        opening parenthesis caused either an internal compiling error, or a
1691        reference to the wrong subpattern.
1692    
1693    
1694    Version 8.02 19-Mar-2010
1695    ------------------------
1696    
1697    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1698    
1699    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1700        configured.
1701    
1702    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1703        original author of that file, following a query about its status.
1704    
1705    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1706        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1707    
1708    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1709        quantifier applied to a forward-referencing subroutine call, could compile
1710        incorrect code or give the error "internal error: previously-checked
1711        referenced subpattern not found".
1712    
1713    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1714        variables to point to external functions. For these systems, therefore,
1715        pcre_malloc etc. are now initialized to local functions that call the
1716        relevant global functions.
1717    
1718    7.  There were two entries missing in the vectors called coptable and poptable
1719        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1720        I've fixed the data, and added a kludgy way of testing at compile time that
1721        the lengths are correct (equal to the number of opcodes).
1722    
1723    8.  Following on from 7, I added a similar kludge to check the length of the
1724        eint vector in pcreposix.c.
1725    
1726    9.  Error texts for pcre_compile() are held as one long string to avoid too
1727        much relocation at load time. To find a text, the string is searched,
1728        counting zeros. There was no check for running off the end of the string,
1729        which could happen if a new error number was added without updating the
1730        string.
1731    
1732    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1733    
1734    11. \K was not working if it appeared in an atomic group or in a group that
1735        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1736        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1737        the assertion is positive, but not if it is negative.
1738    
1739    12. Change 11 fortuitously reduced the size of the stack frame used in the
1740        "match()" function of pcre_exec.c by one pointer. Forthcoming
1741        implementation of support for (*MARK) will need an extra pointer on the
1742        stack; I have reserved it now, so that the stack frame size does not
1743        decrease.
1744    
1745    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1746        item in branch that calls a recursion is a subroutine call - as in the
1747        second branch in the above example - was incorrectly given the compile-
1748        time error "recursive call could loop indefinitely" because pcre_compile()
1749        was not correctly checking the subroutine for matching a non-empty string.
1750    
1751    14. The checks for overrunning compiling workspace could trigger after an
1752        overrun had occurred. This is a "should never occur" error, but it can be
1753        triggered by pathological patterns such as hundreds of nested parentheses.
1754        The checks now trigger 100 bytes before the end of the workspace.
1755    
1756    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1757    
1758    
1759    Version 8.01 19-Jan-2010
1760    ------------------------
1761    
1762    1.  If a pattern contained a conditional subpattern with only one branch (in
1763        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1764        computed the wrong minimum data length (which is of course zero for such
1765        subpatterns). This could cause incorrect "no match" results.
1766    
1767    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1768        the pattern is reset in the first branch, pcre_compile() failed with
1769        "internal error: code overflow at offset...". This happened only when
1770        the reset was to the original external option setting. (An optimization
1771        abstracts leading options settings into an external setting, which was the
1772        cause of this.)
1773    
1774    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1775        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1776        assertion pattern did not match (meaning that the assertion was true), it
1777        was incorrectly treated as false if the SKIP had been reached during the
1778        matching. This also applied to assertions used as conditions.
1779    
1780    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1781        assertion subpattern, including such a pattern used as a condition,
1782        unpredictable results occurred, instead of the error return
1783        PCRE_ERROR_DFA_UITEM.
1784    
1785    5.  The C++ GlobalReplace function was not working like Perl for the special
1786        situation when an empty string is matched. It now does the fancy magic
1787        stuff that is necessary.
1788    
1789    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1790        removed. (These were left over from very, very early versions of PCRE.)
1791    
1792    7.  Some cosmetic changes to the code to make life easier when compiling it
1793        as part of something else:
1794    
1795        (a) Change DEBUG to PCRE_DEBUG.
1796    
1797        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1798            called "current" as "current_branch", to prevent a collision with the
1799            Linux macro when compiled as a kernel module.
1800    
1801        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1802            prevent a collision with the Linux macro when compiled as a kernel
1803            module.
1804    
1805    8.  In pcre_compile() there are some checks for integer overflows that used to
1806        cast potentially large values to (double). This has been changed to that
1807        when building, a check for int64_t is made, and if it is found, it is used
1808        instead, thus avoiding the use of floating point arithmetic. (There is no
1809        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1810        double.
1811    
1812    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1813        2005 (difference between two addresses compared to an unsigned value).
1814    
1815    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1816        custom one, because of the following reported problem in Windows:
1817    
1818          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1819              under Win32.
1820          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1821              therefore missing the function definition.
1822          - The compiler thus generates a "C" signature for the test function.
1823          - The linker fails to find the "C" function.
1824          - PCRE fails to configure if asked to do so against libbz2.
1825    
1826    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1827        messages were output:
1828    
1829          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1830          rerunning libtoolize, to keep the correct libtool macros in-tree.
1831          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1832    
1833        I have done both of these things.
1834    
1835    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1836        most of the time, it *can* run out if it is given a pattern that contains a
1837        runaway infinite recursion. I updated the discussion in the pcrestack man
1838        page.
1839    
1840    13. Now that we have gone to the x.xx style of version numbers, the minor
1841        version may start with zero. Using 08 or 09 is a bad idea because users
1842        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1843        interpreted as invalid octal numbers. I've updated the previous comment in
1844        configure.ac, and also added a check that gives an error if 08 or 09 are
1845        used.
