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

Diff of /code/trunk/ChangeLog

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

revision 221 by ph10, Fri Aug 17 09:25:08 2007 UTC revision 1520 by ph10, Sun Feb 8 16:29:23 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 17-Aug-07  Version 8.37 xx-xxx-2015
5    ------------------------
6    
7    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8        for those parentheses to be closed with whatever has been captured so far.
9        However, it was failing to mark any other groups between the hightest
10        capture so far and the currrent group as "unset". Thus, the ovector for
11        those groups contained whatever was previously there. An example is the
12        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13    
14    2.  If an assertion condition was quantified with a minimum of zero (an odd
15        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16    
17    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
18        unrecognized modifier, a crash could occur.
19    
20    4.  An attempt to do global matching in pcretest with a zero-length ovector
21        caused a crash.
22    
23    5.  Fixed a memory leak during matching that could occur for a subpattern
24        subroutine call (recursive or otherwise) if the number of captured groups
25        that had to be saved was greater than ten.
26    
27    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
28        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
29        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
30    
31    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
32        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
33    
34    8. When a pattern is compiled, it remembers the highest back reference so that
35       when matching, if the ovector is too small, extra memory can be obtained to
36       use instead. A conditional subpattern whose condition is a check on a
37       capture having happened, such as, for example in the pattern
38       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
39       setting the highest backreference number. This mattered only if pcre_exec()
40       was called with an ovector that was too small to hold the capture, and there
41       was no other kind of back reference (a situation which is probably quite
42       rare). The effect of the bug was that the condition was always treated as
43       FALSE when the capture could not be consulted, leading to a incorrect
44       behaviour by pcre2_match(). This bug has been fixed.
45    
46    
47    Version 8.36 26-September-2014
48    ------------------------------
49    
50    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
51        -Wmissing-field-initializers and -Wunused-parameter.
52    
53    2.  The tests for quantifiers being too big (greater than 65535) were being
54        applied after reading the number, and stupidly assuming that integer
55        overflow would give a negative number. The tests are now applied as the
56        numbers are read.
57    
58    3.  Tidy code in pcre_exec.c where two branches that used to be different are
59        now the same.
60    
61    4.  The JIT compiler did not generate match limit checks for certain
62        bracketed expressions with quantifiers. This may lead to exponential
63        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
64        issue should be resolved now.
65    
66    5.  Fixed an issue, which occures when nested alternatives are optimized
67        with table jumps.
68    
69    6.  Inserted two casts and changed some ints to size_t in the light of some
70        reported 64-bit compiler warnings (Bugzilla 1477).
71    
72    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
73        an empty string, which sometimes were behaving incorrectly in the
74        interpreter (though correctly in the JIT matcher). This pcretest input is
75        an example:
76    
77          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
78          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
79    
80        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
81        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
82        for an empty string was breaking the inner loop and carrying on at a lower
83        level, when possessive repeated groups should always return to a higher
84        level as they have no backtrack points in them. The empty string test now
85        occurs at the outer level.
86    
87    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
88        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
89    
90    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
91        doesn't).
92    
93    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
94        to hold a pattern to be incorrectly computed (too small) when there were
95        named back references to duplicated names. This could cause "internal
96        error: code overflow" or "double free or corruption" or other memory
97        handling errors.
98    
99    11. When named subpatterns had the same prefixes, back references could be
100        confused. For example, in this pattern:
101    
102          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
103    
104        the reference to 'Name' was incorrectly treated as a reference to a
105        duplicate name.
106    
107    12. A pattern such as /^s?c/mi8 where the optional character has more than
108        one "other case" was incorrectly compiled such that it would only try to
109        match starting at "c".
110    
111    13. When a pattern starting with \s was studied, VT was not included in the
112        list of possible starting characters; this should have been part of the
113        8.34/18 patch.
114    
115    14. If a character class started [\Qx]... where x is any character, the class
116        was incorrectly terminated at the ].
117    
118    15. If a pattern that started with a caseless match for a character with more
119        than one "other case" was studied, PCRE did not set up the starting code
120        unit bit map for the list of possible characters. Now it does. This is an
121        optimization improvement, not a bug fix.
122    
123    16. The Unicode data tables have been updated to Unicode 7.0.0.
124    
125    17. Fixed a number of memory leaks in pcregrep.
126    
127    18. Avoid a compiler warning (from some compilers) for a function call with
128        a cast that removes "const" from an lvalue by using an intermediate
129        variable (to which the compiler does not object).
130    
131    19. Incorrect code was compiled if a group that contained an internal recursive
132        back reference was optional (had quantifier with a minimum of zero). This
133        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
134        caused segmentation faults because of stack overflows at compile time.
135    
136    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
137        group that is quantified with an indefinite repeat, caused a compile-time
138        loop which used up all the system stack and provoked a segmentation fault.
139        This was not the same bug as 19 above.
140    
141    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
142        Patch by Mike Frysinger.
143    
144    
145    Version 8.35 04-April-2014
146    --------------------------
147    
148    1.  A new flag is set, when property checks are present in an XCLASS.
149        When this flag is not set, PCRE can perform certain optimizations
150        such as studying these XCLASS-es.
151    
152    2.  The auto-possessification of character sets were improved: a normal
153        and an extended character set can be compared now. Furthermore
154        the JIT compiler optimizes more character set checks.
155    
156    3.  Got rid of some compiler warnings for potentially uninitialized variables
157        that show up only when compiled with -O2.
158    
159    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
160        of a match later then the end of the match. The pcretest program was not
161        handling the case sensibly - it was outputting from the start to the next
162        binary zero. It now reports this situation in a message, and outputs the
163        text from the end to the start.
164    
165    5.  Fast forward search is improved in JIT. Instead of the first three
166        characters, any three characters with fixed position can be searched.
167        Search order: first, last, middle.
168    
169    6.  Improve character range checks in JIT. Characters are read by an inprecise
170        function now, which returns with an unknown value if the character code is
171        above a certain threshold (e.g: 256). The only limitation is that the value
172        must be bigger than the threshold as well. This function is useful when
173        the characters above the threshold are handled in the same way.
174    
175    7.  The macros whose names start with RAWUCHAR are placeholders for a future
176        mode in which only the bottom 21 bits of 32-bit data items are used. To
177        make this more memorable for those maintaining the code, the names have
178        been changed to start with UCHAR21, and an extensive comment has been added
179        to their definition.
180    
181    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
182        to the export list in Makefile.am (they were accidentally omitted from the
183        8.34 tarball).
184    
185    9.  The informational output from pcretest used the phrase "starting byte set"
186        which is inappropriate for the 16-bit and 32-bit libraries. As the output
187        for "first char" and "need char" really means "non-UTF-char", I've changed
188        "byte" to "char", and slightly reworded the output. The documentation about
189        these values has also been (I hope) clarified.
190    
191    10. Another JIT related optimization: use table jumps for selecting the correct
192        backtracking path, when more than four alternatives are present inside a
193        bracket.
194    
195    11. Empty match is not possible, when the minimum length is greater than zero,
196        and there is no \K in the pattern. JIT should avoid empty match checks in
197        such cases.
198    
199    12. In a caseless character class with UCP support, when a character with more
200        than one alternative case was not the first character of a range, not all
201        the alternative cases were added to the class. For example, s and \x{17f}
202        are both alternative cases for S: the class [RST] was handled correctly,
203        but [R-T] was not.
204    
205    13. The configure.ac file always checked for pthread support when JIT was
206        enabled. This is not used in Windows, so I have put this test inside a
207        check for the presence of windows.h (which was already tested for).
208    
209    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
210        The algorithm provides a way to skip certain starting offsets, and usually
211        faster than linear prefix searches.
212    
213    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
214        as for 'fr_FR' and 'french'. For some reason, however, it then used the
215        Windows-specific input and output files, which have 'french' screwed in.
216        So this could never have worked. One of the problems with locales is that
217        they aren't always the same. I have now updated RunTest so that it checks
218        the output of the locale test (test 3) against three different output
219        files, and it allows the test to pass if any one of them matches. With luck
220        this should make the test pass on some versions of Solaris where it was
221        failing. Because of the uncertainty, the script did not used to stop if
222        test 3 failed; it now does. If further versions of a French locale ever
223        come to light, they can now easily be added.
224    
225    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
226        there was a message during ./configure, but it did not stop. This now
227        provokes an error. The invalid example in README has been corrected.
228        If a value less than the minimum is given, the minimum value has always
229        been used, but now a warning is given.
230    
231    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
232        was a bug in the test system, which is now fixed. Also, the list of various
233        configurations that are tested for each release did not have one with both
234        16/32 bits and --enable-bar-anycrlf. It now does.
235    
236    18. pcretest was missing "-C bsr" for displaying the \R default setting.
237    
238    19. Little endian PowerPC systems are supported now by the JIT compiler.
239    
240    20. The fast forward newline mechanism could enter to an infinite loop on
241        certain invalid UTF-8 input. Although we don't support these cases
242        this issue can be fixed by a performance optimization.
243    
244    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
245        not take account if existing stack usage. There is now a new global
246        variable called pcre_stack_guard that can be set to point to an external
247        function to check stack availability. It is called at the start of
248        processing every parenthesized group.
249    
250    22. A typo in the code meant that in ungreedy mode the max/min qualifier
251        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
252        not match "ab".
253    
254    23. When UTF was disabled, the JIT program reported some incorrect compile
255        errors. These messages are silenced now.
256    
257    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
258        compiler.
259    
260    25. Change all the temporary files used in RunGrepTest to be different to those
261        used by RunTest so that the tests can be run simultaneously, for example by
262        "make -j check".
263    
264    
265    Version 8.34 15-December-2013
266    -----------------------------
267    
268    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
269        executable memory. Patch inspired by Carsten Klein.
270    
271    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
272        this macro is never tested and has no effect, because the work to support
273        coverage involves only compiling and linking options and special targets in
274        the Makefile. The comment in config.h implied that defining the macro would
275        enable coverage support, which is totally false. There was also support for
276        setting this macro in the CMake files (my fault, I just copied it from
277        configure). SUPPORT_GCOV has now been removed.
278    
279    3.  Make a small performance improvement in strlen16() and strlen32() in
280        pcretest.
281    
282    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
283        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
284        these cases). There was also one in pcretest.c.
285    
286    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
287    
288    6.  In UTF mode, the code for checking whether a group could match an empty
289        string (which is used for indefinitely repeated groups to allow for
290        breaking an infinite loop) was broken when the group contained a repeated
291        negated single-character class with a character that occupied more than one
292        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
293        UTF-8 mode). The effect was undefined: the group might or might not be
294        deemed as matching an empty string, or the program might have crashed.
295    
296    7.  The code for checking whether a group could match an empty string was not
297        recognizing that \h, \H, \v, \V, and \R must match a character.
298    
299    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
300        an empty string. If it can, pcretest shows this in its information output.
301    
302    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
303        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
304        matched by pcre_exec() without using JIT:
305    
306        (a) If the rest of the pattern did not match after a maximal run of
307            grapheme clusters, the code for backing up to try with fewer of them
308            did not always back up over a full grapheme when characters that do not
309            have the modifier quality were involved, e.g. Hangul syllables.
310    
311        (b) If the match point in a subject started with modifier character, and
312            there was no match, the code could incorrectly back up beyond the match
313            point, and potentially beyond the first character in the subject,
314            leading to a segfault or an incorrect match result.
315    
316    10. A conditional group with an assertion condition could lead to PCRE
317        recording an incorrect first data item for a match if no other first data
318        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
319        first data item, and therefore matched "ca" after "c" instead of at the
320        start.
321    
322    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
323        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
324    
325    12. The source of pcregrep now includes z/OS-specific code so that it can be
326        compiled for z/OS as part of the special z/OS distribution.
