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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1396 - (show annotations)
Sun Nov 10 19:04:34 2013 UTC (5 years, 11 months ago) by ph10
File size: 256799 byte(s)
In /x mode, allow white space before a possessive + character.
1 ChangeLog for PCRE
2 ------------------
3
4 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
182 Version 8.33 28-May-2013
183 ------------------------
184
185 1. Added 'U' to some constants that are compared to unsigned integers, to
186 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
187 variables that are added to signed variables, to ensure the result is
188 signed and can be negated.
189
190 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
191 CMake config files.
192
193 3. Revise the creation of config.h.generic so that all boolean macros are
194 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
195 overriding via -D on the command line possible.
196
197 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
198 to unsigned int is reported to make a quite noticeable speed difference in
199 a specific Windows environment. Testing on Linux did also appear to show
200 some benefit (and it is clearly not harmful). Also fixed the definition of
201 Xop which should be unsigned.
202
203 5. Related to (4), changing the definition of the intermediate variable cc
204 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
205 improvements.
206
207 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
208 unnecessary spaces.
209
210 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
211 and later.
212
213 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
214
215 9. Optimizing fast_forward_start_bits in JIT.
216
217 10. Adding support for callouts in JIT, and fixing some issues revealed
218 during this work. Namely:
219
220 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
221
222 (b) Minimum length was not checked before the matching is started.
223
224 11. The value of capture_last that is passed to callouts was incorrect in some
225 cases when there was a capture on one path that was subsequently abandoned
226 after a backtrack. Also, the capture_last value is now reset after a
227 recursion, since all captures are also reset in this case.
228
229 12. The interpreter no longer returns the "too many substrings" error in the
230 case when an overflowing capture is in a branch that is subsequently
231 abandoned after a backtrack.
232
233 13. In the pathological case when an offset vector of size 2 is used, pcretest
234 now prints out the matched string after a yield of 0 or 1.
235
236 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
237 Only supported by the JIT compiler at the moment.
238
239 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
240
241 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
242 offset of the starting point of the matching process, provided the offsets
243 vector is large enough.
244
245 17. The \A escape now records a lookbehind value of 1, though its execution
246 does not actually inspect the previous character. This is to ensure that,
247 in partial multi-segment matching, at least one character from the old
248 segment is retained when a new segment is processed. Otherwise, if there
249 are no lookbehinds in the pattern, \A might match incorrectly at the start
250 of a new segment.
251
252 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
253
254 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
255 gives some modest performance improvement in 8-bit mode.
256
257 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
258 be expressed in certain programming languages using Universal Character
259 Names.
260
261 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
262 which points out that "non characters" are not "characters that may not
263 appear in Unicode strings" but rather "characters that are reserved for
264 internal use and have only local meaning".
265
266 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
267 there was a conditional group that depended on an assertion, if the
268 assertion was false, the callout that immediately followed the alternation
269 in the condition was skipped when pcre_exec() was used for matching.
270
271 23. Allow an explicit callout to be inserted before an assertion that is the
272 condition for a conditional group, for compatibility with automatic
273 callouts, which always insert a callout at this point.
274
275 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
276 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
277
278 25. (*PRUNE) is now supported by the JIT compiler.
279
280 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
281
282 27. Fix the case where there are two or more SKIPs with arguments that may be
283 ignored.
284
285 28. (*SKIP) is now supported by the JIT compiler.
286
287 29. (*THEN) is now supported by the JIT compiler.
288
289 30. Update RunTest with additional test selector options.
290
291 31. The way PCRE handles backtracking verbs has been changed in two ways.
292
293 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
294 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
295 backtracking. In some cases this makes it more Perl-compatible, but Perl's
296 rather obscure rules do not always do the same thing.
297
298 (2) Previously, backtracking verbs were confined within assertions. This is
299 no longer the case for positive assertions, except for (*ACCEPT). Again,
300 this sometimes improves Perl compatibility, and sometimes does not.
301
302 32. A number of tests that were in test 2 because Perl did things differently
303 have been moved to test 1, because either Perl or PCRE has changed, and
304 these tests are now compatible.
305
306 32. Backtracking control verbs are now handled in the same way in JIT and
307 interpreter.
308
309 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
310 contained a forward subroutine reference caused a compile error.
311
312 34. Auto-detect and optimize limited repetitions in JIT.
313
314 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
315 blocking (*UTF) etc.
316
317 36. In the interpreter, maximizing pattern repetitions for characters and
318 character types now use tail recursion, which reduces stack usage.
319
320 37. The value of the max lookbehind was not correctly preserved if a compiled
321 and saved regex was reloaded on a host of different endianness.
322
323 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
324 of the compiled pattern block, expand the flags field from 16 to 32 bits
325 because it was almost full.
326
327 39. Try madvise first before posix_madvise.
328
329 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
330 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
331 original change locked out all matches of empty strings. This has been
332 changed so that one match of an empty string per line is recognized.
333 Subsequent searches on the same line (for colouring or for --only-matching,
334 for example) do not recognize empty strings.
335
336 41. Applied a user patch to fix a number of spelling mistakes in comments.
337
338 42. Data lines longer than 65536 caused pcretest to crash.
339
340 43. Clarified the data type for length and startoffset arguments for pcre_exec
341 and pcre_dfa_exec in the function-specific man pages, where they were
342 explicitly stated to be in bytes, never having been updated. I also added
343 some clarification to the pcreapi man page.
344
345 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
346 a segmentation fault.
347
348
349 Version 8.32 30-November-2012
350 -----------------------------
351
352 1. Improved JIT compiler optimizations for first character search and single
353 character iterators.
354
355 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
356 Patch by Daniel Richard G.
357
358 3. Single character iterator optimizations in the JIT compiler.
359
360 4. Improved JIT compiler optimizations for character ranges.
361
362 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
363 Reported by Giuseppe D'Angelo.
364
365 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
366 of a line, was being set incorrectly in cases where .* appeared inside
367 atomic brackets at the start of a pattern, or where there was a subsequent
368 *PRUNE or *SKIP.
369
370 7. Improved instruction cache flush for POWER/PowerPC.
371 Patch by Daniel Richard G.
372
373 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
374 grep:
375
376 (a) There is now no limit to the number of patterns to be matched.
377
378 (b) An error is given if a pattern is too long.
379
380 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
381 are now supported.
382
383 (d) --exclude-from and --include-from (multiple use) have been added.
384
385 (e) Exclusions and inclusions now apply to all files and directories, not
386 just to those obtained from scanning a directory recursively.
387
388 (f) Multiple uses of -f and --file-list are now supported.
389
390 (g) In a Windows environment, the default for -d has been changed from
391 "read" (the GNU grep default) to "skip", because otherwise the presence
392 of a directory in the file list provokes an error.
393
394 (h) The documentation has been revised and clarified in places.
395
396 9. Improve the matching speed of capturing brackets.
397
398 10. Changed the meaning of \X so that it now matches a Unicode extended
399 grapheme cluster.
400
401 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
402 out POSIX threads when JIT support is configured.
403
404 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
405
406 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
407 structure could go wrong in environments where size_t is not the same size
408 as int.
409
410 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
411 set.
412
413 15. The EBCDIC support had decayed; later updates to the code had included
414 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
415 general tidy up of EBCDIC-related issues, and the documentation was also
416 not quite right. There is now a test that can be run on ASCII systems to
417 check some of the EBCDIC-related things (but is it not a full test).
418
419 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
420 in a small tidy to the code.
421
422 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
423
424 18. If the --only-matching (-o) option in pcregrep is specified multiple
425 times, each one causes appropriate output. For example, -o1 -o2 outputs the
426 substrings matched by the 1st and 2nd capturing parentheses. A separating
427 string can be specified by --om-separator (default empty).
428
429 19. Improving the first n character searches.
430
431 20. Turn case lists for horizontal and vertical white space into macros so that
432 they are defined only once.
433
434 21. This set of changes together give more compatible Unicode case-folding
435 behaviour for characters that have more than one other case when UCP
436 support is available.
437
438 (a) The Unicode property table now has offsets into a new table of sets of
439 three or more characters that are case-equivalent. The MultiStage2.py
440 script that generates these tables (the pcre_ucd.c file) now scans
441 CaseFolding.txt instead of UnicodeData.txt for character case
442 information.
443
444 (b) The code for adding characters or ranges of characters to a character
445 class has been abstracted into a generalized function that also handles
446 case-independence. In UTF-mode with UCP support, this uses the new data
447 to handle characters with more than one other case.
448
449 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
450 whose other case is greater than 256 are now correctly matched
451 caselessly. Previously, the high codepoint matched the low one, but not
452 vice versa.
453
454 (d) The processing of \h, \H, \v, and \ in character classes now makes use
455 of the new class addition function, using character lists defined as
456 macros alongside the case definitions of 20 above.
457
458 (e) Caseless back references now work with characters that have more than
459 one other case.
460
461 (f) General caseless matching of characters with more than one other case
462 is supported.
463
464 22. Unicode character properties were updated from Unicode 6.2.0
465
466 23. Improved CMake support under Windows. Patch by Daniel Richard G.
467
468 24. Add support for 32-bit character strings, and UTF-32
469
470 25. Major JIT compiler update (code refactoring and bugfixing).
471 Experimental Sparc 32 support is added.
472
473 26. Applied a modified version of Daniel Richard G's patch to create
474 pcre.h.generic and config.h.generic by "make" instead of in the
475 PrepareRelease script.
476
477 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
478 pcre_compile.c when checking for a zero character.
479
480 28. Introducing a native interface for JIT. Through this interface, the compiled
481 machine code can be directly executed. The purpose of this interface is to
482 provide fast pattern matching, so several sanity checks are not performed.
483 However, feature tests are still performed. The new interface provides
484 1.4x speedup compared to the old one.
485
486 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
487 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
488 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
489
490 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
491 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
492 the "old" RFC 2279). Instead, it ended up passing a negative length to
493 pcre_exec().
494
495 31. Add support for GCC's visibility feature to hide internal functions.
496
497 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
498 "unknown -C option" after outputting 0 or 1.
499
500 33. There is now support for generating a code coverage report for the test
501 suite in environments where gcc is the compiler and lcov is installed. This
502 is mainly for the benefit of the developers.
503
504 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
505 unaddressable using valgrind annotations, allowing valgrind to detect
506 invalid memory accesses. This is mainly for the benefit of the developers.
507
508 25. (*UTF) can now be used to start a pattern in any of the three libraries.
509
510 26. Give configure error if --enable-cpp but no C++ compiler found.
511
512
513 Version 8.31 06-July-2012
514 -------------------------
515
516 1. Fixing a wrong JIT test case and some compiler warnings.
517
518 2. Removed a bashism from the RunTest script.
519
520 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
521 to unsigned type, result still unsigned" that was given by an MS compiler
522 on encountering the code "-sizeof(xxx)".
523
524 4. Partial matching support is added to the JIT compiler.
525
526 5. Fixed several bugs concerned with partial matching of items that consist
527 of more than one character:
528
529 (a) /^(..)\1/ did not partially match "aba" because checking references was
530 done on an "all or nothing" basis. This also applied to repeated
531 references.
532
533 (b) \R did not give a hard partial match if \r was found at the end of the
534 subject.
535
536 (c) \X did not give a hard partial match after matching one or more
537 characters at the end of the subject.
538
539 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
540 a partial match for the string "\r".
541
542 (e) When newline was set to CRLF, the metacharacter "." did not recognize
543 a partial match for a CR character at the end of the subject string.
544
545 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
546 running pcretest, the text "(JIT)" added to the output whenever JIT is
547 actually used to run the match.
548
549 7. Individual JIT compile options can be set in pcretest by following -s+[+]
550 or /S+[+] with a digit between 1 and 7.
551
552 8. OP_NOT now supports any UTF character not just single-byte ones.
553
554 9. (*MARK) control verb is now supported by the JIT compiler.
555
556 10. The command "./RunTest list" lists the available tests without actually
557 running any of them. (Because I keep forgetting what they all are.)
558
559 11. Add PCRE_INFO_MAXLOOKBEHIND.
560
561 12. Applied a (slightly modified) user-supplied patch that improves performance
562 when the heap is used for recursion (compiled with --disable-stack-for-
563 recursion). Instead of malloc and free for each heap frame each time a
564 logical recursion happens, frames are retained on a chain and re-used where
565 possible. This sometimes gives as much as 30% improvement.
566
567 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
568 call.
569
570 14. As documented, (*COMMIT) is now confined to within a positive assertion.
571
572 15. It is now possible to link pcretest with libedit as an alternative to
573 libreadline.
574
575 16. (*COMMIT) control verb is now supported by the JIT compiler.
576
577 17. The Unicode data tables have been updated to Unicode 6.1.0.
578
579 18. Added --file-list option to pcregrep.
580
581 19. Added binary file support to pcregrep, including the -a, --binary-files,
582 -I, and --text options.
583
584 20. The madvise function is renamed for posix_madvise for QNX compatibility
585 reasons. Fixed by Giuseppe D'Angelo.
586
587 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
588
589 22. Changed the option for creating symbolic links for 16-bit man pages from
590 -s to -sf so that re-installing does not cause issues.
591
592 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
593
594 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
595 matches in certain environments (the workspace was not being correctly
596 retained). Also added to pcre_dfa_exec() a simple plausibility check on
597 some of the workspace data at the beginning of a restart.
598
599 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
600 was not doing so when it should - probably a typo introduced by SVN 528
601 (change 8.10/14).
602
603 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
604 \w+ when the character tables indicated that \x{c4} was a word character.
605 There were several related cases, all because the tests for doing a table
606 lookup were testing for characters less than 127 instead of 255.