1846    
1847    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1848        causing partial matching to fail when the end of the subject matched \W
1849        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1850    
1851    15. There were some discrepancies between the declarations in pcre_internal.h
1852        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1853        their definitions. The declarations used "const uschar *" and the
1854        definitions used USPTR. Even though USPTR is normally defined as "const
1855        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1856        reported that: "This difference in casting confuses some C++ compilers, for
1857        example, SunCC recognizes above declarations as different functions and
1858        generates broken code for hbpcre." I have changed the declarations to use
1859        USPTR.
1860    
1861    16. GNU libtool is named differently on some systems. The autogen.sh script now
1862        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1863        (FreeBSD).
1864    
1865    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1866        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1867        comment: "Figure out how to create a longlong from a string: strtoll and
1868        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1869        instance, but it only takes 2 args instead of 3!"
1870    
1871    18. A subtle bug concerned with back references has been fixed by a change of
1872        specification, with a corresponding code fix. A pattern such as
1873        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1874        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1875        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1876        same bug. Such groups have to be quantified to be useful, or contained
1877        inside another quantified group. (If there's no repetition, the reference
1878        can never match.) The problem arises because, having left the group and
1879        moved on to the rest of the pattern, a later failure that backtracks into
1880        the group uses the captured value from the final iteration of the group
1881        rather than the correct earlier one. I have fixed this in PCRE by forcing
1882        any group that contains a reference to itself to be an atomic group; that
1883        is, there cannot be any backtracking into it once it has completed. This is
1884        similar to recursive and subroutine calls.
1885    
1886    
1887    Version 8.00 19-Oct-09
1888    ----------------------
1889    
1890    1.  The table for translating pcre_compile() error codes into POSIX error codes
1891        was out-of-date, and there was no check on the pcre_compile() error code
1892        being within the table. This could lead to an OK return being given in
1893        error.
1894    
1895    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1896        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1897        in a Windows environment.
1898    
1899    3.  The pcregrep --count option prints the count for each file even when it is
1900        zero, as does GNU grep. However, pcregrep was also printing all files when
1901        --files-with-matches was added. Now, when both options are given, it prints
1902        counts only for those files that have at least one match. (GNU grep just
1903        prints the file name in this circumstance, but including the count seems
1904        more useful - otherwise, why use --count?) Also ensured that the
1905        combination -clh just lists non-zero counts, with no names.
1906    
1907    4.  The long form of the pcregrep -F option was incorrectly implemented as
1908        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1909        but it seems right to fix it, and I didn't think it was worth preserving
1910        the old behaviour.
1911    
1912    5.  The command line items --regex=pattern and --regexp=pattern were not
1913        recognized by pcregrep, which required --regex pattern or --regexp pattern
1914        (with a space rather than an '='). The man page documented the '=' forms,
1915        which are compatible with GNU grep; these now work.
1916    
1917    6.  No libpcreposix.pc file was created for pkg-config; there was just
1918        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1919    
1920    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1921        when UCP support is not needed, by modifying the Python script that
1922        generates it from Unicode data files. This should not matter if the module
1923        is correctly used as a library, but I received one complaint about 50K of
1924        unwanted data. My guess is that the person linked everything into his
1925        program rather than using a library. Anyway, it does no harm.
1926    
1927    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1928        was a minimum greater than 1 for a wide character in a possessive
1929        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1930        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1931        character. Chaos in the form of incorrect output or a compiling loop could
1932        result.
1933    
1934    9.  The restrictions on what a pattern can contain when partial matching is
1935        requested for pcre_exec() have been removed. All patterns can now be
1936        partially matched by this function. In addition, if there are at least two
1937        slots in the offset vector, the offset of the earliest inspected character
1938        for the match and the offset of the end of the subject are set in them when
1939        PCRE_ERROR_PARTIAL is returned.
1940    
1941    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1942        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1943        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1944        and may be more useful for multi-segment matching.
1945    
1946    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1947        used to be given if ever the end of the subject was reached; now it is
1948        given only if matching could not proceed because another character was
1949        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1950        string "Z", which now yields "no match" instead of "partial match". In the
1951        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1952        final character ended with (*FAIL).
1953    
1954    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1955        if the pattern had a "must contain" character that was already found in the
1956        earlier partial match, unless partial matching was again requested. For
1957        example, with the pattern /dog.(body)?/, the "must contain" character is
1958        "g". If the first part-match was for the string "dog", restarting with
1959        "sbody" failed. This bug has been fixed.
1960    
1961    13. The string returned by pcre_dfa_exec() after a partial match has been
1962        changed so that it starts at the first inspected character rather than the
1963        first character of the match. This makes a difference only if the pattern
1964        starts with a lookbehind assertion or \b or \B (\K is not supported by
1965        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1966        matching functions compatible, and I think it's the right thing to do.
1967    
1968    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1969        so that the demonstration program is easily available in environments where
1970        PCRE has not been installed from source.
1971    
1972    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1973        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1974        library.
1975    
1976    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1977        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1978        is not the first non-POSIX option to be added. Clearly some people find
1979        these options useful.
1980    
1981    17. If a caller to the POSIX matching function regexec() passes a non-zero
1982        value for nmatch with a NULL value for pmatch, the value of
1983        nmatch is forced to zero.