327    
328    13. Added the -T and -TM options to pcretest.
329    
330    14. The code in pcre_compile.c for creating the table of named capturing groups
331        has been refactored. Instead of creating the table dynamically during the
332        actual compiling pass, the information is remembered during the pre-compile
333        pass (on the stack unless there are more than 20 named groups, in which
334        case malloc() is used) and the whole table is created before the actual
335        compile happens. This has simplified the code (it is now nearly 150 lines
336        shorter) and prepared the way for better handling of references to groups
337        with duplicate names.
338    
339    15. A back reference to a named subpattern when there is more than one of the
340        same name now checks them in the order in which they appear in the pattern.
341        The first one that is set is used for the reference. Previously only the
342        first one was inspected. This change makes PCRE more compatible with Perl.
343    
344    16. Unicode character properties were updated from Unicode 6.3.0.
345    
346    17. The compile-time code for auto-possessification has been refactored, based
347        on a patch by Zoltan Herczeg. It now happens after instead of during
348        compilation. The code is cleaner, and more cases are handled. The option
349        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
350        options in pcretest are provided to set it. It can also be set by
351        (*NO_AUTO_POSSESS) at the start of a pattern.
352    
353    18. The character VT has been added to the default ("C" locale) set of
354        characters that match \s and are generally treated as white space,
355        following this same change in Perl 5.18. There is now no difference between
356        "Perl space" and "POSIX space". Whether VT is treated as white space in
357        other locales depends on the locale.
358    
359    19. The code for checking named groups as conditions, either for being set or
360        for being recursed, has been refactored (this is related to 14 and 15
361        above). Processing unduplicated named groups should now be as fast at
362        numerical groups, and processing duplicated groups should be faster than
363        before.
364    
365    20. Two patches to the CMake build system, by Alexander Barkov:
366    
367          (1) Replace the "source" command by "." in CMakeLists.txt because
368              "source" is a bash-ism.
369    
370          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
371              without these the CMake build does not work on Solaris.
372    
373    21. Perl has changed its handling of \8 and \9. If there is no previously
374        encountered capturing group of those numbers, they are treated as the
375        literal characters 8 and 9 instead of a binary zero followed by the
376        literals. PCRE now does the same.
377    
378    22. Following Perl, added \o{} to specify codepoints in octal, making it
379        possible to specify values greater than 0777 and also making them
380        unambiguous.
381    
382    23. Perl now gives an error for missing closing braces after \x{... instead of
383        treating the string as literal. PCRE now does the same.
384    
385    24. RunTest used to grumble if an inappropriate test was selected explicitly,
386        but just skip it when running all tests. This make it awkward to run ranges
387        of tests when one of them was inappropriate. Now it just skips any
388        inappropriate tests, as it always did when running all tests.
389    
390    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
391        character types such as \d or \w, too many callouts were inserted, and the
392        data that they returned was rubbish.
393    
394    26. In UCP mode, \s was not matching two of the characters that Perl matches,
395        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
396        were matched by \h. The code has now been refactored so that the lists of
397        the horizontal and vertical whitespace characters used for \h and \v (which
398        are defined only in one place) are now also used for \s.
399    
400    27. Add JIT support for the 64 bit TileGX architecture.
401        Patch by Jiong Wang (Tilera Corporation).
402    
403    28. Possessive quantifiers for classes (both explicit and automatically
404        generated) now use special opcodes instead of wrapping in ONCE brackets.
405    
406    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
407        (because it's meaningless), this was not happening when PCRE_CASELESS was
408        set. Not wrong, but inefficient.
409    
410    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
411        properties for \w, \d, etc) is present in a test regex. Otherwise if the
412        test contains no characters greater than 255, Perl doesn't realise it
413        should be using Unicode semantics.
414    
415    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
416        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
417        does in Unicode mode.
418    
419    32. Added the "forbid" facility to pcretest so that putting tests into the
420        wrong test files can sometimes be quickly detected.
421    
422    33. There is now a limit (default 250) on the depth of nesting of parentheses.
423        This limit is imposed to control the amount of system stack used at compile
424        time. It can be changed at build time by --with-parens-nest-limit=xxx or
425        the equivalent in CMake.
426    
427    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
428        errors. Perl warns for these when in warning mode, but PCRE has no facility
429        for giving warnings.
430    
431    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
432        However, this was not working for (?!) because it is optimized to (*FAIL),
433        for which PCRE does not allow quantifiers. The optimization is now disabled
434        when a quantifier follows (?!). I can't see any use for this, but it makes
435        things uniform.
436    
437    36. Perl no longer allows group names to start with digits, so I have made this
438        change also in PCRE. It simplifies the code a bit.
439    
440    37. In extended mode, Perl ignores spaces before a + that indicates a
441        possessive quantifier. PCRE allowed a space before the quantifier, but not
442        before the possessive +. It now does.
443    
444    38. The use of \K (reset reported match start) within a repeated possessive
445        group such as (a\Kb)*+ was not working.
446    
447    40. Document that the same character tables must be used at compile time and
448        run time, and that the facility to pass tables to pcre_exec() and
449        pcre_dfa_exec() is for use only with saved/restored patterns.
450    
451    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
452        features for Builds with MSVC:
453    
454        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
455           builds). The .rc files can be used to set FileDescription and many other
456           attributes.
457    
458        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
459           This allows higher-level build scripts which want .pdb files to avoid
460           hard-coding the exact files needed."
461    
462    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
463        mean "start of word" and "end of word", respectively, as a transition aid.
464    
465    43. A minimizing repeat of a class containing codepoints greater than 255 in
466        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
467        compiled to use the heap for recursion.
468    
469    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
470        is configured.
471    
472    
473    Version 8.33 28-May-2013
474    ------------------------
475    
476    1.  Added 'U' to some constants that are compared to unsigned integers, to
477        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
478        variables that are added to signed variables, to ensure the result is
479        signed and can be negated.
480    
481    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
482        CMake config files.
483    
484    3.  Revise the creation of config.h.generic so that all boolean macros are
485        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
486        overriding via -D on the command line possible.
487    
488    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
489        to unsigned int is reported to make a quite noticeable speed difference in
490        a specific Windows environment. Testing on Linux did also appear to show
491        some benefit (and it is clearly not harmful). Also fixed the definition of
492        Xop which should be unsigned.
493    
494    5.  Related to (4), changing the definition of the intermediate variable cc
495        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
496        improvements.
497    
498    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
499        unnecessary spaces.
500    
501    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
502        and later.
503    
504    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
505    
506    9.  Optimizing fast_forward_start_bits in JIT.
507    
508    10. Adding support for callouts in JIT, and fixing some issues revealed
509        during this work. Namely:
510    
511        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
512    
513        (b) Minimum length was not checked before the matching is started.
514    
515    11. The value of capture_last that is passed to callouts was incorrect in some
516        cases when there was a capture on one path that was subsequently abandoned
517        after a backtrack. Also, the capture_last value is now reset after a
518        recursion, since all captures are also reset in this case.
519    
520    12. The interpreter no longer returns the "too many substrings" error in the
521        case when an overflowing capture is in a branch that is subsequently
522        abandoned after a backtrack.
523    
524    13. In the pathological case when an offset vector of size 2 is used, pcretest
525        now prints out the matched string after a yield of 0 or 1.
526    
527    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
528        Only supported by the JIT compiler at the moment.
529    
530    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
531    
532    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
533        offset of the starting point of the matching process, provided the offsets
534        vector is large enough.
535    
536    17. The \A escape now records a lookbehind value of 1, though its execution
537        does not actually inspect the previous character. This is to ensure that,
538        in partial multi-segment matching, at least one character from the old
539        segment is retained when a new segment is processed. Otherwise, if there
540        are no lookbehinds in the pattern, \A might match incorrectly at the start
541        of a new segment.
542    
543    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
544    
545    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
546        gives some modest performance improvement in 8-bit mode.
547    
548    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
549        be expressed in certain programming languages using Universal Character
550        Names.
551    
552    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
553        which points out that "non characters" are not "characters that may not
554        appear in Unicode strings" but rather "characters that are reserved for
555        internal use and have only local meaning".
556    
557    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
558        there was a conditional group that depended on an assertion, if the
559        assertion was false, the callout that immediately followed the alternation
560        in the condition was skipped when pcre_exec() was used for matching.
561    
562    23. Allow an explicit callout to be inserted before an assertion that is the
563        condition for a conditional group, for compatibility with automatic
564        callouts, which always insert a callout at this point.
565    
566    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
567        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
568    
569    25. (*PRUNE) is now supported by the JIT compiler.
570    
571    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
572    
573    27. Fix the case where there are two or more SKIPs with arguments that may be
574        ignored.
575    
576    28. (*SKIP) is now supported by the JIT compiler.
577    
578    29. (*THEN) is now supported by the JIT compiler.
579    
580    30. Update RunTest with additional test selector options.
581    
582    31. The way PCRE handles backtracking verbs has been changed in two ways.
583    
584        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
585        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
586        backtracking. In some cases this makes it more Perl-compatible, but Perl's
587        rather obscure rules do not always do the same thing.
588    
589        (2) Previously, backtracking verbs were confined within assertions. This is
590        no longer the case for positive assertions, except for (*ACCEPT). Again,
591        this sometimes improves Perl compatibility, and sometimes does not.
592    
593    32. A number of tests that were in test 2 because Perl did things differently
594        have been moved to test 1, because either Perl or PCRE has changed, and
595        these tests are now compatible.
596    
597    32. Backtracking control verbs are now handled in the same way in JIT and
598        interpreter.
599    
600    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
601        contained a forward subroutine reference caused a compile error.
602    
603    34. Auto-detect and optimize limited repetitions in JIT.
604    
605    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
606        blocking (*UTF) etc.
607    
608    36. In the interpreter, maximizing pattern repetitions for characters and
609        character types now use tail recursion, which reduces stack usage.
610    
611    37. The value of the max lookbehind was not correctly preserved if a compiled
612        and saved regex was reloaded on a host of different endianness.
613    
614    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
615        of the compiled pattern block, expand the flags field from 16 to 32 bits
616        because it was almost full.
617    
618    39. Try madvise first before posix_madvise.
619    
620    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
621        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
622        original change locked out all matches of empty strings. This has been
623        changed so that one match of an empty string per line is recognized.
624        Subsequent searches on the same line (for colouring or for --only-matching,
625        for example) do not recognize empty strings.
626    
627    41. Applied a user patch to fix a number of spelling mistakes in comments.
628    
629    42. Data lines longer than 65536 caused pcretest to crash.
630    
631    43. Clarified the data type for length and startoffset arguments for pcre_exec
632        and pcre_dfa_exec in the function-specific man pages, where they were
633        explicitly stated to be in bytes, never having been updated. I also added
634        some clarification to the pcreapi man page.
635    
636    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
637        a segmentation fault.
638    
639    
640    Version 8.32 30-November-2012
641    -----------------------------
642    
643    1.  Improved JIT compiler optimizations for first character search and single
644        character iterators.
645    
646    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
647        Patch by Daniel Richard G.
648    
649    3.  Single character iterator optimizations in the JIT compiler.
650    
651    4.  Improved JIT compiler optimizations for character ranges.
652    
653    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
654        Reported by Giuseppe D'Angelo.
655    
656    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
657        of a line, was being set incorrectly in cases where .* appeared inside
658        atomic brackets at the start of a pattern, or where there was a subsequent
659        *PRUNE or *SKIP.
660    
661    7.  Improved instruction cache flush for POWER/PowerPC.
662        Patch by Daniel Richard G.
663    
664    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
665        grep:
666    
667        (a) There is now no limit to the number of patterns to be matched.
668    
669        (b) An error is given if a pattern is too long.
