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

Diff of /code/trunk/ChangeLog

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

revision 65 by nigel, Sat Feb 24 21:40:08 2007 UTC revision 1498 by ph10, Fri Aug 8 15:22:51 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.36 xx-xxx-2014
5    ------------------------
6    
7    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
8        -Wmissing-field-initializers and -Wunused-parameter.
9    
10    2.  The tests for quantifiers being too big (greater than 65535) were being
11        applied after reading the number, and stupidly assuming that integer
12        overflow would give a negative number. The tests are now applied as the
13        numbers are read.
14    
15    3.  Tidy code in pcre_exec.c where two branches that used to be different are
16        now the same.
17    
18    4.  The JIT compiler did not generate match limit checks for certain
19        bracketed expressions with quantifiers. This may lead to exponential
20        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
21        issue should be resolved now.
22    
23    5.  Fixed an issue, which occures when nested alternatives are optimized
24        with table jumps.
25    
26    6.  Inserted two casts and changed some ints to size_t in the light of some
27        reported 64-bit compiler warnings (Bugzilla 1477).
28    
29    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
30        an empty string, which sometimes were behaving incorrectly in the
31        interpreter (though correctly in the JIT matcher). This pcretest input is
32        an example:
33    
34          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
35          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
36    
37        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
38        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
39        for an empty string was breaking the inner loop and carrying on at a lower
40        level, when possessive repeated groups should always return to a higher
41        level as they have no backtrack points in them. The empty string test now
42        occurs at the outer level.
43    
44    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
45        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
46    
47    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
48        doesn't).
49    
50    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
51        to hold a pattern to be incorrectly computed (too small) when there were
52        named back references to duplicated names. This could cause "internal
53        error: code overflow" or "double free or corruption" or other memory
54        handling errors.
55    
56    11. When named subpatterns had the same prefixes, back references could be
57        confused. For example, in this pattern:
58    
59          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
60    
61        the reference to 'Name' was incorrectly treated as a reference to a
62        duplicate name.
63    
64    12. A pattern such as /^s?c/mi8 where the optional character has more than
65        one "other case" was incorrectly compiled such that it would only try to
66        match starting at "c".
67    
68    13. When a pattern starting with \s was studied, VT was not included in the
69        list of possible starting characters; this should have been part of the
70        8.34/18 patch.
71    
72    14. If a character class started [\Qx]... where x is any character, the class
73        was incorrectly terminated at the ].
74    
75    15. If a pattern that started with a caseless match for a character with more
76        than one "other case" was studied, PCRE did not set up the starting code
77        unit bit map for the list of possible characters. Now it does. This is an
78        optimization improvement, not a bug fix.
79    
80    16. The Unicode data tables have been updated to Unicode 7.0.0.
81    
82    17. Fixed a number of memory leaks in pcregrep.
83    
84    18. Avoid a compiler warning (from some compilers) for a function call with
85        a cast that removes "const" from an lvalue by using an intermediate
86        variable (to which the compiler does not object).
87    
88    19. Incorrect code was compiled if a group that contained an internal recursive
89        back reference was optional (had quantifier with a minimum of zero). This
90        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
91        caused segmentation faults because of stack overflows at compile time.
92    
93    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
94        group that is quantified with an indefinite repeat, caused a compile-time
95        loop which used up all the system stack and provoked a segmentation fault.
96        This was not the same bug as 19 above.
97    
98    
99    Version 8.35 04-April-2014
100    --------------------------
101    
102    1.  A new flag is set, when property checks are present in an XCLASS.
103        When this flag is not set, PCRE can perform certain optimizations
104        such as studying these XCLASS-es.
105    
106    2.  The auto-possessification of character sets were improved: a normal
107        and an extended character set can be compared now. Furthermore
108        the JIT compiler optimizes more character set checks.
109    
110    3.  Got rid of some compiler warnings for potentially uninitialized variables
111        that show up only when compiled with -O2.
112    
113    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
114        of a match later then the end of the match. The pcretest program was not
115        handling the case sensibly - it was outputting from the start to the next
116        binary zero. It now reports this situation in a message, and outputs the
117        text from the end to the start.
118    
119    5.  Fast forward search is improved in JIT. Instead of the first three
120        characters, any three characters with fixed position can be searched.
121        Search order: first, last, middle.
122    
123    6.  Improve character range checks in JIT. Characters are read by an inprecise
124        function now, which returns with an unknown value if the character code is
125        above a certain threshold (e.g: 256). The only limitation is that the value
126        must be bigger than the threshold as well. This function is useful when
127        the characters above the threshold are handled in the same way.
128    
129    7.  The macros whose names start with RAWUCHAR are placeholders for a future
130        mode in which only the bottom 21 bits of 32-bit data items are used. To
131        make this more memorable for those maintaining the code, the names have
132        been changed to start with UCHAR21, and an extensive comment has been added
133        to their definition.
134    
135    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
136        to the export list in Makefile.am (they were accidentally omitted from the
137        8.34 tarball).
138    
139    9.  The informational output from pcretest used the phrase "starting byte set"
140        which is inappropriate for the 16-bit and 32-bit libraries. As the output
141        for "first char" and "need char" really means "non-UTF-char", I've changed
142        "byte" to "char", and slightly reworded the output. The documentation about
143        these values has also been (I hope) clarified.
144    
145    10. Another JIT related optimization: use table jumps for selecting the correct
146        backtracking path, when more than four alternatives are present inside a
147        bracket.
148    
149    11. Empty match is not possible, when the minimum length is greater than zero,
150        and there is no \K in the pattern. JIT should avoid empty match checks in
151        such cases.
152    
153    12. In a caseless character class with UCP support, when a character with more
154        than one alternative case was not the first character of a range, not all
155        the alternative cases were added to the class. For example, s and \x{17f}
156        are both alternative cases for S: the class [RST] was handled correctly,
157        but [R-T] was not.
158    
159    13. The configure.ac file always checked for pthread support when JIT was
160        enabled. This is not used in Windows, so I have put this test inside a
161        check for the presence of windows.h (which was already tested for).
162    
163    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
164        The algorithm provides a way to skip certain starting offsets, and usually
165        faster than linear prefix searches.
166    
167    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
168        as for 'fr_FR' and 'french'. For some reason, however, it then used the
169        Windows-specific input and output files, which have 'french' screwed in.
170        So this could never have worked. One of the problems with locales is that
171        they aren't always the same. I have now updated RunTest so that it checks
172        the output of the locale test (test 3) against three different output
173        files, and it allows the test to pass if any one of them matches. With luck
174        this should make the test pass on some versions of Solaris where it was
175        failing. Because of the uncertainty, the script did not used to stop if
176        test 3 failed; it now does. If further versions of a French locale ever
177        come to light, they can now easily be added.
178    
179    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
180        there was a message during ./configure, but it did not stop. This now
181        provokes an error. The invalid example in README has been corrected.
182        If a value less than the minimum is given, the minimum value has always
183        been used, but now a warning is given.
184    
185    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
186        was a bug in the test system, which is now fixed. Also, the list of various
187        configurations that are tested for each release did not have one with both
188        16/32 bits and --enable-bar-anycrlf. It now does.
189    
190    18. pcretest was missing "-C bsr" for displaying the \R default setting.
191    
192    19. Little endian PowerPC systems are supported now by the JIT compiler.
193    
194    20. The fast forward newline mechanism could enter to an infinite loop on
195        certain invalid UTF-8 input. Although we don't support these cases
196        this issue can be fixed by a performance optimization.
197    
198    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
199        not take account if existing stack usage. There is now a new global
200        variable called pcre_stack_guard that can be set to point to an external
201        function to check stack availability. It is called at the start of
202        processing every parenthesized group.
203    
204    22. A typo in the code meant that in ungreedy mode the max/min qualifier
205        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
206        not match "ab".
207    
208    23. When UTF was disabled, the JIT program reported some incorrect compile
209        errors. These messages are silenced now.
210    
211    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
212        compiler.
213    
214    25. Change all the temporary files used in RunGrepTest to be different to those
215        used by RunTest so that the tests can be run simultaneously, for example by
216        "make -j check".
217    
218    
219    Version 8.34 15-December-2013
220    -----------------------------
221    
222    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
223        executable memory. Patch inspired by Carsten Klein.
224    
225    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
226        this macro is never tested and has no effect, because the work to support
227        coverage involves only compiling and linking options and special targets in
228        the Makefile. The comment in config.h implied that defining the macro would
229        enable coverage support, which is totally false. There was also support for
230        setting this macro in the CMake files (my fault, I just copied it from
231        configure). SUPPORT_GCOV has now been removed.
232    
233    3.  Make a small performance improvement in strlen16() and strlen32() in
234        pcretest.
235    
236    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
237        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
238        these cases). There was also one in pcretest.c.
239    
240    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
241    
242    6.  In UTF mode, the code for checking whether a group could match an empty
243        string (which is used for indefinitely repeated groups to allow for
244        breaking an infinite loop) was broken when the group contained a repeated
245        negated single-character class with a character that occupied more than one
246        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
247        UTF-8 mode). The effect was undefined: the group might or might not be
248        deemed as matching an empty string, or the program might have crashed.
249    
250    7.  The code for checking whether a group could match an empty string was not
251        recognizing that \h, \H, \v, \V, and \R must match a character.
252    
253    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
254        an empty string. If it can, pcretest shows this in its information output.
255    
256    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
257        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
258        matched by pcre_exec() without using JIT:
259    
260        (a) If the rest of the pattern did not match after a maximal run of
261            grapheme clusters, the code for backing up to try with fewer of them
262            did not always back up over a full grapheme when characters that do not
263            have the modifier quality were involved, e.g. Hangul syllables.
264    
265        (b) If the match point in a subject started with modifier character, and
266            there was no match, the code could incorrectly back up beyond the match
267            point, and potentially beyond the first character in the subject,
268            leading to a segfault or an incorrect match result.
269    
270    10. A conditional group with an assertion condition could lead to PCRE
271        recording an incorrect first data item for a match if no other first data
272        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
273        first data item, and therefore matched "ca" after "c" instead of at the
274        start.
275    
276    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
277        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
278    
279    12. The source of pcregrep now includes z/OS-specific code so that it can be
280        compiled for z/OS as part of the special z/OS distribution.
281    
282    13. Added the -T and -TM options to pcretest.
283    
284    14. The code in pcre_compile.c for creating the table of named capturing groups
285        has been refactored. Instead of creating the table dynamically during the
286        actual compiling pass, the information is remembered during the pre-compile
287        pass (on the stack unless there are more than 20 named groups, in which
288        case malloc() is used) and the whole table is created before the actual
289        compile happens. This has simplified the code (it is now nearly 150 lines
290        shorter) and prepared the way for better handling of references to groups
291        with duplicate names.
292    
293    15. A back reference to a named subpattern when there is more than one of the
294        same name now checks them in the order in which they appear in the pattern.
295        The first one that is set is used for the reference. Previously only the
296        first one was inspected. This change makes PCRE more compatible with Perl.
297    
298    16. Unicode character properties were updated from Unicode 6.3.0.
299    
300    17. The compile-time code for auto-possessification has been refactored, based
301        on a patch by Zoltan Herczeg. It now happens after instead of during
302        compilation. The code is cleaner, and more cases are handled. The option
303        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
304        options in pcretest are provided to set it. It can also be set by
305        (*NO_AUTO_POSSESS) at the start of a pattern.
306    
307    18. The character VT has been added to the default ("C" locale) set of
308        characters that match \s and are generally treated as white space,
309        following this same change in Perl 5.18. There is now no difference between
310        "Perl space" and "POSIX space". Whether VT is treated as white space in
311        other locales depends on the locale.
312    
313    19. The code for checking named groups as conditions, either for being set or
314        for being recursed, has been refactored (this is related to 14 and 15
315        above). Processing unduplicated named groups should now be as fast at
316        numerical groups, and processing duplicated groups should be faster than
317        before.
318    
319    20. Two patches to the CMake build system, by Alexander Barkov:
320    
321          (1) Replace the "source" command by "." in CMakeLists.txt because
322              "source" is a bash-ism.
323    
324          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
325              without these the CMake build does not work on Solaris.
326    
327    21. Perl has changed its handling of \8 and \9. If there is no previously
328        encountered capturing group of those numbers, they are treated as the
329        literal characters 8 and 9 instead of a binary zero followed by the
330        literals. PCRE now does the same.
331    
332    22. Following Perl, added \o{} to specify codepoints in octal, making it
333        possible to specify values greater than 0777 and also making them
334        unambiguous.
335    
336    23. Perl now gives an error for missing closing braces after \x{... instead of
337        treating the string as literal. PCRE now does the same.
338    
339    24. RunTest used to grumble if an inappropriate test was selected explicitly,
340        but just skip it when running all tests. This make it awkward to run ranges
341        of tests when one of them was inappropriate. Now it just skips any
342        inappropriate tests, as it always did when running all tests.
343    
344    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
345        character types such as \d or \w, too many callouts were inserted, and the
346        data that they returned was rubbish.
347    
348    26. In UCP mode, \s was not matching two of the characters that Perl matches,
349        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
350        were matched by \h. The code has now been refactored so that the lists of
351        the horizontal and vertical whitespace characters used for \h and \v (which
352        are defined only in one place) are now also used for \s.
353    
354    27. Add JIT support for the 64 bit TileGX architecture.
355        Patch by Jiong Wang (Tilera Corporation).
356    
357    28. Possessive quantifiers for classes (both explicit and automatically
358        generated) now use special opcodes instead of wrapping in ONCE brackets.
359    
360    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
361        (because it's meaningless), this was not happening when PCRE_CASELESS was
362        set. Not wrong, but inefficient.
363    
364    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
365        properties for \w, \d, etc) is present in a test regex. Otherwise if the
366        test contains no characters greater than 255, Perl doesn't realise it
367        should be using Unicode semantics.
368    
369    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
370        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
371        does in Unicode mode.
372    
373    32. Added the "forbid" facility to pcretest so that putting tests into the
374        wrong test files can sometimes be quickly detected.
375    
376    33. There is now a limit (default 250) on the depth of nesting of parentheses.
377        This limit is imposed to control the amount of system stack used at compile
378        time. It can be changed at build time by --with-parens-nest-limit=xxx or
379        the equivalent in CMake.
380    
381    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
382        errors. Perl warns for these when in warning mode, but PCRE has no facility
383        for giving warnings.
384    
385    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
386        However, this was not working for (?!) because it is optimized to (*FAIL),
387        for which PCRE does not allow quantifiers. The optimization is now disabled
388        when a quantifier follows (?!). I can't see any use for this, but it makes
389        things uniform.
390    
391    36. Perl no longer allows group names to start with digits, so I have made this
392        change also in PCRE. It simplifies the code a bit.
393    
394    37. In extended mode, Perl ignores spaces before a + that indicates a
395        possessive quantifier. PCRE allowed a space before the quantifier, but not
396        before the possessive +. It now does.
397    
398    38. The use of \K (reset reported match start) within a repeated possessive
399        group such as (a\Kb)*+ was not working.
400    
401    40. Document that the same character tables must be used at compile time and
402        run time, and that the facility to pass tables to pcre_exec() and
403        pcre_dfa_exec() is for use only with saved/restored patterns.
404    
405    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
406        features for Builds with MSVC:
407    
408        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
409           builds). The .rc files can be used to set FileDescription and many other
410           attributes.
411    
412        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
413           This allows higher-level build scripts which want .pdb files to avoid
414           hard-coding the exact files needed."
415    
416    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
417        mean "start of word" and "end of word", respectively, as a transition aid.
418    
419    43. A minimizing repeat of a class containing codepoints greater than 255 in
420        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
421        compiled to use the heap for recursion.
422    
423    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
424        is configured.
425    
426    
427    Version 8.33 28-May-2013
428    ------------------------
429    
430    1.  Added 'U' to some constants that are compared to unsigned integers, to
431        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
432        variables that are added to signed variables, to ensure the result is
433        signed and can be negated.
434    
435    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
436        CMake config files.
437    
438    3.  Revise the creation of config.h.generic so that all boolean macros are
439        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
440        overriding via -D on the command line possible.
441    
442    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
443        to unsigned int is reported to make a quite noticeable speed difference in
444        a specific Windows environment. Testing on Linux did also appear to show
445        some benefit (and it is clearly not harmful). Also fixed the definition of
446        Xop which should be unsigned.
447    
448    5.  Related to (4), changing the definition of the intermediate variable cc
449        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
450        improvements.
451    
452    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
453        unnecessary spaces.
