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

Diff of /code/trunk/ChangeLog

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

revision 200 by ph10, Wed Aug 1 09:10:40 2007 UTC revision 1397 by ph10, Mon Nov 11 18:33:23 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 05-Jul-07  Version 8.34 xx-xxxx-201x
5    --------------------------
6    
7    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8        executable memory. Patch inspired by Carsten Klein.
9    
10    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11        this macro is never tested and has no effect, because the work to support
12        coverage involves only compiling and linking options and special targets in
13        the Makefile. The comment in config.h implied that defining the macro would
14        enable coverage support, which is totally false. There was also support for
15        setting this macro in the CMake files (my fault, I just copied it from
16        configure). SUPPORT_GCOV has now been removed.
17    
18    3.  Make a small performance improvement in strlen16() and strlen32() in
19        pcretest.
20    
21    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23        these cases). There was also one in pcretest.c.
24    
25    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26    
27    6.  In UTF mode, the code for checking whether a group could match an empty
28        string (which is used for indefinitely repeated groups to allow for
29        breaking an infinite loop) was broken when the group contained a repeated
30        negated single-character class with a character that occupied more than one
31        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32        UTF-8 mode). The effect was undefined: the group might or might not be
33        deemed as matching an empty string, or the program might have crashed.
34    
35    7.  The code for checking whether a group could match an empty string was not
36        recognizing that \h, \H, \v, \V, and \R must match a character.
37    
38    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39        an empty string. If it can, pcretest shows this in its information output.
40    
41    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
42        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43        matched by pcre_exec() without using JIT:
44    
45        (a) If the rest of the pattern did not match after a maximal run of
46            grapheme clusters, the code for backing up to try with fewer of them
47            did not always back up over a full grapheme when characters that do not
48            have the modifier quality were involved, e.g. Hangul syllables.
49    
50        (b) If the match point in a subject started with modifier character, and
51            there was no match, the code could incorrectly back up beyond the match
52            point, and potentially beyond the first character in the subject,
53            leading to a segfault or an incorrect match result.
54    
55    10. A conditional group with an assertion condition could lead to PCRE
56        recording an incorrect first data item for a match if no other first data
57        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58        first data item, and therefore matched "ca" after "c" instead of at the
59        start.
60    
61    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
63    
64    12. The source of pcregrep now includes z/OS-specific code so that it can be
65        compiled for z/OS as part of the special z/OS distribution.
66    
67    13. Added the -T and -TM options to pcretest.
68    
69    14. The code in pcre_compile.c for creating the table of named capturing groups
70        has been refactored. Instead of creating the table dynamically during the
71        actual compiling pass, the information is remembered during the pre-compile
72        pass (on the stack unless there are more than 20 named groups, in which
73        case malloc() is used) and the whole table is created before the actual
74        compile happens. This has simplified the code (it is now nearly 150 lines
75        shorter) and prepared the way for better handling of references to groups
76        with duplicate names.
77    
78    15. A back reference to a named subpattern when there is more than one of the
79        same name now checks them in the order in which they appear in the pattern.
80        The first one that is set is used for the reference. Previously only the
81        first one was inspected. This change makes PCRE more compatible with Perl.
82    
83    16. Unicode character properties were updated from Unicode 6.3.0.
84    
85    17. The compile-time code for auto-possessification has been refactored, based
86        on a patch by Zoltan Herczeg. It now happens after instead of during
87        compilation. The code is cleaner, and more cases are handled. The option
88        PCRE_NO_AUTO_POSSESSIFY is added for testing purposes, and the -O and /O
89        options in pcretest are provided to set it. It can also be set by
90        (*NO_AUTO_POSSESS) at the start of a pattern.
91    
92    18. The character VT has been added to the set of characters that match \s and
93        are generally treated as white space, following this same change in Perl
94        5.18. There is now no difference between "Perl space" and "POSIX space".
95    
96    19. The code for checking named groups as conditions, either for being set or
97        for being recursed, has been refactored (this is related to 14 and 15
98        above). Processing unduplicated named groups should now be as fast at
99        numerical groups, and processing duplicated groups should be faster than
100        before.
101    
102    20. Two patches to the CMake build system, by Alexander Barkov:
103    
104          (1) Replace the "source" command by "." in CMakeLists.txt because
105              "source" is a bash-ism.
106    
107          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
108              without these the CMake build does not work on Solaris.
109    
110    21. Perl has changed its handling of \8 and \9. If there is no previously
111        encountered capturing group of those numbers, they are treated as the
112        literal characters 8 and 9 instead of a binary zero followed by the
113        literals. PCRE now does the same.
114    
115    22. Following Perl, added \o{} to specify codepoints in octal, making it
116        possible to specify values greater than 0777 and also making them
117        unambiguous.
118    
119    23. Perl now gives an error for missing closing braces after \x{... instead of
120        treating the string as literal. PCRE now does the same.
121    
122    24. RunTest used to grumble if an inappropriate test was selected explicitly,
123        but just skip it when running all tests. This make it awkward to run ranges
124        of tests when one of them was inappropriate. Now it just skips any
125        inappropriate tests, as it always did when running all tests.
126    
127    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
128        character types such as \d or \w, too many callouts were inserted, and the
129        data that they returned was rubbish.
130    
131    26. In UCP mode, \s was not matching two of the characters that Perl matches,
132        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
133        were matched by \h. The code has now been refactored so that the lists of
134        the horizontal and vertical whitespace characters used for \h and \v (which
135        are defined only in one place) are now also used for \s.
136    
137    27. Add JIT support for the 64 bit TileGX architecture.
138        Patch by Jiong Wang (Tilera Corporation).
139    
140    28. Possessive quantifiers for classes (both explicit and automatically
141        generated) now use special opcodes instead of wrapping in ONCE brackets.
142    
143    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
144        (because it's meaningless), this was not happening when PCRE_CASELESS was
145        set. Not wrong, but inefficient.
146    
147    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
148        properties for \w, \d, etc) is present in a test regex. Otherwise if the
149        test contains no characters greater than 255, Perl doesn't realise it
150        should be using Unicode semantics.
151    
152    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
153        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
154        does in Unicode mode.
155    
156    32. Added the "forbid" facility to pcretest so that putting tests into the
157        wrong test files can sometimes be quickly detected.
158    
159    33. There is now a limit (default 250) on the depth of nesting of parentheses.
160        This limit is imposed to control the amount of system stack used at compile
161        time. It can be changed at build time by --with-parens-nest-limit=xxx or
162        the equivalent in CMake.
163    
164    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
165        errors. Perl warns for these when in warning mode, but PCRE has no facility
166        for giving warnings.
167    
168    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
169        However, this was not working for (?!) because it is optimized to (*FAIL),
170        for which PCRE does not allow quantifiers. The optimization is now disabled
171        when a quantifier follows (?!). I can't see any use for this, but it makes
172        things uniform.
173    
174    36. Perl no longer allows group names to start with digits, so I have made this
175        change also in PCRE. It simplifies the code a bit.
176    
177    37. In extended mode, Perl ignores spaces before a + that indicates a
178        possessive quantifier. PCRE allowed a space before the quantifier, but not
179        before the possessive +. It now does.
180    
181    38. The use of \K (reset reported match start) within a repeated possessive
182        group such as (a\Kb)*+ was not working.
183    
184    
185    Version 8.33 28-May-2013
186    ------------------------
187    
188    1.  Added 'U' to some constants that are compared to unsigned integers, to
189        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
190        variables that are added to signed variables, to ensure the result is
191        signed and can be negated.
192    
193    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
194        CMake config files.
195    
196    3.  Revise the creation of config.h.generic so that all boolean macros are
197        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
198        overriding via -D on the command line possible.
199    
200    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
201        to unsigned int is reported to make a quite noticeable speed difference in
202        a specific Windows environment. Testing on Linux did also appear to show
203        some benefit (and it is clearly not harmful). Also fixed the definition of
204        Xop which should be unsigned.
205    
206    5.  Related to (4), changing the definition of the intermediate variable cc
207        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
208        improvements.
209    
210    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
211        unnecessary spaces.
212    
213    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
214        and later.
215    
216    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
217    
218    9.  Optimizing fast_forward_start_bits in JIT.
219    
220    10. Adding support for callouts in JIT, and fixing some issues revealed
221        during this work. Namely:
222    
223        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
224    
225        (b) Minimum length was not checked before the matching is started.
226    
227    11. The value of capture_last that is passed to callouts was incorrect in some
228        cases when there was a capture on one path that was subsequently abandoned
229        after a backtrack. Also, the capture_last value is now reset after a
230        recursion, since all captures are also reset in this case.
231    
232    12. The interpreter no longer returns the "too many substrings" error in the
233        case when an overflowing capture is in a branch that is subsequently
234        abandoned after a backtrack.
235    
236    13. In the pathological case when an offset vector of size 2 is used, pcretest
237        now prints out the matched string after a yield of 0 or 1.
238    
239    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
240        Only supported by the JIT compiler at the moment.
241    
242    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
243    
244    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
245        offset of the starting point of the matching process, provided the offsets
246        vector is large enough.
247    
248    17. The \A escape now records a lookbehind value of 1, though its execution
249        does not actually inspect the previous character. This is to ensure that,
250        in partial multi-segment matching, at least one character from the old
251        segment is retained when a new segment is processed. Otherwise, if there
252        are no lookbehinds in the pattern, \A might match incorrectly at the start
253        of a new segment.
254    
255    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
256    
257    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
258        gives some modest performance improvement in 8-bit mode.
259    
260    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
261        be expressed in certain programming languages using Universal Character
262        Names.
263    
264    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
265        which points out that "non characters" are not "characters that may not
266        appear in Unicode strings" but rather "characters that are reserved for
267        internal use and have only local meaning".
268    
269    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
270        there was a conditional group that depended on an assertion, if the
271        assertion was false, the callout that immediately followed the alternation
272        in the condition was skipped when pcre_exec() was used for matching.
273    
274    23. Allow an explicit callout to be inserted before an assertion that is the
275        condition for a conditional group, for compatibility with automatic
276        callouts, which always insert a callout at this point.
277    
278    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
279        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
280    
281    25. (*PRUNE) is now supported by the JIT compiler.
282    
283    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
284    
285    27. Fix the case where there are two or more SKIPs with arguments that may be
286        ignored.
287    
288    28. (*SKIP) is now supported by the JIT compiler.
289    
290    29. (*THEN) is now supported by the JIT compiler.
291    
292    30. Update RunTest with additional test selector options.
293    
294    31. The way PCRE handles backtracking verbs has been changed in two ways.
295    
296        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
297        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
298        backtracking. In some cases this makes it more Perl-compatible, but Perl's
299        rather obscure rules do not always do the same thing.
300    
301        (2) Previously, backtracking verbs were confined within assertions. This is
302        no longer the case for positive assertions, except for (*ACCEPT). Again,
303        this sometimes improves Perl compatibility, and sometimes does not.