670    
671        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
672            are now supported.
673    
674        (d) --exclude-from and --include-from (multiple use) have been added.
675    
676        (e) Exclusions and inclusions now apply to all files and directories, not
677            just to those obtained from scanning a directory recursively.
678    
679        (f) Multiple uses of -f and --file-list are now supported.
680    
681        (g) In a Windows environment, the default for -d has been changed from
682            "read" (the GNU grep default) to "skip", because otherwise the presence
683            of a directory in the file list provokes an error.
684    
685        (h) The documentation has been revised and clarified in places.
686    
687    9.  Improve the matching speed of capturing brackets.
688    
689    10. Changed the meaning of \X so that it now matches a Unicode extended
690        grapheme cluster.
691    
692    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
693        out POSIX threads when JIT support is configured.
694    
695    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
696    
697    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
698        structure could go wrong in environments where size_t is not the same size
699        as int.
700    
701    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
702        set.
703    
704    15. The EBCDIC support had decayed; later updates to the code had included
705        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
706        general tidy up of EBCDIC-related issues, and the documentation was also
707        not quite right. There is now a test that can be run on ASCII systems to
708        check some of the EBCDIC-related things (but is it not a full test).
709    
710    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
711        in a small tidy to the code.
712    
713    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
714    
715    18. If the --only-matching (-o) option in pcregrep is specified multiple
716        times, each one causes appropriate output. For example, -o1 -o2 outputs the
717        substrings matched by the 1st and 2nd capturing parentheses. A separating
718        string can be specified by --om-separator (default empty).
719    
720    19. Improving the first n character searches.
721    
722    20. Turn case lists for horizontal and vertical white space into macros so that
723        they are defined only once.
724    
725    21. This set of changes together give more compatible Unicode case-folding
726        behaviour for characters that have more than one other case when UCP
727        support is available.
728    
729        (a) The Unicode property table now has offsets into a new table of sets of
730            three or more characters that are case-equivalent. The MultiStage2.py
731            script that generates these tables (the pcre_ucd.c file) now scans
732            CaseFolding.txt instead of UnicodeData.txt for character case
733            information.
734    
735        (b) The code for adding characters or ranges of characters to a character
736            class has been abstracted into a generalized function that also handles
737            case-independence. In UTF-mode with UCP support, this uses the new data
738            to handle characters with more than one other case.
739    
740        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
741            whose other case is greater than 256 are now correctly matched
742            caselessly. Previously, the high codepoint matched the low one, but not
743            vice versa.
744    
745        (d) The processing of \h, \H, \v, and \ in character classes now makes use
746            of the new class addition function, using character lists defined as
747            macros alongside the case definitions of 20 above.
748    
749        (e) Caseless back references now work with characters that have more than
750            one other case.
751    
752        (f) General caseless matching of characters with more than one other case
753            is supported.
754    
755    22. Unicode character properties were updated from Unicode 6.2.0
756    
757    23. Improved CMake support under Windows. Patch by Daniel Richard G.
758    
759    24. Add support for 32-bit character strings, and UTF-32
760    
761    25. Major JIT compiler update (code refactoring and bugfixing).
762        Experimental Sparc 32 support is added.
763    
764    26. Applied a modified version of Daniel Richard G's patch to create
765        pcre.h.generic and config.h.generic by "make" instead of in the
766        PrepareRelease script.
767    
768    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
769        pcre_compile.c when checking for a zero character.
770    
771    28. Introducing a native interface for JIT. Through this interface, the compiled
772        machine code can be directly executed. The purpose of this interface is to
773        provide fast pattern matching, so several sanity checks are not performed.
774        However, feature tests are still performed. The new interface provides
775        1.4x speedup compared to the old one.
776    
777    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
778        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
779        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
780    
781    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
782        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
783        the "old" RFC 2279). Instead, it ended up passing a negative length to
784        pcre_exec().
785    
786    31. Add support for GCC's visibility feature to hide internal functions.
787    
788    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
789        "unknown -C option" after outputting 0 or 1.
790    
791    33. There is now support for generating a code coverage report for the test
792        suite in environments where gcc is the compiler and lcov is installed. This
793        is mainly for the benefit of the developers.
794    
795    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
796        unaddressable using valgrind annotations, allowing valgrind to detect
797        invalid memory accesses. This is mainly for the benefit of the developers.
798    
799    25. (*UTF) can now be used to start a pattern in any of the three libraries.
800    
801    26. Give configure error if --enable-cpp but no C++ compiler found.
802    
803    
804    Version 8.31 06-July-2012
805    -------------------------
806    
807    1.  Fixing a wrong JIT test case and some compiler warnings.
808    
809    2.  Removed a bashism from the RunTest script.
810    
811    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
812        to unsigned type, result still unsigned" that was given by an MS compiler
813        on encountering the code "-sizeof(xxx)".
814    
815    4.  Partial matching support is added to the JIT compiler.
816    
817    5.  Fixed several bugs concerned with partial matching of items that consist
818        of more than one character:
819    
820        (a) /^(..)\1/ did not partially match "aba" because checking references was
821            done on an "all or nothing" basis. This also applied to repeated
822            references.
823    
824        (b) \R did not give a hard partial match if \r was found at the end of the
825            subject.
826    
827        (c) \X did not give a hard partial match after matching one or more
828            characters at the end of the subject.
829    
830        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
831            a partial match for the string "\r".
832    
833        (e) When newline was set to CRLF, the metacharacter "." did not recognize
834            a partial match for a CR character at the end of the subject string.
835    
836    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
837        running pcretest, the text "(JIT)" added to the output whenever JIT is
838        actually used to run the match.
839    
840    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
841        or /S+[+] with a digit between 1 and 7.
842    
843    8.  OP_NOT now supports any UTF character not just single-byte ones.
844    
845    9.  (*MARK) control verb is now supported by the JIT compiler.
846    
847    10. The command "./RunTest list" lists the available tests without actually
848        running any of them. (Because I keep forgetting what they all are.)
849    
850    11. Add PCRE_INFO_MAXLOOKBEHIND.
851    
852    12. Applied a (slightly modified) user-supplied patch that improves performance
853        when the heap is used for recursion (compiled with --disable-stack-for-
854        recursion). Instead of malloc and free for each heap frame each time a
855        logical recursion happens, frames are retained on a chain and re-used where
856        possible. This sometimes gives as much as 30% improvement.
857    
858    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
859        call.
860    
861    14. As documented, (*COMMIT) is now confined to within a positive assertion.
862    
863    15. It is now possible to link pcretest with libedit as an alternative to
864        libreadline.
865    
866    16. (*COMMIT) control verb is now supported by the JIT compiler.
867    
868    17. The Unicode data tables have been updated to Unicode 6.1.0.
869    
870    18. Added --file-list option to pcregrep.
871    
872    19. Added binary file support to pcregrep, including the -a, --binary-files,
873        -I, and --text options.
874    
875    20. The madvise function is renamed for posix_madvise for QNX compatibility
876        reasons. Fixed by Giuseppe D'Angelo.
877    
878    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
879    
880    22. Changed the option for creating symbolic links for 16-bit man pages from
881        -s to -sf so that re-installing does not cause issues.
882    
883    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
884    
885    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
886        matches in certain environments (the workspace was not being correctly
887        retained). Also added to pcre_dfa_exec() a simple plausibility check on
888        some of the workspace data at the beginning of a restart.
889    
890    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
891        was not doing so when it should - probably a typo introduced by SVN 528
892        (change 8.10/14).
893    
894    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
895        \w+ when the character tables indicated that \x{c4} was a word character.
896        There were several related cases, all because the tests for doing a table
897        lookup were testing for characters less than 127 instead of 255.
898    
899    27. If a pattern contains capturing parentheses that are not used in a match,
900        their slots in the ovector are set to -1. For those that are higher than
901        any matched groups, this happens at the end of processing. In the case when
902        there were back references that the ovector was too small to contain
903        (causing temporary malloc'd memory to be used during matching), and the
904        highest capturing number was not used, memory off the end of the ovector
905        was incorrectly being set to -1. (It was using the size of the temporary
906        memory instead of the true size.)
907    
908    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
909        ovector size, it uses memory at the end of the block that it has got.
910    
911    29. Check for an overlong MARK name and give an error at compile time. The
912        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
913    
914    30. JIT compiler update.
915    
916    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
917        Rill for the patch.
918    
919    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
920    
921    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
922    
923    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
924        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
925        was enabled.
926    
927    35. Improve JIT code generation for greedy plus quantifier.
928    
929    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
930        1 to "aa" instead of to an empty string. The bug affected repeated groups
931        that could potentially match an empty string.
932    
933    37. Optimizing single character iterators in JIT.
934    
935    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
936        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
937        codepoints that are too big for the mode are faulted, and in a UTF mode,
938        disallowed codepoints are also faulted.
939    
940    39. If PCRE was compiled with UTF support, in three places in the DFA
941        matcher there was code that should only have been obeyed in UTF mode, but
942        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
943        processing when bytes with values greater than 127 were present. In 16-bit
944        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
945        both cases the values are those that cannot be the first data item in a UTF
946        character. The three items that might have provoked this were recursions,
947        possessively repeated groups, and atomic groups.
948    
949    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
950        and pcregrep, because some OS require shared objects to be explicitly
951        passed to ld, causing the link step to fail if they are not.
952    
953    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
954        mode, patterns that started with \h* or \R* might be incorrectly matched.
955    
956    
957    Version 8.30 04-February-2012
958    -----------------------------
959    
960    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
961        name is defined in ctype.h.
962    
963    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
964        only in quite long subpatterns.
965    
966    3.  Removed the function pcre_info(), which has been obsolete and deprecated
967        since it was replaced by pcre_fullinfo() in February 2000.
968    
969    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
970        match a (*MARK), and the match failed at the start of the subject, a
971        reference to memory before the start of the subject could occur. This bug
972        was introduced by fix 17 of release 8.21.
973    
974    5.  A reference to an unset group with zero minimum repetition was giving
975        totally wrong answers (in non-JavaScript-compatibility mode). For example,
976        /(another)?(\1?)test/ matched against "hello world test". This bug was
977        introduced in release 8.13.
978    
979    6.  Add support for 16-bit character strings (a large amount of work involving
980        many changes and refactorings).
981    
982    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
983        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
984        from a file.
985    
986    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
987        rounding is not applied in this particular case).
988    
989    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
990        if they appear, or are escaped, in patterns.
991    
992    10. Get rid of a number of -Wunused-but-set-variable warnings.
993    
994    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
995        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
996        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
997        also returns the mark "x". This bug applied to capturing parentheses,
998        non-capturing parentheses, and atomic parentheses. It also applied to some
999        assertions.
1000    
1001    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1002        information out of configure.ac instead of relying on pcre.h.generic, which
1003        is not stored in the repository.
1004    
1005    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1006        -lreadline.
1007    
1008    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1009    
1010    15. Applied Graycode's patch to put the top-level frame on the stack rather
1011        than the heap when not using the stack for recursion. This gives a
1012        performance improvement in many cases when recursion is not deep.
1013    
1014    16. Experimental code added to "pcretest -C" to output the stack frame size.
1015    
1016    
1017    Version 8.21 12-Dec-2011
1018    ------------------------
1019    
1020    1.  Updating the JIT compiler.
1021    
1022    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1023        are added as well.
1024    
1025    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1026        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1027        calling _pcre_jit_exec. Some extra comments are added.
1028    
1029    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1030        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1031        was introduced by change 18 for 8.20.
1032    
1033    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1034        ECMA-262 standard.
1035    
1036    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1037        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1038        This bug was probably introduced by change 9 of 8.13.