454    
455    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
456        and later.
457    
458    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
459    
460    9.  Optimizing fast_forward_start_bits in JIT.
461    
462    10. Adding support for callouts in JIT, and fixing some issues revealed
463        during this work. Namely:
464    
465        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
466    
467        (b) Minimum length was not checked before the matching is started.
468    
469    11. The value of capture_last that is passed to callouts was incorrect in some
470        cases when there was a capture on one path that was subsequently abandoned
471        after a backtrack. Also, the capture_last value is now reset after a
472        recursion, since all captures are also reset in this case.
473    
474    12. The interpreter no longer returns the "too many substrings" error in the
475        case when an overflowing capture is in a branch that is subsequently
476        abandoned after a backtrack.
477    
478    13. In the pathological case when an offset vector of size 2 is used, pcretest
479        now prints out the matched string after a yield of 0 or 1.
480    
481    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
482        Only supported by the JIT compiler at the moment.
483    
484    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
485    
486    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
487        offset of the starting point of the matching process, provided the offsets
488        vector is large enough.
489    
490    17. The \A escape now records a lookbehind value of 1, though its execution
491        does not actually inspect the previous character. This is to ensure that,
492        in partial multi-segment matching, at least one character from the old
493        segment is retained when a new segment is processed. Otherwise, if there
494        are no lookbehinds in the pattern, \A might match incorrectly at the start
495        of a new segment.
496    
497    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
498    
499    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
500        gives some modest performance improvement in 8-bit mode.
501    
502    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
503        be expressed in certain programming languages using Universal Character
504        Names.
505    
506    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
507        which points out that "non characters" are not "characters that may not
508        appear in Unicode strings" but rather "characters that are reserved for
509        internal use and have only local meaning".
510    
511    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
512        there was a conditional group that depended on an assertion, if the
513        assertion was false, the callout that immediately followed the alternation
514        in the condition was skipped when pcre_exec() was used for matching.
515    
516    23. Allow an explicit callout to be inserted before an assertion that is the
517        condition for a conditional group, for compatibility with automatic
518        callouts, which always insert a callout at this point.
519    
520    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
521        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
522    
523    25. (*PRUNE) is now supported by the JIT compiler.
524    
525    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
526    
527    27. Fix the case where there are two or more SKIPs with arguments that may be
528        ignored.
529    
530    28. (*SKIP) is now supported by the JIT compiler.
531    
532    29. (*THEN) is now supported by the JIT compiler.
533    
534    30. Update RunTest with additional test selector options.
535    
536    31. The way PCRE handles backtracking verbs has been changed in two ways.
537    
538        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
539        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
540        backtracking. In some cases this makes it more Perl-compatible, but Perl's
541        rather obscure rules do not always do the same thing.
542    
543        (2) Previously, backtracking verbs were confined within assertions. This is
544        no longer the case for positive assertions, except for (*ACCEPT). Again,
545        this sometimes improves Perl compatibility, and sometimes does not.
546    
547    32. A number of tests that were in test 2 because Perl did things differently
548        have been moved to test 1, because either Perl or PCRE has changed, and
549        these tests are now compatible.
550    
551    32. Backtracking control verbs are now handled in the same way in JIT and
552        interpreter.
553    
554    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
555        contained a forward subroutine reference caused a compile error.
556    
557    34. Auto-detect and optimize limited repetitions in JIT.
558    
559    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
560        blocking (*UTF) etc.
561    
562    36. In the interpreter, maximizing pattern repetitions for characters and
563        character types now use tail recursion, which reduces stack usage.
564    
565    37. The value of the max lookbehind was not correctly preserved if a compiled
566        and saved regex was reloaded on a host of different endianness.
567    
568    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
569        of the compiled pattern block, expand the flags field from 16 to 32 bits
570        because it was almost full.
571    
572    39. Try madvise first before posix_madvise.
573    
574    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
575        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
576        original change locked out all matches of empty strings. This has been
577        changed so that one match of an empty string per line is recognized.
578        Subsequent searches on the same line (for colouring or for --only-matching,
579        for example) do not recognize empty strings.
580    
581    41. Applied a user patch to fix a number of spelling mistakes in comments.
582    
583    42. Data lines longer than 65536 caused pcretest to crash.
584    
585    43. Clarified the data type for length and startoffset arguments for pcre_exec
586        and pcre_dfa_exec in the function-specific man pages, where they were
587        explicitly stated to be in bytes, never having been updated. I also added
588        some clarification to the pcreapi man page.
589    
590    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
591        a segmentation fault.
592    
593    
594    Version 8.32 30-November-2012
595    -----------------------------
596    
597    1.  Improved JIT compiler optimizations for first character search and single
598        character iterators.
599    
600    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
601        Patch by Daniel Richard G.
602    
603    3.  Single character iterator optimizations in the JIT compiler.
604    
605    4.  Improved JIT compiler optimizations for character ranges.
606    
607    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
608        Reported by Giuseppe D'Angelo.
609    
610    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
611        of a line, was being set incorrectly in cases where .* appeared inside
612        atomic brackets at the start of a pattern, or where there was a subsequent
613        *PRUNE or *SKIP.
614    
615    7.  Improved instruction cache flush for POWER/PowerPC.
616        Patch by Daniel Richard G.
617    
618    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
619        grep:
620    
621        (a) There is now no limit to the number of patterns to be matched.
622    
623        (b) An error is given if a pattern is too long.
624    
625        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
626            are now supported.
627    
628        (d) --exclude-from and --include-from (multiple use) have been added.
629    
630        (e) Exclusions and inclusions now apply to all files and directories, not
631            just to those obtained from scanning a directory recursively.
632    
633        (f) Multiple uses of -f and --file-list are now supported.
634    
635        (g) In a Windows environment, the default for -d has been changed from
636            "read" (the GNU grep default) to "skip", because otherwise the presence
637            of a directory in the file list provokes an error.
638    
639        (h) The documentation has been revised and clarified in places.
640    
641    9.  Improve the matching speed of capturing brackets.
642    
643    10. Changed the meaning of \X so that it now matches a Unicode extended
644        grapheme cluster.
645    
646    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
647        out POSIX threads when JIT support is configured.
648    
649    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
650    
651    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
652        structure could go wrong in environments where size_t is not the same size
653        as int.
654    
655    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
656        set.
657    
658    15. The EBCDIC support had decayed; later updates to the code had included
659        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
660        general tidy up of EBCDIC-related issues, and the documentation was also
661        not quite right. There is now a test that can be run on ASCII systems to
662        check some of the EBCDIC-related things (but is it not a full test).
663    
664    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
665        in a small tidy to the code.
666    
667    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
668    
669    18. If the --only-matching (-o) option in pcregrep is specified multiple
670        times, each one causes appropriate output. For example, -o1 -o2 outputs the
671        substrings matched by the 1st and 2nd capturing parentheses. A separating
672        string can be specified by --om-separator (default empty).
673    
674    19. Improving the first n character searches.
675    
676    20. Turn case lists for horizontal and vertical white space into macros so that
677        they are defined only once.
678    
679    21. This set of changes together give more compatible Unicode case-folding
680        behaviour for characters that have more than one other case when UCP
681        support is available.
682    
683        (a) The Unicode property table now has offsets into a new table of sets of
684            three or more characters that are case-equivalent. The MultiStage2.py
685            script that generates these tables (the pcre_ucd.c file) now scans
686            CaseFolding.txt instead of UnicodeData.txt for character case
687            information.
688    
689        (b) The code for adding characters or ranges of characters to a character
690            class has been abstracted into a generalized function that also handles
691            case-independence. In UTF-mode with UCP support, this uses the new data
692            to handle characters with more than one other case.
693    
694        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
695            whose other case is greater than 256 are now correctly matched
696            caselessly. Previously, the high codepoint matched the low one, but not
697            vice versa.
698    
699        (d) The processing of \h, \H, \v, and \ in character classes now makes use
700            of the new class addition function, using character lists defined as
701            macros alongside the case definitions of 20 above.
702    
703        (e) Caseless back references now work with characters that have more than
704            one other case.
705    
706        (f) General caseless matching of characters with more than one other case
707            is supported.
708    
709    22. Unicode character properties were updated from Unicode 6.2.0
710    
711    23. Improved CMake support under Windows. Patch by Daniel Richard G.
712    
713    24. Add support for 32-bit character strings, and UTF-32
714    
715    25. Major JIT compiler update (code refactoring and bugfixing).
716        Experimental Sparc 32 support is added.
717    
718    26. Applied a modified version of Daniel Richard G's patch to create
719        pcre.h.generic and config.h.generic by "make" instead of in the
720        PrepareRelease script.
721    
722    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
723        pcre_compile.c when checking for a zero character.
724    
725    28. Introducing a native interface for JIT. Through this interface, the compiled
726        machine code can be directly executed. The purpose of this interface is to
727        provide fast pattern matching, so several sanity checks are not performed.
728        However, feature tests are still performed. The new interface provides
729        1.4x speedup compared to the old one.
730    
731    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
732        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
733        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
734    
735    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
736        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
737        the "old" RFC 2279). Instead, it ended up passing a negative length to
738        pcre_exec().
739    
740    31. Add support for GCC's visibility feature to hide internal functions.
741    
742    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
743        "unknown -C option" after outputting 0 or 1.
744    
745    33. There is now support for generating a code coverage report for the test
746        suite in environments where gcc is the compiler and lcov is installed. This
747        is mainly for the benefit of the developers.
748    
749    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
750        unaddressable using valgrind annotations, allowing valgrind to detect
751        invalid memory accesses. This is mainly for the benefit of the developers.
752    
753    25. (*UTF) can now be used to start a pattern in any of the three libraries.
754    
755    26. Give configure error if --enable-cpp but no C++ compiler found.
756    
757    
758    Version 8.31 06-July-2012
759    -------------------------
760    
761    1.  Fixing a wrong JIT test case and some compiler warnings.
762    
763    2.  Removed a bashism from the RunTest script.
764    
765    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
766        to unsigned type, result still unsigned" that was given by an MS compiler
767        on encountering the code "-sizeof(xxx)".
768    
769    4.  Partial matching support is added to the JIT compiler.
770    
771    5.  Fixed several bugs concerned with partial matching of items that consist
772        of more than one character:
773    
774        (a) /^(..)\1/ did not partially match "aba" because checking references was
775            done on an "all or nothing" basis. This also applied to repeated
776            references.
777    
778        (b) \R did not give a hard partial match if \r was found at the end of the
779            subject.
780    
781        (c) \X did not give a hard partial match after matching one or more
782            characters at the end of the subject.
783    
784        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
785            a partial match for the string "\r".
786    
787        (e) When newline was set to CRLF, the metacharacter "." did not recognize
788            a partial match for a CR character at the end of the subject string.
789    
790    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
791        running pcretest, the text "(JIT)" added to the output whenever JIT is
792        actually used to run the match.
793    
794    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
795        or /S+[+] with a digit between 1 and 7.
796    
797    8.  OP_NOT now supports any UTF character not just single-byte ones.
798    
799    9.  (*MARK) control verb is now supported by the JIT compiler.
800    
801    10. The command "./RunTest list" lists the available tests without actually
802        running any of them. (Because I keep forgetting what they all are.)
803    
804    11. Add PCRE_INFO_MAXLOOKBEHIND.
805    
806    12. Applied a (slightly modified) user-supplied patch that improves performance
807        when the heap is used for recursion (compiled with --disable-stack-for-
808        recursion). Instead of malloc and free for each heap frame each time a
809        logical recursion happens, frames are retained on a chain and re-used where
810        possible. This sometimes gives as much as 30% improvement.
811    
812    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
813        call.
814    
815    14. As documented, (*COMMIT) is now confined to within a positive assertion.
816    
817    15. It is now possible to link pcretest with libedit as an alternative to
818        libreadline.
819    
820    16. (*COMMIT) control verb is now supported by the JIT compiler.
821    
822    17. The Unicode data tables have been updated to Unicode 6.1.0.
823    
824    18. Added --file-list option to pcregrep.
825    
826    19. Added binary file support to pcregrep, including the -a, --binary-files,
827        -I, and --text options.
828    
829    20. The madvise function is renamed for posix_madvise for QNX compatibility
830        reasons. Fixed by Giuseppe D'Angelo.
831    
832    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
833    
834    22. Changed the option for creating symbolic links for 16-bit man pages from
835        -s to -sf so that re-installing does not cause issues.
836    
837    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
838    
839    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
840        matches in certain environments (the workspace was not being correctly
841        retained). Also added to pcre_dfa_exec() a simple plausibility check on
842        some of the workspace data at the beginning of a restart.
843    
844    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
845        was not doing so when it should - probably a typo introduced by SVN 528
846        (change 8.10/14).
847    
848    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
849        \w+ when the character tables indicated that \x{c4} was a word character.
850        There were several related cases, all because the tests for doing a table
851        lookup were testing for characters less than 127 instead of 255.
852    
853    27. If a pattern contains capturing parentheses that are not used in a match,
854        their slots in the ovector are set to -1. For those that are higher than
855        any matched groups, this happens at the end of processing. In the case when
856        there were back references that the ovector was too small to contain
857        (causing temporary malloc'd memory to be used during matching), and the
858        highest capturing number was not used, memory off the end of the ovector
859        was incorrectly being set to -1. (It was using the size of the temporary
860        memory instead of the true size.)
861    
862    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
863        ovector size, it uses memory at the end of the block that it has got.
864    
865    29. Check for an overlong MARK name and give an error at compile time. The
866        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
867    
868    30. JIT compiler update.
869    
870    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
871        Rill for the patch.
872    
873    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
874    
875    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
876    
877    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
878        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
879        was enabled.
880    
881    35. Improve JIT code generation for greedy plus quantifier.
882    
883    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
884        1 to "aa" instead of to an empty string. The bug affected repeated groups
885        that could potentially match an empty string.
886    
887    37. Optimizing single character iterators in JIT.
888    
889    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
890        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
891        codepoints that are too big for the mode are faulted, and in a UTF mode,
892        disallowed codepoints are also faulted.
893    
894    39. If PCRE was compiled with UTF support, in three places in the DFA
895        matcher there was code that should only have been obeyed in UTF mode, but
896        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
897        processing when bytes with values greater than 127 were present. In 16-bit
898        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
899        both cases the values are those that cannot be the first data item in a UTF
900        character. The three items that might have provoked this were recursions,
901        possessively repeated groups, and atomic groups.
902    
903    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
904        and pcregrep, because some OS require shared objects to be explicitly
905        passed to ld, causing the link step to fail if they are not.
906    
907    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
908        mode, patterns that started with \h* or \R* might be incorrectly matched.
909    
910    
911    Version 8.30 04-February-2012
912    -----------------------------
913    
914    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
915        name is defined in ctype.h.
916    
917    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
918        only in quite long subpatterns.
919    
920    3.  Removed the function pcre_info(), which has been obsolete and deprecated
921        since it was replaced by pcre_fullinfo() in February 2000.
922    
923    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
924        match a (*MARK), and the match failed at the start of the subject, a
925        reference to memory before the start of the subject could occur. This bug
926        was introduced by fix 17 of release 8.21.
927    
928    5.  A reference to an unset group with zero minimum repetition was giving
929        totally wrong answers (in non-JavaScript-compatibility mode). For example,
930        /(another)?(\1?)test/ matched against "hello world test". This bug was
931        introduced in release 8.13.
932    
933    6.  Add support for 16-bit character strings (a large amount of work involving
934        many changes and refactorings).
935    
936    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
937        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
938        from a file.
939    
940    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
941        rounding is not applied in this particular case).
942    
943    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
944        if they appear, or are escaped, in patterns.
945    
946    10. Get rid of a number of -Wunused-but-set-variable warnings.
947    
948    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
949        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
950        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
951        also returns the mark "x". This bug applied to capturing parentheses,
952        non-capturing parentheses, and atomic parentheses. It also applied to some
953        assertions.
954    
955    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
956        information out of configure.ac instead of relying on pcre.h.generic, which
957        is not stored in the repository.
958    
959    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
960        -lreadline.
961    
962    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
963    
964    15. Applied Graycode's patch to put the top-level frame on the stack rather
965        than the heap when not using the stack for recursion. This gives a
966        performance improvement in many cases when recursion is not deep.
967    
968    16. Experimental code added to "pcretest -C" to output the stack frame size.
969    
970    
971    Version 8.21 12-Dec-2011
972    ------------------------
973    
974    1.  Updating the JIT compiler.
975    
976    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
977        are added as well.
978    
979    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
980        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
981        calling _pcre_jit_exec. Some extra comments are added.