607
608 27. If a pattern contains capturing parentheses that are not used in a match,
609 their slots in the ovector are set to -1. For those that are higher than
610 any matched groups, this happens at the end of processing. In the case when
611 there were back references that the ovector was too small to contain
612 (causing temporary malloc'd memory to be used during matching), and the
613 highest capturing number was not used, memory off the end of the ovector
614 was incorrectly being set to -1. (It was using the size of the temporary
615 memory instead of the true size.)
616
617 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
618 ovector size, it uses memory at the end of the block that it has got.
619
620 29. Check for an overlong MARK name and give an error at compile time. The
621 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
622
623 30. JIT compiler update.
624
625 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
626 Rill for the patch.
627
628 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
629
630 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
631
632 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
633 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
634 was enabled.
635
636 35. Improve JIT code generation for greedy plus quantifier.
637
638 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
639 1 to "aa" instead of to an empty string. The bug affected repeated groups
640 that could potentially match an empty string.
641
642 37. Optimizing single character iterators in JIT.
643
644 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
645 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
646 codepoints that are too big for the mode are faulted, and in a UTF mode,
647 disallowed codepoints are also faulted.
648
649 39. If PCRE was compiled with UTF support, in three places in the DFA
650 matcher there was code that should only have been obeyed in UTF mode, but
651 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
652 processing when bytes with values greater than 127 were present. In 16-bit
653 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
654 both cases the values are those that cannot be the first data item in a UTF
655 character. The three items that might have provoked this were recursions,
656 possessively repeated groups, and atomic groups.
657
658 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
659 and pcregrep, because some OS require shared objects to be explicitly
660 passed to ld, causing the link step to fail if they are not.
661
662 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
663 mode, patterns that started with \h* or \R* might be incorrectly matched.
664
665
666 Version 8.30 04-February-2012
667 -----------------------------
668
669 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
670 name is defined in ctype.h.
671
672 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
673 only in quite long subpatterns.
674
675 3. Removed the function pcre_info(), which has been obsolete and deprecated
676 since it was replaced by pcre_fullinfo() in February 2000.
677
678 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
679 match a (*MARK), and the match failed at the start of the subject, a
680 reference to memory before the start of the subject could occur. This bug
681 was introduced by fix 17 of release 8.21.
682
683 5. A reference to an unset group with zero minimum repetition was giving
684 totally wrong answers (in non-JavaScript-compatibility mode). For example,
685 /(another)?(\1?)test/ matched against "hello world test". This bug was
686 introduced in release 8.13.
687
688 6. Add support for 16-bit character strings (a large amount of work involving
689 many changes and refactorings).
690
691 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
692 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
693 from a file.
694
695 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
696 rounding is not applied in this particular case).
697
698 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
699 if they appear, or are escaped, in patterns.
700
701 10. Get rid of a number of -Wunused-but-set-variable warnings.
702
703 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
704 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
705 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
706 also returns the mark "x". This bug applied to capturing parentheses,
707 non-capturing parentheses, and atomic parentheses. It also applied to some
708 assertions.
709
710 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
711 information out of configure.ac instead of relying on pcre.h.generic, which
712 is not stored in the repository.
713
714 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
715 -lreadline.
716
717 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
718
719 15. Applied Graycode's patch to put the top-level frame on the stack rather
720 than the heap when not using the stack for recursion. This gives a
721 performance improvement in many cases when recursion is not deep.
722
723 16. Experimental code added to "pcretest -C" to output the stack frame size.
724
725
726 Version 8.21 12-Dec-2011
727 ------------------------
728
729 1. Updating the JIT compiler.
730
731 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
732 are added as well.
733
734 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
735 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
736 calling _pcre_jit_exec. Some extra comments are added.
737
738 4. (*MARK) settings inside atomic groups that do not contain any capturing
739 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
740 was introduced by change 18 for 8.20.
741
742 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
743 ECMA-262 standard.
744
745 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
746 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
747 This bug was probably introduced by change 9 of 8.13.
748
749 7. While fixing 6 above, I noticed that a number of other items were being
750 incorrectly rejected as "not fixed length". This arose partly because newer
751 opcodes had not been added to the fixed-length checking code. I have (a)
752 corrected the bug and added tests for these items, and (b) arranged for an
753 error to occur if an unknown opcode is encountered while checking for fixed
754 length instead of just assuming "not fixed length". The items that were
755 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
756 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
757 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
758
759 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
760 being incorrectly compiled and would have given unpredicatble results.
761
762 9. A possessively repeated subpattern with minimum repeat count greater than
763 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
764 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
765 the first (A) could occur when it should not.
766
767 10. Add a cast and remove a redundant test from the code.
768
769 11. JIT should use pcre_malloc/pcre_free for allocation.
770
771 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
772 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
773 is anything other than /usr, -L is still shown).
774
775 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
776
777 14. Perl does not support \N without a following name in a [] class; PCRE now
778 also gives an error.
779
780 15. If a forward reference was repeated with an upper limit of around 2000,
781 it caused the error "internal error: overran compiling workspace". The
782 maximum number of forward references (including repeats) was limited by the
783 internal workspace, and dependent on the LINK_SIZE. The code has been
784 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
785 the default depends on LINK_SIZE. There is a new upper limit (for safety)
786 of around 200,000 forward references. While doing this, I also speeded up
787 the filling in of repeated forward references.
788
789 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
790 incorrectly expecting the subject to contain another "a" after the start.
791
792 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
793 in the match, the SKIP should be ignored. This was not happening; instead
794 the SKIP was being treated as NOMATCH. For patterns such as
795 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
796 tested.
797
798 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
799 now much more compatible with Perl, in particular in cases where the result
800 is a non-match for a non-anchored pattern. For example, if
801 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
802 "m", where previously it did not return a name. A side effect of this
803 change is that for partial matches, the last encountered mark name is
804 returned, as for non matches. A number of tests that were previously not
805 Perl-compatible have been moved into the Perl-compatible test files. The
806 refactoring has had the pleasing side effect of removing one argument from
807 the match() function, thus reducing its stack requirements.
808
809 19. If the /S+ option was used in pcretest to study a pattern using JIT,
810 subsequent uses of /S (without +) incorrectly behaved like /S+.
811
812 21. Retrieve executable code size support for the JIT compiler and fixing
813 some warnings.
814
815 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
816 not work when the shorter character appeared right at the end of the
817 subject string.
818
819 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
820 systems.
821
822 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
823 output it when the /M option is used in pcretest.
824
825 25. The CheckMan script was not being included in the distribution. Also, added
826 an explicit "perl" to run Perl scripts from the PrepareRelease script
827 because this is reportedly needed in Windows.
828
829 26. If study data was being save in a file and studying had not found a set of
830 "starts with" bytes for the pattern, the data written to the file (though
831 never used) was taken from uninitialized memory and so caused valgrind to
832 complain.
833
834 27. Updated RunTest.bat as provided by Sheri Pierce.
835
836 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
837
838 29. Computation of memory usage for the table of capturing group names was
839 giving an unnecessarily large value.
840
841
842 Version 8.20 21-Oct-2011
843 ------------------------
844
845 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
846 a POSIX class. After further experiments with Perl, which convinced me that
847 Perl has bugs and confusions, a closing square bracket is no longer allowed
848 in a POSIX name. This bug also affected patterns with classes that started
849 with full stops.
850
851 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
852 captured substring, but while checking the failing first alternative,
853 substring 1 is temporarily captured. If the output vector supplied to
854 pcre_exec() was not big enough for this capture, the yield of the function
855 was still zero ("insufficient space for captured substrings"). This cannot
856 be totally fixed without adding another stack variable, which seems a lot
857 of expense for a edge case. However, I have improved the situation in cases
858 such as /(a)(b)x|abc/ matched against "abc", where the return code
859 indicates that fewer than the maximum number of slots in the ovector have
860 been set.
861
862 3. Related to (2) above: when there are more back references in a pattern than
863 slots in the output vector, pcre_exec() uses temporary memory during
864 matching, and copies in the captures as far as possible afterwards. It was
865 using the entire output vector, but this conflicts with the specification
866 that only 2/3 is used for passing back captured substrings. Now it uses
867 only the first 2/3, for compatibility. This is, of course, another edge
868 case.
869
870 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
871 main code base, and can be used by building with --enable-jit. When this is
872 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
873 runtime --no-jit option is given.
874
875 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
876 ovector, the return from the function was zero, implying that there were
877 other matches that did not fit. The correct "exactly full" value is now
878 returned.
879
880 6. If a subpattern that was called recursively or as a subroutine contained
881 (*PRUNE) or any other control that caused it to give a non-standard return,
882 invalid errors such as "Error -26 (nested recursion at the same subject
883 position)" or even infinite loops could occur.
884
885 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
886 computing the minimum length on reaching *ACCEPT, and so ended up with the
887 wrong value of 1 rather than 0. Further investigation indicates that
888 computing a minimum subject length in the presence of *ACCEPT is difficult
889 (think back references, subroutine calls), and so I have changed the code
890 so that no minimum is registered for a pattern that contains *ACCEPT.
891
892 8. If (*THEN) was present in the first (true) branch of a conditional group,
893 it was not handled as intended. [But see 16 below.]
894
895 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
896 Sheri Pierce.
897
898 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
899 the first byte in a match must be "a".
900
901 11. Change 17 for 8.13 increased the recursion depth for patterns like
902 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
903 pattern contains any instances of (*THEN). If it does not, the old
904 optimizations are restored. It would be nice to do this on a per-group
905 basis, but at the moment that is not feasible.
906
907 12. In some environments, the output of pcretest -C is CRLF terminated. This
908 broke RunTest's code that checks for the link size. A single white space
909 character after the value is now allowed for.
910
911 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
912 For "fr", it uses the Windows-specific input and output files.
913
914 14. If (*THEN) appeared in a group that was called recursively or as a
915 subroutine, it did not work as intended. [But see next item.]
916
917 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
918 pattern fragments (but not containing any | characters). If A and B are
919 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
920 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
921 D. In other words, Perl considers parentheses that do not contain any |
922 characters to be part of a surrounding alternative, whereas PCRE was
923 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
924 differently. PCRE now behaves in the same way as Perl, except in the case
925 of subroutine/recursion calls such as (?1) which have in any case always
926 been different (but PCRE had them first :-).
927
928 16. Related to 15 above: Perl does not treat the | in a conditional group as
929 creating alternatives. Such a group is treated in the same way as an
930 ordinary group without any | characters when processing (*THEN). PCRE has
931 been changed to match Perl's behaviour.
932
933 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
934 RunGrepTest script failed.
935
936 18. Change 22 for version 13 caused atomic groups to use more stack. This is
937 inevitable for groups that contain captures, but it can lead to a lot of
938 stack use in large patterns. The old behaviour has been restored for atomic
939 groups that do not contain any capturing parentheses.
940
941 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
942 suppress the check for a minimum subject length at run time. (If it was
943 given to pcre_exec() or pcre_dfa_exec() it did work.)
944
945 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
946 fail to work when decoding hex characters in data strings in EBCDIC
947 environments.
948
949 21. It appears that in at least one Mac OS environment, the isxdigit() function
950 is implemented as a macro that evaluates to its argument more than once,
951 contravening the C 90 Standard (I haven't checked a later standard). There
952 was an instance in pcretest which caused it to go wrong when processing
953 \x{...} escapes in subject strings. The has been rewritten to avoid using
954 things like p++ in the argument of isxdigit().
955
956
957 Version 8.13 16-Aug-2011
958 ------------------------
959
960 1. The Unicode data tables have been updated to Unicode 6.0.0.
961
962 2. Two minor typos in pcre_internal.h have been fixed.
963
964 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
965 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
966 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
967
968 4. There were a number of related bugs in the code for matching backrefences
969 caselessly in UTF-8 mode when codes for the characters concerned were
970 different numbers of bytes. For example, U+023A and U+2C65 are an upper
971 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
972 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
973 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
974 2-byte code at the end of the subject (it thought there wasn't enough data
975 left).
976
977 5. Comprehensive information about what went wrong is now returned by
978 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
979 as the output vector has at least 2 elements. The offset of the start of
980 the failing character and a reason code are placed in the vector.
981
982 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
983 now returned is for the first byte of the failing character, instead of the
984 last byte inspected. This is an incompatible change, but I hope it is small
985 enough not to be a problem. It makes the returned offset consistent with
986 pcre_exec() and pcre_dfa_exec().
987
988 7. pcretest now gives a text phrase as well as the error number when
989 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
990 failure, the offset and reason code are output.
991
992 8. When \R was used with a maximizing quantifier it failed to skip backwards
993 over a \r\n pair if the subsequent match failed. Instead, it just skipped
994 back over a single character (\n). This seems wrong (because it treated the
995 two characters as a single entity when going forwards), conflicts with the
996 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
997 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
998 has been changed.
999
1000 9. Some internal refactoring has changed the processing so that the handling
1001 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1002 time (the PCRE_DOTALL option was changed this way some time ago: version
1003 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1004 which was always a bit of a fudge. It also means that there is one less
1005 argument for the match() function, which reduces its stack requirements
1006 slightly. This change also fixes an incompatibility with Perl: the pattern
1007 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1008
1009 10. More internal refactoring has drastically reduced the number of recursive
1010 calls to match() for possessively repeated groups such as (abc)++ when
1011 using pcre_exec().
1012
1013 11. While implementing 10, a number of bugs in the handling of groups were
1014 discovered and fixed:
1015
1016 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1017 (a|)*(?1) gave a compile-time internal error.
1018 ((a|)+)+ did not notice that the outer group could match an empty string.
1019 (^a|^)+ was not marked as anchored.
1020 (.*a|.*)+ was not marked as matching at start or after a newline.
1021
1022 12. Yet more internal refactoring has removed another argument from the match()
1023 function. Special calls to this function are now indicated by setting a
1024 value in a variable in the "match data" data block.
