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

Diff of /code/trunk/ChangeLog

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

revision 37 by nigel, Sat Feb 24 21:39:09 2007 UTC revision 1465 by ph10, Fri Mar 14 15:58:19 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.35-RC1 14-March-2014
5    ------------------------------
6    
7    1.  A new flag is set, when property checks are present in an XCLASS.
8        When this flag is not set, PCRE can perform certain optimizations
9        such as studying these XCLASS-es.
10    
11    2.  The auto-possessification of character sets were improved: a normal
12        and an extended character set can be compared now. Furthermore
13        the JIT compiler optimizes more character set checks.
14    
15    3.  Got rid of some compiler warnings for potentially uninitialized variables
16        that show up only when compiled with -O2.
17    
18    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
19        of a match later then the end of the match. The pcretest program was not
20        handling the case sensibly - it was outputting from the start to the next
21        binary zero. It now reports this situation in a message, and outputs the
22        text from the end to the start.
23    
24    5.  Fast forward search is improved in JIT. Instead of the first three
25        characters, any three characters with fixed position can be searched.
26        Search order: first, last, middle.
27    
28    6.  Improve character range checks in JIT. Characters are read by an inprecise
29        function now, which returns with an unknown value if the character code is
30        above a certain treshold (e.g: 256). The only limitation is that the value
31        must be bigger than the treshold as well. This function is useful, when
32        the characters above the treshold are handled in the same way.
33    
34    7.  The macros whose names start with RAWUCHAR are placeholders for a future
35        mode in which only the bottom 21 bits of 32-bit data items are used. To
36        make this more memorable for those maintaining the code, the names have
37        been changed to start with UCHAR21, and an extensive comment has been added
38        to their definition.
39    
40    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
41        to the export list in Makefile.am (they were accidentally omitted from the
42        8.34 tarball).
43    
44    9.  The informational output from pcretest used the phrase "starting byte set"
45        which is inappropriate for the 16-bit and 32-bit libraries. As the output
46        for "first char" and "need char" really means "non-UTF-char", I've changed
47        "byte" to "char", and slightly reworded the output. The documentation about
48        these values has also been (I hope) clarified.
49    
50    10. Another JIT related optimization: use table jumps for selecting the correct
51        backtracking path, when more than four alternatives are present inside a
52        bracket.
53    
54    11. Empty match is not possible, when the minimum length is greater than zero,
55        and there is no \K in the pattern. JIT should avoid empty match checks in
56        such cases.
57    
58    12. In a caseless character class with UCP support, when a character with more
59        than one alternative case was not the first character of a range, not all
60        the alternative cases were added to the class. For example, s and \x{17f}
61        are both alternative cases for S: the class [RST] was handled correctly,
62        but [R-T] was not.
63    
64    13. The configure.ac file always checked for pthread support when JIT was
65        enabled. This is not used in Windows, so I have put this test inside a
66        check for the presence of windows.h (which was already tested for).
67    
68    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
69        The algorithm provides a way to skip certain starting offsets, and usually
70        faster than linear prefix searches.
71    
72    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
73        as for 'fr_FR' and 'french'. For some reason, however, it then used the
74        Windows-specific input and output files, which have 'french' screwed in.
75        So this could never have worked. One of the problems with locales is that
76        they aren't always the same. I have now updated RunTest so that it checks
77        the output of the locale test (test 3) against three different output
78        files, and it allows the test to pass if any one of them matches. With luck
79        this should make the test pass on some versions of Solaris where it was
80        failing. Because of the uncertainty, the script did not used to stop if
81        test 3 failed; it now does. If further versions of a French locale ever
82        come to light, they can now easily be added.
83    
84    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
85        there was a message during ./configure, but it did not stop. This now
86        provokes an error. The invalid example in README has been corrected.
87        If a value less than the minimum is given, the minimum value has always
88        been used, but now a warning is given.
89    
90    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
91        was a bug in the test system, which is now fixed. Also, the list of various
92        configurations that are tested for each release did not have one with both
93        16/32 bits and --enable-bar-anycrlf. It now does.
94    
95    18. pcretest was missing "-C bsr" for displaying the \R default setting.
96    
97    19. Little endian PowerPC systems are supported now by the JIT compiler.
98    
99    20. The fast forward newline mechanism could enter to an infinite loop on
100        certain invalid UTF-8 input. Although we don't support these cases
101        this issue can be fixed by a performance optimization.
102    
103    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
104        not take account if existing stack usage. There is now a new global
105        variable called pcre_stack_guard that can be set to point to an external
106        function to check stack availability. It is called at the start of
107        processing every parenthesized group.
108    
109    22. A typo in the code meant that in ungreedy mode the max/min qualifier
110        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
111        not match "ab".
112    
113    23. When UTF was disabled, the JIT program reported some incorrect compile
114        errors. These messages are silenced now.
115    
116    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
117        compiler.
118    
119    
120    Version 8.34 15-December-2013
121    -----------------------------
122    
123    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
124        executable memory. Patch inspired by Carsten Klein.
125    
126    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
127        this macro is never tested and has no effect, because the work to support
128        coverage involves only compiling and linking options and special targets in
129        the Makefile. The comment in config.h implied that defining the macro would
130        enable coverage support, which is totally false. There was also support for
131        setting this macro in the CMake files (my fault, I just copied it from
132        configure). SUPPORT_GCOV has now been removed.
133    
134    3.  Make a small performance improvement in strlen16() and strlen32() in
135        pcretest.
136    
137    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
138        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
139        these cases). There was also one in pcretest.c.
140    
141    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
142    
143    6.  In UTF mode, the code for checking whether a group could match an empty
144        string (which is used for indefinitely repeated groups to allow for
145        breaking an infinite loop) was broken when the group contained a repeated
146        negated single-character class with a character that occupied more than one
147        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
148        UTF-8 mode). The effect was undefined: the group might or might not be
149        deemed as matching an empty string, or the program might have crashed.
150    
151    7.  The code for checking whether a group could match an empty string was not
152        recognizing that \h, \H, \v, \V, and \R must match a character.
153    
154    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
155        an empty string. If it can, pcretest shows this in its information output.
156    
157    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
158        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
159        matched by pcre_exec() without using JIT:
160    
161        (a) If the rest of the pattern did not match after a maximal run of
162            grapheme clusters, the code for backing up to try with fewer of them
163            did not always back up over a full grapheme when characters that do not
164            have the modifier quality were involved, e.g. Hangul syllables.
165    
166        (b) If the match point in a subject started with modifier character, and
167            there was no match, the code could incorrectly back up beyond the match
168            point, and potentially beyond the first character in the subject,
169            leading to a segfault or an incorrect match result.
170    
171    10. A conditional group with an assertion condition could lead to PCRE
172        recording an incorrect first data item for a match if no other first data
173        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
174        first data item, and therefore matched "ca" after "c" instead of at the
175        start.
176    
177    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
178        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
179    
180    12. The source of pcregrep now includes z/OS-specific code so that it can be
181        compiled for z/OS as part of the special z/OS distribution.
182    
183    13. Added the -T and -TM options to pcretest.
184    
185    14. The code in pcre_compile.c for creating the table of named capturing groups
186        has been refactored. Instead of creating the table dynamically during the
187        actual compiling pass, the information is remembered during the pre-compile
188        pass (on the stack unless there are more than 20 named groups, in which
189        case malloc() is used) and the whole table is created before the actual
190        compile happens. This has simplified the code (it is now nearly 150 lines
191        shorter) and prepared the way for better handling of references to groups
192        with duplicate names.
193    
194    15. A back reference to a named subpattern when there is more than one of the
195        same name now checks them in the order in which they appear in the pattern.
196        The first one that is set is used for the reference. Previously only the
197        first one was inspected. This change makes PCRE more compatible with Perl.
198    
199    16. Unicode character properties were updated from Unicode 6.3.0.
200    
201    17. The compile-time code for auto-possessification has been refactored, based
202        on a patch by Zoltan Herczeg. It now happens after instead of during
203        compilation. The code is cleaner, and more cases are handled. The option
204        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
205        options in pcretest are provided to set it. It can also be set by
206        (*NO_AUTO_POSSESS) at the start of a pattern.
207    
208    18. The character VT has been added to the default ("C" locale) set of
209        characters that match \s and are generally treated as white space,
210        following this same change in Perl 5.18. There is now no difference between
211        "Perl space" and "POSIX space". Whether VT is treated as white space in
212        other locales depends on the locale.
213    
214    19. The code for checking named groups as conditions, either for being set or
215        for being recursed, has been refactored (this is related to 14 and 15
216        above). Processing unduplicated named groups should now be as fast at
217        numerical groups, and processing duplicated groups should be faster than
218        before.
219    
220    20. Two patches to the CMake build system, by Alexander Barkov:
221    
222          (1) Replace the "source" command by "." in CMakeLists.txt because
223              "source" is a bash-ism.
224    
225          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
226              without these the CMake build does not work on Solaris.
227    
228    21. Perl has changed its handling of \8 and \9. If there is no previously
229        encountered capturing group of those numbers, they are treated as the
230        literal characters 8 and 9 instead of a binary zero followed by the
231        literals. PCRE now does the same.
232    
233    22. Following Perl, added \o{} to specify codepoints in octal, making it
234        possible to specify values greater than 0777 and also making them
235        unambiguous.
236    
237    23. Perl now gives an error for missing closing braces after \x{... instead of
238        treating the string as literal. PCRE now does the same.
239    
240    24. RunTest used to grumble if an inappropriate test was selected explicitly,
241        but just skip it when running all tests. This make it awkward to run ranges
242        of tests when one of them was inappropriate. Now it just skips any
243        inappropriate tests, as it always did when running all tests.
244    
245    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
246        character types such as \d or \w, too many callouts were inserted, and the
247        data that they returned was rubbish.
248    
249    26. In UCP mode, \s was not matching two of the characters that Perl matches,
250        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
251        were matched by \h. The code has now been refactored so that the lists of
252        the horizontal and vertical whitespace characters used for \h and \v (which
253        are defined only in one place) are now also used for \s.
254    
255    27. Add JIT support for the 64 bit TileGX architecture.
256        Patch by Jiong Wang (Tilera Corporation).
257    
258    28. Possessive quantifiers for classes (both explicit and automatically
259        generated) now use special opcodes instead of wrapping in ONCE brackets.
260    
261    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
262        (because it's meaningless), this was not happening when PCRE_CASELESS was
263        set. Not wrong, but inefficient.
264    
265    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
266        properties for \w, \d, etc) is present in a test regex. Otherwise if the
267        test contains no characters greater than 255, Perl doesn't realise it
268        should be using Unicode semantics.
269    
270    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
271        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
272        does in Unicode mode.
273    
274    32. Added the "forbid" facility to pcretest so that putting tests into the
275        wrong test files can sometimes be quickly detected.
276    
277    33. There is now a limit (default 250) on the depth of nesting of parentheses.
278        This limit is imposed to control the amount of system stack used at compile
279        time. It can be changed at build time by --with-parens-nest-limit=xxx or
280        the equivalent in CMake.
281    
282    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
283        errors. Perl warns for these when in warning mode, but PCRE has no facility
284        for giving warnings.
285    
286    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
287        However, this was not working for (?!) because it is optimized to (*FAIL),
288        for which PCRE does not allow quantifiers. The optimization is now disabled
289        when a quantifier follows (?!). I can't see any use for this, but it makes
290        things uniform.
291    
292    36. Perl no longer allows group names to start with digits, so I have made this
293        change also in PCRE. It simplifies the code a bit.
294    
295    37. In extended mode, Perl ignores spaces before a + that indicates a
296        possessive quantifier. PCRE allowed a space before the quantifier, but not
297        before the possessive +. It now does.
298    
299    38. The use of \K (reset reported match start) within a repeated possessive
300        group such as (a\Kb)*+ was not working.
301    
302    40. Document that the same character tables must be used at compile time and
303        run time, and that the facility to pass tables to pcre_exec() and
304        pcre_dfa_exec() is for use only with saved/restored patterns.
305    
306    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
307        features for Builds with MSVC:
308    
309        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
310           builds). The .rc files can be used to set FileDescription and many other
311           attributes.
312    
313        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
314           This allows higher-level build scripts which want .pdb files to avoid
315           hard-coding the exact files needed."
316    
317    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
318        mean "start of word" and "end of word", respectively, as a transition aid.
319    
320    43. A minimizing repeat of a class containing codepoints greater than 255 in
321        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
322        compiled to use the heap for recursion.
323    
324    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
325        is configured.
326    
327    
328    Version 8.33 28-May-2013
329    ------------------------
330    
331    1.  Added 'U' to some constants that are compared to unsigned integers, to
332        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
333        variables that are added to signed variables, to ensure the result is
334        signed and can be negated.
335    
336    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
337        CMake config files.
338    
339    3.  Revise the creation of config.h.generic so that all boolean macros are
340        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
341        overriding via -D on the command line possible.
342    
343    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
344        to unsigned int is reported to make a quite noticeable speed difference in
345        a specific Windows environment. Testing on Linux did also appear to show
346        some benefit (and it is clearly not harmful). Also fixed the definition of
347        Xop which should be unsigned.
348    
349    5.  Related to (4), changing the definition of the intermediate variable cc
350        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
351        improvements.
352    
353    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
354        unnecessary spaces.
355    
356    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
357        and later.
358    
359    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
360    
361    9.  Optimizing fast_forward_start_bits in JIT.
362    
363    10. Adding support for callouts in JIT, and fixing some issues revealed
364        during this work. Namely:
365    
366        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
367    
368        (b) Minimum length was not checked before the matching is started.
369    
370    11. The value of capture_last that is passed to callouts was incorrect in some
371        cases when there was a capture on one path that was subsequently abandoned
372        after a backtrack. Also, the capture_last value is now reset after a
373        recursion, since all captures are also reset in this case.
374    
375    12. The interpreter no longer returns the "too many substrings" error in the
376        case when an overflowing capture is in a branch that is subsequently
377        abandoned after a backtrack.
378    
379    13. In the pathological case when an offset vector of size 2 is used, pcretest
380        now prints out the matched string after a yield of 0 or 1.
381    
382    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
383        Only supported by the JIT compiler at the moment.
384    
385    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
386    
387    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
388        offset of the starting point of the matching process, provided the offsets
389        vector is large enough.
390    
391    17. The \A escape now records a lookbehind value of 1, though its execution
392        does not actually inspect the previous character. This is to ensure that,
393        in partial multi-segment matching, at least one character from the old
394        segment is retained when a new segment is processed. Otherwise, if there
395        are no lookbehinds in the pattern, \A might match incorrectly at the start
396        of a new segment.
397    
398    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
399    
400    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
401        gives some modest performance improvement in 8-bit mode.
402    
403    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
404        be expressed in certain programming languages using Universal Character
405        Names.
406    
407    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
408        which points out that "non characters" are not "characters that may not
409        appear in Unicode strings" but rather "characters that are reserved for
410        internal use and have only local meaning".
411    
412    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
413        there was a conditional group that depended on an assertion, if the
414        assertion was false, the callout that immediately followed the alternation
415        in the condition was skipped when pcre_exec() was used for matching.
416    
417    23. Allow an explicit callout to be inserted before an assertion that is the
418        condition for a conditional group, for compatibility with automatic
419        callouts, which always insert a callout at this point.
420    
421    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
422        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
423    
424    25. (*PRUNE) is now supported by the JIT compiler.
425    
426    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
427    
428    27. Fix the case where there are two or more SKIPs with arguments that may be
429        ignored.
430    
431    28. (*SKIP) is now supported by the JIT compiler.
432    
433    29. (*THEN) is now supported by the JIT compiler.
434    
435    30. Update RunTest with additional test selector options.
436    
437    31. The way PCRE handles backtracking verbs has been changed in two ways.
438    
439        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
440        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
441        backtracking. In some cases this makes it more Perl-compatible, but Perl's
442        rather obscure rules do not always do the same thing.
443    
444        (2) Previously, backtracking verbs were confined within assertions. This is
445        no longer the case for positive assertions, except for (*ACCEPT). Again,
446        this sometimes improves Perl compatibility, and sometimes does not.
447    
448    32. A number of tests that were in test 2 because Perl did things differently
449        have been moved to test 1, because either Perl or PCRE has changed, and
450        these tests are now compatible.
451    
452    32. Backtracking control verbs are now handled in the same way in JIT and
453        interpreter.
454    
455    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
456        contained a forward subroutine reference caused a compile error.
457    
458    34. Auto-detect and optimize limited repetitions in JIT.
459    
460    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
461        blocking (*UTF) etc.
