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

Diff of /code/trunk/ChangeLog

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

revision 5 by nigel, Sat Feb 24 21:38:05 2007 UTC revision 1523 by ph10, Sun Feb 8 18:00:45 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    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    9. A reference to a duplicated named group (either a back reference or a test
47       for being set in a conditional) that occurred in a part of the pattern where
48       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
49       to be incorrectly calculated, leading to overwriting.
50    
51    10. A mutually recursive set of back references such as (\2)(\1) caused a
52        segfault at study time (while trying to find the minimum matching length).
53        The infinite loop is now broken (with the minimum length unset, that is,
54        zero).
55    
56    11. If an assertion that was used as a condition was quantified with a minimum
57        of zero, matching went wrong. In particular, if the whole group had
58        unlimited repetition and could match an empty string, a segfault was
59        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
60        assertions to be quantified, but not if they are being used as conditions,
61        so the above pattern is faulted by Perl. PCRE has now been changed so that
62        it also rejects such patterns.
63    
64    
65    Version 8.36 26-September-2014
66    ------------------------------
67    
68    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
69        -Wmissing-field-initializers and -Wunused-parameter.
70    
71    2.  The tests for quantifiers being too big (greater than 65535) were being
72        applied after reading the number, and stupidly assuming that integer
73        overflow would give a negative number. The tests are now applied as the
74        numbers are read.
75    
76    3.  Tidy code in pcre_exec.c where two branches that used to be different are
77        now the same.
78    
79    4.  The JIT compiler did not generate match limit checks for certain
80        bracketed expressions with quantifiers. This may lead to exponential
81        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
82        issue should be resolved now.
83    
84    5.  Fixed an issue, which occures when nested alternatives are optimized
85        with table jumps.
86    
87    6.  Inserted two casts and changed some ints to size_t in the light of some
88        reported 64-bit compiler warnings (Bugzilla 1477).
89    
90    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
91        an empty string, which sometimes were behaving incorrectly in the
92        interpreter (though correctly in the JIT matcher). This pcretest input is
93        an example:
94    
95          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
96          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
97    
98        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
99        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
100        for an empty string was breaking the inner loop and carrying on at a lower
101        level, when possessive repeated groups should always return to a higher
102        level as they have no backtrack points in them. The empty string test now
103        occurs at the outer level.
104    
105    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
106        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
107    
108    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
109        doesn't).
110    
111    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
112        to hold a pattern to be incorrectly computed (too small) when there were
113        named back references to duplicated names. This could cause "internal
114        error: code overflow" or "double free or corruption" or other memory
115        handling errors.
116    
117    11. When named subpatterns had the same prefixes, back references could be
118        confused. For example, in this pattern:
119    
120          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
121    
122        the reference to 'Name' was incorrectly treated as a reference to a
123        duplicate name.
124    
125    12. A pattern such as /^s?c/mi8 where the optional character has more than
126        one "other case" was incorrectly compiled such that it would only try to
127        match starting at "c".
128    
129    13. When a pattern starting with \s was studied, VT was not included in the
130        list of possible starting characters; this should have been part of the
131        8.34/18 patch.
132    
133    14. If a character class started [\Qx]... where x is any character, the class
134        was incorrectly terminated at the ].
135    
136    15. If a pattern that started with a caseless match for a character with more
137        than one "other case" was studied, PCRE did not set up the starting code
138        unit bit map for the list of possible characters. Now it does. This is an
139        optimization improvement, not a bug fix.
140    
141    16. The Unicode data tables have been updated to Unicode 7.0.0.
142    
143    17. Fixed a number of memory leaks in pcregrep.
144    
145    18. Avoid a compiler warning (from some compilers) for a function call with
146        a cast that removes "const" from an lvalue by using an intermediate
147        variable (to which the compiler does not object).
148    
149    19. Incorrect code was compiled if a group that contained an internal recursive
150        back reference was optional (had quantifier with a minimum of zero). This
151        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
152        caused segmentation faults because of stack overflows at compile time.
153    
154    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
155        group that is quantified with an indefinite repeat, caused a compile-time
156        loop which used up all the system stack and provoked a segmentation fault.
157        This was not the same bug as 19 above.
158    
159    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
160        Patch by Mike Frysinger.
161    
162    
163    Version 8.35 04-April-2014
164    --------------------------
165    
166    1.  A new flag is set, when property checks are present in an XCLASS.
167        When this flag is not set, PCRE can perform certain optimizations
168        such as studying these XCLASS-es.
169    
170    2.  The auto-possessification of character sets were improved: a normal
171        and an extended character set can be compared now. Furthermore
172        the JIT compiler optimizes more character set checks.
173    
174    3.  Got rid of some compiler warnings for potentially uninitialized variables
175        that show up only when compiled with -O2.
176    
177    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
178        of a match later then the end of the match. The pcretest program was not
179        handling the case sensibly - it was outputting from the start to the next
180        binary zero. It now reports this situation in a message, and outputs the
181        text from the end to the start.
182    
183    5.  Fast forward search is improved in JIT. Instead of the first three
184        characters, any three characters with fixed position can be searched.
185        Search order: first, last, middle.
186    
187    6.  Improve character range checks in JIT. Characters are read by an inprecise
188        function now, which returns with an unknown value if the character code is
189        above a certain threshold (e.g: 256). The only limitation is that the value
190        must be bigger than the threshold as well. This function is useful when
191        the characters above the threshold are handled in the same way.
192    
193    7.  The macros whose names start with RAWUCHAR are placeholders for a future
194        mode in which only the bottom 21 bits of 32-bit data items are used. To
195        make this more memorable for those maintaining the code, the names have
196        been changed to start with UCHAR21, and an extensive comment has been added
197        to their definition.
198    
199    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
200        to the export list in Makefile.am (they were accidentally omitted from the
201        8.34 tarball).
202    
203    9.  The informational output from pcretest used the phrase "starting byte set"
204        which is inappropriate for the 16-bit and 32-bit libraries. As the output
205        for "first char" and "need char" really means "non-UTF-char", I've changed
206        "byte" to "char", and slightly reworded the output. The documentation about
207        these values has also been (I hope) clarified.
208    
209    10. Another JIT related optimization: use table jumps for selecting the correct
210        backtracking path, when more than four alternatives are present inside a
211        bracket.
212    
213    11. Empty match is not possible, when the minimum length is greater than zero,
214        and there is no \K in the pattern. JIT should avoid empty match checks in
215        such cases.
216    
217    12. In a caseless character class with UCP support, when a character with more
218        than one alternative case was not the first character of a range, not all
219        the alternative cases were added to the class. For example, s and \x{17f}
220        are both alternative cases for S: the class [RST] was handled correctly,
221        but [R-T] was not.
222    
223    13. The configure.ac file always checked for pthread support when JIT was
224        enabled. This is not used in Windows, so I have put this test inside a
225        check for the presence of windows.h (which was already tested for).
226    
227    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
228        The algorithm provides a way to skip certain starting offsets, and usually
229        faster than linear prefix searches.
230    
231    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
232        as for 'fr_FR' and 'french'. For some reason, however, it then used the
233        Windows-specific input and output files, which have 'french' screwed in.
234        So this could never have worked. One of the problems with locales is that
235        they aren't always the same. I have now updated RunTest so that it checks
236        the output of the locale test (test 3) against three different output
237        files, and it allows the test to pass if any one of them matches. With luck
238        this should make the test pass on some versions of Solaris where it was
239        failing. Because of the uncertainty, the script did not used to stop if
240        test 3 failed; it now does. If further versions of a French locale ever
241        come to light, they can now easily be added.
242    
243    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
244        there was a message during ./configure, but it did not stop. This now
245        provokes an error. The invalid example in README has been corrected.
246        If a value less than the minimum is given, the minimum value has always
247        been used, but now a warning is given.
248    
249    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
250        was a bug in the test system, which is now fixed. Also, the list of various
251        configurations that are tested for each release did not have one with both
252        16/32 bits and --enable-bar-anycrlf. It now does.
253    
254    18. pcretest was missing "-C bsr" for displaying the \R default setting.
255    
256    19. Little endian PowerPC systems are supported now by the JIT compiler.
257    
258    20. The fast forward newline mechanism could enter to an infinite loop on
259        certain invalid UTF-8 input. Although we don't support these cases
260        this issue can be fixed by a performance optimization.
261    
262    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
263        not take account if existing stack usage. There is now a new global
264        variable called pcre_stack_guard that can be set to point to an external
265        function to check stack availability. It is called at the start of
266        processing every parenthesized group.
267    
268    22. A typo in the code meant that in ungreedy mode the max/min qualifier
269        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
270        not match "ab".
271    
272    23. When UTF was disabled, the JIT program reported some incorrect compile
273        errors. These messages are silenced now.
274    
275    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
276        compiler.
277    
278    25. Change all the temporary files used in RunGrepTest to be different to those
279        used by RunTest so that the tests can be run simultaneously, for example by
280        "make -j check".
281    
282    
283    Version 8.34 15-December-2013
284    -----------------------------
285    
286    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
287        executable memory. Patch inspired by Carsten Klein.
288    
289    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
290        this macro is never tested and has no effect, because the work to support
291        coverage involves only compiling and linking options and special targets in
292        the Makefile. The comment in config.h implied that defining the macro would
293        enable coverage support, which is totally false. There was also support for
294        setting this macro in the CMake files (my fault, I just copied it from
295        configure). SUPPORT_GCOV has now been removed.
296    
297    3.  Make a small performance improvement in strlen16() and strlen32() in
298        pcretest.
299    
300    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
301        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
302        these cases). There was also one in pcretest.c.
303    
304    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
305    
306    6.  In UTF mode, the code for checking whether a group could match an empty
307        string (which is used for indefinitely repeated groups to allow for
308        breaking an infinite loop) was broken when the group contained a repeated
309        negated single-character class with a character that occupied more than one
310        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
311        UTF-8 mode). The effect was undefined: the group might or might not be
312        deemed as matching an empty string, or the program might have crashed.
313    
314    7.  The code for checking whether a group could match an empty string was not
315        recognizing that \h, \H, \v, \V, and \R must match a character.
316    
317    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
318        an empty string. If it can, pcretest shows this in its information output.
319    
320    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
321        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
322        matched by pcre_exec() without using JIT:
323    
324        (a) If the rest of the pattern did not match after a maximal run of
325            grapheme clusters, the code for backing up to try with fewer of them
326            did not always back up over a full grapheme when characters that do not
327            have the modifier quality were involved, e.g. Hangul syllables.
328    
329        (b) If the match point in a subject started with modifier character, and
330            there was no match, the code could incorrectly back up beyond the match
331            point, and potentially beyond the first character in the subject,
332            leading to a segfault or an incorrect match result.
333    
334    10. A conditional group with an assertion condition could lead to PCRE
335        recording an incorrect first data item for a match if no other first data
336        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
337        first data item, and therefore matched "ca" after "c" instead of at the
338        start.
339    
340    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
341        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
342    
343    12. The source of pcregrep now includes z/OS-specific code so that it can be
344        compiled for z/OS as part of the special z/OS distribution.
345    
346    13. Added the -T and -TM options to pcretest.
347    
348    14. The code in pcre_compile.c for creating the table of named capturing groups
349        has been refactored. Instead of creating the table dynamically during the
350        actual compiling pass, the information is remembered during the pre-compile
351        pass (on the stack unless there are more than 20 named groups, in which
352        case malloc() is used) and the whole table is created before the actual
353        compile happens. This has simplified the code (it is now nearly 150 lines
354        shorter) and prepared the way for better handling of references to groups
355        with duplicate names.
356    
357    15. A back reference to a named subpattern when there is more than one of the
358        same name now checks them in the order in which they appear in the pattern.
359        The first one that is set is used for the reference. Previously only the
360        first one was inspected. This change makes PCRE more compatible with Perl.
361    
362    16. Unicode character properties were updated from Unicode 6.3.0.
363    
364    17. The compile-time code for auto-possessification has been refactored, based
365        on a patch by Zoltan Herczeg. It now happens after instead of during
366        compilation. The code is cleaner, and more cases are handled. The option
367        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
368        options in pcretest are provided to set it. It can also be set by
369        (*NO_AUTO_POSSESS) at the start of a pattern.
370    
371    18. The character VT has been added to the default ("C" locale) set of
372        characters that match \s and are generally treated as white space,
373        following this same change in Perl 5.18. There is now no difference between
374        "Perl space" and "POSIX space". Whether VT is treated as white space in
375        other locales depends on the locale.
376    
377    19. The code for checking named groups as conditions, either for being set or
378        for being recursed, has been refactored (this is related to 14 and 15
379        above). Processing unduplicated named groups should now be as fast at
380        numerical groups, and processing duplicated groups should be faster than
381        before.
382    
383    20. Two patches to the CMake build system, by Alexander Barkov:
384    
385          (1) Replace the "source" command by "." in CMakeLists.txt because
386              "source" is a bash-ism.
387    
388          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
389              without these the CMake build does not work on Solaris.
390    
391    21. Perl has changed its handling of \8 and \9. If there is no previously
392        encountered capturing group of those numbers, they are treated as the
393        literal characters 8 and 9 instead of a binary zero followed by the
394        literals. PCRE now does the same.
395    
396    22. Following Perl, added \o{} to specify codepoints in octal, making it
397        possible to specify values greater than 0777 and also making them
398        unambiguous.
399    
400    23. Perl now gives an error for missing closing braces after \x{... instead of
401        treating the string as literal. PCRE now does the same.
402    
403    24. RunTest used to grumble if an inappropriate test was selected explicitly,
404        but just skip it when running all tests. This make it awkward to run ranges
405        of tests when one of them was inappropriate. Now it just skips any
406        inappropriate tests, as it always did when running all tests.
407    
408    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
409        character types such as \d or \w, too many callouts were inserted, and the
410        data that they returned was rubbish.
411    
412    26. In UCP mode, \s was not matching two of the characters that Perl matches,
413        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
414        were matched by \h. The code has now been refactored so that the lists of
415        the horizontal and vertical whitespace characters used for \h and \v (which
416        are defined only in one place) are now also used for \s.
417    
418    27. Add JIT support for the 64 bit TileGX architecture.
419        Patch by Jiong Wang (Tilera Corporation).
420    
421    28. Possessive quantifiers for classes (both explicit and automatically
422        generated) now use special opcodes instead of wrapping in ONCE brackets.
423    
424    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
425        (because it's meaningless), this was not happening when PCRE_CASELESS was
426        set. Not wrong, but inefficient.
427    
428    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
429        properties for \w, \d, etc) is present in a test regex. Otherwise if the
430        test contains no characters greater than 255, Perl doesn't realise it
431        should be using Unicode semantics.
432    
433    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
434        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
435        does in Unicode mode.
436    
437    32. Added the "forbid" facility to pcretest so that putting tests into the
438        wrong test files can sometimes be quickly detected.
439    
440    33. There is now a limit (default 250) on the depth of nesting of parentheses.
441        This limit is imposed to control the amount of system stack used at compile
442        time. It can be changed at build time by --with-parens-nest-limit=xxx or
443        the equivalent in CMake.
444    
445    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
446        errors. Perl warns for these when in warning mode, but PCRE has no facility
447        for giving warnings.
448    
449    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
450        However, this was not working for (?!) because it is optimized to (*FAIL),
451        for which PCRE does not allow quantifiers. The optimization is now disabled
452        when a quantifier follows (?!). I can't see any use for this, but it makes
453        things uniform.
454    
455    36. Perl no longer allows group names to start with digits, so I have made this
456        change also in PCRE. It simplifies the code a bit.
457    
458    37. In extended mode, Perl ignores spaces before a + that indicates a
459        possessive quantifier. PCRE allowed a space before the quantifier, but not
460        before the possessive +. It now does.
461    
462    38. The use of \K (reset reported match start) within a repeated possessive
463        group such as (a\Kb)*+ was not working.
464    
465    40. Document that the same character tables must be used at compile time and
466        run time, and that the facility to pass tables to pcre_exec() and
467        pcre_dfa_exec() is for use only with saved/restored patterns.
468    
469    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
470        features for Builds with MSVC:
471    
472        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
473           builds). The .rc files can be used to set FileDescription and many other
474           attributes.
475    
476        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
477           This allows higher-level build scripts which want .pdb files to avoid
478           hard-coding the exact files needed."
479    
480    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
481        mean "start of word" and "end of word", respectively, as a transition aid.