1039    
1040    7.  While fixing 6 above, I noticed that a number of other items were being
1041        incorrectly rejected as "not fixed length". This arose partly because newer
1042        opcodes had not been added to the fixed-length checking code. I have (a)
1043        corrected the bug and added tests for these items, and (b) arranged for an
1044        error to occur if an unknown opcode is encountered while checking for fixed
1045        length instead of just assuming "not fixed length". The items that were
1046        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1047        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1048        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1049    
1050    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1051        being incorrectly compiled and would have given unpredicatble results.
1052    
1053    9.  A possessively repeated subpattern with minimum repeat count greater than
1054        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1055        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1056        the first (A) could occur when it should not.
1057    
1058    10. Add a cast and remove a redundant test from the code.
1059    
1060    11. JIT should use pcre_malloc/pcre_free for allocation.
1061    
1062    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1063        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1064        is anything other than /usr, -L is still shown).
1065    
1066    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1067    
1068    14. Perl does not support \N without a following name in a [] class; PCRE now
1069        also gives an error.
1070    
1071    15. If a forward reference was repeated with an upper limit of around 2000,
1072        it caused the error "internal error: overran compiling workspace". The
1073        maximum number of forward references (including repeats) was limited by the
1074        internal workspace, and dependent on the LINK_SIZE. The code has been
1075        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1076        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1077        of around 200,000 forward references. While doing this, I also speeded up
1078        the filling in of repeated forward references.
1079    
1080    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1081        incorrectly expecting the subject to contain another "a" after the start.
1082    
1083    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1084        in the match, the SKIP should be ignored. This was not happening; instead
1085        the SKIP was being treated as NOMATCH. For patterns such as
1086        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1087        tested.
1088    
1089    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1090        now much more compatible with Perl, in particular in cases where the result
1091        is a non-match for a non-anchored pattern. For example, if
1092        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1093        "m", where previously it did not return a name. A side effect of this
1094        change is that for partial matches, the last encountered mark name is
1095        returned, as for non matches. A number of tests that were previously not
1096        Perl-compatible have been moved into the Perl-compatible test files. The
1097        refactoring has had the pleasing side effect of removing one argument from
1098        the match() function, thus reducing its stack requirements.
1099    
1100    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1101        subsequent uses of /S (without +) incorrectly behaved like /S+.
1102    
1103    21. Retrieve executable code size support for the JIT compiler and fixing
1104        some warnings.
1105    
1106    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1107        not work when the shorter character appeared right at the end of the
1108        subject string.
1109    
1110    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1111        systems.
1112    
1113    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1114        output it when the /M option is used in pcretest.
1115    
1116    25. The CheckMan script was not being included in the distribution. Also, added
1117        an explicit "perl" to run Perl scripts from the PrepareRelease script
1118        because this is reportedly needed in Windows.
1119    
1120    26. If study data was being save in a file and studying had not found a set of
1121        "starts with" bytes for the pattern, the data written to the file (though
1122        never used) was taken from uninitialized memory and so caused valgrind to
1123        complain.
1124    
1125    27. Updated RunTest.bat as provided by Sheri Pierce.
1126    
1127    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1128    
1129    29. Computation of memory usage for the table of capturing group names was
1130        giving an unnecessarily large value.
1131    
1132    
1133    Version 8.20 21-Oct-2011
1134    ------------------------
1135    
1136    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1137        a POSIX class. After further experiments with Perl, which convinced me that
1138        Perl has bugs and confusions, a closing square bracket is no longer allowed
1139        in a POSIX name. This bug also affected patterns with classes that started
1140        with full stops.
1141    
1142    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1143        captured substring, but while checking the failing first alternative,
1144        substring 1 is temporarily captured. If the output vector supplied to
1145        pcre_exec() was not big enough for this capture, the yield of the function
1146        was still zero ("insufficient space for captured substrings"). This cannot
1147        be totally fixed without adding another stack variable, which seems a lot
1148        of expense for a edge case. However, I have improved the situation in cases
1149        such as /(a)(b)x|abc/ matched against "abc", where the return code
1150        indicates that fewer than the maximum number of slots in the ovector have
1151        been set.
1152    
1153    3.  Related to (2) above: when there are more back references in a pattern than
1154        slots in the output vector, pcre_exec() uses temporary memory during
1155        matching, and copies in the captures as far as possible afterwards. It was
1156        using the entire output vector, but this conflicts with the specification
1157        that only 2/3 is used for passing back captured substrings. Now it uses
1158        only the first 2/3, for compatibility. This is, of course, another edge
1159        case.
1160    
1161    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1162        main code base, and can be used by building with --enable-jit. When this is
1163        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1164        runtime --no-jit option is given.
1165    
1166    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1167        ovector, the return from the function was zero, implying that there were
1168        other matches that did not fit. The correct "exactly full" value is now
1169        returned.
1170    
1171    6.  If a subpattern that was called recursively or as a subroutine contained
1172        (*PRUNE) or any other control that caused it to give a non-standard return,
1173        invalid errors such as "Error -26 (nested recursion at the same subject
1174        position)" or even infinite loops could occur.
1175    
1176    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1177        computing the minimum length on reaching *ACCEPT, and so ended up with the
1178        wrong value of 1 rather than 0. Further investigation indicates that
1179        computing a minimum subject length in the presence of *ACCEPT is difficult
1180        (think back references, subroutine calls), and so I have changed the code
1181        so that no minimum is registered for a pattern that contains *ACCEPT.
1182    
1183    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1184        it was not handled as intended. [But see 16 below.]
1185    
1186    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1187        Sheri Pierce.
1188    
1189    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1190        the first byte in a match must be "a".
1191    
1192    11. Change 17 for 8.13 increased the recursion depth for patterns like
1193        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1194        pattern contains any instances of (*THEN). If it does not, the old
1195        optimizations are restored. It would be nice to do this on a per-group
1196        basis, but at the moment that is not feasible.
1197    
1198    12. In some environments, the output of pcretest -C is CRLF terminated. This
1199        broke RunTest's code that checks for the link size. A single white space
1200        character after the value is now allowed for.
1201    
1202    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1203        For "fr", it uses the Windows-specific input and output files.
1204    
1205    14. If (*THEN) appeared in a group that was called recursively or as a
1206        subroutine, it did not work as intended. [But see next item.]
1207    
1208    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1209        pattern fragments (but not containing any | characters). If A and B are
1210        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1211        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1212        D. In other words, Perl considers parentheses that do not contain any |
1213        characters to be part of a surrounding alternative, whereas PCRE was
1214        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1215        differently. PCRE now behaves in the same way as Perl, except in the case
1216        of subroutine/recursion calls such as (?1) which have in any case always
1217        been different (but PCRE had them first :-).
1218    
1219    16. Related to 15 above: Perl does not treat the | in a conditional group as
1220        creating alternatives. Such a group is treated in the same way as an
1221        ordinary group without any | characters when processing (*THEN). PCRE has
1222        been changed to match Perl's behaviour.
1223    
1224    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1225        RunGrepTest script failed.
1226    
1227    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1228        inevitable for groups that contain captures, but it can lead to a lot of
1229        stack use in large patterns. The old behaviour has been restored for atomic
1230        groups that do not contain any capturing parentheses.
1231    
1232    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1233        suppress the check for a minimum subject length at run time. (If it was
1234        given to pcre_exec() or pcre_dfa_exec() it did work.)
1235    
1236    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1237        fail to work when decoding hex characters in data strings in EBCDIC
1238        environments.
1239    
1240    21. It appears that in at least one Mac OS environment, the isxdigit() function
1241        is implemented as a macro that evaluates to its argument more than once,
1242        contravening the C 90 Standard (I haven't checked a later standard). There
1243        was an instance in pcretest which caused it to go wrong when processing
1244        \x{...} escapes in subject strings. The has been rewritten to avoid using
1245        things like p++ in the argument of isxdigit().
1246    
1247    
1248    Version 8.13 16-Aug-2011
1249    ------------------------
1250    
1251    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1252    
1253    2.  Two minor typos in pcre_internal.h have been fixed.
1254    
1255    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1256        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1257        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1258    
1259    4.  There were a number of related bugs in the code for matching backrefences
1260        caselessly in UTF-8 mode when codes for the characters concerned were
1261        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1262        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1263        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1264        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1265        2-byte code at the end of the subject (it thought there wasn't enough data
1266        left).
1267    
1268    5.  Comprehensive information about what went wrong is now returned by
1269        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1270        as the output vector has at least 2 elements. The offset of the start of
1271        the failing character and a reason code are placed in the vector.
1272    
1273    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1274        now returned is for the first byte of the failing character, instead of the
1275        last byte inspected. This is an incompatible change, but I hope it is small
1276        enough not to be a problem. It makes the returned offset consistent with
1277        pcre_exec() and pcre_dfa_exec().
1278    
1279    7.  pcretest now gives a text phrase as well as the error number when
1280        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1281        failure, the offset and reason code are output.
1282    
1283    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1284        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1285        back over a single character (\n). This seems wrong (because it treated the
1286        two characters as a single entity when going forwards), conflicts with the
1287        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1288        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1289        has been changed.
1290    
1291    9.  Some internal refactoring has changed the processing so that the handling
1292        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1293        time (the PCRE_DOTALL option was changed this way some time ago: version
1294        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1295        which was always a bit of a fudge. It also means that there is one less
1296        argument for the match() function, which reduces its stack requirements
1297        slightly. This change also fixes an incompatibility with Perl: the pattern
1298        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1299    
1300    10. More internal refactoring has drastically reduced the number of recursive
1301        calls to match() for possessively repeated groups such as (abc)++ when
1302        using pcre_exec().
1303    
1304    11. While implementing 10, a number of bugs in the handling of groups were
1305        discovered and fixed:
1306    
1307        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1308        (a|)*(?1) gave a compile-time internal error.
1309        ((a|)+)+  did not notice that the outer group could match an empty string.
1310        (^a|^)+   was not marked as anchored.
1311        (.*a|.*)+ was not marked as matching at start or after a newline.
1312    
1313    12. Yet more internal refactoring has removed another argument from the match()
1314        function. Special calls to this function are now indicated by setting a
1315        value in a variable in the "match data" data block.
1316    
1317    13. Be more explicit in pcre_study() instead of relying on "default" for
1318        opcodes that mean there is no starting character; this means that when new
1319        ones are added and accidentally left out of pcre_study(), testing should
1320        pick them up.
1321    
1322    14. The -s option of pcretest has been documented for ages as being an old
1323        synonym of -m (show memory usage). I have changed it to mean "force study
1324        for every regex", that is, assume /S for every regex. This is similar to -i
1325        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1326        using it. It makes it easier to run collections of tests with and without
1327        study enabled, and thereby test pcre_study() more easily. All the standard
1328        tests are now run with and without -s (but some patterns can be marked as
1329        "never study" - see 20 below).
1330    
1331    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1332        restoration of the capturing data to the outer values was not happening
1333        correctly.
1334    
1335    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1336        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1337        pattern had matched an empty string, and so incorrectly returned a no
1338        match.
1339    
1340    17. There was optimizing code for the last branch of non-capturing parentheses,
1341        and also for the obeyed branch of a conditional subexpression, which used
1342        tail recursion to cut down on stack usage. Unfortunately, now that there is
1343        the possibility of (*THEN) occurring in these branches, tail recursion is
1344        no longer possible because the return has to be checked for (*THEN). These
1345        two optimizations have therefore been removed. [But see 8.20/11 above.]
1346    
1347    18. If a pattern containing \R was studied, it was assumed that \R always
1348        matched two bytes, thus causing the minimum subject length to be
1349        incorrectly computed because \R can also match just one byte.
1350    
1351    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1352        was incorrectly computed.