1025
1026 13. Be more explicit in pcre_study() instead of relying on "default" for
1027 opcodes that mean there is no starting character; this means that when new
1028 ones are added and accidentally left out of pcre_study(), testing should
1029 pick them up.
1030
1031 14. The -s option of pcretest has been documented for ages as being an old
1032 synonym of -m (show memory usage). I have changed it to mean "force study
1033 for every regex", that is, assume /S for every regex. This is similar to -i
1034 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1035 using it. It makes it easier to run collections of tests with and without
1036 study enabled, and thereby test pcre_study() more easily. All the standard
1037 tests are now run with and without -s (but some patterns can be marked as
1038 "never study" - see 20 below).
1039
1040 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1041 restoration of the capturing data to the outer values was not happening
1042 correctly.
1043
1044 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1045 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1046 pattern had matched an empty string, and so incorrectly returned a no
1047 match.
1048
1049 17. There was optimizing code for the last branch of non-capturing parentheses,
1050 and also for the obeyed branch of a conditional subexpression, which used
1051 tail recursion to cut down on stack usage. Unfortunately, now that there is
1052 the possibility of (*THEN) occurring in these branches, tail recursion is
1053 no longer possible because the return has to be checked for (*THEN). These
1054 two optimizations have therefore been removed. [But see 8.20/11 above.]
1055
1056 18. If a pattern containing \R was studied, it was assumed that \R always
1057 matched two bytes, thus causing the minimum subject length to be
1058 incorrectly computed because \R can also match just one byte.
1059
1060 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1061 was incorrectly computed.
1062
1063 20. If /S is present twice on a test pattern in pcretest input, it now
1064 *disables* studying, thereby overriding the use of -s on the command line
1065 (see 14 above). This is necessary for one or two tests to keep the output
1066 identical in both cases.
1067
1068 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1069 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1070
1071 22. When an atomic group that contained a capturing parenthesis was
1072 successfully matched, but the branch in which it appeared failed, the
1073 capturing was not being forgotten if a higher numbered group was later
1074 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1075 group 1 to "a", when in fact it should be unset. This applied to multi-
1076 branched capturing and non-capturing groups, repeated or not, and also to
1077 positive assertions (capturing in negative assertions does not happen
1078 in PCRE) and also to nested atomic groups.
1079
1080 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1081 subject after a captured substring, to make it easier to tell which of a
1082 number of identical substrings has been captured.
1083
1084 24. The way atomic groups are processed by pcre_exec() has been changed so that
1085 if they are repeated, backtracking one repetition now resets captured
1086 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1087 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1088 "aaa". Previously, it would have been "a". As part of this code
1089 refactoring, the way recursive calls are handled has also been changed.
1090
1091 25. If an assertion condition captured any substrings, they were not passed
1092 back unless some other capturing happened later. For example, if
1093 (?(?=(a))a) was matched against "a", no capturing was returned.
1094
1095 26. When studying a pattern that contained subroutine calls or assertions,
1096 the code for finding the minimum length of a possible match was handling
1097 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1098 group 1 called group 2 while simultaneously a separate group 2 called group
1099 1). A stack overflow occurred in this case. I have fixed this by limiting
1100 the recursion depth to 10.
1101
1102 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1103 Fortmann. This supports explicit test numbers on the command line, and has
1104 argument validation and error reporting.
1105
1106 28. An instance of \X with an unlimited repeat could fail if at any point the
1107 first character it looked at was a mark character.
1108
1109 29. Some minor code refactoring concerning Unicode properties and scripts
1110 should reduce the stack requirement of match() slightly.
1111
1112 30. Added the '=' option to pcretest to check the setting of unused capturing
1113 slots at the end of the pattern, which are documented as being -1, but are
1114 not included in the return count.
1115
1116 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1117 compiled something random. Now it gives a compile-time error (as does
1118 Perl).
1119
1120 32. A *MARK encountered during the processing of a positive assertion is now
1121 recorded and passed back (compatible with Perl).
1122
1123 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1124 had alternative anchored branches, the search for a second match in a line
1125 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1126 matched the line "0102" twice. The same bug affected patterns that started
1127 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1128 twice.
1129
1130 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1131 does, and because of capturing effects, quantifying parenthesized
1132 assertions may at times be useful. Quantifiers are now allowed for
1133 parenthesized assertions.
1134
1135 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1136
1137 36. \g was being checked for fancy things in a character class, when it should
1138 just be a literal "g".
1139
1140 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1141 appearance of a nested POSIX class supersedes an apparent external class.
1142 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1143 unescaped square brackets may also appear as part of class names. For
1144 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1145 more like Perl. (But see 8.20/1 above.)
1146
1147 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1148 was because it thought it was \N{name}, which is not supported).
1149
1150 39. Add minix to OS list not supporting the -S option in pcretest.
1151
1152 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1153 cannot analyze patterns in sufficient detail to catch mutual recursions
1154 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1155 subgroup is called recursively as a subpattern for a second time at the
1156 same position in the subject string. In previous releases this might have
1157 been caught by the recursion limit, or it might have run out of stack.
1158
1159 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1160 happen only once. PCRE was, however incorrectly giving a compile time error
1161 "recursive call could loop indefinitely" because it cannot analyze the
1162 pattern in sufficient detail. The compile time test no longer happens when
1163 PCRE is compiling a conditional subpattern, but actual runaway loops are
1164 now caught at runtime (see 40 above).
1165
1166 42. It seems that Perl allows any characters other than a closing parenthesis
1167 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1168 has been changed to be the same.
1169
1170 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1171 as not to get warnings when autogen.sh is called. Also changed
1172 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1173
1174 44. To help people who use pcregrep to scan files containing exceedingly long
1175 lines, the following changes have been made:
1176
1177 (a) The default value of the buffer size parameter has been increased from
1178 8K to 20K. (The actual buffer used is three times this size.)
1179
1180 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1181 PCRE is built.
1182
1183 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1184 to be set at run time.
1185
1186 (d) Numerical values in pcregrep options can be followed by K or M, for
1187 example --buffer-size=50K.
1188
1189 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1190 given and the return code is set to 2.
1191
1192 45. Add a pointer to the latest mark to the callout data block.
1193
1194 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1195 partial match of an empty string instead of no match. This was specific to
1196 the use of ".".
1197
1198 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1199 complete match instead of a partial match. This bug was dependent on both
1200 the PCRE_UTF8 and PCRE_DOTALL options being set.
1201
1202 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1203 starting byte set, because \b was not being ignored.
1204
1205
1206 Version 8.12 15-Jan-2011
1207 ------------------------
1208
1209 1. Fixed some typos in the markup of the man pages, and wrote a script that
1210 checks for such things as part of the documentation building process.
1211
1212 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1213 --match-limit and --recursion-limit options (added for 8.11). In
1214 particular, this made one of the standard tests fail. (The integer value
1215 went into the wrong half of a long int.)
1216
1217 3. If the --colour option was given to pcregrep with -v (invert match), it
1218 did strange things, either producing crazy output, or crashing. It should,
1219 of course, ignore a request for colour when reporting lines that do not
1220 match.
1221
1222 4. Another pcregrep bug caused similar problems if --colour was specified with
1223 -M (multiline) and the pattern match finished with a line ending.
1224
1225 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1226 matched in multiline mode, the following line was shown as part of the
1227 match. This seems wrong, so I have changed it.
1228
1229 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1230 the check for further matches in the same line (so they could be coloured)
1231 to overrun the end of the current line. If another match was found, it was
1232 incorrectly shown (and then shown again when found in the next line).
1233
1234 7. If pcregrep was compiled under Windows, there was a reference to the
1235 function pcregrep_exit() before it was defined. I am assuming this was
1236 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1237 reported by a user. I've moved the definition above the reference.
1238
1239
1240 Version 8.11 10-Dec-2010
1241 ------------------------
1242
1243 1. (*THEN) was not working properly if there were untried alternatives prior
1244 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1245 backtracked to try for "b" instead of moving to the next alternative branch
1246 at the same level (in this case, to look for "c"). The Perl documentation
1247 is clear that when (*THEN) is backtracked onto, it goes to the "next
1248 alternative in the innermost enclosing group".
1249
1250 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1251 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1252 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1253 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1254 (*THEN).
1255
1256 3. If \s appeared in a character class, it removed the VT character from
1257 the class, even if it had been included by some previous item, for example
1258 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1259 of \s, but is part of the POSIX "space" class.)
1260
1261 4. A partial match never returns an empty string (because you can always
1262 match an empty string at the end of the subject); however the checking for
1263 an empty string was starting at the "start of match" point. This has been
1264 changed to the "earliest inspected character" point, because the returned
1265 data for a partial match starts at this character. This means that, for
1266 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1267 (previously it gave "no match").
1268
1269 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1270 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1271 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1272 has an implication that the given string is incomplete (because a partial
1273 match is preferred over a full match). For this reason, these items now
1274 give a partial match in this situation. [Aside: previously, the one case
1275 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1276 match rather than a full match, which was wrong by the old rules, but is
1277 now correct.]
1278
1279 6. There was a bug in the handling of #-introduced comments, recognized when
1280 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1281 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1282 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1283 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1284 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1285 places in pcre_compile().
1286
1287 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1288 comments when looking ahead for named forward references to subpatterns,
1289 the only newline sequence it recognized was NL. It now handles newlines
1290 according to the set newline convention.
1291
1292 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1293 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1294 cater for a lack of strerror(). These oversights have been fixed.
1295
1296 9. Added --match-limit and --recursion-limit to pcregrep.
1297
1298 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1299
1300 11. When the -o option was used, pcregrep was setting a return code of 1, even
1301 when matches were found, and --line-buffered was not being honoured.
1302
1303 12. Added an optional parentheses number to the -o and --only-matching options
1304 of pcregrep.
1305
1306 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1307 can match an empty string. The code to do it in pcretest and pcredemo
1308 needed fixing:
1309
1310 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1311 only one byte after an empty string match just before CRLF (this case
1312 just got forgotten; "any" and "anycrlf" were OK).
1313
1314 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1315 mode when an empty string match preceded an ASCII character followed by
1316 a non-ASCII character. (The code for advancing by one character rather
1317 than one byte was nonsense.)
1318
1319 (c) The pcredemo.c sample program did not have any code at all to handle
1320 the cases when CRLF is a valid newline sequence.
1321
1322 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1323 as a starting offset was within the subject string. There is now a new
1324 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1325 negative or greater than the length of the string. In order to test this,
1326 pcretest is extended to allow the setting of negative starting offsets.
1327
1328 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1329 starting offset points to the beginning of a UTF-8 character was
1330 unnecessarily clumsy. I tidied it up.
1331
1332 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1333 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1334
1335 17. Nobody had reported that the --include_dir option, which was added in
1336 release 7.7 should have been called --include-dir (hyphen, not underscore)
1337 for compatibility with GNU grep. I have changed it to --include-dir, but
1338 left --include_dir as an undocumented synonym, and the same for
1339 --exclude-dir, though that is not available in GNU grep, at least as of
1340 release 2.5.4.
1341
1342 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1343 characters from a string of bytes) have been redefined so as not to use
1344 loops, in order to improve performance in some environments. At the same
1345 time, I abstracted some of the common code into auxiliary macros to save
1346 repetition (this should not affect the compiled code).
1347
1348 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1349 compile-time error is now given if \c is not followed by an ASCII
1350 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1351 different, and any byte value is allowed.)
1352
1353 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1354 START_OPTIMIZE option, which is now allowed at compile time - but just
1355 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1356 to pcregrep and other applications that have no direct access to PCRE
1357 options. The new /Y option in pcretest sets this option when calling
1358 pcre_compile().
1359
1360 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1361 back references. Groups containing recursive back references were forced to
1362 be atomic by that change, but in the case of named groups, the amount of
1363 memory required was incorrectly computed, leading to "Failed: internal
1364 error: code overflow". This has been fixed.
1365
1366 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1367 pcretest.c, to avoid build problems in some Borland environments.
1368
1369
1370 Version 8.10 25-Jun-2010
1371 ------------------------
1372
1373 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1374 THEN.
1375
1376 2. (*ACCEPT) was not working when inside an atomic group.
1377
1378 3. Inside a character class, \B is treated as a literal by default, but
1379 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1380 causes the error). The code is unchanged, but I tidied the documentation.
1381
1382 4. Inside a character class, PCRE always treated \R and \X as literals,
1383 whereas Perl faults them if its -w option is set. I have changed PCRE so
1384 that it faults them when PCRE_EXTRA is set.
1385
1386 5. Added support for \N, which always matches any character other than
1387 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1388
1389 6. When compiling pcregrep with newer versions of gcc which may have
1390 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1391 declared with attribute warn_unused_result" were given. Just casting the
1392 result to (void) does not stop the warnings; a more elaborate fudge is
1393 needed. I've used a macro to implement this.
1394
1395 7. Minor change to pcretest.c to avoid a compiler warning.
1396
1397 8. Added four artifical Unicode properties to help with an option to make
1398 \s etc use properties (see next item). The new properties are: Xan
1399 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1400
1401 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1402 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1403 this option. Modified pcretest to add /W to test this facility. Added
1404 REG_UCP to make it available via the POSIX interface.
1405
1406 10. Added --line-buffered to pcregrep.
1407
1408 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1409 studied, and the match started with a letter with a code point greater than
1410 127 whose first byte was different to the first byte of the other case of
1411 the letter, the other case of this starting letter was not recognized
1412 (#976).
1413
1414 12. If a pattern that was studied started with a repeated Unicode property
1415 test, for example, \p{Nd}+, there was the theoretical possibility of
1416 setting up an incorrect bitmap of starting bytes, but fortunately it could
1417 not have actually happened in practice until change 8 above was made (it
1418 added property types that matched character-matching opcodes).
1419
1420 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1421 possible starting bytes for non-anchored patterns.