482    
483    43. A minimizing repeat of a class containing codepoints greater than 255 in
484        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
485        compiled to use the heap for recursion.
486    
487    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
488        is configured.
489    
490    
491    Version 8.33 28-May-2013
492    ------------------------
493    
494    1.  Added 'U' to some constants that are compared to unsigned integers, to
495        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
496        variables that are added to signed variables, to ensure the result is
497        signed and can be negated.
498    
499    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
500        CMake config files.
501    
502    3.  Revise the creation of config.h.generic so that all boolean macros are
503        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
504        overriding via -D on the command line possible.
505    
506    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
507        to unsigned int is reported to make a quite noticeable speed difference in
508        a specific Windows environment. Testing on Linux did also appear to show
509        some benefit (and it is clearly not harmful). Also fixed the definition of
510        Xop which should be unsigned.
511    
512    5.  Related to (4), changing the definition of the intermediate variable cc
513        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
514        improvements.
515    
516    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
517        unnecessary spaces.
518    
519    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
520        and later.
521    
522    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
523    
524    9.  Optimizing fast_forward_start_bits in JIT.
525    
526    10. Adding support for callouts in JIT, and fixing some issues revealed
527        during this work. Namely:
528    
529        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
530    
531        (b) Minimum length was not checked before the matching is started.
532    
533    11. The value of capture_last that is passed to callouts was incorrect in some
534        cases when there was a capture on one path that was subsequently abandoned
535        after a backtrack. Also, the capture_last value is now reset after a
536        recursion, since all captures are also reset in this case.
537    
538    12. The interpreter no longer returns the "too many substrings" error in the
539        case when an overflowing capture is in a branch that is subsequently
540        abandoned after a backtrack.
541    
542    13. In the pathological case when an offset vector of size 2 is used, pcretest
543        now prints out the matched string after a yield of 0 or 1.
544    
545    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
546        Only supported by the JIT compiler at the moment.
547    
548    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
549    
550    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
551        offset of the starting point of the matching process, provided the offsets
552        vector is large enough.
553    
554    17. The \A escape now records a lookbehind value of 1, though its execution
555        does not actually inspect the previous character. This is to ensure that,
556        in partial multi-segment matching, at least one character from the old
557        segment is retained when a new segment is processed. Otherwise, if there
558        are no lookbehinds in the pattern, \A might match incorrectly at the start
559        of a new segment.
560    
561    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
562    
563    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
564        gives some modest performance improvement in 8-bit mode.
565    
566    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
567        be expressed in certain programming languages using Universal Character
568        Names.
569    
570    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
571        which points out that "non characters" are not "characters that may not
572        appear in Unicode strings" but rather "characters that are reserved for
573        internal use and have only local meaning".
574    
575    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
576        there was a conditional group that depended on an assertion, if the
577        assertion was false, the callout that immediately followed the alternation
578        in the condition was skipped when pcre_exec() was used for matching.
579    
580    23. Allow an explicit callout to be inserted before an assertion that is the
581        condition for a conditional group, for compatibility with automatic
582        callouts, which always insert a callout at this point.
583    
584    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
585        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
586    
587    25. (*PRUNE) is now supported by the JIT compiler.
588    
589    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
590    
591    27. Fix the case where there are two or more SKIPs with arguments that may be
592        ignored.
593    
594    28. (*SKIP) is now supported by the JIT compiler.
595    
596    29. (*THEN) is now supported by the JIT compiler.
597    
598    30. Update RunTest with additional test selector options.
599    
600    31. The way PCRE handles backtracking verbs has been changed in two ways.
601    
602        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
603        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
604        backtracking. In some cases this makes it more Perl-compatible, but Perl's
605        rather obscure rules do not always do the same thing.
606    
607        (2) Previously, backtracking verbs were confined within assertions. This is
608        no longer the case for positive assertions, except for (*ACCEPT). Again,
609        this sometimes improves Perl compatibility, and sometimes does not.
610    
611    32. A number of tests that were in test 2 because Perl did things differently
612        have been moved to test 1, because either Perl or PCRE has changed, and
613        these tests are now compatible.
614    
615    32. Backtracking control verbs are now handled in the same way in JIT and
616        interpreter.
617    
618    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
619        contained a forward subroutine reference caused a compile error.
620    
621    34. Auto-detect and optimize limited repetitions in JIT.
622    
623    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
624        blocking (*UTF) etc.
625    
626    36. In the interpreter, maximizing pattern repetitions for characters and
627        character types now use tail recursion, which reduces stack usage.
628    
629    37. The value of the max lookbehind was not correctly preserved if a compiled
630        and saved regex was reloaded on a host of different endianness.
631    
632    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
633        of the compiled pattern block, expand the flags field from 16 to 32 bits
634        because it was almost full.
635    
636    39. Try madvise first before posix_madvise.
637    
638    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
639        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
640        original change locked out all matches of empty strings. This has been
641        changed so that one match of an empty string per line is recognized.
642        Subsequent searches on the same line (for colouring or for --only-matching,
643        for example) do not recognize empty strings.
644    
645    41. Applied a user patch to fix a number of spelling mistakes in comments.
646    
647    42. Data lines longer than 65536 caused pcretest to crash.
648    
649    43. Clarified the data type for length and startoffset arguments for pcre_exec
650        and pcre_dfa_exec in the function-specific man pages, where they were
651        explicitly stated to be in bytes, never having been updated. I also added
652        some clarification to the pcreapi man page.
653    
654    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
655        a segmentation fault.
656    
657    
658    Version 8.32 30-November-2012
659    -----------------------------
660    
661    1.  Improved JIT compiler optimizations for first character search and single
662        character iterators.
663    
664    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
665        Patch by Daniel Richard G.
666    
667    3.  Single character iterator optimizations in the JIT compiler.
668    
669    4.  Improved JIT compiler optimizations for character ranges.
670    
671    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
672        Reported by Giuseppe D'Angelo.
673    
674    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
675        of a line, was being set incorrectly in cases where .* appeared inside
676        atomic brackets at the start of a pattern, or where there was a subsequent
677        *PRUNE or *SKIP.
678    
679    7.  Improved instruction cache flush for POWER/PowerPC.
680        Patch by Daniel Richard G.
681    
682    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
683        grep:
684    
685        (a) There is now no limit to the number of patterns to be matched.
686    
687        (b) An error is given if a pattern is too long.
688    
689        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
690            are now supported.
691    
692        (d) --exclude-from and --include-from (multiple use) have been added.
693    
694        (e) Exclusions and inclusions now apply to all files and directories, not
695            just to those obtained from scanning a directory recursively.
696    
697        (f) Multiple uses of -f and --file-list are now supported.
698    
699        (g) In a Windows environment, the default for -d has been changed from
700            "read" (the GNU grep default) to "skip", because otherwise the presence
701            of a directory in the file list provokes an error.
702    
703        (h) The documentation has been revised and clarified in places.
704    
705    9.  Improve the matching speed of capturing brackets.
706    
707    10. Changed the meaning of \X so that it now matches a Unicode extended
708        grapheme cluster.
709    
710    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
711        out POSIX threads when JIT support is configured.
712    
713    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
714    
715    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
716        structure could go wrong in environments where size_t is not the same size
717        as int.
718    
719    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
720        set.
721    
722    15. The EBCDIC support had decayed; later updates to the code had included
723        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
724        general tidy up of EBCDIC-related issues, and the documentation was also
725        not quite right. There is now a test that can be run on ASCII systems to
726        check some of the EBCDIC-related things (but is it not a full test).
727    
728    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
729        in a small tidy to the code.
730    
731    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
732    
733    18. If the --only-matching (-o) option in pcregrep is specified multiple
734        times, each one causes appropriate output. For example, -o1 -o2 outputs the
735        substrings matched by the 1st and 2nd capturing parentheses. A separating
736        string can be specified by --om-separator (default empty).
737    
738    19. Improving the first n character searches.
739    
740    20. Turn case lists for horizontal and vertical white space into macros so that
741        they are defined only once.
742    
743    21. This set of changes together give more compatible Unicode case-folding
744        behaviour for characters that have more than one other case when UCP
745        support is available.
746    
747        (a) The Unicode property table now has offsets into a new table of sets of
748            three or more characters that are case-equivalent. The MultiStage2.py
749            script that generates these tables (the pcre_ucd.c file) now scans
750            CaseFolding.txt instead of UnicodeData.txt for character case
751            information.
752    
753        (b) The code for adding characters or ranges of characters to a character
754            class has been abstracted into a generalized function that also handles
755            case-independence. In UTF-mode with UCP support, this uses the new data
756            to handle characters with more than one other case.
757    
758        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
759            whose other case is greater than 256 are now correctly matched
760            caselessly. Previously, the high codepoint matched the low one, but not
761            vice versa.
762    
763        (d) The processing of \h, \H, \v, and \ in character classes now makes use
764            of the new class addition function, using character lists defined as
765            macros alongside the case definitions of 20 above.
766    
767        (e) Caseless back references now work with characters that have more than
768            one other case.
769    
770        (f) General caseless matching of characters with more than one other case
771            is supported.
772    
773    22. Unicode character properties were updated from Unicode 6.2.0
774    
775    23. Improved CMake support under Windows. Patch by Daniel Richard G.
776    
777    24. Add support for 32-bit character strings, and UTF-32
778    
779    25. Major JIT compiler update (code refactoring and bugfixing).
780        Experimental Sparc 32 support is added.
781    
782    26. Applied a modified version of Daniel Richard G's patch to create
783        pcre.h.generic and config.h.generic by "make" instead of in the
784        PrepareRelease script.
785    
786    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
787        pcre_compile.c when checking for a zero character.
788    
789    28. Introducing a native interface for JIT. Through this interface, the compiled
790        machine code can be directly executed. The purpose of this interface is to
791        provide fast pattern matching, so several sanity checks are not performed.
792        However, feature tests are still performed. The new interface provides
793        1.4x speedup compared to the old one.
794    
795    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
796        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
797        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
798    
799    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
800        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
801        the "old" RFC 2279). Instead, it ended up passing a negative length to
802        pcre_exec().
803    
804    31. Add support for GCC's visibility feature to hide internal functions.
805    
806    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
807        "unknown -C option" after outputting 0 or 1.
808    
809    33. There is now support for generating a code coverage report for the test
810        suite in environments where gcc is the compiler and lcov is installed. This
811        is mainly for the benefit of the developers.
812    
813    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
814        unaddressable using valgrind annotations, allowing valgrind to detect
815        invalid memory accesses. This is mainly for the benefit of the developers.
816    
817    25. (*UTF) can now be used to start a pattern in any of the three libraries.
818    
819    26. Give configure error if --enable-cpp but no C++ compiler found.
820    
821    
822    Version 8.31 06-July-2012
823    -------------------------
824    
825    1.  Fixing a wrong JIT test case and some compiler warnings.
826    
827    2.  Removed a bashism from the RunTest script.
828    
829    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
830        to unsigned type, result still unsigned" that was given by an MS compiler
831        on encountering the code "-sizeof(xxx)".
832    
833    4.  Partial matching support is added to the JIT compiler.
834    
835    5.  Fixed several bugs concerned with partial matching of items that consist
836        of more than one character:
837    
838        (a) /^(..)\1/ did not partially match "aba" because checking references was
839            done on an "all or nothing" basis. This also applied to repeated
840            references.
841    
842        (b) \R did not give a hard partial match if \r was found at the end of the
843            subject.
844    
845        (c) \X did not give a hard partial match after matching one or more
846            characters at the end of the subject.
847    
848        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
849            a partial match for the string "\r".
850    
851        (e) When newline was set to CRLF, the metacharacter "." did not recognize
852            a partial match for a CR character at the end of the subject string.
853    
854    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
855        running pcretest, the text "(JIT)" added to the output whenever JIT is
856        actually used to run the match.
857    
858    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
859        or /S+[+] with a digit between 1 and 7.
860    
861    8.  OP_NOT now supports any UTF character not just single-byte ones.
862    
863    9.  (*MARK) control verb is now supported by the JIT compiler.
864    
865    10. The command "./RunTest list" lists the available tests without actually
866        running any of them. (Because I keep forgetting what they all are.)
867    
868    11. Add PCRE_INFO_MAXLOOKBEHIND.
869    
870    12. Applied a (slightly modified) user-supplied patch that improves performance
871        when the heap is used for recursion (compiled with --disable-stack-for-
872        recursion). Instead of malloc and free for each heap frame each time a
873        logical recursion happens, frames are retained on a chain and re-used where
874        possible. This sometimes gives as much as 30% improvement.
875    
876    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
877        call.
878    
879    14. As documented, (*COMMIT) is now confined to within a positive assertion.
880    
881    15. It is now possible to link pcretest with libedit as an alternative to
882        libreadline.
883    
884    16. (*COMMIT) control verb is now supported by the JIT compiler.
885    
886    17. The Unicode data tables have been updated to Unicode 6.1.0.
887    
888    18. Added --file-list option to pcregrep.
889    
890    19. Added binary file support to pcregrep, including the -a, --binary-files,
891        -I, and --text options.
892    
893    20. The madvise function is renamed for posix_madvise for QNX compatibility
894        reasons. Fixed by Giuseppe D'Angelo.
895    
896    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
897    
898    22. Changed the option for creating symbolic links for 16-bit man pages from
899        -s to -sf so that re-installing does not cause issues.
900    
901    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
902    
903    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
904        matches in certain environments (the workspace was not being correctly
905        retained). Also added to pcre_dfa_exec() a simple plausibility check on
906        some of the workspace data at the beginning of a restart.
907    
908    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
909        was not doing so when it should - probably a typo introduced by SVN 528
910        (change 8.10/14).
911    
912    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
913        \w+ when the character tables indicated that \x{c4} was a word character.
914        There were several related cases, all because the tests for doing a table
915        lookup were testing for characters less than 127 instead of 255.
916    
917    27. If a pattern contains capturing parentheses that are not used in a match,
918        their slots in the ovector are set to -1. For those that are higher than
919        any matched groups, this happens at the end of processing. In the case when
920        there were back references that the ovector was too small to contain
921        (causing temporary malloc'd memory to be used during matching), and the
922        highest capturing number was not used, memory off the end of the ovector
923        was incorrectly being set to -1. (It was using the size of the temporary
924        memory instead of the true size.)
925    
926    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
927        ovector size, it uses memory at the end of the block that it has got.
928    
929    29. Check for an overlong MARK name and give an error at compile time. The
930        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
931    
932    30. JIT compiler update.
933    
934    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
935        Rill for the patch.
936    
937    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
938    
939    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
940    
941    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
942        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
943        was enabled.
944    
945    35. Improve JIT code generation for greedy plus quantifier.
946    
947    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
948        1 to "aa" instead of to an empty string. The bug affected repeated groups
949        that could potentially match an empty string.
950    
951    37. Optimizing single character iterators in JIT.
952    
953    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
954        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
955        codepoints that are too big for the mode are faulted, and in a UTF mode,
956        disallowed codepoints are also faulted.
957    
958    39. If PCRE was compiled with UTF support, in three places in the DFA
959        matcher there was code that should only have been obeyed in UTF mode, but
960        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
961        processing when bytes with values greater than 127 were present. In 16-bit
962        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
963        both cases the values are those that cannot be the first data item in a UTF
964        character. The three items that might have provoked this were recursions,
965        possessively repeated groups, and atomic groups.
966    
967    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
968        and pcregrep, because some OS require shared objects to be explicitly
969        passed to ld, causing the link step to fail if they are not.
970    
971    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
972        mode, patterns that started with \h* or \R* might be incorrectly matched.
973    
974    
975    Version 8.30 04-February-2012
976    -----------------------------
977    
978    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
979        name is defined in ctype.h.
980    
981    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
982        only in quite long subpatterns.
983    
984    3.  Removed the function pcre_info(), which has been obsolete and deprecated
985        since it was replaced by pcre_fullinfo() in February 2000.
986    
987    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
988        match a (*MARK), and the match failed at the start of the subject, a
989        reference to memory before the start of the subject could occur. This bug
990        was introduced by fix 17 of release 8.21.
991    
992    5.  A reference to an unset group with zero minimum repetition was giving
993        totally wrong answers (in non-JavaScript-compatibility mode). For example,
994        /(another)?(\1?)test/ matched against "hello world test". This bug was
995        introduced in release 8.13.