982    
983    4.  (*MARK) settings inside atomic groups that do not contain any capturing
984        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
985        was introduced by change 18 for 8.20.
986    
987    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
988        ECMA-262 standard.
989    
990    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
991        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
992        This bug was probably introduced by change 9 of 8.13.
993    
994    7.  While fixing 6 above, I noticed that a number of other items were being
995        incorrectly rejected as "not fixed length". This arose partly because newer
996        opcodes had not been added to the fixed-length checking code. I have (a)
997        corrected the bug and added tests for these items, and (b) arranged for an
998        error to occur if an unknown opcode is encountered while checking for fixed
999        length instead of just assuming "not fixed length". The items that were
1000        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1001        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1002        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1003    
1004    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1005        being incorrectly compiled and would have given unpredicatble results.
1006    
1007    9.  A possessively repeated subpattern with minimum repeat count greater than
1008        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1009        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1010        the first (A) could occur when it should not.
1011    
1012    10. Add a cast and remove a redundant test from the code.
1013    
1014    11. JIT should use pcre_malloc/pcre_free for allocation.
1015    
1016    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1017        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1018        is anything other than /usr, -L is still shown).
1019    
1020    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1021    
1022    14. Perl does not support \N without a following name in a [] class; PCRE now
1023        also gives an error.
1024    
1025    15. If a forward reference was repeated with an upper limit of around 2000,
1026        it caused the error "internal error: overran compiling workspace". The
1027        maximum number of forward references (including repeats) was limited by the
1028        internal workspace, and dependent on the LINK_SIZE. The code has been
1029        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1030        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1031        of around 200,000 forward references. While doing this, I also speeded up
1032        the filling in of repeated forward references.
1033    
1034    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1035        incorrectly expecting the subject to contain another "a" after the start.
1036    
1037    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1038        in the match, the SKIP should be ignored. This was not happening; instead
1039        the SKIP was being treated as NOMATCH. For patterns such as
1040        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1041        tested.
1042    
1043    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1044        now much more compatible with Perl, in particular in cases where the result
1045        is a non-match for a non-anchored pattern. For example, if
1046        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1047        "m", where previously it did not return a name. A side effect of this
1048        change is that for partial matches, the last encountered mark name is
1049        returned, as for non matches. A number of tests that were previously not
1050        Perl-compatible have been moved into the Perl-compatible test files. The
1051        refactoring has had the pleasing side effect of removing one argument from
1052        the match() function, thus reducing its stack requirements.
1053    
1054    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1055        subsequent uses of /S (without +) incorrectly behaved like /S+.
1056    
1057    21. Retrieve executable code size support for the JIT compiler and fixing
1058        some warnings.
1059    
1060    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1061        not work when the shorter character appeared right at the end of the
1062        subject string.
1063    
1064    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1065        systems.
1066    
1067    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1068        output it when the /M option is used in pcretest.
1069    
1070    25. The CheckMan script was not being included in the distribution. Also, added
1071        an explicit "perl" to run Perl scripts from the PrepareRelease script
1072        because this is reportedly needed in Windows.
1073    
1074    26. If study data was being save in a file and studying had not found a set of
1075        "starts with" bytes for the pattern, the data written to the file (though
1076        never used) was taken from uninitialized memory and so caused valgrind to
1077        complain.
1078    
1079    27. Updated RunTest.bat as provided by Sheri Pierce.
1080    
1081    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1082    
1083    29. Computation of memory usage for the table of capturing group names was
1084        giving an unnecessarily large value.
1085    
1086    
1087    Version 8.20 21-Oct-2011
1088    ------------------------
1089    
1090    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1091        a POSIX class. After further experiments with Perl, which convinced me that
1092        Perl has bugs and confusions, a closing square bracket is no longer allowed
1093        in a POSIX name. This bug also affected patterns with classes that started
1094        with full stops.
1095    
1096    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1097        captured substring, but while checking the failing first alternative,
1098        substring 1 is temporarily captured. If the output vector supplied to
1099        pcre_exec() was not big enough for this capture, the yield of the function
1100        was still zero ("insufficient space for captured substrings"). This cannot
1101        be totally fixed without adding another stack variable, which seems a lot
1102        of expense for a edge case. However, I have improved the situation in cases
1103        such as /(a)(b)x|abc/ matched against "abc", where the return code
1104        indicates that fewer than the maximum number of slots in the ovector have
1105        been set.
1106    
1107    3.  Related to (2) above: when there are more back references in a pattern than
1108        slots in the output vector, pcre_exec() uses temporary memory during
1109        matching, and copies in the captures as far as possible afterwards. It was
1110        using the entire output vector, but this conflicts with the specification
1111        that only 2/3 is used for passing back captured substrings. Now it uses
1112        only the first 2/3, for compatibility. This is, of course, another edge
1113        case.
1114    
1115    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1116        main code base, and can be used by building with --enable-jit. When this is
1117        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1118        runtime --no-jit option is given.
1119    
1120    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1121        ovector, the return from the function was zero, implying that there were
1122        other matches that did not fit. The correct "exactly full" value is now
1123        returned.
1124    
1125    6.  If a subpattern that was called recursively or as a subroutine contained
1126        (*PRUNE) or any other control that caused it to give a non-standard return,
1127        invalid errors such as "Error -26 (nested recursion at the same subject
1128        position)" or even infinite loops could occur.
1129    
1130    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1131        computing the minimum length on reaching *ACCEPT, and so ended up with the
1132        wrong value of 1 rather than 0. Further investigation indicates that
1133        computing a minimum subject length in the presence of *ACCEPT is difficult
1134        (think back references, subroutine calls), and so I have changed the code
1135        so that no minimum is registered for a pattern that contains *ACCEPT.
1136    
1137    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1138        it was not handled as intended. [But see 16 below.]
1139    
1140    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1141        Sheri Pierce.
1142    
1143    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1144        the first byte in a match must be "a".
1145    
1146    11. Change 17 for 8.13 increased the recursion depth for patterns like
1147        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1148        pattern contains any instances of (*THEN). If it does not, the old
1149        optimizations are restored. It would be nice to do this on a per-group
1150        basis, but at the moment that is not feasible.
1151    
1152    12. In some environments, the output of pcretest -C is CRLF terminated. This
1153        broke RunTest's code that checks for the link size. A single white space
1154        character after the value is now allowed for.
1155    
1156    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1157        For "fr", it uses the Windows-specific input and output files.
1158    
1159    14. If (*THEN) appeared in a group that was called recursively or as a
1160        subroutine, it did not work as intended. [But see next item.]
1161    
1162    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1163        pattern fragments (but not containing any | characters). If A and B are
1164        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1165        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1166        D. In other words, Perl considers parentheses that do not contain any |
1167        characters to be part of a surrounding alternative, whereas PCRE was
1168        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1169        differently. PCRE now behaves in the same way as Perl, except in the case
1170        of subroutine/recursion calls such as (?1) which have in any case always
1171        been different (but PCRE had them first :-).
1172    
1173    16. Related to 15 above: Perl does not treat the | in a conditional group as
1174        creating alternatives. Such a group is treated in the same way as an
1175        ordinary group without any | characters when processing (*THEN). PCRE has
1176        been changed to match Perl's behaviour.
1177    
1178    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1179        RunGrepTest script failed.
1180    
1181    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1182        inevitable for groups that contain captures, but it can lead to a lot of
1183        stack use in large patterns. The old behaviour has been restored for atomic
1184        groups that do not contain any capturing parentheses.
1185    
1186    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1187        suppress the check for a minimum subject length at run time. (If it was
1188        given to pcre_exec() or pcre_dfa_exec() it did work.)
1189    
1190    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1191        fail to work when decoding hex characters in data strings in EBCDIC
1192        environments.
1193    
1194    21. It appears that in at least one Mac OS environment, the isxdigit() function
1195        is implemented as a macro that evaluates to its argument more than once,
1196        contravening the C 90 Standard (I haven't checked a later standard). There
1197        was an instance in pcretest which caused it to go wrong when processing
1198        \x{...} escapes in subject strings. The has been rewritten to avoid using
1199        things like p++ in the argument of isxdigit().
1200    
1201    
1202    Version 8.13 16-Aug-2011
1203    ------------------------
1204    
1205    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1206    
1207    2.  Two minor typos in pcre_internal.h have been fixed.
1208    
1209    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1210        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1211        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1212    
1213    4.  There were a number of related bugs in the code for matching backrefences
1214        caselessly in UTF-8 mode when codes for the characters concerned were
1215        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1216        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1217        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1218        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1219        2-byte code at the end of the subject (it thought there wasn't enough data
1220        left).
1221    
1222    5.  Comprehensive information about what went wrong is now returned by
1223        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1224        as the output vector has at least 2 elements. The offset of the start of
1225        the failing character and a reason code are placed in the vector.
1226    
1227    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1228        now returned is for the first byte of the failing character, instead of the
1229        last byte inspected. This is an incompatible change, but I hope it is small
1230        enough not to be a problem. It makes the returned offset consistent with
1231        pcre_exec() and pcre_dfa_exec().
1232    
1233    7.  pcretest now gives a text phrase as well as the error number when
1234        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1235        failure, the offset and reason code are output.
1236    
1237    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1238        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1239        back over a single character (\n). This seems wrong (because it treated the
1240        two characters as a single entity when going forwards), conflicts with the
1241        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1242        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1243        has been changed.
1244    
1245    9.  Some internal refactoring has changed the processing so that the handling
1246        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1247        time (the PCRE_DOTALL option was changed this way some time ago: version
1248        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1249        which was always a bit of a fudge. It also means that there is one less
1250        argument for the match() function, which reduces its stack requirements
1251        slightly. This change also fixes an incompatibility with Perl: the pattern
1252        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1253    
1254    10. More internal refactoring has drastically reduced the number of recursive
1255        calls to match() for possessively repeated groups such as (abc)++ when
1256        using pcre_exec().
1257    
1258    11. While implementing 10, a number of bugs in the handling of groups were
1259        discovered and fixed:
1260    
1261        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1262        (a|)*(?1) gave a compile-time internal error.
1263        ((a|)+)+  did not notice that the outer group could match an empty string.
1264        (^a|^)+   was not marked as anchored.
1265        (.*a|.*)+ was not marked as matching at start or after a newline.
1266    
1267    12. Yet more internal refactoring has removed another argument from the match()
1268        function. Special calls to this function are now indicated by setting a
1269        value in a variable in the "match data" data block.
1270    
1271    13. Be more explicit in pcre_study() instead of relying on "default" for
1272        opcodes that mean there is no starting character; this means that when new
1273        ones are added and accidentally left out of pcre_study(), testing should
1274        pick them up.
1275    
1276    14. The -s option of pcretest has been documented for ages as being an old
1277        synonym of -m (show memory usage). I have changed it to mean "force study
1278        for every regex", that is, assume /S for every regex. This is similar to -i
1279        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1280        using it. It makes it easier to run collections of tests with and without
1281        study enabled, and thereby test pcre_study() more easily. All the standard
1282        tests are now run with and without -s (but some patterns can be marked as
1283        "never study" - see 20 below).
1284    
1285    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1286        restoration of the capturing data to the outer values was not happening
1287        correctly.
1288    
1289    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1290        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1291        pattern had matched an empty string, and so incorrectly returned a no
1292        match.
1293    
1294    17. There was optimizing code for the last branch of non-capturing parentheses,
1295        and also for the obeyed branch of a conditional subexpression, which used
1296        tail recursion to cut down on stack usage. Unfortunately, now that there is
1297        the possibility of (*THEN) occurring in these branches, tail recursion is
1298        no longer possible because the return has to be checked for (*THEN). These
1299        two optimizations have therefore been removed. [But see 8.20/11 above.]
1300    
1301    18. If a pattern containing \R was studied, it was assumed that \R always
1302        matched two bytes, thus causing the minimum subject length to be
1303        incorrectly computed because \R can also match just one byte.
1304    
1305    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1306        was incorrectly computed.
1307    
1308    20. If /S is present twice on a test pattern in pcretest input, it now
1309        *disables* studying, thereby overriding the use of -s on the command line
1310        (see 14 above). This is necessary for one or two tests to keep the output
1311        identical in both cases.
1312    
1313    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1314        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1315    
1316    22. When an atomic group that contained a capturing parenthesis was
1317        successfully matched, but the branch in which it appeared failed, the
1318        capturing was not being forgotten if a higher numbered group was later
1319        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1320        group 1 to "a", when in fact it should be unset. This applied to multi-
1321        branched capturing and non-capturing groups, repeated or not, and also to
1322        positive assertions (capturing in negative assertions does not happen
1323        in PCRE) and also to nested atomic groups.
1324    
1325    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1326        subject after a captured substring, to make it easier to tell which of a
1327        number of identical substrings has been captured.
1328    
1329    24. The way atomic groups are processed by pcre_exec() has been changed so that
1330        if they are repeated, backtracking one repetition now resets captured
1331        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1332        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1333        "aaa". Previously, it would have been "a". As part of this code
1334        refactoring, the way recursive calls are handled has also been changed.
1335    
1336    25. If an assertion condition captured any substrings, they were not passed
1337        back unless some other capturing happened later. For example, if
1338        (?(?=(a))a) was matched against "a", no capturing was returned.
1339    
1340    26. When studying a pattern that contained subroutine calls or assertions,
1341        the code for finding the minimum length of a possible match was handling
1342        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1343        group 1 called group 2 while simultaneously a separate group 2 called group
1344        1). A stack overflow occurred in this case. I have fixed this by limiting
1345        the recursion depth to 10.
1346    
1347    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1348        Fortmann. This supports explicit test numbers on the command line, and has
1349        argument validation and error reporting.
1350    
1351    28. An instance of \X with an unlimited repeat could fail if at any point the
1352        first character it looked at was a mark character.
1353    
1354    29. Some minor code refactoring concerning Unicode properties and scripts
1355        should reduce the stack requirement of match() slightly.
1356    
1357    30. Added the '=' option to pcretest to check the setting of unused capturing
1358        slots at the end of the pattern, which are documented as being -1, but are
1359        not included in the return count.
1360    
1361    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1362        compiled something random. Now it gives a compile-time error (as does
1363        Perl).
1364    
1365    32. A *MARK encountered during the processing of a positive assertion is now
1366        recorded and passed back (compatible with Perl).
1367    
1368    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1369        had alternative anchored branches, the search for a second match in a line
1370        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1371        matched the line "0102" twice. The same bug affected patterns that started
1372        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1373        twice.
1374    
1375    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1376        does, and because of capturing effects, quantifying parenthesized
1377        assertions may at times be useful. Quantifiers are now allowed for
1378        parenthesized assertions.
1379    
1380    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1381    
1382    36. \g was being checked for fancy things in a character class, when it should
1383        just be a literal "g".
1384    
1385    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1386        appearance of a nested POSIX class supersedes an apparent external class.
1387        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1388        unescaped square brackets may also appear as part of class names. For
1389        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1390        more like Perl. (But see 8.20/1 above.)
1391    
1392    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1393        was because it thought it was \N{name}, which is not supported).
1394    
1395    39. Add minix to OS list not supporting the -S option in pcretest.
1396    
1397    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1398        cannot analyze patterns in sufficient detail to catch mutual recursions
1399        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1400        subgroup is called recursively as a subpattern for a second time at the
1401        same position in the subject string. In previous releases this might have
1402        been caught by the recursion limit, or it might have run out of stack.
1403    
1404    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1405        happen only once. PCRE was, however incorrectly giving a compile time error
1406        "recursive call could loop indefinitely" because it cannot analyze the
1407        pattern in sufficient detail. The compile time test no longer happens when
1408        PCRE is compiling a conditional subpattern, but actual runaway loops are
1409        now caught at runtime (see 40 above).
1410    
1411    42. It seems that Perl allows any characters other than a closing parenthesis
1412        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1413        has been changed to be the same.
1414    
1415    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1416        as not to get warnings when autogen.sh is called. Also changed
1417        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1418    
1419    44. To help people who use pcregrep to scan files containing exceedingly long
1420        lines, the following changes have been made:
1421    
1422        (a) The default value of the buffer size parameter has been increased from
1423            8K to 20K. (The actual buffer used is three times this size.)
1424    
1425        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1426            PCRE is built.
1427    
1428        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1429            to be set at run time.
1430    
1431        (d) Numerical values in pcregrep options can be followed by K or M, for
1432            example --buffer-size=50K.
1433    
1434        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1435            given and the return code is set to 2.
1436    
1437    45. Add a pointer to the latest mark to the callout data block.