462    
463    36. In the interpreter, maximizing pattern repetitions for characters and
464        character types now use tail recursion, which reduces stack usage.
465    
466    37. The value of the max lookbehind was not correctly preserved if a compiled
467        and saved regex was reloaded on a host of different endianness.
468    
469    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
470        of the compiled pattern block, expand the flags field from 16 to 32 bits
471        because it was almost full.
472    
473    39. Try madvise first before posix_madvise.
474    
475    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
476        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
477        original change locked out all matches of empty strings. This has been
478        changed so that one match of an empty string per line is recognized.
479        Subsequent searches on the same line (for colouring or for --only-matching,
480        for example) do not recognize empty strings.
481    
482    41. Applied a user patch to fix a number of spelling mistakes in comments.
483    
484    42. Data lines longer than 65536 caused pcretest to crash.
485    
486    43. Clarified the data type for length and startoffset arguments for pcre_exec
487        and pcre_dfa_exec in the function-specific man pages, where they were
488        explicitly stated to be in bytes, never having been updated. I also added
489        some clarification to the pcreapi man page.
490    
491    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
492        a segmentation fault.
493    
494    
495    Version 8.32 30-November-2012
496    -----------------------------
497    
498    1.  Improved JIT compiler optimizations for first character search and single
499        character iterators.
500    
501    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
502        Patch by Daniel Richard G.
503    
504    3.  Single character iterator optimizations in the JIT compiler.
505    
506    4.  Improved JIT compiler optimizations for character ranges.
507    
508    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
509        Reported by Giuseppe D'Angelo.
510    
511    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
512        of a line, was being set incorrectly in cases where .* appeared inside
513        atomic brackets at the start of a pattern, or where there was a subsequent
514        *PRUNE or *SKIP.
515    
516    7.  Improved instruction cache flush for POWER/PowerPC.
517        Patch by Daniel Richard G.
518    
519    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
520        grep:
521    
522        (a) There is now no limit to the number of patterns to be matched.
523    
524        (b) An error is given if a pattern is too long.
525    
526        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
527            are now supported.
528    
529        (d) --exclude-from and --include-from (multiple use) have been added.
530    
531        (e) Exclusions and inclusions now apply to all files and directories, not
532            just to those obtained from scanning a directory recursively.
533    
534        (f) Multiple uses of -f and --file-list are now supported.
535    
536        (g) In a Windows environment, the default for -d has been changed from
537            "read" (the GNU grep default) to "skip", because otherwise the presence
538            of a directory in the file list provokes an error.
539    
540        (h) The documentation has been revised and clarified in places.
541    
542    9.  Improve the matching speed of capturing brackets.
543    
544    10. Changed the meaning of \X so that it now matches a Unicode extended
545        grapheme cluster.
546    
547    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
548        out POSIX threads when JIT support is configured.
549    
550    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
551    
552    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
553        structure could go wrong in environments where size_t is not the same size
554        as int.
555    
556    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
557        set.
558    
559    15. The EBCDIC support had decayed; later updates to the code had included
560        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
561        general tidy up of EBCDIC-related issues, and the documentation was also
562        not quite right. There is now a test that can be run on ASCII systems to
563        check some of the EBCDIC-related things (but is it not a full test).
564    
565    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
566        in a small tidy to the code.
567    
568    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
569    
570    18. If the --only-matching (-o) option in pcregrep is specified multiple
571        times, each one causes appropriate output. For example, -o1 -o2 outputs the
572        substrings matched by the 1st and 2nd capturing parentheses. A separating
573        string can be specified by --om-separator (default empty).
574    
575    19. Improving the first n character searches.
576    
577    20. Turn case lists for horizontal and vertical white space into macros so that
578        they are defined only once.
579    
580    21. This set of changes together give more compatible Unicode case-folding
581        behaviour for characters that have more than one other case when UCP
582        support is available.
583    
584        (a) The Unicode property table now has offsets into a new table of sets of
585            three or more characters that are case-equivalent. The MultiStage2.py
586            script that generates these tables (the pcre_ucd.c file) now scans
587            CaseFolding.txt instead of UnicodeData.txt for character case
588            information.
589    
590        (b) The code for adding characters or ranges of characters to a character
591            class has been abstracted into a generalized function that also handles
592            case-independence. In UTF-mode with UCP support, this uses the new data
593            to handle characters with more than one other case.
594    
595        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
596            whose other case is greater than 256 are now correctly matched
597            caselessly. Previously, the high codepoint matched the low one, but not
598            vice versa.
599    
600        (d) The processing of \h, \H, \v, and \ in character classes now makes use
601            of the new class addition function, using character lists defined as
602            macros alongside the case definitions of 20 above.
603    
604        (e) Caseless back references now work with characters that have more than
605            one other case.
606    
607        (f) General caseless matching of characters with more than one other case
608            is supported.
609    
610    22. Unicode character properties were updated from Unicode 6.2.0
611    
612    23. Improved CMake support under Windows. Patch by Daniel Richard G.
613    
614    24. Add support for 32-bit character strings, and UTF-32
615    
616    25. Major JIT compiler update (code refactoring and bugfixing).
617        Experimental Sparc 32 support is added.
618    
619    26. Applied a modified version of Daniel Richard G's patch to create
620        pcre.h.generic and config.h.generic by "make" instead of in the
621        PrepareRelease script.
622    
623    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
624        pcre_compile.c when checking for a zero character.
625    
626    28. Introducing a native interface for JIT. Through this interface, the compiled
627        machine code can be directly executed. The purpose of this interface is to
628        provide fast pattern matching, so several sanity checks are not performed.
629        However, feature tests are still performed. The new interface provides
630        1.4x speedup compared to the old one.
631    
632    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
633        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
634        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
635    
636    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
637        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
638        the "old" RFC 2279). Instead, it ended up passing a negative length to
639        pcre_exec().
640    
641    31. Add support for GCC's visibility feature to hide internal functions.
642    
643    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
644        "unknown -C option" after outputting 0 or 1.
645    
646    33. There is now support for generating a code coverage report for the test
647        suite in environments where gcc is the compiler and lcov is installed. This
648        is mainly for the benefit of the developers.
649    
650    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
651        unaddressable using valgrind annotations, allowing valgrind to detect
652        invalid memory accesses. This is mainly for the benefit of the developers.
653    
654    25. (*UTF) can now be used to start a pattern in any of the three libraries.
655    
656    26. Give configure error if --enable-cpp but no C++ compiler found.
657    
658    
659    Version 8.31 06-July-2012
660    -------------------------
661    
662    1.  Fixing a wrong JIT test case and some compiler warnings.
663    
664    2.  Removed a bashism from the RunTest script.
665    
666    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
667        to unsigned type, result still unsigned" that was given by an MS compiler
668        on encountering the code "-sizeof(xxx)".
669    
670    4.  Partial matching support is added to the JIT compiler.
671    
672    5.  Fixed several bugs concerned with partial matching of items that consist
673        of more than one character:
674    
675        (a) /^(..)\1/ did not partially match "aba" because checking references was
676            done on an "all or nothing" basis. This also applied to repeated
677            references.
678    
679        (b) \R did not give a hard partial match if \r was found at the end of the
680            subject.
681    
682        (c) \X did not give a hard partial match after matching one or more
683            characters at the end of the subject.
684    
685        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
686            a partial match for the string "\r".
687    
688        (e) When newline was set to CRLF, the metacharacter "." did not recognize
689            a partial match for a CR character at the end of the subject string.
690    
691    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
692        running pcretest, the text "(JIT)" added to the output whenever JIT is
693        actually used to run the match.
694    
695    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
696        or /S+[+] with a digit between 1 and 7.
697    
698    8.  OP_NOT now supports any UTF character not just single-byte ones.
699    
700    9.  (*MARK) control verb is now supported by the JIT compiler.
701    
702    10. The command "./RunTest list" lists the available tests without actually
703        running any of them. (Because I keep forgetting what they all are.)
704    
705    11. Add PCRE_INFO_MAXLOOKBEHIND.
706    
707    12. Applied a (slightly modified) user-supplied patch that improves performance
708        when the heap is used for recursion (compiled with --disable-stack-for-
709        recursion). Instead of malloc and free for each heap frame each time a
710        logical recursion happens, frames are retained on a chain and re-used where
711        possible. This sometimes gives as much as 30% improvement.
712    
713    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
714        call.
715    
716    14. As documented, (*COMMIT) is now confined to within a positive assertion.
717    
718    15. It is now possible to link pcretest with libedit as an alternative to
719        libreadline.
720    
721    16. (*COMMIT) control verb is now supported by the JIT compiler.
722    
723    17. The Unicode data tables have been updated to Unicode 6.1.0.
724    
725    18. Added --file-list option to pcregrep.
726    
727    19. Added binary file support to pcregrep, including the -a, --binary-files,
728        -I, and --text options.
729    
730    20. The madvise function is renamed for posix_madvise for QNX compatibility
731        reasons. Fixed by Giuseppe D'Angelo.
732    
733    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
734    
735    22. Changed the option for creating symbolic links for 16-bit man pages from
736        -s to -sf so that re-installing does not cause issues.
737    
738    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
739    
740    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
741        matches in certain environments (the workspace was not being correctly
742        retained). Also added to pcre_dfa_exec() a simple plausibility check on
743        some of the workspace data at the beginning of a restart.
744    
745    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
746        was not doing so when it should - probably a typo introduced by SVN 528
747        (change 8.10/14).
748    
749    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
750        \w+ when the character tables indicated that \x{c4} was a word character.
751        There were several related cases, all because the tests for doing a table
752        lookup were testing for characters less than 127 instead of 255.
753    
754    27. If a pattern contains capturing parentheses that are not used in a match,
755        their slots in the ovector are set to -1. For those that are higher than
756        any matched groups, this happens at the end of processing. In the case when
757        there were back references that the ovector was too small to contain
758        (causing temporary malloc'd memory to be used during matching), and the
759        highest capturing number was not used, memory off the end of the ovector
760        was incorrectly being set to -1. (It was using the size of the temporary
761        memory instead of the true size.)
762    
763    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
764        ovector size, it uses memory at the end of the block that it has got.
765    
766    29. Check for an overlong MARK name and give an error at compile time. The
767        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
768    
769    30. JIT compiler update.
770    
771    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
772        Rill for the patch.
773    
774    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
775    
776    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
777    
778    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
779        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
780        was enabled.
781    
782    35. Improve JIT code generation for greedy plus quantifier.
783    
784    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
785        1 to "aa" instead of to an empty string. The bug affected repeated groups
786        that could potentially match an empty string.
787    
788    37. Optimizing single character iterators in JIT.
789    
790    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
791        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
792        codepoints that are too big for the mode are faulted, and in a UTF mode,
793        disallowed codepoints are also faulted.
794    
795    39. If PCRE was compiled with UTF support, in three places in the DFA
796        matcher there was code that should only have been obeyed in UTF mode, but
797        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
798        processing when bytes with values greater than 127 were present. In 16-bit
799        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
800        both cases the values are those that cannot be the first data item in a UTF
801        character. The three items that might have provoked this were recursions,
802        possessively repeated groups, and atomic groups.
803    
804    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
805        and pcregrep, because some OS require shared objects to be explicitly
806        passed to ld, causing the link step to fail if they are not.
807    
808    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
809        mode, patterns that started with \h* or \R* might be incorrectly matched.
810    
811    
812    Version 8.30 04-February-2012
813    -----------------------------
814    
815    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
816        name is defined in ctype.h.
817    
818    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
819        only in quite long subpatterns.
820    
821    3.  Removed the function pcre_info(), which has been obsolete and deprecated
822        since it was replaced by pcre_fullinfo() in February 2000.
823    
824    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
825        match a (*MARK), and the match failed at the start of the subject, a
826        reference to memory before the start of the subject could occur. This bug
827        was introduced by fix 17 of release 8.21.
828    
829    5.  A reference to an unset group with zero minimum repetition was giving
830        totally wrong answers (in non-JavaScript-compatibility mode). For example,
831        /(another)?(\1?)test/ matched against "hello world test". This bug was
832        introduced in release 8.13.
833    
834    6.  Add support for 16-bit character strings (a large amount of work involving
835        many changes and refactorings).
836    
837    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
838        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
839        from a file.
840    
841    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
842        rounding is not applied in this particular case).
843    
844    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
845        if they appear, or are escaped, in patterns.
846    
847    10. Get rid of a number of -Wunused-but-set-variable warnings.
848    
849    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
850        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
851        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
852        also returns the mark "x". This bug applied to capturing parentheses,
853        non-capturing parentheses, and atomic parentheses. It also applied to some
854        assertions.
855    
856    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
857        information out of configure.ac instead of relying on pcre.h.generic, which
858        is not stored in the repository.
859    
860    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
861        -lreadline.
862    
863    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
864    
865    15. Applied Graycode's patch to put the top-level frame on the stack rather
866        than the heap when not using the stack for recursion. This gives a
867        performance improvement in many cases when recursion is not deep.
868    
869    16. Experimental code added to "pcretest -C" to output the stack frame size.
870    
871    
872    Version 8.21 12-Dec-2011
873    ------------------------
874    
875    1.  Updating the JIT compiler.
876    
877    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
878        are added as well.
879    
880    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
881        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
882        calling _pcre_jit_exec. Some extra comments are added.
883    
884    4.  (*MARK) settings inside atomic groups that do not contain any capturing
885        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
886        was introduced by change 18 for 8.20.
887    
888    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
889        ECMA-262 standard.
890    
891    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
892        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
893        This bug was probably introduced by change 9 of 8.13.
894    
895    7.  While fixing 6 above, I noticed that a number of other items were being
896        incorrectly rejected as "not fixed length". This arose partly because newer
897        opcodes had not been added to the fixed-length checking code. I have (a)
898        corrected the bug and added tests for these items, and (b) arranged for an
899        error to occur if an unknown opcode is encountered while checking for fixed
900        length instead of just assuming "not fixed length". The items that were
901        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
902        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
903        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
904    
905    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
906        being incorrectly compiled and would have given unpredicatble results.
907    
908    9.  A possessively repeated subpattern with minimum repeat count greater than
909        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
910        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
911        the first (A) could occur when it should not.
912    
913    10. Add a cast and remove a redundant test from the code.
914    
915    11. JIT should use pcre_malloc/pcre_free for allocation.
916    
917    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
918        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
919        is anything other than /usr, -L is still shown).
920    
921    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
922    
923    14. Perl does not support \N without a following name in a [] class; PCRE now
924        also gives an error.
925    
926    15. If a forward reference was repeated with an upper limit of around 2000,
927        it caused the error "internal error: overran compiling workspace". The
928        maximum number of forward references (including repeats) was limited by the
929        internal workspace, and dependent on the LINK_SIZE. The code has been
930        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
931        the default depends on LINK_SIZE. There is a new upper limit (for safety)
932        of around 200,000 forward references. While doing this, I also speeded up
933        the filling in of repeated forward references.
934    
935    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
936        incorrectly expecting the subject to contain another "a" after the start.
937    
938    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
939        in the match, the SKIP should be ignored. This was not happening; instead
940        the SKIP was being treated as NOMATCH. For patterns such as
941        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
942        tested.
943    
944    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
945        now much more compatible with Perl, in particular in cases where the result
946        is a non-match for a non-anchored pattern. For example, if
947        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
948        "m", where previously it did not return a name. A side effect of this
949        change is that for partial matches, the last encountered mark name is
950        returned, as for non matches. A number of tests that were previously not
951        Perl-compatible have been moved into the Perl-compatible test files. The
952        refactoring has had the pleasing side effect of removing one argument from
953        the match() function, thus reducing its stack requirements.
954    
955    19. If the /S+ option was used in pcretest to study a pattern using JIT,
956        subsequent uses of /S (without +) incorrectly behaved like /S+.
957    
958    21. Retrieve executable code size support for the JIT compiler and fixing
959        some warnings.
960    
961    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
962        not work when the shorter character appeared right at the end of the
963        subject string.
964    
965    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
966        systems.
967    
968    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
969        output it when the /M option is used in pcretest.
970    
971    25. The CheckMan script was not being included in the distribution. Also, added
972        an explicit "perl" to run Perl scripts from the PrepareRelease script
973        because this is reportedly needed in Windows.
974    
975    26. If study data was being save in a file and studying had not found a set of
976        "starts with" bytes for the pattern, the data written to the file (though
977        never used) was taken from uninitialized memory and so caused valgrind to
978        complain.
979    
980    27. Updated RunTest.bat as provided by Sheri Pierce.