996    
997    6.  Add support for 16-bit character strings (a large amount of work involving
998        many changes and refactorings).
999    
1000    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1001        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1002        from a file.
1003    
1004    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1005        rounding is not applied in this particular case).
1006    
1007    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1008        if they appear, or are escaped, in patterns.
1009    
1010    10. Get rid of a number of -Wunused-but-set-variable warnings.
1011    
1012    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1013        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1014        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1015        also returns the mark "x". This bug applied to capturing parentheses,
1016        non-capturing parentheses, and atomic parentheses. It also applied to some
1017        assertions.
1018    
1019    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1020        information out of configure.ac instead of relying on pcre.h.generic, which
1021        is not stored in the repository.
1022    
1023    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1024        -lreadline.
1025    
1026    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1027    
1028    15. Applied Graycode's patch to put the top-level frame on the stack rather
1029        than the heap when not using the stack for recursion. This gives a
1030        performance improvement in many cases when recursion is not deep.
1031    
1032    16. Experimental code added to "pcretest -C" to output the stack frame size.
1033    
1034    
1035    Version 8.21 12-Dec-2011
1036    ------------------------
1037    
1038    1.  Updating the JIT compiler.
1039    
1040    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1041        are added as well.
1042    
1043    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1044        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1045        calling _pcre_jit_exec. Some extra comments are added.
1046    
1047    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1048        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1049        was introduced by change 18 for 8.20.
1050    
1051    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1052        ECMA-262 standard.
1053    
1054    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1055        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1056        This bug was probably introduced by change 9 of 8.13.
1057    
1058    7.  While fixing 6 above, I noticed that a number of other items were being
1059        incorrectly rejected as "not fixed length". This arose partly because newer
1060        opcodes had not been added to the fixed-length checking code. I have (a)
1061        corrected the bug and added tests for these items, and (b) arranged for an
1062        error to occur if an unknown opcode is encountered while checking for fixed
1063        length instead of just assuming "not fixed length". The items that were
1064        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1065        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1066        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1067    
1068    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1069        being incorrectly compiled and would have given unpredicatble results.
1070    
1071    9.  A possessively repeated subpattern with minimum repeat count greater than
1072        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1073        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1074        the first (A) could occur when it should not.
1075    
1076    10. Add a cast and remove a redundant test from the code.
1077    
1078    11. JIT should use pcre_malloc/pcre_free for allocation.
1079    
1080    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1081        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1082        is anything other than /usr, -L is still shown).
1083    
1084    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1085    
1086    14. Perl does not support \N without a following name in a [] class; PCRE now
1087        also gives an error.
1088    
1089    15. If a forward reference was repeated with an upper limit of around 2000,
1090        it caused the error "internal error: overran compiling workspace". The
1091        maximum number of forward references (including repeats) was limited by the
1092        internal workspace, and dependent on the LINK_SIZE. The code has been
1093        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1094        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1095        of around 200,000 forward references. While doing this, I also speeded up
1096        the filling in of repeated forward references.
1097    
1098    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1099        incorrectly expecting the subject to contain another "a" after the start.
1100    
1101    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1102        in the match, the SKIP should be ignored. This was not happening; instead
1103        the SKIP was being treated as NOMATCH. For patterns such as
1104        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1105        tested.
1106    
1107    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1108        now much more compatible with Perl, in particular in cases where the result
1109        is a non-match for a non-anchored pattern. For example, if
1110        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1111        "m", where previously it did not return a name. A side effect of this
1112        change is that for partial matches, the last encountered mark name is
1113        returned, as for non matches. A number of tests that were previously not
1114        Perl-compatible have been moved into the Perl-compatible test files. The
1115        refactoring has had the pleasing side effect of removing one argument from
1116        the match() function, thus reducing its stack requirements.
1117    
1118    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1119        subsequent uses of /S (without +) incorrectly behaved like /S+.
1120    
1121    21. Retrieve executable code size support for the JIT compiler and fixing
1122        some warnings.
1123    
1124    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1125        not work when the shorter character appeared right at the end of the
1126        subject string.
1127    
1128    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1129        systems.
1130    
1131    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1132        output it when the /M option is used in pcretest.
1133    
1134    25. The CheckMan script was not being included in the distribution. Also, added
1135        an explicit "perl" to run Perl scripts from the PrepareRelease script
1136        because this is reportedly needed in Windows.
1137    
1138    26. If study data was being save in a file and studying had not found a set of
1139        "starts with" bytes for the pattern, the data written to the file (though
1140        never used) was taken from uninitialized memory and so caused valgrind to
1141        complain.
1142    
1143    27. Updated RunTest.bat as provided by Sheri Pierce.
1144    
1145    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1146    
1147    29. Computation of memory usage for the table of capturing group names was
1148        giving an unnecessarily large value.
1149    
1150    
1151    Version 8.20 21-Oct-2011
1152    ------------------------
1153    
1154    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1155        a POSIX class. After further experiments with Perl, which convinced me that
1156        Perl has bugs and confusions, a closing square bracket is no longer allowed
1157        in a POSIX name. This bug also affected patterns with classes that started
1158        with full stops.
1159    
1160    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1161        captured substring, but while checking the failing first alternative,
1162        substring 1 is temporarily captured. If the output vector supplied to
1163        pcre_exec() was not big enough for this capture, the yield of the function
1164        was still zero ("insufficient space for captured substrings"). This cannot
1165        be totally fixed without adding another stack variable, which seems a lot
1166        of expense for a edge case. However, I have improved the situation in cases
1167        such as /(a)(b)x|abc/ matched against "abc", where the return code
1168        indicates that fewer than the maximum number of slots in the ovector have
1169        been set.
1170    
1171    3.  Related to (2) above: when there are more back references in a pattern than
1172        slots in the output vector, pcre_exec() uses temporary memory during
1173        matching, and copies in the captures as far as possible afterwards. It was
1174        using the entire output vector, but this conflicts with the specification
1175        that only 2/3 is used for passing back captured substrings. Now it uses
1176        only the first 2/3, for compatibility. This is, of course, another edge
1177        case.
1178    
1179    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1180        main code base, and can be used by building with --enable-jit. When this is
1181        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1182        runtime --no-jit option is given.
1183    
1184    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1185        ovector, the return from the function was zero, implying that there were
1186        other matches that did not fit. The correct "exactly full" value is now
1187        returned.
1188    
1189    6.  If a subpattern that was called recursively or as a subroutine contained
1190        (*PRUNE) or any other control that caused it to give a non-standard return,
1191        invalid errors such as "Error -26 (nested recursion at the same subject
1192        position)" or even infinite loops could occur.
1193    
1194    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1195        computing the minimum length on reaching *ACCEPT, and so ended up with the
1196        wrong value of 1 rather than 0. Further investigation indicates that
1197        computing a minimum subject length in the presence of *ACCEPT is difficult
1198        (think back references, subroutine calls), and so I have changed the code
1199        so that no minimum is registered for a pattern that contains *ACCEPT.
1200    
1201    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1202        it was not handled as intended. [But see 16 below.]
1203    
1204    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1205        Sheri Pierce.
1206    
1207    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1208        the first byte in a match must be "a".
1209    
1210    11. Change 17 for 8.13 increased the recursion depth for patterns like
1211        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1212        pattern contains any instances of (*THEN). If it does not, the old
1213        optimizations are restored. It would be nice to do this on a per-group
1214        basis, but at the moment that is not feasible.
1215    
1216    12. In some environments, the output of pcretest -C is CRLF terminated. This
1217        broke RunTest's code that checks for the link size. A single white space
1218        character after the value is now allowed for.
1219    
1220    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1221        For "fr", it uses the Windows-specific input and output files.
1222    
1223    14. If (*THEN) appeared in a group that was called recursively or as a
1224        subroutine, it did not work as intended. [But see next item.]
1225    
1226    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1227        pattern fragments (but not containing any | characters). If A and B are
1228        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1229        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1230        D. In other words, Perl considers parentheses that do not contain any |
1231        characters to be part of a surrounding alternative, whereas PCRE was
1232        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1233        differently. PCRE now behaves in the same way as Perl, except in the case
1234        of subroutine/recursion calls such as (?1) which have in any case always
1235        been different (but PCRE had them first :-).
1236    
1237    16. Related to 15 above: Perl does not treat the | in a conditional group as
1238        creating alternatives. Such a group is treated in the same way as an
1239        ordinary group without any | characters when processing (*THEN). PCRE has
1240        been changed to match Perl's behaviour.
1241    
1242    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1243        RunGrepTest script failed.
1244    
1245    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1246        inevitable for groups that contain captures, but it can lead to a lot of
1247        stack use in large patterns. The old behaviour has been restored for atomic
1248        groups that do not contain any capturing parentheses.
1249    
1250    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1251        suppress the check for a minimum subject length at run time. (If it was
1252        given to pcre_exec() or pcre_dfa_exec() it did work.)
1253    
1254    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1255        fail to work when decoding hex characters in data strings in EBCDIC
1256        environments.
1257    
1258    21. It appears that in at least one Mac OS environment, the isxdigit() function
1259        is implemented as a macro that evaluates to its argument more than once,
1260        contravening the C 90 Standard (I haven't checked a later standard). There
1261        was an instance in pcretest which caused it to go wrong when processing
1262        \x{...} escapes in subject strings. The has been rewritten to avoid using
1263        things like p++ in the argument of isxdigit().
1264    
1265    
1266    Version 8.13 16-Aug-2011
1267    ------------------------
1268    
1269    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1270    
1271    2.  Two minor typos in pcre_internal.h have been fixed.
1272    
1273    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1274        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1275        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1276    
1277    4.  There were a number of related bugs in the code for matching backrefences
1278        caselessly in UTF-8 mode when codes for the characters concerned were
1279        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1280        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1281        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1282        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1283        2-byte code at the end of the subject (it thought there wasn't enough data
1284        left).
1285    
1286    5.  Comprehensive information about what went wrong is now returned by
1287        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1288        as the output vector has at least 2 elements. The offset of the start of
1289        the failing character and a reason code are placed in the vector.
1290    
1291    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1292        now returned is for the first byte of the failing character, instead of the
1293        last byte inspected. This is an incompatible change, but I hope it is small
1294        enough not to be a problem. It makes the returned offset consistent with
1295        pcre_exec() and pcre_dfa_exec().
1296    
1297    7.  pcretest now gives a text phrase as well as the error number when
1298        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1299        failure, the offset and reason code are output.
1300    
1301    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1302        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1303        back over a single character (\n). This seems wrong (because it treated the
1304        two characters as a single entity when going forwards), conflicts with the
1305        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1306        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1307        has been changed.
1308    
1309    9.  Some internal refactoring has changed the processing so that the handling
1310        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1311        time (the PCRE_DOTALL option was changed this way some time ago: version
1312        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1313        which was always a bit of a fudge. It also means that there is one less
1314        argument for the match() function, which reduces its stack requirements
1315        slightly. This change also fixes an incompatibility with Perl: the pattern
1316        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1317    
1318    10. More internal refactoring has drastically reduced the number of recursive
1319        calls to match() for possessively repeated groups such as (abc)++ when
1320        using pcre_exec().
1321    
1322    11. While implementing 10, a number of bugs in the handling of groups were
1323        discovered and fixed:
1324    
1325        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1326        (a|)*(?1) gave a compile-time internal error.
1327        ((a|)+)+  did not notice that the outer group could match an empty string.
1328        (^a|^)+   was not marked as anchored.
1329        (.*a|.*)+ was not marked as matching at start or after a newline.
1330    
1331    12. Yet more internal refactoring has removed another argument from the match()
1332        function. Special calls to this function are now indicated by setting a
1333        value in a variable in the "match data" data block.
1334    
1335    13. Be more explicit in pcre_study() instead of relying on "default" for
1336        opcodes that mean there is no starting character; this means that when new
1337        ones are added and accidentally left out of pcre_study(), testing should
1338        pick them up.
1339    
1340    14. The -s option of pcretest has been documented for ages as being an old
1341        synonym of -m (show memory usage). I have changed it to mean "force study
1342        for every regex", that is, assume /S for every regex. This is similar to -i
1343        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1344        using it. It makes it easier to run collections of tests with and without
1345        study enabled, and thereby test pcre_study() more easily. All the standard
1346        tests are now run with and without -s (but some patterns can be marked as
1347        "never study" - see 20 below).
1348    
1349    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1350        restoration of the capturing data to the outer values was not happening
1351        correctly.
1352    
1353    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1354        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1355        pattern had matched an empty string, and so incorrectly returned a no
1356        match.
1357    
1358    17. There was optimizing code for the last branch of non-capturing parentheses,
1359        and also for the obeyed branch of a conditional subexpression, which used
1360        tail recursion to cut down on stack usage. Unfortunately, now that there is
1361        the possibility of (*THEN) occurring in these branches, tail recursion is
1362        no longer possible because the return has to be checked for (*THEN). These
1363        two optimizations have therefore been removed. [But see 8.20/11 above.]
1364    
1365    18. If a pattern containing \R was studied, it was assumed that \R always
1366        matched two bytes, thus causing the minimum subject length to be
1367        incorrectly computed because \R can also match just one byte.
1368    
1369    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1370        was incorrectly computed.
1371    
1372    20. If /S is present twice on a test pattern in pcretest input, it now
1373        *disables* studying, thereby overriding the use of -s on the command line
1374        (see 14 above). This is necessary for one or two tests to keep the output
1375        identical in both cases.
1376    
1377    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1378        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1379    
1380    22. When an atomic group that contained a capturing parenthesis was
1381        successfully matched, but the branch in which it appeared failed, the
1382        capturing was not being forgotten if a higher numbered group was later
1383        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1384        group 1 to "a", when in fact it should be unset. This applied to multi-
1385        branched capturing and non-capturing groups, repeated or not, and also to
1386        positive assertions (capturing in negative assertions does not happen
1387        in PCRE) and also to nested atomic groups.
1388    
1389    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1390        subject after a captured substring, to make it easier to tell which of a
1391        number of identical substrings has been captured.
1392    
1393    24. The way atomic groups are processed by pcre_exec() has been changed so that
1394        if they are repeated, backtracking one repetition now resets captured
1395        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1396        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1397        "aaa". Previously, it would have been "a". As part of this code
1398        refactoring, the way recursive calls are handled has also been changed.
1399    
1400    25. If an assertion condition captured any substrings, they were not passed
1401        back unless some other capturing happened later. For example, if
1402        (?(?=(a))a) was matched against "a", no capturing was returned.
1403    
1404    26. When studying a pattern that contained subroutine calls or assertions,
1405        the code for finding the minimum length of a possible match was handling
1406        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1407        group 1 called group 2 while simultaneously a separate group 2 called group
1408        1). A stack overflow occurred in this case. I have fixed this by limiting
1409        the recursion depth to 10.
1410    
1411    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1412        Fortmann. This supports explicit test numbers on the command line, and has
1413        argument validation and error reporting.
1414    
1415    28. An instance of \X with an unlimited repeat could fail if at any point the
1416        first character it looked at was a mark character.
1417    
1418    29. Some minor code refactoring concerning Unicode properties and scripts
1419        should reduce the stack requirement of match() slightly.
1420    
1421    30. Added the '=' option to pcretest to check the setting of unused capturing
1422        slots at the end of the pattern, which are documented as being -1, but are
1423        not included in the return count.
1424    
1425    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1426        compiled something random. Now it gives a compile-time error (as does
1427        Perl).
1428    
1429    32. A *MARK encountered during the processing of a positive assertion is now
1430        recorded and passed back (compatible with Perl).
1431    
1432    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1433        had alternative anchored branches, the search for a second match in a line
1434        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1435        matched the line "0102" twice. The same bug affected patterns that started
1436        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1437        twice.
1438    
1439    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1440        does, and because of capturing effects, quantifying parenthesized
1441        assertions may at times be useful. Quantifiers are now allowed for
1442        parenthesized assertions.
1443    
1444    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1445    
1446    36. \g was being checked for fancy things in a character class, when it should
1447        just be a literal "g".
1448    
1449    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1450        appearance of a nested POSIX class supersedes an apparent external class.
1451        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1452        unescaped square brackets may also appear as part of class names. For
1453        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1454        more like Perl. (But see 8.20/1 above.)
1455    
1456    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1457        was because it thought it was \N{name}, which is not supported).
1458    
1459    39. Add minix to OS list not supporting the -S option in pcretest.