1438    
1439    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1440        partial match of an empty string instead of no match. This was specific to
1441        the use of ".".
1442    
1443    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1444        complete match instead of a partial match. This bug was dependent on both
1445        the PCRE_UTF8 and PCRE_DOTALL options being set.
1446    
1447    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1448        starting byte set, because \b was not being ignored.
1449    
1450    
1451    Version 8.12 15-Jan-2011
1452    ------------------------
1453    
1454    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1455        checks for such things as part of the documentation building process.
1456    
1457    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1458        --match-limit and --recursion-limit options (added for 8.11). In
1459        particular, this made one of the standard tests fail. (The integer value
1460        went into the wrong half of a long int.)
1461    
1462    3.  If the --colour option was given to pcregrep with -v (invert match), it
1463        did strange things, either producing crazy output, or crashing. It should,
1464        of course, ignore a request for colour when reporting lines that do not
1465        match.
1466    
1467    4.  Another pcregrep bug caused similar problems if --colour was specified with
1468        -M (multiline) and the pattern match finished with a line ending.
1469    
1470    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1471        matched in multiline mode, the following line was shown as part of the
1472        match. This seems wrong, so I have changed it.
1473    
1474    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1475        the check for further matches in the same line (so they could be coloured)
1476        to overrun the end of the current line. If another match was found, it was
1477        incorrectly shown (and then shown again when found in the next line).
1478    
1479    7.  If pcregrep was compiled under Windows, there was a reference to the
1480        function pcregrep_exit() before it was defined. I am assuming this was
1481        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1482        reported by a user. I've moved the definition above the reference.
1483    
1484    
1485    Version 8.11 10-Dec-2010
1486    ------------------------
1487    
1488    1.  (*THEN) was not working properly if there were untried alternatives prior
1489        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1490        backtracked to try for "b" instead of moving to the next alternative branch
1491        at the same level (in this case, to look for "c"). The Perl documentation
1492        is clear that when (*THEN) is backtracked onto, it goes to the "next
1493        alternative in the innermost enclosing group".
1494    
1495    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1496        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1497        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1498        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1499        (*THEN).
1500    
1501    3.  If \s appeared in a character class, it removed the VT character from
1502        the class, even if it had been included by some previous item, for example
1503        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1504        of \s, but is part of the POSIX "space" class.)
1505    
1506    4.  A partial match never returns an empty string (because you can always
1507        match an empty string at the end of the subject); however the checking for
1508        an empty string was starting at the "start of match" point. This has been
1509        changed to the "earliest inspected character" point, because the returned
1510        data for a partial match starts at this character. This means that, for
1511        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1512        (previously it gave "no match").
1513    
1514    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1515        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1516        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1517        has an implication that the given string is incomplete (because a partial
1518        match is preferred over a full match). For this reason, these items now
1519        give a partial match in this situation. [Aside: previously, the one case
1520        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1521        match rather than a full match, which was wrong by the old rules, but is
1522        now correct.]
1523    
1524    6.  There was a bug in the handling of #-introduced comments, recognized when
1525        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1526        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1527        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1528        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1529        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1530        places in pcre_compile().
1531    
1532    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1533        comments when looking ahead for named forward references to subpatterns,
1534        the only newline sequence it recognized was NL. It now handles newlines
1535        according to the set newline convention.
1536    
1537    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1538        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1539        cater for a lack of strerror(). These oversights have been fixed.
1540    
1541    9.  Added --match-limit and --recursion-limit to pcregrep.
1542    
1543    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1544    
1545    11. When the -o option was used, pcregrep was setting a return code of 1, even
1546        when matches were found, and --line-buffered was not being honoured.
1547    
1548    12. Added an optional parentheses number to the -o and --only-matching options
1549        of pcregrep.
1550    
1551    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1552        can match an empty string. The code to do it in pcretest and pcredemo
1553        needed fixing:
1554    
1555        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1556            only one byte after an empty string match just before CRLF (this case
1557            just got forgotten; "any" and "anycrlf" were OK).
1558    
1559        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1560            mode when an empty string match preceded an ASCII character followed by
1561            a non-ASCII character. (The code for advancing by one character rather
1562            than one byte was nonsense.)
1563    
1564        (c) The pcredemo.c sample program did not have any code at all to handle
1565            the cases when CRLF is a valid newline sequence.
1566    
1567    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1568        as a starting offset was within the subject string. There is now a new
1569        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1570        negative or greater than the length of the string. In order to test this,
1571        pcretest is extended to allow the setting of negative starting offsets.
1572    
1573    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1574        starting offset points to the beginning of a UTF-8 character was
1575        unnecessarily clumsy. I tidied it up.
1576    
1577    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1578        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1579    
1580    17. Nobody had reported that the --include_dir option, which was added in
1581        release 7.7 should have been called --include-dir (hyphen, not underscore)
1582        for compatibility with GNU grep. I have changed it to --include-dir, but
1583        left --include_dir as an undocumented synonym, and the same for
1584        --exclude-dir, though that is not available in GNU grep, at least as of
1585        release 2.5.4.
1586    
1587    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1588        characters from a string of bytes) have been redefined so as not to use
1589        loops, in order to improve performance in some environments. At the same
1590        time, I abstracted some of the common code into auxiliary macros to save
1591        repetition (this should not affect the compiled code).
1592    
1593    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1594        compile-time error is now given if \c is not followed by an ASCII
1595        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1596        different, and any byte value is allowed.)
1597    
1598    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1599        START_OPTIMIZE option, which is now allowed at compile time - but just
1600        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1601        to pcregrep and other applications that have no direct access to PCRE
1602        options. The new /Y option in pcretest sets this option when calling
1603        pcre_compile().
1604    
1605    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1606        back references. Groups containing recursive back references were forced to
1607        be atomic by that change, but in the case of named groups, the amount of
1608        memory required was incorrectly computed, leading to "Failed: internal
1609        error: code overflow". This has been fixed.
1610    
1611    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1612        pcretest.c, to avoid build problems in some Borland environments.
1613    
1614    
1615    Version 8.10 25-Jun-2010
1616    ------------------------
1617    
1618    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1619        THEN.
1620    
1621    2.  (*ACCEPT) was not working when inside an atomic group.
1622    
1623    3.  Inside a character class, \B is treated as a literal by default, but
1624        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1625        causes the error). The code is unchanged, but I tidied the documentation.
1626    
1627    4.  Inside a character class, PCRE always treated \R and \X as literals,
1628        whereas Perl faults them if its -w option is set. I have changed PCRE so
1629        that it faults them when PCRE_EXTRA is set.
1630    
1631    5.  Added support for \N, which always matches any character other than
1632        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1633    
1634    6.  When compiling pcregrep with newer versions of gcc which may have
1635        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1636        declared with attribute warn_unused_result" were given. Just casting the
1637        result to (void) does not stop the warnings; a more elaborate fudge is
1638        needed. I've used a macro to implement this.
1639    
1640    7.  Minor change to pcretest.c to avoid a compiler warning.
1641    
1642    8.  Added four artifical Unicode properties to help with an option to make
1643        \s etc use properties (see next item). The new properties are: Xan
1644        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1645    
1646    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1647        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1648        this option. Modified pcretest to add /W to test this facility. Added
1649        REG_UCP to make it available via the POSIX interface.
1650    
1651    10. Added --line-buffered to pcregrep.
1652    
1653    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1654        studied, and the match started with a letter with a code point greater than
1655        127 whose first byte was different to the first byte of the other case of
1656        the letter, the other case of this starting letter was not recognized
1657        (#976).
1658    
1659    12. If a pattern that was studied started with a repeated Unicode property
1660        test, for example, \p{Nd}+, there was the theoretical possibility of
1661        setting up an incorrect bitmap of starting bytes, but fortunately it could
1662        not have actually happened in practice until change 8 above was made (it
1663        added property types that matched character-matching opcodes).
1664    
1665    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1666        possible starting bytes for non-anchored patterns.
1667    
1668    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1669        \R, and also a number of cases that involve Unicode properties, both
1670        explicit and implicit when PCRE_UCP is set.
1671    
1672    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1673        input, it could crash or give wrong results if characters with values
1674        greater than 0xc0 were present in the subject string. (Detail: it assumed
1675        UTF-8 input when processing these items.)
1676    
1677    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1678        size_t is 64-bit (#991).
1679    
1680    17. Added a check for running out of memory when PCRE is compiled with
1681        --disable-stack-for-recursion (#990).
1682    
1683    18. If the last data line in a file for pcretest does not have a newline on
1684        the end, a newline was missing in the output.
1685    
1686    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1687        less than 128) in its various bitmaps. However, there is a facility for
1688        generating tables according to the current locale when PCRE is compiled. It
1689        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1690        characters, are recognized by isspace() and therefore were getting set in
1691        these tables, and indeed these tables seem to approximate to ISO 8859. This
1692        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1693        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1694        which of course cannot start UTF-8 characters. I have changed the code so
1695        that only real ASCII characters (less than 128) and the correct starting
1696        bytes for UTF-8 encodings are set for characters greater than 127 when in
1697        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1698        altogether.)
1699    
1700    20. Added the /T option to pcretest so as to be able to run tests with non-
1701        standard character tables, thus making it possible to include the tests
1702        used for 19 above in the standard set of tests.
1703    
1704    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1705        reference to a subpattern the other side of a comment that contains an
1706        opening parenthesis caused either an internal compiling error, or a
1707        reference to the wrong subpattern.
1708    
1709    
1710    Version 8.02 19-Mar-2010
1711    ------------------------
1712    
1713    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1714    
1715    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1716        configured.
1717    
1718    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1719        original author of that file, following a query about its status.
1720    
1721    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1722        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1723    
1724    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1725        quantifier applied to a forward-referencing subroutine call, could compile
1726        incorrect code or give the error "internal error: previously-checked
1727        referenced subpattern not found".
1728    
1729    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1730        variables to point to external functions. For these systems, therefore,
1731        pcre_malloc etc. are now initialized to local functions that call the
1732        relevant global functions.
1733    
1734    7.  There were two entries missing in the vectors called coptable and poptable
1735        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1736        I've fixed the data, and added a kludgy way of testing at compile time that
1737        the lengths are correct (equal to the number of opcodes).
1738    
1739    8.  Following on from 7, I added a similar kludge to check the length of the
1740        eint vector in pcreposix.c.
1741    
1742    9.  Error texts for pcre_compile() are held as one long string to avoid too
1743        much relocation at load time. To find a text, the string is searched,
1744        counting zeros. There was no check for running off the end of the string,
1745        which could happen if a new error number was added without updating the
1746        string.
1747    
1748    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1749    
1750    11. \K was not working if it appeared in an atomic group or in a group that
1751        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1752        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1753        the assertion is positive, but not if it is negative.
1754    
1755    12. Change 11 fortuitously reduced the size of the stack frame used in the
1756        "match()" function of pcre_exec.c by one pointer. Forthcoming
1757        implementation of support for (*MARK) will need an extra pointer on the
1758        stack; I have reserved it now, so that the stack frame size does not
1759        decrease.
1760    
1761    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1762        item in branch that calls a recursion is a subroutine call - as in the
1763        second branch in the above example - was incorrectly given the compile-
1764        time error "recursive call could loop indefinitely" because pcre_compile()
1765        was not correctly checking the subroutine for matching a non-empty string.
1766    
1767    14. The checks for overrunning compiling workspace could trigger after an
1768        overrun had occurred. This is a "should never occur" error, but it can be
1769        triggered by pathological patterns such as hundreds of nested parentheses.
1770        The checks now trigger 100 bytes before the end of the workspace.
1771    
1772    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1773    
1774    
1775    Version 8.01 19-Jan-2010
1776    ------------------------
1777    
1778    1.  If a pattern contained a conditional subpattern with only one branch (in
1779        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1780        computed the wrong minimum data length (which is of course zero for such
1781        subpatterns). This could cause incorrect "no match" results.
1782    
1783    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1784        the pattern is reset in the first branch, pcre_compile() failed with
1785        "internal error: code overflow at offset...". This happened only when
1786        the reset was to the original external option setting. (An optimization
1787        abstracts leading options settings into an external setting, which was the
1788        cause of this.)
1789    
1790    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1791        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1792        assertion pattern did not match (meaning that the assertion was true), it
1793        was incorrectly treated as false if the SKIP had been reached during the
1794        matching. This also applied to assertions used as conditions.
1795    
1796    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1797        assertion subpattern, including such a pattern used as a condition,
1798        unpredictable results occurred, instead of the error return
1799        PCRE_ERROR_DFA_UITEM.
1800    
1801    5.  The C++ GlobalReplace function was not working like Perl for the special
1802        situation when an empty string is matched. It now does the fancy magic
1803        stuff that is necessary.
1804    
1805    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1806        removed. (These were left over from very, very early versions of PCRE.)
1807    
1808    7.  Some cosmetic changes to the code to make life easier when compiling it
1809        as part of something else:
1810    
1811        (a) Change DEBUG to PCRE_DEBUG.
1812    
1813        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1814            called "current" as "current_branch", to prevent a collision with the
1815            Linux macro when compiled as a kernel module.
1816    
1817        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1818            prevent a collision with the Linux macro when compiled as a kernel
1819            module.
1820    
1821    8.  In pcre_compile() there are some checks for integer overflows that used to
1822        cast potentially large values to (double). This has been changed to that
1823        when building, a check for int64_t is made, and if it is found, it is used
1824        instead, thus avoiding the use of floating point arithmetic. (There is no
1825        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1826        double.
1827    
1828    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1829        2005 (difference between two addresses compared to an unsigned value).
1830    
1831    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1832        custom one, because of the following reported problem in Windows:
1833    
1834          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1835              under Win32.
1836          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1837              therefore missing the function definition.
1838          - The compiler thus generates a "C" signature for the test function.
1839          - The linker fails to find the "C" function.
1840          - PCRE fails to configure if asked to do so against libbz2.
1841    
1842    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1843        messages were output:
1844    
1845          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1846          rerunning libtoolize, to keep the correct libtool macros in-tree.
1847          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1848    
1849        I have done both of these things.
1850    
1851    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1852        most of the time, it *can* run out if it is given a pattern that contains a
1853        runaway infinite recursion. I updated the discussion in the pcrestack man
1854        page.
1855    
1856    13. Now that we have gone to the x.xx style of version numbers, the minor
1857        version may start with zero. Using 08 or 09 is a bad idea because users
1858        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1859        interpreted as invalid octal numbers. I've updated the previous comment in
1860        configure.ac, and also added a check that gives an error if 08 or 09 are
1861        used.
1862    
1863    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1864        causing partial matching to fail when the end of the subject matched \W
1865        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1866    
1867    15. There were some discrepancies between the declarations in pcre_internal.h
1868        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1869        their definitions. The declarations used "const uschar *" and the
1870        definitions used USPTR. Even though USPTR is normally defined as "const
1871        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1872        reported that: "This difference in casting confuses some C++ compilers, for
1873        example, SunCC recognizes above declarations as different functions and
1874        generates broken code for hbpcre." I have changed the declarations to use
1875        USPTR.
1876    
1877    16. GNU libtool is named differently on some systems. The autogen.sh script now
1878        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1879        (FreeBSD).
1880    
1881    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1882        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1883        comment: "Figure out how to create a longlong from a string: strtoll and
1884        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1885        instance, but it only takes 2 args instead of 3!"
1886    
1887    18. A subtle bug concerned with back references has been fixed by a change of
1888        specification, with a corresponding code fix. A pattern such as
1889        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1890        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1891        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1892        same bug. Such groups have to be quantified to be useful, or contained
1893        inside another quantified group. (If there's no repetition, the reference
1894        can never match.) The problem arises because, having left the group and
1895        moved on to the rest of the pattern, a later failure that backtracks into
1896        the group uses the captured value from the final iteration of the group
1897        rather than the correct earlier one. I have fixed this in PCRE by forcing
1898        any group that contains a reference to itself to be an atomic group; that
1899        is, there cannot be any backtracking into it once it has completed. This is
1900        similar to recursive and subroutine calls.
1901    
1902    
1903    Version 8.00 19-Oct-09
1904    ----------------------
1905    
1906    1.  The table for translating pcre_compile() error codes into POSIX error codes
1907        was out-of-date, and there was no check on the pcre_compile() error code
1908        being within the table. This could lead to an OK return being given in
1909        error.
1910    
1911    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1912        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1913        in a Windows environment.