1353    
1354    20. If /S is present twice on a test pattern in pcretest input, it now
1355        *disables* studying, thereby overriding the use of -s on the command line
1356        (see 14 above). This is necessary for one or two tests to keep the output
1357        identical in both cases.
1358    
1359    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1360        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1361    
1362    22. When an atomic group that contained a capturing parenthesis was
1363        successfully matched, but the branch in which it appeared failed, the
1364        capturing was not being forgotten if a higher numbered group was later
1365        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1366        group 1 to "a", when in fact it should be unset. This applied to multi-
1367        branched capturing and non-capturing groups, repeated or not, and also to
1368        positive assertions (capturing in negative assertions does not happen
1369        in PCRE) and also to nested atomic groups.
1370    
1371    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1372        subject after a captured substring, to make it easier to tell which of a
1373        number of identical substrings has been captured.
1374    
1375    24. The way atomic groups are processed by pcre_exec() has been changed so that
1376        if they are repeated, backtracking one repetition now resets captured
1377        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1378        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1379        "aaa". Previously, it would have been "a". As part of this code
1380        refactoring, the way recursive calls are handled has also been changed.
1381    
1382    25. If an assertion condition captured any substrings, they were not passed
1383        back unless some other capturing happened later. For example, if
1384        (?(?=(a))a) was matched against "a", no capturing was returned.
1385    
1386    26. When studying a pattern that contained subroutine calls or assertions,
1387        the code for finding the minimum length of a possible match was handling
1388        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1389        group 1 called group 2 while simultaneously a separate group 2 called group
1390        1). A stack overflow occurred in this case. I have fixed this by limiting
1391        the recursion depth to 10.
1392    
1393    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1394        Fortmann. This supports explicit test numbers on the command line, and has
1395        argument validation and error reporting.
1396    
1397    28. An instance of \X with an unlimited repeat could fail if at any point the
1398        first character it looked at was a mark character.
1399    
1400    29. Some minor code refactoring concerning Unicode properties and scripts
1401        should reduce the stack requirement of match() slightly.
1402    
1403    30. Added the '=' option to pcretest to check the setting of unused capturing
1404        slots at the end of the pattern, which are documented as being -1, but are
1405        not included in the return count.
1406    
1407    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1408        compiled something random. Now it gives a compile-time error (as does
1409        Perl).
1410    
1411    32. A *MARK encountered during the processing of a positive assertion is now
1412        recorded and passed back (compatible with Perl).
1413    
1414    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1415        had alternative anchored branches, the search for a second match in a line
1416        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1417        matched the line "0102" twice. The same bug affected patterns that started
1418        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1419        twice.
1420    
1421    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1422        does, and because of capturing effects, quantifying parenthesized
1423        assertions may at times be useful. Quantifiers are now allowed for
1424        parenthesized assertions.
1425    
1426    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1427    
1428    36. \g was being checked for fancy things in a character class, when it should
1429        just be a literal "g".
1430    
1431    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1432        appearance of a nested POSIX class supersedes an apparent external class.
1433        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1434        unescaped square brackets may also appear as part of class names. For
1435        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1436        more like Perl. (But see 8.20/1 above.)
1437    
1438    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1439        was because it thought it was \N{name}, which is not supported).
1440    
1441    39. Add minix to OS list not supporting the -S option in pcretest.
1442    
1443    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1444        cannot analyze patterns in sufficient detail to catch mutual recursions
1445        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1446        subgroup is called recursively as a subpattern for a second time at the
1447        same position in the subject string. In previous releases this might have
1448        been caught by the recursion limit, or it might have run out of stack.
1449    
1450    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1451        happen only once. PCRE was, however incorrectly giving a compile time error
1452        "recursive call could loop indefinitely" because it cannot analyze the
1453        pattern in sufficient detail. The compile time test no longer happens when
1454        PCRE is compiling a conditional subpattern, but actual runaway loops are
1455        now caught at runtime (see 40 above).
1456    
1457    42. It seems that Perl allows any characters other than a closing parenthesis
1458        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1459        has been changed to be the same.
1460    
1461    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1462        as not to get warnings when autogen.sh is called. Also changed
1463        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1464    
1465    44. To help people who use pcregrep to scan files containing exceedingly long
1466        lines, the following changes have been made:
1467    
1468        (a) The default value of the buffer size parameter has been increased from
1469            8K to 20K. (The actual buffer used is three times this size.)
1470    
1471        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1472            PCRE is built.
1473    
1474        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1475            to be set at run time.
1476    
1477        (d) Numerical values in pcregrep options can be followed by K or M, for
1478            example --buffer-size=50K.
1479    
1480        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1481            given and the return code is set to 2.
1482    
1483    45. Add a pointer to the latest mark to the callout data block.
1484    
1485    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1486        partial match of an empty string instead of no match. This was specific to
1487        the use of ".".
1488    
1489    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1490        complete match instead of a partial match. This bug was dependent on both
1491        the PCRE_UTF8 and PCRE_DOTALL options being set.
1492    
1493    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1494        starting byte set, because \b was not being ignored.
1495    
1496    
1497    Version 8.12 15-Jan-2011
1498    ------------------------
1499    
1500    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1501        checks for such things as part of the documentation building process.
1502    
1503    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1504        --match-limit and --recursion-limit options (added for 8.11). In
1505        particular, this made one of the standard tests fail. (The integer value
1506        went into the wrong half of a long int.)
1507    
1508    3.  If the --colour option was given to pcregrep with -v (invert match), it
1509        did strange things, either producing crazy output, or crashing. It should,
1510        of course, ignore a request for colour when reporting lines that do not
1511        match.
1512    
1513    4.  Another pcregrep bug caused similar problems if --colour was specified with
1514        -M (multiline) and the pattern match finished with a line ending.
1515    
1516    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1517        matched in multiline mode, the following line was shown as part of the
1518        match. This seems wrong, so I have changed it.
1519    
1520    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1521        the check for further matches in the same line (so they could be coloured)
1522        to overrun the end of the current line. If another match was found, it was
1523        incorrectly shown (and then shown again when found in the next line).
1524    
1525    7.  If pcregrep was compiled under Windows, there was a reference to the
1526        function pcregrep_exit() before it was defined. I am assuming this was
1527        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1528        reported by a user. I've moved the definition above the reference.
1529    
1530    
1531    Version 8.11 10-Dec-2010
1532    ------------------------
1533    
1534    1.  (*THEN) was not working properly if there were untried alternatives prior
1535        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1536        backtracked to try for "b" instead of moving to the next alternative branch
1537        at the same level (in this case, to look for "c"). The Perl documentation
1538        is clear that when (*THEN) is backtracked onto, it goes to the "next
1539        alternative in the innermost enclosing group".
1540    
1541    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1542        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1543        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1544        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1545        (*THEN).
1546    
1547    3.  If \s appeared in a character class, it removed the VT character from
1548        the class, even if it had been included by some previous item, for example
1549        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1550        of \s, but is part of the POSIX "space" class.)
1551    
1552    4.  A partial match never returns an empty string (because you can always
1553        match an empty string at the end of the subject); however the checking for
1554        an empty string was starting at the "start of match" point. This has been
1555        changed to the "earliest inspected character" point, because the returned
1556        data for a partial match starts at this character. This means that, for
1557        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1558        (previously it gave "no match").
1559    
1560    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1561        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1562        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1563        has an implication that the given string is incomplete (because a partial
1564        match is preferred over a full match). For this reason, these items now
1565        give a partial match in this situation. [Aside: previously, the one case
1566        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1567        match rather than a full match, which was wrong by the old rules, but is
1568        now correct.]
1569    
1570    6.  There was a bug in the handling of #-introduced comments, recognized when
1571        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1572        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1573        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1574        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1575        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1576        places in pcre_compile().
1577    
1578    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1579        comments when looking ahead for named forward references to subpatterns,
1580        the only newline sequence it recognized was NL. It now handles newlines
1581        according to the set newline convention.
1582    
1583    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1584        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1585        cater for a lack of strerror(). These oversights have been fixed.
1586    
1587    9.  Added --match-limit and --recursion-limit to pcregrep.
1588    
1589    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1590    
1591    11. When the -o option was used, pcregrep was setting a return code of 1, even
1592        when matches were found, and --line-buffered was not being honoured.
1593    
1594    12. Added an optional parentheses number to the -o and --only-matching options
1595        of pcregrep.
1596    
1597    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1598        can match an empty string. The code to do it in pcretest and pcredemo
1599        needed fixing:
1600    
1601        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1602            only one byte after an empty string match just before CRLF (this case
1603            just got forgotten; "any" and "anycrlf" were OK).
1604    
1605        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1606            mode when an empty string match preceded an ASCII character followed by
1607            a non-ASCII character. (The code for advancing by one character rather
1608            than one byte was nonsense.)
1609    
1610        (c) The pcredemo.c sample program did not have any code at all to handle
1611            the cases when CRLF is a valid newline sequence.
1612    
1613    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1614        as a starting offset was within the subject string. There is now a new
1615        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1616        negative or greater than the length of the string. In order to test this,
1617        pcretest is extended to allow the setting of negative starting offsets.
1618    
1619    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1620        starting offset points to the beginning of a UTF-8 character was
1621        unnecessarily clumsy. I tidied it up.
1622    
1623    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1624        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1625    
1626    17. Nobody had reported that the --include_dir option, which was added in
1627        release 7.7 should have been called --include-dir (hyphen, not underscore)
1628        for compatibility with GNU grep. I have changed it to --include-dir, but
1629        left --include_dir as an undocumented synonym, and the same for
1630        --exclude-dir, though that is not available in GNU grep, at least as of
1631        release 2.5.4.
1632    
1633    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1634        characters from a string of bytes) have been redefined so as not to use
1635        loops, in order to improve performance in some environments. At the same
1636        time, I abstracted some of the common code into auxiliary macros to save
1637        repetition (this should not affect the compiled code).
1638    
1639    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1640        compile-time error is now given if \c is not followed by an ASCII
1641        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1642        different, and any byte value is allowed.)
1643    
1644    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1645        START_OPTIMIZE option, which is now allowed at compile time - but just
1646        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1647        to pcregrep and other applications that have no direct access to PCRE
1648        options. The new /Y option in pcretest sets this option when calling
1649        pcre_compile().
1650    
1651    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1652        back references. Groups containing recursive back references were forced to
1653        be atomic by that change, but in the case of named groups, the amount of
1654        memory required was incorrectly computed, leading to "Failed: internal
1655        error: code overflow". This has been fixed.
1656    
1657    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1658        pcretest.c, to avoid build problems in some Borland environments.
1659    
1660    
1661    Version 8.10 25-Jun-2010
1662    ------------------------
1663    
1664    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1665        THEN.
1666    
1667    2.  (*ACCEPT) was not working when inside an atomic group.
1668    
1669    3.  Inside a character class, \B is treated as a literal by default, but
1670        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1671        causes the error). The code is unchanged, but I tidied the documentation.
1672    
1673    4.  Inside a character class, PCRE always treated \R and \X as literals,
1674        whereas Perl faults them if its -w option is set. I have changed PCRE so
1675        that it faults them when PCRE_EXTRA is set.
1676    
1677    5.  Added support for \N, which always matches any character other than
1678        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1679    
1680    6.  When compiling pcregrep with newer versions of gcc which may have
1681        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1682        declared with attribute warn_unused_result" were given. Just casting the
1683        result to (void) does not stop the warnings; a more elaborate fudge is
1684        needed. I've used a macro to implement this.
1685    
1686    7.  Minor change to pcretest.c to avoid a compiler warning.
1687    
1688    8.  Added four artifical Unicode properties to help with an option to make
1689        \s etc use properties (see next item). The new properties are: Xan
1690        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1691    
1692    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1693        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1694        this option. Modified pcretest to add /W to test this facility. Added
1695        REG_UCP to make it available via the POSIX interface.
1696    
1697    10. Added --line-buffered to pcregrep.