1984    
1985    18. RunGrepTest did not have a test for the availability of the -u option of
1986        the diff command, as RunTest does. It now checks in the same way as
1987        RunTest, and also checks for the -b option.
1988    
1989    19. If an odd number of negated classes containing just a single character
1990        interposed, within parentheses, between a forward reference to a named
1991        subpattern and the definition of the subpattern, compilation crashed with
1992        an internal error, complaining that it could not find the referenced
1993        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1994        [The bug was that it was starting one character too far in when skipping
1995        over the character class, thus treating the ] as data rather than
1996        terminating the class. This meant it could skip too much.]
1997    
1998    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1999        /g option in pcretest when the pattern contains \K, which makes it possible
2000        to have an empty string match not at the start, even when the pattern is
2001        anchored. Updated pcretest and pcredemo to use this option.
2002    
2003    21. If the maximum number of capturing subpatterns in a recursion was greater
2004        than the maximum at the outer level, the higher number was returned, but
2005        with unset values at the outer level. The correct (outer level) value is
2006        now given.
2007    
2008    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2009        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2010        make it do so. The string so far is captured, making this feature
2011        compatible with Perl.
2012    
2013    23. The tests have been re-organized, adding tests 11 and 12, to make it
2014        possible to check the Perl 5.10 features against Perl 5.10.
2015    
2016    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2017        pattern matches a fixed length string. PCRE did not allow this; now it
2018        does. Neither allows recursion.
2019    
2020    25. I finally figured out how to implement a request to provide the minimum
2021        length of subject string that was needed in order to match a given pattern.
2022        (It was back references and recursion that I had previously got hung up
2023        on.) This code has now been added to pcre_study(); it finds a lower bound
2024        to the length of subject needed. It is not necessarily the greatest lower
2025        bound, but using it to avoid searching strings that are too short does give
2026        some useful speed-ups. The value is available to calling programs via
2027        pcre_fullinfo().
2028    
2029    26. While implementing 25, I discovered to my embarrassment that pcretest had
2030        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2031        study optimizations had never been tested with that matching function.
2032        Oops. What is worse, even when it was passed study data, there was a bug in
2033        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2034        were also very few tests of studied patterns with pcre_dfa_exec().
2035    
2036    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2037        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2038        on the other side of the coin, they are no longer allowed to have different
2039        names, because these cannot be distinguished in PCRE, and this has caused
2040        confusion. (This is a difference from Perl.)
2041    
2042    28. When duplicate subpattern names are present (necessarily with different
2043        numbers, as required by 27 above), and a test is made by name in a
2044        conditional pattern, either for a subpattern having been matched, or for
2045        recursion in such a pattern, all the associated numbered subpatterns are
2046        tested, and the overall condition is true if the condition is true for any
2047        one of them. This is the way Perl works, and is also more like the way
2048        testing by number works.
2049    
2050    
2051    Version 7.9 11-Apr-09
2052    ---------------------
2053    
2054    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2055        (pcretest), all targets were linked against these libraries. This included
2056        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2057        libraries. This caused unwanted dependencies to be created. This problem
2058        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2059        pcretest is linked with readline.
2060    
2061    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2062        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2063        moved outside it again, because FALSE and TRUE are already defined in AIX,
2064        but BOOL is not.
2065    
2066    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2067        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2068    
2069    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2070        hyphens) following file names and line numbers when outputting matching
2071        lines. This is not true; no spaces are inserted. I have also clarified the
2072        wording for the --colour (or --color) option.
2073    
2074    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2075        was not coloured; this is different to GNU grep, so I have changed it to be
2076        the same.
2077    
2078    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2079        each matching line was coloured. Now it goes on to look for further matches
2080        of any of the test patterns, which is the same behaviour as GNU grep.
2081    
2082    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2083        doesn't make sense to accept an empty string match in pcregrep, so I have
2084        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2085        seems to be how GNU grep behaves. [But see later change 40 for release
2086        8.33.]
2087    
2088    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2089        start or after a newline", because the conditional assertion was not being
2090        correctly handled. The rule now is that both the assertion and what follows
2091        in the first alternative must satisfy the test.
2092    
2093    9.  If auto-callout was enabled in a pattern with a conditional group whose
2094        condition was an assertion, PCRE could crash during matching, both with
2095        pcre_exec() and pcre_dfa_exec().
2096    
2097    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2098        used for matching.
2099    
2100    11. Unicode property support in character classes was not working for
2101        characters (bytes) greater than 127 when not in UTF-8 mode.
2102    
2103    12. Added the -M command line option to pcretest.
2104    
2105    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2106    
2107    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2108    
2109    16. Added comments and documentation about mis-use of no_arg in the C++
2110        wrapper.
2111    
2112    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2113        from Martin Jerabek that uses macro names for all relevant character and
2114        string constants.
2115    
2116    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2117        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2118        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2119        these, but not everybody uses configure.
2120    
2121    19. A conditional group that had only one branch was not being correctly
2122        recognized as an item that could match an empty string. This meant that an
2123        enclosing group might also not be so recognized, causing infinite looping
2124        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2125        with the subject "ab", where knowledge that the repeated group can match
2126        nothing is needed in order to break the loop.
2127    
2128    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2129        exec(), but without supplying a callout function, matching went wrong.