1422
1423 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1424 \R, and also a number of cases that involve Unicode properties, both
1425 explicit and implicit when PCRE_UCP is set.
1426
1427 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1428 input, it could crash or give wrong results if characters with values
1429 greater than 0xc0 were present in the subject string. (Detail: it assumed
1430 UTF-8 input when processing these items.)
1431
1432 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1433 size_t is 64-bit (#991).
1434
1435 17. Added a check for running out of memory when PCRE is compiled with
1436 --disable-stack-for-recursion (#990).
1437
1438 18. If the last data line in a file for pcretest does not have a newline on
1439 the end, a newline was missing in the output.
1440
1441 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1442 less than 128) in its various bitmaps. However, there is a facility for
1443 generating tables according to the current locale when PCRE is compiled. It
1444 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1445 characters, are recognized by isspace() and therefore were getting set in
1446 these tables, and indeed these tables seem to approximate to ISO 8859. This
1447 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1448 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1449 which of course cannot start UTF-8 characters. I have changed the code so
1450 that only real ASCII characters (less than 128) and the correct starting
1451 bytes for UTF-8 encodings are set for characters greater than 127 when in
1452 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1453 altogether.)
1454
1455 20. Added the /T option to pcretest so as to be able to run tests with non-
1456 standard character tables, thus making it possible to include the tests
1457 used for 19 above in the standard set of tests.
1458
1459 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1460 reference to a subpattern the other side of a comment that contains an
1461 opening parenthesis caused either an internal compiling error, or a
1462 reference to the wrong subpattern.
1463
1464
1465 Version 8.02 19-Mar-2010
1466 ------------------------
1467
1468 1. The Unicode data tables have been updated to Unicode 5.2.0.
1469
1470 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1471 configured.
1472
1473 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1474 original author of that file, following a query about its status.
1475
1476 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1477 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1478
1479 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1480 quantifier applied to a forward-referencing subroutine call, could compile
1481 incorrect code or give the error "internal error: previously-checked
1482 referenced subpattern not found".
1483
1484 6. Both MS Visual Studio and Symbian OS have problems with initializing
1485 variables to point to external functions. For these systems, therefore,
1486 pcre_malloc etc. are now initialized to local functions that call the
1487 relevant global functions.
1488
1489 7. There were two entries missing in the vectors called coptable and poptable
1490 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1491 I've fixed the data, and added a kludgy way of testing at compile time that
1492 the lengths are correct (equal to the number of opcodes).
1493
1494 8. Following on from 7, I added a similar kludge to check the length of the
1495 eint vector in pcreposix.c.
1496
1497 9. Error texts for pcre_compile() are held as one long string to avoid too
1498 much relocation at load time. To find a text, the string is searched,
1499 counting zeros. There was no check for running off the end of the string,
1500 which could happen if a new error number was added without updating the
1501 string.
1502
1503 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1504
1505 11. \K was not working if it appeared in an atomic group or in a group that
1506 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1507 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1508 the assertion is positive, but not if it is negative.
1509
1510 12. Change 11 fortuitously reduced the size of the stack frame used in the
1511 "match()" function of pcre_exec.c by one pointer. Forthcoming
1512 implementation of support for (*MARK) will need an extra pointer on the
1513 stack; I have reserved it now, so that the stack frame size does not
1514 decrease.
1515
1516 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1517 item in branch that calls a recursion is a subroutine call - as in the
1518 second branch in the above example - was incorrectly given the compile-
1519 time error "recursive call could loop indefinitely" because pcre_compile()
1520 was not correctly checking the subroutine for matching a non-empty string.
1521
1522 14. The checks for overrunning compiling workspace could trigger after an
1523 overrun had occurred. This is a "should never occur" error, but it can be
1524 triggered by pathological patterns such as hundreds of nested parentheses.
1525 The checks now trigger 100 bytes before the end of the workspace.
1526
1527 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1528
1529
1530 Version 8.01 19-Jan-2010
1531 ------------------------
1532
1533 1. If a pattern contained a conditional subpattern with only one branch (in
1534 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1535 computed the wrong minimum data length (which is of course zero for such
1536 subpatterns). This could cause incorrect "no match" results.
1537
1538 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1539 the pattern is reset in the first branch, pcre_compile() failed with
1540 "internal error: code overflow at offset...". This happened only when
1541 the reset was to the original external option setting. (An optimization
1542 abstracts leading options settings into an external setting, which was the
1543 cause of this.)
1544
1545 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1546 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1547 assertion pattern did not match (meaning that the assertion was true), it
1548 was incorrectly treated as false if the SKIP had been reached during the
1549 matching. This also applied to assertions used as conditions.
1550
1551 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1552 assertion subpattern, including such a pattern used as a condition,
1553 unpredictable results occurred, instead of the error return
1554 PCRE_ERROR_DFA_UITEM.
1555
1556 5. The C++ GlobalReplace function was not working like Perl for the special
1557 situation when an empty string is matched. It now does the fancy magic
1558 stuff that is necessary.
1559
1560 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1561 removed. (These were left over from very, very early versions of PCRE.)
1562
1563 7. Some cosmetic changes to the code to make life easier when compiling it
1564 as part of something else:
1565
1566 (a) Change DEBUG to PCRE_DEBUG.
1567
1568 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1569 called "current" as "current_branch", to prevent a collision with the
1570 Linux macro when compiled as a kernel module.
1571
1572 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1573 prevent a collision with the Linux macro when compiled as a kernel
1574 module.
1575
1576 8. In pcre_compile() there are some checks for integer overflows that used to
1577 cast potentially large values to (double). This has been changed to that
1578 when building, a check for int64_t is made, and if it is found, it is used
1579 instead, thus avoiding the use of floating point arithmetic. (There is no
1580 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1581 double.
1582
1583 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1584 2005 (difference between two addresses compared to an unsigned value).
1585
1586 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1587 custom one, because of the following reported problem in Windows:
1588
1589 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1590 under Win32.
1591 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1592 therefore missing the function definition.
1593 - The compiler thus generates a "C" signature for the test function.
1594 - The linker fails to find the "C" function.
1595 - PCRE fails to configure if asked to do so against libbz2.
1596
1597 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1598 messages were output:
1599
1600 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1601 rerunning libtoolize, to keep the correct libtool macros in-tree.
1602 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1603
1604 I have done both of these things.
1605
1606 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1607 most of the time, it *can* run out if it is given a pattern that contains a
1608 runaway infinite recursion. I updated the discussion in the pcrestack man
1609 page.
1610
1611 13. Now that we have gone to the x.xx style of version numbers, the minor
1612 version may start with zero. Using 08 or 09 is a bad idea because users
1613 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1614 interpreted as invalid octal numbers. I've updated the previous comment in
1615 configure.ac, and also added a check that gives an error if 08 or 09 are
1616 used.
1617
1618 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1619 causing partial matching to fail when the end of the subject matched \W
1620 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1621
1622 15. There were some discrepancies between the declarations in pcre_internal.h
1623 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1624 their definitions. The declarations used "const uschar *" and the
1625 definitions used USPTR. Even though USPTR is normally defined as "const
1626 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1627 reported that: "This difference in casting confuses some C++ compilers, for
1628 example, SunCC recognizes above declarations as different functions and
1629 generates broken code for hbpcre." I have changed the declarations to use
1630 USPTR.
1631
1632 16. GNU libtool is named differently on some systems. The autogen.sh script now
1633 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1634 (FreeBSD).
1635
1636 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1637 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1638 comment: "Figure out how to create a longlong from a string: strtoll and
1639 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1640 instance, but it only takes 2 args instead of 3!"
1641
1642 18. A subtle bug concerned with back references has been fixed by a change of
1643 specification, with a corresponding code fix. A pattern such as
1644 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1645 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1646 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1647 same bug. Such groups have to be quantified to be useful, or contained
1648 inside another quantified group. (If there's no repetition, the reference
1649 can never match.) The problem arises because, having left the group and
1650 moved on to the rest of the pattern, a later failure that backtracks into
1651 the group uses the captured value from the final iteration of the group
1652 rather than the correct earlier one. I have fixed this in PCRE by forcing
1653 any group that contains a reference to itself to be an atomic group; that
1654 is, there cannot be any backtracking into it once it has completed. This is
1655 similar to recursive and subroutine calls.
1656
1657
1658 Version 8.00 19-Oct-09
1659 ----------------------
1660
1661 1. The table for translating pcre_compile() error codes into POSIX error codes
1662 was out-of-date, and there was no check on the pcre_compile() error code
1663 being within the table. This could lead to an OK return being given in
1664 error.
1665
1666 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1667 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1668 in a Windows environment.
1669
1670 3. The pcregrep --count option prints the count for each file even when it is
1671 zero, as does GNU grep. However, pcregrep was also printing all files when
1672 --files-with-matches was added. Now, when both options are given, it prints
1673 counts only for those files that have at least one match. (GNU grep just
1674 prints the file name in this circumstance, but including the count seems
1675 more useful - otherwise, why use --count?) Also ensured that the
1676 combination -clh just lists non-zero counts, with no names.
1677
1678 4. The long form of the pcregrep -F option was incorrectly implemented as
1679 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1680 but it seems right to fix it, and I didn't think it was worth preserving
1681 the old behaviour.
1682
1683 5. The command line items --regex=pattern and --regexp=pattern were not
1684 recognized by pcregrep, which required --regex pattern or --regexp pattern
1685 (with a space rather than an '='). The man page documented the '=' forms,
1686 which are compatible with GNU grep; these now work.
1687
1688 6. No libpcreposix.pc file was created for pkg-config; there was just
1689 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1690
1691 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1692 when UCP support is not needed, by modifying the Python script that
1693 generates it from Unicode data files. This should not matter if the module
1694 is correctly used as a library, but I received one complaint about 50K of
1695 unwanted data. My guess is that the person linked everything into his
1696 program rather than using a library. Anyway, it does no harm.
1697
1698 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1699 was a minimum greater than 1 for a wide character in a possessive
1700 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1701 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1702 character. Chaos in the form of incorrect output or a compiling loop could
1703 result.
1704
1705 9. The restrictions on what a pattern can contain when partial matching is
1706 requested for pcre_exec() have been removed. All patterns can now be
1707 partially matched by this function. In addition, if there are at least two
1708 slots in the offset vector, the offset of the earliest inspected character
1709 for the match and the offset of the end of the subject are set in them when
1710 PCRE_ERROR_PARTIAL is returned.
1711
1712 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1713 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1714 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1715 and may be more useful for multi-segment matching.
1716
1717 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1718 used to be given if ever the end of the subject was reached; now it is
1719 given only if matching could not proceed because another character was
1720 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1721 string "Z", which now yields "no match" instead of "partial match". In the
1722 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1723 final character ended with (*FAIL).
1724
1725 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1726 if the pattern had a "must contain" character that was already found in the
1727 earlier partial match, unless partial matching was again requested. For
1728 example, with the pattern /dog.(body)?/, the "must contain" character is
1729 "g". If the first part-match was for the string "dog", restarting with
1730 "sbody" failed. This bug has been fixed.
1731
1732 13. The string returned by pcre_dfa_exec() after a partial match has been
1733 changed so that it starts at the first inspected character rather than the
1734 first character of the match. This makes a difference only if the pattern
1735 starts with a lookbehind assertion or \b or \B (\K is not supported by
1736 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1737 matching functions compatible, and I think it's the right thing to do.
1738
1739 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1740 so that the demonstration program is easily available in environments where
1741 PCRE has not been installed from source.
1742
1743 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1744 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1745 library.
1746
1747 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1748 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1749 is not the first non-POSIX option to be added. Clearly some people find
1750 these options useful.
1751
1752 17. If a caller to the POSIX matching function regexec() passes a non-zero
1753 value for nmatch with a NULL value for pmatch, the value of
1754 nmatch is forced to zero.
1755
1756 18. RunGrepTest did not have a test for the availability of the -u option of
1757 the diff command, as RunTest does. It now checks in the same way as
1758 RunTest, and also checks for the -b option.
1759
1760 19. If an odd number of negated classes containing just a single character
1761 interposed, within parentheses, between a forward reference to a named
1762 subpattern and the definition of the subpattern, compilation crashed with
1763 an internal error, complaining that it could not find the referenced
1764 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1765 [The bug was that it was starting one character too far in when skipping
1766 over the character class, thus treating the ] as data rather than
1767 terminating the class. This meant it could skip too much.]
1768
1769 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1770 /g option in pcretest when the pattern contains \K, which makes it possible
1771 to have an empty string match not at the start, even when the pattern is
1772 anchored. Updated pcretest and pcredemo to use this option.
1773
1774 21. If the maximum number of capturing subpatterns in a recursion was greater
1775 than the maximum at the outer level, the higher number was returned, but
1776 with unset values at the outer level. The correct (outer level) value is
1777 now given.
1778
1779 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1780 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1781 make it do so. The string so far is captured, making this feature
1782 compatible with Perl.
1783
1784 23. The tests have been re-organized, adding tests 11 and 12, to make it
1785 possible to check the Perl 5.10 features against Perl 5.10.
1786
1787 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1788 pattern matches a fixed length string. PCRE did not allow this; now it
1789 does. Neither allows recursion.
1790
1791 25. I finally figured out how to implement a request to provide the minimum
1792 length of subject string that was needed in order to match a given pattern.
1793 (It was back references and recursion that I had previously got hung up
1794 on.) This code has now been added to pcre_study(); it finds a lower bound
1795 to the length of subject needed. It is not necessarily the greatest lower
1796 bound, but using it to avoid searching strings that are too short does give
1797 some useful speed-ups. The value is available to calling programs via
1798 pcre_fullinfo().
1799
1800 26. While implementing 25, I discovered to my embarrassment that pcretest had
1801 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1802 study optimizations had never been tested with that matching function.