304    
305    32. A number of tests that were in test 2 because Perl did things differently
306        have been moved to test 1, because either Perl or PCRE has changed, and
307        these tests are now compatible.
308    
309    32. Backtracking control verbs are now handled in the same way in JIT and
310        interpreter.
311    
312    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
313        contained a forward subroutine reference caused a compile error.
314    
315    34. Auto-detect and optimize limited repetitions in JIT.
316    
317    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
318        blocking (*UTF) etc.
319    
320    36. In the interpreter, maximizing pattern repetitions for characters and
321        character types now use tail recursion, which reduces stack usage.
322    
323    37. The value of the max lookbehind was not correctly preserved if a compiled
324        and saved regex was reloaded on a host of different endianness.
325    
326    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
327        of the compiled pattern block, expand the flags field from 16 to 32 bits
328        because it was almost full.
329    
330    39. Try madvise first before posix_madvise.
331    
332    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
333        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
334        original change locked out all matches of empty strings. This has been
335        changed so that one match of an empty string per line is recognized.
336        Subsequent searches on the same line (for colouring or for --only-matching,
337        for example) do not recognize empty strings.
338    
339    41. Applied a user patch to fix a number of spelling mistakes in comments.
340    
341    42. Data lines longer than 65536 caused pcretest to crash.
342    
343    43. Clarified the data type for length and startoffset arguments for pcre_exec
344        and pcre_dfa_exec in the function-specific man pages, where they were
345        explicitly stated to be in bytes, never having been updated. I also added
346        some clarification to the pcreapi man page.
347    
348    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
349        a segmentation fault.
350    
351    
352    Version 8.32 30-November-2012
353    -----------------------------
354    
355    1.  Improved JIT compiler optimizations for first character search and single
356        character iterators.
357    
358    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
359        Patch by Daniel Richard G.
360    
361    3.  Single character iterator optimizations in the JIT compiler.
362    
363    4.  Improved JIT compiler optimizations for character ranges.
364    
365    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
366        Reported by Giuseppe D'Angelo.
367    
368    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
369        of a line, was being set incorrectly in cases where .* appeared inside
370        atomic brackets at the start of a pattern, or where there was a subsequent
371        *PRUNE or *SKIP.
372    
373    7.  Improved instruction cache flush for POWER/PowerPC.
374        Patch by Daniel Richard G.
375    
376    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
377        grep:
378    
379        (a) There is now no limit to the number of patterns to be matched.
380    
381        (b) An error is given if a pattern is too long.
382    
383        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
384            are now supported.
385    
386        (d) --exclude-from and --include-from (multiple use) have been added.
387    
388        (e) Exclusions and inclusions now apply to all files and directories, not
389            just to those obtained from scanning a directory recursively.
390    
391        (f) Multiple uses of -f and --file-list are now supported.
392    
393        (g) In a Windows environment, the default for -d has been changed from
394            "read" (the GNU grep default) to "skip", because otherwise the presence
395            of a directory in the file list provokes an error.
396    
397        (h) The documentation has been revised and clarified in places.
398    
399    9.  Improve the matching speed of capturing brackets.
400    
401    10. Changed the meaning of \X so that it now matches a Unicode extended
402        grapheme cluster.
403    
404    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
405        out POSIX threads when JIT support is configured.
406    
407    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
408    
409    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
410        structure could go wrong in environments where size_t is not the same size
411        as int.
412    
413    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
414        set.
415    
416    15. The EBCDIC support had decayed; later updates to the code had included
417        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
418        general tidy up of EBCDIC-related issues, and the documentation was also
419        not quite right. There is now a test that can be run on ASCII systems to
420        check some of the EBCDIC-related things (but is it not a full test).
421    
422    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
423        in a small tidy to the code.
424    
425    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
426    
427    18. If the --only-matching (-o) option in pcregrep is specified multiple
428        times, each one causes appropriate output. For example, -o1 -o2 outputs the
429        substrings matched by the 1st and 2nd capturing parentheses. A separating
430        string can be specified by --om-separator (default empty).
431    
432    19. Improving the first n character searches.
433    
434    20. Turn case lists for horizontal and vertical white space into macros so that
435        they are defined only once.
436    
437    21. This set of changes together give more compatible Unicode case-folding
438        behaviour for characters that have more than one other case when UCP
439        support is available.
440    
441        (a) The Unicode property table now has offsets into a new table of sets of
442            three or more characters that are case-equivalent. The MultiStage2.py
443            script that generates these tables (the pcre_ucd.c file) now scans
444            CaseFolding.txt instead of UnicodeData.txt for character case
445            information.
446    
447        (b) The code for adding characters or ranges of characters to a character
448            class has been abstracted into a generalized function that also handles
449            case-independence. In UTF-mode with UCP support, this uses the new data
450            to handle characters with more than one other case.
451    
452        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
453            whose other case is greater than 256 are now correctly matched
454            caselessly. Previously, the high codepoint matched the low one, but not
455            vice versa.
456    
457        (d) The processing of \h, \H, \v, and \ in character classes now makes use
458            of the new class addition function, using character lists defined as
459            macros alongside the case definitions of 20 above.
460    
461        (e) Caseless back references now work with characters that have more than
462            one other case.
463    
464        (f) General caseless matching of characters with more than one other case
465            is supported.
466    
467    22. Unicode character properties were updated from Unicode 6.2.0
468    
469    23. Improved CMake support under Windows. Patch by Daniel Richard G.
470    
471    24. Add support for 32-bit character strings, and UTF-32
472    
473    25. Major JIT compiler update (code refactoring and bugfixing).
474        Experimental Sparc 32 support is added.
475    
476    26. Applied a modified version of Daniel Richard G's patch to create
477        pcre.h.generic and config.h.generic by "make" instead of in the
478        PrepareRelease script.
479    
480    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
481        pcre_compile.c when checking for a zero character.
482    
483    28. Introducing a native interface for JIT. Through this interface, the compiled
484        machine code can be directly executed. The purpose of this interface is to
485        provide fast pattern matching, so several sanity checks are not performed.
486        However, feature tests are still performed. The new interface provides
487        1.4x speedup compared to the old one.
488    
489    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
490        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
491        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
492    
493    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
494        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
495        the "old" RFC 2279). Instead, it ended up passing a negative length to
496        pcre_exec().
497    
498    31. Add support for GCC's visibility feature to hide internal functions.
499    
500    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
501        "unknown -C option" after outputting 0 or 1.
502    
503    33. There is now support for generating a code coverage report for the test
504        suite in environments where gcc is the compiler and lcov is installed. This
505        is mainly for the benefit of the developers.
506    
507    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
508        unaddressable using valgrind annotations, allowing valgrind to detect
509        invalid memory accesses. This is mainly for the benefit of the developers.
510    
511    25. (*UTF) can now be used to start a pattern in any of the three libraries.
512    
513    26. Give configure error if --enable-cpp but no C++ compiler found.
514    
515    
516    Version 8.31 06-July-2012
517    -------------------------
518    
519    1.  Fixing a wrong JIT test case and some compiler warnings.
520    
521    2.  Removed a bashism from the RunTest script.
522    
523    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
524        to unsigned type, result still unsigned" that was given by an MS compiler
525        on encountering the code "-sizeof(xxx)".
526    
527    4.  Partial matching support is added to the JIT compiler.
528    
529    5.  Fixed several bugs concerned with partial matching of items that consist
530        of more than one character:
531    
532        (a) /^(..)\1/ did not partially match "aba" because checking references was
533            done on an "all or nothing" basis. This also applied to repeated
534            references.
535    
536        (b) \R did not give a hard partial match if \r was found at the end of the
537            subject.
538    
539        (c) \X did not give a hard partial match after matching one or more
540            characters at the end of the subject.
541    
542        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
543            a partial match for the string "\r".
544    
545        (e) When newline was set to CRLF, the metacharacter "." did not recognize
546            a partial match for a CR character at the end of the subject string.
547    
548    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
549        running pcretest, the text "(JIT)" added to the output whenever JIT is
550        actually used to run the match.
551    
552    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
553        or /S+[+] with a digit between 1 and 7.
554    
555    8.  OP_NOT now supports any UTF character not just single-byte ones.
556    
557    9.  (*MARK) control verb is now supported by the JIT compiler.
558    
559    10. The command "./RunTest list" lists the available tests without actually
560        running any of them. (Because I keep forgetting what they all are.)
561    
562    11. Add PCRE_INFO_MAXLOOKBEHIND.
563    
564    12. Applied a (slightly modified) user-supplied patch that improves performance
565        when the heap is used for recursion (compiled with --disable-stack-for-
566        recursion). Instead of malloc and free for each heap frame each time a
567        logical recursion happens, frames are retained on a chain and re-used where
568        possible. This sometimes gives as much as 30% improvement.
569    
570    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
571        call.
572    
573    14. As documented, (*COMMIT) is now confined to within a positive assertion.
574    
575    15. It is now possible to link pcretest with libedit as an alternative to
576        libreadline.
577    
578    16. (*COMMIT) control verb is now supported by the JIT compiler.
579    
580    17. The Unicode data tables have been updated to Unicode 6.1.0.
581    
582    18. Added --file-list option to pcregrep.
583    
584    19. Added binary file support to pcregrep, including the -a, --binary-files,
585        -I, and --text options.
586    
587    20. The madvise function is renamed for posix_madvise for QNX compatibility
588        reasons. Fixed by Giuseppe D'Angelo.
589    
590    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
591    
592    22. Changed the option for creating symbolic links for 16-bit man pages from
593        -s to -sf so that re-installing does not cause issues.
594    
595    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
596    
597    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
598        matches in certain environments (the workspace was not being correctly
599        retained). Also added to pcre_dfa_exec() a simple plausibility check on
600        some of the workspace data at the beginning of a restart.
601    
602    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
603        was not doing so when it should - probably a typo introduced by SVN 528
604        (change 8.10/14).
605    
606    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
607        \w+ when the character tables indicated that \x{c4} was a word character.
608        There were several related cases, all because the tests for doing a table
609        lookup were testing for characters less than 127 instead of 255.
610    
611    27. If a pattern contains capturing parentheses that are not used in a match,
612        their slots in the ovector are set to -1. For those that are higher than
613        any matched groups, this happens at the end of processing. In the case when
614        there were back references that the ovector was too small to contain
615        (causing temporary malloc'd memory to be used during matching), and the
616        highest capturing number was not used, memory off the end of the ovector
617        was incorrectly being set to -1. (It was using the size of the temporary
618        memory instead of the true size.)
619    
620    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
621        ovector size, it uses memory at the end of the block that it has got.
622    
623    29. Check for an overlong MARK name and give an error at compile time. The
624        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
625    
626    30. JIT compiler update.
627    
628    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
629        Rill for the patch.
630    
631    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
632    
633    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
634    
635    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
636        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
637        was enabled.
638    
639    35. Improve JIT code generation for greedy plus quantifier.