2130    
2131    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2132        leak if the size of the offset vector was greater than 30. When the vector
2133        is smaller, the saved offsets during recursion go onto a local stack
2134        vector, but for larger vectors malloc() is used. It was failing to free
2135        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2136        error, in fact).
2137    
2138    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2139        heapframe that is used only when UTF-8 support is enabled. This caused no
2140        problem, but was untidy.
2141    
2142    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2143        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2144        included within another project.
2145    
2146    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2147        slightly modified by me:
2148    
2149          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2150              not building pcregrep.
2151    
2152          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2153              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2154    
2155    25. Forward references, both numeric and by name, in patterns that made use of
2156        duplicate group numbers, could behave incorrectly or give incorrect errors,
2157        because when scanning forward to find the reference group, PCRE was not
2158        taking into account the duplicate group numbers. A pattern such as
2159        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2160    
2161    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2162        the feature of a custom pointer more persuasive (as requested by a user).
2163    
2164    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2165        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2166        pre-defined.
2167    
2168    28. Added support for (*UTF8) at the start of a pattern.
2169    
2170    29. Arrange for flags added by the "release type" setting in CMake to be shown
2171        in the configuration summary.
2172    
2173    
2174    Version 7.8 05-Sep-08
2175    ---------------------
2176    
2177    1.  Replaced UCP searching code with optimized version as implemented for Ad
2178        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2179        stage table and inline lookup instead of a function, giving speed ups of 2
2180        to 5 times on some simple patterns that I tested. Permission was given to
2181        distribute the MultiStage2.py script that generates the tables (it's not in
2182        the tarball, but is in the Subversion repository).
2183    
2184    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2185        scripts.
2186    
2187    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2188        a group with a zero qualifier. The result of the study could be incorrect,
2189        or the function might crash, depending on the pattern.
2190    
2191    4.  Caseless matching was not working for non-ASCII characters in back
2192        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2193        It now works when Unicode Property Support is available.
2194    
2195    5.  In pcretest, an escape such as \x{de} in the data was always generating
2196        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2197        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2198        truncation.
2199    
2200    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2201    
2202    7.  Added two (int) casts to pcregrep when printing the difference of two
2203        pointers, in case they are 64-bit values.
2204    
2205    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2206        test 2 if it fails.
2207    
2208    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2209        and a #define of that name to empty if it is not externally set. This is to
2210        allow users of MSVC to set it if necessary.
2211    
2212    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2213        the convenience functions in the pcre_get.c source file.
2214    
2215    11. An option change at the start of a pattern that had top-level alternatives
2216        could cause overwriting and/or a crash. This command provoked a crash in
2217        some environments:
2218    
2219          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2220    
2221        This potential security problem was recorded as CVE-2008-2371.
2222    
2223    12. For a pattern where the match had to start at the beginning or immediately
2224        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2225        pcre_dfa_exec() could read past the end of the passed subject if there was
2226        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2227        pcretest so that it places the subject at the end of its malloc-ed buffer.
2228    
2229    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2230        exec() might read past the end of the data buffer in UTF-8 mode.
2231    
2232    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2233        the data contained the byte 0x85 as part of a UTF-8 character within its
2234        first line. This applied both to normal and DFA matching.
2235    
2236    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2237        /^[^d]*?$/8 failed to match "abc".
2238    
2239    16. Added a missing copyright notice to pcrecpp_internal.h.
2240    
2241    17. Make it more clear in the documentation that values returned from
2242        pcre_exec() in ovector are byte offsets, not character counts.
2243    
2244    18. Tidied a few places to stop certain compilers from issuing warnings.
2245    
2246    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2247        supplied by Stefan Weber. I made a further small update for 7.8 because
2248        there is a change of source arrangements: the pcre_searchfuncs.c module is
2249        replaced by pcre_ucd.c.
2250    
2251    
2252    Version 7.7 07-May-08
2253    ---------------------
2254    
2255    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2256        a string to a long long, pretend we don't even have a long long." This is
2257        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2258    
2259    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2260        pre-7.6 versions, which defined a global no_arg variable instead of putting
2261        it in the RE class. (See also #8 below.)
2262    
2263    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2264        Lopes.
2265    
2266    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2267    
2268        (1) The include/exclude patterns were being applied to the whole pathnames
2269            of files, instead of just to the final components.
2270    
2271        (2) If there was more than one level of directory, the subdirectories were
2272            skipped unless they satisfied the include/exclude conditions. This is
2273            inconsistent with GNU grep (and could even be seen as contrary to the
2274            pcregrep specification - which I improved to make it absolutely clear).
2275            The action now is always to scan all levels of directory, and just
2276            apply the include/exclude patterns to regular files.
2277    
2278    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2279        --exclude_dir in the tests to avoid scanning .svn directories.
2280    
2281    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2282        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2283        doesn't support NULs in patterns.
2284    
2285    7.  Added some missing "const"s to declarations of static tables in
2286        pcre_compile.c and pcre_dfa_exec.c.
2287    
2288    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2289        caused by fix #2  above. (Subsequently also a second patch to fix the
2290        first patch. And a third patch - this was a messy problem.)
2291    
2292    9.  Applied Craig's patch to remove the use of push_back().
2293    
2294    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2295        matching function regexec().