1914    
1915    3.  The pcregrep --count option prints the count for each file even when it is
1916        zero, as does GNU grep. However, pcregrep was also printing all files when
1917        --files-with-matches was added. Now, when both options are given, it prints
1918        counts only for those files that have at least one match. (GNU grep just
1919        prints the file name in this circumstance, but including the count seems
1920        more useful - otherwise, why use --count?) Also ensured that the
1921        combination -clh just lists non-zero counts, with no names.
1922    
1923    4.  The long form of the pcregrep -F option was incorrectly implemented as
1924        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1925        but it seems right to fix it, and I didn't think it was worth preserving
1926        the old behaviour.
1927    
1928    5.  The command line items --regex=pattern and --regexp=pattern were not
1929        recognized by pcregrep, which required --regex pattern or --regexp pattern
1930        (with a space rather than an '='). The man page documented the '=' forms,
1931        which are compatible with GNU grep; these now work.
1932    
1933    6.  No libpcreposix.pc file was created for pkg-config; there was just
1934        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1935    
1936    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1937        when UCP support is not needed, by modifying the Python script that
1938        generates it from Unicode data files. This should not matter if the module
1939        is correctly used as a library, but I received one complaint about 50K of
1940        unwanted data. My guess is that the person linked everything into his
1941        program rather than using a library. Anyway, it does no harm.
1942    
1943    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1944        was a minimum greater than 1 for a wide character in a possessive
1945        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1946        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1947        character. Chaos in the form of incorrect output or a compiling loop could
1948        result.
1949    
1950    9.  The restrictions on what a pattern can contain when partial matching is
1951        requested for pcre_exec() have been removed. All patterns can now be
1952        partially matched by this function. In addition, if there are at least two
1953        slots in the offset vector, the offset of the earliest inspected character
1954        for the match and the offset of the end of the subject are set in them when
1955        PCRE_ERROR_PARTIAL is returned.
1956    
1957    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1958        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1959        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1960        and may be more useful for multi-segment matching.
1961    
1962    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1963        used to be given if ever the end of the subject was reached; now it is
1964        given only if matching could not proceed because another character was
1965        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1966        string "Z", which now yields "no match" instead of "partial match". In the
1967        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1968        final character ended with (*FAIL).
1969    
1970    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1971        if the pattern had a "must contain" character that was already found in the
1972        earlier partial match, unless partial matching was again requested. For
1973        example, with the pattern /dog.(body)?/, the "must contain" character is
1974        "g". If the first part-match was for the string "dog", restarting with
1975        "sbody" failed. This bug has been fixed.
1976    
1977    13. The string returned by pcre_dfa_exec() after a partial match has been
1978        changed so that it starts at the first inspected character rather than the
1979        first character of the match. This makes a difference only if the pattern
1980        starts with a lookbehind assertion or \b or \B (\K is not supported by
1981        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1982        matching functions compatible, and I think it's the right thing to do.
1983    
1984    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1985        so that the demonstration program is easily available in environments where
1986        PCRE has not been installed from source.
1987    
1988    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1989        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1990        library.
1991    
1992    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1993        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1994        is not the first non-POSIX option to be added. Clearly some people find
1995        these options useful.
1996    
1997    17. If a caller to the POSIX matching function regexec() passes a non-zero
1998        value for nmatch with a NULL value for pmatch, the value of
1999        nmatch is forced to zero.
2000    
2001    18. RunGrepTest did not have a test for the availability of the -u option of
2002        the diff command, as RunTest does. It now checks in the same way as
2003        RunTest, and also checks for the -b option.
2004    
2005    19. If an odd number of negated classes containing just a single character
2006        interposed, within parentheses, between a forward reference to a named
2007        subpattern and the definition of the subpattern, compilation crashed with
2008        an internal error, complaining that it could not find the referenced
2009        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2010        [The bug was that it was starting one character too far in when skipping
2011        over the character class, thus treating the ] as data rather than
2012        terminating the class. This meant it could skip too much.]
2013    
2014    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2015        /g option in pcretest when the pattern contains \K, which makes it possible
2016        to have an empty string match not at the start, even when the pattern is
2017        anchored. Updated pcretest and pcredemo to use this option.
2018    
2019    21. If the maximum number of capturing subpatterns in a recursion was greater
2020        than the maximum at the outer level, the higher number was returned, but
2021        with unset values at the outer level. The correct (outer level) value is
2022        now given.
2023    
2024    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2025        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2026        make it do so. The string so far is captured, making this feature
2027        compatible with Perl.
2028    
2029    23. The tests have been re-organized, adding tests 11 and 12, to make it
2030        possible to check the Perl 5.10 features against Perl 5.10.
2031    
2032    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2033        pattern matches a fixed length string. PCRE did not allow this; now it
2034        does. Neither allows recursion.
2035    
2036    25. I finally figured out how to implement a request to provide the minimum
2037        length of subject string that was needed in order to match a given pattern.
2038        (It was back references and recursion that I had previously got hung up
2039        on.) This code has now been added to pcre_study(); it finds a lower bound
2040        to the length of subject needed. It is not necessarily the greatest lower
2041        bound, but using it to avoid searching strings that are too short does give
2042        some useful speed-ups. The value is available to calling programs via
2043        pcre_fullinfo().
2044    
2045    26. While implementing 25, I discovered to my embarrassment that pcretest had
2046        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2047        study optimizations had never been tested with that matching function.
2048        Oops. What is worse, even when it was passed study data, there was a bug in
2049        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2050        were also very few tests of studied patterns with pcre_dfa_exec().
2051    
2052    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2053        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2054        on the other side of the coin, they are no longer allowed to have different
2055        names, because these cannot be distinguished in PCRE, and this has caused
2056        confusion. (This is a difference from Perl.)
2057    
2058    28. When duplicate subpattern names are present (necessarily with different
2059        numbers, as required by 27 above), and a test is made by name in a
2060        conditional pattern, either for a subpattern having been matched, or for
2061        recursion in such a pattern, all the associated numbered subpatterns are
2062        tested, and the overall condition is true if the condition is true for any
2063        one of them. This is the way Perl works, and is also more like the way
2064        testing by number works.
2065    
2066    
2067    Version 7.9 11-Apr-09
2068    ---------------------
2069    
2070    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2071        (pcretest), all targets were linked against these libraries. This included
2072        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2073        libraries. This caused unwanted dependencies to be created. This problem
2074        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2075        pcretest is linked with readline.
2076    
2077    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2078        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2079        moved outside it again, because FALSE and TRUE are already defined in AIX,
2080        but BOOL is not.
2081    
2082    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2083        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2084    
2085    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2086        hyphens) following file names and line numbers when outputting matching
2087        lines. This is not true; no spaces are inserted. I have also clarified the
2088        wording for the --colour (or --color) option.
2089    
2090    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2091        was not coloured; this is different to GNU grep, so I have changed it to be
2092        the same.
2093    
2094    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2095        each matching line was coloured. Now it goes on to look for further matches
2096        of any of the test patterns, which is the same behaviour as GNU grep.
2097    
2098    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2099        doesn't make sense to accept an empty string match in pcregrep, so I have
2100        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2101        seems to be how GNU grep behaves. [But see later change 40 for release
2102        8.33.]
2103    
2104    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2105        start or after a newline", because the conditional assertion was not being
2106        correctly handled. The rule now is that both the assertion and what follows
2107        in the first alternative must satisfy the test.
2108    
2109    9.  If auto-callout was enabled in a pattern with a conditional group whose
2110        condition was an assertion, PCRE could crash during matching, both with
2111        pcre_exec() and pcre_dfa_exec().
2112    
2113    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2114        used for matching.
2115    
2116    11. Unicode property support in character classes was not working for
2117        characters (bytes) greater than 127 when not in UTF-8 mode.
2118    
2119    12. Added the -M command line option to pcretest.
2120    
2121    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2122    
2123    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2124    
2125    16. Added comments and documentation about mis-use of no_arg in the C++
2126        wrapper.
2127    
2128    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2129        from Martin Jerabek that uses macro names for all relevant character and
2130        string constants.
2131    
2132    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2133        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2134        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2135        these, but not everybody uses configure.
2136    
2137    19. A conditional group that had only one branch was not being correctly
2138        recognized as an item that could match an empty string. This meant that an
2139        enclosing group might also not be so recognized, causing infinite looping
2140        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2141        with the subject "ab", where knowledge that the repeated group can match
2142        nothing is needed in order to break the loop.
2143    
2144    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2145        exec(), but without supplying a callout function, matching went wrong.
2146    
2147    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2148        leak if the size of the offset vector was greater than 30. When the vector
2149        is smaller, the saved offsets during recursion go onto a local stack
2150        vector, but for larger vectors malloc() is used. It was failing to free
2151        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2152        error, in fact).
2153    
2154    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2155        heapframe that is used only when UTF-8 support is enabled. This caused no
2156        problem, but was untidy.
2157    
2158    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2159        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2160        included within another project.
2161    
2162    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2163        slightly modified by me:
2164    
2165          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2166              not building pcregrep.
2167    
2168          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2169              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2170    
2171    25. Forward references, both numeric and by name, in patterns that made use of
2172        duplicate group numbers, could behave incorrectly or give incorrect errors,
2173        because when scanning forward to find the reference group, PCRE was not
2174        taking into account the duplicate group numbers. A pattern such as
2175        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2176    
2177    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2178        the feature of a custom pointer more persuasive (as requested by a user).
2179    
2180    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2181        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2182        pre-defined.
2183    
2184    28. Added support for (*UTF8) at the start of a pattern.
2185    
2186    29. Arrange for flags added by the "release type" setting in CMake to be shown
2187        in the configuration summary.
2188    
2189    
2190    Version 7.8 05-Sep-08
2191    ---------------------
2192    
2193    1.  Replaced UCP searching code with optimized version as implemented for Ad
2194        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2195        stage table and inline lookup instead of a function, giving speed ups of 2
2196        to 5 times on some simple patterns that I tested. Permission was given to
2197        distribute the MultiStage2.py script that generates the tables (it's not in
2198        the tarball, but is in the Subversion repository).
2199    
2200    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2201        scripts.
2202    
2203    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2204        a group with a zero qualifier. The result of the study could be incorrect,
2205        or the function might crash, depending on the pattern.
2206    
2207    4.  Caseless matching was not working for non-ASCII characters in back
2208        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2209        It now works when Unicode Property Support is available.
2210    
2211    5.  In pcretest, an escape such as \x{de} in the data was always generating
2212        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2213        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2214        truncation.
2215    
2216    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2217    
2218    7.  Added two (int) casts to pcregrep when printing the difference of two
2219        pointers, in case they are 64-bit values.
2220    
2221    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2222        test 2 if it fails.
2223    
2224    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2225        and a #define of that name to empty if it is not externally set. This is to
2226        allow users of MSVC to set it if necessary.
2227    
2228    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2229        the convenience functions in the pcre_get.c source file.
2230    
2231    11. An option change at the start of a pattern that had top-level alternatives
2232        could cause overwriting and/or a crash. This command provoked a crash in
2233        some environments:
2234    
2235          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2236    
2237        This potential security problem was recorded as CVE-2008-2371.
2238    
2239    12. For a pattern where the match had to start at the beginning or immediately
2240        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2241        pcre_dfa_exec() could read past the end of the passed subject if there was
2242        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2243        pcretest so that it places the subject at the end of its malloc-ed buffer.
2244    
2245    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2246        exec() might read past the end of the data buffer in UTF-8 mode.
2247    
2248    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2249        the data contained the byte 0x85 as part of a UTF-8 character within its
2250        first line. This applied both to normal and DFA matching.
2251    
2252    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2253        /^[^d]*?$/8 failed to match "abc".
2254    
2255    16. Added a missing copyright notice to pcrecpp_internal.h.
2256    
2257    17. Make it more clear in the documentation that values returned from
2258        pcre_exec() in ovector are byte offsets, not character counts.
2259    
2260    18. Tidied a few places to stop certain compilers from issuing warnings.
2261    
2262    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2263        supplied by Stefan Weber. I made a further small update for 7.8 because
2264        there is a change of source arrangements: the pcre_searchfuncs.c module is
2265        replaced by pcre_ucd.c.
2266    
2267    
2268    Version 7.7 07-May-08
2269    ---------------------
2270    
2271    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2272        a string to a long long, pretend we don't even have a long long." This is
2273        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2274    
2275    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2276        pre-7.6 versions, which defined a global no_arg variable instead of putting
2277        it in the RE class. (See also #8 below.)
2278    
2279    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2280        Lopes.
2281    
2282    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2283    
2284        (1) The include/exclude patterns were being applied to the whole pathnames
2285            of files, instead of just to the final components.
2286    
2287        (2) If there was more than one level of directory, the subdirectories were
2288            skipped unless they satisfied the include/exclude conditions. This is
2289            inconsistent with GNU grep (and could even be seen as contrary to the
2290            pcregrep specification - which I improved to make it absolutely clear).
2291            The action now is always to scan all levels of directory, and just
2292            apply the include/exclude patterns to regular files.
2293    
2294    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2295        --exclude_dir in the tests to avoid scanning .svn directories.
2296    
2297    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2298        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2299        doesn't support NULs in patterns.
2300    
2301    7.  Added some missing "const"s to declarations of static tables in
2302        pcre_compile.c and pcre_dfa_exec.c.
2303    
2304    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2305        caused by fix #2  above. (Subsequently also a second patch to fix the
2306        first patch. And a third patch - this was a messy problem.)
2307    
2308    9.  Applied Craig's patch to remove the use of push_back().
2309    
2310    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2311        matching function regexec().
2312    
2313    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2314        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2315        references. PCRE supports relative numbers with this syntax (I don't think
2316        Oniguruma does).
2317    
2318    12. Previously, a group with a zero repeat such as (...){0} was completely
2319        omitted from the compiled regex. However, this means that if the group
2320        was called as a subroutine from elsewhere in the pattern, things went wrong
2321        (an internal error was given). Such groups are now left in the compiled
2322        pattern, with a new opcode that causes them to be skipped at execution
2323        time.
2324    
2325    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2326        to the way PCRE behaves:
2327    
2328        (a) A lone ] character is dis-allowed (Perl treats it as data).
2329    
2330        (b) A back reference to an unmatched subpattern matches an empty string
2331            (Perl fails the current match path).
2332    
2333        (c) A data ] in a character class must be notated as \] because if the
2334            first data character in a class is ], it defines an empty class. (In
2335            Perl it is not possible to have an empty class.) The empty class []
2336            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2337            The negative empty class [^] matches any one character, independently
2338            of the DOTALL setting.
2339    
2340    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2341        non-existent subpattern following a character class starting with ']' and
2342        containing () gave an internal compiling error instead of "reference to
2343        non-existent subpattern". Fortunately, when the pattern did exist, the
2344        compiled code was correct. (When scanning forwards to check for the
2345        existence of the subpattern, it was treating the data ']' as terminating
2346        the class, so got the count wrong. When actually compiling, the reference
2347        was subsequently set up correctly.)
2348    
2349    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2350        it was being rejected as not supported by pcre_dfa_exec(), even though
2351        other assertions are supported. I have made pcre_dfa_exec() support
2352        (*FAIL).
2353    
2354    16. The implementation of 13c above involved the invention of a new opcode,
2355        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2356        cannot be changed at match time, I realized I could make a small
2357        improvement to matching performance by compiling OP_ALLANY instead of
2358        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2359        on the OP_ANY path.
2360    
2361    17. Compiling pcretest on Windows with readline support failed without the
2362        following two fixes: (1) Make the unistd.h include conditional on
2363        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2364    
2365    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2366        ncurses library to be included for pcretest when ReadLine support is
2367        requested, but also to allow for it to be overridden. This patch came from
2368        Daniel Bergström.
2369    
2370    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2371        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2372        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2373        spotting this.
2374    
2375    
2376    Version 7.6 28-Jan-08
2377    ---------------------
2378    
2379    1.  A character class containing a very large number of characters with
2380        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2381        overflow.
2382    
2383    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2384        HAVE_LONG_LONG is not defined.
2385    
2386    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2387        bring them up to date and include new features. This patch includes:
2388    
2389        - Fixed PH's badly added libz and libbz2 support.
2390        - Fixed a problem with static linking.
2391        - Added pcredemo. [But later removed - see 7 below.]
2392        - Fixed dftables problem and added an option.
2393        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2394            HAVE_LONG_LONG.
2395        - Added readline support for pcretest.
2396        - Added an listing of the option settings after cmake has run.