640    
641    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
642        1 to "aa" instead of to an empty string. The bug affected repeated groups
643        that could potentially match an empty string.
644    
645    37. Optimizing single character iterators in JIT.
646    
647    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
648        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
649        codepoints that are too big for the mode are faulted, and in a UTF mode,
650        disallowed codepoints are also faulted.
651    
652    39. If PCRE was compiled with UTF support, in three places in the DFA
653        matcher there was code that should only have been obeyed in UTF mode, but
654        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
655        processing when bytes with values greater than 127 were present. In 16-bit
656        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
657        both cases the values are those that cannot be the first data item in a UTF
658        character. The three items that might have provoked this were recursions,
659        possessively repeated groups, and atomic groups.
660    
661    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
662        and pcregrep, because some OS require shared objects to be explicitly
663        passed to ld, causing the link step to fail if they are not.
664    
665    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
666        mode, patterns that started with \h* or \R* might be incorrectly matched.
667    
668    
669    Version 8.30 04-February-2012
670    -----------------------------
671    
672    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
673        name is defined in ctype.h.
674    
675    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
676        only in quite long subpatterns.
677    
678    3.  Removed the function pcre_info(), which has been obsolete and deprecated
679        since it was replaced by pcre_fullinfo() in February 2000.
680    
681    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
682        match a (*MARK), and the match failed at the start of the subject, a
683        reference to memory before the start of the subject could occur. This bug
684        was introduced by fix 17 of release 8.21.
685    
686    5.  A reference to an unset group with zero minimum repetition was giving
687        totally wrong answers (in non-JavaScript-compatibility mode). For example,
688        /(another)?(\1?)test/ matched against "hello world test". This bug was
689        introduced in release 8.13.
690    
691    6.  Add support for 16-bit character strings (a large amount of work involving
692        many changes and refactorings).
693    
694    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
695        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
696        from a file.
697    
698    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
699        rounding is not applied in this particular case).
700    
701    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
702        if they appear, or are escaped, in patterns.
703    
704    10. Get rid of a number of -Wunused-but-set-variable warnings.
705    
706    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
707        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
708        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
709        also returns the mark "x". This bug applied to capturing parentheses,
710        non-capturing parentheses, and atomic parentheses. It also applied to some
711        assertions.
712    
713    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
714        information out of configure.ac instead of relying on pcre.h.generic, which
715        is not stored in the repository.
716    
717    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
718        -lreadline.
719    
720    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
721    
722    15. Applied Graycode's patch to put the top-level frame on the stack rather
723        than the heap when not using the stack for recursion. This gives a
724        performance improvement in many cases when recursion is not deep.
725    
726    16. Experimental code added to "pcretest -C" to output the stack frame size.
727    
728    
729    Version 8.21 12-Dec-2011
730    ------------------------
731    
732    1.  Updating the JIT compiler.
733    
734    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
735        are added as well.
736    
737    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
738        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
739        calling _pcre_jit_exec. Some extra comments are added.
740    
741    4.  (*MARK) settings inside atomic groups that do not contain any capturing
742        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
743        was introduced by change 18 for 8.20.
744    
745    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
746        ECMA-262 standard.
747    
748    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
749        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
750        This bug was probably introduced by change 9 of 8.13.
751    
752    7.  While fixing 6 above, I noticed that a number of other items were being
753        incorrectly rejected as "not fixed length". This arose partly because newer
754        opcodes had not been added to the fixed-length checking code. I have (a)
755        corrected the bug and added tests for these items, and (b) arranged for an
756        error to occur if an unknown opcode is encountered while checking for fixed
757        length instead of just assuming "not fixed length". The items that were
758        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
759        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
760        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
761    
762    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
763        being incorrectly compiled and would have given unpredicatble results.
764    
765    9.  A possessively repeated subpattern with minimum repeat count greater than
766        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
767        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
768        the first (A) could occur when it should not.
769    
770    10. Add a cast and remove a redundant test from the code.
771    
772    11. JIT should use pcre_malloc/pcre_free for allocation.
773    
774    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
775        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
776        is anything other than /usr, -L is still shown).
777    
778    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
779    
780    14. Perl does not support \N without a following name in a [] class; PCRE now
781        also gives an error.
782    
783    15. If a forward reference was repeated with an upper limit of around 2000,
784        it caused the error "internal error: overran compiling workspace". The
785        maximum number of forward references (including repeats) was limited by the
786        internal workspace, and dependent on the LINK_SIZE. The code has been
787        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
788        the default depends on LINK_SIZE. There is a new upper limit (for safety)
789        of around 200,000 forward references. While doing this, I also speeded up
790        the filling in of repeated forward references.
791    
792    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
793        incorrectly expecting the subject to contain another "a" after the start.
794    
795    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
796        in the match, the SKIP should be ignored. This was not happening; instead
797        the SKIP was being treated as NOMATCH. For patterns such as
798        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
799        tested.
800    
801    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
802        now much more compatible with Perl, in particular in cases where the result
803        is a non-match for a non-anchored pattern. For example, if
804        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
805        "m", where previously it did not return a name. A side effect of this
806        change is that for partial matches, the last encountered mark name is
807        returned, as for non matches. A number of tests that were previously not
808        Perl-compatible have been moved into the Perl-compatible test files. The
809        refactoring has had the pleasing side effect of removing one argument from
810        the match() function, thus reducing its stack requirements.
811    
812    19. If the /S+ option was used in pcretest to study a pattern using JIT,
813        subsequent uses of /S (without +) incorrectly behaved like /S+.
814    
815    21. Retrieve executable code size support for the JIT compiler and fixing
816        some warnings.
817    
818    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
819        not work when the shorter character appeared right at the end of the
820        subject string.
821    
822    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
823        systems.
824    
825    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
826        output it when the /M option is used in pcretest.
827    
828    25. The CheckMan script was not being included in the distribution. Also, added
829        an explicit "perl" to run Perl scripts from the PrepareRelease script
830        because this is reportedly needed in Windows.
831    
832    26. If study data was being save in a file and studying had not found a set of
833        "starts with" bytes for the pattern, the data written to the file (though
834        never used) was taken from uninitialized memory and so caused valgrind to
835        complain.
836    
837    27. Updated RunTest.bat as provided by Sheri Pierce.
838    
839    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
840    
841    29. Computation of memory usage for the table of capturing group names was
842        giving an unnecessarily large value.
843    
844    
845    Version 8.20 21-Oct-2011
846    ------------------------
847    
848    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
849        a POSIX class. After further experiments with Perl, which convinced me that
850        Perl has bugs and confusions, a closing square bracket is no longer allowed
851        in a POSIX name. This bug also affected patterns with classes that started
852        with full stops.
853    
854    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
855        captured substring, but while checking the failing first alternative,
856        substring 1 is temporarily captured. If the output vector supplied to
857        pcre_exec() was not big enough for this capture, the yield of the function
858        was still zero ("insufficient space for captured substrings"). This cannot
859        be totally fixed without adding another stack variable, which seems a lot
860        of expense for a edge case. However, I have improved the situation in cases
861        such as /(a)(b)x|abc/ matched against "abc", where the return code
862        indicates that fewer than the maximum number of slots in the ovector have
863        been set.
864    
865    3.  Related to (2) above: when there are more back references in a pattern than
866        slots in the output vector, pcre_exec() uses temporary memory during
867        matching, and copies in the captures as far as possible afterwards. It was
868        using the entire output vector, but this conflicts with the specification
869        that only 2/3 is used for passing back captured substrings. Now it uses
870        only the first 2/3, for compatibility. This is, of course, another edge
871        case.
872    
873    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
874        main code base, and can be used by building with --enable-jit. When this is
875        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
876        runtime --no-jit option is given.
877    
878    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
879        ovector, the return from the function was zero, implying that there were
880        other matches that did not fit. The correct "exactly full" value is now
881        returned.
882    
883    6.  If a subpattern that was called recursively or as a subroutine contained
884        (*PRUNE) or any other control that caused it to give a non-standard return,
885        invalid errors such as "Error -26 (nested recursion at the same subject
886        position)" or even infinite loops could occur.
887    
888    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
889        computing the minimum length on reaching *ACCEPT, and so ended up with the
890        wrong value of 1 rather than 0. Further investigation indicates that
891        computing a minimum subject length in the presence of *ACCEPT is difficult
892        (think back references, subroutine calls), and so I have changed the code
893        so that no minimum is registered for a pattern that contains *ACCEPT.
894    
895    8.  If (*THEN) was present in the first (true) branch of a conditional group,
896        it was not handled as intended. [But see 16 below.]
897    
898    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
899        Sheri Pierce.
900    
901    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
902        the first byte in a match must be "a".
903    
904    11. Change 17 for 8.13 increased the recursion depth for patterns like
905        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
906        pattern contains any instances of (*THEN). If it does not, the old
907        optimizations are restored. It would be nice to do this on a per-group
908        basis, but at the moment that is not feasible.
909    
910    12. In some environments, the output of pcretest -C is CRLF terminated. This
911        broke RunTest's code that checks for the link size. A single white space
912        character after the value is now allowed for.
913    
914    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
915        For "fr", it uses the Windows-specific input and output files.
916    
917    14. If (*THEN) appeared in a group that was called recursively or as a
918        subroutine, it did not work as intended. [But see next item.]
919    
920    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
921        pattern fragments (but not containing any | characters). If A and B are
922        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
923        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
924        D. In other words, Perl considers parentheses that do not contain any |
925        characters to be part of a surrounding alternative, whereas PCRE was
926        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
927        differently. PCRE now behaves in the same way as Perl, except in the case
928        of subroutine/recursion calls such as (?1) which have in any case always
929        been different (but PCRE had them first :-).
930    
931    16. Related to 15 above: Perl does not treat the | in a conditional group as
932        creating alternatives. Such a group is treated in the same way as an
933        ordinary group without any | characters when processing (*THEN). PCRE has
934        been changed to match Perl's behaviour.
935    
936    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
937        RunGrepTest script failed.
938    
939    18. Change 22 for version 13 caused atomic groups to use more stack. This is
940        inevitable for groups that contain captures, but it can lead to a lot of
941        stack use in large patterns. The old behaviour has been restored for atomic
942        groups that do not contain any capturing parentheses.
943    
944    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
945        suppress the check for a minimum subject length at run time. (If it was
946        given to pcre_exec() or pcre_dfa_exec() it did work.)
947    
948    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
949        fail to work when decoding hex characters in data strings in EBCDIC
950        environments.
951    
952    21. It appears that in at least one Mac OS environment, the isxdigit() function
953        is implemented as a macro that evaluates to its argument more than once,
954        contravening the C 90 Standard (I haven't checked a later standard). There
955        was an instance in pcretest which caused it to go wrong when processing
956        \x{...} escapes in subject strings. The has been rewritten to avoid using
957        things like p++ in the argument of isxdigit().
958    
959    
960    Version 8.13 16-Aug-2011
961    ------------------------
962    
963    1.  The Unicode data tables have been updated to Unicode 6.0.0.