2296    
2297    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2298        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2299        references. PCRE supports relative numbers with this syntax (I don't think
2300        Oniguruma does).
2301    
2302    12. Previously, a group with a zero repeat such as (...){0} was completely
2303        omitted from the compiled regex. However, this means that if the group
2304        was called as a subroutine from elsewhere in the pattern, things went wrong
2305        (an internal error was given). Such groups are now left in the compiled
2306        pattern, with a new opcode that causes them to be skipped at execution
2307        time.
2308    
2309    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2310        to the way PCRE behaves:
2311    
2312        (a) A lone ] character is dis-allowed (Perl treats it as data).
2313    
2314        (b) A back reference to an unmatched subpattern matches an empty string
2315            (Perl fails the current match path).
2316    
2317        (c) A data ] in a character class must be notated as \] because if the
2318            first data character in a class is ], it defines an empty class. (In
2319            Perl it is not possible to have an empty class.) The empty class []
2320            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2321            The negative empty class [^] matches any one character, independently
2322            of the DOTALL setting.
2323    
2324    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2325        non-existent subpattern following a character class starting with ']' and
2326        containing () gave an internal compiling error instead of "reference to
2327        non-existent subpattern". Fortunately, when the pattern did exist, the
2328        compiled code was correct. (When scanning forwards to check for the
2329        existence of the subpattern, it was treating the data ']' as terminating
2330        the class, so got the count wrong. When actually compiling, the reference
2331        was subsequently set up correctly.)
2332    
2333    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2334        it was being rejected as not supported by pcre_dfa_exec(), even though
2335        other assertions are supported. I have made pcre_dfa_exec() support
2336        (*FAIL).
2337    
2338    16. The implementation of 13c above involved the invention of a new opcode,
2339        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2340        cannot be changed at match time, I realized I could make a small
2341        improvement to matching performance by compiling OP_ALLANY instead of
2342        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2343        on the OP_ANY path.
2344    
2345    17. Compiling pcretest on Windows with readline support failed without the
2346        following two fixes: (1) Make the unistd.h include conditional on
2347        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2348    
2349    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2350        ncurses library to be included for pcretest when ReadLine support is
2351        requested, but also to allow for it to be overridden. This patch came from
2352        Daniel Bergström.
2353    
2354    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2355        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2356        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2357        spotting this.
2358    
2359    
2360    Version 7.6 28-Jan-08
2361    ---------------------
2362    
2363    1.  A character class containing a very large number of characters with
2364        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2365        overflow.
2366    
2367    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2368        HAVE_LONG_LONG is not defined.
2369    
2370    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2371        bring them up to date and include new features. This patch includes:
2372    
2373        - Fixed PH's badly added libz and libbz2 support.
2374        - Fixed a problem with static linking.
2375        - Added pcredemo. [But later removed - see 7 below.]
2376        - Fixed dftables problem and added an option.
2377        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2378            HAVE_LONG_LONG.
2379        - Added readline support for pcretest.
2380        - Added an listing of the option settings after cmake has run.
2381    
2382    4.  A user submitted a patch to Makefile that makes it easy to create
2383        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2384        Makefile.am that cause it to include this special target, without
2385        affecting anything else. Note that the same mingw target plus all
2386        the other distribution libraries and programs are now supported
2387        when configuring with CMake (see 6 below) instead of with
2388        Configure/Make.
2389    
2390    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2391        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2392        exported in the Windows port". It has not yet been confirmed that the patch
2393        solves the problem, but it does no harm.
2394    
2395    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2396        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2397        with CMake, and also correct the comment about stack recursion.
2398    
2399    7.  Remove the automatic building of pcredemo from the ./configure system and
2400        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2401        of a program that users should build themselves after PCRE is installed, so
2402        building it automatically is not really right. What is more, it gave
2403        trouble in some build environments.
2404    
2405    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2406    
2407    
2408    Version 7.5 10-Jan-08
2409    ---------------------
2410    
2411    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2412        values in parens when parsing an RE using the C++ wrapper."
2413    
2414    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2415        Characters greater than 255 were excluded from the class instead of being
2416        included.
2417    
2418    3.  The same bug as (2) above applied to negated POSIX classes such as
2419        [:^space:].
2420    
2421    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2422        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2423        I have changed it.
2424    
2425    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2426        first named subpattern) and a construct such as (?&a) would reference the
2427        first named subpattern whose name started with "a" (in other words, the
2428        length check was missing). Both these problems are fixed. "Subpattern name
2429        expected" is now given for (?&) (a zero-length name), and this patch also
2430        makes it give the same error for \k'' (previously it complained that that
2431        was a reference to a non-existent subpattern).
2432    
2433    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2434        this is right because (?- can be followed by option settings as well as by
2435        digits. I have, however, made the messages clearer.
2436    
2437    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2438        than the number used in the conditional) now cause a compile-time error.
2439        This is actually not compatible with Perl, which accepts such patterns, but
2440        treats the conditional as always being FALSE (as PCRE used to), but it
2441        seems to me that giving a diagnostic is better.
2442    
2443    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2444        and messages.
2445    
2446    9.  Fix two occurrences of "backslash" in comments that should have been
2447        "backspace".
2448    
2449    10. Remove two redundant lines of code that can never be obeyed (their function
2450        was moved elsewhere).