1698    
1699    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1700        studied, and the match started with a letter with a code point greater than
1701        127 whose first byte was different to the first byte of the other case of
1702        the letter, the other case of this starting letter was not recognized
1703        (#976).
1704    
1705    12. If a pattern that was studied started with a repeated Unicode property
1706        test, for example, \p{Nd}+, there was the theoretical possibility of
1707        setting up an incorrect bitmap of starting bytes, but fortunately it could
1708        not have actually happened in practice until change 8 above was made (it
1709        added property types that matched character-matching opcodes).
1710    
1711    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1712        possible starting bytes for non-anchored patterns.
1713    
1714    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1715        \R, and also a number of cases that involve Unicode properties, both
1716        explicit and implicit when PCRE_UCP is set.
1717    
1718    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1719        input, it could crash or give wrong results if characters with values
1720        greater than 0xc0 were present in the subject string. (Detail: it assumed
1721        UTF-8 input when processing these items.)
1722    
1723    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1724        size_t is 64-bit (#991).
1725    
1726    17. Added a check for running out of memory when PCRE is compiled with
1727        --disable-stack-for-recursion (#990).
1728    
1729    18. If the last data line in a file for pcretest does not have a newline on
1730        the end, a newline was missing in the output.
1731    
1732    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1733        less than 128) in its various bitmaps. However, there is a facility for
1734        generating tables according to the current locale when PCRE is compiled. It
1735        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1736        characters, are recognized by isspace() and therefore were getting set in
1737        these tables, and indeed these tables seem to approximate to ISO 8859. This
1738        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1739        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1740        which of course cannot start UTF-8 characters. I have changed the code so
1741        that only real ASCII characters (less than 128) and the correct starting
1742        bytes for UTF-8 encodings are set for characters greater than 127 when in
1743        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1744        altogether.)
1745    
1746    20. Added the /T option to pcretest so as to be able to run tests with non-
1747        standard character tables, thus making it possible to include the tests
1748        used for 19 above in the standard set of tests.
1749    
1750    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1751        reference to a subpattern the other side of a comment that contains an
1752        opening parenthesis caused either an internal compiling error, or a
1753        reference to the wrong subpattern.
1754    
1755    
1756    Version 8.02 19-Mar-2010
1757    ------------------------
1758    
1759    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1760    
1761    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1762        configured.
1763    
1764    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1765        original author of that file, following a query about its status.
1766    
1767    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1768        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1769    
1770    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1771        quantifier applied to a forward-referencing subroutine call, could compile
1772        incorrect code or give the error "internal error: previously-checked
1773        referenced subpattern not found".
1774    
1775    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1776        variables to point to external functions. For these systems, therefore,
1777        pcre_malloc etc. are now initialized to local functions that call the
1778        relevant global functions.
1779    
1780    7.  There were two entries missing in the vectors called coptable and poptable
1781        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1782        I've fixed the data, and added a kludgy way of testing at compile time that
1783        the lengths are correct (equal to the number of opcodes).
1784    
1785    8.  Following on from 7, I added a similar kludge to check the length of the
1786        eint vector in pcreposix.c.
1787    
1788    9.  Error texts for pcre_compile() are held as one long string to avoid too
1789        much relocation at load time. To find a text, the string is searched,
1790        counting zeros. There was no check for running off the end of the string,
1791        which could happen if a new error number was added without updating the
1792        string.
1793    
1794    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1795    
1796    11. \K was not working if it appeared in an atomic group or in a group that
1797        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1798        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1799        the assertion is positive, but not if it is negative.
1800    
1801    12. Change 11 fortuitously reduced the size of the stack frame used in the
1802        "match()" function of pcre_exec.c by one pointer. Forthcoming
1803        implementation of support for (*MARK) will need an extra pointer on the
1804        stack; I have reserved it now, so that the stack frame size does not
1805        decrease.
1806    
1807    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1808        item in branch that calls a recursion is a subroutine call - as in the
1809        second branch in the above example - was incorrectly given the compile-
1810        time error "recursive call could loop indefinitely" because pcre_compile()
1811        was not correctly checking the subroutine for matching a non-empty string.
1812    
1813    14. The checks for overrunning compiling workspace could trigger after an
1814        overrun had occurred. This is a "should never occur" error, but it can be
1815        triggered by pathological patterns such as hundreds of nested parentheses.
1816        The checks now trigger 100 bytes before the end of the workspace.
1817    
1818    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1819    
1820    
1821    Version 8.01 19-Jan-2010
1822    ------------------------
1823    
1824    1.  If a pattern contained a conditional subpattern with only one branch (in
1825        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1826        computed the wrong minimum data length (which is of course zero for such
1827        subpatterns). This could cause incorrect "no match" results.
1828    
1829    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1830        the pattern is reset in the first branch, pcre_compile() failed with
1831        "internal error: code overflow at offset...". This happened only when
1832        the reset was to the original external option setting. (An optimization
1833        abstracts leading options settings into an external setting, which was the
1834        cause of this.)
1835    
1836    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1837        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1838        assertion pattern did not match (meaning that the assertion was true), it
1839        was incorrectly treated as false if the SKIP had been reached during the
1840        matching. This also applied to assertions used as conditions.
1841    
1842    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1843        assertion subpattern, including such a pattern used as a condition,
1844        unpredictable results occurred, instead of the error return
1845        PCRE_ERROR_DFA_UITEM.
1846    
1847    5.  The C++ GlobalReplace function was not working like Perl for the special
1848        situation when an empty string is matched. It now does the fancy magic
1849        stuff that is necessary.
1850    
1851    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1852        removed. (These were left over from very, very early versions of PCRE.)
1853    
1854    7.  Some cosmetic changes to the code to make life easier when compiling it
1855        as part of something else:
1856    
1857        (a) Change DEBUG to PCRE_DEBUG.
1858    
1859        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1860            called "current" as "current_branch", to prevent a collision with the
1861            Linux macro when compiled as a kernel module.
1862    
1863        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1864            prevent a collision with the Linux macro when compiled as a kernel
1865            module.
1866    
1867    8.  In pcre_compile() there are some checks for integer overflows that used to
1868        cast potentially large values to (double). This has been changed to that
1869        when building, a check for int64_t is made, and if it is found, it is used
1870        instead, thus avoiding the use of floating point arithmetic. (There is no
1871        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1872        double.
1873    
1874    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1875        2005 (difference between two addresses compared to an unsigned value).
1876    
1877    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1878        custom one, because of the following reported problem in Windows:
1879    
1880          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1881              under Win32.
1882          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1883              therefore missing the function definition.
1884          - The compiler thus generates a "C" signature for the test function.
1885          - The linker fails to find the "C" function.
1886          - PCRE fails to configure if asked to do so against libbz2.
1887    
1888    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1889        messages were output:
1890    
1891          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1892          rerunning libtoolize, to keep the correct libtool macros in-tree.
1893          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1894    
1895        I have done both of these things.
1896    
1897    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1898        most of the time, it *can* run out if it is given a pattern that contains a
1899        runaway infinite recursion. I updated the discussion in the pcrestack man
1900        page.
1901    
1902    13. Now that we have gone to the x.xx style of version numbers, the minor
1903        version may start with zero. Using 08 or 09 is a bad idea because users
1904        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1905        interpreted as invalid octal numbers. I've updated the previous comment in
1906        configure.ac, and also added a check that gives an error if 08 or 09 are
1907        used.
1908    
1909    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1910        causing partial matching to fail when the end of the subject matched \W
1911        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1912    
1913    15. There were some discrepancies between the declarations in pcre_internal.h
1914        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1915        their definitions. The declarations used "const uschar *" and the
1916        definitions used USPTR. Even though USPTR is normally defined as "const
1917        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1918        reported that: "This difference in casting confuses some C++ compilers, for
1919        example, SunCC recognizes above declarations as different functions and
1920        generates broken code for hbpcre." I have changed the declarations to use
1921        USPTR.
1922    
1923    16. GNU libtool is named differently on some systems. The autogen.sh script now
1924        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1925        (FreeBSD).
1926    
1927    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1928        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1929        comment: "Figure out how to create a longlong from a string: strtoll and
1930        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1931        instance, but it only takes 2 args instead of 3!"
1932    
1933    18. A subtle bug concerned with back references has been fixed by a change of
1934        specification, with a corresponding code fix. A pattern such as
1935        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1936        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1937        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1938        same bug. Such groups have to be quantified to be useful, or contained
1939        inside another quantified group. (If there's no repetition, the reference
1940        can never match.) The problem arises because, having left the group and
1941        moved on to the rest of the pattern, a later failure that backtracks into
1942        the group uses the captured value from the final iteration of the group
1943        rather than the correct earlier one. I have fixed this in PCRE by forcing
1944        any group that contains a reference to itself to be an atomic group; that
1945        is, there cannot be any backtracking into it once it has completed. This is
1946        similar to recursive and subroutine calls.
1947    
1948    
1949    Version 8.00 19-Oct-09
1950    ----------------------
1951    
1952    1.  The table for translating pcre_compile() error codes into POSIX error codes
1953        was out-of-date, and there was no check on the pcre_compile() error code
1954        being within the table. This could lead to an OK return being given in
1955        error.
1956    
1957    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1958        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1959        in a Windows environment.
1960    
1961    3.  The pcregrep --count option prints the count for each file even when it is
1962        zero, as does GNU grep. However, pcregrep was also printing all files when
1963        --files-with-matches was added. Now, when both options are given, it prints
1964        counts only for those files that have at least one match. (GNU grep just
1965        prints the file name in this circumstance, but including the count seems
1966        more useful - otherwise, why use --count?) Also ensured that the
1967        combination -clh just lists non-zero counts, with no names.
1968    
1969    4.  The long form of the pcregrep -F option was incorrectly implemented as
1970        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1971        but it seems right to fix it, and I didn't think it was worth preserving
1972        the old behaviour.
1973    
1974    5.  The command line items --regex=pattern and --regexp=pattern were not
1975        recognized by pcregrep, which required --regex pattern or --regexp pattern
1976        (with a space rather than an '='). The man page documented the '=' forms,
1977        which are compatible with GNU grep; these now work.
1978    
1979    6.  No libpcreposix.pc file was created for pkg-config; there was just
1980        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1981    
1982    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1983        when UCP support is not needed, by modifying the Python script that
1984        generates it from Unicode data files. This should not matter if the module
1985        is correctly used as a library, but I received one complaint about 50K of
1986        unwanted data. My guess is that the person linked everything into his
1987        program rather than using a library. Anyway, it does no harm.
1988    
1989    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1990        was a minimum greater than 1 for a wide character in a possessive
1991        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1992        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1993        character. Chaos in the form of incorrect output or a compiling loop could
1994        result.
1995    
1996    9.  The restrictions on what a pattern can contain when partial matching is
1997        requested for pcre_exec() have been removed. All patterns can now be
1998        partially matched by this function. In addition, if there are at least two
1999        slots in the offset vector, the offset of the earliest inspected character
2000        for the match and the offset of the end of the subject are set in them when
2001        PCRE_ERROR_PARTIAL is returned.
2002    
2003    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2004        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2005        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2006        and may be more useful for multi-segment matching.
2007    
2008    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2009        used to be given if ever the end of the subject was reached; now it is
2010        given only if matching could not proceed because another character was
2011        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2012        string "Z", which now yields "no match" instead of "partial match". In the
2013        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2014        final character ended with (*FAIL).
2015    
2016    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2017        if the pattern had a "must contain" character that was already found in the
2018        earlier partial match, unless partial matching was again requested. For
2019        example, with the pattern /dog.(body)?/, the "must contain" character is
2020        "g". If the first part-match was for the string "dog", restarting with
2021        "sbody" failed. This bug has been fixed.