964    
965    2.  Two minor typos in pcre_internal.h have been fixed.
966    
967    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
968        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
969        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
970    
971    4.  There were a number of related bugs in the code for matching backrefences
972        caselessly in UTF-8 mode when codes for the characters concerned were
973        different numbers of bytes. For example, U+023A and U+2C65 are an upper
974        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
975        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
976        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
977        2-byte code at the end of the subject (it thought there wasn't enough data
978        left).
979    
980    5.  Comprehensive information about what went wrong is now returned by
981        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
982        as the output vector has at least 2 elements. The offset of the start of
983        the failing character and a reason code are placed in the vector.
984    
985    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
986        now returned is for the first byte of the failing character, instead of the
987        last byte inspected. This is an incompatible change, but I hope it is small
988        enough not to be a problem. It makes the returned offset consistent with
989        pcre_exec() and pcre_dfa_exec().
990    
991    7.  pcretest now gives a text phrase as well as the error number when
992        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
993        failure, the offset and reason code are output.
994    
995    8.  When \R was used with a maximizing quantifier it failed to skip backwards
996        over a \r\n pair if the subsequent match failed. Instead, it just skipped
997        back over a single character (\n). This seems wrong (because it treated the
998        two characters as a single entity when going forwards), conflicts with the
999        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1000        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1001        has been changed.
1002    
1003    9.  Some internal refactoring has changed the processing so that the handling
1004        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1005        time (the PCRE_DOTALL option was changed this way some time ago: version
1006        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1007        which was always a bit of a fudge. It also means that there is one less
1008        argument for the match() function, which reduces its stack requirements
1009        slightly. This change also fixes an incompatibility with Perl: the pattern
1010        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1011    
1012    10. More internal refactoring has drastically reduced the number of recursive
1013        calls to match() for possessively repeated groups such as (abc)++ when
1014        using pcre_exec().
1015    
1016    11. While implementing 10, a number of bugs in the handling of groups were
1017        discovered and fixed:
1018    
1019        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1020        (a|)*(?1) gave a compile-time internal error.
1021        ((a|)+)+  did not notice that the outer group could match an empty string.
1022        (^a|^)+   was not marked as anchored.
1023        (.*a|.*)+ was not marked as matching at start or after a newline.
1024    
1025    12. Yet more internal refactoring has removed another argument from the match()
1026        function. Special calls to this function are now indicated by setting a
1027        value in a variable in the "match data" data block.
1028    
1029    13. Be more explicit in pcre_study() instead of relying on "default" for
1030        opcodes that mean there is no starting character; this means that when new
1031        ones are added and accidentally left out of pcre_study(), testing should
1032        pick them up.
1033    
1034    14. The -s option of pcretest has been documented for ages as being an old
1035        synonym of -m (show memory usage). I have changed it to mean "force study
1036        for every regex", that is, assume /S for every regex. This is similar to -i
1037        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1038        using it. It makes it easier to run collections of tests with and without
1039        study enabled, and thereby test pcre_study() more easily. All the standard
1040        tests are now run with and without -s (but some patterns can be marked as
1041        "never study" - see 20 below).
1042    
1043    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1044        restoration of the capturing data to the outer values was not happening
1045        correctly.
1046    
1047    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1048        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1049        pattern had matched an empty string, and so incorrectly returned a no
1050        match.
1051    
1052    17. There was optimizing code for the last branch of non-capturing parentheses,
1053        and also for the obeyed branch of a conditional subexpression, which used
1054        tail recursion to cut down on stack usage. Unfortunately, now that there is
1055        the possibility of (*THEN) occurring in these branches, tail recursion is
1056        no longer possible because the return has to be checked for (*THEN). These
1057        two optimizations have therefore been removed. [But see 8.20/11 above.]
1058    
1059    18. If a pattern containing \R was studied, it was assumed that \R always
1060        matched two bytes, thus causing the minimum subject length to be
1061        incorrectly computed because \R can also match just one byte.
1062    
1063    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1064        was incorrectly computed.
1065    
1066    20. If /S is present twice on a test pattern in pcretest input, it now
1067        *disables* studying, thereby overriding the use of -s on the command line
1068        (see 14 above). This is necessary for one or two tests to keep the output
1069        identical in both cases.
1070    
1071    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1072        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1073    
1074    22. When an atomic group that contained a capturing parenthesis was
1075        successfully matched, but the branch in which it appeared failed, the
1076        capturing was not being forgotten if a higher numbered group was later
1077        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1078        group 1 to "a", when in fact it should be unset. This applied to multi-
1079        branched capturing and non-capturing groups, repeated or not, and also to
1080        positive assertions (capturing in negative assertions does not happen
1081        in PCRE) and also to nested atomic groups.
1082    
1083    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1084        subject after a captured substring, to make it easier to tell which of a
1085        number of identical substrings has been captured.
1086    
1087    24. The way atomic groups are processed by pcre_exec() has been changed so that
1088        if they are repeated, backtracking one repetition now resets captured
1089        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1090        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1091        "aaa". Previously, it would have been "a". As part of this code
1092        refactoring, the way recursive calls are handled has also been changed.
1093    
1094    25. If an assertion condition captured any substrings, they were not passed
1095        back unless some other capturing happened later. For example, if
1096        (?(?=(a))a) was matched against "a", no capturing was returned.
1097    
1098    26. When studying a pattern that contained subroutine calls or assertions,
1099        the code for finding the minimum length of a possible match was handling
1100        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1101        group 1 called group 2 while simultaneously a separate group 2 called group
1102        1). A stack overflow occurred in this case. I have fixed this by limiting
1103        the recursion depth to 10.
1104    
1105    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1106        Fortmann. This supports explicit test numbers on the command line, and has
1107        argument validation and error reporting.
1108    
1109    28. An instance of \X with an unlimited repeat could fail if at any point the
1110        first character it looked at was a mark character.
1111    
1112    29. Some minor code refactoring concerning Unicode properties and scripts
1113        should reduce the stack requirement of match() slightly.
1114    
1115    30. Added the '=' option to pcretest to check the setting of unused capturing
1116        slots at the end of the pattern, which are documented as being -1, but are
1117        not included in the return count.
1118    
1119    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1120        compiled something random. Now it gives a compile-time error (as does
1121        Perl).
1122    
1123    32. A *MARK encountered during the processing of a positive assertion is now
1124        recorded and passed back (compatible with Perl).
1125    
1126    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1127        had alternative anchored branches, the search for a second match in a line
1128        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1129        matched the line "0102" twice. The same bug affected patterns that started
1130        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1131        twice.
1132    
1133    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1134        does, and because of capturing effects, quantifying parenthesized
1135        assertions may at times be useful. Quantifiers are now allowed for
1136        parenthesized assertions.
1137    
1138    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1139    
1140    36. \g was being checked for fancy things in a character class, when it should
1141        just be a literal "g".
1142    
1143    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1144        appearance of a nested POSIX class supersedes an apparent external class.
1145        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1146        unescaped square brackets may also appear as part of class names. For
1147        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1148        more like Perl. (But see 8.20/1 above.)
1149    
1150    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1151        was because it thought it was \N{name}, which is not supported).
1152    
1153    39. Add minix to OS list not supporting the -S option in pcretest.
1154    
1155    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1156        cannot analyze patterns in sufficient detail to catch mutual recursions
1157        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1158        subgroup is called recursively as a subpattern for a second time at the
1159        same position in the subject string. In previous releases this might have
1160        been caught by the recursion limit, or it might have run out of stack.
1161    
1162    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1163        happen only once. PCRE was, however incorrectly giving a compile time error
1164        "recursive call could loop indefinitely" because it cannot analyze the
1165        pattern in sufficient detail. The compile time test no longer happens when
1166        PCRE is compiling a conditional subpattern, but actual runaway loops are
1167        now caught at runtime (see 40 above).
1168    
1169    42. It seems that Perl allows any characters other than a closing parenthesis
1170        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1171        has been changed to be the same.
1172    
1173    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1174        as not to get warnings when autogen.sh is called. Also changed
1175        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1176    
1177    44. To help people who use pcregrep to scan files containing exceedingly long
1178        lines, the following changes have been made:
1179    
1180        (a) The default value of the buffer size parameter has been increased from
1181            8K to 20K. (The actual buffer used is three times this size.)
1182    
1183        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1184            PCRE is built.
1185    
1186        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1187            to be set at run time.
1188    
1189        (d) Numerical values in pcregrep options can be followed by K or M, for
1190            example --buffer-size=50K.
1191    
1192        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1193            given and the return code is set to 2.
1194    
1195    45. Add a pointer to the latest mark to the callout data block.
1196    
1197    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1198        partial match of an empty string instead of no match. This was specific to
1199        the use of ".".
1200    
1201    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1202        complete match instead of a partial match. This bug was dependent on both
1203        the PCRE_UTF8 and PCRE_DOTALL options being set.
1204    
1205    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1206        starting byte set, because \b was not being ignored.
1207    
1208    
1209    Version 8.12 15-Jan-2011
1210    ------------------------
1211    
1212    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1213        checks for such things as part of the documentation building process.
1214    
1215    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1216        --match-limit and --recursion-limit options (added for 8.11). In
1217        particular, this made one of the standard tests fail. (The integer value
1218        went into the wrong half of a long int.)
1219    
1220    3.  If the --colour option was given to pcregrep with -v (invert match), it
1221        did strange things, either producing crazy output, or crashing. It should,
1222        of course, ignore a request for colour when reporting lines that do not
1223        match.
1224    
1225    4.  Another pcregrep bug caused similar problems if --colour was specified with
1226        -M (multiline) and the pattern match finished with a line ending.
1227    
1228    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1229        matched in multiline mode, the following line was shown as part of the
1230        match. This seems wrong, so I have changed it.
1231    
1232    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1233        the check for further matches in the same line (so they could be coloured)
1234        to overrun the end of the current line. If another match was found, it was
1235        incorrectly shown (and then shown again when found in the next line).
1236    
1237    7.  If pcregrep was compiled under Windows, there was a reference to the
1238        function pcregrep_exit() before it was defined. I am assuming this was
1239        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1240        reported by a user. I've moved the definition above the reference.
1241    
1242    
1243    Version 8.11 10-Dec-2010
1244    ------------------------
1245    
1246    1.  (*THEN) was not working properly if there were untried alternatives prior
1247        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1248        backtracked to try for "b" instead of moving to the next alternative branch
1249        at the same level (in this case, to look for "c"). The Perl documentation
1250        is clear that when (*THEN) is backtracked onto, it goes to the "next
1251        alternative in the innermost enclosing group".
1252    
1253    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1254        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1255        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1256        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1257        (*THEN).
1258    
1259    3.  If \s appeared in a character class, it removed the VT character from
1260        the class, even if it had been included by some previous item, for example
1261        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1262        of \s, but is part of the POSIX "space" class.)