2451    
2452    11. The program that makes PCRE's Unicode character property table had a bug
2453        which caused it to generate incorrect table entries for sequences of
2454        characters that have the same character type, but are in different scripts.
2455        It amalgamated them into a single range, with the script of the first of
2456        them. In other words, some characters were in the wrong script. There were
2457        thirteen such cases, affecting characters in the following ranges:
2458    
2459          U+002b0 - U+002c1
2460          U+0060c - U+0060d
2461          U+0061e - U+00612
2462          U+0064b - U+0065e
2463          U+0074d - U+0076d
2464          U+01800 - U+01805
2465          U+01d00 - U+01d77
2466          U+01d9b - U+01dbf
2467          U+0200b - U+0200f
2468          U+030fc - U+030fe
2469          U+03260 - U+0327f
2470          U+0fb46 - U+0fbb1
2471          U+10450 - U+1049d
2472    
2473    12. The -o option (show only the matching part of a line) for pcregrep was not
2474        compatible with GNU grep in that, if there was more than one match in a
2475        line, it showed only the first of them. It now behaves in the same way as
2476        GNU grep.
2477    
2478    13. If the -o and -v options were combined for pcregrep, it printed a blank
2479        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2480        does the same. The return code can be used to tell if there were any
2481        non-matching lines.
2482    
2483    14. Added --file-offsets and --line-offsets to pcregrep.
2484    
2485    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2486        infinitely looping recursion. The bug was that positive lookaheads were not
2487        being skipped when checking for a possible empty match (negative lookaheads
2488        and both kinds of lookbehind were skipped).
2489    
2490    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2491        inclusion of <windows.h> to before rather than after the definition of
2492        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2493    
2494    17. Specifying a possessive quantifier with a specific limit for a Unicode
2495        character property caused pcre_compile() to compile bad code, which led at
2496        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2497        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2498        caused the error; without that there was no problem.
2499    
2500    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2501    
2502    19. Added --enable-pcretest-libreadline.
2503    
2504    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2505        RE::GlobalReplace(). As a result, the number of replacements returned was
2506        double what it should be. I removed one of the increments, but Craig sent a
2507        later patch that removed the other one (the right fix) and added unit tests
2508        that check the return values (which was not done before).
2509    
2510    21. Several CMake things:
2511    
2512        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2513            the names libpcre and libpcreposix, not just pcre and pcreposix.
2514    
2515        (2) The above change means that pcretest and pcregrep are now correctly
2516            linked with the newly-built libraries, not previously installed ones.
2517    
2518        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2519    
2520    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2521        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2522        UTF-8 newline character). The key issue is that the pattern starts .*;
2523        this means that the match must be either at the beginning, or after a
2524        newline. The bug was in the code for advancing after a failed match and
2525        checking that the new position followed a newline. It was not taking
2526        account of UTF-8 characters correctly.
2527    
2528    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2529        character classes. PCRE was not treating the sequence [:...:] as a
2530        character class unless the ... were all letters. Perl, however, seems to
2531        allow any characters between [: and :], though of course it rejects as
2532        unknown any "names" that contain non-letters, because all the known class
2533        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2534        for example, whereas PCRE did not - it did not recognize a POSIX character
2535        class. This seemed a bit dangerous, so the code has been changed to be
2536        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2537        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2538        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2539        Perl does, and where it didn't before.
2540    
2541    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2542        Windows environments %n is disabled by default.
2543    
2544    
2545    Version 7.4 21-Sep-07
2546    ---------------------
2547    
2548    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2549        means that a class such as [\s] counted as "explicit reference to CR or
2550        LF". That isn't really right - the whole point of the change was to try to
2551        help when there was an actual mention of one of the two characters. So now
2552        the change happens only if \r or \n (or a literal CR or LF) character is
2553        encountered.
2554    
2555    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2556        of both had grown to the point where there were only 3 bits left.
2557        Fortunately, there was spare space in the data structure, and so I have
2558        moved the internal flags into a new 16-bit field to free up more option
2559        bits.
2560    
2561    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2562        but did not set the internal JCHANGED flag - either of these is enough to
2563        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2564        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2565        start sets both bits.
2566    
2567    4.  Added options (at build time, compile time, exec time) to change \R from
2568        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2569    
2570    5.  doc/pcresyntax.html was missing from the distribution.
2571    
2572    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2573        compatibility, even though it is no longer used.
2574    
2575    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2576        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2577        windows.h file is present (where different names are used). [This was
2578        reversed later after testing - see 16 below.]
2579    
2580    8.  Changed all #include <config.h> to #include "config.h". There were also
2581        some further <pcre.h> cases that I changed to "pcre.h".
2582    
2583    9.  When pcregrep was used with the --colour option, it missed the line ending
2584        sequence off the lines that it output.
2585    
2586    10. It was pointed out to me that arrays of string pointers cause lots of
2587        relocations when a shared library is dynamically loaded. A technique of
2588        using a single long string with a table of offsets can drastically reduce
2589        these. I have refactored PCRE in four places to do this. The result is
2590        dramatic:
2591    
2592          Originally:                          290
2593          After changing UCP table:            187
2594          After changing error message table:   43
2595          After changing table of "verbs"       36
2596          After changing table of Posix names   22
2597    
2598        Thanks to the folks working on Gregex for glib for this insight.
2599    
2600    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2601        unicode-properties was also set.