981    
982    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
983    
984    29. Computation of memory usage for the table of capturing group names was
985        giving an unnecessarily large value.
986    
987    
988    Version 8.20 21-Oct-2011
989    ------------------------
990    
991    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
992        a POSIX class. After further experiments with Perl, which convinced me that
993        Perl has bugs and confusions, a closing square bracket is no longer allowed
994        in a POSIX name. This bug also affected patterns with classes that started
995        with full stops.
996    
997    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
998        captured substring, but while checking the failing first alternative,
999        substring 1 is temporarily captured. If the output vector supplied to
1000        pcre_exec() was not big enough for this capture, the yield of the function
1001        was still zero ("insufficient space for captured substrings"). This cannot
1002        be totally fixed without adding another stack variable, which seems a lot
1003        of expense for a edge case. However, I have improved the situation in cases
1004        such as /(a)(b)x|abc/ matched against "abc", where the return code
1005        indicates that fewer than the maximum number of slots in the ovector have
1006        been set.
1007    
1008    3.  Related to (2) above: when there are more back references in a pattern than
1009        slots in the output vector, pcre_exec() uses temporary memory during
1010        matching, and copies in the captures as far as possible afterwards. It was
1011        using the entire output vector, but this conflicts with the specification
1012        that only 2/3 is used for passing back captured substrings. Now it uses
1013        only the first 2/3, for compatibility. This is, of course, another edge
1014        case.
1015    
1016    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1017        main code base, and can be used by building with --enable-jit. When this is
1018        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1019        runtime --no-jit option is given.
1020    
1021    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1022        ovector, the return from the function was zero, implying that there were
1023        other matches that did not fit. The correct "exactly full" value is now
1024        returned.
1025    
1026    6.  If a subpattern that was called recursively or as a subroutine contained
1027        (*PRUNE) or any other control that caused it to give a non-standard return,
1028        invalid errors such as "Error -26 (nested recursion at the same subject
1029        position)" or even infinite loops could occur.
1030    
1031    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1032        computing the minimum length on reaching *ACCEPT, and so ended up with the
1033        wrong value of 1 rather than 0. Further investigation indicates that
1034        computing a minimum subject length in the presence of *ACCEPT is difficult
1035        (think back references, subroutine calls), and so I have changed the code
1036        so that no minimum is registered for a pattern that contains *ACCEPT.
1037    
1038    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1039        it was not handled as intended. [But see 16 below.]
1040    
1041    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1042        Sheri Pierce.
1043    
1044    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1045        the first byte in a match must be "a".
1046    
1047    11. Change 17 for 8.13 increased the recursion depth for patterns like
1048        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1049        pattern contains any instances of (*THEN). If it does not, the old
1050        optimizations are restored. It would be nice to do this on a per-group
1051        basis, but at the moment that is not feasible.
1052    
1053    12. In some environments, the output of pcretest -C is CRLF terminated. This
1054        broke RunTest's code that checks for the link size. A single white space
1055        character after the value is now allowed for.
1056    
1057    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1058        For "fr", it uses the Windows-specific input and output files.
1059    
1060    14. If (*THEN) appeared in a group that was called recursively or as a
1061        subroutine, it did not work as intended. [But see next item.]
1062    
1063    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1064        pattern fragments (but not containing any | characters). If A and B are
1065        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1066        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1067        D. In other words, Perl considers parentheses that do not contain any |
1068        characters to be part of a surrounding alternative, whereas PCRE was
1069        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1070        differently. PCRE now behaves in the same way as Perl, except in the case
1071        of subroutine/recursion calls such as (?1) which have in any case always
1072        been different (but PCRE had them first :-).
1073    
1074    16. Related to 15 above: Perl does not treat the | in a conditional group as
1075        creating alternatives. Such a group is treated in the same way as an
1076        ordinary group without any | characters when processing (*THEN). PCRE has
1077        been changed to match Perl's behaviour.
1078    
1079    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1080        RunGrepTest script failed.
1081    
1082    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1083        inevitable for groups that contain captures, but it can lead to a lot of
1084        stack use in large patterns. The old behaviour has been restored for atomic
1085        groups that do not contain any capturing parentheses.
1086    
1087    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1088        suppress the check for a minimum subject length at run time. (If it was
1089        given to pcre_exec() or pcre_dfa_exec() it did work.)
1090    
1091    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1092        fail to work when decoding hex characters in data strings in EBCDIC
1093        environments.
1094    
1095    21. It appears that in at least one Mac OS environment, the isxdigit() function
1096        is implemented as a macro that evaluates to its argument more than once,
1097        contravening the C 90 Standard (I haven't checked a later standard). There
1098        was an instance in pcretest which caused it to go wrong when processing
1099        \x{...} escapes in subject strings. The has been rewritten to avoid using
1100        things like p++ in the argument of isxdigit().
1101    
1102    
1103    Version 8.13 16-Aug-2011
1104    ------------------------
1105    
1106    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1107    
1108    2.  Two minor typos in pcre_internal.h have been fixed.
1109    
1110    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1111        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1112        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1113    
1114    4.  There were a number of related bugs in the code for matching backrefences
1115        caselessly in UTF-8 mode when codes for the characters concerned were
1116        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1117        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1118        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1119        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1120        2-byte code at the end of the subject (it thought there wasn't enough data
1121        left).
1122    
1123    5.  Comprehensive information about what went wrong is now returned by
1124        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1125        as the output vector has at least 2 elements. The offset of the start of
1126        the failing character and a reason code are placed in the vector.
1127    
1128    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1129        now returned is for the first byte of the failing character, instead of the
1130        last byte inspected. This is an incompatible change, but I hope it is small
1131        enough not to be a problem. It makes the returned offset consistent with
1132        pcre_exec() and pcre_dfa_exec().
1133    
1134    7.  pcretest now gives a text phrase as well as the error number when
1135        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1136        failure, the offset and reason code are output.
1137    
1138    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1139        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1140        back over a single character (\n). This seems wrong (because it treated the
1141        two characters as a single entity when going forwards), conflicts with the
1142        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1143        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1144        has been changed.
1145    
1146    9.  Some internal refactoring has changed the processing so that the handling
1147        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1148        time (the PCRE_DOTALL option was changed this way some time ago: version
1149        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1150        which was always a bit of a fudge. It also means that there is one less
1151        argument for the match() function, which reduces its stack requirements
1152        slightly. This change also fixes an incompatibility with Perl: the pattern
1153        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1154    
1155    10. More internal refactoring has drastically reduced the number of recursive
1156        calls to match() for possessively repeated groups such as (abc)++ when
1157        using pcre_exec().
1158    
1159    11. While implementing 10, a number of bugs in the handling of groups were
1160        discovered and fixed:
1161    
1162        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1163        (a|)*(?1) gave a compile-time internal error.
1164        ((a|)+)+  did not notice that the outer group could match an empty string.
1165        (^a|^)+   was not marked as anchored.
1166        (.*a|.*)+ was not marked as matching at start or after a newline.
1167    
1168    12. Yet more internal refactoring has removed another argument from the match()
1169        function. Special calls to this function are now indicated by setting a
1170        value in a variable in the "match data" data block.
1171    
1172    13. Be more explicit in pcre_study() instead of relying on "default" for
1173        opcodes that mean there is no starting character; this means that when new
1174        ones are added and accidentally left out of pcre_study(), testing should
1175        pick them up.
1176    
1177    14. The -s option of pcretest has been documented for ages as being an old
1178        synonym of -m (show memory usage). I have changed it to mean "force study
1179        for every regex", that is, assume /S for every regex. This is similar to -i
1180        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1181        using it. It makes it easier to run collections of tests with and without
1182        study enabled, and thereby test pcre_study() more easily. All the standard
1183        tests are now run with and without -s (but some patterns can be marked as
1184        "never study" - see 20 below).
1185    
1186    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1187        restoration of the capturing data to the outer values was not happening
1188        correctly.
1189    
1190    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1191        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1192        pattern had matched an empty string, and so incorrectly returned a no
1193        match.
1194    
1195    17. There was optimizing code for the last branch of non-capturing parentheses,
1196        and also for the obeyed branch of a conditional subexpression, which used
1197        tail recursion to cut down on stack usage. Unfortunately, now that there is
1198        the possibility of (*THEN) occurring in these branches, tail recursion is
1199        no longer possible because the return has to be checked for (*THEN). These
1200        two optimizations have therefore been removed. [But see 8.20/11 above.]
1201    
1202    18. If a pattern containing \R was studied, it was assumed that \R always
1203        matched two bytes, thus causing the minimum subject length to be
1204        incorrectly computed because \R can also match just one byte.
1205    
1206    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1207        was incorrectly computed.
1208    
1209    20. If /S is present twice on a test pattern in pcretest input, it now
1210        *disables* studying, thereby overriding the use of -s on the command line
1211        (see 14 above). This is necessary for one or two tests to keep the output
1212        identical in both cases.
1213    
1214    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1215        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1216    
1217    22. When an atomic group that contained a capturing parenthesis was
1218        successfully matched, but the branch in which it appeared failed, the
1219        capturing was not being forgotten if a higher numbered group was later
1220        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1221        group 1 to "a", when in fact it should be unset. This applied to multi-
1222        branched capturing and non-capturing groups, repeated or not, and also to
1223        positive assertions (capturing in negative assertions does not happen
1224        in PCRE) and also to nested atomic groups.
1225    
1226    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1227        subject after a captured substring, to make it easier to tell which of a
1228        number of identical substrings has been captured.
1229    
1230    24. The way atomic groups are processed by pcre_exec() has been changed so that
1231        if they are repeated, backtracking one repetition now resets captured
1232        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1233        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1234        "aaa". Previously, it would have been "a". As part of this code
1235        refactoring, the way recursive calls are handled has also been changed.
1236    
1237    25. If an assertion condition captured any substrings, they were not passed
1238        back unless some other capturing happened later. For example, if
1239        (?(?=(a))a) was matched against "a", no capturing was returned.
1240    
1241    26. When studying a pattern that contained subroutine calls or assertions,
1242        the code for finding the minimum length of a possible match was handling
1243        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1244        group 1 called group 2 while simultaneously a separate group 2 called group
1245        1). A stack overflow occurred in this case. I have fixed this by limiting
1246        the recursion depth to 10.
1247    
1248    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1249        Fortmann. This supports explicit test numbers on the command line, and has
1250        argument validation and error reporting.
1251    
1252    28. An instance of \X with an unlimited repeat could fail if at any point the
1253        first character it looked at was a mark character.
1254    
1255    29. Some minor code refactoring concerning Unicode properties and scripts
1256        should reduce the stack requirement of match() slightly.
1257    
1258    30. Added the '=' option to pcretest to check the setting of unused capturing
1259        slots at the end of the pattern, which are documented as being -1, but are
1260        not included in the return count.
1261    
1262    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1263        compiled something random. Now it gives a compile-time error (as does
1264        Perl).
1265    
1266    32. A *MARK encountered during the processing of a positive assertion is now
1267        recorded and passed back (compatible with Perl).
1268    
1269    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1270        had alternative anchored branches, the search for a second match in a line
1271        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1272        matched the line "0102" twice. The same bug affected patterns that started
1273        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1274        twice.
1275    
1276    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1277        does, and because of capturing effects, quantifying parenthesized
1278        assertions may at times be useful. Quantifiers are now allowed for
1279        parenthesized assertions.
1280    
1281    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1282    
1283    36. \g was being checked for fancy things in a character class, when it should
1284        just be a literal "g".
1285    
1286    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1287        appearance of a nested POSIX class supersedes an apparent external class.
1288        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1289        unescaped square brackets may also appear as part of class names. For
1290        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1291        more like Perl. (But see 8.20/1 above.)
1292    
1293    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1294        was because it thought it was \N{name}, which is not supported).
1295    
1296    39. Add minix to OS list not supporting the -S option in pcretest.
1297    
1298    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1299        cannot analyze patterns in sufficient detail to catch mutual recursions
1300        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1301        subgroup is called recursively as a subpattern for a second time at the
1302        same position in the subject string. In previous releases this might have
1303        been caught by the recursion limit, or it might have run out of stack.
1304    
1305    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1306        happen only once. PCRE was, however incorrectly giving a compile time error
1307        "recursive call could loop indefinitely" because it cannot analyze the
1308        pattern in sufficient detail. The compile time test no longer happens when
1309        PCRE is compiling a conditional subpattern, but actual runaway loops are
1310        now caught at runtime (see 40 above).
1311    
1312    42. It seems that Perl allows any characters other than a closing parenthesis
1313        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1314        has been changed to be the same.
1315    
1316    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1317        as not to get warnings when autogen.sh is called. Also changed
1318        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1319    
1320    44. To help people who use pcregrep to scan files containing exceedingly long
1321        lines, the following changes have been made:
1322    
1323        (a) The default value of the buffer size parameter has been increased from
1324            8K to 20K. (The actual buffer used is three times this size.)
1325    
1326        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1327            PCRE is built.
1328    
1329        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1330            to be set at run time.
1331    
1332        (d) Numerical values in pcregrep options can be followed by K or M, for
1333            example --buffer-size=50K.
1334    
1335        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1336            given and the return code is set to 2.
1337    
1338    45. Add a pointer to the latest mark to the callout data block.
1339    
1340    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1341        partial match of an empty string instead of no match. This was specific to
1342        the use of ".".
1343    
1344    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1345        complete match instead of a partial match. This bug was dependent on both
1346        the PCRE_UTF8 and PCRE_DOTALL options being set.
1347    
1348    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1349        starting byte set, because \b was not being ignored.
1350    
1351    
1352    Version 8.12 15-Jan-2011
1353    ------------------------
1354    
1355    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1356        checks for such things as part of the documentation building process.
1357    
1358    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1359        --match-limit and --recursion-limit options (added for 8.11). In
1360        particular, this made one of the standard tests fail. (The integer value
1361        went into the wrong half of a long int.)
1362    
1363    3.  If the --colour option was given to pcregrep with -v (invert match), it
1364        did strange things, either producing crazy output, or crashing. It should,
1365        of course, ignore a request for colour when reporting lines that do not
1366        match.
1367    
1368    4.  Another pcregrep bug caused similar problems if --colour was specified with
1369        -M (multiline) and the pattern match finished with a line ending.
1370    
1371    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1372        matched in multiline mode, the following line was shown as part of the
1373        match. This seems wrong, so I have changed it.
1374    
1375    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1376        the check for further matches in the same line (so they could be coloured)
1377        to overrun the end of the current line. If another match was found, it was
1378        incorrectly shown (and then shown again when found in the next line).
1379    
1380    7.  If pcregrep was compiled under Windows, there was a reference to the
1381        function pcregrep_exit() before it was defined. I am assuming this was
1382        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1383        reported by a user. I've moved the definition above the reference.
1384    
1385    
1386    Version 8.11 10-Dec-2010
1387    ------------------------
1388    
1389    1.  (*THEN) was not working properly if there were untried alternatives prior
1390        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1391        backtracked to try for "b" instead of moving to the next alternative branch
1392        at the same level (in this case, to look for "c"). The Perl documentation
1393        is clear that when (*THEN) is backtracked onto, it goes to the "next
1394        alternative in the innermost enclosing group".
1395    
1396    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1397        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1398        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1399        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1400        (*THEN).
1401    
1402    3.  If \s appeared in a character class, it removed the VT character from
1403        the class, even if it had been included by some previous item, for example
1404        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1405        of \s, but is part of the POSIX "space" class.)
1406    
1407    4.  A partial match never returns an empty string (because you can always
1408        match an empty string at the end of the subject); however the checking for
1409        an empty string was starting at the "start of match" point. This has been
1410        changed to the "earliest inspected character" point, because the returned
1411        data for a partial match starts at this character. This means that, for
1412        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1413        (previously it gave "no match").
1414    
1415    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1416        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1417        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1418        has an implication that the given string is incomplete (because a partial
1419        match is preferred over a full match). For this reason, these items now
1420        give a partial match in this situation. [Aside: previously, the one case
1421        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1422        match rather than a full match, which was wrong by the old rules, but is
1423        now correct.]
1424    
1425    6.  There was a bug in the handling of #-introduced comments, recognized when
1426        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1427        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1428        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1429        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1430        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1431        places in pcre_compile().
1432    
1433    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1434        comments when looking ahead for named forward references to subpatterns,
1435        the only newline sequence it recognized was NL. It now handles newlines
1436        according to the set newline convention.
1437    
1438    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1439        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1440        cater for a lack of strerror(). These oversights have been fixed.
1441    
1442    9.  Added --match-limit and --recursion-limit to pcregrep.