1803 Oops. What is worse, even when it was passed study data, there was a bug in
1804 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1805 were also very few tests of studied patterns with pcre_dfa_exec().
1806
1807 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1808 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1809 on the other side of the coin, they are no longer allowed to have different
1810 names, because these cannot be distinguished in PCRE, and this has caused
1811 confusion. (This is a difference from Perl.)
1812
1813 28. When duplicate subpattern names are present (necessarily with different
1814 numbers, as required by 27 above), and a test is made by name in a
1815 conditional pattern, either for a subpattern having been matched, or for
1816 recursion in such a pattern, all the associated numbered subpatterns are
1817 tested, and the overall condition is true if the condition is true for any
1818 one of them. This is the way Perl works, and is also more like the way
1819 testing by number works.
1820
1821
1822 Version 7.9 11-Apr-09
1823 ---------------------
1824
1825 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1826 (pcretest), all targets were linked against these libraries. This included
1827 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1828 libraries. This caused unwanted dependencies to be created. This problem
1829 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1830 pcretest is linked with readline.
1831
1832 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1833 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1834 moved outside it again, because FALSE and TRUE are already defined in AIX,
1835 but BOOL is not.
1836
1837 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1838 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1839
1840 4. The pcregrep documentation said spaces were inserted as well as colons (or
1841 hyphens) following file names and line numbers when outputting matching
1842 lines. This is not true; no spaces are inserted. I have also clarified the
1843 wording for the --colour (or --color) option.
1844
1845 5. In pcregrep, when --colour was used with -o, the list of matching strings
1846 was not coloured; this is different to GNU grep, so I have changed it to be
1847 the same.
1848
1849 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1850 each matching line was coloured. Now it goes on to look for further matches
1851 of any of the test patterns, which is the same behaviour as GNU grep.
1852
1853 7. A pattern that could match an empty string could cause pcregrep to loop; it
1854 doesn't make sense to accept an empty string match in pcregrep, so I have
1855 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1856 seems to be how GNU grep behaves. [But see later change 40 for release
1857 8.33.]
1858
1859 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1860 start or after a newline", because the conditional assertion was not being
1861 correctly handled. The rule now is that both the assertion and what follows
1862 in the first alternative must satisfy the test.
1863
1864 9. If auto-callout was enabled in a pattern with a conditional group whose
1865 condition was an assertion, PCRE could crash during matching, both with
1866 pcre_exec() and pcre_dfa_exec().
1867
1868 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1869 used for matching.
1870
1871 11. Unicode property support in character classes was not working for
1872 characters (bytes) greater than 127 when not in UTF-8 mode.
1873
1874 12. Added the -M command line option to pcretest.
1875
1876 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1877
1878 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1879
1880 16. Added comments and documentation about mis-use of no_arg in the C++
1881 wrapper.
1882
1883 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1884 from Martin Jerabek that uses macro names for all relevant character and
1885 string constants.
1886
1887 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1888 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1889 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1890 these, but not everybody uses configure.
1891
1892 19. A conditional group that had only one branch was not being correctly
1893 recognized as an item that could match an empty string. This meant that an
1894 enclosing group might also not be so recognized, causing infinite looping
1895 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1896 with the subject "ab", where knowledge that the repeated group can match
1897 nothing is needed in order to break the loop.
1898
1899 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1900 exec(), but without supplying a callout function, matching went wrong.
1901
1902 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1903 leak if the size of the offset vector was greater than 30. When the vector
1904 is smaller, the saved offsets during recursion go onto a local stack
1905 vector, but for larger vectors malloc() is used. It was failing to free
1906 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1907 error, in fact).
1908
1909 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1910 heapframe that is used only when UTF-8 support is enabled. This caused no
1911 problem, but was untidy.
1912
1913 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1914 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1915 included within another project.
1916
1917 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1918 slightly modified by me:
1919
1920 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1921 not building pcregrep.
1922
1923 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1924 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1925
1926 25. Forward references, both numeric and by name, in patterns that made use of
1927 duplicate group numbers, could behave incorrectly or give incorrect errors,
1928 because when scanning forward to find the reference group, PCRE was not
1929 taking into account the duplicate group numbers. A pattern such as
1930 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1931
1932 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1933 the feature of a custom pointer more persuasive (as requested by a user).
1934
1935 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1936 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1937 pre-defined.
1938
1939 28. Added support for (*UTF8) at the start of a pattern.
1940
1941 29. Arrange for flags added by the "release type" setting in CMake to be shown
1942 in the configuration summary.
1943
1944
1945 Version 7.8 05-Sep-08
1946 ---------------------
1947
1948 1. Replaced UCP searching code with optimized version as implemented for Ad
1949 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1950 stage table and inline lookup instead of a function, giving speed ups of 2
1951 to 5 times on some simple patterns that I tested. Permission was given to
1952 distribute the MultiStage2.py script that generates the tables (it's not in
1953 the tarball, but is in the Subversion repository).
1954
1955 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1956 scripts.
1957
1958 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1959 a group with a zero qualifier. The result of the study could be incorrect,
1960 or the function might crash, depending on the pattern.
1961
1962 4. Caseless matching was not working for non-ASCII characters in back
1963 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1964 It now works when Unicode Property Support is available.
1965
1966 5. In pcretest, an escape such as \x{de} in the data was always generating
1967 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1968 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1969 truncation.
1970
1971 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1972
1973 7. Added two (int) casts to pcregrep when printing the difference of two
1974 pointers, in case they are 64-bit values.
1975
1976 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1977 test 2 if it fails.
1978
1979 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1980 and a #define of that name to empty if it is not externally set. This is to
1981 allow users of MSVC to set it if necessary.
1982
1983 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1984 the convenience functions in the pcre_get.c source file.
1985
1986 11. An option change at the start of a pattern that had top-level alternatives
1987 could cause overwriting and/or a crash. This command provoked a crash in
1988 some environments:
1989
1990 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1991
1992 This potential security problem was recorded as CVE-2008-2371.
1993
1994 12. For a pattern where the match had to start at the beginning or immediately
1995 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1996 pcre_dfa_exec() could read past the end of the passed subject if there was
1997 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1998 pcretest so that it places the subject at the end of its malloc-ed buffer.
1999
2000 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2001 exec() might read past the end of the data buffer in UTF-8 mode.
2002
2003 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2004 the data contained the byte 0x85 as part of a UTF-8 character within its
2005 first line. This applied both to normal and DFA matching.
2006
2007 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2008 /^[^d]*?$/8 failed to match "abc".
2009
2010 16. Added a missing copyright notice to pcrecpp_internal.h.
2011
2012 17. Make it more clear in the documentation that values returned from
2013 pcre_exec() in ovector are byte offsets, not character counts.
2014
2015 18. Tidied a few places to stop certain compilers from issuing warnings.
2016
2017 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2018 supplied by Stefan Weber. I made a further small update for 7.8 because
2019 there is a change of source arrangements: the pcre_searchfuncs.c module is
2020 replaced by pcre_ucd.c.
2021
2022
2023 Version 7.7 07-May-08
2024 ---------------------
2025
2026 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2027 a string to a long long, pretend we don't even have a long long." This is
2028 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2029
2030 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2031 pre-7.6 versions, which defined a global no_arg variable instead of putting
2032 it in the RE class. (See also #8 below.)
2033
2034 3. Remove a line of dead code, identified by coverity and reported by Nuno
2035 Lopes.
2036
2037 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2038
2039 (1) The include/exclude patterns were being applied to the whole pathnames
2040 of files, instead of just to the final components.
2041
2042 (2) If there was more than one level of directory, the subdirectories were
2043 skipped unless they satisfied the include/exclude conditions. This is
2044 inconsistent with GNU grep (and could even be seen as contrary to the
2045 pcregrep specification - which I improved to make it absolutely clear).
2046 The action now is always to scan all levels of directory, and just
2047 apply the include/exclude patterns to regular files.
2048
2049 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2050 --exclude_dir in the tests to avoid scanning .svn directories.
2051
2052 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2053 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2054 doesn't support NULs in patterns.
2055
2056 7. Added some missing "const"s to declarations of static tables in
2057 pcre_compile.c and pcre_dfa_exec.c.
2058
2059 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2060 caused by fix #2 above. (Subsequently also a second patch to fix the
2061 first patch. And a third patch - this was a messy problem.)
2062
2063 9. Applied Craig's patch to remove the use of push_back().
2064
2065 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2066 matching function regexec().
2067
2068 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2069 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2070 references. PCRE supports relative numbers with this syntax (I don't think
2071 Oniguruma does).
2072
2073 12. Previously, a group with a zero repeat such as (...){0} was completely
2074 omitted from the compiled regex. However, this means that if the group
2075 was called as a subroutine from elsewhere in the pattern, things went wrong
2076 (an internal error was given). Such groups are now left in the compiled
2077 pattern, with a new opcode that causes them to be skipped at execution
2078 time.
2079
2080 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2081 to the way PCRE behaves:
2082
2083 (a) A lone ] character is dis-allowed (Perl treats it as data).
2084
2085 (b) A back reference to an unmatched subpattern matches an empty string
2086 (Perl fails the current match path).
2087
2088 (c) A data ] in a character class must be notated as \] because if the
2089 first data character in a class is ], it defines an empty class. (In
2090 Perl it is not possible to have an empty class.) The empty class []
2091 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2092 The negative empty class [^] matches any one character, independently
2093 of the DOTALL setting.
2094
2095 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2096 non-existent subpattern following a character class starting with ']' and
2097 containing () gave an internal compiling error instead of "reference to
2098 non-existent subpattern". Fortunately, when the pattern did exist, the
2099 compiled code was correct. (When scanning forwards to check for the
2100 existence of the subpattern, it was treating the data ']' as terminating
2101 the class, so got the count wrong. When actually compiling, the reference
2102 was subsequently set up correctly.)
2103
2104 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2105 it was being rejected as not supported by pcre_dfa_exec(), even though
2106 other assertions are supported. I have made pcre_dfa_exec() support
2107 (*FAIL).
2108
2109 16. The implementation of 13c above involved the invention of a new opcode,
2110 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2111 cannot be changed at match time, I realized I could make a small
2112 improvement to matching performance by compiling OP_ALLANY instead of
2113 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2114 on the OP_ANY path.
2115
2116 17. Compiling pcretest on Windows with readline support failed without the
2117 following two fixes: (1) Make the unistd.h include conditional on
2118 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2119
2120 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2121 ncurses library to be included for pcretest when ReadLine support is
2122 requested, but also to allow for it to be overridden. This patch came from
2123 Daniel Bergström.
2124
2125 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2126 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2127 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2128 spotting this.
2129
2130
2131 Version 7.6 28-Jan-08
2132 ---------------------
2133
2134 1. A character class containing a very large number of characters with
2135 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2136 overflow.
2137
2138 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2139 HAVE_LONG_LONG is not defined.
2140
2141 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2142 bring them up to date and include new features. This patch includes:
2143
2144 - Fixed PH's badly added libz and libbz2 support.
2145 - Fixed a problem with static linking.
2146 - Added pcredemo. [But later removed - see 7 below.]
2147 - Fixed dftables problem and added an option.
2148 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2149 HAVE_LONG_LONG.
2150 - Added readline support for pcretest.
2151 - Added an listing of the option settings after cmake has run.
2152
2153 4. A user submitted a patch to Makefile that makes it easy to create
2154 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2155 Makefile.am that cause it to include this special target, without
2156 affecting anything else. Note that the same mingw target plus all
2157 the other distribution libraries and programs are now supported
2158 when configuring with CMake (see 6 below) instead of with
2159 Configure/Make.
2160
2161 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2162 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2163 exported in the Windows port". It has not yet been confirmed that the patch
2164 solves the problem, but it does no harm.
2165
2166 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2167 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2168 with CMake, and also correct the comment about stack recursion.
2169
2170 7. Remove the automatic building of pcredemo from the ./configure system and
2171 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2172 of a program that users should build themselves after PCRE is installed, so
2173 building it automatically is not really right. What is more, it gave
2174 trouble in some build environments.
2175
2176 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2177
2178
2179 Version 7.5 10-Jan-08
2180 ---------------------
2181
2182 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2183 values in parens when parsing an RE using the C++ wrapper."
2184
2185 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2186 Characters greater than 255 were excluded from the class instead of being
2187 included.
2188
2189 3. The same bug as (2) above applied to negated POSIX classes such as
2190 [:^space:].
2191
2192 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2193 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2194 I have changed it.
2195
2196 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2197 first named subpattern) and a construct such as (?&a) would reference the
2198 first named subpattern whose name started with "a" (in other words, the
2199 length check was missing). Both these problems are fixed. "Subpattern name
2200 expected" is now given for (?&) (a zero-length name), and this patch also
2201 makes it give the same error for \k'' (previously it complained that that
2202 was a reference to a non-existent subpattern).
2203
2204 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2205 this is right because (?- can be followed by option settings as well as by
2206 digits. I have, however, made the messages clearer.
2207
2208 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2209 than the number used in the conditional) now cause a compile-time error.
2210 This is actually not compatible with Perl, which accepts such patterns, but
2211 treats the conditional as always being FALSE (as PCRE used to), but it
2212 seems to me that giving a diagnostic is better.
2213
2214 8. Change "alphameric" to the more common word "alphanumeric" in comments
2215 and messages.
2216
2217 9. Fix two occurrences of "backslash" in comments that should have been
2218 "backspace".
2219
2220 10. Remove two redundant lines of code that can never be obeyed (their function
2221 was moved elsewhere).
2222
2223 11. The program that makes PCRE's Unicode character property table had a bug
2224 which caused it to generate incorrect table entries for sequences of
2225 characters that have the same character type, but are in different scripts.