2602    
2603    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2604    
2605    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2606        checked only for CRLF.
2607    
2608    14. Added casts to pcretest.c to avoid compiler warnings.
2609    
2610    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2611    
2612    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2613        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2614        entirely. This removes changes made in 7 above.
2615    
2616    17. The CMake files have been updated, and there is now more information about
2617        building with CMake in the NON-UNIX-USE document.
2618    
2619    
2620    Version 7.3 28-Aug-07
2621    ---------------------
2622    
2623     1. In the rejigging of the build system that eventually resulted in 7.1, the
2624        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2625        brackets there is not right, since it causes compilers to look for an
2626        installed pcre.h, not the version that is in the source that is being
2627        compiled (which of course may be different). I have changed it back to:
2628    
2629          #include "pcre.h"
2630    
2631        I have a vague recollection that the change was concerned with compiling in
2632        different directories, but in the new build system, that is taken care of
2633        by the VPATH setting the Makefile.
2634    
2635     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2636        when the subject happened to end in the byte 0x85 (e.g. if the last
2637        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2638        characters but of course it shouldn't be taken as a newline when it is part
2639        of another character. The bug was that, for an unlimited repeat of . in
2640        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2641        characters when looking for a newline.
2642    
2643     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2644    
2645     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2646        in debug output.
2647    
2648     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2649        long printing in the pcrecpp unittest when running under MinGW.
2650    
2651     6. ESC_K was left out of the EBCDIC table.
2652    
2653     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2654        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2655        limit also applies to "virtual nesting" when a pattern is recursive, and in
2656        this case 1000 isn't so big. I have been able to remove this limit at the
2657        expense of backing off one optimization in certain circumstances. Normally,
2658        when pcre_exec() would call its internal match() function recursively and
2659        immediately return the result unconditionally, it uses a "tail recursion"
2660        feature to save stack. However, when a subpattern that can match an empty
2661        string has an unlimited repetition quantifier, it no longer makes this
2662        optimization. That gives it a stack frame in which to save the data for
2663        checking that an empty string has been matched. Previously this was taken
2664        from the 1000-entry workspace that had been reserved. So now there is no
2665        explicit limit, but more stack is used.
2666    
2667     8. Applied Daniel's patches to solve problems with the import/export magic
2668        syntax that is required for Windows, and which was going wrong for the
2669        pcreposix and pcrecpp parts of the library. These were overlooked when this
2670        problem was solved for the main library.
2671    
2672     9. There were some crude static tests to avoid integer overflow when computing
2673        the size of patterns that contain repeated groups with explicit upper
2674        limits. As the maximum quantifier is 65535, the maximum group length was
2675        set at 30,000 so that the product of these two numbers did not overflow a
2676        32-bit integer. However, it turns out that people want to use groups that
2677        are longer than 30,000 bytes (though not repeat them that many times).
2678        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2679        made it possible to implement the integer overflow checks in a much more
2680        dynamic way, which I have now done. The artificial limitation on group
2681        length has been removed - we now have only the limit on the total length of
2682        the compiled pattern, which depends on the LINK_SIZE setting.
2683    
2684    10. Fixed a bug in the documentation for get/copy named substring when
2685        duplicate names are permitted. If none of the named substrings are set, the
2686        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2687        empty string.
2688    
2689    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2690        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2691        because the ] is interpreted as the first data character and the
2692        terminating ] is not found. PCRE has been made compatible with Perl in this
2693        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2694        cause memory overwriting.
2695    
2696    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2697        string has been matched (to stop an infinite loop). It was not recognizing
2698        a conditional subpattern that could match an empty string if that
2699        subpattern was within another subpattern. For example, it looped when
2700        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2701        condition was not nested. This bug has been fixed.
2702    
2703    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2704        past the start of the subject in the presence of bytes with the top bit
2705        set, for example "\x8aBCD".
2706    
2707    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2708        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2709    
2710    14. Optimized (?!) to (*FAIL).
2711    
2712    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2713        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2714        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2715        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2716        does: it's just the validity check that is more restrictive.
2717    
2718    16. Inserted checks for integer overflows during escape sequence (backslash)
2719        processing, and also fixed erroneous offset values for syntax errors during
2720        backslash processing.
2721    
2722    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2723        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2724    
2725    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2726        caused an overrun.
2727    
2728    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2729        something other than just ASCII characters) inside a group that had an
2730        unlimited repeat caused a loop at compile time (while checking to see
2731        whether the group could match an empty string).
2732    
2733    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2734        [\P{Any}] did so. (Error in the code for printing property names.)
2735    
2736    21. An orphan \E inside a character class could cause a crash.
2737    
2738    22. A repeated capturing bracket such as (A)? could cause a wild memory
2739        reference during compilation.
2740    
2741    23. There are several functions in pcre_compile() that scan along a compiled
2742        expression for various reasons (e.g. to see if it's fixed length for look
2743        behind). There were bugs in these functions when a repeated \p or \P was
2744        present in the pattern. These operators have additional parameters compared
2745        with \d, etc, and these were not being taken into account when moving along
2746        the compiled data. Specifically:
2747    
2748        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2749            length.
2750    
2751        (b) An item such as \pL+ within a repeated group could cause crashes or
2752            loops.
2753    
2754        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2755            "reference to non-existent subpattern" error.