1443    
1444    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1445    
1446    11. When the -o option was used, pcregrep was setting a return code of 1, even
1447        when matches were found, and --line-buffered was not being honoured.
1448    
1449    12. Added an optional parentheses number to the -o and --only-matching options
1450        of pcregrep.
1451    
1452    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1453        can match an empty string. The code to do it in pcretest and pcredemo
1454        needed fixing:
1455    
1456        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1457            only one byte after an empty string match just before CRLF (this case
1458            just got forgotten; "any" and "anycrlf" were OK).
1459    
1460        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1461            mode when an empty string match preceded an ASCII character followed by
1462            a non-ASCII character. (The code for advancing by one character rather
1463            than one byte was nonsense.)
1464    
1465        (c) The pcredemo.c sample program did not have any code at all to handle
1466            the cases when CRLF is a valid newline sequence.
1467    
1468    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1469        as a starting offset was within the subject string. There is now a new
1470        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1471        negative or greater than the length of the string. In order to test this,
1472        pcretest is extended to allow the setting of negative starting offsets.
1473    
1474    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1475        starting offset points to the beginning of a UTF-8 character was
1476        unnecessarily clumsy. I tidied it up.
1477    
1478    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1479        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1480    
1481    17. Nobody had reported that the --include_dir option, which was added in
1482        release 7.7 should have been called --include-dir (hyphen, not underscore)
1483        for compatibility with GNU grep. I have changed it to --include-dir, but
1484        left --include_dir as an undocumented synonym, and the same for
1485        --exclude-dir, though that is not available in GNU grep, at least as of
1486        release 2.5.4.
1487    
1488    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1489        characters from a string of bytes) have been redefined so as not to use
1490        loops, in order to improve performance in some environments. At the same
1491        time, I abstracted some of the common code into auxiliary macros to save
1492        repetition (this should not affect the compiled code).
1493    
1494    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1495        compile-time error is now given if \c is not followed by an ASCII
1496        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1497        different, and any byte value is allowed.)
1498    
1499    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1500        START_OPTIMIZE option, which is now allowed at compile time - but just
1501        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1502        to pcregrep and other applications that have no direct access to PCRE
1503        options. The new /Y option in pcretest sets this option when calling
1504        pcre_compile().
1505    
1506    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1507        back references. Groups containing recursive back references were forced to
1508        be atomic by that change, but in the case of named groups, the amount of
1509        memory required was incorrectly computed, leading to "Failed: internal
1510        error: code overflow". This has been fixed.
1511    
1512    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1513        pcretest.c, to avoid build problems in some Borland environments.
1514    
1515    
1516    Version 8.10 25-Jun-2010
1517    ------------------------
1518    
1519    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1520        THEN.
1521    
1522    2.  (*ACCEPT) was not working when inside an atomic group.
1523    
1524    3.  Inside a character class, \B is treated as a literal by default, but
1525        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1526        causes the error). The code is unchanged, but I tidied the documentation.
1527    
1528    4.  Inside a character class, PCRE always treated \R and \X as literals,
1529        whereas Perl faults them if its -w option is set. I have changed PCRE so
1530        that it faults them when PCRE_EXTRA is set.
1531    
1532    5.  Added support for \N, which always matches any character other than
1533        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1534    
1535    6.  When compiling pcregrep with newer versions of gcc which may have
1536        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1537        declared with attribute warn_unused_result" were given. Just casting the
1538        result to (void) does not stop the warnings; a more elaborate fudge is
1539        needed. I've used a macro to implement this.
1540    
1541    7.  Minor change to pcretest.c to avoid a compiler warning.
1542    
1543    8.  Added four artifical Unicode properties to help with an option to make
1544        \s etc use properties (see next item). The new properties are: Xan
1545        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1546    
1547    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1548        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1549        this option. Modified pcretest to add /W to test this facility. Added
1550        REG_UCP to make it available via the POSIX interface.
1551    
1552    10. Added --line-buffered to pcregrep.
1553    
1554    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1555        studied, and the match started with a letter with a code point greater than
1556        127 whose first byte was different to the first byte of the other case of
1557        the letter, the other case of this starting letter was not recognized
1558        (#976).
1559    
1560    12. If a pattern that was studied started with a repeated Unicode property
1561        test, for example, \p{Nd}+, there was the theoretical possibility of
1562        setting up an incorrect bitmap of starting bytes, but fortunately it could
1563        not have actually happened in practice until change 8 above was made (it
1564        added property types that matched character-matching opcodes).
1565    
1566    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1567        possible starting bytes for non-anchored patterns.
1568    
1569    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1570        \R, and also a number of cases that involve Unicode properties, both
1571        explicit and implicit when PCRE_UCP is set.
1572    
1573    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1574        input, it could crash or give wrong results if characters with values
1575        greater than 0xc0 were present in the subject string. (Detail: it assumed
1576        UTF-8 input when processing these items.)
1577    
1578    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1579        size_t is 64-bit (#991).
1580    
1581    17. Added a check for running out of memory when PCRE is compiled with
1582        --disable-stack-for-recursion (#990).
1583    
1584    18. If the last data line in a file for pcretest does not have a newline on
1585        the end, a newline was missing in the output.
1586    
1587    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1588        less than 128) in its various bitmaps. However, there is a facility for
1589        generating tables according to the current locale when PCRE is compiled. It
1590        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1591        characters, are recognized by isspace() and therefore were getting set in
1592        these tables, and indeed these tables seem to approximate to ISO 8859. This
1593        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1594        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1595        which of course cannot start UTF-8 characters. I have changed the code so
1596        that only real ASCII characters (less than 128) and the correct starting
1597        bytes for UTF-8 encodings are set for characters greater than 127 when in
1598        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1599        altogether.)
1600    
1601    20. Added the /T option to pcretest so as to be able to run tests with non-
1602        standard character tables, thus making it possible to include the tests
1603        used for 19 above in the standard set of tests.
1604    
1605    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1606        reference to a subpattern the other side of a comment that contains an
1607        opening parenthesis caused either an internal compiling error, or a
1608        reference to the wrong subpattern.
1609    
1610    
1611    Version 8.02 19-Mar-2010
1612    ------------------------
1613    
1614    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1615    
1616    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1617        configured.
1618    
1619    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1620        original author of that file, following a query about its status.
1621    
1622    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1623        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1624    
1625    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1626        quantifier applied to a forward-referencing subroutine call, could compile
1627        incorrect code or give the error "internal error: previously-checked
1628        referenced subpattern not found".
1629    
1630    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1631        variables to point to external functions. For these systems, therefore,
1632        pcre_malloc etc. are now initialized to local functions that call the
1633        relevant global functions.
1634    
1635    7.  There were two entries missing in the vectors called coptable and poptable
1636        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1637        I've fixed the data, and added a kludgy way of testing at compile time that
1638        the lengths are correct (equal to the number of opcodes).
1639    
1640    8.  Following on from 7, I added a similar kludge to check the length of the
1641        eint vector in pcreposix.c.
1642    
1643    9.  Error texts for pcre_compile() are held as one long string to avoid too
1644        much relocation at load time. To find a text, the string is searched,
1645        counting zeros. There was no check for running off the end of the string,
1646        which could happen if a new error number was added without updating the
1647        string.
1648    
1649    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1650    
1651    11. \K was not working if it appeared in an atomic group or in a group that
1652        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1653        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1654        the assertion is positive, but not if it is negative.
1655    
1656    12. Change 11 fortuitously reduced the size of the stack frame used in the
1657        "match()" function of pcre_exec.c by one pointer. Forthcoming
1658        implementation of support for (*MARK) will need an extra pointer on the
1659        stack; I have reserved it now, so that the stack frame size does not
1660        decrease.
1661    
1662    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1663        item in branch that calls a recursion is a subroutine call - as in the
1664        second branch in the above example - was incorrectly given the compile-
1665        time error "recursive call could loop indefinitely" because pcre_compile()
1666        was not correctly checking the subroutine for matching a non-empty string.
1667    
1668    14. The checks for overrunning compiling workspace could trigger after an
1669        overrun had occurred. This is a "should never occur" error, but it can be
1670        triggered by pathological patterns such as hundreds of nested parentheses.
1671        The checks now trigger 100 bytes before the end of the workspace.
1672    
1673    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1674    
1675    
1676    Version 8.01 19-Jan-2010
1677    ------------------------
1678    
1679    1.  If a pattern contained a conditional subpattern with only one branch (in
1680        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1681        computed the wrong minimum data length (which is of course zero for such
1682        subpatterns). This could cause incorrect "no match" results.
1683    
1684    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1685        the pattern is reset in the first branch, pcre_compile() failed with
1686        "internal error: code overflow at offset...". This happened only when
1687        the reset was to the original external option setting. (An optimization
1688        abstracts leading options settings into an external setting, which was the
1689        cause of this.)
1690    
1691    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1692        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1693        assertion pattern did not match (meaning that the assertion was true), it
1694        was incorrectly treated as false if the SKIP had been reached during the
1695        matching. This also applied to assertions used as conditions.
1696    
1697    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1698        assertion subpattern, including such a pattern used as a condition,
1699        unpredictable results occurred, instead of the error return
1700        PCRE_ERROR_DFA_UITEM.
1701    
1702    5.  The C++ GlobalReplace function was not working like Perl for the special
1703        situation when an empty string is matched. It now does the fancy magic
1704        stuff that is necessary.
1705    
1706    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1707        removed. (These were left over from very, very early versions of PCRE.)
1708    
1709    7.  Some cosmetic changes to the code to make life easier when compiling it
1710        as part of something else:
1711    
1712        (a) Change DEBUG to PCRE_DEBUG.
1713    
1714        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1715            called "current" as "current_branch", to prevent a collision with the
1716            Linux macro when compiled as a kernel module.
1717    
1718        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1719            prevent a collision with the Linux macro when compiled as a kernel
1720            module.
1721    
1722    8.  In pcre_compile() there are some checks for integer overflows that used to
1723        cast potentially large values to (double). This has been changed to that
1724        when building, a check for int64_t is made, and if it is found, it is used
1725        instead, thus avoiding the use of floating point arithmetic. (There is no
1726        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1727        double.
1728    
1729    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1730        2005 (difference between two addresses compared to an unsigned value).
1731    
1732    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1733        custom one, because of the following reported problem in Windows:
1734    
1735          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1736              under Win32.
1737          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1738              therefore missing the function definition.
1739          - The compiler thus generates a "C" signature for the test function.
1740          - The linker fails to find the "C" function.
1741          - PCRE fails to configure if asked to do so against libbz2.
1742    
1743    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1744        messages were output:
1745    
1746          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1747          rerunning libtoolize, to keep the correct libtool macros in-tree.
1748          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1749    
1750        I have done both of these things.
1751    
1752    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1753        most of the time, it *can* run out if it is given a pattern that contains a
1754        runaway infinite recursion. I updated the discussion in the pcrestack man
1755        page.
1756    
1757    13. Now that we have gone to the x.xx style of version numbers, the minor
1758        version may start with zero. Using 08 or 09 is a bad idea because users
1759        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1760        interpreted as invalid octal numbers. I've updated the previous comment in
1761        configure.ac, and also added a check that gives an error if 08 or 09 are
1762        used.
1763    
1764    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1765        causing partial matching to fail when the end of the subject matched \W
1766        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1767    
1768    15. There were some discrepancies between the declarations in pcre_internal.h
1769        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1770        their definitions. The declarations used "const uschar *" and the
1771        definitions used USPTR. Even though USPTR is normally defined as "const
1772        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1773        reported that: "This difference in casting confuses some C++ compilers, for
1774        example, SunCC recognizes above declarations as different functions and
1775        generates broken code for hbpcre." I have changed the declarations to use
1776        USPTR.
1777    
1778    16. GNU libtool is named differently on some systems. The autogen.sh script now
1779        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1780        (FreeBSD).
1781    
1782    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1783        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1784        comment: "Figure out how to create a longlong from a string: strtoll and
1785        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1786        instance, but it only takes 2 args instead of 3!"
1787    
1788    18. A subtle bug concerned with back references has been fixed by a change of
1789        specification, with a corresponding code fix. A pattern such as
1790        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1791        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1792        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1793        same bug. Such groups have to be quantified to be useful, or contained
1794        inside another quantified group. (If there's no repetition, the reference
1795        can never match.) The problem arises because, having left the group and
1796        moved on to the rest of the pattern, a later failure that backtracks into
1797        the group uses the captured value from the final iteration of the group
1798        rather than the correct earlier one. I have fixed this in PCRE by forcing
1799        any group that contains a reference to itself to be an atomic group; that
1800        is, there cannot be any backtracking into it once it has completed. This is
1801        similar to recursive and subroutine calls.
1802    
1803    
1804    Version 8.00 19-Oct-09
1805    ----------------------
1806    
1807    1.  The table for translating pcre_compile() error codes into POSIX error codes
1808        was out-of-date, and there was no check on the pcre_compile() error code
1809        being within the table. This could lead to an OK return being given in
1810        error.
1811    
1812    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1813        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1814        in a Windows environment.
1815    
1816    3.  The pcregrep --count option prints the count for each file even when it is
1817        zero, as does GNU grep. However, pcregrep was also printing all files when
1818        --files-with-matches was added. Now, when both options are given, it prints
1819        counts only for those files that have at least one match. (GNU grep just
1820        prints the file name in this circumstance, but including the count seems
1821        more useful - otherwise, why use --count?) Also ensured that the
1822        combination -clh just lists non-zero counts, with no names.
1823    
1824    4.  The long form of the pcregrep -F option was incorrectly implemented as
1825        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1826        but it seems right to fix it, and I didn't think it was worth preserving
1827        the old behaviour.
1828    
1829    5.  The command line items --regex=pattern and --regexp=pattern were not
1830        recognized by pcregrep, which required --regex pattern or --regexp pattern
1831        (with a space rather than an '='). The man page documented the '=' forms,
1832        which are compatible with GNU grep; these now work.
1833    
1834    6.  No libpcreposix.pc file was created for pkg-config; there was just
1835        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1836    
1837    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1838        when UCP support is not needed, by modifying the Python script that
1839        generates it from Unicode data files. This should not matter if the module
1840        is correctly used as a library, but I received one complaint about 50K of
1841        unwanted data. My guess is that the person linked everything into his
1842        program rather than using a library. Anyway, it does no harm.
1843    
1844    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1845        was a minimum greater than 1 for a wide character in a possessive
1846        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1847        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1848        character. Chaos in the form of incorrect output or a compiling loop could
1849        result.
1850    
1851    9.  The restrictions on what a pattern can contain when partial matching is
1852        requested for pcre_exec() have been removed. All patterns can now be
1853        partially matched by this function. In addition, if there are at least two
1854        slots in the offset vector, the offset of the earliest inspected character
1855        for the match and the offset of the end of the subject are set in them when
1856        PCRE_ERROR_PARTIAL is returned.
1857    
1858    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1859        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1860        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1861        and may be more useful for multi-segment matching.
1862    
1863    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1864        used to be given if ever the end of the subject was reached; now it is
1865        given only if matching could not proceed because another character was
1866        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1867        string "Z", which now yields "no match" instead of "partial match". In the
1868        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1869        final character ended with (*FAIL).
1870    
1871    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1872        if the pattern had a "must contain" character that was already found in the
1873        earlier partial match, unless partial matching was again requested. For
1874        example, with the pattern /dog.(body)?/, the "must contain" character is
1875        "g". If the first part-match was for the string "dog", restarting with
1876        "sbody" failed. This bug has been fixed.
1877    
1878    13. The string returned by pcre_dfa_exec() after a partial match has been
1879        changed so that it starts at the first inspected character rather than the
1880        first character of the match. This makes a difference only if the pattern
1881        starts with a lookbehind assertion or \b or \B (\K is not supported by
1882        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1883        matching functions compatible, and I think it's the right thing to do.
1884    
1885    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1886        so that the demonstration program is easily available in environments where
1887        PCRE has not been installed from source.
1888    
1889    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1890        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1891        library.
1892    
1893    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1894        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1895        is not the first non-POSIX option to be added. Clearly some people find
1896        these options useful.
1897    
1898    17. If a caller to the POSIX matching function regexec() passes a non-zero
1899        value for nmatch with a NULL value for pmatch, the value of
1900        nmatch is forced to zero.
1901    
1902    18. RunGrepTest did not have a test for the availability of the -u option of
1903        the diff command, as RunTest does. It now checks in the same way as
1904        RunTest, and also checks for the -b option.
1905    
1906    19. If an odd number of negated classes containing just a single character
1907        interposed, within parentheses, between a forward reference to a named
1908        subpattern and the definition of the subpattern, compilation crashed with
1909        an internal error, complaining that it could not find the referenced
1910        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1911        [The bug was that it was starting one character too far in when skipping
1912        over the character class, thus treating the ] as data rather than
1913        terminating the class. This meant it could skip too much.]