1460    
1461    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1462        cannot analyze patterns in sufficient detail to catch mutual recursions
1463        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1464        subgroup is called recursively as a subpattern for a second time at the
1465        same position in the subject string. In previous releases this might have
1466        been caught by the recursion limit, or it might have run out of stack.
1467    
1468    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1469        happen only once. PCRE was, however incorrectly giving a compile time error
1470        "recursive call could loop indefinitely" because it cannot analyze the
1471        pattern in sufficient detail. The compile time test no longer happens when
1472        PCRE is compiling a conditional subpattern, but actual runaway loops are
1473        now caught at runtime (see 40 above).
1474    
1475    42. It seems that Perl allows any characters other than a closing parenthesis
1476        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1477        has been changed to be the same.
1478    
1479    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1480        as not to get warnings when autogen.sh is called. Also changed
1481        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1482    
1483    44. To help people who use pcregrep to scan files containing exceedingly long
1484        lines, the following changes have been made:
1485    
1486        (a) The default value of the buffer size parameter has been increased from
1487            8K to 20K. (The actual buffer used is three times this size.)
1488    
1489        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1490            PCRE is built.
1491    
1492        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1493            to be set at run time.
1494    
1495        (d) Numerical values in pcregrep options can be followed by K or M, for
1496            example --buffer-size=50K.
1497    
1498        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1499            given and the return code is set to 2.
1500    
1501    45. Add a pointer to the latest mark to the callout data block.
1502    
1503    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1504        partial match of an empty string instead of no match. This was specific to
1505        the use of ".".
1506    
1507    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1508        complete match instead of a partial match. This bug was dependent on both
1509        the PCRE_UTF8 and PCRE_DOTALL options being set.
1510    
1511    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1512        starting byte set, because \b was not being ignored.
1513    
1514    
1515    Version 8.12 15-Jan-2011
1516    ------------------------
1517    
1518    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1519        checks for such things as part of the documentation building process.
1520    
1521    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1522        --match-limit and --recursion-limit options (added for 8.11). In
1523        particular, this made one of the standard tests fail. (The integer value
1524        went into the wrong half of a long int.)
1525    
1526    3.  If the --colour option was given to pcregrep with -v (invert match), it
1527        did strange things, either producing crazy output, or crashing. It should,
1528        of course, ignore a request for colour when reporting lines that do not
1529        match.
1530    
1531    4.  Another pcregrep bug caused similar problems if --colour was specified with
1532        -M (multiline) and the pattern match finished with a line ending.
1533    
1534    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1535        matched in multiline mode, the following line was shown as part of the
1536        match. This seems wrong, so I have changed it.
1537    
1538    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1539        the check for further matches in the same line (so they could be coloured)
1540        to overrun the end of the current line. If another match was found, it was
1541        incorrectly shown (and then shown again when found in the next line).
1542    
1543    7.  If pcregrep was compiled under Windows, there was a reference to the
1544        function pcregrep_exit() before it was defined. I am assuming this was
1545        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1546        reported by a user. I've moved the definition above the reference.
1547    
1548    
1549    Version 8.11 10-Dec-2010
1550    ------------------------
1551    
1552    1.  (*THEN) was not working properly if there were untried alternatives prior
1553        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1554        backtracked to try for "b" instead of moving to the next alternative branch
1555        at the same level (in this case, to look for "c"). The Perl documentation
1556        is clear that when (*THEN) is backtracked onto, it goes to the "next
1557        alternative in the innermost enclosing group".
1558    
1559    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1560        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1561        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1562        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1563        (*THEN).
1564    
1565    3.  If \s appeared in a character class, it removed the VT character from
1566        the class, even if it had been included by some previous item, for example
1567        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1568        of \s, but is part of the POSIX "space" class.)
1569    
1570    4.  A partial match never returns an empty string (because you can always
1571        match an empty string at the end of the subject); however the checking for
1572        an empty string was starting at the "start of match" point. This has been
1573        changed to the "earliest inspected character" point, because the returned
1574        data for a partial match starts at this character. This means that, for
1575        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1576        (previously it gave "no match").
1577    
1578    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1579        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1580        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1581        has an implication that the given string is incomplete (because a partial
1582        match is preferred over a full match). For this reason, these items now
1583        give a partial match in this situation. [Aside: previously, the one case
1584        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1585        match rather than a full match, which was wrong by the old rules, but is
1586        now correct.]
1587    
1588    6.  There was a bug in the handling of #-introduced comments, recognized when
1589        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1590        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1591        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1592        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1593        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1594        places in pcre_compile().
1595    
1596    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1597        comments when looking ahead for named forward references to subpatterns,
1598        the only newline sequence it recognized was NL. It now handles newlines
1599        according to the set newline convention.
1600    
1601    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1602        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1603        cater for a lack of strerror(). These oversights have been fixed.
1604    
1605    9.  Added --match-limit and --recursion-limit to pcregrep.
1606    
1607    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1608    
1609    11. When the -o option was used, pcregrep was setting a return code of 1, even
1610        when matches were found, and --line-buffered was not being honoured.
1611    
1612    12. Added an optional parentheses number to the -o and --only-matching options
1613        of pcregrep.
1614    
1615    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1616        can match an empty string. The code to do it in pcretest and pcredemo
1617        needed fixing:
1618    
1619        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1620            only one byte after an empty string match just before CRLF (this case
1621            just got forgotten; "any" and "anycrlf" were OK).
1622    
1623        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1624            mode when an empty string match preceded an ASCII character followed by
1625            a non-ASCII character. (The code for advancing by one character rather
1626            than one byte was nonsense.)
1627    
1628        (c) The pcredemo.c sample program did not have any code at all to handle
1629            the cases when CRLF is a valid newline sequence.
1630    
1631    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1632        as a starting offset was within the subject string. There is now a new
1633        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1634        negative or greater than the length of the string. In order to test this,
1635        pcretest is extended to allow the setting of negative starting offsets.
1636    
1637    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1638        starting offset points to the beginning of a UTF-8 character was
1639        unnecessarily clumsy. I tidied it up.
1640    
1641    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1642        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1643    
1644    17. Nobody had reported that the --include_dir option, which was added in
1645        release 7.7 should have been called --include-dir (hyphen, not underscore)
1646        for compatibility with GNU grep. I have changed it to --include-dir, but
1647        left --include_dir as an undocumented synonym, and the same for
1648        --exclude-dir, though that is not available in GNU grep, at least as of
1649        release 2.5.4.
1650    
1651    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1652        characters from a string of bytes) have been redefined so as not to use
1653        loops, in order to improve performance in some environments. At the same
1654        time, I abstracted some of the common code into auxiliary macros to save
1655        repetition (this should not affect the compiled code).
1656    
1657    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1658        compile-time error is now given if \c is not followed by an ASCII
1659        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1660        different, and any byte value is allowed.)
1661    
1662    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1663        START_OPTIMIZE option, which is now allowed at compile time - but just
1664        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1665        to pcregrep and other applications that have no direct access to PCRE
1666        options. The new /Y option in pcretest sets this option when calling
1667        pcre_compile().
1668    
1669    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1670        back references. Groups containing recursive back references were forced to
1671        be atomic by that change, but in the case of named groups, the amount of
1672        memory required was incorrectly computed, leading to "Failed: internal
1673        error: code overflow". This has been fixed.
1674    
1675    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1676        pcretest.c, to avoid build problems in some Borland environments.
1677    
1678    
1679    Version 8.10 25-Jun-2010
1680    ------------------------
1681    
1682    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1683        THEN.
1684    
1685    2.  (*ACCEPT) was not working when inside an atomic group.
1686    
1687    3.  Inside a character class, \B is treated as a literal by default, but
1688        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1689        causes the error). The code is unchanged, but I tidied the documentation.
1690    
1691    4.  Inside a character class, PCRE always treated \R and \X as literals,
1692        whereas Perl faults them if its -w option is set. I have changed PCRE so
1693        that it faults them when PCRE_EXTRA is set.
1694    
1695    5.  Added support for \N, which always matches any character other than
1696        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1697    
1698    6.  When compiling pcregrep with newer versions of gcc which may have
1699        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1700        declared with attribute warn_unused_result" were given. Just casting the
1701        result to (void) does not stop the warnings; a more elaborate fudge is
1702        needed. I've used a macro to implement this.
1703    
1704    7.  Minor change to pcretest.c to avoid a compiler warning.
1705    
1706    8.  Added four artifical Unicode properties to help with an option to make
1707        \s etc use properties (see next item). The new properties are: Xan
1708        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1709    
1710    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1711        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1712        this option. Modified pcretest to add /W to test this facility. Added
1713        REG_UCP to make it available via the POSIX interface.
1714    
1715    10. Added --line-buffered to pcregrep.
1716    
1717    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1718        studied, and the match started with a letter with a code point greater than
1719        127 whose first byte was different to the first byte of the other case of
1720        the letter, the other case of this starting letter was not recognized
1721        (#976).
1722    
1723    12. If a pattern that was studied started with a repeated Unicode property
1724        test, for example, \p{Nd}+, there was the theoretical possibility of
1725        setting up an incorrect bitmap of starting bytes, but fortunately it could
1726        not have actually happened in practice until change 8 above was made (it
1727        added property types that matched character-matching opcodes).
1728    
1729    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1730        possible starting bytes for non-anchored patterns.
1731    
1732    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1733        \R, and also a number of cases that involve Unicode properties, both
1734        explicit and implicit when PCRE_UCP is set.
1735    
1736    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1737        input, it could crash or give wrong results if characters with values
1738        greater than 0xc0 were present in the subject string. (Detail: it assumed
1739        UTF-8 input when processing these items.)
1740    
1741    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1742        size_t is 64-bit (#991).
1743    
1744    17. Added a check for running out of memory when PCRE is compiled with
1745        --disable-stack-for-recursion (#990).
1746    
1747    18. If the last data line in a file for pcretest does not have a newline on
1748        the end, a newline was missing in the output.
1749    
1750    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1751        less than 128) in its various bitmaps. However, there is a facility for
1752        generating tables according to the current locale when PCRE is compiled. It
1753        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1754        characters, are recognized by isspace() and therefore were getting set in
1755        these tables, and indeed these tables seem to approximate to ISO 8859. This
1756        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1757        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1758        which of course cannot start UTF-8 characters. I have changed the code so
1759        that only real ASCII characters (less than 128) and the correct starting
1760        bytes for UTF-8 encodings are set for characters greater than 127 when in
1761        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1762        altogether.)
1763    
1764    20. Added the /T option to pcretest so as to be able to run tests with non-
1765        standard character tables, thus making it possible to include the tests
1766        used for 19 above in the standard set of tests.
1767    
1768    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1769        reference to a subpattern the other side of a comment that contains an
1770        opening parenthesis caused either an internal compiling error, or a
1771        reference to the wrong subpattern.
1772    
1773    
1774    Version 8.02 19-Mar-2010
1775    ------------------------
1776    
1777    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1778    
1779    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1780        configured.
1781    
1782    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1783        original author of that file, following a query about its status.
1784    
1785    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1786        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1787    
1788    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1789        quantifier applied to a forward-referencing subroutine call, could compile
1790        incorrect code or give the error "internal error: previously-checked
1791        referenced subpattern not found".
1792    
1793    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1794        variables to point to external functions. For these systems, therefore,
1795        pcre_malloc etc. are now initialized to local functions that call the
1796        relevant global functions.
1797    
1798    7.  There were two entries missing in the vectors called coptable and poptable
1799        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1800        I've fixed the data, and added a kludgy way of testing at compile time that
1801        the lengths are correct (equal to the number of opcodes).
1802    
1803    8.  Following on from 7, I added a similar kludge to check the length of the
1804        eint vector in pcreposix.c.
1805    
1806    9.  Error texts for pcre_compile() are held as one long string to avoid too
1807        much relocation at load time. To find a text, the string is searched,
1808        counting zeros. There was no check for running off the end of the string,
1809        which could happen if a new error number was added without updating the
1810        string.
1811    
1812    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1813    
1814    11. \K was not working if it appeared in an atomic group or in a group that
1815        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1816        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1817        the assertion is positive, but not if it is negative.
1818    
1819    12. Change 11 fortuitously reduced the size of the stack frame used in the
1820        "match()" function of pcre_exec.c by one pointer. Forthcoming
1821        implementation of support for (*MARK) will need an extra pointer on the
1822        stack; I have reserved it now, so that the stack frame size does not
1823        decrease.
1824    
1825    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1826        item in branch that calls a recursion is a subroutine call - as in the
1827        second branch in the above example - was incorrectly given the compile-
1828        time error "recursive call could loop indefinitely" because pcre_compile()
1829        was not correctly checking the subroutine for matching a non-empty string.
1830    
1831    14. The checks for overrunning compiling workspace could trigger after an
1832        overrun had occurred. This is a "should never occur" error, but it can be
1833        triggered by pathological patterns such as hundreds of nested parentheses.
1834        The checks now trigger 100 bytes before the end of the workspace.
1835    
1836    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1837    
1838    
1839    Version 8.01 19-Jan-2010
1840    ------------------------
1841    
1842    1.  If a pattern contained a conditional subpattern with only one branch (in
1843        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1844        computed the wrong minimum data length (which is of course zero for such
1845        subpatterns). This could cause incorrect "no match" results.
1846    
1847    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1848        the pattern is reset in the first branch, pcre_compile() failed with
1849        "internal error: code overflow at offset...". This happened only when
1850        the reset was to the original external option setting. (An optimization
1851        abstracts leading options settings into an external setting, which was the
1852        cause of this.)
1853    
1854    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1855        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1856        assertion pattern did not match (meaning that the assertion was true), it
1857        was incorrectly treated as false if the SKIP had been reached during the
1858        matching. This also applied to assertions used as conditions.
1859    
1860    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1861        assertion subpattern, including such a pattern used as a condition,
1862        unpredictable results occurred, instead of the error return
1863        PCRE_ERROR_DFA_UITEM.
1864    
1865    5.  The C++ GlobalReplace function was not working like Perl for the special
1866        situation when an empty string is matched. It now does the fancy magic
1867        stuff that is necessary.
1868    
1869    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1870        removed. (These were left over from very, very early versions of PCRE.)
1871    
1872    7.  Some cosmetic changes to the code to make life easier when compiling it
1873        as part of something else:
1874    
1875        (a) Change DEBUG to PCRE_DEBUG.
1876    
1877        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1878            called "current" as "current_branch", to prevent a collision with the
1879            Linux macro when compiled as a kernel module.
1880    
1881        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1882            prevent a collision with the Linux macro when compiled as a kernel
1883            module.
1884    
1885    8.  In pcre_compile() there are some checks for integer overflows that used to
1886        cast potentially large values to (double). This has been changed to that
1887        when building, a check for int64_t is made, and if it is found, it is used
1888        instead, thus avoiding the use of floating point arithmetic. (There is no
1889        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1890        double.
1891    
1892    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1893        2005 (difference between two addresses compared to an unsigned value).
1894    
1895    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1896        custom one, because of the following reported problem in Windows:
1897    
1898          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1899              under Win32.
1900          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1901              therefore missing the function definition.
1902          - The compiler thus generates a "C" signature for the test function.
1903          - The linker fails to find the "C" function.
1904          - PCRE fails to configure if asked to do so against libbz2.
1905    
1906    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1907        messages were output:
1908    
1909          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1910          rerunning libtoolize, to keep the correct libtool macros in-tree.
1911          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1912    
1913        I have done both of these things.
1914    
1915    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1916        most of the time, it *can* run out if it is given a pattern that contains a
1917        runaway infinite recursion. I updated the discussion in the pcrestack man
1918        page.
1919    
1920    13. Now that we have gone to the x.xx style of version numbers, the minor
1921        version may start with zero. Using 08 or 09 is a bad idea because users
1922        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1923        interpreted as invalid octal numbers. I've updated the previous comment in
1924        configure.ac, and also added a check that gives an error if 08 or 09 are
1925        used.
1926    
1927    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1928        causing partial matching to fail when the end of the subject matched \W
1929        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1930    
1931    15. There were some discrepancies between the declarations in pcre_internal.h
1932        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1933        their definitions. The declarations used "const uschar *" and the
1934        definitions used USPTR. Even though USPTR is normally defined as "const
1935        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1936        reported that: "This difference in casting confuses some C++ compilers, for
1937        example, SunCC recognizes above declarations as different functions and
1938        generates broken code for hbpcre." I have changed the declarations to use
1939        USPTR.
1940    
1941    16. GNU libtool is named differently on some systems. The autogen.sh script now
1942        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1943        (FreeBSD).