2226 It amalgamated them into a single range, with the script of the first of
2227 them. In other words, some characters were in the wrong script. There were
2228 thirteen such cases, affecting characters in the following ranges:
2229
2230 U+002b0 - U+002c1
2231 U+0060c - U+0060d
2232 U+0061e - U+00612
2233 U+0064b - U+0065e
2234 U+0074d - U+0076d
2235 U+01800 - U+01805
2236 U+01d00 - U+01d77
2237 U+01d9b - U+01dbf
2238 U+0200b - U+0200f
2239 U+030fc - U+030fe
2240 U+03260 - U+0327f
2241 U+0fb46 - U+0fbb1
2242 U+10450 - U+1049d
2243
2244 12. The -o option (show only the matching part of a line) for pcregrep was not
2245 compatible with GNU grep in that, if there was more than one match in a
2246 line, it showed only the first of them. It now behaves in the same way as
2247 GNU grep.
2248
2249 13. If the -o and -v options were combined for pcregrep, it printed a blank
2250 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2251 does the same. The return code can be used to tell if there were any
2252 non-matching lines.
2253
2254 14. Added --file-offsets and --line-offsets to pcregrep.
2255
2256 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2257 infinitely looping recursion. The bug was that positive lookaheads were not
2258 being skipped when checking for a possible empty match (negative lookaheads
2259 and both kinds of lookbehind were skipped).
2260
2261 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2262 inclusion of <windows.h> to before rather than after the definition of
2263 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2264
2265 17. Specifying a possessive quantifier with a specific limit for a Unicode
2266 character property caused pcre_compile() to compile bad code, which led at
2267 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2268 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2269 caused the error; without that there was no problem.
2270
2271 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2272
2273 19. Added --enable-pcretest-libreadline.
2274
2275 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2276 RE::GlobalReplace(). As a result, the number of replacements returned was
2277 double what it should be. I removed one of the increments, but Craig sent a
2278 later patch that removed the other one (the right fix) and added unit tests
2279 that check the return values (which was not done before).
2280
2281 21. Several CMake things:
2282
2283 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2284 the names libpcre and libpcreposix, not just pcre and pcreposix.
2285
2286 (2) The above change means that pcretest and pcregrep are now correctly
2287 linked with the newly-built libraries, not previously installed ones.
2288
2289 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2290
2291 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2292 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2293 UTF-8 newline character). The key issue is that the pattern starts .*;
2294 this means that the match must be either at the beginning, or after a
2295 newline. The bug was in the code for advancing after a failed match and
2296 checking that the new position followed a newline. It was not taking
2297 account of UTF-8 characters correctly.
2298
2299 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2300 character classes. PCRE was not treating the sequence [:...:] as a
2301 character class unless the ... were all letters. Perl, however, seems to
2302 allow any characters between [: and :], though of course it rejects as
2303 unknown any "names" that contain non-letters, because all the known class
2304 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2305 for example, whereas PCRE did not - it did not recognize a POSIX character
2306 class. This seemed a bit dangerous, so the code has been changed to be
2307 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2308 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2309 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2310 Perl does, and where it didn't before.
2311
2312 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2313 Windows environments %n is disabled by default.
2314
2315
2316 Version 7.4 21-Sep-07
2317 ---------------------
2318
2319 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2320 means that a class such as [\s] counted as "explicit reference to CR or
2321 LF". That isn't really right - the whole point of the change was to try to
2322 help when there was an actual mention of one of the two characters. So now
2323 the change happens only if \r or \n (or a literal CR or LF) character is
2324 encountered.
2325
2326 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2327 of both had grown to the point where there were only 3 bits left.
2328 Fortunately, there was spare space in the data structure, and so I have
2329 moved the internal flags into a new 16-bit field to free up more option
2330 bits.
2331
2332 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2333 but did not set the internal JCHANGED flag - either of these is enough to
2334 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2335 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2336 start sets both bits.
2337
2338 4. Added options (at build time, compile time, exec time) to change \R from
2339 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2340
2341 5. doc/pcresyntax.html was missing from the distribution.
2342
2343 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2344 compatibility, even though it is no longer used.
2345
2346 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2347 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2348 windows.h file is present (where different names are used). [This was
2349 reversed later after testing - see 16 below.]
2350
2351 8. Changed all #include <config.h> to #include "config.h". There were also
2352 some further <pcre.h> cases that I changed to "pcre.h".
2353
2354 9. When pcregrep was used with the --colour option, it missed the line ending
2355 sequence off the lines that it output.
2356
2357 10. It was pointed out to me that arrays of string pointers cause lots of
2358 relocations when a shared library is dynamically loaded. A technique of
2359 using a single long string with a table of offsets can drastically reduce
2360 these. I have refactored PCRE in four places to do this. The result is
2361 dramatic:
2362
2363 Originally: 290
2364 After changing UCP table: 187
2365 After changing error message table: 43
2366 After changing table of "verbs" 36
2367 After changing table of Posix names 22
2368
2369 Thanks to the folks working on Gregex for glib for this insight.
2370
2371 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2372 unicode-properties was also set.
2373
2374 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2375
2376 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2377 checked only for CRLF.
2378
2379 14. Added casts to pcretest.c to avoid compiler warnings.
2380
2381 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2382
2383 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2384 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2385 entirely. This removes changes made in 7 above.
2386
2387 17. The CMake files have been updated, and there is now more information about
2388 building with CMake in the NON-UNIX-USE document.
2389
2390
2391 Version 7.3 28-Aug-07
2392 ---------------------
2393
2394 1. In the rejigging of the build system that eventually resulted in 7.1, the
2395 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2396 brackets there is not right, since it causes compilers to look for an
2397 installed pcre.h, not the version that is in the source that is being
2398 compiled (which of course may be different). I have changed it back to:
2399
2400 #include "pcre.h"
2401
2402 I have a vague recollection that the change was concerned with compiling in
2403 different directories, but in the new build system, that is taken care of
2404 by the VPATH setting the Makefile.
2405
2406 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2407 when the subject happened to end in the byte 0x85 (e.g. if the last
2408 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2409 characters but of course it shouldn't be taken as a newline when it is part
2410 of another character. The bug was that, for an unlimited repeat of . in
2411 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2412 characters when looking for a newline.
2413
2414 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2415
2416 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2417 in debug output.
2418
2419 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2420 long printing in the pcrecpp unittest when running under MinGW.
2421
2422 6. ESC_K was left out of the EBCDIC table.
2423
2424 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2425 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2426 limit also applies to "virtual nesting" when a pattern is recursive, and in
2427 this case 1000 isn't so big. I have been able to remove this limit at the
2428 expense of backing off one optimization in certain circumstances. Normally,
2429 when pcre_exec() would call its internal match() function recursively and
2430 immediately return the result unconditionally, it uses a "tail recursion"
2431 feature to save stack. However, when a subpattern that can match an empty
2432 string has an unlimited repetition quantifier, it no longer makes this
2433 optimization. That gives it a stack frame in which to save the data for
2434 checking that an empty string has been matched. Previously this was taken
2435 from the 1000-entry workspace that had been reserved. So now there is no
2436 explicit limit, but more stack is used.
2437
2438 8. Applied Daniel's patches to solve problems with the import/export magic
2439 syntax that is required for Windows, and which was going wrong for the
2440 pcreposix and pcrecpp parts of the library. These were overlooked when this
2441 problem was solved for the main library.
2442
2443 9. There were some crude static tests to avoid integer overflow when computing
2444 the size of patterns that contain repeated groups with explicit upper
2445 limits. As the maximum quantifier is 65535, the maximum group length was
2446 set at 30,000 so that the product of these two numbers did not overflow a
2447 32-bit integer. However, it turns out that people want to use groups that
2448 are longer than 30,000 bytes (though not repeat them that many times).
2449 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2450 made it possible to implement the integer overflow checks in a much more
2451 dynamic way, which I have now done. The artificial limitation on group
2452 length has been removed - we now have only the limit on the total length of
2453 the compiled pattern, which depends on the LINK_SIZE setting.
2454
2455 10. Fixed a bug in the documentation for get/copy named substring when
2456 duplicate names are permitted. If none of the named substrings are set, the
2457 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2458 empty string.
2459
2460 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2461 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2462 because the ] is interpreted as the first data character and the
2463 terminating ] is not found. PCRE has been made compatible with Perl in this
2464 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2465 cause memory overwriting.
2466
2467 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2468 string has been matched (to stop an infinite loop). It was not recognizing
2469 a conditional subpattern that could match an empty string if that
2470 subpattern was within another subpattern. For example, it looped when
2471 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2472 condition was not nested. This bug has been fixed.
2473
2474 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2475 past the start of the subject in the presence of bytes with the top bit
2476 set, for example "\x8aBCD".
2477
2478 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2479 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2480
2481 14. Optimized (?!) to (*FAIL).
2482
2483 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2484 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2485 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2486 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2487 does: it's just the validity check that is more restrictive.
2488
2489 16. Inserted checks for integer overflows during escape sequence (backslash)
2490 processing, and also fixed erroneous offset values for syntax errors during
2491 backslash processing.
2492
2493 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2494 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2495
2496 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2497 caused an overrun.
2498
2499 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2500 something other than just ASCII characters) inside a group that had an
2501 unlimited repeat caused a loop at compile time (while checking to see
2502 whether the group could match an empty string).
2503
2504 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2505 [\P{Any}] did so. (Error in the code for printing property names.)
2506
2507 21. An orphan \E inside a character class could cause a crash.
2508
2509 22. A repeated capturing bracket such as (A)? could cause a wild memory
2510 reference during compilation.
2511
2512 23. There are several functions in pcre_compile() that scan along a compiled
2513 expression for various reasons (e.g. to see if it's fixed length for look
2514 behind). There were bugs in these functions when a repeated \p or \P was
2515 present in the pattern. These operators have additional parameters compared
2516 with \d, etc, and these were not being taken into account when moving along
2517 the compiled data. Specifically:
2518
2519 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2520 length.
2521
2522 (b) An item such as \pL+ within a repeated group could cause crashes or
2523 loops.
2524
2525 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2526 "reference to non-existent subpattern" error.
2527
2528 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2529
2530 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2531 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2532
2533 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2534
2535 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2536 character were causing crashes (broken optimization).
2537
2538 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2539 \p or \P) caused a compile-time loop.
2540
2541 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2542 break. For example, the unstudied pattern [\r\n]A does not match the string
2543 "\r\nA" because change 7.0/46 below moves the current point on by two
2544 characters after failing to match at the start. However, the pattern \nA
2545 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2546 the same is true. There doesn't seem any very clean way out of this, but
2547 what I have chosen to do makes the common cases work: PCRE now takes note
2548 of whether there can be an explicit match for \r or \n anywhere in the
2549 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2550 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2551 pattern has explicit CR or LF references.
2552
2553 29. Added (*CR) etc for changing newline setting at start of pattern.
2554
2555
2556 Version 7.2 19-Jun-07
2557 ---------------------
2558
2559 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2560 which is apparently normally available under Windows.
2561
2562 2. Re-jig the pcregrep tests with different newline settings in an attempt
2563 to make them independent of the local environment's newline setting.
2564
2565 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2566
2567 4. Some of the "internals" tests were previously cut out when the link size
2568 was not 2, because the output contained actual offsets. The recent new
2569 "Z" feature of pcretest means that these can be cut out, making the tests
2570 usable with all link sizes.
2571
2572 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2573 stack recursion. This gives a massive performance boost under BSD, but just
2574 a small improvement under Linux. However, it saves one field in the frame
2575 in all cases.
2576
2577 6. Added more features from the forthcoming Perl 5.10:
2578
2579 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2580 recursion call. It refers to the nth most recently opened parentheses.
2581
2582 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2583 to be opened parentheses.
2584
2585 (c) Conditions that refer to capturing parentheses can be specified
2586 relatively, for example, (?(-2)... or (?(+3)...
2587
2588 (d) \K resets the start of the current match so that everything before
2589 is not part of it.
2590
2591 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2592
2593 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2594 reference syntax.
2595
2596 (g) (?| introduces a group in which the numbering of parentheses in each
2597 alternative starts with the same number.
2598
2599 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2600
2601 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2602 PCRE_INFO_JCHANGED.
2603
2604 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2605 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2606 for detecting groups that can match an empty string.
2607
2608 9. A pattern with a very large number of alternatives (more than several
2609 hundred) was running out of internal workspace during the pre-compile
2610 phase, where pcre_compile() figures out how much memory will be needed. A
2611 bit of new cunning has reduced the workspace needed for groups with
2612 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2613 workspace instead of running out of the 4096 that are available.
2614
2615 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2616
2617 11. Applied patch from Google to remove an optimization that didn't quite work.
2618 The report of the bug said:
2619
2620 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2621 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2622 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2623
2624 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2625 it matched the wrong number of bytes.
2626
2627
2628 Version 7.1 24-Apr-07
2629 ---------------------
2630
2631 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2632 that is more "standard", making use of automake and other Autotools. There
2633 is some re-arrangement of the files and adjustment of comments consequent
2634 on this.
2635
2636 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2637 for recursive directory scanning broke on some systems because the files
2638 are not scanned in any specific order and on different systems the order
2639 was different. A call to "sort" has been inserted into RunGrepTest for the
2640 approprate test as a short-term fix. In the longer term there may be an
2641 alternative.
2642
2643 3. I had an email from Eric Raymond about problems translating some of PCRE's
2644 man pages to HTML (despite the fact that I distribute HTML pages, some
2645 people do their own conversions for various reasons). The problems
2646 concerned the use of low-level troff macros .br and .in. I have therefore
2647 removed all such uses from the man pages (some were redundant, some could
2648 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2649 HTML has been updated to handle .nf/.fi and to complain if it encounters
2650 .br or .in.