1914    
1915    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1916        /g option in pcretest when the pattern contains \K, which makes it possible
1917        to have an empty string match not at the start, even when the pattern is
1918        anchored. Updated pcretest and pcredemo to use this option.
1919    
1920    21. If the maximum number of capturing subpatterns in a recursion was greater
1921        than the maximum at the outer level, the higher number was returned, but
1922        with unset values at the outer level. The correct (outer level) value is
1923        now given.
1924    
1925    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1926        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1927        make it do so. The string so far is captured, making this feature
1928        compatible with Perl.
1929    
1930    23. The tests have been re-organized, adding tests 11 and 12, to make it
1931        possible to check the Perl 5.10 features against Perl 5.10.
1932    
1933    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1934        pattern matches a fixed length string. PCRE did not allow this; now it
1935        does. Neither allows recursion.
1936    
1937    25. I finally figured out how to implement a request to provide the minimum
1938        length of subject string that was needed in order to match a given pattern.
1939        (It was back references and recursion that I had previously got hung up
1940        on.) This code has now been added to pcre_study(); it finds a lower bound
1941        to the length of subject needed. It is not necessarily the greatest lower
1942        bound, but using it to avoid searching strings that are too short does give
1943        some useful speed-ups. The value is available to calling programs via
1944        pcre_fullinfo().
1945    
1946    26. While implementing 25, I discovered to my embarrassment that pcretest had
1947        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1948        study optimizations had never been tested with that matching function.
1949        Oops. What is worse, even when it was passed study data, there was a bug in
1950        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1951        were also very few tests of studied patterns with pcre_dfa_exec().
1952    
1953    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1954        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1955        on the other side of the coin, they are no longer allowed to have different
1956        names, because these cannot be distinguished in PCRE, and this has caused
1957        confusion. (This is a difference from Perl.)
1958    
1959    28. When duplicate subpattern names are present (necessarily with different
1960        numbers, as required by 27 above), and a test is made by name in a
1961        conditional pattern, either for a subpattern having been matched, or for
1962        recursion in such a pattern, all the associated numbered subpatterns are
1963        tested, and the overall condition is true if the condition is true for any
1964        one of them. This is the way Perl works, and is also more like the way
1965        testing by number works.
1966    
1967    
1968    Version 7.9 11-Apr-09
1969    ---------------------
1970    
1971    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1972        (pcretest), all targets were linked against these libraries. This included
1973        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1974        libraries. This caused unwanted dependencies to be created. This problem
1975        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1976        pcretest is linked with readline.
1977    
1978    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1979        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1980        moved outside it again, because FALSE and TRUE are already defined in AIX,
1981        but BOOL is not.
1982    
1983    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1984        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1985    
1986    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1987        hyphens) following file names and line numbers when outputting matching
1988        lines. This is not true; no spaces are inserted. I have also clarified the
1989        wording for the --colour (or --color) option.
1990    
1991    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1992        was not coloured; this is different to GNU grep, so I have changed it to be
1993        the same.
1994    
1995    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1996        each matching line was coloured. Now it goes on to look for further matches
1997        of any of the test patterns, which is the same behaviour as GNU grep.
1998    
1999    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2000        doesn't make sense to accept an empty string match in pcregrep, so I have
2001        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2002        seems to be how GNU grep behaves. [But see later change 40 for release
2003        8.33.]
2004    
2005    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2006        start or after a newline", because the conditional assertion was not being
2007        correctly handled. The rule now is that both the assertion and what follows
2008        in the first alternative must satisfy the test.
2009    
2010    9.  If auto-callout was enabled in a pattern with a conditional group whose
2011        condition was an assertion, PCRE could crash during matching, both with
2012        pcre_exec() and pcre_dfa_exec().
2013    
2014    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2015        used for matching.
2016    
2017    11. Unicode property support in character classes was not working for
2018        characters (bytes) greater than 127 when not in UTF-8 mode.
2019    
2020    12. Added the -M command line option to pcretest.
2021    
2022    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2023    
2024    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2025    
2026    16. Added comments and documentation about mis-use of no_arg in the C++
2027        wrapper.
2028    
2029    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2030        from Martin Jerabek that uses macro names for all relevant character and
2031        string constants.
2032    
2033    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2034        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2035        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2036        these, but not everybody uses configure.
2037    
2038    19. A conditional group that had only one branch was not being correctly
2039        recognized as an item that could match an empty string. This meant that an
2040        enclosing group might also not be so recognized, causing infinite looping
2041        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2042        with the subject "ab", where knowledge that the repeated group can match
2043        nothing is needed in order to break the loop.
2044    
2045    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2046        exec(), but without supplying a callout function, matching went wrong.
2047    
2048    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2049        leak if the size of the offset vector was greater than 30. When the vector
2050        is smaller, the saved offsets during recursion go onto a local stack
2051        vector, but for larger vectors malloc() is used. It was failing to free
2052        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2053        error, in fact).
2054    
2055    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2056        heapframe that is used only when UTF-8 support is enabled. This caused no
2057        problem, but was untidy.
2058    
2059    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2060        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2061        included within another project.
2062    
2063    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2064        slightly modified by me:
2065    
2066          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2067              not building pcregrep.
2068    
2069          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2070              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2071    
2072    25. Forward references, both numeric and by name, in patterns that made use of
2073        duplicate group numbers, could behave incorrectly or give incorrect errors,
2074        because when scanning forward to find the reference group, PCRE was not
2075        taking into account the duplicate group numbers. A pattern such as
2076        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2077    
2078    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2079        the feature of a custom pointer more persuasive (as requested by a user).
2080    
2081    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2082        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2083        pre-defined.
2084    
2085    28. Added support for (*UTF8) at the start of a pattern.
2086    
2087    29. Arrange for flags added by the "release type" setting in CMake to be shown
2088        in the configuration summary.
2089    
2090    
2091    Version 7.8 05-Sep-08
2092    ---------------------
2093    
2094    1.  Replaced UCP searching code with optimized version as implemented for Ad
2095        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2096        stage table and inline lookup instead of a function, giving speed ups of 2
2097        to 5 times on some simple patterns that I tested. Permission was given to
2098        distribute the MultiStage2.py script that generates the tables (it's not in
2099        the tarball, but is in the Subversion repository).
2100    
2101    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2102        scripts.
2103    
2104    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2105        a group with a zero qualifier. The result of the study could be incorrect,
2106        or the function might crash, depending on the pattern.
2107    
2108    4.  Caseless matching was not working for non-ASCII characters in back
2109        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2110        It now works when Unicode Property Support is available.
2111    
2112    5.  In pcretest, an escape such as \x{de} in the data was always generating
2113        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2114        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2115        truncation.
2116    
2117    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2118    
2119    7.  Added two (int) casts to pcregrep when printing the difference of two
2120        pointers, in case they are 64-bit values.
2121    
2122    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2123        test 2 if it fails.
2124    
2125    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2126        and a #define of that name to empty if it is not externally set. This is to
2127        allow users of MSVC to set it if necessary.
2128    
2129    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2130        the convenience functions in the pcre_get.c source file.
2131    
2132    11. An option change at the start of a pattern that had top-level alternatives
2133        could cause overwriting and/or a crash. This command provoked a crash in
2134        some environments:
2135    
2136          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2137    
2138        This potential security problem was recorded as CVE-2008-2371.
2139    
2140    12. For a pattern where the match had to start at the beginning or immediately
2141        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2142        pcre_dfa_exec() could read past the end of the passed subject if there was
2143        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2144        pcretest so that it places the subject at the end of its malloc-ed buffer.
2145    
2146    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2147        exec() might read past the end of the data buffer in UTF-8 mode.
2148    
2149    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2150        the data contained the byte 0x85 as part of a UTF-8 character within its
2151        first line. This applied both to normal and DFA matching.
2152    
2153    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2154        /^[^d]*?$/8 failed to match "abc".
2155    
2156    16. Added a missing copyright notice to pcrecpp_internal.h.
2157    
2158    17. Make it more clear in the documentation that values returned from
2159        pcre_exec() in ovector are byte offsets, not character counts.
2160    
2161    18. Tidied a few places to stop certain compilers from issuing warnings.
2162    
2163    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2164        supplied by Stefan Weber. I made a further small update for 7.8 because
2165        there is a change of source arrangements: the pcre_searchfuncs.c module is
2166        replaced by pcre_ucd.c.
2167    
2168    
2169    Version 7.7 07-May-08
2170    ---------------------
2171    
2172    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2173        a string to a long long, pretend we don't even have a long long." This is
2174        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2175    
2176    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2177        pre-7.6 versions, which defined a global no_arg variable instead of putting
2178        it in the RE class. (See also #8 below.)
2179    
2180    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2181        Lopes.
2182    
2183    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2184    
2185        (1) The include/exclude patterns were being applied to the whole pathnames
2186            of files, instead of just to the final components.
2187    
2188        (2) If there was more than one level of directory, the subdirectories were
2189            skipped unless they satisfied the include/exclude conditions. This is
2190            inconsistent with GNU grep (and could even be seen as contrary to the
2191            pcregrep specification - which I improved to make it absolutely clear).
2192            The action now is always to scan all levels of directory, and just
2193            apply the include/exclude patterns to regular files.
2194    
2195    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2196        --exclude_dir in the tests to avoid scanning .svn directories.
2197    
2198    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2199        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2200        doesn't support NULs in patterns.
2201    
2202    7.  Added some missing "const"s to declarations of static tables in
2203        pcre_compile.c and pcre_dfa_exec.c.
2204    
2205    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2206        caused by fix #2  above. (Subsequently also a second patch to fix the
2207        first patch. And a third patch - this was a messy problem.)
2208    
2209    9.  Applied Craig's patch to remove the use of push_back().
2210    
2211    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2212        matching function regexec().
2213    
2214    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2215        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2216        references. PCRE supports relative numbers with this syntax (I don't think
2217        Oniguruma does).
2218    
2219    12. Previously, a group with a zero repeat such as (...){0} was completely
2220        omitted from the compiled regex. However, this means that if the group
2221        was called as a subroutine from elsewhere in the pattern, things went wrong
2222        (an internal error was given). Such groups are now left in the compiled
2223        pattern, with a new opcode that causes them to be skipped at execution
2224        time.
2225    
2226    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2227        to the way PCRE behaves:
2228    
2229        (a) A lone ] character is dis-allowed (Perl treats it as data).
2230    
2231        (b) A back reference to an unmatched subpattern matches an empty string
2232            (Perl fails the current match path).
2233    
2234        (c) A data ] in a character class must be notated as \] because if the
2235            first data character in a class is ], it defines an empty class. (In
2236            Perl it is not possible to have an empty class.) The empty class []
2237            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2238            The negative empty class [^] matches any one character, independently
2239            of the DOTALL setting.
2240    
2241    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2242        non-existent subpattern following a character class starting with ']' and
2243        containing () gave an internal compiling error instead of "reference to
2244        non-existent subpattern". Fortunately, when the pattern did exist, the
2245        compiled code was correct. (When scanning forwards to check for the
2246        existence of the subpattern, it was treating the data ']' as terminating
2247        the class, so got the count wrong. When actually compiling, the reference
2248        was subsequently set up correctly.)
2249    
2250    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2251        it was being rejected as not supported by pcre_dfa_exec(), even though
2252        other assertions are supported. I have made pcre_dfa_exec() support
2253        (*FAIL).
2254    
2255    16. The implementation of 13c above involved the invention of a new opcode,
2256        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2257        cannot be changed at match time, I realized I could make a small
2258        improvement to matching performance by compiling OP_ALLANY instead of
2259        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2260        on the OP_ANY path.
2261    
2262    17. Compiling pcretest on Windows with readline support failed without the
2263        following two fixes: (1) Make the unistd.h include conditional on
2264        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2265    
2266    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2267        ncurses library to be included for pcretest when ReadLine support is
2268        requested, but also to allow for it to be overridden. This patch came from
2269        Daniel Bergström.
2270    
2271    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2272        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2273        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2274        spotting this.
2275    
2276    
2277    Version 7.6 28-Jan-08
2278    ---------------------
2279    
2280    1.  A character class containing a very large number of characters with
2281        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2282        overflow.
2283    
2284    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2285        HAVE_LONG_LONG is not defined.
2286    
2287    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2288        bring them up to date and include new features. This patch includes:
2289    
2290        - Fixed PH's badly added libz and libbz2 support.
2291        - Fixed a problem with static linking.
2292        - Added pcredemo. [But later removed - see 7 below.]
2293        - Fixed dftables problem and added an option.
2294        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2295            HAVE_LONG_LONG.
2296        - Added readline support for pcretest.
2297        - Added an listing of the option settings after cmake has run.
2298    
2299    4.  A user submitted a patch to Makefile that makes it easy to create
2300        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2301        Makefile.am that cause it to include this special target, without
2302        affecting anything else. Note that the same mingw target plus all
2303        the other distribution libraries and programs are now supported
2304        when configuring with CMake (see 6 below) instead of with
2305        Configure/Make.
2306    
2307    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2308        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2309        exported in the Windows port". It has not yet been confirmed that the patch
2310        solves the problem, but it does no harm.
2311    
2312    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2313        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2314        with CMake, and also correct the comment about stack recursion.
2315    
2316    7.  Remove the automatic building of pcredemo from the ./configure system and
2317        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2318        of a program that users should build themselves after PCRE is installed, so
2319        building it automatically is not really right. What is more, it gave
2320        trouble in some build environments.
2321    
2322    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2323    
2324    
2325    Version 7.5 10-Jan-08
2326    ---------------------
2327    
2328    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2329        values in parens when parsing an RE using the C++ wrapper."
2330    
2331    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2332        Characters greater than 255 were excluded from the class instead of being
2333        included.
2334    
2335    3.  The same bug as (2) above applied to negated POSIX classes such as
2336        [:^space:].
2337    
2338    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2339        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2340        I have changed it.
2341    
2342    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2343        first named subpattern) and a construct such as (?&a) would reference the
2344        first named subpattern whose name started with "a" (in other words, the
2345        length check was missing). Both these problems are fixed. "Subpattern name
2346        expected" is now given for (?&) (a zero-length name), and this patch also
2347        makes it give the same error for \k'' (previously it complained that that
2348        was a reference to a non-existent subpattern).
2349    
2350    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2351        this is right because (?- can be followed by option settings as well as by
2352        digits. I have, however, made the messages clearer.
2353    
2354    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2355        than the number used in the conditional) now cause a compile-time error.
2356        This is actually not compatible with Perl, which accepts such patterns, but
2357        treats the conditional as always being FALSE (as PCRE used to), but it
2358        seems to me that giving a diagnostic is better.
2359    
2360    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2361        and messages.
2362    
2363    9.  Fix two occurrences of "backslash" in comments that should have been
2364        "backspace".
2365    
2366    10. Remove two redundant lines of code that can never be obeyed (their function
2367        was moved elsewhere).
2368    
2369    11. The program that makes PCRE's Unicode character property table had a bug
2370        which caused it to generate incorrect table entries for sequences of
2371        characters that have the same character type, but are in different scripts.
2372        It amalgamated them into a single range, with the script of the first of
2373        them. In other words, some characters were in the wrong script. There were
2374        thirteen such cases, affecting characters in the following ranges:
2375    
2376          U+002b0 - U+002c1
2377          U+0060c - U+0060d
2378          U+0061e - U+00612
2379          U+0064b - U+0065e
2380          U+0074d - U+0076d
2381          U+01800 - U+01805
2382          U+01d00 - U+01d77
2383          U+01d9b - U+01dbf
2384          U+0200b - U+0200f
2385          U+030fc - U+030fe
2386          U+03260 - U+0327f
2387          U+0fb46 - U+0fbb1
2388          U+10450 - U+1049d
2389    
2390    12. The -o option (show only the matching part of a line) for pcregrep was not
2391        compatible with GNU grep in that, if there was more than one match in a
2392        line, it showed only the first of them. It now behaves in the same way as
2393        GNU grep.
2394    
2395    13. If the -o and -v options were combined for pcregrep, it printed a blank
2396        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2397        does the same. The return code can be used to tell if there were any
2398        non-matching lines.
2399    
2400    14. Added --file-offsets and --line-offsets to pcregrep.
2401    
2402    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2403        infinitely looping recursion. The bug was that positive lookaheads were not
2404        being skipped when checking for a possible empty match (negative lookaheads
2405        and both kinds of lookbehind were skipped).