1944    
1945    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1946        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1947        comment: "Figure out how to create a longlong from a string: strtoll and
1948        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1949        instance, but it only takes 2 args instead of 3!"
1950    
1951    18. A subtle bug concerned with back references has been fixed by a change of
1952        specification, with a corresponding code fix. A pattern such as
1953        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1954        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1955        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1956        same bug. Such groups have to be quantified to be useful, or contained
1957        inside another quantified group. (If there's no repetition, the reference
1958        can never match.) The problem arises because, having left the group and
1959        moved on to the rest of the pattern, a later failure that backtracks into
1960        the group uses the captured value from the final iteration of the group
1961        rather than the correct earlier one. I have fixed this in PCRE by forcing
1962        any group that contains a reference to itself to be an atomic group; that
1963        is, there cannot be any backtracking into it once it has completed. This is
1964        similar to recursive and subroutine calls.
1965    
1966    
1967    Version 8.00 19-Oct-09
1968    ----------------------
1969    
1970    1.  The table for translating pcre_compile() error codes into POSIX error codes
1971        was out-of-date, and there was no check on the pcre_compile() error code
1972        being within the table. This could lead to an OK return being given in
1973        error.
1974    
1975    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1976        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1977        in a Windows environment.
1978    
1979    3.  The pcregrep --count option prints the count for each file even when it is
1980        zero, as does GNU grep. However, pcregrep was also printing all files when
1981        --files-with-matches was added. Now, when both options are given, it prints
1982        counts only for those files that have at least one match. (GNU grep just
1983        prints the file name in this circumstance, but including the count seems
1984        more useful - otherwise, why use --count?) Also ensured that the
1985        combination -clh just lists non-zero counts, with no names.
1986    
1987    4.  The long form of the pcregrep -F option was incorrectly implemented as
1988        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1989        but it seems right to fix it, and I didn't think it was worth preserving
1990        the old behaviour.
1991    
1992    5.  The command line items --regex=pattern and --regexp=pattern were not
1993        recognized by pcregrep, which required --regex pattern or --regexp pattern
1994        (with a space rather than an '='). The man page documented the '=' forms,
1995        which are compatible with GNU grep; these now work.
1996    
1997    6.  No libpcreposix.pc file was created for pkg-config; there was just
1998        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1999    
2000    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2001        when UCP support is not needed, by modifying the Python script that
2002        generates it from Unicode data files. This should not matter if the module
2003        is correctly used as a library, but I received one complaint about 50K of
2004        unwanted data. My guess is that the person linked everything into his
2005        program rather than using a library. Anyway, it does no harm.
2006    
2007    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2008        was a minimum greater than 1 for a wide character in a possessive
2009        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2010        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2011        character. Chaos in the form of incorrect output or a compiling loop could
2012        result.
2013    
2014    9.  The restrictions on what a pattern can contain when partial matching is
2015        requested for pcre_exec() have been removed. All patterns can now be
2016        partially matched by this function. In addition, if there are at least two
2017        slots in the offset vector, the offset of the earliest inspected character
2018        for the match and the offset of the end of the subject are set in them when
2019        PCRE_ERROR_PARTIAL is returned.
2020    
2021    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2022        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2023        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2024        and may be more useful for multi-segment matching.
2025    
2026    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2027        used to be given if ever the end of the subject was reached; now it is
2028        given only if matching could not proceed because another character was
2029        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2030        string "Z", which now yields "no match" instead of "partial match". In the
2031        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2032        final character ended with (*FAIL).
2033    
2034    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2035        if the pattern had a "must contain" character that was already found in the
2036        earlier partial match, unless partial matching was again requested. For
2037        example, with the pattern /dog.(body)?/, the "must contain" character is
2038        "g". If the first part-match was for the string "dog", restarting with
2039        "sbody" failed. This bug has been fixed.
2040    
2041    13. The string returned by pcre_dfa_exec() after a partial match has been
2042        changed so that it starts at the first inspected character rather than the
2043        first character of the match. This makes a difference only if the pattern
2044        starts with a lookbehind assertion or \b or \B (\K is not supported by
2045        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2046        matching functions compatible, and I think it's the right thing to do.
2047    
2048    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2049        so that the demonstration program is easily available in environments where
2050        PCRE has not been installed from source.
2051    
2052    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2053        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2054        library.
2055    
2056    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2057        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2058        is not the first non-POSIX option to be added. Clearly some people find
2059        these options useful.
2060    
2061    17. If a caller to the POSIX matching function regexec() passes a non-zero
2062        value for nmatch with a NULL value for pmatch, the value of
2063        nmatch is forced to zero.
2064    
2065    18. RunGrepTest did not have a test for the availability of the -u option of
2066        the diff command, as RunTest does. It now checks in the same way as
2067        RunTest, and also checks for the -b option.
2068    
2069    19. If an odd number of negated classes containing just a single character
2070        interposed, within parentheses, between a forward reference to a named
2071        subpattern and the definition of the subpattern, compilation crashed with
2072        an internal error, complaining that it could not find the referenced
2073        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2074        [The bug was that it was starting one character too far in when skipping
2075        over the character class, thus treating the ] as data rather than
2076        terminating the class. This meant it could skip too much.]
2077    
2078    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2079        /g option in pcretest when the pattern contains \K, which makes it possible
2080        to have an empty string match not at the start, even when the pattern is
2081        anchored. Updated pcretest and pcredemo to use this option.
2082    
2083    21. If the maximum number of capturing subpatterns in a recursion was greater
2084        than the maximum at the outer level, the higher number was returned, but
2085        with unset values at the outer level. The correct (outer level) value is
2086        now given.
2087    
2088    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2089        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2090        make it do so. The string so far is captured, making this feature
2091        compatible with Perl.
2092    
2093    23. The tests have been re-organized, adding tests 11 and 12, to make it
2094        possible to check the Perl 5.10 features against Perl 5.10.
2095    
2096    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2097        pattern matches a fixed length string. PCRE did not allow this; now it
2098        does. Neither allows recursion.
2099    
2100    25. I finally figured out how to implement a request to provide the minimum
2101        length of subject string that was needed in order to match a given pattern.
2102        (It was back references and recursion that I had previously got hung up
2103        on.) This code has now been added to pcre_study(); it finds a lower bound
2104        to the length of subject needed. It is not necessarily the greatest lower
2105        bound, but using it to avoid searching strings that are too short does give
2106        some useful speed-ups. The value is available to calling programs via
2107        pcre_fullinfo().
2108    
2109    26. While implementing 25, I discovered to my embarrassment that pcretest had
2110        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2111        study optimizations had never been tested with that matching function.
2112        Oops. What is worse, even when it was passed study data, there was a bug in
2113        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2114        were also very few tests of studied patterns with pcre_dfa_exec().
2115    
2116    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2117        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2118        on the other side of the coin, they are no longer allowed to have different
2119        names, because these cannot be distinguished in PCRE, and this has caused
2120        confusion. (This is a difference from Perl.)
2121    
2122    28. When duplicate subpattern names are present (necessarily with different
2123        numbers, as required by 27 above), and a test is made by name in a
2124        conditional pattern, either for a subpattern having been matched, or for
2125        recursion in such a pattern, all the associated numbered subpatterns are
2126        tested, and the overall condition is true if the condition is true for any
2127        one of them. This is the way Perl works, and is also more like the way
2128        testing by number works.
2129    
2130    
2131    Version 7.9 11-Apr-09
2132    ---------------------
2133    
2134    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2135        (pcretest), all targets were linked against these libraries. This included
2136        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2137        libraries. This caused unwanted dependencies to be created. This problem
2138        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2139        pcretest is linked with readline.
2140    
2141    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2142        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2143        moved outside it again, because FALSE and TRUE are already defined in AIX,
2144        but BOOL is not.
2145    
2146    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2147        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2148    
2149    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2150        hyphens) following file names and line numbers when outputting matching
2151        lines. This is not true; no spaces are inserted. I have also clarified the
2152        wording for the --colour (or --color) option.
2153    
2154    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2155        was not coloured; this is different to GNU grep, so I have changed it to be
2156        the same.
2157    
2158    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2159        each matching line was coloured. Now it goes on to look for further matches
2160        of any of the test patterns, which is the same behaviour as GNU grep.
2161    
2162    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2163        doesn't make sense to accept an empty string match in pcregrep, so I have
2164        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2165        seems to be how GNU grep behaves. [But see later change 40 for release
2166        8.33.]
2167    
2168    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2169        start or after a newline", because the conditional assertion was not being
2170        correctly handled. The rule now is that both the assertion and what follows
2171        in the first alternative must satisfy the test.
2172    
2173    9.  If auto-callout was enabled in a pattern with a conditional group whose
2174        condition was an assertion, PCRE could crash during matching, both with
2175        pcre_exec() and pcre_dfa_exec().
2176    
2177    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2178        used for matching.
2179    
2180    11. Unicode property support in character classes was not working for
2181        characters (bytes) greater than 127 when not in UTF-8 mode.
2182    
2183    12. Added the -M command line option to pcretest.
2184    
2185    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2186    
2187    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2188    
2189    16. Added comments and documentation about mis-use of no_arg in the C++
2190        wrapper.
2191    
2192    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2193        from Martin Jerabek that uses macro names for all relevant character and
2194        string constants.
2195    
2196    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2197        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2198        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2199        these, but not everybody uses configure.
2200    
2201    19. A conditional group that had only one branch was not being correctly
2202        recognized as an item that could match an empty string. This meant that an
2203        enclosing group might also not be so recognized, causing infinite looping
2204        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2205        with the subject "ab", where knowledge that the repeated group can match
2206        nothing is needed in order to break the loop.
2207    
2208    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2209        exec(), but without supplying a callout function, matching went wrong.
2210    
2211    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2212        leak if the size of the offset vector was greater than 30. When the vector
2213        is smaller, the saved offsets during recursion go onto a local stack
2214        vector, but for larger vectors malloc() is used. It was failing to free
2215        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2216        error, in fact).
2217    
2218    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2219        heapframe that is used only when UTF-8 support is enabled. This caused no
2220        problem, but was untidy.
2221    
2222    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2223        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2224        included within another project.
2225    
2226    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2227        slightly modified by me:
2228    
2229          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2230              not building pcregrep.
2231    
2232          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2233              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2234    
2235    25. Forward references, both numeric and by name, in patterns that made use of
2236        duplicate group numbers, could behave incorrectly or give incorrect errors,
2237        because when scanning forward to find the reference group, PCRE was not
2238        taking into account the duplicate group numbers. A pattern such as
2239        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2240    
2241    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2242        the feature of a custom pointer more persuasive (as requested by a user).
2243    
2244    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2245        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2246        pre-defined.
2247    
2248    28. Added support for (*UTF8) at the start of a pattern.
2249    
2250    29. Arrange for flags added by the "release type" setting in CMake to be shown
2251        in the configuration summary.
2252    
2253    
2254    Version 7.8 05-Sep-08
2255    ---------------------
2256    
2257    1.  Replaced UCP searching code with optimized version as implemented for Ad
2258        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2259        stage table and inline lookup instead of a function, giving speed ups of 2
2260        to 5 times on some simple patterns that I tested. Permission was given to
2261        distribute the MultiStage2.py script that generates the tables (it's not in
2262        the tarball, but is in the Subversion repository).
2263    
2264    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2265        scripts.
2266    
2267    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2268        a group with a zero qualifier. The result of the study could be incorrect,
2269        or the function might crash, depending on the pattern.
2270    
2271    4.  Caseless matching was not working for non-ASCII characters in back
2272        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2273        It now works when Unicode Property Support is available.
2274    
2275    5.  In pcretest, an escape such as \x{de} in the data was always generating
2276        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2277        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2278        truncation.
2279    
2280    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2281    
2282    7.  Added two (int) casts to pcregrep when printing the difference of two
2283        pointers, in case they are 64-bit values.
2284    
2285    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2286        test 2 if it fails.
2287    
2288    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2289        and a #define of that name to empty if it is not externally set. This is to
2290        allow users of MSVC to set it if necessary.
2291    
2292    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2293        the convenience functions in the pcre_get.c source file.
2294    
2295    11. An option change at the start of a pattern that had top-level alternatives
2296        could cause overwriting and/or a crash. This command provoked a crash in
2297        some environments:
2298    
2299          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2300    
2301        This potential security problem was recorded as CVE-2008-2371.
2302    
2303    12. For a pattern where the match had to start at the beginning or immediately
2304        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2305        pcre_dfa_exec() could read past the end of the passed subject if there was
2306        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2307        pcretest so that it places the subject at the end of its malloc-ed buffer.
2308    
2309    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2310        exec() might read past the end of the data buffer in UTF-8 mode.
2311    
2312    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2313        the data contained the byte 0x85 as part of a UTF-8 character within its
2314        first line. This applied both to normal and DFA matching.
2315    
2316    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2317        /^[^d]*?$/8 failed to match "abc".
2318    
2319    16. Added a missing copyright notice to pcrecpp_internal.h.
2320    
2321    17. Make it more clear in the documentation that values returned from
2322        pcre_exec() in ovector are byte offsets, not character counts.
2323    
2324    18. Tidied a few places to stop certain compilers from issuing warnings.
2325    
2326    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2327        supplied by Stefan Weber. I made a further small update for 7.8 because
2328        there is a change of source arrangements: the pcre_searchfuncs.c module is
2329        replaced by pcre_ucd.c.
2330    
2331    
2332    Version 7.7 07-May-08
2333    ---------------------
2334    
2335    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2336        a string to a long long, pretend we don't even have a long long." This is
2337        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2338    
2339    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2340        pre-7.6 versions, which defined a global no_arg variable instead of putting
2341        it in the RE class. (See also #8 below.)
2342    
2343    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2344        Lopes.
2345    
2346    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2347    
2348        (1) The include/exclude patterns were being applied to the whole pathnames
2349            of files, instead of just to the final components.
2350    
2351        (2) If there was more than one level of directory, the subdirectories were
2352            skipped unless they satisfied the include/exclude conditions. This is
2353            inconsistent with GNU grep (and could even be seen as contrary to the
2354            pcregrep specification - which I improved to make it absolutely clear).
2355            The action now is always to scan all levels of directory, and just
2356            apply the include/exclude patterns to regular files.
2357    
2358    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2359        --exclude_dir in the tests to avoid scanning .svn directories.
2360    
2361    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2362        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2363        doesn't support NULs in patterns.
2364    
2365    7.  Added some missing "const"s to declarations of static tables in
2366        pcre_compile.c and pcre_dfa_exec.c.
2367    
2368    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2369        caused by fix #2  above. (Subsequently also a second patch to fix the
2370        first patch. And a third patch - this was a messy problem.)
2371    
2372    9.  Applied Craig's patch to remove the use of push_back().
2373    
2374    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2375        matching function regexec().
2376    
2377    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2378        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2379        references. PCRE supports relative numbers with this syntax (I don't think
2380        Oniguruma does).
2381    
2382    12. Previously, a group with a zero repeat such as (...){0} was completely
2383        omitted from the compiled regex. However, this means that if the group
2384        was called as a subroutine from elsewhere in the pattern, things went wrong
2385        (an internal error was given). Such groups are now left in the compiled
2386        pattern, with a new opcode that causes them to be skipped at execution
2387        time.
2388    
2389    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2390        to the way PCRE behaves:
2391    
2392        (a) A lone ] character is dis-allowed (Perl treats it as data).
2393    
2394        (b) A back reference to an unmatched subpattern matches an empty string
2395            (Perl fails the current match path).
2396    
2397        (c) A data ] in a character class must be notated as \] because if the
2398            first data character in a class is ], it defines an empty class. (In
2399            Perl it is not possible to have an empty class.) The empty class []
2400            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2401            The negative empty class [^] matches any one character, independently
2402            of the DOTALL setting.
2403    
2404    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2405        non-existent subpattern following a character class starting with ']' and
2406        containing () gave an internal compiling error instead of "reference to
2407        non-existent subpattern". Fortunately, when the pattern did exist, the
2408        compiled code was correct. (When scanning forwards to check for the
2409        existence of the subpattern, it was treating the data ']' as terminating
2410        the class, so got the count wrong. When actually compiling, the reference
2411        was subsequently set up correctly.)