1263    
1264    4.  A partial match never returns an empty string (because you can always
1265        match an empty string at the end of the subject); however the checking for
1266        an empty string was starting at the "start of match" point. This has been
1267        changed to the "earliest inspected character" point, because the returned
1268        data for a partial match starts at this character. This means that, for
1269        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1270        (previously it gave "no match").
1271    
1272    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1273        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1274        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1275        has an implication that the given string is incomplete (because a partial
1276        match is preferred over a full match). For this reason, these items now
1277        give a partial match in this situation. [Aside: previously, the one case
1278        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1279        match rather than a full match, which was wrong by the old rules, but is
1280        now correct.]
1281    
1282    6.  There was a bug in the handling of #-introduced comments, recognized when
1283        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1284        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1285        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1286        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1287        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1288        places in pcre_compile().
1289    
1290    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1291        comments when looking ahead for named forward references to subpatterns,
1292        the only newline sequence it recognized was NL. It now handles newlines
1293        according to the set newline convention.
1294    
1295    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1296        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1297        cater for a lack of strerror(). These oversights have been fixed.
1298    
1299    9.  Added --match-limit and --recursion-limit to pcregrep.
1300    
1301    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1302    
1303    11. When the -o option was used, pcregrep was setting a return code of 1, even
1304        when matches were found, and --line-buffered was not being honoured.
1305    
1306    12. Added an optional parentheses number to the -o and --only-matching options
1307        of pcregrep.
1308    
1309    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1310        can match an empty string. The code to do it in pcretest and pcredemo
1311        needed fixing:
1312    
1313        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1314            only one byte after an empty string match just before CRLF (this case
1315            just got forgotten; "any" and "anycrlf" were OK).
1316    
1317        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1318            mode when an empty string match preceded an ASCII character followed by
1319            a non-ASCII character. (The code for advancing by one character rather
1320            than one byte was nonsense.)
1321    
1322        (c) The pcredemo.c sample program did not have any code at all to handle
1323            the cases when CRLF is a valid newline sequence.
1324    
1325    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1326        as a starting offset was within the subject string. There is now a new
1327        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1328        negative or greater than the length of the string. In order to test this,
1329        pcretest is extended to allow the setting of negative starting offsets.
1330    
1331    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1332        starting offset points to the beginning of a UTF-8 character was
1333        unnecessarily clumsy. I tidied it up.
1334    
1335    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1336        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1337    
1338    17. Nobody had reported that the --include_dir option, which was added in
1339        release 7.7 should have been called --include-dir (hyphen, not underscore)
1340        for compatibility with GNU grep. I have changed it to --include-dir, but
1341        left --include_dir as an undocumented synonym, and the same for
1342        --exclude-dir, though that is not available in GNU grep, at least as of
1343        release 2.5.4.
1344    
1345    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1346        characters from a string of bytes) have been redefined so as not to use
1347        loops, in order to improve performance in some environments. At the same
1348        time, I abstracted some of the common code into auxiliary macros to save
1349        repetition (this should not affect the compiled code).
1350    
1351    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1352        compile-time error is now given if \c is not followed by an ASCII
1353        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1354        different, and any byte value is allowed.)
1355    
1356    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1357        START_OPTIMIZE option, which is now allowed at compile time - but just
1358        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1359        to pcregrep and other applications that have no direct access to PCRE
1360        options. The new /Y option in pcretest sets this option when calling
1361        pcre_compile().
1362    
1363    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1364        back references. Groups containing recursive back references were forced to
1365        be atomic by that change, but in the case of named groups, the amount of
1366        memory required was incorrectly computed, leading to "Failed: internal
1367        error: code overflow". This has been fixed.
1368    
1369    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1370        pcretest.c, to avoid build problems in some Borland environments.
1371    
1372    
1373    Version 8.10 25-Jun-2010
1374    ------------------------
1375    
1376    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1377        THEN.
1378    
1379    2.  (*ACCEPT) was not working when inside an atomic group.
1380    
1381    3.  Inside a character class, \B is treated as a literal by default, but
1382        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1383        causes the error). The code is unchanged, but I tidied the documentation.
1384    
1385    4.  Inside a character class, PCRE always treated \R and \X as literals,
1386        whereas Perl faults them if its -w option is set. I have changed PCRE so
1387        that it faults them when PCRE_EXTRA is set.
1388    
1389    5.  Added support for \N, which always matches any character other than
1390        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1391    
1392    6.  When compiling pcregrep with newer versions of gcc which may have
1393        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1394        declared with attribute warn_unused_result" were given. Just casting the
1395        result to (void) does not stop the warnings; a more elaborate fudge is
1396        needed. I've used a macro to implement this.
1397    
1398    7.  Minor change to pcretest.c to avoid a compiler warning.
1399    
1400    8.  Added four artifical Unicode properties to help with an option to make
1401        \s etc use properties (see next item). The new properties are: Xan
1402        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1403    
1404    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1405        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1406        this option. Modified pcretest to add /W to test this facility. Added
1407        REG_UCP to make it available via the POSIX interface.
1408    
1409    10. Added --line-buffered to pcregrep.
1410    
1411    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1412        studied, and the match started with a letter with a code point greater than
1413        127 whose first byte was different to the first byte of the other case of
1414        the letter, the other case of this starting letter was not recognized
1415        (#976).
1416    
1417    12. If a pattern that was studied started with a repeated Unicode property
1418        test, for example, \p{Nd}+, there was the theoretical possibility of
1419        setting up an incorrect bitmap of starting bytes, but fortunately it could
1420        not have actually happened in practice until change 8 above was made (it
1421        added property types that matched character-matching opcodes).
1422    
1423    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1424        possible starting bytes for non-anchored patterns.
1425    
1426    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1427        \R, and also a number of cases that involve Unicode properties, both
1428        explicit and implicit when PCRE_UCP is set.
1429    
1430    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1431        input, it could crash or give wrong results if characters with values
1432        greater than 0xc0 were present in the subject string. (Detail: it assumed
1433        UTF-8 input when processing these items.)
1434    
1435    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1436        size_t is 64-bit (#991).
1437    
1438    17. Added a check for running out of memory when PCRE is compiled with
1439        --disable-stack-for-recursion (#990).
1440    
1441    18. If the last data line in a file for pcretest does not have a newline on
1442        the end, a newline was missing in the output.
1443    
1444    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1445        less than 128) in its various bitmaps. However, there is a facility for
1446        generating tables according to the current locale when PCRE is compiled. It
1447        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1448        characters, are recognized by isspace() and therefore were getting set in
1449        these tables, and indeed these tables seem to approximate to ISO 8859. This
1450        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1451        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1452        which of course cannot start UTF-8 characters. I have changed the code so
1453        that only real ASCII characters (less than 128) and the correct starting
1454        bytes for UTF-8 encodings are set for characters greater than 127 when in
1455        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1456        altogether.)
1457    
1458    20. Added the /T option to pcretest so as to be able to run tests with non-
1459        standard character tables, thus making it possible to include the tests
1460        used for 19 above in the standard set of tests.
1461    
1462    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1463        reference to a subpattern the other side of a comment that contains an
1464        opening parenthesis caused either an internal compiling error, or a
1465        reference to the wrong subpattern.
1466    
1467    
1468    Version 8.02 19-Mar-2010
1469    ------------------------
1470    
1471    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1472    
1473    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1474        configured.
1475    
1476    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1477        original author of that file, following a query about its status.
1478    
1479    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1480        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1481    
1482    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1483        quantifier applied to a forward-referencing subroutine call, could compile
1484        incorrect code or give the error "internal error: previously-checked
1485        referenced subpattern not found".
1486    
1487    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1488        variables to point to external functions. For these systems, therefore,
1489        pcre_malloc etc. are now initialized to local functions that call the
1490        relevant global functions.
1491    
1492    7.  There were two entries missing in the vectors called coptable and poptable
1493        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1494        I've fixed the data, and added a kludgy way of testing at compile time that
1495        the lengths are correct (equal to the number of opcodes).
1496    
1497    8.  Following on from 7, I added a similar kludge to check the length of the
1498        eint vector in pcreposix.c.
1499    
1500    9.  Error texts for pcre_compile() are held as one long string to avoid too
1501        much relocation at load time. To find a text, the string is searched,
1502        counting zeros. There was no check for running off the end of the string,
1503        which could happen if a new error number was added without updating the
1504        string.
1505    
1506    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1507    
1508    11. \K was not working if it appeared in an atomic group or in a group that
1509        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1510        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1511        the assertion is positive, but not if it is negative.
1512    
1513    12. Change 11 fortuitously reduced the size of the stack frame used in the
1514        "match()" function of pcre_exec.c by one pointer. Forthcoming
1515        implementation of support for (*MARK) will need an extra pointer on the
1516        stack; I have reserved it now, so that the stack frame size does not
1517        decrease.
1518    
1519    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1520        item in branch that calls a recursion is a subroutine call - as in the
1521        second branch in the above example - was incorrectly given the compile-
1522        time error "recursive call could loop indefinitely" because pcre_compile()
1523        was not correctly checking the subroutine for matching a non-empty string.
1524    
1525    14. The checks for overrunning compiling workspace could trigger after an
1526        overrun had occurred. This is a "should never occur" error, but it can be
1527        triggered by pathological patterns such as hundreds of nested parentheses.
1528        The checks now trigger 100 bytes before the end of the workspace.
1529    
1530    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1531    
1532    
1533    Version 8.01 19-Jan-2010
1534    ------------------------
1535    
1536    1.  If a pattern contained a conditional subpattern with only one branch (in
1537        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1538        computed the wrong minimum data length (which is of course zero for such
1539        subpatterns). This could cause incorrect "no match" results.
1540    
1541    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1542        the pattern is reset in the first branch, pcre_compile() failed with
1543        "internal error: code overflow at offset...". This happened only when
1544        the reset was to the original external option setting. (An optimization
1545        abstracts leading options settings into an external setting, which was the
1546        cause of this.)
1547    
1548    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1549        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1550        assertion pattern did not match (meaning that the assertion was true), it
1551        was incorrectly treated as false if the SKIP had been reached during the
1552        matching. This also applied to assertions used as conditions.
1553    
1554    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1555        assertion subpattern, including such a pattern used as a condition,
1556        unpredictable results occurred, instead of the error return
1557        PCRE_ERROR_DFA_UITEM.
1558    
1559    5.  The C++ GlobalReplace function was not working like Perl for the special
1560        situation when an empty string is matched. It now does the fancy magic
1561        stuff that is necessary.
1562    
1563    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1564        removed. (These were left over from very, very early versions of PCRE.)
1565    
1566    7.  Some cosmetic changes to the code to make life easier when compiling it
1567        as part of something else:
1568    
1569        (a) Change DEBUG to PCRE_DEBUG.
1570    
1571        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1572            called "current" as "current_branch", to prevent a collision with the
1573            Linux macro when compiled as a kernel module.
1574    
1575        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1576            prevent a collision with the Linux macro when compiled as a kernel
1577            module.