2397    
2398    4.  A user submitted a patch to Makefile that makes it easy to create
2399        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2400        Makefile.am that cause it to include this special target, without
2401        affecting anything else. Note that the same mingw target plus all
2402        the other distribution libraries and programs are now supported
2403        when configuring with CMake (see 6 below) instead of with
2404        Configure/Make.
2405    
2406    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2407        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2408        exported in the Windows port". It has not yet been confirmed that the patch
2409        solves the problem, but it does no harm.
2410    
2411    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2412        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2413        with CMake, and also correct the comment about stack recursion.
2414    
2415    7.  Remove the automatic building of pcredemo from the ./configure system and
2416        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2417        of a program that users should build themselves after PCRE is installed, so
2418        building it automatically is not really right. What is more, it gave
2419        trouble in some build environments.
2420    
2421    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2422    
2423    
2424    Version 7.5 10-Jan-08
2425    ---------------------
2426    
2427    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2428        values in parens when parsing an RE using the C++ wrapper."
2429    
2430    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2431        Characters greater than 255 were excluded from the class instead of being
2432        included.
2433    
2434    3.  The same bug as (2) above applied to negated POSIX classes such as
2435        [:^space:].
2436    
2437    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2438        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2439        I have changed it.
2440    
2441    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2442        first named subpattern) and a construct such as (?&a) would reference the
2443        first named subpattern whose name started with "a" (in other words, the
2444        length check was missing). Both these problems are fixed. "Subpattern name
2445        expected" is now given for (?&) (a zero-length name), and this patch also
2446        makes it give the same error for \k'' (previously it complained that that
2447        was a reference to a non-existent subpattern).
2448    
2449    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2450        this is right because (?- can be followed by option settings as well as by
2451        digits. I have, however, made the messages clearer.
2452    
2453    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2454        than the number used in the conditional) now cause a compile-time error.
2455        This is actually not compatible with Perl, which accepts such patterns, but
2456        treats the conditional as always being FALSE (as PCRE used to), but it
2457        seems to me that giving a diagnostic is better.
2458    
2459    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2460        and messages.
2461    
2462    9.  Fix two occurrences of "backslash" in comments that should have been
2463        "backspace".
2464    
2465    10. Remove two redundant lines of code that can never be obeyed (their function
2466        was moved elsewhere).
2467    
2468    11. The program that makes PCRE's Unicode character property table had a bug
2469        which caused it to generate incorrect table entries for sequences of
2470        characters that have the same character type, but are in different scripts.
2471        It amalgamated them into a single range, with the script of the first of
2472        them. In other words, some characters were in the wrong script. There were
2473        thirteen such cases, affecting characters in the following ranges:
2474    
2475          U+002b0 - U+002c1
2476          U+0060c - U+0060d
2477          U+0061e - U+00612
2478          U+0064b - U+0065e
2479          U+0074d - U+0076d
2480          U+01800 - U+01805
2481          U+01d00 - U+01d77
2482          U+01d9b - U+01dbf
2483          U+0200b - U+0200f
2484          U+030fc - U+030fe
2485          U+03260 - U+0327f
2486          U+0fb46 - U+0fbb1
2487          U+10450 - U+1049d
2488    
2489    12. The -o option (show only the matching part of a line) for pcregrep was not
2490        compatible with GNU grep in that, if there was more than one match in a
2491        line, it showed only the first of them. It now behaves in the same way as
2492        GNU grep.
2493    
2494    13. If the -o and -v options were combined for pcregrep, it printed a blank
2495        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2496        does the same. The return code can be used to tell if there were any
2497        non-matching lines.
2498    
2499    14. Added --file-offsets and --line-offsets to pcregrep.
2500    
2501    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2502        infinitely looping recursion. The bug was that positive lookaheads were not
2503        being skipped when checking for a possible empty match (negative lookaheads
2504        and both kinds of lookbehind were skipped).
2505    
2506    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2507        inclusion of <windows.h> to before rather than after the definition of
2508        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2509    
2510    17. Specifying a possessive quantifier with a specific limit for a Unicode
2511        character property caused pcre_compile() to compile bad code, which led at
2512        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2513        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2514        caused the error; without that there was no problem.
2515    
2516    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2517    
2518    19. Added --enable-pcretest-libreadline.
2519    
2520    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2521        RE::GlobalReplace(). As a result, the number of replacements returned was
2522        double what it should be. I removed one of the increments, but Craig sent a
2523        later patch that removed the other one (the right fix) and added unit tests
2524        that check the return values (which was not done before).
2525    
2526    21. Several CMake things:
2527    
2528        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2529            the names libpcre and libpcreposix, not just pcre and pcreposix.
2530    
2531        (2) The above change means that pcretest and pcregrep are now correctly
2532            linked with the newly-built libraries, not previously installed ones.
2533    
2534        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2535    
2536    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2537        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2538        UTF-8 newline character). The key issue is that the pattern starts .*;
2539        this means that the match must be either at the beginning, or after a
2540        newline. The bug was in the code for advancing after a failed match and
2541        checking that the new position followed a newline. It was not taking
2542        account of UTF-8 characters correctly.
2543    
2544    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2545        character classes. PCRE was not treating the sequence [:...:] as a
2546        character class unless the ... were all letters. Perl, however, seems to
2547        allow any characters between [: and :], though of course it rejects as
2548        unknown any "names" that contain non-letters, because all the known class
2549        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2550        for example, whereas PCRE did not - it did not recognize a POSIX character
2551        class. This seemed a bit dangerous, so the code has been changed to be
2552        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2553        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2554        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2555        Perl does, and where it didn't before.
2556    
2557    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2558        Windows environments %n is disabled by default.
2559    
2560    
2561    Version 7.4 21-Sep-07
2562    ---------------------
2563    
2564    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2565        means that a class such as [\s] counted as "explicit reference to CR or
2566        LF". That isn't really right - the whole point of the change was to try to
2567        help when there was an actual mention of one of the two characters. So now
2568        the change happens only if \r or \n (or a literal CR or LF) character is
2569        encountered.
2570    
2571    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2572        of both had grown to the point where there were only 3 bits left.
2573        Fortunately, there was spare space in the data structure, and so I have
2574        moved the internal flags into a new 16-bit field to free up more option
2575        bits.
2576    
2577    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2578        but did not set the internal JCHANGED flag - either of these is enough to
2579        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2580        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2581        start sets both bits.
2582    
2583    4.  Added options (at build time, compile time, exec time) to change \R from
2584        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2585    
2586    5.  doc/pcresyntax.html was missing from the distribution.
2587    
2588    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2589        compatibility, even though it is no longer used.
2590    
2591    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2592        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2593        windows.h file is present (where different names are used). [This was
2594        reversed later after testing - see 16 below.]
2595    
2596    8.  Changed all #include <config.h> to #include "config.h". There were also
2597        some further <pcre.h> cases that I changed to "pcre.h".
2598    
2599    9.  When pcregrep was used with the --colour option, it missed the line ending
2600        sequence off the lines that it output.
2601    
2602    10. It was pointed out to me that arrays of string pointers cause lots of
2603        relocations when a shared library is dynamically loaded. A technique of
2604        using a single long string with a table of offsets can drastically reduce
2605        these. I have refactored PCRE in four places to do this. The result is
2606        dramatic:
2607    
2608          Originally:                          290
2609          After changing UCP table:            187
2610          After changing error message table:   43
2611          After changing table of "verbs"       36
2612          After changing table of Posix names   22
2613    
2614        Thanks to the folks working on Gregex for glib for this insight.
2615    
2616    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2617        unicode-properties was also set.
2618    
2619    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2620    
2621    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2622        checked only for CRLF.
2623    
2624    14. Added casts to pcretest.c to avoid compiler warnings.
2625    
2626    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2627    
2628    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2629        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2630        entirely. This removes changes made in 7 above.
2631    
2632    17. The CMake files have been updated, and there is now more information about
2633        building with CMake in the NON-UNIX-USE document.
2634    
2635    
2636    Version 7.3 28-Aug-07
2637    ---------------------
2638    
2639     1. In the rejigging of the build system that eventually resulted in 7.1, the
2640        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2641        brackets there is not right, since it causes compilers to look for an
2642        installed pcre.h, not the version that is in the source that is being
2643        compiled (which of course may be different). I have changed it back to:
2644    
2645          #include "pcre.h"
2646    
2647        I have a vague recollection that the change was concerned with compiling in
2648        different directories, but in the new build system, that is taken care of
2649        by the VPATH setting the Makefile.
2650    
2651     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2652        when the subject happened to end in the byte 0x85 (e.g. if the last
2653        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2654        characters but of course it shouldn't be taken as a newline when it is part
2655        of another character. The bug was that, for an unlimited repeat of . in
2656        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2657        characters when looking for a newline.
2658    
2659     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2660    
2661     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2662        in debug output.
2663    
2664     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2665        long printing in the pcrecpp unittest when running under MinGW.
2666    
2667     6. ESC_K was left out of the EBCDIC table.
2668    
2669     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2670        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2671        limit also applies to "virtual nesting" when a pattern is recursive, and in
2672        this case 1000 isn't so big. I have been able to remove this limit at the
2673        expense of backing off one optimization in certain circumstances. Normally,
2674        when pcre_exec() would call its internal match() function recursively and
2675        immediately return the result unconditionally, it uses a "tail recursion"
2676        feature to save stack. However, when a subpattern that can match an empty
2677        string has an unlimited repetition quantifier, it no longer makes this
2678        optimization. That gives it a stack frame in which to save the data for
2679        checking that an empty string has been matched. Previously this was taken
2680        from the 1000-entry workspace that had been reserved. So now there is no
2681        explicit limit, but more stack is used.
2682    
2683     8. Applied Daniel's patches to solve problems with the import/export magic
2684        syntax that is required for Windows, and which was going wrong for the
2685        pcreposix and pcrecpp parts of the library. These were overlooked when this
2686        problem was solved for the main library.
2687    
2688     9. There were some crude static tests to avoid integer overflow when computing
2689        the size of patterns that contain repeated groups with explicit upper
2690        limits. As the maximum quantifier is 65535, the maximum group length was
2691        set at 30,000 so that the product of these two numbers did not overflow a
2692        32-bit integer. However, it turns out that people want to use groups that
2693        are longer than 30,000 bytes (though not repeat them that many times).
2694        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2695        made it possible to implement the integer overflow checks in a much more
2696        dynamic way, which I have now done. The artificial limitation on group
2697        length has been removed - we now have only the limit on the total length of
2698        the compiled pattern, which depends on the LINK_SIZE setting.
2699    
2700    10. Fixed a bug in the documentation for get/copy named substring when
2701        duplicate names are permitted. If none of the named substrings are set, the
2702        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2703        empty string.
2704    
2705    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2706        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2707        because the ] is interpreted as the first data character and the
2708        terminating ] is not found. PCRE has been made compatible with Perl in this
2709        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2710        cause memory overwriting.
2711    
2712    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2713        string has been matched (to stop an infinite loop). It was not recognizing
2714        a conditional subpattern that could match an empty string if that
2715        subpattern was within another subpattern. For example, it looped when
2716        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2717        condition was not nested. This bug has been fixed.
2718    
2719    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2720        past the start of the subject in the presence of bytes with the top bit
2721        set, for example "\x8aBCD".
2722    
2723    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2724        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2725    
2726    14. Optimized (?!) to (*FAIL).
2727    
2728    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2729        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2730        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2731        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2732        does: it's just the validity check that is more restrictive.
2733    
2734    16. Inserted checks for integer overflows during escape sequence (backslash)
2735        processing, and also fixed erroneous offset values for syntax errors during
2736        backslash processing.
2737    
2738    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2739        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2740    
2741    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2742        caused an overrun.
2743    
2744    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2745        something other than just ASCII characters) inside a group that had an
2746        unlimited repeat caused a loop at compile time (while checking to see
2747        whether the group could match an empty string).
2748    
2749    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2750        [\P{Any}] did so. (Error in the code for printing property names.)
2751    
2752    21. An orphan \E inside a character class could cause a crash.
2753    
2754    22. A repeated capturing bracket such as (A)? could cause a wild memory
2755        reference during compilation.
2756    
2757    23. There are several functions in pcre_compile() that scan along a compiled
2758        expression for various reasons (e.g. to see if it's fixed length for look
2759        behind). There were bugs in these functions when a repeated \p or \P was
2760        present in the pattern. These operators have additional parameters compared
2761        with \d, etc, and these were not being taken into account when moving along
2762        the compiled data. Specifically:
2763    
2764        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2765            length.
2766    
2767        (b) An item such as \pL+ within a repeated group could cause crashes or
2768            loops.
2769    
2770        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2771            "reference to non-existent subpattern" error.
2772    
2773        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2774    
2775    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2776        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2777    
2778    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2779    
2780    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2781        character were causing crashes (broken optimization).
2782    
2783    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2784        \p or \P) caused a compile-time loop.
2785    
2786    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2787        break. For example, the unstudied pattern [\r\n]A does not match the string
2788        "\r\nA" because change 7.0/46 below moves the current point on by two
2789        characters after failing to match at the start. However, the pattern \nA
2790        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2791        the same is true. There doesn't seem any very clean way out of this, but
2792        what I have chosen to do makes the common cases work: PCRE now takes note
2793        of whether there can be an explicit match for \r or \n anywhere in the
2794        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2795        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2796        pattern has explicit CR or LF references.
2797    
2798    29. Added (*CR) etc for changing newline setting at start of pattern.
2799    
2800    
2801    Version 7.2 19-Jun-07
2802    ---------------------
2803    
2804     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2805        which is apparently normally available under Windows.
2806    
2807     2. Re-jig the pcregrep tests with different newline settings in an attempt
2808        to make them independent of the local environment's newline setting.
2809    
2810     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2811    
2812     4. Some of the "internals" tests were previously cut out when the link size
2813        was not 2, because the output contained actual offsets. The recent new
2814        "Z" feature of pcretest means that these can be cut out, making the tests
2815        usable with all link sizes.
2816    
2817     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2818        stack recursion. This gives a massive performance boost under BSD, but just
2819        a small improvement under Linux. However, it saves one field in the frame
2820        in all cases.
2821    
2822     6. Added more features from the forthcoming Perl 5.10:
2823    
2824        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2825            recursion call. It refers to the nth most recently opened parentheses.
2826    
2827        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2828            to be opened parentheses.
2829    
2830        (c) Conditions that refer to capturing parentheses can be specified
2831            relatively, for example, (?(-2)... or (?(+3)...
2832    
2833        (d) \K resets the start of the current match so that everything before
2834            is not part of it.
2835    
2836        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2837    
2838        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2839            reference syntax.
2840    
2841        (g) (?| introduces a group in which the numbering of parentheses in each
2842            alternative starts with the same number.
2843    
2844        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2845    
2846     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2847        PCRE_INFO_JCHANGED.
2848    
2849     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2850        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2851        for detecting groups that can match an empty string.
2852    
2853     9. A pattern with a very large number of alternatives (more than several
2854        hundred) was running out of internal workspace during the pre-compile
2855        phase, where pcre_compile() figures out how much memory will be needed. A
2856        bit of new cunning has reduced the workspace needed for groups with
2857        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2858        workspace instead of running out of the 4096 that are available.
2859    
2860    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2861    
2862    11. Applied patch from Google to remove an optimization that didn't quite work.
2863        The report of the bug said:
2864    
2865          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2866          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2867          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2868    
2869    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2870        it matched the wrong number of bytes.
2871    
2872    
2873    Version 7.1 24-Apr-07
2874    ---------------------
2875    
2876     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2877        that is more "standard", making use of automake and other Autotools. There
2878        is some re-arrangement of the files and adjustment of comments consequent
2879        on this.
2880    
2881     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2882        for recursive directory scanning broke on some systems because the files
2883        are not scanned in any specific order and on different systems the order
2884        was different. A call to "sort" has been inserted into RunGrepTest for the
2885        approprate test as a short-term fix. In the longer term there may be an
2886        alternative.
2887    
2888     3. I had an email from Eric Raymond about problems translating some of PCRE's
2889        man pages to HTML (despite the fact that I distribute HTML pages, some
2890        people do their own conversions for various reasons). The problems
2891        concerned the use of low-level troff macros .br and .in. I have therefore
2892        removed all such uses from the man pages (some were redundant, some could
2893        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2894        HTML has been updated to handle .nf/.fi and to complain if it encounters
2895        .br or .in.
2896    
2897     4. Updated comments in configure.ac that get placed in config.h.in and also
2898        arranged for config.h to be included in the distribution, with the name
2899        config.h.generic, for the benefit of those who have to compile without
2900        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2901    
2902     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2903        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2904        makevp.bat for the current PCRE, using the additional files
2905        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2906    
2907     6. A Windows user reported a minor discrepancy with test 2, which turned out
2908        to be caused by a trailing space on an input line that had got lost in his
2909        copy. The trailing space was an accident, so I've just removed it.