2412    
2413    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2414        it was being rejected as not supported by pcre_dfa_exec(), even though
2415        other assertions are supported. I have made pcre_dfa_exec() support
2416        (*FAIL).
2417    
2418    16. The implementation of 13c above involved the invention of a new opcode,
2419        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2420        cannot be changed at match time, I realized I could make a small
2421        improvement to matching performance by compiling OP_ALLANY instead of
2422        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2423        on the OP_ANY path.
2424    
2425    17. Compiling pcretest on Windows with readline support failed without the
2426        following two fixes: (1) Make the unistd.h include conditional on
2427        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2428    
2429    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2430        ncurses library to be included for pcretest when ReadLine support is
2431        requested, but also to allow for it to be overridden. This patch came from
2432        Daniel Bergström.
2433    
2434    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2435        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2436        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2437        spotting this.
2438    
2439    
2440    Version 7.6 28-Jan-08
2441    ---------------------
2442    
2443    1.  A character class containing a very large number of characters with
2444        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2445        overflow.
2446    
2447    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2448        HAVE_LONG_LONG is not defined.
2449    
2450    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2451        bring them up to date and include new features. This patch includes:
2452    
2453        - Fixed PH's badly added libz and libbz2 support.
2454        - Fixed a problem with static linking.
2455        - Added pcredemo. [But later removed - see 7 below.]
2456        - Fixed dftables problem and added an option.
2457        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2458            HAVE_LONG_LONG.
2459        - Added readline support for pcretest.
2460        - Added an listing of the option settings after cmake has run.
2461    
2462    4.  A user submitted a patch to Makefile that makes it easy to create
2463        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2464        Makefile.am that cause it to include this special target, without
2465        affecting anything else. Note that the same mingw target plus all
2466        the other distribution libraries and programs are now supported
2467        when configuring with CMake (see 6 below) instead of with
2468        Configure/Make.
2469    
2470    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2471        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2472        exported in the Windows port". It has not yet been confirmed that the patch
2473        solves the problem, but it does no harm.
2474    
2475    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2476        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2477        with CMake, and also correct the comment about stack recursion.
2478    
2479    7.  Remove the automatic building of pcredemo from the ./configure system and
2480        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2481        of a program that users should build themselves after PCRE is installed, so
2482        building it automatically is not really right. What is more, it gave
2483        trouble in some build environments.
2484    
2485    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2486    
2487    
2488    Version 7.5 10-Jan-08
2489    ---------------------
2490    
2491    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2492        values in parens when parsing an RE using the C++ wrapper."
2493    
2494    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2495        Characters greater than 255 were excluded from the class instead of being
2496        included.
2497    
2498    3.  The same bug as (2) above applied to negated POSIX classes such as
2499        [:^space:].
2500    
2501    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2502        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2503        I have changed it.
2504    
2505    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2506        first named subpattern) and a construct such as (?&a) would reference the
2507        first named subpattern whose name started with "a" (in other words, the
2508        length check was missing). Both these problems are fixed. "Subpattern name
2509        expected" is now given for (?&) (a zero-length name), and this patch also
2510        makes it give the same error for \k'' (previously it complained that that
2511        was a reference to a non-existent subpattern).
2512    
2513    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2514        this is right because (?- can be followed by option settings as well as by
2515        digits. I have, however, made the messages clearer.
2516    
2517    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2518        than the number used in the conditional) now cause a compile-time error.
2519        This is actually not compatible with Perl, which accepts such patterns, but
2520        treats the conditional as always being FALSE (as PCRE used to), but it
2521        seems to me that giving a diagnostic is better.
2522    
2523    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2524        and messages.
2525    
2526    9.  Fix two occurrences of "backslash" in comments that should have been
2527        "backspace".
2528    
2529    10. Remove two redundant lines of code that can never be obeyed (their function
2530        was moved elsewhere).
2531    
2532    11. The program that makes PCRE's Unicode character property table had a bug
2533        which caused it to generate incorrect table entries for sequences of
2534        characters that have the same character type, but are in different scripts.
2535        It amalgamated them into a single range, with the script of the first of
2536        them. In other words, some characters were in the wrong script. There were
2537        thirteen such cases, affecting characters in the following ranges:
2538    
2539          U+002b0 - U+002c1
2540          U+0060c - U+0060d
2541          U+0061e - U+00612
2542          U+0064b - U+0065e
2543          U+0074d - U+0076d
2544          U+01800 - U+01805
2545          U+01d00 - U+01d77
2546          U+01d9b - U+01dbf
2547          U+0200b - U+0200f
2548          U+030fc - U+030fe
2549          U+03260 - U+0327f
2550          U+0fb46 - U+0fbb1
2551          U+10450 - U+1049d
2552    
2553    12. The -o option (show only the matching part of a line) for pcregrep was not
2554        compatible with GNU grep in that, if there was more than one match in a
2555        line, it showed only the first of them. It now behaves in the same way as
2556        GNU grep.
2557    
2558    13. If the -o and -v options were combined for pcregrep, it printed a blank
2559        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2560        does the same. The return code can be used to tell if there were any
2561        non-matching lines.
2562    
2563    14. Added --file-offsets and --line-offsets to pcregrep.
2564    
2565    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2566        infinitely looping recursion. The bug was that positive lookaheads were not
2567        being skipped when checking for a possible empty match (negative lookaheads
2568        and both kinds of lookbehind were skipped).
2569    
2570    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2571        inclusion of <windows.h> to before rather than after the definition of
2572        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2573    
2574    17. Specifying a possessive quantifier with a specific limit for a Unicode
2575        character property caused pcre_compile() to compile bad code, which led at
2576        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2577        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2578        caused the error; without that there was no problem.
2579    
2580    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2581    
2582    19. Added --enable-pcretest-libreadline.
2583    
2584    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2585        RE::GlobalReplace(). As a result, the number of replacements returned was
2586        double what it should be. I removed one of the increments, but Craig sent a
2587        later patch that removed the other one (the right fix) and added unit tests
2588        that check the return values (which was not done before).
2589    
2590    21. Several CMake things:
2591    
2592        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2593            the names libpcre and libpcreposix, not just pcre and pcreposix.
2594    
2595        (2) The above change means that pcretest and pcregrep are now correctly
2596            linked with the newly-built libraries, not previously installed ones.
2597    
2598        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2599    
2600    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2601        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2602        UTF-8 newline character). The key issue is that the pattern starts .*;
2603        this means that the match must be either at the beginning, or after a
2604        newline. The bug was in the code for advancing after a failed match and
2605        checking that the new position followed a newline. It was not taking
2606        account of UTF-8 characters correctly.
2607    
2608    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2609        character classes. PCRE was not treating the sequence [:...:] as a
2610        character class unless the ... were all letters. Perl, however, seems to
2611        allow any characters between [: and :], though of course it rejects as
2612        unknown any "names" that contain non-letters, because all the known class
2613        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2614        for example, whereas PCRE did not - it did not recognize a POSIX character
2615        class. This seemed a bit dangerous, so the code has been changed to be
2616        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2617        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2618        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2619        Perl does, and where it didn't before.
2620    
2621    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2622        Windows environments %n is disabled by default.
2623    
2624    
2625    Version 7.4 21-Sep-07
2626    ---------------------
2627    
2628    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2629        means that a class such as [\s] counted as "explicit reference to CR or
2630        LF". That isn't really right - the whole point of the change was to try to
2631        help when there was an actual mention of one of the two characters. So now
2632        the change happens only if \r or \n (or a literal CR or LF) character is
2633        encountered.
2634    
2635    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2636        of both had grown to the point where there were only 3 bits left.
2637        Fortunately, there was spare space in the data structure, and so I have
2638        moved the internal flags into a new 16-bit field to free up more option
2639        bits.
2640    
2641    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2642        but did not set the internal JCHANGED flag - either of these is enough to
2643        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2644        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2645        start sets both bits.
2646    
2647    4.  Added options (at build time, compile time, exec time) to change \R from
2648        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2649    
2650    5.  doc/pcresyntax.html was missing from the distribution.
2651    
2652    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2653        compatibility, even though it is no longer used.
2654    
2655    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2656        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2657        windows.h file is present (where different names are used). [This was
2658        reversed later after testing - see 16 below.]
2659    
2660    8.  Changed all #include <config.h> to #include "config.h". There were also
2661        some further <pcre.h> cases that I changed to "pcre.h".
2662    
2663    9.  When pcregrep was used with the --colour option, it missed the line ending
2664        sequence off the lines that it output.
2665    
2666    10. It was pointed out to me that arrays of string pointers cause lots of
2667        relocations when a shared library is dynamically loaded. A technique of
2668        using a single long string with a table of offsets can drastically reduce
2669        these. I have refactored PCRE in four places to do this. The result is
2670        dramatic:
2671    
2672          Originally:                          290
2673          After changing UCP table:            187
2674          After changing error message table:   43
2675          After changing table of "verbs"       36
2676          After changing table of Posix names   22
2677    
2678        Thanks to the folks working on Gregex for glib for this insight.
2679    
2680    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2681        unicode-properties was also set.
2682    
2683    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2684    
2685    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2686        checked only for CRLF.
2687    
2688    14. Added casts to pcretest.c to avoid compiler warnings.
2689    
2690    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2691    
2692    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2693        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2694        entirely. This removes changes made in 7 above.
2695    
2696    17. The CMake files have been updated, and there is now more information about
2697        building with CMake in the NON-UNIX-USE document.
2698    
2699    
2700    Version 7.3 28-Aug-07
2701    ---------------------
2702    
2703     1. In the rejigging of the build system that eventually resulted in 7.1, the
2704        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2705        brackets there is not right, since it causes compilers to look for an
2706        installed pcre.h, not the version that is in the source that is being
2707        compiled (which of course may be different). I have changed it back to:
2708    
2709          #include "pcre.h"
2710    
2711        I have a vague recollection that the change was concerned with compiling in
2712        different directories, but in the new build system, that is taken care of
2713        by the VPATH setting the Makefile.
2714    
2715     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2716        when the subject happened to end in the byte 0x85 (e.g. if the last
2717        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2718        characters but of course it shouldn't be taken as a newline when it is part
2719        of another character. The bug was that, for an unlimited repeat of . in
2720        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2721        characters when looking for a newline.
2722    
2723     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2724    
2725     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2726        in debug output.
2727    
2728     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2729        long printing in the pcrecpp unittest when running under MinGW.
2730    
2731     6. ESC_K was left out of the EBCDIC table.
2732    
2733     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2734        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2735        limit also applies to "virtual nesting" when a pattern is recursive, and in
2736        this case 1000 isn't so big. I have been able to remove this limit at the
2737        expense of backing off one optimization in certain circumstances. Normally,
2738        when pcre_exec() would call its internal match() function recursively and
2739        immediately return the result unconditionally, it uses a "tail recursion"
2740        feature to save stack. However, when a subpattern that can match an empty
2741        string has an unlimited repetition quantifier, it no longer makes this
2742        optimization. That gives it a stack frame in which to save the data for
2743        checking that an empty string has been matched. Previously this was taken
2744        from the 1000-entry workspace that had been reserved. So now there is no
2745        explicit limit, but more stack is used.
2746    
2747     8. Applied Daniel's patches to solve problems with the import/export magic
2748        syntax that is required for Windows, and which was going wrong for the
2749        pcreposix and pcrecpp parts of the library. These were overlooked when this
2750        problem was solved for the main library.
2751    
2752     9. There were some crude static tests to avoid integer overflow when computing
2753        the size of patterns that contain repeated groups with explicit upper
2754        limits. As the maximum quantifier is 65535, the maximum group length was
2755        set at 30,000 so that the product of these two numbers did not overflow a
2756        32-bit integer. However, it turns out that people want to use groups that
2757        are longer than 30,000 bytes (though not repeat them that many times).
2758        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2759        made it possible to implement the integer overflow checks in a much more
2760        dynamic way, which I have now done. The artificial limitation on group
2761        length has been removed - we now have only the limit on the total length of
2762        the compiled pattern, which depends on the LINK_SIZE setting.
2763    
2764    10. Fixed a bug in the documentation for get/copy named substring when
2765        duplicate names are permitted. If none of the named substrings are set, the
2766        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2767        empty string.
2768    
2769    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2770        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2771        because the ] is interpreted as the first data character and the
2772        terminating ] is not found. PCRE has been made compatible with Perl in this
2773        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2774        cause memory overwriting.
2775    
2776    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2777        string has been matched (to stop an infinite loop). It was not recognizing
2778        a conditional subpattern that could match an empty string if that
2779        subpattern was within another subpattern. For example, it looped when
2780        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2781        condition was not nested. This bug has been fixed.
2782    
2783    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2784        past the start of the subject in the presence of bytes with the top bit
2785        set, for example "\x8aBCD".
2786    
2787    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2788        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2789    
2790    14. Optimized (?!) to (*FAIL).
2791    
2792    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2793        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2794        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2795        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2796        does: it's just the validity check that is more restrictive.
2797    
2798    16. Inserted checks for integer overflows during escape sequence (backslash)
2799        processing, and also fixed erroneous offset values for syntax errors during
2800        backslash processing.
2801    
2802    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2803        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2804    
2805    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2806        caused an overrun.
2807    
2808    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2809        something other than just ASCII characters) inside a group that had an
2810        unlimited repeat caused a loop at compile time (while checking to see
2811        whether the group could match an empty string).
2812    
2813    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2814        [\P{Any}] did so. (Error in the code for printing property names.)
2815    
2816    21. An orphan \E inside a character class could cause a crash.
2817    
2818    22. A repeated capturing bracket such as (A)? could cause a wild memory
2819        reference during compilation.
2820    
2821    23. There are several functions in pcre_compile() that scan along a compiled
2822        expression for various reasons (e.g. to see if it's fixed length for look
2823        behind). There were bugs in these functions when a repeated \p or \P was
2824        present in the pattern. These operators have additional parameters compared
2825        with \d, etc, and these were not being taken into account when moving along
2826        the compiled data. Specifically:
2827    
2828        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2829            length.
2830    
2831        (b) An item such as \pL+ within a repeated group could cause crashes or
2832            loops.
2833    
2834        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2835            "reference to non-existent subpattern" error.
2836    
2837        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2838    
2839    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2840        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2841    
2842    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2843    
2844    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2845        character were causing crashes (broken optimization).
2846    
2847    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2848        \p or \P) caused a compile-time loop.
2849    
2850    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2851        break. For example, the unstudied pattern [\r\n]A does not match the string
2852        "\r\nA" because change 7.0/46 below moves the current point on by two
2853        characters after failing to match at the start. However, the pattern \nA
2854        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2855        the same is true. There doesn't seem any very clean way out of this, but
2856        what I have chosen to do makes the common cases work: PCRE now takes note
2857        of whether there can be an explicit match for \r or \n anywhere in the
2858        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2859        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2860        pattern has explicit CR or LF references.
2861    
2862    29. Added (*CR) etc for changing newline setting at start of pattern.
2863    
2864    
2865    Version 7.2 19-Jun-07
2866    ---------------------
2867    
2868     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2869        which is apparently normally available under Windows.
2870    
2871     2. Re-jig the pcregrep tests with different newline settings in an attempt
2872        to make them independent of the local environment's newline setting.
2873    
2874     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2875    
2876     4. Some of the "internals" tests were previously cut out when the link size
2877        was not 2, because the output contained actual offsets. The recent new
2878        "Z" feature of pcretest means that these can be cut out, making the tests
2879        usable with all link sizes.
2880    
2881     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2882        stack recursion. This gives a massive performance boost under BSD, but just
2883        a small improvement under Linux. However, it saves one field in the frame
2884        in all cases.
2885    
2886     6. Added more features from the forthcoming Perl 5.10:
2887    
2888        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2889            recursion call. It refers to the nth most recently opened parentheses.
2890    
2891        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2892            to be opened parentheses.
2893    
2894        (c) Conditions that refer to capturing parentheses can be specified
2895            relatively, for example, (?(-2)... or (?(+3)...
2896    
2897        (d) \K resets the start of the current match so that everything before
2898            is not part of it.
2899    
2900        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2901    
2902        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2903            reference syntax.
2904    
2905        (g) (?| introduces a group in which the numbering of parentheses in each
2906            alternative starts with the same number.
2907    
2908        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2909    
2910     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2911        PCRE_INFO_JCHANGED.
2912    
2913     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2914        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2915        for detecting groups that can match an empty string.