2406    
2407    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2408        inclusion of <windows.h> to before rather than after the definition of
2409        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2410    
2411    17. Specifying a possessive quantifier with a specific limit for a Unicode
2412        character property caused pcre_compile() to compile bad code, which led at
2413        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2414        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2415        caused the error; without that there was no problem.
2416    
2417    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2418    
2419    19. Added --enable-pcretest-libreadline.
2420    
2421    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2422        RE::GlobalReplace(). As a result, the number of replacements returned was
2423        double what it should be. I removed one of the increments, but Craig sent a
2424        later patch that removed the other one (the right fix) and added unit tests
2425        that check the return values (which was not done before).
2426    
2427    21. Several CMake things:
2428    
2429        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2430            the names libpcre and libpcreposix, not just pcre and pcreposix.
2431    
2432        (2) The above change means that pcretest and pcregrep are now correctly
2433            linked with the newly-built libraries, not previously installed ones.
2434    
2435        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2436    
2437    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2438        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2439        UTF-8 newline character). The key issue is that the pattern starts .*;
2440        this means that the match must be either at the beginning, or after a
2441        newline. The bug was in the code for advancing after a failed match and
2442        checking that the new position followed a newline. It was not taking
2443        account of UTF-8 characters correctly.
2444    
2445    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2446        character classes. PCRE was not treating the sequence [:...:] as a
2447        character class unless the ... were all letters. Perl, however, seems to
2448        allow any characters between [: and :], though of course it rejects as
2449        unknown any "names" that contain non-letters, because all the known class
2450        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2451        for example, whereas PCRE did not - it did not recognize a POSIX character
2452        class. This seemed a bit dangerous, so the code has been changed to be
2453        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2454        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2455        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2456        Perl does, and where it didn't before.
2457    
2458    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2459        Windows environments %n is disabled by default.
2460    
2461    
2462    Version 7.4 21-Sep-07
2463    ---------------------
2464    
2465    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2466        means that a class such as [\s] counted as "explicit reference to CR or
2467        LF". That isn't really right - the whole point of the change was to try to
2468        help when there was an actual mention of one of the two characters. So now
2469        the change happens only if \r or \n (or a literal CR or LF) character is
2470        encountered.
2471    
2472    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2473        of both had grown to the point where there were only 3 bits left.
2474        Fortunately, there was spare space in the data structure, and so I have
2475        moved the internal flags into a new 16-bit field to free up more option
2476        bits.
2477    
2478    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2479        but did not set the internal JCHANGED flag - either of these is enough to
2480        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2481        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2482        start sets both bits.
2483    
2484    4.  Added options (at build time, compile time, exec time) to change \R from
2485        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2486    
2487    5.  doc/pcresyntax.html was missing from the distribution.
2488    
2489    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2490        compatibility, even though it is no longer used.
2491    
2492    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2493        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2494        windows.h file is present (where different names are used). [This was
2495        reversed later after testing - see 16 below.]
2496    
2497    8.  Changed all #include <config.h> to #include "config.h". There were also
2498        some further <pcre.h> cases that I changed to "pcre.h".
2499    
2500    9.  When pcregrep was used with the --colour option, it missed the line ending
2501        sequence off the lines that it output.
2502    
2503    10. It was pointed out to me that arrays of string pointers cause lots of
2504        relocations when a shared library is dynamically loaded. A technique of
2505        using a single long string with a table of offsets can drastically reduce
2506        these. I have refactored PCRE in four places to do this. The result is
2507        dramatic:
2508    
2509          Originally:                          290
2510          After changing UCP table:            187
2511          After changing error message table:   43
2512          After changing table of "verbs"       36
2513          After changing table of Posix names   22
2514    
2515        Thanks to the folks working on Gregex for glib for this insight.
2516    
2517    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2518        unicode-properties was also set.
2519    
2520    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2521    
2522    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2523        checked only for CRLF.
2524    
2525    14. Added casts to pcretest.c to avoid compiler warnings.
2526    
2527    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2528    
2529    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2530        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2531        entirely. This removes changes made in 7 above.
2532    
2533    17. The CMake files have been updated, and there is now more information about
2534        building with CMake in the NON-UNIX-USE document.
2535    
2536    
2537    Version 7.3 28-Aug-07
2538    ---------------------
2539    
2540     1. In the rejigging of the build system that eventually resulted in 7.1, the
2541        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2542        brackets there is not right, since it causes compilers to look for an
2543        installed pcre.h, not the version that is in the source that is being
2544        compiled (which of course may be different). I have changed it back to:
2545    
2546          #include "pcre.h"
2547    
2548        I have a vague recollection that the change was concerned with compiling in
2549        different directories, but in the new build system, that is taken care of
2550        by the VPATH setting the Makefile.
2551    
2552     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2553        when the subject happened to end in the byte 0x85 (e.g. if the last
2554        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2555        characters but of course it shouldn't be taken as a newline when it is part
2556        of another character. The bug was that, for an unlimited repeat of . in
2557        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2558        characters when looking for a newline.
2559    
2560     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2561    
2562     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2563        in debug output.
2564    
2565     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2566        long printing in the pcrecpp unittest when running under MinGW.
2567    
2568     6. ESC_K was left out of the EBCDIC table.
2569    
2570     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2571        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2572        limit also applies to "virtual nesting" when a pattern is recursive, and in
2573        this case 1000 isn't so big. I have been able to remove this limit at the
2574        expense of backing off one optimization in certain circumstances. Normally,
2575        when pcre_exec() would call its internal match() function recursively and
2576        immediately return the result unconditionally, it uses a "tail recursion"
2577        feature to save stack. However, when a subpattern that can match an empty
2578        string has an unlimited repetition quantifier, it no longer makes this
2579        optimization. That gives it a stack frame in which to save the data for
2580        checking that an empty string has been matched. Previously this was taken
2581        from the 1000-entry workspace that had been reserved. So now there is no
2582        explicit limit, but more stack is used.
2583    
2584     8. Applied Daniel's patches to solve problems with the import/export magic
2585        syntax that is required for Windows, and which was going wrong for the
2586        pcreposix and pcrecpp parts of the library. These were overlooked when this
2587        problem was solved for the main library.
2588    
2589     9. There were some crude static tests to avoid integer overflow when computing
2590        the size of patterns that contain repeated groups with explicit upper
2591        limits. As the maximum quantifier is 65535, the maximum group length was
2592        set at 30,000 so that the product of these two numbers did not overflow a
2593        32-bit integer. However, it turns out that people want to use groups that
2594        are longer than 30,000 bytes (though not repeat them that many times).
2595        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2596        made it possible to implement the integer overflow checks in a much more
2597        dynamic way, which I have now done. The artificial limitation on group
2598        length has been removed - we now have only the limit on the total length of
2599        the compiled pattern, which depends on the LINK_SIZE setting.
2600    
2601    10. Fixed a bug in the documentation for get/copy named substring when
2602        duplicate names are permitted. If none of the named substrings are set, the
2603        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2604        empty string.
2605    
2606    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2607        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2608        because the ] is interpreted as the first data character and the
2609        terminating ] is not found. PCRE has been made compatible with Perl in this
2610        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2611        cause memory overwriting.
2612    
2613    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2614        string has been matched (to stop an infinite loop). It was not recognizing
2615        a conditional subpattern that could match an empty string if that
2616        subpattern was within another subpattern. For example, it looped when
2617        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2618        condition was not nested. This bug has been fixed.
2619    
2620    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2621        past the start of the subject in the presence of bytes with the top bit
2622        set, for example "\x8aBCD".
2623    
2624    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2625        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2626    
2627    14. Optimized (?!) to (*FAIL).
2628    
2629    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2630        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2631        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2632        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2633        does: it's just the validity check that is more restrictive.
2634    
2635    16. Inserted checks for integer overflows during escape sequence (backslash)
2636        processing, and also fixed erroneous offset values for syntax errors during
2637        backslash processing.
2638    
2639    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2640        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2641    
2642    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2643        caused an overrun.
2644    
2645    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2646        something other than just ASCII characters) inside a group that had an
2647        unlimited repeat caused a loop at compile time (while checking to see
2648        whether the group could match an empty string).
2649    
2650    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2651        [\P{Any}] did so. (Error in the code for printing property names.)
2652    
2653    21. An orphan \E inside a character class could cause a crash.
2654    
2655    22. A repeated capturing bracket such as (A)? could cause a wild memory
2656        reference during compilation.
2657    
2658    23. There are several functions in pcre_compile() that scan along a compiled
2659        expression for various reasons (e.g. to see if it's fixed length for look
2660        behind). There were bugs in these functions when a repeated \p or \P was
2661        present in the pattern. These operators have additional parameters compared
2662        with \d, etc, and these were not being taken into account when moving along
2663        the compiled data. Specifically:
2664    
2665        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2666            length.
2667    
2668        (b) An item such as \pL+ within a repeated group could cause crashes or
2669            loops.
2670    
2671        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2672            "reference to non-existent subpattern" error.
2673    
2674        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2675    
2676    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2677        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2678    
2679    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2680    
2681    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2682        character were causing crashes (broken optimization).
2683    
2684    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2685        \p or \P) caused a compile-time loop.
2686    
2687    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2688        break. For example, the unstudied pattern [\r\n]A does not match the string
2689        "\r\nA" because change 7.0/46 below moves the current point on by two
2690        characters after failing to match at the start. However, the pattern \nA
2691        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2692        the same is true. There doesn't seem any very clean way out of this, but
2693        what I have chosen to do makes the common cases work: PCRE now takes note
2694        of whether there can be an explicit match for \r or \n anywhere in the
2695        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2696        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2697        pattern has explicit CR or LF references.
2698    
2699    29. Added (*CR) etc for changing newline setting at start of pattern.
2700    
2701    
2702    Version 7.2 19-Jun-07
2703    ---------------------
2704    
2705     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2706        which is apparently normally available under Windows.
2707    
2708     2. Re-jig the pcregrep tests with different newline settings in an attempt
2709        to make them independent of the local environment's newline setting.
2710    
2711     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2712    
2713     4. Some of the "internals" tests were previously cut out when the link size
2714        was not 2, because the output contained actual offsets. The recent new
2715        "Z" feature of pcretest means that these can be cut out, making the tests
2716        usable with all link sizes.
2717    
2718     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2719        stack recursion. This gives a massive performance boost under BSD, but just
2720        a small improvement under Linux. However, it saves one field in the frame
2721        in all cases.
2722    
2723     6. Added more features from the forthcoming Perl 5.10:
2724    
2725        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2726            recursion call. It refers to the nth most recently opened parentheses.
2727    
2728        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2729            to be opened parentheses.
2730    
2731        (c) Conditions that refer to capturing parentheses can be specified
2732            relatively, for example, (?(-2)... or (?(+3)...
2733    
2734        (d) \K resets the start of the current match so that everything before
2735            is not part of it.
2736    
2737        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2738    
2739        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2740            reference syntax.
2741    
2742        (g) (?| introduces a group in which the numbering of parentheses in each
2743            alternative starts with the same number.
2744    
2745        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2746    
2747     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2748        PCRE_INFO_JCHANGED.
2749    
2750     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2751        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2752        for detecting groups that can match an empty string.
2753    
2754     9. A pattern with a very large number of alternatives (more than several
2755        hundred) was running out of internal workspace during the pre-compile
2756        phase, where pcre_compile() figures out how much memory will be needed. A
2757        bit of new cunning has reduced the workspace needed for groups with
2758        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2759        workspace instead of running out of the 4096 that are available.
2760    
2761    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2762    
2763    11. Applied patch from Google to remove an optimization that didn't quite work.
2764        The report of the bug said:
2765    
2766          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2767          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2768          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2769    
2770    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2771        it matched the wrong number of bytes.
2772    
2773    
2774    Version 7.1 24-Apr-07
2775    ---------------------
2776    
2777     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2778        that is more "standard", making use of automake and other Autotools. There
2779        is some re-arrangement of the files and adjustment of comments consequent
2780        on this.
2781    
2782     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2783        for recursive directory scanning broke on some systems because the files
2784        are not scanned in any specific order and on different systems the order
2785        was different. A call to "sort" has been inserted into RunGrepTest for the
2786        approprate test as a short-term fix. In the longer term there may be an
2787        alternative.
2788    
2789     3. I had an email from Eric Raymond about problems translating some of PCRE's
2790        man pages to HTML (despite the fact that I distribute HTML pages, some
2791        people do their own conversions for various reasons). The problems
2792        concerned the use of low-level troff macros .br and .in. I have therefore
2793        removed all such uses from the man pages (some were redundant, some could
2794        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2795        HTML has been updated to handle .nf/.fi and to complain if it encounters
2796        .br or .in.
2797    
2798     4. Updated comments in configure.ac that get placed in config.h.in and also
2799        arranged for config.h to be included in the distribution, with the name
2800        config.h.generic, for the benefit of those who have to compile without
2801        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2802    
2803     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2804        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2805        makevp.bat for the current PCRE, using the additional files
2806        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2807    
2808     6. A Windows user reported a minor discrepancy with test 2, which turned out
2809        to be caused by a trailing space on an input line that had got lost in his
2810        copy. The trailing space was an accident, so I've just removed it.
2811    
2812     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2813        that is needed.
2814    
2815     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2816        as "const" (a) because they are and (b) because it helps the PHP
2817        maintainers who have recently made a script to detect big data structures
2818        in the php code that should be moved to the .rodata section. I remembered
2819        to update Builducptable as well, so it won't revert if ucptable.h is ever
2820        re-created.
2821    
2822     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2823        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2824        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2825        support is not required. This saves 1.5-2K of code, which is important in
2826        some applications.
2827    
2828        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2829        so as not to refer to the tables, even though these functions will never be
2830        called when UTF-8 support is disabled. Otherwise there are problems with a
2831        shared library.
2832    
2833    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2834    
2835        (a) It was defining its arguments as char * instead of void *.
2836    
2837        (b) It was assuming that all moves were upwards in memory; this was true
2838            a long time ago when I wrote it, but is no longer the case.
2839    
2840        The emulated memove() is provided for those environments that have neither
2841        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2842        is clearly not the case, as these two bugs were recently reported.
2843    
2844    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2845        and Detrail to create the HTML documentation, the .txt form of the man
2846        pages, and it removes trailing spaces from listed files. It also creates
2847        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2848        case, it wraps all the #defines with #ifndefs. This script should be run
2849        before "make dist".
2850    
2851    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2852        with Unicode property support.
2853    
2854        (a) For a maximizing quantifier, if the two different cases of the
2855            character were of different lengths in their UTF-8 codings (there are
2856            some cases like this - I found 11), and the matching function had to
2857            back up over a mixture of the two cases, it incorrectly assumed they
2858            were both the same length.
2859    
2860        (b) When PCRE was configured to use the heap rather than the stack for
2861            recursion during matching, it was not correctly preserving the data for
2862            the other case of a UTF-8 character when checking ahead for a match
2863            while processing a minimizing repeat. If the check also involved
2864            matching a wide character, but failed, corruption could cause an
2865            erroneous result when trying to check for a repeat of the original
2866            character.
2867    
2868    13. Some tidying changes to the testing mechanism:
2869    
2870        (a) The RunTest script now detects the internal link size and whether there
2871            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2872            values substituted by "configure". (The RunGrepTest script already did
2873            this for UTF-8.) The configure.ac script no longer substitutes the
2874            relevant variables.
2875    
2876        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2877            with length and offset values. This means that the output is different
2878            for different internal link sizes. Test 2 is skipped for link sizes
2879            other than 2 because of this, bypassing the problem. Unfortunately,
2880            there was also a test in test 3 (the locale tests) that used /B and
2881            failed for link sizes other than 2. Rather than cut the whole test out,
2882            I have added a new /Z option to pcretest that replaces the length and
2883            offset values with spaces. This is now used to make test 3 independent
2884            of link size. (Test 2 will be tidied up later.)
2885    
2886    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2887        segmentation fault instead of returning the appropriate error message.
2888    
2889    15. In multiline mode when the newline sequence was set to "any", the pattern
2890        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2891        This doesn't seem right; it now treats the CRLF combination as the line
2892        ending, and so does not match in that case. It's only a pattern such as ^$
2893        that would hit this one: something like ^ABC$ would have failed after \r
2894        and then tried again after \r\n.
2895    
2896    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2897        in an attempt to make files that differ only in their line terminators
2898        compare equal. This works on Linux.
2899    
2900    17. Under certain error circumstances pcregrep might try to free random memory
2901        as it exited. This is now fixed, thanks to valgrind.