1578    
1579    8.  In pcre_compile() there are some checks for integer overflows that used to
1580        cast potentially large values to (double). This has been changed to that
1581        when building, a check for int64_t is made, and if it is found, it is used
1582        instead, thus avoiding the use of floating point arithmetic. (There is no
1583        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1584        double.
1585    
1586    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1587        2005 (difference between two addresses compared to an unsigned value).
1588    
1589    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1590        custom one, because of the following reported problem in Windows:
1591    
1592          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1593              under Win32.
1594          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1595              therefore missing the function definition.
1596          - The compiler thus generates a "C" signature for the test function.
1597          - The linker fails to find the "C" function.
1598          - PCRE fails to configure if asked to do so against libbz2.
1599    
1600    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1601        messages were output:
1602    
1603          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1604          rerunning libtoolize, to keep the correct libtool macros in-tree.
1605          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1606    
1607        I have done both of these things.
1608    
1609    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1610        most of the time, it *can* run out if it is given a pattern that contains a
1611        runaway infinite recursion. I updated the discussion in the pcrestack man
1612        page.
1613    
1614    13. Now that we have gone to the x.xx style of version numbers, the minor
1615        version may start with zero. Using 08 or 09 is a bad idea because users
1616        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1617        interpreted as invalid octal numbers. I've updated the previous comment in
1618        configure.ac, and also added a check that gives an error if 08 or 09 are
1619        used.
1620    
1621    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1622        causing partial matching to fail when the end of the subject matched \W
1623        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1624    
1625    15. There were some discrepancies between the declarations in pcre_internal.h
1626        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1627        their definitions. The declarations used "const uschar *" and the
1628        definitions used USPTR. Even though USPTR is normally defined as "const
1629        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1630        reported that: "This difference in casting confuses some C++ compilers, for
1631        example, SunCC recognizes above declarations as different functions and
1632        generates broken code for hbpcre." I have changed the declarations to use
1633        USPTR.
1634    
1635    16. GNU libtool is named differently on some systems. The autogen.sh script now
1636        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1637        (FreeBSD).
1638    
1639    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1640        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1641        comment: "Figure out how to create a longlong from a string: strtoll and
1642        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1643        instance, but it only takes 2 args instead of 3!"
1644    
1645    18. A subtle bug concerned with back references has been fixed by a change of
1646        specification, with a corresponding code fix. A pattern such as
1647        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1648        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1649        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1650        same bug. Such groups have to be quantified to be useful, or contained
1651        inside another quantified group. (If there's no repetition, the reference
1652        can never match.) The problem arises because, having left the group and
1653        moved on to the rest of the pattern, a later failure that backtracks into
1654        the group uses the captured value from the final iteration of the group
1655        rather than the correct earlier one. I have fixed this in PCRE by forcing
1656        any group that contains a reference to itself to be an atomic group; that
1657        is, there cannot be any backtracking into it once it has completed. This is
1658        similar to recursive and subroutine calls.
1659    
1660    
1661    Version 8.00 19-Oct-09
1662    ----------------------
1663    
1664    1.  The table for translating pcre_compile() error codes into POSIX error codes
1665        was out-of-date, and there was no check on the pcre_compile() error code
1666        being within the table. This could lead to an OK return being given in
1667        error.
1668    
1669    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1670        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1671        in a Windows environment.
1672    
1673    3.  The pcregrep --count option prints the count for each file even when it is
1674        zero, as does GNU grep. However, pcregrep was also printing all files when
1675        --files-with-matches was added. Now, when both options are given, it prints
1676        counts only for those files that have at least one match. (GNU grep just
1677        prints the file name in this circumstance, but including the count seems
1678        more useful - otherwise, why use --count?) Also ensured that the
1679        combination -clh just lists non-zero counts, with no names.
1680    
1681    4.  The long form of the pcregrep -F option was incorrectly implemented as
1682        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1683        but it seems right to fix it, and I didn't think it was worth preserving
1684        the old behaviour.
1685    
1686    5.  The command line items --regex=pattern and --regexp=pattern were not
1687        recognized by pcregrep, which required --regex pattern or --regexp pattern
1688        (with a space rather than an '='). The man page documented the '=' forms,
1689        which are compatible with GNU grep; these now work.
1690    
1691    6.  No libpcreposix.pc file was created for pkg-config; there was just
1692        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1693    
1694    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1695        when UCP support is not needed, by modifying the Python script that
1696        generates it from Unicode data files. This should not matter if the module
1697        is correctly used as a library, but I received one complaint about 50K of
1698        unwanted data. My guess is that the person linked everything into his
1699        program rather than using a library. Anyway, it does no harm.
1700    
1701    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1702        was a minimum greater than 1 for a wide character in a possessive
1703        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1704        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1705        character. Chaos in the form of incorrect output or a compiling loop could
1706        result.
1707    
1708    9.  The restrictions on what a pattern can contain when partial matching is
1709        requested for pcre_exec() have been removed. All patterns can now be
1710        partially matched by this function. In addition, if there are at least two
1711        slots in the offset vector, the offset of the earliest inspected character
1712        for the match and the offset of the end of the subject are set in them when
1713        PCRE_ERROR_PARTIAL is returned.
1714    
1715    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1716        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1717        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1718        and may be more useful for multi-segment matching.
1719    
1720    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1721        used to be given if ever the end of the subject was reached; now it is
1722        given only if matching could not proceed because another character was
1723        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1724        string "Z", which now yields "no match" instead of "partial match". In the
1725        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1726        final character ended with (*FAIL).
1727    
1728    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1729        if the pattern had a "must contain" character that was already found in the
1730        earlier partial match, unless partial matching was again requested. For
1731        example, with the pattern /dog.(body)?/, the "must contain" character is
1732        "g". If the first part-match was for the string "dog", restarting with
1733        "sbody" failed. This bug has been fixed.
1734    
1735    13. The string returned by pcre_dfa_exec() after a partial match has been
1736        changed so that it starts at the first inspected character rather than the
1737        first character of the match. This makes a difference only if the pattern
1738        starts with a lookbehind assertion or \b or \B (\K is not supported by
1739        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1740        matching functions compatible, and I think it's the right thing to do.
1741    
1742    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1743        so that the demonstration program is easily available in environments where
1744        PCRE has not been installed from source.
1745    
1746    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1747        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1748        library.
1749    
1750    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1751        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1752        is not the first non-POSIX option to be added. Clearly some people find
1753        these options useful.
1754    
1755    17. If a caller to the POSIX matching function regexec() passes a non-zero
1756        value for nmatch with a NULL value for pmatch, the value of
1757        nmatch is forced to zero.
1758    
1759    18. RunGrepTest did not have a test for the availability of the -u option of
1760        the diff command, as RunTest does. It now checks in the same way as
1761        RunTest, and also checks for the -b option.
1762    
1763    19. If an odd number of negated classes containing just a single character
1764        interposed, within parentheses, between a forward reference to a named
1765        subpattern and the definition of the subpattern, compilation crashed with
1766        an internal error, complaining that it could not find the referenced
1767        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1768        [The bug was that it was starting one character too far in when skipping
1769        over the character class, thus treating the ] as data rather than
1770        terminating the class. This meant it could skip too much.]
1771    
1772    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1773        /g option in pcretest when the pattern contains \K, which makes it possible
1774        to have an empty string match not at the start, even when the pattern is
1775        anchored. Updated pcretest and pcredemo to use this option.
1776    
1777    21. If the maximum number of capturing subpatterns in a recursion was greater
1778        than the maximum at the outer level, the higher number was returned, but
1779        with unset values at the outer level. The correct (outer level) value is
1780        now given.
1781    
1782    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1783        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1784        make it do so. The string so far is captured, making this feature
1785        compatible with Perl.
1786    
1787    23. The tests have been re-organized, adding tests 11 and 12, to make it
1788        possible to check the Perl 5.10 features against Perl 5.10.
1789    
1790    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1791        pattern matches a fixed length string. PCRE did not allow this; now it
1792        does. Neither allows recursion.
1793    
1794    25. I finally figured out how to implement a request to provide the minimum
1795        length of subject string that was needed in order to match a given pattern.
1796        (It was back references and recursion that I had previously got hung up
1797        on.) This code has now been added to pcre_study(); it finds a lower bound
1798        to the length of subject needed. It is not necessarily the greatest lower
1799        bound, but using it to avoid searching strings that are too short does give
1800        some useful speed-ups. The value is available to calling programs via
1801        pcre_fullinfo().
1802    
1803    26. While implementing 25, I discovered to my embarrassment that pcretest had
1804        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1805        study optimizations had never been tested with that matching function.
1806        Oops. What is worse, even when it was passed study data, there was a bug in
1807        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1808        were also very few tests of studied patterns with pcre_dfa_exec().
1809    
1810    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1811        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1812        on the other side of the coin, they are no longer allowed to have different
1813        names, because these cannot be distinguished in PCRE, and this has caused
1814        confusion. (This is a difference from Perl.)
1815    
1816    28. When duplicate subpattern names are present (necessarily with different
1817        numbers, as required by 27 above), and a test is made by name in a
1818        conditional pattern, either for a subpattern having been matched, or for
1819        recursion in such a pattern, all the associated numbered subpatterns are
1820        tested, and the overall condition is true if the condition is true for any
1821        one of them. This is the way Perl works, and is also more like the way
1822        testing by number works.
1823    
1824    
1825    Version 7.9 11-Apr-09
1826    ---------------------
1827    
1828    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1829        (pcretest), all targets were linked against these libraries. This included
1830        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1831        libraries. This caused unwanted dependencies to be created. This problem
1832        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1833        pcretest is linked with readline.
1834    
1835    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1836        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1837        moved outside it again, because FALSE and TRUE are already defined in AIX,
1838        but BOOL is not.
1839    
1840    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1841        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1842    
1843    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1844        hyphens) following file names and line numbers when outputting matching
1845        lines. This is not true; no spaces are inserted. I have also clarified the
1846        wording for the --colour (or --color) option.
1847    
1848    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1849        was not coloured; this is different to GNU grep, so I have changed it to be
1850        the same.
1851    
1852    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1853        each matching line was coloured. Now it goes on to look for further matches
1854        of any of the test patterns, which is the same behaviour as GNU grep.
1855    
1856    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1857        doesn't make sense to accept an empty string match in pcregrep, so I have
1858        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1859        seems to be how GNU grep behaves. [But see later change 40 for release
1860        8.33.]
1861    
1862    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1863        start or after a newline", because the conditional assertion was not being
1864        correctly handled. The rule now is that both the assertion and what follows
1865        in the first alternative must satisfy the test.
1866    
1867    9.  If auto-callout was enabled in a pattern with a conditional group whose
1868        condition was an assertion, PCRE could crash during matching, both with
1869        pcre_exec() and pcre_dfa_exec().
1870    
1871    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1872        used for matching.