2916    
2917     9. A pattern with a very large number of alternatives (more than several
2918        hundred) was running out of internal workspace during the pre-compile
2919        phase, where pcre_compile() figures out how much memory will be needed. A
2920        bit of new cunning has reduced the workspace needed for groups with
2921        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2922        workspace instead of running out of the 4096 that are available.
2923    
2924    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2925    
2926    11. Applied patch from Google to remove an optimization that didn't quite work.
2927        The report of the bug said:
2928    
2929          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2930          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2931          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2932    
2933    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2934        it matched the wrong number of bytes.
2935    
2936    
2937    Version 7.1 24-Apr-07
2938    ---------------------
2939    
2940     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2941        that is more "standard", making use of automake and other Autotools. There
2942        is some re-arrangement of the files and adjustment of comments consequent
2943        on this.
2944    
2945     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2946        for recursive directory scanning broke on some systems because the files
2947        are not scanned in any specific order and on different systems the order
2948        was different. A call to "sort" has been inserted into RunGrepTest for the
2949        approprate test as a short-term fix. In the longer term there may be an
2950        alternative.
2951    
2952     3. I had an email from Eric Raymond about problems translating some of PCRE's
2953        man pages to HTML (despite the fact that I distribute HTML pages, some
2954        people do their own conversions for various reasons). The problems
2955        concerned the use of low-level troff macros .br and .in. I have therefore
2956        removed all such uses from the man pages (some were redundant, some could
2957        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2958        HTML has been updated to handle .nf/.fi and to complain if it encounters
2959        .br or .in.
2960    
2961     4. Updated comments in configure.ac that get placed in config.h.in and also
2962        arranged for config.h to be included in the distribution, with the name
2963        config.h.generic, for the benefit of those who have to compile without
2964        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2965    
2966     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2967        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2968        makevp.bat for the current PCRE, using the additional files
2969        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2970    
2971     6. A Windows user reported a minor discrepancy with test 2, which turned out
2972        to be caused by a trailing space on an input line that had got lost in his
2973        copy. The trailing space was an accident, so I've just removed it.
2974    
2975     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2976        that is needed.
2977    
2978     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2979        as "const" (a) because they are and (b) because it helps the PHP
2980        maintainers who have recently made a script to detect big data structures
2981        in the php code that should be moved to the .rodata section. I remembered
2982        to update Builducptable as well, so it won't revert if ucptable.h is ever
2983        re-created.
2984    
2985     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2986        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2987        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2988        support is not required. This saves 1.5-2K of code, which is important in
2989        some applications.
2990    
2991        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2992        so as not to refer to the tables, even though these functions will never be
2993        called when UTF-8 support is disabled. Otherwise there are problems with a
2994        shared library.
2995    
2996    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2997    
2998        (a) It was defining its arguments as char * instead of void *.
2999    
3000        (b) It was assuming that all moves were upwards in memory; this was true
3001            a long time ago when I wrote it, but is no longer the case.
3002    
3003        The emulated memove() is provided for those environments that have neither
3004        memmove() nor bcopy(). I didn't think anyone used it these days, but that
3005        is clearly not the case, as these two bugs were recently reported.
3006    
3007    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3008        and Detrail to create the HTML documentation, the .txt form of the man
3009        pages, and it removes trailing spaces from listed files. It also creates
3010        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3011        case, it wraps all the #defines with #ifndefs. This script should be run
3012        before "make dist".
3013    
3014    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3015        with Unicode property support.
3016    
3017        (a) For a maximizing quantifier, if the two different cases of the
3018            character were of different lengths in their UTF-8 codings (there are
3019            some cases like this - I found 11), and the matching function had to
3020            back up over a mixture of the two cases, it incorrectly assumed they
3021            were both the same length.
3022    
3023        (b) When PCRE was configured to use the heap rather than the stack for
3024            recursion during matching, it was not correctly preserving the data for
3025            the other case of a UTF-8 character when checking ahead for a match
3026            while processing a minimizing repeat. If the check also involved
3027            matching a wide character, but failed, corruption could cause an
3028            erroneous result when trying to check for a repeat of the original
3029            character.
3030    
3031    13. Some tidying changes to the testing mechanism:
3032    
3033        (a) The RunTest script now detects the internal link size and whether there
3034            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3035            values substituted by "configure". (The RunGrepTest script already did
3036            this for UTF-8.) The configure.ac script no longer substitutes the
3037            relevant variables.
3038    
3039        (b) The debugging options /B and /D in pcretest show the compiled bytecode
3040            with length and offset values. This means that the output is different
3041            for different internal link sizes. Test 2 is skipped for link sizes
3042            other than 2 because of this, bypassing the problem. Unfortunately,
3043            there was also a test in test 3 (the locale tests) that used /B and
3044            failed for link sizes other than 2. Rather than cut the whole test out,
3045            I have added a new /Z option to pcretest that replaces the length and
3046            offset values with spaces. This is now used to make test 3 independent
3047            of link size. (Test 2 will be tidied up later.)
3048    
3049    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3050        segmentation fault instead of returning the appropriate error message.
3051    
3052    15. In multiline mode when the newline sequence was set to "any", the pattern
3053        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3054        This doesn't seem right; it now treats the CRLF combination as the line
3055        ending, and so does not match in that case. It's only a pattern such as ^$
3056        that would hit this one: something like ^ABC$ would have failed after \r
3057        and then tried again after \r\n.
3058    
3059    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3060        in an attempt to make files that differ only in their line terminators
3061        compare equal. This works on Linux.
3062    
3063    17. Under certain error circumstances pcregrep might try to free random memory
3064        as it exited. This is now fixed, thanks to valgrind.
3065    
3066    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3067        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3068        was because its rules for how to advance for /g after matching an empty
3069        string at the end of a line did not allow for this case. They now check for
3070        it specially.
3071    
3072    20. pcretest is supposed to handle patterns and data of any length, by
3073        extending its buffers when necessary. It was getting this wrong when the
3074        buffer for a data line had to be extended.
3075    
3076    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3077        CRLF as a newline sequence.
3078    
3079    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3080        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3081        I have nevertheless tidied it up.
3082    
3083    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3084    
3085    24. Added a man page for pcre-config.
3086    
3087    
3088    Version 7.0 19-Dec-06
3089    ---------------------
3090    
3091     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3092        moving to gcc 4.1.1.
3093    
3094     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3095        sys/time.h, which is documented as needed for this function. It doesn't
3096        seem to matter on Linux, but it showed up on some releases of OS X.
3097    
3098     3. It seems that there are systems where bytes whose values are greater than
3099        127 match isprint() in the "C" locale. The "C" locale should be the
3100        default when a C program starts up. In most systems, only ASCII printing
3101        characters match isprint(). This difference caused the output from pcretest
3102        to vary, making some of the tests fail. I have changed pcretest so that:
3103    
3104        (a) When it is outputting text in the compiled version of a pattern, bytes
3105            other than 32-126 are always shown as hex escapes.
3106    
3107        (b) When it is outputting text that is a matched part of a subject string,
3108            it does the same, unless a different locale has been set for the match
3109            (using the /L modifier). In this case, it uses isprint() to decide.
3110    
3111     4. Fixed a major bug that caused incorrect computation of the amount of memory
3112        required for a compiled pattern when options that changed within the
3113        pattern affected the logic of the preliminary scan that determines the
3114        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3115        that the computed length was too small. The symptoms of this bug were
3116        either the PCRE error "internal error: code overflow" from pcre_compile(),
3117        or a glibc crash with a message such as "pcretest: free(): invalid next
3118        size (fast)". Examples of patterns that provoked this bug (shown in
3119        pcretest format) are:
3120    
3121          /(?-x: )/x
3122          /(?x)(?-x: \s*#\s*)/
3123          /((?i)[\x{c0}])/8
3124          /(?i:[\x{c0}])/8
3125    
3126        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3127        is now done differently.
3128    
3129     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3130        wrapper classes; (b) implement a new function in the C++ scanner that is
3131        more efficient than the old way of doing things because it avoids levels of
3132        recursion in the regex matching; (c) add a paragraph to the documentation
3133        for the FullMatch() function.
3134    
3135     6. The escape sequence \n was being treated as whatever was defined as
3136        "newline". Not only was this contrary to the documentation, which states
3137        that \n is character 10 (hex 0A), but it also went horribly wrong when
3138        "newline" was defined as CRLF. This has been fixed.
3139    
3140     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3141        was being set to -1 for the "end of line" case (supposedly a value that no
3142        character can have). Though this value is never used (the check for end of
3143        line is "zero bytes in current character"), it caused compiler complaints.
3144        I've changed it to 0xffffffff.
3145    
3146     8. In pcre_version.c, the version string was being built by a sequence of
3147        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3148        string (as it is for production releases) called a macro with an empty
3149        argument. The C standard says the result of this is undefined. The gcc
3150        compiler treats it as an empty string (which was what was wanted) but it is
3151        reported that Visual C gives an error. The source has been hacked around to
3152        avoid this problem.
3153    
3154     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3155        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3156        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3157        of them did).
3158    
3159    10. Originally, pcretest opened its input and output without "b"; then I was
3160        told that "b" was needed in some environments, so it was added for release
3161        5.0 to both the input and output. (It makes no difference on Unix-like
3162        systems.) Later I was told that it is wrong for the input on Windows. I've
3163        now abstracted the modes into two macros, to make it easier to fiddle with
3164        them, and removed "b" from the input mode under Windows.
3165    
3166    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3167    
3168    12. Added -help and --help to pcretest as an official way of being reminded
3169        of the options.
3170    
3171    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3172        and pcrecpp.cc because they annoy compilers at high warning levels.
3173    
3174    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3175    
3176    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3177        scripts are not C programs :-) and which was not noticed because it works
3178        on Linux.
3179    
3180    16. pcretest is supposed to handle any length of pattern and data line (as one
3181        line or as a continued sequence of lines) by extending its input buffer if
3182        necessary. This feature was broken for very long pattern lines, leading to
3183        a string of junk being passed to pcre_compile() if the pattern was longer
3184        than about 50K.
3185    
3186    17. I have done a major re-factoring of the way pcre_compile() computes the
3187        amount of memory needed for a compiled pattern. Previously, there was code
3188        that made a preliminary scan of the pattern in order to do this. That was
3189        OK when PCRE was new, but as the facilities have expanded, it has become
3190        harder and harder to keep it in step with the real compile phase, and there
3191        have been a number of bugs (see for example, 4 above). I have now found a
3192        cunning way of running the real compile function in a "fake" mode that
3193        enables it to compute how much memory it would need, while actually only
3194        ever using a few hundred bytes of working memory and without too many
3195        tests of the mode. This should make future maintenance and development
3196        easier. A side effect of this work is that the limit of 200 on the nesting
3197        depth of parentheses has been removed (though this was never a serious
3198        limitation, I suspect). However, there is a downside: pcre_compile() now
3199        runs more slowly than before (30% or more, depending on the pattern). I
3200        hope this isn't a big issue. There is no effect on runtime performance.
3201    
3202    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3203        newline (only possible for the last line of a file) and it was a
3204        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3205    
3206    19. Added additional timing features to pcretest. (1) The -tm option now times
3207        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3208        separate command line item, by a number that specifies the number of
3209        repeats to use when timing. The default is 50000; this gives better
3210        precision, but takes uncomfortably long for very large patterns.
3211    
3212    20. Extended pcre_study() to be more clever in cases where a branch of a
3213        subpattern has no definite first character. For example, (a*|b*)[cd] would
3214        previously give no result from pcre_study(). Now it recognizes that the
3215        first character must be a, b, c, or d.
3216    
3217    21. There was an incorrect error "recursive call could loop indefinitely" if
3218        a subpattern (or the entire pattern) that was being tested for matching an
3219        empty string contained only one non-empty item after a nested subpattern.
3220        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3221        incorrectly, because the \d was being skipped in the check.
3222    
3223    22. The pcretest program now has a new pattern option /B and a command line
3224        option -b, which is equivalent to adding /B to every pattern. This causes
3225        it to show the compiled bytecode, without the additional information that
3226        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3227        is the same as /B/I).
3228    
3229    23. A new optimization is now able automatically to treat some sequences such
3230        as a*b as a*+b. More specifically, if something simple (such as a character
3231        or a simple class like \d) has an unlimited quantifier, and is followed by
3232        something that cannot possibly match the quantified thing, the quantifier
3233        is automatically "possessified".
3234    
3235    24. A recursive reference to a subpattern whose number was greater than 39
3236        went wrong under certain circumstances in UTF-8 mode. This bug could also
3237        have affected the operation of pcre_study().
3238    
3239    25. Realized that a little bit of performance could be had by replacing
3240        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3241    
3242    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3243    
3244    27. Possessive quantifiers such as a++ were previously implemented by turning
3245        them into atomic groups such as ($>a+). Now they have their own opcodes,
3246        which improves performance. This includes the automatically created ones
3247        from 23 above.
3248    
3249    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3250        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3251        the first matched character to be a colon. This applied both to named and
3252        numbered groups.
3253    
3254    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3255    
3256    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3257        building PCRE on Windows easier, so I have included it in the distribution.
3258    
3259    31. There is now a check in pcretest against a ridiculously large number being
3260        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3261        loop, the loop is abandoned.
3262    
3263    32. Forward references to subpatterns in conditions such as (?(2)...) where
3264        subpattern 2 is defined later cause pcre_compile() to search forwards in
3265        the pattern for the relevant set of parentheses. This search went wrong
3266        when there were unescaped parentheses in a character class, parentheses
3267        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3268    
3269    33. "Subroutine" calls and backreferences were previously restricted to
3270        referencing subpatterns earlier in the regex. This restriction has now
3271        been removed.
3272    
3273    34. Added a number of extra features that are going to be in Perl 5.10. On the
3274        whole, these are just syntactic alternatives for features that PCRE had
3275        previously implemented using the Python syntax or my own invention. The
3276        other formats are all retained for compatibility.
3277    
3278        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3279            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3280            also .NET compatible.
3281    
3282        (b) A recursion or subroutine call to a named group can now be defined as
3283            (?&name) as well as (?P>name).
3284    
3285        (c) A backreference to a named group can now be defined as \k<name> or
3286            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3287            5.10, are also .NET compatible.
3288    
3289        (d) A conditional reference to a named group can now use the syntax
3290            (?(<name>) or (?('name') as well as (?(name).
3291    
3292        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3293            groups (named and numbered) that are never evaluated inline, but can be
3294            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3295            is always false. There may be only one alternative in such a group.
3296    
3297        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3298            as the simple (?(R). The condition is true only if the most recent
3299            recursion is that of the given number or name. It does not search out
3300            through the entire recursion stack.
3301    
3302        (g) The escape \gN or \g{N} has been added, where N is a positive or
3303            negative number, specifying an absolute or relative reference.
3304    
3305    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3306        some "unreachable code" warnings.
3307    
3308    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3309        things, this adds five new scripts.
3310    
3311    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3312        There were also incompatibilities regarding the handling of \Q..\E inside
3313        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3314        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3315    
3316    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3317        matches an empty string, and forcibly breaks the loop. There were bugs in
3318        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3319        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3320        separate and independent bugs (that affected different cases) have been
3321        fixed.
3322    
3323    39. Refactored the code to abolish the use of different opcodes for small
3324        capturing bracket numbers. This is a tidy that I avoided doing when I
3325        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3326        The new approach is not only tidier, it makes it possible to reduce the
3327        memory needed to fix the previous bug (38).
3328    
3329    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3330        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3331        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3332        mode.
3333    
3334    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3335        report.
3336    
3337    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3338        copy construction and assignment in the C++ wrapper.
3339    
3340    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3341        couple of bugs that could have given wrong results in the "--newline=crlf"
3342        case.
3343    
3344    44. Added a number of casts and did some reorganization of signed/unsigned int
3345        variables following suggestions from Dair Grant. Also renamed the variable
3346        "this" as "item" because it is a C++ keyword.
3347    
3348    45. Arranged for dftables to add
3349    
3350          #include "pcre_internal.h"
3351    
3352        to pcre_chartables.c because without it, gcc 4.x may remove the array
3353        definition from the final binary if PCRE is built into a static library and
3354        dead code stripping is activated.
3355    
3356    46. For an unanchored pattern, if a match attempt fails at the start of a
3357        newline sequence, and the newline setting is CRLF or ANY, and the next two
3358        characters are CRLF, advance by two characters instead of one.