2022    
2023    13. The string returned by pcre_dfa_exec() after a partial match has been
2024        changed so that it starts at the first inspected character rather than the
2025        first character of the match. This makes a difference only if the pattern
2026        starts with a lookbehind assertion or \b or \B (\K is not supported by
2027        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2028        matching functions compatible, and I think it's the right thing to do.
2029    
2030    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2031        so that the demonstration program is easily available in environments where
2032        PCRE has not been installed from source.
2033    
2034    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2035        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2036        library.
2037    
2038    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2039        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2040        is not the first non-POSIX option to be added. Clearly some people find
2041        these options useful.
2042    
2043    17. If a caller to the POSIX matching function regexec() passes a non-zero
2044        value for nmatch with a NULL value for pmatch, the value of
2045        nmatch is forced to zero.
2046    
2047    18. RunGrepTest did not have a test for the availability of the -u option of
2048        the diff command, as RunTest does. It now checks in the same way as
2049        RunTest, and also checks for the -b option.
2050    
2051    19. If an odd number of negated classes containing just a single character
2052        interposed, within parentheses, between a forward reference to a named
2053        subpattern and the definition of the subpattern, compilation crashed with
2054        an internal error, complaining that it could not find the referenced
2055        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2056        [The bug was that it was starting one character too far in when skipping
2057        over the character class, thus treating the ] as data rather than
2058        terminating the class. This meant it could skip too much.]
2059    
2060    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2061        /g option in pcretest when the pattern contains \K, which makes it possible
2062        to have an empty string match not at the start, even when the pattern is
2063        anchored. Updated pcretest and pcredemo to use this option.
2064    
2065    21. If the maximum number of capturing subpatterns in a recursion was greater
2066        than the maximum at the outer level, the higher number was returned, but
2067        with unset values at the outer level. The correct (outer level) value is
2068        now given.
2069    
2070    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2071        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2072        make it do so. The string so far is captured, making this feature
2073        compatible with Perl.
2074    
2075    23. The tests have been re-organized, adding tests 11 and 12, to make it
2076        possible to check the Perl 5.10 features against Perl 5.10.
2077    
2078    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2079        pattern matches a fixed length string. PCRE did not allow this; now it
2080        does. Neither allows recursion.
2081    
2082    25. I finally figured out how to implement a request to provide the minimum
2083        length of subject string that was needed in order to match a given pattern.
2084        (It was back references and recursion that I had previously got hung up
2085        on.) This code has now been added to pcre_study(); it finds a lower bound
2086        to the length of subject needed. It is not necessarily the greatest lower
2087        bound, but using it to avoid searching strings that are too short does give
2088        some useful speed-ups. The value is available to calling programs via
2089        pcre_fullinfo().
2090    
2091    26. While implementing 25, I discovered to my embarrassment that pcretest had
2092        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2093        study optimizations had never been tested with that matching function.
2094        Oops. What is worse, even when it was passed study data, there was a bug in
2095        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2096        were also very few tests of studied patterns with pcre_dfa_exec().
2097    
2098    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2099        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2100        on the other side of the coin, they are no longer allowed to have different
2101        names, because these cannot be distinguished in PCRE, and this has caused
2102        confusion. (This is a difference from Perl.)
2103    
2104    28. When duplicate subpattern names are present (necessarily with different
2105        numbers, as required by 27 above), and a test is made by name in a
2106        conditional pattern, either for a subpattern having been matched, or for
2107        recursion in such a pattern, all the associated numbered subpatterns are
2108        tested, and the overall condition is true if the condition is true for any
2109        one of them. This is the way Perl works, and is also more like the way
2110        testing by number works.
2111    
2112    
2113    Version 7.9 11-Apr-09
2114    ---------------------
2115    
2116    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2117        (pcretest), all targets were linked against these libraries. This included
2118        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2119        libraries. This caused unwanted dependencies to be created. This problem
2120        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2121        pcretest is linked with readline.
2122    
2123    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2124        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2125        moved outside it again, because FALSE and TRUE are already defined in AIX,
2126        but BOOL is not.
2127    
2128    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2129        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2130    
2131    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2132        hyphens) following file names and line numbers when outputting matching
2133        lines. This is not true; no spaces are inserted. I have also clarified the
2134        wording for the --colour (or --color) option.
2135    
2136    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2137        was not coloured; this is different to GNU grep, so I have changed it to be
2138        the same.
2139    
2140    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2141        each matching line was coloured. Now it goes on to look for further matches
2142        of any of the test patterns, which is the same behaviour as GNU grep.
2143    
2144    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2145        doesn't make sense to accept an empty string match in pcregrep, so I have
2146        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2147        seems to be how GNU grep behaves. [But see later change 40 for release
2148        8.33.]
2149    
2150    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2151        start or after a newline", because the conditional assertion was not being
2152        correctly handled. The rule now is that both the assertion and what follows
2153        in the first alternative must satisfy the test.
2154    
2155    9.  If auto-callout was enabled in a pattern with a conditional group whose
2156        condition was an assertion, PCRE could crash during matching, both with
2157        pcre_exec() and pcre_dfa_exec().
2158    
2159    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2160        used for matching.
2161    
2162    11. Unicode property support in character classes was not working for
2163        characters (bytes) greater than 127 when not in UTF-8 mode.
2164    
2165    12. Added the -M command line option to pcretest.
2166    
2167    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2168    
2169    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2170    
2171    16. Added comments and documentation about mis-use of no_arg in the C++
2172        wrapper.
2173    
2174    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2175        from Martin Jerabek that uses macro names for all relevant character and
2176        string constants.
2177    
2178    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2179        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2180        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2181        these, but not everybody uses configure.
2182    
2183    19. A conditional group that had only one branch was not being correctly
2184        recognized as an item that could match an empty string. This meant that an
2185        enclosing group might also not be so recognized, causing infinite looping
2186        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2187        with the subject "ab", where knowledge that the repeated group can match
2188        nothing is needed in order to break the loop.
2189    
2190    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2191        exec(), but without supplying a callout function, matching went wrong.
2192    
2193    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2194        leak if the size of the offset vector was greater than 30. When the vector
2195        is smaller, the saved offsets during recursion go onto a local stack
2196        vector, but for larger vectors malloc() is used. It was failing to free
2197        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2198        error, in fact).
2199    
2200    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2201        heapframe that is used only when UTF-8 support is enabled. This caused no
2202        problem, but was untidy.
2203    
2204    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2205        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2206        included within another project.
2207    
2208    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2209        slightly modified by me:
2210    
2211          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2212              not building pcregrep.
2213    
2214          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2215              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2216    
2217    25. Forward references, both numeric and by name, in patterns that made use of
2218        duplicate group numbers, could behave incorrectly or give incorrect errors,
2219        because when scanning forward to find the reference group, PCRE was not
2220        taking into account the duplicate group numbers. A pattern such as
2221        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2222    
2223    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2224        the feature of a custom pointer more persuasive (as requested by a user).
2225    
2226    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2227        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2228        pre-defined.
2229    
2230    28. Added support for (*UTF8) at the start of a pattern.
2231    
2232    29. Arrange for flags added by the "release type" setting in CMake to be shown
2233        in the configuration summary.
2234    
2235    
2236    Version 7.8 05-Sep-08
2237    ---------------------
2238    
2239    1.  Replaced UCP searching code with optimized version as implemented for Ad
2240        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2241        stage table and inline lookup instead of a function, giving speed ups of 2
2242        to 5 times on some simple patterns that I tested. Permission was given to
2243        distribute the MultiStage2.py script that generates the tables (it's not in
2244        the tarball, but is in the Subversion repository).
2245    
2246    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2247        scripts.
2248    
2249    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2250        a group with a zero qualifier. The result of the study could be incorrect,
2251        or the function might crash, depending on the pattern.
2252    
2253    4.  Caseless matching was not working for non-ASCII characters in back
2254        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2255        It now works when Unicode Property Support is available.
2256    
2257    5.  In pcretest, an escape such as \x{de} in the data was always generating
2258        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2259        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2260        truncation.
2261    
2262    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2263    
2264    7.  Added two (int) casts to pcregrep when printing the difference of two
2265        pointers, in case they are 64-bit values.
2266    
2267    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2268        test 2 if it fails.
2269    
2270    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2271        and a #define of that name to empty if it is not externally set. This is to
2272        allow users of MSVC to set it if necessary.
2273    
2274    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2275        the convenience functions in the pcre_get.c source file.
2276    
2277    11. An option change at the start of a pattern that had top-level alternatives
2278        could cause overwriting and/or a crash. This command provoked a crash in
2279        some environments:
2280    
2281          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2282    
2283        This potential security problem was recorded as CVE-2008-2371.
2284    
2285    12. For a pattern where the match had to start at the beginning or immediately
2286        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2287        pcre_dfa_exec() could read past the end of the passed subject if there was
2288        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2289        pcretest so that it places the subject at the end of its malloc-ed buffer.
2290    
2291    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2292        exec() might read past the end of the data buffer in UTF-8 mode.
2293    
2294    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2295        the data contained the byte 0x85 as part of a UTF-8 character within its
2296        first line. This applied both to normal and DFA matching.
2297    
2298    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2299        /^[^d]*?$/8 failed to match "abc".
2300    
2301    16. Added a missing copyright notice to pcrecpp_internal.h.
2302    
2303    17. Make it more clear in the documentation that values returned from
2304        pcre_exec() in ovector are byte offsets, not character counts.
2305    
2306    18. Tidied a few places to stop certain compilers from issuing warnings.
2307    
2308    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2309        supplied by Stefan Weber. I made a further small update for 7.8 because
2310        there is a change of source arrangements: the pcre_searchfuncs.c module is
2311        replaced by pcre_ucd.c.
2312    
2313    
2314    Version 7.7 07-May-08
2315    ---------------------
2316    
2317    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2318        a string to a long long, pretend we don't even have a long long." This is
2319        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2320    
2321    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2322        pre-7.6 versions, which defined a global no_arg variable instead of putting
2323        it in the RE class. (See also #8 below.)
2324    
2325    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2326        Lopes.
2327    
2328    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2329    
2330        (1) The include/exclude patterns were being applied to the whole pathnames
2331            of files, instead of just to the final components.
2332    
2333        (2) If there was more than one level of directory, the subdirectories were
2334            skipped unless they satisfied the include/exclude conditions. This is
2335            inconsistent with GNU grep (and could even be seen as contrary to the
2336            pcregrep specification - which I improved to make it absolutely clear).
2337            The action now is always to scan all levels of directory, and just
2338            apply the include/exclude patterns to regular files.
2339    
2340    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2341        --exclude_dir in the tests to avoid scanning .svn directories.
2342    
2343    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2344        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2345        doesn't support NULs in patterns.
2346    
2347    7.  Added some missing "const"s to declarations of static tables in
2348        pcre_compile.c and pcre_dfa_exec.c.
2349    
2350    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2351        caused by fix #2  above. (Subsequently also a second patch to fix the
2352        first patch. And a third patch - this was a messy problem.)
2353    
2354    9.  Applied Craig's patch to remove the use of push_back().
2355    
2356    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2357        matching function regexec().
2358    
2359    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2360        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2361        references. PCRE supports relative numbers with this syntax (I don't think
2362        Oniguruma does).
2363    
2364    12. Previously, a group with a zero repeat such as (...){0} was completely
2365        omitted from the compiled regex. However, this means that if the group
2366        was called as a subroutine from elsewhere in the pattern, things went wrong
2367        (an internal error was given). Such groups are now left in the compiled
2368        pattern, with a new opcode that causes them to be skipped at execution
2369        time.