1873    
1874    11. Unicode property support in character classes was not working for
1875        characters (bytes) greater than 127 when not in UTF-8 mode.
1876    
1877    12. Added the -M command line option to pcretest.
1878    
1879    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1880    
1881    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1882    
1883    16. Added comments and documentation about mis-use of no_arg in the C++
1884        wrapper.
1885    
1886    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1887        from Martin Jerabek that uses macro names for all relevant character and
1888        string constants.
1889    
1890    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1891        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1892        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1893        these, but not everybody uses configure.
1894    
1895    19. A conditional group that had only one branch was not being correctly
1896        recognized as an item that could match an empty string. This meant that an
1897        enclosing group might also not be so recognized, causing infinite looping
1898        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1899        with the subject "ab", where knowledge that the repeated group can match
1900        nothing is needed in order to break the loop.
1901    
1902    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1903        exec(), but without supplying a callout function, matching went wrong.
1904    
1905    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1906        leak if the size of the offset vector was greater than 30. When the vector
1907        is smaller, the saved offsets during recursion go onto a local stack
1908        vector, but for larger vectors malloc() is used. It was failing to free
1909        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1910        error, in fact).
1911    
1912    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1913        heapframe that is used only when UTF-8 support is enabled. This caused no
1914        problem, but was untidy.
1915    
1916    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1917        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1918        included within another project.
1919    
1920    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1921        slightly modified by me:
1922    
1923          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1924              not building pcregrep.
1925    
1926          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1927              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1928    
1929    25. Forward references, both numeric and by name, in patterns that made use of
1930        duplicate group numbers, could behave incorrectly or give incorrect errors,
1931        because when scanning forward to find the reference group, PCRE was not
1932        taking into account the duplicate group numbers. A pattern such as
1933        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1934    
1935    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1936        the feature of a custom pointer more persuasive (as requested by a user).
1937    
1938    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1939        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1940        pre-defined.
1941    
1942    28. Added support for (*UTF8) at the start of a pattern.
1943    
1944    29. Arrange for flags added by the "release type" setting in CMake to be shown
1945        in the configuration summary.
1946    
1947    
1948    Version 7.8 05-Sep-08
1949    ---------------------
1950    
1951    1.  Replaced UCP searching code with optimized version as implemented for Ad
1952        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1953        stage table and inline lookup instead of a function, giving speed ups of 2
1954        to 5 times on some simple patterns that I tested. Permission was given to
1955        distribute the MultiStage2.py script that generates the tables (it's not in
1956        the tarball, but is in the Subversion repository).
1957    
1958    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1959        scripts.
1960    
1961    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1962        a group with a zero qualifier. The result of the study could be incorrect,
1963        or the function might crash, depending on the pattern.
1964    
1965    4.  Caseless matching was not working for non-ASCII characters in back
1966        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1967        It now works when Unicode Property Support is available.
1968    
1969    5.  In pcretest, an escape such as \x{de} in the data was always generating
1970        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1971        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1972        truncation.
1973    
1974    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1975    
1976    7.  Added two (int) casts to pcregrep when printing the difference of two
1977        pointers, in case they are 64-bit values.
1978    
1979    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1980        test 2 if it fails.
1981    
1982    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1983        and a #define of that name to empty if it is not externally set. This is to
1984        allow users of MSVC to set it if necessary.
1985    
1986    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1987        the convenience functions in the pcre_get.c source file.
1988    
1989    11. An option change at the start of a pattern that had top-level alternatives
1990        could cause overwriting and/or a crash. This command provoked a crash in
1991        some environments:
1992    
1993          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1994    
1995        This potential security problem was recorded as CVE-2008-2371.
1996    
1997    12. For a pattern where the match had to start at the beginning or immediately
1998        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1999        pcre_dfa_exec() could read past the end of the passed subject if there was
2000        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2001        pcretest so that it places the subject at the end of its malloc-ed buffer.
2002    
2003    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2004        exec() might read past the end of the data buffer in UTF-8 mode.
2005    
2006    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2007        the data contained the byte 0x85 as part of a UTF-8 character within its
2008        first line. This applied both to normal and DFA matching.
2009    
2010    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2011        /^[^d]*?$/8 failed to match "abc".
2012    
2013    16. Added a missing copyright notice to pcrecpp_internal.h.
2014    
2015    17. Make it more clear in the documentation that values returned from
2016        pcre_exec() in ovector are byte offsets, not character counts.
2017    
2018    18. Tidied a few places to stop certain compilers from issuing warnings.
2019    
2020    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2021        supplied by Stefan Weber. I made a further small update for 7.8 because
2022        there is a change of source arrangements: the pcre_searchfuncs.c module is
2023        replaced by pcre_ucd.c.
2024    
2025    
2026    Version 7.7 07-May-08
2027    ---------------------
2028    
2029    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2030        a string to a long long, pretend we don't even have a long long." This is
2031        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2032    
2033    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2034        pre-7.6 versions, which defined a global no_arg variable instead of putting
2035        it in the RE class. (See also #8 below.)
2036    
2037    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2038        Lopes.
2039    
2040    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2041    
2042        (1) The include/exclude patterns were being applied to the whole pathnames
2043            of files, instead of just to the final components.
2044    
2045        (2) If there was more than one level of directory, the subdirectories were
2046            skipped unless they satisfied the include/exclude conditions. This is
2047            inconsistent with GNU grep (and could even be seen as contrary to the
2048            pcregrep specification - which I improved to make it absolutely clear).
2049            The action now is always to scan all levels of directory, and just
2050            apply the include/exclude patterns to regular files.
2051    
2052    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2053        --exclude_dir in the tests to avoid scanning .svn directories.
2054    
2055    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2056        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2057        doesn't support NULs in patterns.
2058    
2059    7.  Added some missing "const"s to declarations of static tables in
2060        pcre_compile.c and pcre_dfa_exec.c.
2061    
2062    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2063        caused by fix #2  above. (Subsequently also a second patch to fix the
2064        first patch. And a third patch - this was a messy problem.)
2065    
2066    9.  Applied Craig's patch to remove the use of push_back().
2067    
2068    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2069        matching function regexec().
2070    
2071    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2072        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2073        references. PCRE supports relative numbers with this syntax (I don't think
2074        Oniguruma does).
2075    
2076    12. Previously, a group with a zero repeat such as (...){0} was completely
2077        omitted from the compiled regex. However, this means that if the group
2078        was called as a subroutine from elsewhere in the pattern, things went wrong
2079        (an internal error was given). Such groups are now left in the compiled
2080        pattern, with a new opcode that causes them to be skipped at execution
2081        time.
2082    
2083    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2084        to the way PCRE behaves:
2085    
2086        (a) A lone ] character is dis-allowed (Perl treats it as data).
2087    
2088        (b) A back reference to an unmatched subpattern matches an empty string
2089            (Perl fails the current match path).
2090    
2091        (c) A data ] in a character class must be notated as \] because if the
2092            first data character in a class is ], it defines an empty class. (In
2093            Perl it is not possible to have an empty class.) The empty class []
2094            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2095            The negative empty class [^] matches any one character, independently
2096            of the DOTALL setting.
2097    
2098    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2099        non-existent subpattern following a character class starting with ']' and
2100        containing () gave an internal compiling error instead of "reference to
2101        non-existent subpattern". Fortunately, when the pattern did exist, the
2102        compiled code was correct. (When scanning forwards to check for the
2103        existence of the subpattern, it was treating the data ']' as terminating
2104        the class, so got the count wrong. When actually compiling, the reference
2105        was subsequently set up correctly.)
2106    
2107    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2108        it was being rejected as not supported by pcre_dfa_exec(), even though
2109        other assertions are supported. I have made pcre_dfa_exec() support
2110        (*FAIL).
2111    
2112    16. The implementation of 13c above involved the invention of a new opcode,
2113        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2114        cannot be changed at match time, I realized I could make a small
2115        improvement to matching performance by compiling OP_ALLANY instead of
2116        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2117        on the OP_ANY path.
2118    
2119    17. Compiling pcretest on Windows with readline support failed without the
2120        following two fixes: (1) Make the unistd.h include conditional on
2121        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2122    
2123    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2124        ncurses library to be included for pcretest when ReadLine support is
2125        requested, but also to allow for it to be overridden. This patch came from
2126        Daniel Bergström.
2127    
2128    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2129        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2130        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2131        spotting this.
2132    
2133    
2134    Version 7.6 28-Jan-08
2135    ---------------------
2136    
2137    1.  A character class containing a very large number of characters with
2138        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2139        overflow.
2140    
2141    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2142        HAVE_LONG_LONG is not defined.
2143    
2144    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2145        bring them up to date and include new features. This patch includes:
2146    
2147        - Fixed PH's badly added libz and libbz2 support.
2148        - Fixed a problem with static linking.
2149        - Added pcredemo. [But later removed - see 7 below.]
2150        - Fixed dftables problem and added an option.
2151        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2152            HAVE_LONG_LONG.
2153        - Added readline support for pcretest.
2154        - Added an listing of the option settings after cmake has run.
2155    
2156    4.  A user submitted a patch to Makefile that makes it easy to create
2157        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2158        Makefile.am that cause it to include this special target, without
2159        affecting anything else. Note that the same mingw target plus all
2160        the other distribution libraries and programs are now supported
2161        when configuring with CMake (see 6 below) instead of with
2162        Configure/Make.
2163    
2164    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2165        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2166        exported in the Windows port". It has not yet been confirmed that the patch
2167        solves the problem, but it does no harm.
2168    
2169    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2170        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2171        with CMake, and also correct the comment about stack recursion.
2172    
2173    7.  Remove the automatic building of pcredemo from the ./configure system and
2174        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2175        of a program that users should build themselves after PCRE is installed, so
2176        building it automatically is not really right. What is more, it gave
2177        trouble in some build environments.
2178    
2179    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2180    
2181    
2182    Version 7.5 10-Jan-08
2183    ---------------------
2184    
2185    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2186        values in parens when parsing an RE using the C++ wrapper."
2187    
2188    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2189        Characters greater than 255 were excluded from the class instead of being
2190        included.
2191    
2192    3.  The same bug as (2) above applied to negated POSIX classes such as
2193        [:^space:].
2194    
2195    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2196        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2197        I have changed it.
2198    
2199    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2200        first named subpattern) and a construct such as (?&a) would reference the
2201        first named subpattern whose name started with "a" (in other words, the
2202        length check was missing). Both these problems are fixed. "Subpattern name
2203        expected" is now given for (?&) (a zero-length name), and this patch also
2204        makes it give the same error for \k'' (previously it complained that that
2205        was a reference to a non-existent subpattern).
2206    
2207    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2208        this is right because (?- can be followed by option settings as well as by
2209        digits. I have, however, made the messages clearer.
2210    
2211    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2212        than the number used in the conditional) now cause a compile-time error.
2213        This is actually not compatible with Perl, which accepts such patterns, but
2214        treats the conditional as always being FALSE (as PCRE used to), but it
2215        seems to me that giving a diagnostic is better.