2910    
2911     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2912        that is needed.
2913    
2914     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2915        as "const" (a) because they are and (b) because it helps the PHP
2916        maintainers who have recently made a script to detect big data structures
2917        in the php code that should be moved to the .rodata section. I remembered
2918        to update Builducptable as well, so it won't revert if ucptable.h is ever
2919        re-created.
2920    
2921     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2922        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2923        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2924        support is not required. This saves 1.5-2K of code, which is important in
2925        some applications.
2926    
2927        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2928        so as not to refer to the tables, even though these functions will never be
2929        called when UTF-8 support is disabled. Otherwise there are problems with a
2930        shared library.
2931    
2932    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2933    
2934        (a) It was defining its arguments as char * instead of void *.
2935    
2936        (b) It was assuming that all moves were upwards in memory; this was true
2937            a long time ago when I wrote it, but is no longer the case.
2938    
2939        The emulated memove() is provided for those environments that have neither
2940        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2941        is clearly not the case, as these two bugs were recently reported.
2942    
2943    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2944        and Detrail to create the HTML documentation, the .txt form of the man
2945        pages, and it removes trailing spaces from listed files. It also creates
2946        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2947        case, it wraps all the #defines with #ifndefs. This script should be run
2948        before "make dist".
2949    
2950    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2951        with Unicode property support.
2952    
2953        (a) For a maximizing quantifier, if the two different cases of the
2954            character were of different lengths in their UTF-8 codings (there are
2955            some cases like this - I found 11), and the matching function had to
2956            back up over a mixture of the two cases, it incorrectly assumed they
2957            were both the same length.
2958    
2959        (b) When PCRE was configured to use the heap rather than the stack for
2960            recursion during matching, it was not correctly preserving the data for
2961            the other case of a UTF-8 character when checking ahead for a match
2962            while processing a minimizing repeat. If the check also involved
2963            matching a wide character, but failed, corruption could cause an
2964            erroneous result when trying to check for a repeat of the original
2965            character.
2966    
2967    13. Some tidying changes to the testing mechanism:
2968    
2969        (a) The RunTest script now detects the internal link size and whether there
2970            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2971            values substituted by "configure". (The RunGrepTest script already did
2972            this for UTF-8.) The configure.ac script no longer substitutes the
2973            relevant variables.
2974    
2975        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2976            with length and offset values. This means that the output is different
2977            for different internal link sizes. Test 2 is skipped for link sizes
2978            other than 2 because of this, bypassing the problem. Unfortunately,
2979            there was also a test in test 3 (the locale tests) that used /B and
2980            failed for link sizes other than 2. Rather than cut the whole test out,
2981            I have added a new /Z option to pcretest that replaces the length and
2982            offset values with spaces. This is now used to make test 3 independent
2983            of link size. (Test 2 will be tidied up later.)
2984    
2985    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2986        segmentation fault instead of returning the appropriate error message.
2987    
2988    15. In multiline mode when the newline sequence was set to "any", the pattern
2989        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2990        This doesn't seem right; it now treats the CRLF combination as the line
2991        ending, and so does not match in that case. It's only a pattern such as ^$
2992        that would hit this one: something like ^ABC$ would have failed after \r
2993        and then tried again after \r\n.
2994    
2995    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2996        in an attempt to make files that differ only in their line terminators
2997        compare equal. This works on Linux.
2998    
2999    17. Under certain error circumstances pcregrep might try to free random memory
3000        as it exited. This is now fixed, thanks to valgrind.
3001    
3002    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3003        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3004        was because its rules for how to advance for /g after matching an empty
3005        string at the end of a line did not allow for this case. They now check for
3006        it specially.
3007    
3008    20. pcretest is supposed to handle patterns and data of any length, by
3009        extending its buffers when necessary. It was getting this wrong when the
3010        buffer for a data line had to be extended.
3011    
3012    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3013        CRLF as a newline sequence.
3014    
3015    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3016        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3017        I have nevertheless tidied it up.
3018    
3019    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3020    
3021    24. Added a man page for pcre-config.
3022    
3023    
3024    Version 7.0 19-Dec-06
3025    ---------------------
3026    
3027     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3028        moving to gcc 4.1.1.
3029    
3030     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3031        sys/time.h, which is documented as needed for this function. It doesn't
3032        seem to matter on Linux, but it showed up on some releases of OS X.
3033    
3034     3. It seems that there are systems where bytes whose values are greater than
3035        127 match isprint() in the "C" locale. The "C" locale should be the
3036        default when a C program starts up. In most systems, only ASCII printing
3037        characters match isprint(). This difference caused the output from pcretest
3038        to vary, making some of the tests fail. I have changed pcretest so that:
3039    
3040        (a) When it is outputting text in the compiled version of a pattern, bytes
3041            other than 32-126 are always shown as hex escapes.
3042    
3043        (b) When it is outputting text that is a matched part of a subject string,
3044            it does the same, unless a different locale has been set for the match
3045            (using the /L modifier). In this case, it uses isprint() to decide.
3046    
3047     4. Fixed a major bug that caused incorrect computation of the amount of memory
3048        required for a compiled pattern when options that changed within the
3049        pattern affected the logic of the preliminary scan that determines the
3050        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3051        that the computed length was too small. The symptoms of this bug were
3052        either the PCRE error "internal error: code overflow" from pcre_compile(),
3053        or a glibc crash with a message such as "pcretest: free(): invalid next
3054        size (fast)". Examples of patterns that provoked this bug (shown in
3055        pcretest format) are:
3056    
3057          /(?-x: )/x
3058          /(?x)(?-x: \s*#\s*)/
3059          /((?i)[\x{c0}])/8
3060          /(?i:[\x{c0}])/8
3061    
3062        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3063        is now done differently.
3064    
3065     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3066        wrapper classes; (b) implement a new function in the C++ scanner that is
3067        more efficient than the old way of doing things because it avoids levels of
3068        recursion in the regex matching; (c) add a paragraph to the documentation
3069        for the FullMatch() function.
3070    
3071     6. The escape sequence \n was being treated as whatever was defined as
3072        "newline". Not only was this contrary to the documentation, which states
3073        that \n is character 10 (hex 0A), but it also went horribly wrong when
3074        "newline" was defined as CRLF. This has been fixed.
3075    
3076     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3077        was being set to -1 for the "end of line" case (supposedly a value that no
3078        character can have). Though this value is never used (the check for end of
3079        line is "zero bytes in current character"), it caused compiler complaints.
3080        I've changed it to 0xffffffff.
3081    
3082     8. In pcre_version.c, the version string was being built by a sequence of
3083        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3084        string (as it is for production releases) called a macro with an empty
3085        argument. The C standard says the result of this is undefined. The gcc
3086        compiler treats it as an empty string (which was what was wanted) but it is
3087        reported that Visual C gives an error. The source has been hacked around to
3088        avoid this problem.
3089    
3090     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3091        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3092        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3093        of them did).
3094    
3095    10. Originally, pcretest opened its input and output without "b"; then I was
3096        told that "b" was needed in some environments, so it was added for release
3097        5.0 to both the input and output. (It makes no difference on Unix-like
3098        systems.) Later I was told that it is wrong for the input on Windows. I've
3099        now abstracted the modes into two macros, to make it easier to fiddle with
3100        them, and removed "b" from the input mode under Windows.
3101    
3102    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3103    
3104    12. Added -help and --help to pcretest as an official way of being reminded
3105        of the options.
3106    
3107    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3108        and pcrecpp.cc because they annoy compilers at high warning levels.
3109    
3110    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3111    
3112    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3113        scripts are not C programs :-) and which was not noticed because it works
3114        on Linux.
3115    
3116    16. pcretest is supposed to handle any length of pattern and data line (as one
3117        line or as a continued sequence of lines) by extending its input buffer if
3118        necessary. This feature was broken for very long pattern lines, leading to
3119        a string of junk being passed to pcre_compile() if the pattern was longer
3120        than about 50K.
3121    
3122    17. I have done a major re-factoring of the way pcre_compile() computes the
3123        amount of memory needed for a compiled pattern. Previously, there was code
3124        that made a preliminary scan of the pattern in order to do this. That was
3125        OK when PCRE was new, but as the facilities have expanded, it has become
3126        harder and harder to keep it in step with the real compile phase, and there
3127        have been a number of bugs (see for example, 4 above). I have now found a
3128        cunning way of running the real compile function in a "fake" mode that
3129        enables it to compute how much memory it would need, while actually only
3130        ever using a few hundred bytes of working memory and without too many
3131        tests of the mode. This should make future maintenance and development
3132        easier. A side effect of this work is that the limit of 200 on the nesting
3133        depth of parentheses has been removed (though this was never a serious
3134        limitation, I suspect). However, there is a downside: pcre_compile() now
3135        runs more slowly than before (30% or more, depending on the pattern). I
3136        hope this isn't a big issue. There is no effect on runtime performance.
3137    
3138    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3139        newline (only possible for the last line of a file) and it was a
3140        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3141    
3142    19. Added additional timing features to pcretest. (1) The -tm option now times
3143        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3144        separate command line item, by a number that specifies the number of
3145        repeats to use when timing. The default is 50000; this gives better
3146        precision, but takes uncomfortably long for very large patterns.
3147    
3148    20. Extended pcre_study() to be more clever in cases where a branch of a
3149        subpattern has no definite first character. For example, (a*|b*)[cd] would
3150        previously give no result from pcre_study(). Now it recognizes that the
3151        first character must be a, b, c, or d.
3152    
3153    21. There was an incorrect error "recursive call could loop indefinitely" if
3154        a subpattern (or the entire pattern) that was being tested for matching an
3155        empty string contained only one non-empty item after a nested subpattern.
3156        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3157        incorrectly, because the \d was being skipped in the check.
3158    
3159    22. The pcretest program now has a new pattern option /B and a command line
3160        option -b, which is equivalent to adding /B to every pattern. This causes
3161        it to show the compiled bytecode, without the additional information that
3162        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3163        is the same as /B/I).
3164    
3165    23. A new optimization is now able automatically to treat some sequences such
3166        as a*b as a*+b. More specifically, if something simple (such as a character
3167        or a simple class like \d) has an unlimited quantifier, and is followed by
3168        something that cannot possibly match the quantified thing, the quantifier
3169        is automatically "possessified".
3170    
3171    24. A recursive reference to a subpattern whose number was greater than 39
3172        went wrong under certain circumstances in UTF-8 mode. This bug could also
3173        have affected the operation of pcre_study().
3174    
3175    25. Realized that a little bit of performance could be had by replacing
3176        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3177    
3178    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3179    
3180    27. Possessive quantifiers such as a++ were previously implemented by turning
3181        them into atomic groups such as ($>a+). Now they have their own opcodes,
3182        which improves performance. This includes the automatically created ones
3183        from 23 above.
3184    
3185    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3186        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3187        the first matched character to be a colon. This applied both to named and
3188        numbered groups.
3189    
3190    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3191    
3192    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3193        building PCRE on Windows easier, so I have included it in the distribution.
3194    
3195    31. There is now a check in pcretest against a ridiculously large number being
3196        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3197        loop, the loop is abandoned.
3198    
3199    32. Forward references to subpatterns in conditions such as (?(2)...) where
3200        subpattern 2 is defined later cause pcre_compile() to search forwards in
3201        the pattern for the relevant set of parentheses. This search went wrong
3202        when there were unescaped parentheses in a character class, parentheses
3203        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3204    
3205    33. "Subroutine" calls and backreferences were previously restricted to
3206        referencing subpatterns earlier in the regex. This restriction has now
3207        been removed.
3208    
3209    34. Added a number of extra features that are going to be in Perl 5.10. On the
3210        whole, these are just syntactic alternatives for features that PCRE had
3211        previously implemented using the Python syntax or my own invention. The
3212        other formats are all retained for compatibility.
3213    
3214        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3215            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3216            also .NET compatible.
3217    
3218        (b) A recursion or subroutine call to a named group can now be defined as
3219            (?&name) as well as (?P>name).
3220    
3221        (c) A backreference to a named group can now be defined as \k<name> or
3222            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3223            5.10, are also .NET compatible.
3224    
3225        (d) A conditional reference to a named group can now use the syntax
3226            (?(<name>) or (?('name') as well as (?(name).
3227    
3228        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3229            groups (named and numbered) that are never evaluated inline, but can be
3230            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3231            is always false. There may be only one alternative in such a group.
3232    
3233        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3234            as the simple (?(R). The condition is true only if the most recent
3235            recursion is that of the given number or name. It does not search out
3236            through the entire recursion stack.
3237    
3238        (g) The escape \gN or \g{N} has been added, where N is a positive or
3239            negative number, specifying an absolute or relative reference.
3240    
3241    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3242        some "unreachable code" warnings.
3243    
3244    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3245        things, this adds five new scripts.
3246    
3247    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3248        There were also incompatibilities regarding the handling of \Q..\E inside
3249        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3250        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3251    
3252    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3253        matches an empty string, and forcibly breaks the loop. There were bugs in
3254        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3255        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3256        separate and independent bugs (that affected different cases) have been
3257        fixed.
3258    
3259    39. Refactored the code to abolish the use of different opcodes for small
3260        capturing bracket numbers. This is a tidy that I avoided doing when I
3261        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3262        The new approach is not only tidier, it makes it possible to reduce the
3263        memory needed to fix the previous bug (38).
3264    
3265    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3266        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3267        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3268        mode.
3269    
3270    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3271        report.
3272    
3273    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3274        copy construction and assignment in the C++ wrapper.
3275    
3276    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3277        couple of bugs that could have given wrong results in the "--newline=crlf"
3278        case.
3279    
3280    44. Added a number of casts and did some reorganization of signed/unsigned int
3281        variables following suggestions from Dair Grant. Also renamed the variable
3282        "this" as "item" because it is a C++ keyword.
3283    
3284    45. Arranged for dftables to add
3285    
3286          #include "pcre_internal.h"
3287    
3288        to pcre_chartables.c because without it, gcc 4.x may remove the array
3289        definition from the final binary if PCRE is built into a static library and
3290        dead code stripping is activated.
3291    
3292    46. For an unanchored pattern, if a match attempt fails at the start of a
3293        newline sequence, and the newline setting is CRLF or ANY, and the next two
3294        characters are CRLF, advance by two characters instead of one.
3295    
3296    
3297    Version 6.7 04-Jul-06
3298    ---------------------
3299    
3300     1. In order to handle tests when input lines are enormously long, pcretest has
3301        been re-factored so that it automatically extends its buffers when
3302        necessary. The code is crude, but this _is_ just a test program. The
3303        default size has been increased from 32K to 50K.
3304    
3305     2. The code in pcre_study() was using the value of the re argument before
3306        testing it for NULL. (Of course, in any sensible call of the function, it
3307        won't be NULL.)
3308    
3309     3. The memmove() emulation function in pcre_internal.h, which is used on
3310        systems that lack both memmove() and bcopy() - that is, hardly ever -
3311        was missing a "static" storage class specifier.
3312    
3313     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3314        containing an extended class (one that cannot be represented by a bitmap
3315        because it contains high-valued characters or Unicode property items, e.g.
3316        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3317        pattern, but PCRE should not loop if you do not (it no longer does).
3318        [Detail: two cases were found: (a) a repeated subpattern containing an
3319        extended class; (b) a recursive reference to a subpattern that followed a
3320        previous extended class. It wasn't skipping over the extended class
3321        correctly when UTF-8 mode was not set.]
3322    
3323     5. A negated single-character class was not being recognized as fixed-length
3324        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3325        compile error "lookbehind assertion is not fixed length".
3326    
3327     6. The RunPerlTest auxiliary script was showing an unexpected difference
3328        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3329        write a Perl script that can interpret lines of an input file either as
3330        byte characters or as UTF-8, which is what "perltest" was being required to
3331        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3332        can't do is switch easily at run time between having the "use utf8;" pragma
3333        or not. In the end, I fudged it by using the RunPerlTest script to insert
3334        "use utf8;" explicitly for the UTF-8 tests.
3335    
3336     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3337        the end of the subject string, contrary to the documentation and to what
3338        Perl does. This was true of both matching functions. Now it matches only at
3339        the start of the subject and immediately after *internal* newlines.
3340    
3341     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3342        a pointer to an int instead of a pointer to an unsigned long int. This
3343        caused problems on 64-bit systems.