2370    
2371    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2372        to the way PCRE behaves:
2373    
2374        (a) A lone ] character is dis-allowed (Perl treats it as data).
2375    
2376        (b) A back reference to an unmatched subpattern matches an empty string
2377            (Perl fails the current match path).
2378    
2379        (c) A data ] in a character class must be notated as \] because if the
2380            first data character in a class is ], it defines an empty class. (In
2381            Perl it is not possible to have an empty class.) The empty class []
2382            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2383            The negative empty class [^] matches any one character, independently
2384            of the DOTALL setting.
2385    
2386    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2387        non-existent subpattern following a character class starting with ']' and
2388        containing () gave an internal compiling error instead of "reference to
2389        non-existent subpattern". Fortunately, when the pattern did exist, the
2390        compiled code was correct. (When scanning forwards to check for the
2391        existence of the subpattern, it was treating the data ']' as terminating
2392        the class, so got the count wrong. When actually compiling, the reference
2393        was subsequently set up correctly.)
2394    
2395    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2396        it was being rejected as not supported by pcre_dfa_exec(), even though
2397        other assertions are supported. I have made pcre_dfa_exec() support
2398        (*FAIL).
2399    
2400    16. The implementation of 13c above involved the invention of a new opcode,
2401        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2402        cannot be changed at match time, I realized I could make a small
2403        improvement to matching performance by compiling OP_ALLANY instead of
2404        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2405        on the OP_ANY path.
2406    
2407    17. Compiling pcretest on Windows with readline support failed without the
2408        following two fixes: (1) Make the unistd.h include conditional on
2409        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2410    
2411    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2412        ncurses library to be included for pcretest when ReadLine support is
2413        requested, but also to allow for it to be overridden. This patch came from
2414        Daniel Bergström.
2415    
2416    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2417        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2418        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2419        spotting this.
2420    
2421    
2422    Version 7.6 28-Jan-08
2423    ---------------------
2424    
2425    1.  A character class containing a very large number of characters with
2426        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2427        overflow.
2428    
2429    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2430        HAVE_LONG_LONG is not defined.
2431    
2432    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2433        bring them up to date and include new features. This patch includes:
2434    
2435        - Fixed PH's badly added libz and libbz2 support.
2436        - Fixed a problem with static linking.
2437        - Added pcredemo. [But later removed - see 7 below.]
2438        - Fixed dftables problem and added an option.
2439        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2440            HAVE_LONG_LONG.
2441        - Added readline support for pcretest.
2442        - Added an listing of the option settings after cmake has run.
2443    
2444    4.  A user submitted a patch to Makefile that makes it easy to create
2445        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2446        Makefile.am that cause it to include this special target, without
2447        affecting anything else. Note that the same mingw target plus all
2448        the other distribution libraries and programs are now supported
2449        when configuring with CMake (see 6 below) instead of with
2450        Configure/Make.
2451    
2452    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2453        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2454        exported in the Windows port". It has not yet been confirmed that the patch
2455        solves the problem, but it does no harm.
2456    
2457    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2458        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2459        with CMake, and also correct the comment about stack recursion.
2460    
2461    7.  Remove the automatic building of pcredemo from the ./configure system and
2462        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2463        of a program that users should build themselves after PCRE is installed, so
2464        building it automatically is not really right. What is more, it gave
2465        trouble in some build environments.
2466    
2467    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2468    
2469    
2470    Version 7.5 10-Jan-08
2471    ---------------------
2472    
2473    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2474        values in parens when parsing an RE using the C++ wrapper."
2475    
2476    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2477        Characters greater than 255 were excluded from the class instead of being
2478        included.
2479    
2480    3.  The same bug as (2) above applied to negated POSIX classes such as
2481        [:^space:].
2482    
2483    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2484        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2485        I have changed it.
2486    
2487    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2488        first named subpattern) and a construct such as (?&a) would reference the
2489        first named subpattern whose name started with "a" (in other words, the
2490        length check was missing). Both these problems are fixed. "Subpattern name
2491        expected" is now given for (?&) (a zero-length name), and this patch also
2492        makes it give the same error for \k'' (previously it complained that that
2493        was a reference to a non-existent subpattern).
2494    
2495    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2496        this is right because (?- can be followed by option settings as well as by
2497        digits. I have, however, made the messages clearer.
2498    
2499    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2500        than the number used in the conditional) now cause a compile-time error.
2501        This is actually not compatible with Perl, which accepts such patterns, but
2502        treats the conditional as always being FALSE (as PCRE used to), but it
2503        seems to me that giving a diagnostic is better.
2504    
2505    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2506        and messages.
2507    
2508    9.  Fix two occurrences of "backslash" in comments that should have been
2509        "backspace".
2510    
2511    10. Remove two redundant lines of code that can never be obeyed (their function
2512        was moved elsewhere).
2513    
2514    11. The program that makes PCRE's Unicode character property table had a bug
2515        which caused it to generate incorrect table entries for sequences of
2516        characters that have the same character type, but are in different scripts.
2517        It amalgamated them into a single range, with the script of the first of
2518        them. In other words, some characters were in the wrong script. There were
2519        thirteen such cases, affecting characters in the following ranges:
2520    
2521          U+002b0 - U+002c1
2522          U+0060c - U+0060d
2523          U+0061e - U+00612
2524          U+0064b - U+0065e
2525          U+0074d - U+0076d
2526          U+01800 - U+01805
2527          U+01d00 - U+01d77
2528          U+01d9b - U+01dbf
2529          U+0200b - U+0200f
2530          U+030fc - U+030fe
2531          U+03260 - U+0327f
2532          U+0fb46 - U+0fbb1
2533          U+10450 - U+1049d
2534    
2535    12. The -o option (show only the matching part of a line) for pcregrep was not
2536        compatible with GNU grep in that, if there was more than one match in a
2537        line, it showed only the first of them. It now behaves in the same way as
2538        GNU grep.
2539    
2540    13. If the -o and -v options were combined for pcregrep, it printed a blank
2541        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2542        does the same. The return code can be used to tell if there were any
2543        non-matching lines.
2544    
2545    14. Added --file-offsets and --line-offsets to pcregrep.
2546    
2547    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2548        infinitely looping recursion. The bug was that positive lookaheads were not
2549        being skipped when checking for a possible empty match (negative lookaheads
2550        and both kinds of lookbehind were skipped).
2551    
2552    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2553        inclusion of <windows.h> to before rather than after the definition of
2554        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2555    
2556    17. Specifying a possessive quantifier with a specific limit for a Unicode
2557        character property caused pcre_compile() to compile bad code, which led at
2558        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2559        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2560        caused the error; without that there was no problem.
2561    
2562    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2563    
2564    19. Added --enable-pcretest-libreadline.
2565    
2566    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2567        RE::GlobalReplace(). As a result, the number of replacements returned was
2568        double what it should be. I removed one of the increments, but Craig sent a
2569        later patch that removed the other one (the right fix) and added unit tests
2570        that check the return values (which was not done before).
2571    
2572    21. Several CMake things:
2573    
2574        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2575            the names libpcre and libpcreposix, not just pcre and pcreposix.
2576    
2577        (2) The above change means that pcretest and pcregrep are now correctly
2578            linked with the newly-built libraries, not previously installed ones.
2579    
2580        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2581    
2582    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2583        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2584        UTF-8 newline character). The key issue is that the pattern starts .*;
2585        this means that the match must be either at the beginning, or after a
2586        newline. The bug was in the code for advancing after a failed match and
2587        checking that the new position followed a newline. It was not taking
2588        account of UTF-8 characters correctly.
2589    
2590    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2591        character classes. PCRE was not treating the sequence [:...:] as a
2592        character class unless the ... were all letters. Perl, however, seems to
2593        allow any characters between [: and :], though of course it rejects as
2594        unknown any "names" that contain non-letters, because all the known class
2595        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2596        for example, whereas PCRE did not - it did not recognize a POSIX character
2597        class. This seemed a bit dangerous, so the code has been changed to be
2598        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2599        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2600        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2601        Perl does, and where it didn't before.
2602    
2603    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2604        Windows environments %n is disabled by default.
2605    
2606    
2607    Version 7.4 21-Sep-07
2608    ---------------------
2609    
2610    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2611        means that a class such as [\s] counted as "explicit reference to CR or
2612        LF". That isn't really right - the whole point of the change was to try to
2613        help when there was an actual mention of one of the two characters. So now
2614        the change happens only if \r or \n (or a literal CR or LF) character is
2615        encountered.
2616    
2617    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2618        of both had grown to the point where there were only 3 bits left.
2619        Fortunately, there was spare space in the data structure, and so I have
2620        moved the internal flags into a new 16-bit field to free up more option
2621        bits.
2622    
2623    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2624        but did not set the internal JCHANGED flag - either of these is enough to
2625        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2626        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2627        start sets both bits.
2628    
2629    4.  Added options (at build time, compile time, exec time) to change \R from
2630        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2631    
2632    5.  doc/pcresyntax.html was missing from the distribution.
2633    
2634    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2635        compatibility, even though it is no longer used.
2636    
2637    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2638        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2639        windows.h file is present (where different names are used). [This was
2640        reversed later after testing - see 16 below.]
2641    
2642    8.  Changed all #include <config.h> to #include "config.h". There were also
2643        some further <pcre.h> cases that I changed to "pcre.h".
2644    
2645    9.  When pcregrep was used with the --colour option, it missed the line ending
2646        sequence off the lines that it output.
2647    
2648    10. It was pointed out to me that arrays of string pointers cause lots of
2649        relocations when a shared library is dynamically loaded. A technique of
2650        using a single long string with a table of offsets can drastically reduce
2651        these. I have refactored PCRE in four places to do this. The result is
2652        dramatic:
2653    
2654          Originally:                          290
2655          After changing UCP table:            187
2656          After changing error message table:   43
2657          After changing table of "verbs"       36
2658          After changing table of Posix names   22
2659    
2660        Thanks to the folks working on Gregex for glib for this insight.
2661    
2662    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2663        unicode-properties was also set.
2664    
2665    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2666    
2667    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2668        checked only for CRLF.
2669    
2670    14. Added casts to pcretest.c to avoid compiler warnings.
2671    
2672    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2673    
2674    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2675        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2676        entirely. This removes changes made in 7 above.
2677    
2678    17. The CMake files have been updated, and there is now more information about
2679        building with CMake in the NON-UNIX-USE document.
2680    
2681    
2682    Version 7.3 28-Aug-07
2683  ---------------------  ---------------------
2684    
2685   1. In the rejigging of the build system that eventually resulted in 7.1, the   1. In the rejigging of the build system that eventually resulted in 7.1, the
# Line 143  Version 7.3 17-Aug-07 Line 2821  Version 7.3 17-Aug-07
2821  24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte  24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2822      characters were involved (for example /\S{2}/8g with "A\x{a3}BC").      characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2823    
2824    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2825    
2826    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2827        character were causing crashes (broken optimization).
2828    
2829    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2830        \p or \P) caused a compile-time loop.
2831    
2832    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2833        break. For example, the unstudied pattern [\r\n]A does not match the string
2834        "\r\nA" because change 7.0/46 below moves the current point on by two
2835        characters after failing to match at the start. However, the pattern \nA
2836        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2837        the same is true. There doesn't seem any very clean way out of this, but
2838        what I have chosen to do makes the common cases work: PCRE now takes note
2839        of whether there can be an explicit match for \r or \n anywhere in the
2840        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2841        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2842        pattern has explicit CR or LF references.
2843    
2844    29. Added (*CR) etc for changing newline setting at start of pattern.
2845    
2846    
2847  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07
2848  ---------------------  ---------------------

Legend:
Removed from v.221  
changed lines
  Added in v.1520

  ViewVC Help
Powered by ViewVC 1.1.5