2216    
2217    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2218        and messages.
2219    
2220    9.  Fix two occurrences of "backslash" in comments that should have been
2221        "backspace".
2222    
2223    10. Remove two redundant lines of code that can never be obeyed (their function
2224        was moved elsewhere).
2225    
2226    11. The program that makes PCRE's Unicode character property table had a bug
2227        which caused it to generate incorrect table entries for sequences of
2228        characters that have the same character type, but are in different scripts.
2229        It amalgamated them into a single range, with the script of the first of
2230        them. In other words, some characters were in the wrong script. There were
2231        thirteen such cases, affecting characters in the following ranges:
2232    
2233          U+002b0 - U+002c1
2234          U+0060c - U+0060d
2235          U+0061e - U+00612
2236          U+0064b - U+0065e
2237          U+0074d - U+0076d
2238          U+01800 - U+01805
2239          U+01d00 - U+01d77
2240          U+01d9b - U+01dbf
2241          U+0200b - U+0200f
2242          U+030fc - U+030fe
2243          U+03260 - U+0327f
2244          U+0fb46 - U+0fbb1
2245          U+10450 - U+1049d
2246    
2247    12. The -o option (show only the matching part of a line) for pcregrep was not
2248        compatible with GNU grep in that, if there was more than one match in a
2249        line, it showed only the first of them. It now behaves in the same way as
2250        GNU grep.
2251    
2252    13. If the -o and -v options were combined for pcregrep, it printed a blank
2253        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2254        does the same. The return code can be used to tell if there were any
2255        non-matching lines.
2256    
2257    14. Added --file-offsets and --line-offsets to pcregrep.
2258    
2259    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2260        infinitely looping recursion. The bug was that positive lookaheads were not
2261        being skipped when checking for a possible empty match (negative lookaheads
2262        and both kinds of lookbehind were skipped).
2263    
2264    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2265        inclusion of <windows.h> to before rather than after the definition of
2266        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2267    
2268    17. Specifying a possessive quantifier with a specific limit for a Unicode
2269        character property caused pcre_compile() to compile bad code, which led at
2270        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2271        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2272        caused the error; without that there was no problem.
2273    
2274    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2275    
2276    19. Added --enable-pcretest-libreadline.
2277    
2278    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2279        RE::GlobalReplace(). As a result, the number of replacements returned was
2280        double what it should be. I removed one of the increments, but Craig sent a
2281        later patch that removed the other one (the right fix) and added unit tests
2282        that check the return values (which was not done before).
2283    
2284    21. Several CMake things:
2285    
2286        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2287            the names libpcre and libpcreposix, not just pcre and pcreposix.
2288    
2289        (2) The above change means that pcretest and pcregrep are now correctly
2290            linked with the newly-built libraries, not previously installed ones.
2291    
2292        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2293    
2294    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2295        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2296        UTF-8 newline character). The key issue is that the pattern starts .*;
2297        this means that the match must be either at the beginning, or after a
2298        newline. The bug was in the code for advancing after a failed match and
2299        checking that the new position followed a newline. It was not taking
2300        account of UTF-8 characters correctly.
2301    
2302    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2303        character classes. PCRE was not treating the sequence [:...:] as a
2304        character class unless the ... were all letters. Perl, however, seems to
2305        allow any characters between [: and :], though of course it rejects as
2306        unknown any "names" that contain non-letters, because all the known class
2307        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2308        for example, whereas PCRE did not - it did not recognize a POSIX character
2309        class. This seemed a bit dangerous, so the code has been changed to be
2310        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2311        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2312        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2313        Perl does, and where it didn't before.
2314    
2315    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2316        Windows environments %n is disabled by default.
2317    
2318    
2319    Version 7.4 21-Sep-07
2320    ---------------------
2321    
2322    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2323        means that a class such as [\s] counted as "explicit reference to CR or
2324        LF". That isn't really right - the whole point of the change was to try to
2325        help when there was an actual mention of one of the two characters. So now
2326        the change happens only if \r or \n (or a literal CR or LF) character is
2327        encountered.
2328    
2329    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2330        of both had grown to the point where there were only 3 bits left.
2331        Fortunately, there was spare space in the data structure, and so I have
2332        moved the internal flags into a new 16-bit field to free up more option
2333        bits.
2334    
2335    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2336        but did not set the internal JCHANGED flag - either of these is enough to
2337        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2338        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2339        start sets both bits.
2340    
2341    4.  Added options (at build time, compile time, exec time) to change \R from
2342        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2343    
2344    5.  doc/pcresyntax.html was missing from the distribution.
2345    
2346    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2347        compatibility, even though it is no longer used.
2348    
2349    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2350        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2351        windows.h file is present (where different names are used). [This was
2352        reversed later after testing - see 16 below.]
2353    
2354    8.  Changed all #include <config.h> to #include "config.h". There were also
2355        some further <pcre.h> cases that I changed to "pcre.h".
2356    
2357    9.  When pcregrep was used with the --colour option, it missed the line ending
2358        sequence off the lines that it output.
2359    
2360    10. It was pointed out to me that arrays of string pointers cause lots of
2361        relocations when a shared library is dynamically loaded. A technique of
2362        using a single long string with a table of offsets can drastically reduce
2363        these. I have refactored PCRE in four places to do this. The result is
2364        dramatic:
2365    
2366          Originally:                          290
2367          After changing UCP table:            187
2368          After changing error message table:   43
2369          After changing table of "verbs"       36
2370          After changing table of Posix names   22
2371    
2372        Thanks to the folks working on Gregex for glib for this insight.
2373    
2374    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2375        unicode-properties was also set.
2376    
2377    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2378    
2379    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2380        checked only for CRLF.
2381    
2382    14. Added casts to pcretest.c to avoid compiler warnings.
2383    
2384    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2385    
2386    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2387        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2388        entirely. This removes changes made in 7 above.
2389    
2390    17. The CMake files have been updated, and there is now more information about
2391        building with CMake in the NON-UNIX-USE document.
2392    
2393    
2394    Version 7.3 28-Aug-07
2395  ---------------------  ---------------------
2396    
2397   1. In the rejigging of the build system that eventually resulted in 7.1, the   1. In the rejigging of the build system that eventually resulted in 7.1, the
# Line 29  Version 7.3 05-Jul-07 Line 2419  Version 7.3 05-Jul-07
2419   4. Debugging: adjusted the names of opcodes for different kinds of parentheses   4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2420      in debug output.      in debug output.
2421    
2422   5. Arrange to use "%I64d" instead of "%lld" for long printing in the pcrecpp   5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2423      unittest when running under MinGW.      long printing in the pcrecpp unittest when running under MinGW.
2424    
2425   6. ESC_K was left out of the EBCDIC table.   6. ESC_K was left out of the EBCDIC table.
2426    
# Line 53  Version 7.3 05-Jul-07 Line 2443  Version 7.3 05-Jul-07
2443      pcreposix and pcrecpp parts of the library. These were overlooked when this      pcreposix and pcrecpp parts of the library. These were overlooked when this
2444      problem was solved for the main library.      problem was solved for the main library.
2445    
2446     9. There were some crude static tests to avoid integer overflow when computing
2447        the size of patterns that contain repeated groups with explicit upper
2448        limits. As the maximum quantifier is 65535, the maximum group length was
2449        set at 30,000 so that the product of these two numbers did not overflow a
2450        32-bit integer. However, it turns out that people want to use groups that
2451        are longer than 30,000 bytes (though not repeat them that many times).
2452        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2453        made it possible to implement the integer overflow checks in a much more
2454        dynamic way, which I have now done. The artificial limitation on group
2455        length has been removed - we now have only the limit on the total length of
2456        the compiled pattern, which depends on the LINK_SIZE setting.
2457    
2458    10. Fixed a bug in the documentation for get/copy named substring when
2459        duplicate names are permitted. If none of the named substrings are set, the
2460        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2461        empty string.
2462    
2463    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2464        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2465        because the ] is interpreted as the first data character and the
2466        terminating ] is not found. PCRE has been made compatible with Perl in this
2467        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2468        cause memory overwriting.
2469    
2470    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2471        string has been matched (to stop an infinite loop). It was not recognizing
2472        a conditional subpattern that could match an empty string if that
2473        subpattern was within another subpattern. For example, it looped when
2474        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2475        condition was not nested. This bug has been fixed.
2476    
2477    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2478        past the start of the subject in the presence of bytes with the top bit
2479        set, for example "\x8aBCD".
2480    
2481    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2482        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2483    
2484    14. Optimized (?!) to (*FAIL).
2485    
2486    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2487        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2488        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2489        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2490        does: it's just the validity check that is more restrictive.
2491    
2492    16. Inserted checks for integer overflows during escape sequence (backslash)
2493        processing, and also fixed erroneous offset values for syntax errors during
2494        backslash processing.
2495    
2496    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2497        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2498    
2499    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2500        caused an overrun.
2501    
2502    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2503        something other than just ASCII characters) inside a group that had an
2504        unlimited repeat caused a loop at compile time (while checking to see
2505        whether the group could match an empty string).
2506    
2507    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2508        [\P{Any}] did so. (Error in the code for printing property names.)
2509    
2510    21. An orphan \E inside a character class could cause a crash.
2511    
2512    22. A repeated capturing bracket such as (A)? could cause a wild memory
2513        reference during compilation.
2514    
2515    23. There are several functions in pcre_compile() that scan along a compiled
2516        expression for various reasons (e.g. to see if it's fixed length for look
2517        behind). There were bugs in these functions when a repeated \p or \P was
2518        present in the pattern. These operators have additional parameters compared
2519        with \d, etc, and these were not being taken into account when moving along
2520        the compiled data. Specifically:
2521    
2522        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2523            length.
2524    
2525        (b) An item such as \pL+ within a repeated group could cause crashes or
2526            loops.
2527    
2528        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2529            "reference to non-existent subpattern" error.
2530    
2531        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2532    
2533    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2534        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2535    
2536    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2537    
2538    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2539        character were causing crashes (broken optimization).
2540    
2541    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2542        \p or \P) caused a compile-time loop.
2543    
2544    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2545        break. For example, the unstudied pattern [\r\n]A does not match the string
2546        "\r\nA" because change 7.0/46 below moves the current point on by two
2547        characters after failing to match at the start. However, the pattern \nA
2548        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2549        the same is true. There doesn't seem any very clean way out of this, but
2550        what I have chosen to do makes the common cases work: PCRE now takes note
2551        of whether there can be an explicit match for \r or \n anywhere in the
2552        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2553        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2554        pattern has explicit CR or LF references.
2555    
2556    29. Added (*CR) etc for changing newline setting at start of pattern.
2557    
2558    
2559  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07
2560  ---------------------  ---------------------

Legend:
Removed from v.200  
changed lines
  Added in v.1397

  ViewVC Help
Powered by ViewVC 1.1.5