2902    
2903    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2904        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2905        was because its rules for how to advance for /g after matching an empty
2906        string at the end of a line did not allow for this case. They now check for
2907        it specially.
2908    
2909    20. pcretest is supposed to handle patterns and data of any length, by
2910        extending its buffers when necessary. It was getting this wrong when the
2911        buffer for a data line had to be extended.
2912    
2913    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2914        CRLF as a newline sequence.
2915    
2916    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2917        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2918        I have nevertheless tidied it up.
2919    
2920    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2921    
2922    24. Added a man page for pcre-config.
2923    
2924    
2925    Version 7.0 19-Dec-06
2926    ---------------------
2927    
2928     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2929        moving to gcc 4.1.1.
2930    
2931     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2932        sys/time.h, which is documented as needed for this function. It doesn't
2933        seem to matter on Linux, but it showed up on some releases of OS X.
2934    
2935     3. It seems that there are systems where bytes whose values are greater than
2936        127 match isprint() in the "C" locale. The "C" locale should be the
2937        default when a C program starts up. In most systems, only ASCII printing
2938        characters match isprint(). This difference caused the output from pcretest
2939        to vary, making some of the tests fail. I have changed pcretest so that:
2940    
2941        (a) When it is outputting text in the compiled version of a pattern, bytes
2942            other than 32-126 are always shown as hex escapes.
2943    
2944        (b) When it is outputting text that is a matched part of a subject string,
2945            it does the same, unless a different locale has been set for the match
2946            (using the /L modifier). In this case, it uses isprint() to decide.
2947    
2948     4. Fixed a major bug that caused incorrect computation of the amount of memory
2949        required for a compiled pattern when options that changed within the
2950        pattern affected the logic of the preliminary scan that determines the
2951        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2952        that the computed length was too small. The symptoms of this bug were
2953        either the PCRE error "internal error: code overflow" from pcre_compile(),
2954        or a glibc crash with a message such as "pcretest: free(): invalid next
2955        size (fast)". Examples of patterns that provoked this bug (shown in
2956        pcretest format) are:
2957    
2958          /(?-x: )/x
2959          /(?x)(?-x: \s*#\s*)/
2960          /((?i)[\x{c0}])/8
2961          /(?i:[\x{c0}])/8
2962    
2963        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2964        is now done differently.
2965    
2966     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2967        wrapper classes; (b) implement a new function in the C++ scanner that is
2968        more efficient than the old way of doing things because it avoids levels of
2969        recursion in the regex matching; (c) add a paragraph to the documentation
2970        for the FullMatch() function.
2971    
2972     6. The escape sequence \n was being treated as whatever was defined as
2973        "newline". Not only was this contrary to the documentation, which states
2974        that \n is character 10 (hex 0A), but it also went horribly wrong when
2975        "newline" was defined as CRLF. This has been fixed.
2976    
2977     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2978        was being set to -1 for the "end of line" case (supposedly a value that no
2979        character can have). Though this value is never used (the check for end of
2980        line is "zero bytes in current character"), it caused compiler complaints.
2981        I've changed it to 0xffffffff.
2982    
2983     8. In pcre_version.c, the version string was being built by a sequence of
2984        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2985        string (as it is for production releases) called a macro with an empty
2986        argument. The C standard says the result of this is undefined. The gcc
2987        compiler treats it as an empty string (which was what was wanted) but it is
2988        reported that Visual C gives an error. The source has been hacked around to
2989        avoid this problem.
2990    
2991     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2992        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2993        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2994        of them did).
2995    
2996    10. Originally, pcretest opened its input and output without "b"; then I was
2997        told that "b" was needed in some environments, so it was added for release
2998        5.0 to both the input and output. (It makes no difference on Unix-like
2999        systems.) Later I was told that it is wrong for the input on Windows. I've
3000        now abstracted the modes into two macros, to make it easier to fiddle with
3001        them, and removed "b" from the input mode under Windows.
3002    
3003    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3004    
3005    12. Added -help and --help to pcretest as an official way of being reminded
3006        of the options.
3007    
3008    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3009        and pcrecpp.cc because they annoy compilers at high warning levels.
3010    
3011    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3012    
3013    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3014        scripts are not C programs :-) and which was not noticed because it works
3015        on Linux.
3016    
3017    16. pcretest is supposed to handle any length of pattern and data line (as one
3018        line or as a continued sequence of lines) by extending its input buffer if
3019        necessary. This feature was broken for very long pattern lines, leading to
3020        a string of junk being passed to pcre_compile() if the pattern was longer
3021        than about 50K.
3022    
3023    17. I have done a major re-factoring of the way pcre_compile() computes the
3024        amount of memory needed for a compiled pattern. Previously, there was code
3025        that made a preliminary scan of the pattern in order to do this. That was
3026        OK when PCRE was new, but as the facilities have expanded, it has become
3027        harder and harder to keep it in step with the real compile phase, and there
3028        have been a number of bugs (see for example, 4 above). I have now found a
3029        cunning way of running the real compile function in a "fake" mode that
3030        enables it to compute how much memory it would need, while actually only
3031        ever using a few hundred bytes of working memory and without too many
3032        tests of the mode. This should make future maintenance and development
3033        easier. A side effect of this work is that the limit of 200 on the nesting
3034        depth of parentheses has been removed (though this was never a serious
3035        limitation, I suspect). However, there is a downside: pcre_compile() now
3036        runs more slowly than before (30% or more, depending on the pattern). I
3037        hope this isn't a big issue. There is no effect on runtime performance.
3038    
3039    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3040        newline (only possible for the last line of a file) and it was a
3041        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3042    
3043    19. Added additional timing features to pcretest. (1) The -tm option now times
3044        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3045        separate command line item, by a number that specifies the number of
3046        repeats to use when timing. The default is 50000; this gives better
3047        precision, but takes uncomfortably long for very large patterns.
3048    
3049    20. Extended pcre_study() to be more clever in cases where a branch of a
3050        subpattern has no definite first character. For example, (a*|b*)[cd] would
3051        previously give no result from pcre_study(). Now it recognizes that the
3052        first character must be a, b, c, or d.
3053    
3054    21. There was an incorrect error "recursive call could loop indefinitely" if
3055        a subpattern (or the entire pattern) that was being tested for matching an
3056        empty string contained only one non-empty item after a nested subpattern.
3057        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3058        incorrectly, because the \d was being skipped in the check.
3059    
3060    22. The pcretest program now has a new pattern option /B and a command line
3061        option -b, which is equivalent to adding /B to every pattern. This causes
3062        it to show the compiled bytecode, without the additional information that
3063        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3064        is the same as /B/I).
3065    
3066    23. A new optimization is now able automatically to treat some sequences such
3067        as a*b as a*+b. More specifically, if something simple (such as a character
3068        or a simple class like \d) has an unlimited quantifier, and is followed by
3069        something that cannot possibly match the quantified thing, the quantifier
3070        is automatically "possessified".
3071    
3072    24. A recursive reference to a subpattern whose number was greater than 39
3073        went wrong under certain circumstances in UTF-8 mode. This bug could also
3074        have affected the operation of pcre_study().
3075    
3076    25. Realized that a little bit of performance could be had by replacing
3077        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3078    
3079    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3080    
3081    27. Possessive quantifiers such as a++ were previously implemented by turning
3082        them into atomic groups such as ($>a+). Now they have their own opcodes,
3083        which improves performance. This includes the automatically created ones
3084        from 23 above.
3085    
3086    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3087        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3088        the first matched character to be a colon. This applied both to named and
3089        numbered groups.
3090    
3091    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3092    
3093    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3094        building PCRE on Windows easier, so I have included it in the distribution.
3095    
3096    31. There is now a check in pcretest against a ridiculously large number being
3097        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3098        loop, the loop is abandoned.
3099    
3100    32. Forward references to subpatterns in conditions such as (?(2)...) where
3101        subpattern 2 is defined later cause pcre_compile() to search forwards in
3102        the pattern for the relevant set of parentheses. This search went wrong
3103        when there were unescaped parentheses in a character class, parentheses
3104        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3105    
3106    33. "Subroutine" calls and backreferences were previously restricted to
3107        referencing subpatterns earlier in the regex. This restriction has now
3108        been removed.
3109    
3110    34. Added a number of extra features that are going to be in Perl 5.10. On the
3111        whole, these are just syntactic alternatives for features that PCRE had
3112        previously implemented using the Python syntax or my own invention. The
3113        other formats are all retained for compatibility.
3114    
3115        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3116            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3117            also .NET compatible.
3118    
3119        (b) A recursion or subroutine call to a named group can now be defined as
3120            (?&name) as well as (?P>name).
3121    
3122        (c) A backreference to a named group can now be defined as \k<name> or
3123            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3124            5.10, are also .NET compatible.
3125    
3126        (d) A conditional reference to a named group can now use the syntax
3127            (?(<name>) or (?('name') as well as (?(name).
3128    
3129        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3130            groups (named and numbered) that are never evaluated inline, but can be
3131            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3132            is always false. There may be only one alternative in such a group.
3133    
3134        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3135            as the simple (?(R). The condition is true only if the most recent
3136            recursion is that of the given number or name. It does not search out
3137            through the entire recursion stack.
3138    
3139        (g) The escape \gN or \g{N} has been added, where N is a positive or
3140            negative number, specifying an absolute or relative reference.
3141    
3142    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3143        some "unreachable code" warnings.
3144    
3145    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3146        things, this adds five new scripts.
3147    
3148    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3149        There were also incompatibilities regarding the handling of \Q..\E inside
3150        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3151        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3152    
3153    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3154        matches an empty string, and forcibly breaks the loop. There were bugs in
3155        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3156        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3157        separate and independent bugs (that affected different cases) have been
3158        fixed.
3159    
3160    39. Refactored the code to abolish the use of different opcodes for small
3161        capturing bracket numbers. This is a tidy that I avoided doing when I
3162        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3163        The new approach is not only tidier, it makes it possible to reduce the
3164        memory needed to fix the previous bug (38).
3165    
3166    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3167        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3168        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3169        mode.
3170    
3171    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3172        report.
3173    
3174    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3175        copy construction and assignment in the C++ wrapper.
3176    
3177    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3178        couple of bugs that could have given wrong results in the "--newline=crlf"
3179        case.
3180    
3181    44. Added a number of casts and did some reorganization of signed/unsigned int
3182        variables following suggestions from Dair Grant. Also renamed the variable
3183        "this" as "item" because it is a C++ keyword.
3184    
3185    45. Arranged for dftables to add
3186    
3187          #include "pcre_internal.h"
3188    
3189        to pcre_chartables.c because without it, gcc 4.x may remove the array
3190        definition from the final binary if PCRE is built into a static library and
3191        dead code stripping is activated.
3192    
3193    46. For an unanchored pattern, if a match attempt fails at the start of a
3194        newline sequence, and the newline setting is CRLF or ANY, and the next two
3195        characters are CRLF, advance by two characters instead of one.
3196    
3197    
3198    Version 6.7 04-Jul-06
3199    ---------------------
3200    
3201     1. In order to handle tests when input lines are enormously long, pcretest has
3202        been re-factored so that it automatically extends its buffers when
3203        necessary. The code is crude, but this _is_ just a test program. The
3204        default size has been increased from 32K to 50K.
3205    
3206     2. The code in pcre_study() was using the value of the re argument before
3207        testing it for NULL. (Of course, in any sensible call of the function, it
3208        won't be NULL.)
3209    
3210     3. The memmove() emulation function in pcre_internal.h, which is used on
3211        systems that lack both memmove() and bcopy() - that is, hardly ever -
3212        was missing a "static" storage class specifier.
3213    
3214     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3215        containing an extended class (one that cannot be represented by a bitmap
3216        because it contains high-valued characters or Unicode property items, e.g.
3217        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3218        pattern, but PCRE should not loop if you do not (it no longer does).
3219        [Detail: two cases were found: (a) a repeated subpattern containing an
3220        extended class; (b) a recursive reference to a subpattern that followed a
3221        previous extended class. It wasn't skipping over the extended class
3222        correctly when UTF-8 mode was not set.]
3223    
3224     5. A negated single-character class was not being recognized as fixed-length
3225        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3226        compile error "lookbehind assertion is not fixed length".
3227    
3228     6. The RunPerlTest auxiliary script was showing an unexpected difference
3229        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3230        write a Perl script that can interpret lines of an input file either as
3231        byte characters or as UTF-8, which is what "perltest" was being required to
3232        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3233        can't do is switch easily at run time between having the "use utf8;" pragma
3234        or not. In the end, I fudged it by using the RunPerlTest script to insert
3235        "use utf8;" explicitly for the UTF-8 tests.
3236    
3237     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3238        the end of the subject string, contrary to the documentation and to what
3239        Perl does. This was true of both matching functions. Now it matches only at
3240        the start of the subject and immediately after *internal* newlines.
3241    
3242     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3243        a pointer to an int instead of a pointer to an unsigned long int. This
3244        caused problems on 64-bit systems.
3245    
3246     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3247        instance of the 'standard' template library not being so standard".
3248    
3249    10. There was no check on the number of named subpatterns nor the maximum
3250        length of a subpattern name. The product of these values is used to compute
3251        the size of the memory block for a compiled pattern. By supplying a very
3252        long subpattern name and a large number of named subpatterns, the size
3253        computation could be caused to overflow. This is now prevented by limiting
3254        the length of names to 32 characters, and the number of named subpatterns
3255        to 10,000.
3256    
3257    11. Subpatterns that are repeated with specific counts have to be replicated in
3258        the compiled pattern. The size of memory for this was computed from the
3259        length of the subpattern and the repeat count. The latter is limited to
3260        65535, but there was no limit on the former, meaning that integer overflow
3261        could in principle occur. The compiled length of a repeated subpattern is
3262        now limited to 30,000 bytes in order to prevent this.
3263    
3264    12. Added the optional facility to have named substrings with the same name.
3265    
3266    13. Added the ability to use a named substring as a condition, using the
3267        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3268        are numbers (not recommended). Forward references are permitted.
3269    
3270    14. Added forward references in named backreferences (if you see what I mean).
3271    
3272    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3273        pattern could run off the end of the subject. For example, the pattern
3274        "(?s)(.{1,5})"8 did this with the subject "ab".
3275    
3276    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3277        PCRE_CASELESS was set when matching characters that were quantified with ?
3278        or *.
3279    
3280    17. A character class other than a single negated character that had a minimum
3281        but no maximum quantifier - for example [ab]{6,} - was not handled
3282        correctly by pce_dfa_exec(). It would match only one character.
3283    
3284    18. A valid (though odd) pattern that looked like a POSIX character
3285        class but used an invalid character after [ (for example [[,abc,]]) caused
3286        pcre_compile() to give the error "Failed: internal error: code overflow" or
3287        in some cases to crash with a glibc free() error. This could even happen if
3288        the pattern terminated after [[ but there just happened to be a sequence of
3289        letters, a binary zero, and a closing ] in the memory that followed.
3290    
3291    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3292        over the years. Originally (before any Unicode support), just the bottom 8
3293        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3294        output from "man perlunicode" includes this:
3295    
3296          The regular expression compiler produces polymorphic opcodes.  That
3297          is, the pattern adapts to the data and automatically switches to
3298          the Unicode character scheme when presented with Unicode data--or
3299          instead uses a traditional byte scheme when presented with byte
3300          data.
3301    
3302        Sadly, a wide octal escape does not cause a switch, and in a string with
3303        no other multibyte characters, these octal escapes are treated as before.
3304        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3305        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3306        Unicode string.
3307    
3308        I have not perpetrated such confusion in PCRE. Up till now, it took just
3309        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3310        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3311        translate to the appropriate multibyte character.
3312    
3313    29. Applied some refactoring to reduce the number of warnings from Microsoft
3314        and Borland compilers. This has included removing the fudge introduced
3315        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3316        a warning about an unused variable.
3317    
3318    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3319        characters since release 4.0, because Perl (from release 5.004) does not.
3320        [Or at least, is documented not to: some releases seem to be in conflict
3321        with the documentation.] However, when a pattern was studied with
3322        pcre_study() and all its branches started with \s, PCRE still included VT
3323        as a possible starting character. Of course, this did no harm; it just
3324        caused an unnecessary match attempt.
3325    
3326    22. Removed a now-redundant internal flag bit that recorded the fact that case
3327        dependency changed within the pattern. This was once needed for "required
3328        byte" processing, but is no longer used. This recovers a now-scarce options
3329        bit. Also moved the least significant internal flag bit to the most-
3330        significant bit of the word, which was not previously used (hangover from
3331        the days when it was an int rather than a uint) to free up another bit for
3332        the future.