3359    
3360    
3361    Version 6.7 04-Jul-06
3362    ---------------------
3363    
3364     1. In order to handle tests when input lines are enormously long, pcretest has
3365        been re-factored so that it automatically extends its buffers when
3366        necessary. The code is crude, but this _is_ just a test program. The
3367        default size has been increased from 32K to 50K.
3368    
3369     2. The code in pcre_study() was using the value of the re argument before
3370        testing it for NULL. (Of course, in any sensible call of the function, it
3371        won't be NULL.)
3372    
3373     3. The memmove() emulation function in pcre_internal.h, which is used on
3374        systems that lack both memmove() and bcopy() - that is, hardly ever -
3375        was missing a "static" storage class specifier.
3376    
3377     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3378        containing an extended class (one that cannot be represented by a bitmap
3379        because it contains high-valued characters or Unicode property items, e.g.
3380        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3381        pattern, but PCRE should not loop if you do not (it no longer does).
3382        [Detail: two cases were found: (a) a repeated subpattern containing an
3383        extended class; (b) a recursive reference to a subpattern that followed a
3384        previous extended class. It wasn't skipping over the extended class
3385        correctly when UTF-8 mode was not set.]
3386    
3387     5. A negated single-character class was not being recognized as fixed-length
3388        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3389        compile error "lookbehind assertion is not fixed length".
3390    
3391     6. The RunPerlTest auxiliary script was showing an unexpected difference
3392        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3393        write a Perl script that can interpret lines of an input file either as
3394        byte characters or as UTF-8, which is what "perltest" was being required to
3395        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3396        can't do is switch easily at run time between having the "use utf8;" pragma
3397        or not. In the end, I fudged it by using the RunPerlTest script to insert
3398        "use utf8;" explicitly for the UTF-8 tests.
3399    
3400     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3401        the end of the subject string, contrary to the documentation and to what
3402        Perl does. This was true of both matching functions. Now it matches only at
3403        the start of the subject and immediately after *internal* newlines.
3404    
3405     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3406        a pointer to an int instead of a pointer to an unsigned long int. This
3407        caused problems on 64-bit systems.
3408    
3409     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3410        instance of the 'standard' template library not being so standard".
3411    
3412    10. There was no check on the number of named subpatterns nor the maximum
3413        length of a subpattern name. The product of these values is used to compute
3414        the size of the memory block for a compiled pattern. By supplying a very
3415        long subpattern name and a large number of named subpatterns, the size
3416        computation could be caused to overflow. This is now prevented by limiting
3417        the length of names to 32 characters, and the number of named subpatterns
3418        to 10,000.
3419    
3420    11. Subpatterns that are repeated with specific counts have to be replicated in
3421        the compiled pattern. The size of memory for this was computed from the
3422        length of the subpattern and the repeat count. The latter is limited to
3423        65535, but there was no limit on the former, meaning that integer overflow
3424        could in principle occur. The compiled length of a repeated subpattern is
3425        now limited to 30,000 bytes in order to prevent this.
3426    
3427    12. Added the optional facility to have named substrings with the same name.
3428    
3429    13. Added the ability to use a named substring as a condition, using the
3430        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3431        are numbers (not recommended). Forward references are permitted.
3432    
3433    14. Added forward references in named backreferences (if you see what I mean).
3434    
3435    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3436        pattern could run off the end of the subject. For example, the pattern
3437        "(?s)(.{1,5})"8 did this with the subject "ab".
3438    
3439    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3440        PCRE_CASELESS was set when matching characters that were quantified with ?
3441        or *.
3442    
3443    17. A character class other than a single negated character that had a minimum
3444        but no maximum quantifier - for example [ab]{6,} - was not handled
3445        correctly by pce_dfa_exec(). It would match only one character.
3446    
3447    18. A valid (though odd) pattern that looked like a POSIX character
3448        class but used an invalid character after [ (for example [[,abc,]]) caused
3449        pcre_compile() to give the error "Failed: internal error: code overflow" or
3450        in some cases to crash with a glibc free() error. This could even happen if
3451        the pattern terminated after [[ but there just happened to be a sequence of
3452        letters, a binary zero, and a closing ] in the memory that followed.
3453    
3454    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3455        over the years. Originally (before any Unicode support), just the bottom 8
3456        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3457        output from "man perlunicode" includes this:
3458    
3459          The regular expression compiler produces polymorphic opcodes.  That
3460          is, the pattern adapts to the data and automatically switches to
3461          the Unicode character scheme when presented with Unicode data--or
3462          instead uses a traditional byte scheme when presented with byte
3463          data.
3464    
3465        Sadly, a wide octal escape does not cause a switch, and in a string with
3466        no other multibyte characters, these octal escapes are treated as before.
3467        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3468        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3469        Unicode string.
3470    
3471        I have not perpetrated such confusion in PCRE. Up till now, it took just
3472        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3473        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3474        translate to the appropriate multibyte character.
3475    
3476    29. Applied some refactoring to reduce the number of warnings from Microsoft
3477        and Borland compilers. This has included removing the fudge introduced
3478        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3479        a warning about an unused variable.
3480    
3481    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3482        characters since release 4.0, because Perl (from release 5.004) does not.
3483        [Or at least, is documented not to: some releases seem to be in conflict
3484        with the documentation.] However, when a pattern was studied with
3485        pcre_study() and all its branches started with \s, PCRE still included VT
3486        as a possible starting character. Of course, this did no harm; it just
3487        caused an unnecessary match attempt.
3488    
3489    22. Removed a now-redundant internal flag bit that recorded the fact that case
3490        dependency changed within the pattern. This was once needed for "required
3491        byte" processing, but is no longer used. This recovers a now-scarce options
3492        bit. Also moved the least significant internal flag bit to the most-
3493        significant bit of the word, which was not previously used (hangover from
3494        the days when it was an int rather than a uint) to free up another bit for
3495        the future.
3496    
3497    23. Added support for CRLF line endings as well as CR and LF. As well as the
3498        default being selectable at build time, it can now be changed at runtime
3499        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3500        specify that it is scanning data with non-default line endings.
3501    
3502    24. Changed the definition of CXXLINK to make it agree with the definition of
3503        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3504    
3505    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3506        recursions. This makes a big different to stack usage for some patterns.
3507    
3508    26. If a subpattern containing a named recursion or subroutine reference such
3509        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3510        the space required for the compiled pattern went wrong and gave too small a
3511        value. Depending on the environment, this could lead to "Failed: internal
3512        error: code overflow at offset 49" or "glibc detected double free or
3513        corruption" errors.
3514    
3515    27. Applied patches from Google (a) to support the new newline modes and (b) to
3516        advance over multibyte UTF-8 characters in GlobalReplace.
3517    
3518    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3519        difference for some implementation of PCRE in some Windows version.
3520    
3521    29. Added some extra testing facilities to pcretest:
3522    
3523        \q<number>   in a data line sets the "match limit" value
3524        \Q<number>   in a data line sets the "match recursion limt" value
3525        -S <number>  sets the stack size, where <number> is in megabytes
3526    
3527        The -S option isn't available for Windows.
3528    
3529    
3530    Version 6.6 06-Feb-06
3531    ---------------------
3532    
3533     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3534        in pcreposix.h. I have copied the definition from pcre.h.
3535    
3536     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3537        because pcre.h is no longer a built file.
3538    
3539     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3540        not normally included in the compiled code.
3541    
3542    
3543    Version 6.5 01-Feb-06
3544    ---------------------
3545    
3546     1. When using the partial match feature with pcre_dfa_exec(), it was not
3547        anchoring the second and subsequent partial matches at the new starting
3548        point. This could lead to incorrect results. For example, with the pattern
3549        /1234/, partially matching against "123" and then "a4" gave a match.
3550    
3551     2. Changes to pcregrep:
3552    
3553        (a) All non-match returns from pcre_exec() were being treated as failures
3554            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3555            error message is output. Some extra information is given for the
3556            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3557            probably the only errors that are likely to be caused by users (by
3558            specifying a regex that has nested indefinite repeats, for instance).
3559            If there are more than 20 of these errors, pcregrep is abandoned.
3560    
3561        (b) A binary zero was treated as data while matching, but terminated the
3562            output line if it was written out. This has been fixed: binary zeroes
3563            are now no different to any other data bytes.
3564    
3565        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3566            used to set a locale for matching. The --locale=xxxx long option has
3567            been added (no short equivalent) to specify a locale explicitly on the
3568            pcregrep command, overriding the environment variables.
3569    
3570        (d) When -B was used with -n, some line numbers in the output were one less
3571            than they should have been.
3572    
3573        (e) Added the -o (--only-matching) option.
3574    
3575        (f) If -A or -C was used with -c (count only), some lines of context were
3576            accidentally printed for the final match.
3577    
3578        (g) Added the -H (--with-filename) option.
3579    
3580        (h) The combination of options -rh failed to suppress file names for files
3581            that were found from directory arguments.
3582    
3583        (i) Added the -D (--devices) and -d (--directories) options.
3584    
3585        (j) Added the -F (--fixed-strings) option.
3586    
3587        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3588    
3589        (l) Added the --colo(u)r option.
3590    
3591        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3592            is not present by default.
3593    
3594     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3595        items such as (?R) or (?1), when the recursion could match a number of
3596        alternatives. If it matched one of the alternatives, but subsequently,
3597        outside the recursion, there was a failure, the code tried to back up into
3598        the recursion. However, because of the way PCRE is implemented, this is not
3599        possible, and the result was an incorrect result from the match.
3600    
3601        In order to prevent this happening, the specification of recursion has
3602        been changed so that all such subpatterns are automatically treated as
3603        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3604    
3605     4. I had overlooked the fact that, in some locales, there are characters for
3606        which isalpha() is true but neither isupper() nor islower() are true. In
3607        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3608        and ordfeminine) are like this. This affected the treatment of \w and \W
3609        when they appeared in character classes, but not when they appeared outside
3610        a character class. The bit map for "word" characters is now created
3611        separately from the results of isalnum() instead of just taking it from the
3612        upper, lower, and digit maps. (Plus the underscore character, of course.)
3613    
3614     5. The above bug also affected the handling of POSIX character classes such as
3615        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3616        permanent tables. Instead, the bit maps for such a class were previously
3617        created as the appropriate unions of the upper, lower, and digit bitmaps.
3618        Now they are created by subtraction from the [[:word:]] class, which has
3619        its own bitmap.
3620    
3621     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3622        It is created by subtracting the vertical space characters (\x09, \x0a,
3623        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3624        subtraction was done in the overall bitmap for a character class, meaning
3625        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3626        be recognized. This bug has been fixed.
3627    
3628     7. Patches from the folks at Google:
3629    
3630          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3631          real life, but is still worth protecting against".
3632    
3633          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3634          regular expressions".
3635    
3636          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3637          have it.
3638    
3639          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3640          "configure" and the latter not, in order to fix a problem somebody had
3641          with compiling the Arg class on HP-UX.
3642    
3643          (e) Improve the error-handling of the C++ wrapper a little bit.
3644    
3645          (f) New tests for checking recursion limiting.
3646    
3647     8. The pcre_memmove() function, which is used only if the environment does not
3648        have a standard memmove() function (and is therefore rarely compiled),
3649        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3650        returning a result (though PCRE never actually uses the result).
3651    
3652     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3653        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3654        returned instead of calling malloc() with an overflowing number that would
3655        most likely cause subsequent chaos.
3656    
3657    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3658    
3659    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3660        with this option is matched, the nmatch and pmatch options of regexec() are
3661        ignored.
3662    
3663    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3664        provided in case anyone wants to the the POSIX interface with UTF-8
3665        strings.
3666    
3667    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3668        C++ linking (needed for some HP-UX environments).
3669    
3670    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3671        (unused parameter) and in the pcre_printint() function (omitted "default"
3672        switch label when the default is to do nothing).
3673    
3674    15. Added some code to make it possible, when PCRE is compiled as a C++
3675        library, to replace subject pointers for pcre_exec() with a smart pointer
3676        class, thus making it possible to process discontinuous strings.
3677    
3678    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3679        much the same function. They were added by different people who were trying
3680        to make PCRE easy to compile on non-Unix systems. It has been suggested
3681        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3682        for compiling on Windows systems. I have therefore replaced it with
3683        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3684        defaults to "extern" for C or "extern C" for C++, which works fine on
3685        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3686        SCOPE with something explicit in config.h. In addition:
3687    
3688        (a) pcreposix.h still had just "extern" instead of either of these macros;
3689            I have replaced it with PCRE_DATA_SCOPE.
3690    
3691        (b) Functions such as _pcre_xclass(), which are internal to the library,
3692            but external in the C sense, all had PCRE_EXPORT in their definitions.
3693            This is apparently wrong for the Windows case, so I have removed it.
3694            (It makes no difference on Unix-like systems.)
3695    
3696    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3697        of recursive calls to match(). This is different to MATCH_LIMIT because
3698        that limits the total number of calls to match(), not all of which increase
3699        the depth of recursion. Limiting the recursion depth limits the amount of
3700        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3701        when PCRE is compiled, and changed at run time. A patch from Google adds
3702        this functionality to the C++ interface.
3703    
3704    18. Changes to the handling of Unicode character properties:
3705    
3706        (a) Updated the table to Unicode 4.1.0.
3707    
3708        (b) Recognize characters that are not in the table as "Cn" (undefined).
3709    
3710        (c) I revised the way the table is implemented to a much improved format
3711            which includes recognition of ranges. It now supports the ranges that
3712            are defined in UnicodeData.txt, and it also amalgamates other
3713            characters into ranges. This has reduced the number of entries in the
3714            table from around 16,000 to around 3,000, thus reducing its size
3715            considerably. I realized I did not need to use a tree structure after
3716            all - a binary chop search is just as efficient. Having reduced the
3717            number of entries, I extended their size from 6 bytes to 8 bytes to
3718            allow for more data.
3719    
3720        (d) Added support for Unicode script names via properties such as \p{Han}.
3721    
3722    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3723        matching that character.
3724    
3725    20. When matching a repeated Unicode property with a minimum greater than zero,
3726        (for example \pL{2,}), PCRE could look past the end of the subject if it
3727        reached it while seeking the minimum number of characters. This could
3728        happen only if some of the characters were more than one byte long, because
3729        there is a check for at least the minimum number of bytes.
3730    
3731    21. Refactored the implementation of \p and \P so as to be more general, to
3732        allow for more different types of property in future. This has changed the
3733        compiled form incompatibly. Anybody with saved compiled patterns that use
3734        \p or \P will have to recompile them.
3735    
3736    22. Added "Any" and "L&" to the supported property types.
3737    
3738    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3739        but give a compile time error if the value is greater than 0xff.
3740    
3741    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3742        accidentally not being installed or uninstalled.
3743    
3744    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3745        made were to insert the current release number. This seemed silly, because
3746        it made things harder for people building PCRE on systems that don't run
3747        "configure". I have turned pcre.h into a distributed file, no longer built
3748        by "configure", with the version identification directly included. There is
3749        no longer a pcre.h.in file.
3750    
3751        However, this change necessitated a change to the pcre-config script as
3752        well. It is built from pcre-config.in, and one of the substitutions was the
3753        release number. I have updated configure.ac so that ./configure now finds
3754        the release number by grepping pcre.h.
3755    
3756    26. Added the ability to run the tests under valgrind.
3757    
3758    
3759    Version 6.4 05-Sep-05
3760    ---------------------
3761    
3762     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3763        "--" to be printed when multiple files were scanned, even when none of the
3764        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3765        consider it to be a bug, and have restored the previous behaviour.
3766    
3767     2. A couple of code tidies to get rid of compiler warnings.
3768    
3769     3. The pcretest program used to cheat by referring to symbols in the library
3770        whose names begin with _pcre_. These are internal symbols that are not
3771        really supposed to be visible externally, and in some environments it is
3772        possible to suppress them. The cheating is now confined to including
3773        certain files from the library's source, which is a bit cleaner.
3774    
3775     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3776        file's purpose clearer.
3777    
3778     5. Reorganized pcre_ucp_findchar().
3779    
3780    
3781    Version 6.3 15-Aug-05
3782    ---------------------
3783    
3784     1. The file libpcre.pc.in did not have general read permission in the tarball.
3785    
3786     2. There were some problems when building without C++ support:
3787    
3788        (a) If C++ support was not built, "make install" and "make test" still
3789