2651
2652 4. Updated comments in configure.ac that get placed in config.h.in and also
2653 arranged for config.h to be included in the distribution, with the name
2654 config.h.generic, for the benefit of those who have to compile without
2655 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2656
2657 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2658 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2659 makevp.bat for the current PCRE, using the additional files
2660 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2661
2662 6. A Windows user reported a minor discrepancy with test 2, which turned out
2663 to be caused by a trailing space on an input line that had got lost in his
2664 copy. The trailing space was an accident, so I've just removed it.
2665
2666 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2667 that is needed.
2668
2669 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2670 as "const" (a) because they are and (b) because it helps the PHP
2671 maintainers who have recently made a script to detect big data structures
2672 in the php code that should be moved to the .rodata section. I remembered
2673 to update Builducptable as well, so it won't revert if ucptable.h is ever
2674 re-created.
2675
2676 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2677 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2678 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2679 support is not required. This saves 1.5-2K of code, which is important in
2680 some applications.
2681
2682 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2683 so as not to refer to the tables, even though these functions will never be
2684 called when UTF-8 support is disabled. Otherwise there are problems with a
2685 shared library.
2686
2687 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2688
2689 (a) It was defining its arguments as char * instead of void *.
2690
2691 (b) It was assuming that all moves were upwards in memory; this was true
2692 a long time ago when I wrote it, but is no longer the case.
2693
2694 The emulated memove() is provided for those environments that have neither
2695 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2696 is clearly not the case, as these two bugs were recently reported.
2697
2698 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2699 and Detrail to create the HTML documentation, the .txt form of the man
2700 pages, and it removes trailing spaces from listed files. It also creates
2701 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2702 case, it wraps all the #defines with #ifndefs. This script should be run
2703 before "make dist".
2704
2705 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2706 with Unicode property support.
2707
2708 (a) For a maximizing quantifier, if the two different cases of the
2709 character were of different lengths in their UTF-8 codings (there are
2710 some cases like this - I found 11), and the matching function had to
2711 back up over a mixture of the two cases, it incorrectly assumed they
2712 were both the same length.
2713
2714 (b) When PCRE was configured to use the heap rather than the stack for
2715 recursion during matching, it was not correctly preserving the data for
2716 the other case of a UTF-8 character when checking ahead for a match
2717 while processing a minimizing repeat. If the check also involved
2718 matching a wide character, but failed, corruption could cause an
2719 erroneous result when trying to check for a repeat of the original
2720 character.
2721
2722 13. Some tidying changes to the testing mechanism:
2723
2724 (a) The RunTest script now detects the internal link size and whether there
2725 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2726 values substituted by "configure". (The RunGrepTest script already did
2727 this for UTF-8.) The configure.ac script no longer substitutes the
2728 relevant variables.
2729
2730 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2731 with length and offset values. This means that the output is different
2732 for different internal link sizes. Test 2 is skipped for link sizes
2733 other than 2 because of this, bypassing the problem. Unfortunately,
2734 there was also a test in test 3 (the locale tests) that used /B and
2735 failed for link sizes other than 2. Rather than cut the whole test out,
2736 I have added a new /Z option to pcretest that replaces the length and
2737 offset values with spaces. This is now used to make test 3 independent
2738 of link size. (Test 2 will be tidied up later.)
2739
2740 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2741 segmentation fault instead of returning the appropriate error message.
2742
2743 15. In multiline mode when the newline sequence was set to "any", the pattern
2744 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2745 This doesn't seem right; it now treats the CRLF combination as the line
2746 ending, and so does not match in that case. It's only a pattern such as ^$
2747 that would hit this one: something like ^ABC$ would have failed after \r
2748 and then tried again after \r\n.
2749
2750 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2751 in an attempt to make files that differ only in their line terminators
2752 compare equal. This works on Linux.
2753
2754 17. Under certain error circumstances pcregrep might try to free random memory
2755 as it exited. This is now fixed, thanks to valgrind.
2756
2757 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2758 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2759 was because its rules for how to advance for /g after matching an empty
2760 string at the end of a line did not allow for this case. They now check for
2761 it specially.
2762
2763 20. pcretest is supposed to handle patterns and data of any length, by
2764 extending its buffers when necessary. It was getting this wrong when the
2765 buffer for a data line had to be extended.
2766
2767 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2768 CRLF as a newline sequence.
2769
2770 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2771 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2772 I have nevertheless tidied it up.
2773
2774 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2775
2776 24. Added a man page for pcre-config.
2777
2778
2779 Version 7.0 19-Dec-06
2780 ---------------------
2781
2782 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2783 moving to gcc 4.1.1.
2784
2785 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2786 sys/time.h, which is documented as needed for this function. It doesn't
2787 seem to matter on Linux, but it showed up on some releases of OS X.
2788
2789 3. It seems that there are systems where bytes whose values are greater than
2790 127 match isprint() in the "C" locale. The "C" locale should be the
2791 default when a C program starts up. In most systems, only ASCII printing
2792 characters match isprint(). This difference caused the output from pcretest
2793 to vary, making some of the tests fail. I have changed pcretest so that:
2794
2795 (a) When it is outputting text in the compiled version of a pattern, bytes
2796 other than 32-126 are always shown as hex escapes.
2797
2798 (b) When it is outputting text that is a matched part of a subject string,
2799 it does the same, unless a different locale has been set for the match
2800 (using the /L modifier). In this case, it uses isprint() to decide.
2801
2802 4. Fixed a major bug that caused incorrect computation of the amount of memory
2803 required for a compiled pattern when options that changed within the
2804 pattern affected the logic of the preliminary scan that determines the
2805 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2806 that the computed length was too small. The symptoms of this bug were
2807 either the PCRE error "internal error: code overflow" from pcre_compile(),
2808 or a glibc crash with a message such as "pcretest: free(): invalid next
2809 size (fast)". Examples of patterns that provoked this bug (shown in
2810 pcretest format) are:
2811
2812 /(?-x: )/x
2813 /(?x)(?-x: \s*#\s*)/
2814 /((?i)[\x{c0}])/8
2815 /(?i:[\x{c0}])/8
2816
2817 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2818 is now done differently.
2819
2820 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2821 wrapper classes; (b) implement a new function in the C++ scanner that is
2822 more efficient than the old way of doing things because it avoids levels of
2823 recursion in the regex matching; (c) add a paragraph to the documentation
2824 for the FullMatch() function.
2825
2826 6. The escape sequence \n was being treated as whatever was defined as
2827 "newline". Not only was this contrary to the documentation, which states
2828 that \n is character 10 (hex 0A), but it also went horribly wrong when
2829 "newline" was defined as CRLF. This has been fixed.
2830
2831 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2832 was being set to -1 for the "end of line" case (supposedly a value that no
2833 character can have). Though this value is never used (the check for end of
2834 line is "zero bytes in current character"), it caused compiler complaints.
2835 I've changed it to 0xffffffff.
2836
2837 8. In pcre_version.c, the version string was being built by a sequence of
2838 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2839 string (as it is for production releases) called a macro with an empty
2840 argument. The C standard says the result of this is undefined. The gcc
2841 compiler treats it as an empty string (which was what was wanted) but it is
2842 reported that Visual C gives an error. The source has been hacked around to
2843 avoid this problem.
2844
2845 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2846 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2847 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2848 of them did).
2849
2850 10. Originally, pcretest opened its input and output without "b"; then I was
2851 told that "b" was needed in some environments, so it was added for release
2852 5.0 to both the input and output. (It makes no difference on Unix-like
2853 systems.) Later I was told that it is wrong for the input on Windows. I've
2854 now abstracted the modes into two macros, to make it easier to fiddle with
2855 them, and removed "b" from the input mode under Windows.
2856
2857 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2858
2859 12. Added -help and --help to pcretest as an official way of being reminded
2860 of the options.
2861
2862 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2863 and pcrecpp.cc because they annoy compilers at high warning levels.
2864
2865 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2866
2867 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2868 scripts are not C programs :-) and which was not noticed because it works
2869 on Linux.
2870
2871 16. pcretest is supposed to handle any length of pattern and data line (as one
2872 line or as a continued sequence of lines) by extending its input buffer if
2873 necessary. This feature was broken for very long pattern lines, leading to
2874 a string of junk being passed to pcre_compile() if the pattern was longer
2875 than about 50K.
2876
2877 17. I have done a major re-factoring of the way pcre_compile() computes the
2878 amount of memory needed for a compiled pattern. Previously, there was code
2879 that made a preliminary scan of the pattern in order to do this. That was
2880 OK when PCRE was new, but as the facilities have expanded, it has become
2881 harder and harder to keep it in step with the real compile phase, and there
2882 have been a number of bugs (see for example, 4 above). I have now found a
2883 cunning way of running the real compile function in a "fake" mode that
2884 enables it to compute how much memory it would need, while actually only
2885 ever using a few hundred bytes of working memory and without too many
2886 tests of the mode. This should make future maintenance and development
2887 easier. A side effect of this work is that the limit of 200 on the nesting
2888 depth of parentheses has been removed (though this was never a serious
2889 limitation, I suspect). However, there is a downside: pcre_compile() now
2890 runs more slowly than before (30% or more, depending on the pattern). I
2891 hope this isn't a big issue. There is no effect on runtime performance.
2892
2893 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2894 newline (only possible for the last line of a file) and it was a
2895 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2896
2897 19. Added additional timing features to pcretest. (1) The -tm option now times
2898 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2899 separate command line item, by a number that specifies the number of
2900 repeats to use when timing. The default is 50000; this gives better
2901 precision, but takes uncomfortably long for very large patterns.
2902
2903 20. Extended pcre_study() to be more clever in cases where a branch of a
2904 subpattern has no definite first character. For example, (a*|b*)[cd] would
2905 previously give no result from pcre_study(). Now it recognizes that the
2906 first character must be a, b, c, or d.
2907
2908 21. There was an incorrect error "recursive call could loop indefinitely" if
2909 a subpattern (or the entire pattern) that was being tested for matching an
2910 empty string contained only one non-empty item after a nested subpattern.
2911 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2912 incorrectly, because the \d was being skipped in the check.
2913
2914 22. The pcretest program now has a new pattern option /B and a command line
2915 option -b, which is equivalent to adding /B to every pattern. This causes
2916 it to show the compiled bytecode, without the additional information that
2917 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2918 is the same as /B/I).
2919
2920 23. A new optimization is now able automatically to treat some sequences such
2921 as a*b as a*+b. More specifically, if something simple (such as a character
2922 or a simple class like \d) has an unlimited quantifier, and is followed by
2923 something that cannot possibly match the quantified thing, the quantifier
2924 is automatically "possessified".
2925
2926 24. A recursive reference to a subpattern whose number was greater than 39
2927 went wrong under certain circumstances in UTF-8 mode. This bug could also
2928 have affected the operation of pcre_study().
2929
2930 25. Realized that a little bit of performance could be had by replacing
2931 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2932
2933 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2934
2935 27. Possessive quantifiers such as a++ were previously implemented by turning
2936 them into atomic groups such as ($>a+). Now they have their own opcodes,
2937 which improves performance. This includes the automatically created ones
2938 from 23 above.
2939
2940 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2941 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2942 the first matched character to be a colon. This applied both to named and
2943 numbered groups.
2944
2945 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2946
2947 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2948 building PCRE on Windows easier, so I have included it in the distribution.
2949
2950 31. There is now a check in pcretest against a ridiculously large number being
2951 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2952 loop, the loop is abandoned.
2953
2954 32. Forward references to subpatterns in conditions such as (?(2)...) where
2955 subpattern 2 is defined later cause pcre_compile() to search forwards in
2956 the pattern for the relevant set of parentheses. This search went wrong
2957 when there were unescaped parentheses in a character class, parentheses
2958 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2959
2960 33. "Subroutine" calls and backreferences were previously restricted to
2961 referencing subpatterns earlier in the regex. This restriction has now
2962 been removed.
2963
2964 34. Added a number of extra features that are going to be in Perl 5.10. On the
2965 whole, these are just syntactic alternatives for features that PCRE had
2966 previously implemented using the Python syntax or my own invention. The
2967 other formats are all retained for compatibility.
2968
2969 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2970 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2971 also .NET compatible.
2972
2973 (b) A recursion or subroutine call to a named group can now be defined as
2974 (?&name) as well as (?P>name).
2975
2976 (c) A backreference to a named group can now be defined as \k<name> or
2977 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2978 5.10, are also .NET compatible.
2979
2980 (d) A conditional reference to a named group can now use the syntax
2981 (?(<name>) or (?('name') as well as (?(name).
2982
2983 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2984 groups (named and numbered) that are never evaluated inline, but can be
2985 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2986 is always false. There may be only one alternative in such a group.
2987
2988 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2989 as the simple (?(R). The condition is true only if the most recent
2990 recursion is that of the given number or name. It does not search out
2991 through the entire recursion stack.
2992
2993 (g) The escape \gN or \g{N} has been added, where N is a positive or
2994 negative number, specifying an absolute or relative reference.
2995
2996 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2997 some "unreachable code" warnings.
2998
2999 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3000 things, this adds five new scripts.
3001
3002 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3003 There were also incompatibilities regarding the handling of \Q..\E inside
3004 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3005 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3006
3007 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3008 matches an empty string, and forcibly breaks the loop. There were bugs in
3009 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3010 against aaaa the result was just "a" rather than "aaaa", for example. Two
3011 separate and independent bugs (that affected different cases) have been
3012 fixed.
3013
3014 39. Refactored the code to abolish the use of different opcodes for small
3015 capturing bracket numbers. This is a tidy that I avoided doing when I
3016 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3017 The new approach is not only tidier, it makes it possible to reduce the
3018 memory needed to fix the previous bug (38).
3019
3020 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3021 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3022 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3023 mode.
3024
3025 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3026 report.
3027
3028 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3029 copy construction and assignment in the C++ wrapper.
3030
3031 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3032 couple of bugs that could have given wrong results in the "--newline=crlf"
3033 case.