2756    
2757        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2758    
2759    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2760        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2761    
2762    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2763    
2764    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2765        character were causing crashes (broken optimization).
2766    
2767    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2768        \p or \P) caused a compile-time loop.
2769    
2770    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2771        break. For example, the unstudied pattern [\r\n]A does not match the string
2772        "\r\nA" because change 7.0/46 below moves the current point on by two
2773        characters after failing to match at the start. However, the pattern \nA
2774        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2775        the same is true. There doesn't seem any very clean way out of this, but
2776        what I have chosen to do makes the common cases work: PCRE now takes note
2777        of whether there can be an explicit match for \r or \n anywhere in the
2778        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2779        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2780        pattern has explicit CR or LF references.
2781    
2782    29. Added (*CR) etc for changing newline setting at start of pattern.
2783    
2784    
2785    Version 7.2 19-Jun-07
2786    ---------------------
2787    
2788     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2789        which is apparently normally available under Windows.
2790    
2791     2. Re-jig the pcregrep tests with different newline settings in an attempt
2792        to make them independent of the local environment's newline setting.
2793    
2794     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2795    
2796     4. Some of the "internals" tests were previously cut out when the link size
2797        was not 2, because the output contained actual offsets. The recent new
2798        "Z" feature of pcretest means that these can be cut out, making the tests
2799        usable with all link sizes.
2800    
2801     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2802        stack recursion. This gives a massive performance boost under BSD, but just
2803        a small improvement under Linux. However, it saves one field in the frame
2804        in all cases.
2805    
2806     6. Added more features from the forthcoming Perl 5.10:
2807    
2808        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2809            recursion call. It refers to the nth most recently opened parentheses.
2810    
2811        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2812            to be opened parentheses.
2813    
2814        (c) Conditions that refer to capturing parentheses can be specified
2815            relatively, for example, (?(-2)... or (?(+3)...
2816    
2817        (d) \K resets the start of the current match so that everything before
2818            is not part of it.
2819    
2820        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2821    
2822        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2823            reference syntax.
2824    
2825        (g) (?| introduces a group in which the numbering of parentheses in each
2826            alternative starts with the same number.
2827    
2828        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2829    
2830     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2831        PCRE_INFO_JCHANGED.
2832    
2833     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2834        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2835        for detecting groups that can match an empty string.
2836    
2837     9. A pattern with a very large number of alternatives (more than several
2838        hundred) was running out of internal workspace during the pre-compile
2839        phase, where pcre_compile() figures out how much memory will be needed. A
2840        bit of new cunning has reduced the workspace needed for groups with
2841        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2842        workspace instead of running out of the 4096 that are available.
2843    
2844    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2845    
2846    11. Applied patch from Google to remove an optimization that didn't quite work.
2847        The report of the bug said:
2848    
2849          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2850          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2851          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2852    
2853    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2854        it matched the wrong number of bytes.
2855    
2856    
2857    Version 7.1 24-Apr-07
2858  ---------------------  ---------------------
2859    
2860   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
# Line 33  Version 7.1 12-Mar-07 Line 2886  Version 7.1 12-Mar-07
2886   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2887      Weber: (1) pcre_internal.h was missing some function renames; (2) updated      Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2888      makevp.bat for the current PCRE, using the additional files      makevp.bat for the current PCRE, using the additional files
2889      makevp-compile.txt, makevp-linklib.txt, and pcregexp.pas.      makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2890    
2891   6. A Windows user reported a minor discrepancy with test 2, which turned out   6. A Windows user reported a minor discrepancy with test 2, which turned out
2892      to be caused by a trailing space on an input line that had got lost in his      to be caused by a trailing space on an input line that had got lost in his
# Line 116  Version 7.1 12-Mar-07 Line 2969  Version 7.1 12-Mar-07
2969  14. If erroroffset was passed as NULL to pcre_compile, it provoked a  14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2970      segmentation fault instead of returning the appropriate error message.      segmentation fault instead of returning the appropriate error message.
2971    
2972    15. In multiline mode when the newline sequence was set to "any", the pattern
2973        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2974        This doesn't seem right; it now treats the CRLF combination as the line
2975        ending, and so does not match in that case. It's only a pattern such as ^$
2976        that would hit this one: something like ^ABC$ would have failed after \r
2977        and then tried again after \r\n.
2978    
2979    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2980        in an attempt to make files that differ only in their line terminators
2981        compare equal. This works on Linux.
2982    
2983    17. Under certain error circumstances pcregrep might try to free random memory
2984        as it exited. This is now fixed, thanks to valgrind.
2985    
2986    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2987        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2988        was because its rules for how to advance for /g after matching an empty
2989        string at the end of a line did not allow for this case. They now check for
2990        it specially.
2991    
2992    20. pcretest is supposed to handle patterns and data of any length, by
2993        extending its buffers when necessary. It was getting this wrong when the
2994        buffer for a data line had to be extended.
2995    
2996    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2997        CRLF as a newline sequence.
2998    
2999    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3000        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3001        I have nevertheless tidied it up.
3002    
3003    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3004    
3005    24. Added a man page for pcre-config.
3006    
3007    
3008  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06
3009  ---------------------  ---------------------

Legend:
Removed from v.127  
changed lines
  Added in v.1490

  ViewVC Help
Powered by ViewVC 1.1.5