3333    
3334    23. Added support for CRLF line endings as well as CR and LF. As well as the
3335        default being selectable at build time, it can now be changed at runtime
3336        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3337        specify that it is scanning data with non-default line endings.
3338    
3339    24. Changed the definition of CXXLINK to make it agree with the definition of
3340        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3341    
3342    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3343        recursions. This makes a big different to stack usage for some patterns.
3344    
3345    26. If a subpattern containing a named recursion or subroutine reference such
3346        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3347        the space required for the compiled pattern went wrong and gave too small a
3348        value. Depending on the environment, this could lead to "Failed: internal
3349        error: code overflow at offset 49" or "glibc detected double free or
3350        corruption" errors.
3351    
3352    27. Applied patches from Google (a) to support the new newline modes and (b) to
3353        advance over multibyte UTF-8 characters in GlobalReplace.
3354    
3355    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3356        difference for some implementation of PCRE in some Windows version.
3357    
3358    29. Added some extra testing facilities to pcretest:
3359    
3360        \q<number>   in a data line sets the "match limit" value
3361        \Q<number>   in a data line sets the "match recursion limt" value
3362        -S <number>  sets the stack size, where <number> is in megabytes
3363    
3364        The -S option isn't available for Windows.
3365    
3366    
3367    Version 6.6 06-Feb-06
3368    ---------------------
3369    
3370     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3371        in pcreposix.h. I have copied the definition from pcre.h.
3372    
3373     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3374        because pcre.h is no longer a built file.
3375    
3376     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3377        not normally included in the compiled code.
3378    
3379    
3380    Version 6.5 01-Feb-06
3381    ---------------------
3382    
3383     1. When using the partial match feature with pcre_dfa_exec(), it was not
3384        anchoring the second and subsequent partial matches at the new starting
3385        point. This could lead to incorrect results. For example, with the pattern
3386        /1234/, partially matching against "123" and then "a4" gave a match.
3387    
3388     2. Changes to pcregrep:
3389    
3390        (a) All non-match returns from pcre_exec() were being treated as failures
3391            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3392            error message is output. Some extra information is given for the
3393            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3394            probably the only errors that are likely to be caused by users (by
3395            specifying a regex that has nested indefinite repeats, for instance).
3396            If there are more than 20 of these errors, pcregrep is abandoned.
3397    
3398        (b) A binary zero was treated as data while matching, but terminated the
3399            output line if it was written out. This has been fixed: binary zeroes
3400            are now no different to any other data bytes.
3401    
3402        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3403            used to set a locale for matching. The --locale=xxxx long option has
3404            been added (no short equivalent) to specify a locale explicitly on the
3405            pcregrep command, overriding the environment variables.
3406    
3407        (d) When -B was used with -n, some line numbers in the output were one less
3408            than they should have been.
3409    
3410        (e) Added the -o (--only-matching) option.
3411    
3412        (f) If -A or -C was used with -c (count only), some lines of context were
3413            accidentally printed for the final match.
3414    
3415        (g) Added the -H (--with-filename) option.
3416    
3417        (h) The combination of options -rh failed to suppress file names for files
3418            that were found from directory arguments.
3419    
3420        (i) Added the -D (--devices) and -d (--directories) options.
3421    
3422        (j) Added the -F (--fixed-strings) option.
3423    
3424        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3425    
3426        (l) Added the --colo(u)r option.
3427    
3428        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3429            is not present by default.
3430    
3431     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3432        items such as (?R) or (?1), when the recursion could match a number of
3433        alternatives. If it matched one of the alternatives, but subsequently,
3434        outside the recursion, there was a failure, the code tried to back up into
3435        the recursion. However, because of the way PCRE is implemented, this is not
3436        possible, and the result was an incorrect result from the match.
3437    
3438        In order to prevent this happening, the specification of recursion has
3439        been changed so that all such subpatterns are automatically treated as
3440        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3441    
3442     4. I had overlooked the fact that, in some locales, there are characters for
3443        which isalpha() is true but neither isupper() nor islower() are true. In
3444        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3445        and ordfeminine) are like this. This affected the treatment of \w and \W
3446        when they appeared in character classes, but not when they appeared outside
3447        a character class. The bit map for "word" characters is now created
3448        separately from the results of isalnum() instead of just taking it from the
3449        upper, lower, and digit maps. (Plus the underscore character, of course.)
3450    
3451     5. The above bug also affected the handling of POSIX character classes such as
3452        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3453        permanent tables. Instead, the bit maps for such a class were previously
3454        created as the appropriate unions of the upper, lower, and digit bitmaps.
3455        Now they are created by subtraction from the [[:word:]] class, which has
3456        its own bitmap.
3457    
3458     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3459        It is created by subtracting the vertical space characters (\x09, \x0a,
3460        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3461        subtraction was done in the overall bitmap for a character class, meaning
3462        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3463        be recognized. This bug has been fixed.
3464    
3465     7. Patches from the folks at Google:
3466    
3467          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3468          real life, but is still worth protecting against".
3469    
3470          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3471          regular expressions".
3472    
3473          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3474          have it.
3475    
3476          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3477          "configure" and the latter not, in order to fix a problem somebody had
3478          with compiling the Arg class on HP-UX.
3479    
3480          (e) Improve the error-handling of the C++ wrapper a little bit.
3481    
3482          (f) New tests for checking recursion limiting.
3483    
3484     8. The pcre_memmove() function, which is used only if the environment does not
3485        have a standard memmove() function (and is therefore rarely compiled),
3486        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3487        returning a result (though PCRE never actually uses the result).
3488    
3489     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3490        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3491        returned instead of calling malloc() with an overflowing number that would
3492        most likely cause subsequent chaos.
3493    
3494    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3495    
3496    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3497        with this option is matched, the nmatch and pmatch options of regexec() are
3498        ignored.
3499    
3500    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3501        provided in case anyone wants to the the POSIX interface with UTF-8
3502        strings.
3503    
3504    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3505        C++ linking (needed for some HP-UX environments).
3506    
3507    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3508        (unused parameter) and in the pcre_printint() function (omitted "default"
3509        switch label when the default is to do nothing).
3510    
3511    15. Added some code to make it possible, when PCRE is compiled as a C++
3512        library, to replace subject pointers for pcre_exec() with a smart pointer
3513        class, thus making it possible to process discontinuous strings.
3514    
3515    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3516        much the same function. They were added by different people who were trying
3517        to make PCRE easy to compile on non-Unix systems. It has been suggested
3518        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3519        for compiling on Windows systems. I have therefore replaced it with
3520        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3521        defaults to "extern" for C or "extern C" for C++, which works fine on
3522        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3523        SCOPE with something explicit in config.h. In addition:
3524    
3525        (a) pcreposix.h still had just "extern" instead of either of these macros;
3526            I have replaced it with PCRE_DATA_SCOPE.
3527    
3528        (b) Functions such as _pcre_xclass(), which are internal to the library,
3529            but external in the C sense, all had PCRE_EXPORT in their definitions.
3530            This is apparently wrong for the Windows case, so I have removed it.
3531            (It makes no difference on Unix-like systems.)
3532    
3533    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3534        of recursive calls to match(). This is different to MATCH_LIMIT because
3535        that limits the total number of calls to match(), not all of which increase
3536        the depth of recursion. Limiting the recursion depth limits the amount of
3537        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3538        when PCRE is compiled, and changed at run time. A patch from Google adds
3539        this functionality to the C++ interface.
3540    
3541    18. Changes to the handling of Unicode character properties:
3542    
3543        (a) Updated the table to Unicode 4.1.0.
3544    
3545        (b) Recognize characters that are not in the table as "Cn" (undefined).
3546    
3547        (c) I revised the way the table is implemented to a much improved format
3548            which includes recognition of ranges. It now supports the ranges that
3549            are defined in UnicodeData.txt, and it also amalgamates other
3550            characters into ranges. This has reduced the number of entries in the
3551            table from around 16,000 to around 3,000, thus reducing its size
3552            considerably. I realized I did not need to use a tree structure after
3553            all - a binary chop search is just as efficient. Having reduced the
3554            number of entries, I extended their size from 6 bytes to 8 bytes to
3555            allow for more data.
3556    
3557        (d) Added support for Unicode script names via properties such as \p{Han}.
3558    
3559    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3560        matching that character.
3561    
3562    20. When matching a repeated Unicode property with a minimum greater than zero,
3563        (for example \pL{2,}), PCRE could look past the end of the subject if it
3564        reached it while seeking the minimum number of characters. This could
3565        happen only if some of the characters were more than one byte long, because
3566        there is a check for at least the minimum number of bytes.
3567    
3568    21. Refactored the implementation of \p and \P so as to be more general, to
3569        allow for more different types of property in future. This has changed the
3570        compiled form incompatibly. Anybody with saved compiled patterns that use
3571        \p or \P will have to recompile them.
3572    
3573    22. Added "Any" and "L&" to the supported property types.
3574    
3575    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3576        but give a compile time error if the value is greater than 0xff.
3577    
3578    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3579        accidentally not being installed or uninstalled.
3580    
3581    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3582        made were to insert the current release number. This seemed silly, because
3583        it made things harder for people building PCRE on systems that don't run
3584        "configure". I have turned pcre.h into a distributed file, no longer built
3585        by "configure", with the version identification directly included. There is
3586        no longer a pcre.h.in file.
3587    
3588        However, this change necessitated a change to the pcre-config script as
3589        well. It is built from pcre-config.in, and one of the substitutions was the
3590        release number. I have updated configure.ac so that ./configure now finds
3591        the release number by grepping pcre.h.
3592    
3593    26. Added the ability to run the tests under valgrind.
3594    
3595    
3596    Version 6.4 05-Sep-05
3597    ---------------------
3598    
3599     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3600        "--" to be printed when multiple files were scanned, even when none of the
3601        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3602        consider it to be a bug, and have restored the previous behaviour.
3603    
3604     2. A couple of code tidies to get rid of compiler warnings.
3605    
3606     3. The pcretest program used to cheat by referring to symbols in the library
3607        whose names begin with _pcre_. These are internal symbols that are not
3608        really supposed to be visible externally, and in some environments it is
3609        possible to suppress them. The cheating is now confined to including
3610        certain files from the library's source, which is a bit cleaner.
3611    
3612     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3613        file's purpose clearer.
3614    
3615     5. Reorganized pcre_ucp_findchar().
3616    
3617    
3618    Version 6.3 15-Aug-05
3619    ---------------------
3620    
3621     1. The file libpcre.pc.in did not have general read permission in the tarball.
3622    
3623     2. There were some problems when building without C++ support:
3624    
3625        (a) If C++ support was not built, "make install" and "make test" still
3626            tried to test it.
3627    
3628        (b) There were problems when the value of CXX was explicitly set. Some
3629            changes have been made to try to fix these, and ...
3630    
3631        (c) --disable-cpp can now be used to explicitly disable C++ support.
3632    
3633        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3634            backslash in a target when C++ was disabled. This confuses some
3635            versions of "make", apparently. Using an intermediate variable solves
3636            this. (Same for CPP_LOBJ.)
3637    
3638     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3639        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3640        necessary on certain architectures.
3641    
3642     4. Added a setting of -export-symbols-regex to the link command to remove
3643        those symbols that are exported in the C sense, but actually are local
3644        within the library, and not documented. Their names all begin with
3645        "_pcre_". This is not a perfect job, because (a) we have to except some
3646        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3647        available (and never for static libraries). I have made a note to try to
3648        find a way round (a) in the future.
3649    
3650    
3651    Version 6.2 01-Aug-05
3652    ---------------------
3653    
3654     1. There was no test for integer overflow of quantifier values. A construction
3655        such as {1111111111111111} would give undefined results. What is worse, if
3656        a minimum quantifier for a parenthesized subpattern overflowed and became
3657        negative, the calculation of the memory size went wrong. This could have
3658        led to memory overwriting.
3659    
3660     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3661    
3662     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3663        operating environments where this matters.
3664    
3665     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3666        PCRE options from within the C++ wrapper.
3667    
3668     5. Named capturing subpatterns were not being correctly counted when a pattern
3669        was compiled. This caused two problems: (a) If there were more than 100
3670        such subpatterns, the calculation of the memory needed for the whole
3671        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3672        back references of the form \12, where the number was greater than 9, were
3673        not recognized as back references, even though there were sufficient
3674        previous subpatterns.
3675    
3676     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3677        versions of gcc, e.g. 2.95.4.
3678    
3679    
3680    Version 6.1 21-Jun-05
3681    ---------------------
3682    
3683     1. There was one reference to the variable "posix" in pcretest.c that was not
3684        surrounded by "#if !defined NOPOSIX".
3685    
3686     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3687        the cross-check on the old pcre_info() function, for the benefit of the
3688        cut-down version of PCRE that is currently imported into Exim.
3689    
3690     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3691        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3692        patterns that start (?i) but I don't think that matters. The use of (?i) is
3693        just an example; this all applies to the other options as well.
3694    
3695     4. Since libtool seems to echo the compile commands it is issuing, the output
3696        from "make" can be reduced a bit by putting "@" in front of each libtool
3697        compile command.
3698    
3699     5. Patch from the folks at Google for configure.in to be a bit more thorough
3700        in checking for a suitable C++ installation before trying to compile the
3701        C++ stuff. This should fix a reported problem when a compiler was present,
3702        but no suitable headers.
3703    
3704     6. The man pages all had just "PCRE" as their title. I have changed them to
3705        be the relevant file name. I have also arranged that these names are
3706        retained in the file doc/pcre.txt, which is a concatenation in text format
3707        of all the man pages except the little individual ones for each function.
3708    
3709     7. The NON-UNIX-USE file had not been updated for the different set of source
3710        files that come with release 6. I also added a few comments about the C++
3711        wrapper.
3712    
3713    
3714    Version 6.0 07-Jun-05
3715    ---------------------
3716    
3717     1. Some minor internal re-organization to help with my DFA experiments.
3718    
3719     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3720        didn't matter for the library itself when fully configured, but did matter
3721        when compiling without UCP support, or within Exim, where the ucp files are
3722        not imported.
3723    
3724     3. Refactoring of the library code to split up the various functions into
3725        different source modules. The addition of the new DFA matching code (see
3726        below) to a single monolithic source would have made it really too
3727        unwieldy, quite apart from causing all the code to be include in a
3728        statically linked application, when only some functions are used. This is
3729        relevant even without the DFA addition now that patterns can be compiled in
3730        one application and matched in another.
3731    
3732        The downside of splitting up is that there have to be some external
3733        functions and data tables that are used internally in different modules of
3734        the library but which are not part of the API. These have all had their
3735        names changed to start with "_pcre_" so that they are unlikely to clash
3736        with other external names.
3737    
3738     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3739        a different (DFA) algorithm. Although it is slower than the original
3740        function, it does have some advantages for certain types of matching
3741        problem.
3742    
3743     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3744        including restarting after a partial match.
3745    
3746     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3747        defined when compiling for Windows was sent to me. I have put it into the
3748        code, though I have no means of testing or verifying it.
3749    
3750     7. Added the pcre_refcount() auxiliary function.
3751    
3752     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3753        match before or at the first newline in the subject string. In pcretest,
3754        the /f option on a pattern can be used to set this.
3755    
3756     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3757        would behave wrongly. This has been present in PCRE since release 4.0.
3758    
3759    10. A number of changes to the pcregrep command:
3760    
3761        (a) Refactored how -x works; insert ^(...)$ instead of setting
3762            PCRE_ANCHORED and checking the length, in preparation for adding
3763            something similar for -w.
3764    
3765        (b) Added the -w (match as a word) option.
3766    
3767        (c) Refactored the way lines are read and buffered so as to have more
3768            than one at a time available.
3769    
3770        (d) Implemented a pcregrep test script.
3771    
3772        (e) Added the -M (multiline match) option. This allows patterns to match
3773            over several lines of the subject. The buffering ensures that at least
3774            8K, or the rest of the document (whichever is the shorter) is available
3775            for matching (and similarly the previous 8K for lookbehind assertions).
3776    
3777        (f) Changed the --help output so that it now says
3778    
3779              -w, --word-regex(p)
3780    
3781            instead of two lines, one with "regex" and the other with "regexp"
3782            because that confused at least one person since the short forms are the
3783            same. (This required a bit of code, as the output is generated
3784            automatically from a table. It wasn't just a text change.)
3785    
3786        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3787            option but starts with a hyphen. Could be a pattern or a path name
3788            starting with a hyphen, for instance.
3789