3344    
3345     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3346        instance of the 'standard' template library not being so standard".
3347    
3348    10. There was no check on the number of named subpatterns nor the maximum
3349        length of a subpattern name. The product of these values is used to compute
3350        the size of the memory block for a compiled pattern. By supplying a very
3351        long subpattern name and a large number of named subpatterns, the size
3352        computation could be caused to overflow. This is now prevented by limiting
3353        the length of names to 32 characters, and the number of named subpatterns
3354        to 10,000.
3355    
3356    11. Subpatterns that are repeated with specific counts have to be replicated in
3357        the compiled pattern. The size of memory for this was computed from the
3358        length of the subpattern and the repeat count. The latter is limited to
3359        65535, but there was no limit on the former, meaning that integer overflow
3360        could in principle occur. The compiled length of a repeated subpattern is
3361        now limited to 30,000 bytes in order to prevent this.
3362    
3363    12. Added the optional facility to have named substrings with the same name.
3364    
3365    13. Added the ability to use a named substring as a condition, using the
3366        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3367        are numbers (not recommended). Forward references are permitted.
3368    
3369    14. Added forward references in named backreferences (if you see what I mean).
3370    
3371    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3372        pattern could run off the end of the subject. For example, the pattern
3373        "(?s)(.{1,5})"8 did this with the subject "ab".
3374    
3375    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3376        PCRE_CASELESS was set when matching characters that were quantified with ?
3377        or *.
3378    
3379    17. A character class other than a single negated character that had a minimum
3380        but no maximum quantifier - for example [ab]{6,} - was not handled
3381        correctly by pce_dfa_exec(). It would match only one character.
3382    
3383    18. A valid (though odd) pattern that looked like a POSIX character
3384        class but used an invalid character after [ (for example [[,abc,]]) caused
3385        pcre_compile() to give the error "Failed: internal error: code overflow" or
3386        in some cases to crash with a glibc free() error. This could even happen if
3387        the pattern terminated after [[ but there just happened to be a sequence of
3388        letters, a binary zero, and a closing ] in the memory that followed.
3389    
3390    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3391        over the years. Originally (before any Unicode support), just the bottom 8
3392        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3393        output from "man perlunicode" includes this:
3394    
3395          The regular expression compiler produces polymorphic opcodes.  That
3396          is, the pattern adapts to the data and automatically switches to
3397          the Unicode character scheme when presented with Unicode data--or
3398          instead uses a traditional byte scheme when presented with byte
3399          data.
3400    
3401        Sadly, a wide octal escape does not cause a switch, and in a string with
3402        no other multibyte characters, these octal escapes are treated as before.
3403        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3404        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3405        Unicode string.
3406    
3407        I have not perpetrated such confusion in PCRE. Up till now, it took just
3408        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3409        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3410        translate to the appropriate multibyte character.
3411    
3412    29. Applied some refactoring to reduce the number of warnings from Microsoft
3413        and Borland compilers. This has included removing the fudge introduced
3414        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3415        a warning about an unused variable.
3416    
3417    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3418        characters since release 4.0, because Perl (from release 5.004) does not.
3419        [Or at least, is documented not to: some releases seem to be in conflict
3420        with the documentation.] However, when a pattern was studied with
3421        pcre_study() and all its branches started with \s, PCRE still included VT
3422        as a possible starting character. Of course, this did no harm; it just
3423        caused an unnecessary match attempt.
3424    
3425    22. Removed a now-redundant internal flag bit that recorded the fact that case
3426        dependency changed within the pattern. This was once needed for "required
3427        byte" processing, but is no longer used. This recovers a now-scarce options
3428        bit. Also moved the least significant internal flag bit to the most-
3429        significant bit of the word, which was not previously used (hangover from
3430        the days when it was an int rather than a uint) to free up another bit for
3431        the future.
3432    
3433    23. Added support for CRLF line endings as well as CR and LF. As well as the
3434        default being selectable at build time, it can now be changed at runtime
3435        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3436        specify that it is scanning data with non-default line endings.
3437    
3438    24. Changed the definition of CXXLINK to make it agree with the definition of
3439        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3440    
3441    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3442        recursions. This makes a big different to stack usage for some patterns.
3443    
3444    26. If a subpattern containing a named recursion or subroutine reference such
3445        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3446        the space required for the compiled pattern went wrong and gave too small a
3447        value. Depending on the environment, this could lead to "Failed: internal
3448        error: code overflow at offset 49" or "glibc detected double free or
3449        corruption" errors.
3450    
3451    27. Applied patches from Google (a) to support the new newline modes and (b) to
3452        advance over multibyte UTF-8 characters in GlobalReplace.
3453    
3454    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3455        difference for some implementation of PCRE in some Windows version.
3456    
3457    29. Added some extra testing facilities to pcretest:
3458    
3459        \q<number>   in a data line sets the "match limit" value
3460        \Q<number>   in a data line sets the "match recursion limt" value
3461        -S <number>  sets the stack size, where <number> is in megabytes
3462    
3463        The -S option isn't available for Windows.
3464    
3465    
3466    Version 6.6 06-Feb-06
3467    ---------------------
3468    
3469     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3470        in pcreposix.h. I have copied the definition from pcre.h.
3471    
3472     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3473        because pcre.h is no longer a built file.
3474    
3475     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3476        not normally included in the compiled code.
3477    
3478    
3479    Version 6.5 01-Feb-06
3480    ---------------------
3481    
3482     1. When using the partial match feature with pcre_dfa_exec(), it was not
3483        anchoring the second and subsequent partial matches at the new starting
3484        point. This could lead to incorrect results. For example, with the pattern
3485        /1234/, partially matching against "123" and then "a4" gave a match.
3486    
3487     2. Changes to pcregrep:
3488    
3489        (a) All non-match returns from pcre_exec() were being treated as failures
3490            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3491            error message is output. Some extra information is given for the
3492            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3493            probably the only errors that are likely to be caused by users (by
3494            specifying a regex that has nested indefinite repeats, for instance).
3495            If there are more than 20 of these errors, pcregrep is abandoned.
3496    
3497        (b) A binary zero was treated as data while matching, but terminated the
3498            output line if it was written out. This has been fixed: binary zeroes
3499            are now no different to any other data bytes.
3500    
3501        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3502            used to set a locale for matching. The --locale=xxxx long option has
3503            been added (no short equivalent) to specify a locale explicitly on the
3504            pcregrep command, overriding the environment variables.
3505    
3506        (d) When -B was used with -n, some line numbers in the output were one less
3507            than they should have been.
3508    
3509        (e) Added the -o (--only-matching) option.
3510    
3511        (f) If -A or -C was used with -c (count only), some lines of context were
3512            accidentally printed for the final match.
3513    
3514        (g) Added the -H (--with-filename) option.
3515    
3516        (h) The combination of options -rh failed to suppress file names for files
3517            that were found from directory arguments.
3518    
3519        (i) Added the -D (--devices) and -d (--directories) options.
3520    
3521        (j) Added the -F (--fixed-strings) option.
3522    
3523        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3524    
3525        (l) Added the --colo(u)r option.
3526    
3527        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3528            is not present by default.
3529    
3530     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3531        items such as (?R) or (?1), when the recursion could match a number of
3532        alternatives. If it matched one of the alternatives, but subsequently,
3533        outside the recursion, there was a failure, the code tried to back up into
3534        the recursion. However, because of the way PCRE is implemented, this is not
3535        possible, and the result was an incorrect result from the match.
3536    
3537        In order to prevent this happening, the specification of recursion has
3538        been changed so that all such subpatterns are automatically treated as
3539        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3540    
3541     4. I had overlooked the fact that, in some locales, there are characters for
3542        which isalpha() is true but neither isupper() nor islower() are true. In
3543        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3544        and ordfeminine) are like this. This affected the treatment of \w and \W
3545        when they appeared in character classes, but not when they appeared outside
3546        a character class. The bit map for "word" characters is now created
3547        separately from the results of isalnum() instead of just taking it from the
3548        upper, lower, and digit maps. (Plus the underscore character, of course.)
3549    
3550     5. The above bug also affected the handling of POSIX character classes such as
3551        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3552        permanent tables. Instead, the bit maps for such a class were previously
3553        created as the appropriate unions of the upper, lower, and digit bitmaps.
3554        Now they are created by subtraction from the [[:word:]] class, which has
3555        its own bitmap.
3556    
3557     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3558        It is created by subtracting the vertical space characters (\x09, \x0a,
3559        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3560        subtraction was done in the overall bitmap for a character class, meaning
3561        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3562        be recognized. This bug has been fixed.
3563    
3564     7. Patches from the folks at Google:
3565    
3566          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3567          real life, but is still worth protecting against".
3568    
3569          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3570          regular expressions".
3571    
3572          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3573          have it.
3574    
3575          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3576          "configure" and the latter not, in order to fix a problem somebody had
3577          with compiling the Arg class on HP-UX.
3578    
3579          (e) Improve the error-handling of the C++ wrapper a little bit.
3580    
3581          (f) New tests for checking recursion limiting.
3582    
3583     8. The pcre_memmove() function, which is used only if the environment does not
3584        have a standard memmove() function (and is therefore rarely compiled),
3585        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3586        returning a result (though PCRE never actually uses the result).
3587    
3588     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3589        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3590        returned instead of calling malloc() with an overflowing number that would
3591        most likely cause subsequent chaos.
3592    
3593    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3594    
3595    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3596        with this option is matched, the nmatch and pmatch options of regexec() are
3597        ignored.
3598    
3599    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3600        provided in case anyone wants to the the POSIX interface with UTF-8
3601        strings.
3602    
3603    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3604        C++ linking (needed for some HP-UX environments).
3605    
3606    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3607        (unused parameter) and in the pcre_printint() function (omitted "default"
3608        switch label when the default is to do nothing).
3609    
3610    15. Added some code to make it possible, when PCRE is compiled as a C++
3611        library, to replace subject pointers for pcre_exec() with a smart pointer
3612        class, thus making it possible to process discontinuous strings.
3613    
3614    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3615        much the same function. They were added by different people who were trying
3616        to make PCRE easy to compile on non-Unix systems. It has been suggested
3617        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3618        for compiling on Windows systems. I have therefore replaced it with
3619        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3620        defaults to "extern" for C or "extern C" for C++, which works fine on
3621        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3622        SCOPE with something explicit in config.h. In addition:
3623    
3624        (a) pcreposix.h still had just "extern" instead of either of these macros;
3625            I have replaced it with PCRE_DATA_SCOPE.
3626    
3627        (b) Functions such as _pcre_xclass(), which are internal to the library,
3628            but external in the C sense, all had PCRE_EXPORT in their definitions.
3629            This is apparently wrong for the Windows case, so I have removed it.
3630            (It makes no difference on Unix-like systems.)
3631    
3632    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3633        of recursive calls to match(). This is different to MATCH_LIMIT because
3634        that limits the total number of calls to match(), not all of which increase
3635        the depth of recursion. Limiting the recursion depth limits the amount of
3636        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3637        when PCRE is compiled, and changed at run time. A patch from Google adds
3638        this functionality to the C++ interface.
3639    
3640    18. Changes to the handling of Unicode character properties:
3641    
3642        (a) Updated the table to Unicode 4.1.0.
3643    
3644        (b) Recognize characters that are not in the table as "Cn" (undefined).
3645    
3646        (c) I revised the way the table is implemented to a much improved format
3647            which includes recognition of ranges. It now supports the ranges that
3648            are defined in UnicodeData.txt, and it also amalgamates other
3649            characters into ranges. This has reduced the number of entries in the
3650            table from around 16,000 to around 3,000, thus reducing its size
3651            considerably. I realized I did not need to use a tree structure after
3652            all - a binary chop search is just as efficient. Having reduced the
3653            number of entries, I extended their size from 6 bytes to 8 bytes to
3654            allow for more data.
3655    
3656        (d) Added support for Unicode script names via properties such as \p{Han}.
3657    
3658    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3659        matching that character.
3660    
3661    20. When matching a repeated Unicode property with a minimum greater than zero,
3662        (for example \pL{2,}), PCRE could look past the end of the subject if it
3663        reached it while seeking the minimum number of characters. This could
3664        happen only if some of the characters were more than one byte long, because
3665        there is a check for at least the minimum number of bytes.
3666    
3667    21. Refactored the implementation of \p and \P so as to be more general, to
3668        allow for more different types of property in future. This has changed the
3669        compiled form incompatibly. Anybody with saved compiled patterns that use
3670        \p or \P will have to recompile them.
3671    
3672    22. Added "Any" and "L&" to the supported property types.
3673    
3674    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3675        but give a compile time error if the value is greater than 0xff.
3676    
3677    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3678        accidentally not being installed or uninstalled.
3679    
3680    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3681        made were to insert the current release number. This seemed silly, because
3682        it made things harder for people building PCRE on systems that don't run
3683        "configure". I have turned pcre.h into a distributed file, no longer built
3684        by "configure", with the version identification directly included. There is
3685        no longer a pcre.h.in file.
3686    
3687        However, this change necessitated a change to the pcre-config script as
3688        well. It is built from pcre-config.in, and one of the substitutions was the
3689        release number. I have updated configure.ac so that ./configure now finds
3690        the release number by grepping pcre.h.
3691    
3692    26. Added the ability to run the tests under valgrind.
3693    
3694    
3695    Version 6.4 05-Sep-05
3696    ---------------------
3697    
3698     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3699        "--" to be printed when multiple files were scanned, even when none of the
3700        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3701        consider it to be a bug, and have restored the previous behaviour.
3702    
3703     2. A couple of code tidies to get rid of compiler warnings.
3704    
3705     3. The pcretest program used to cheat by referring to symbols in the library
3706        whose names begin with _pcre_. These are internal symbols that are not
3707        really supposed to be visible externally, and in some environments it is
3708        possible to suppress them. The cheating is now confined to including
3709        certain files from the library's source, which is a bit cleaner.
3710    
3711     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3712        file's purpose clearer.
3713    
3714     5. Reorganized pcre_ucp_findchar().
3715    
3716    
3717    Version 6.3 15-Aug-05
3718    ---------------------
3719    
3720     1. The file libpcre.pc.in did not have general read permission in the tarball.
3721    
3722     2. There were some problems when building without C++ support:
3723    
3724        (a) If C++ support was not built, "make install" and "make test" still
3725            tried to test it.
3726    
3727        (b) There were problems when the value of CXX was explicitly set. Some
3728            changes have been made to try to fix these, and ...
3729    
3730        (c) --disable-cpp can now be used to explicitly disable C++ support.
3731    
3732        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3733            backslash in a target when C++ was disabled. This confuses some
3734            versions of "make", apparently. Using an intermediate variable solves
3735            this. (Same for CPP_LOBJ.)
3736    
3737     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3738        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3739        necessary on certain architectures.
3740    
3741     4. Added a setting of -export-symbols-regex to the link command to remove
3742        those symbols that are exported in the C sense, but actually are local
3743        within the library, and not documented. Their names all begin with
3744        "_pcre_". This is not a perfect job, because (a) we have to except some
3745        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3746        available (and never for static libraries). I have made a note to try to
3747        find a way round (a) in the future.
3748    
3749    
3750    Version 6.2 01-Aug-05
3751    ---------------------
3752    
3753     1. There was no test for integer overflow of quantifier values. A construction
3754        such as {1111111111111111} would give undefined results. What is worse, if
3755        a minimum quantifier for a parenthesized subpattern overflowed and became
3756        negative, the calculation of the memory size went wrong. This could have
3757        led to memory overwriting.
3758    
3759     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3760    
3761     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3762        operating environments where this matters.
3763    
3764     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3765        PCRE options from within the C++ wrapper.
3766    
3767     5. Named capturing subpatterns were not being correctly counted when a pattern
3768        was compiled. This caused two problems: (a) If there were more than 100
3769        such subpatterns, the calculation of the memory needed for the whole
3770        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3771        back references of the form \12, where the number was greater than 9, were
3772        not recognized as back references, even though there were sufficient
3773        previous subpatterns.
3774    
3775     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3776        versions of gcc, e.g. 2.95.4.
3777    
3778    
3779    Version 6.1 21-Jun-05
3780    ---------------------
3781    
3782     1. There was one reference to the variable "posix" in pcretest.c that was not
3783        surrounded by "#if !defined NOPOSIX".
3784    
3785     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3786        the cross-check on the old pcre_info() function, for the benefit of the
3787        cut-down version of PCRE that is currently imported into Exim.
3788    
3789     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3790        allocation error. I've done the easy fix, which wastes 2 bytes for sensible