3034
3035 44. Added a number of casts and did some reorganization of signed/unsigned int
3036 variables following suggestions from Dair Grant. Also renamed the variable
3037 "this" as "item" because it is a C++ keyword.
3038
3039 45. Arranged for dftables to add
3040
3041 #include "pcre_internal.h"
3042
3043 to pcre_chartables.c because without it, gcc 4.x may remove the array
3044 definition from the final binary if PCRE is built into a static library and
3045 dead code stripping is activated.
3046
3047 46. For an unanchored pattern, if a match attempt fails at the start of a
3048 newline sequence, and the newline setting is CRLF or ANY, and the next two
3049 characters are CRLF, advance by two characters instead of one.
3050
3051
3052 Version 6.7 04-Jul-06
3053 ---------------------
3054
3055 1. In order to handle tests when input lines are enormously long, pcretest has
3056 been re-factored so that it automatically extends its buffers when
3057 necessary. The code is crude, but this _is_ just a test program. The
3058 default size has been increased from 32K to 50K.
3059
3060 2. The code in pcre_study() was using the value of the re argument before
3061 testing it for NULL. (Of course, in any sensible call of the function, it
3062 won't be NULL.)
3063
3064 3. The memmove() emulation function in pcre_internal.h, which is used on
3065 systems that lack both memmove() and bcopy() - that is, hardly ever -
3066 was missing a "static" storage class specifier.
3067
3068 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3069 containing an extended class (one that cannot be represented by a bitmap
3070 because it contains high-valued characters or Unicode property items, e.g.
3071 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3072 pattern, but PCRE should not loop if you do not (it no longer does).
3073 [Detail: two cases were found: (a) a repeated subpattern containing an
3074 extended class; (b) a recursive reference to a subpattern that followed a
3075 previous extended class. It wasn't skipping over the extended class
3076 correctly when UTF-8 mode was not set.]
3077
3078 5. A negated single-character class was not being recognized as fixed-length
3079 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3080 compile error "lookbehind assertion is not fixed length".
3081
3082 6. The RunPerlTest auxiliary script was showing an unexpected difference
3083 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3084 write a Perl script that can interpret lines of an input file either as
3085 byte characters or as UTF-8, which is what "perltest" was being required to
3086 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3087 can't do is switch easily at run time between having the "use utf8;" pragma
3088 or not. In the end, I fudged it by using the RunPerlTest script to insert
3089 "use utf8;" explicitly for the UTF-8 tests.
3090
3091 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3092 the end of the subject string, contrary to the documentation and to what
3093 Perl does. This was true of both matching functions. Now it matches only at
3094 the start of the subject and immediately after *internal* newlines.
3095
3096 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3097 a pointer to an int instead of a pointer to an unsigned long int. This
3098 caused problems on 64-bit systems.
3099
3100 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3101 instance of the 'standard' template library not being so standard".
3102
3103 10. There was no check on the number of named subpatterns nor the maximum
3104 length of a subpattern name. The product of these values is used to compute
3105 the size of the memory block for a compiled pattern. By supplying a very
3106 long subpattern name and a large number of named subpatterns, the size
3107 computation could be caused to overflow. This is now prevented by limiting
3108 the length of names to 32 characters, and the number of named subpatterns
3109 to 10,000.
3110
3111 11. Subpatterns that are repeated with specific counts have to be replicated in
3112 the compiled pattern. The size of memory for this was computed from the
3113 length of the subpattern and the repeat count. The latter is limited to
3114 65535, but there was no limit on the former, meaning that integer overflow
3115 could in principle occur. The compiled length of a repeated subpattern is
3116 now limited to 30,000 bytes in order to prevent this.
3117
3118 12. Added the optional facility to have named substrings with the same name.
3119
3120 13. Added the ability to use a named substring as a condition, using the
3121 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3122 are numbers (not recommended). Forward references are permitted.
3123
3124 14. Added forward references in named backreferences (if you see what I mean).
3125
3126 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3127 pattern could run off the end of the subject. For example, the pattern
3128 "(?s)(.{1,5})"8 did this with the subject "ab".
3129
3130 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3131 PCRE_CASELESS was set when matching characters that were quantified with ?
3132 or *.
3133
3134 17. A character class other than a single negated character that had a minimum
3135 but no maximum quantifier - for example [ab]{6,} - was not handled
3136 correctly by pce_dfa_exec(). It would match only one character.
3137
3138 18. A valid (though odd) pattern that looked like a POSIX character
3139 class but used an invalid character after [ (for example [[,abc,]]) caused
3140 pcre_compile() to give the error "Failed: internal error: code overflow" or
3141 in some cases to crash with a glibc free() error. This could even happen if
3142 the pattern terminated after [[ but there just happened to be a sequence of
3143 letters, a binary zero, and a closing ] in the memory that followed.
3144
3145 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3146 over the years. Originally (before any Unicode support), just the bottom 8
3147 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3148 output from "man perlunicode" includes this:
3149
3150 The regular expression compiler produces polymorphic opcodes. That
3151 is, the pattern adapts to the data and automatically switches to
3152 the Unicode character scheme when presented with Unicode data--or
3153 instead uses a traditional byte scheme when presented with byte
3154 data.
3155
3156 Sadly, a wide octal escape does not cause a switch, and in a string with
3157 no other multibyte characters, these octal escapes are treated as before.
3158 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3159 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3160 Unicode string.
3161
3162 I have not perpetrated such confusion in PCRE. Up till now, it took just
3163 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3164 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3165 translate to the appropriate multibyte character.
3166
3167 29. Applied some refactoring to reduce the number of warnings from Microsoft
3168 and Borland compilers. This has included removing the fudge introduced
3169 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3170 a warning about an unused variable.
3171
3172 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3173 characters since release 4.0, because Perl (from release 5.004) does not.
3174 [Or at least, is documented not to: some releases seem to be in conflict
3175 with the documentation.] However, when a pattern was studied with
3176 pcre_study() and all its branches started with \s, PCRE still included VT
3177 as a possible starting character. Of course, this did no harm; it just
3178 caused an unnecessary match attempt.
3179
3180 22. Removed a now-redundant internal flag bit that recorded the fact that case
3181 dependency changed within the pattern. This was once needed for "required
3182 byte" processing, but is no longer used. This recovers a now-scarce options
3183 bit. Also moved the least significant internal flag bit to the most-
3184 significant bit of the word, which was not previously used (hangover from
3185 the days when it was an int rather than a uint) to free up another bit for
3186 the future.
3187
3188 23. Added support for CRLF line endings as well as CR and LF. As well as the
3189 default being selectable at build time, it can now be changed at runtime
3190 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3191 specify that it is scanning data with non-default line endings.
3192
3193 24. Changed the definition of CXXLINK to make it agree with the definition of
3194 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3195
3196 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3197 recursions. This makes a big different to stack usage for some patterns.
3198
3199 26. If a subpattern containing a named recursion or subroutine reference such
3200 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3201 the space required for the compiled pattern went wrong and gave too small a
3202 value. Depending on the environment, this could lead to "Failed: internal
3203 error: code overflow at offset 49" or "glibc detected double free or
3204 corruption" errors.
3205
3206 27. Applied patches from Google (a) to support the new newline modes and (b) to
3207 advance over multibyte UTF-8 characters in GlobalReplace.
3208
3209 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3210 difference for some implementation of PCRE in some Windows version.
3211
3212 29. Added some extra testing facilities to pcretest:
3213
3214 \q<number> in a data line sets the "match limit" value
3215 \Q<number> in a data line sets the "match recursion limt" value
3216 -S <number> sets the stack size, where <number> is in megabytes
3217
3218 The -S option isn't available for Windows.
3219
3220
3221 Version 6.6 06-Feb-06
3222 ---------------------
3223
3224 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3225 in pcreposix.h. I have copied the definition from pcre.h.
3226
3227 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3228 because pcre.h is no longer a built file.
3229
3230 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3231 not normally included in the compiled code.
3232
3233
3234 Version 6.5 01-Feb-06
3235 ---------------------
3236
3237 1. When using the partial match feature with pcre_dfa_exec(), it was not
3238 anchoring the second and subsequent partial matches at the new starting
3239 point. This could lead to incorrect results. For example, with the pattern
3240 /1234/, partially matching against "123" and then "a4" gave a match.
3241
3242 2. Changes to pcregrep:
3243
3244 (a) All non-match returns from pcre_exec() were being treated as failures
3245 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3246 error message is output. Some extra information is given for the
3247 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3248 probably the only errors that are likely to be caused by users (by
3249 specifying a regex that has nested indefinite repeats, for instance).
3250 If there are more than 20 of these errors, pcregrep is abandoned.
3251
3252 (b) A binary zero was treated as data while matching, but terminated the
3253 output line if it was written out. This has been fixed: binary zeroes
3254 are now no different to any other data bytes.
3255
3256 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3257 used to set a locale for matching. The --locale=xxxx long option has
3258 been added (no short equivalent) to specify a locale explicitly on the
3259 pcregrep command, overriding the environment variables.
3260
3261 (d) When -B was used with -n, some line numbers in the output were one less
3262 than they should have been.
3263
3264 (e) Added the -o (--only-matching) option.
3265
3266 (f) If -A or -C was used with -c (count only), some lines of context were
3267 accidentally printed for the final match.
3268
3269 (g) Added the -H (--with-filename) option.
3270
3271 (h) The combination of options -rh failed to suppress file names for files
3272 that were found from directory arguments.
3273
3274 (i) Added the -D (--devices) and -d (--directories) options.
3275
3276 (j) Added the -F (--fixed-strings) option.
3277
3278 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3279
3280 (l) Added the --colo(u)r option.
3281
3282 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3283 is not present by default.
3284
3285 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3286 items such as (?R) or (?1), when the recursion could match a number of
3287 alternatives. If it matched one of the alternatives, but subsequently,
3288 outside the recursion, there was a failure, the code tried to back up into
3289 the recursion. However, because of the way PCRE is implemented, this is not
3290 possible, and the result was an incorrect result from the match.
3291
3292 In order to prevent this happening, the specification of recursion has
3293 been changed so that all such subpatterns are automatically treated as
3294 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3295
3296 4. I had overlooked the fact that, in some locales, there are characters for
3297 which isalpha() is true but neither isupper() nor islower() are true. In
3298 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3299 and ordfeminine) are like this. This affected the treatment of \w and \W
3300 when they appeared in character classes, but not when they appeared outside
3301 a character class. The bit map for "word" characters is now created
3302 separately from the results of isalnum() instead of just taking it from the
3303 upper, lower, and digit maps. (Plus the underscore character, of course.)
3304
3305 5. The above bug also affected the handling of POSIX character classes such as
3306 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3307 permanent tables. Instead, the bit maps for such a class were previously
3308 created as the appropriate unions of the upper, lower, and digit bitmaps.
3309 Now they are created by subtraction from the [[:word:]] class, which has
3310 its own bitmap.
3311
3312 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3313 It is created by subtracting the vertical space characters (\x09, \x0a,
3314 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3315 subtraction was done in the overall bitmap for a character class, meaning
3316 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3317 be recognized. This bug has been fixed.
3318
3319 7. Patches from the folks at Google:
3320
3321 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3322 real life, but is still worth protecting against".
3323
3324 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3325 regular expressions".
3326
3327 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3328 have it.
3329
3330 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3331 "configure" and the latter not, in order to fix a problem somebody had
3332 with compiling the Arg class on HP-UX.
3333
3334 (e) Improve the error-handling of the C++ wrapper a little bit.
3335
3336 (f) New tests for checking recursion limiting.
3337
3338 8. The pcre_memmove() function, which is used only if the environment does not
3339 have a standard memmove() function (and is therefore rarely compiled),
3340 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3341 returning a result (though PCRE never actually uses the result).
3342
3343 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3344 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3345 returned instead of calling malloc() with an overflowing number that would
3346 most likely cause subsequent chaos.
3347
3348 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3349
3350 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3351 with this option is matched, the nmatch and pmatch options of regexec() are
3352 ignored.
3353
3354 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3355 provided in case anyone wants to the the POSIX interface with UTF-8
3356 strings.
3357
3358 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3359 C++ linking (needed for some HP-UX environments).
3360
3361 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3362 (unused parameter) and in the pcre_printint() function (omitted "default"
3363 switch label when the default is to do nothing).
3364
3365 15. Added some code to make it possible, when PCRE is compiled as a C++
3366 library, to replace subject pointers for pcre_exec() with a smart pointer
3367 class, thus making it possible to process discontinuous strings.
3368
3369 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3370 much the same function. They were added by different people who were trying
3371 to make PCRE easy to compile on non-Unix systems. It has been suggested
3372 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3373 for compiling on Windows systems. I have therefore replaced it with
3374 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3375 defaults to "extern" for C or "extern C" for C++, which works fine on
3376 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3377 SCOPE with something explicit in config.h. In addition:
3378
3379 (a) pcreposix.h still had just "extern" instead of either of these macros;
3380 I have replaced it with PCRE_DATA_SCOPE.
3381
3382 (b) Functions such as _pcre_xclass(), which are internal to the library,
3383 but external in the C sense, all had PCRE_EXPORT in their definitions.
3384 This is apparently wrong for the Windows case, so I have removed it.
3385 (It makes no difference on Unix-like systems.)
3386
3387 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3388 of recursive calls to match(). This is different to MATCH_LIMIT because
3389 that limits the total number of calls to match(), not all of which increase
3390 the depth of recursion. Limiting the recursion depth limits the amount of
3391 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3392 when PCRE is compiled, and changed at run time. A patch from Google adds
3393 this functionality to the C++ interface.
3394
3395 18. Changes to the handling of Unicode character properties:
3396
3397 (a) Updated the table to Unicode 4.1.0.
3398
3399 (b) Recognize characters that are not in the table as "Cn" (undefined).
3400
3401 (c) I revised the way the table is implemented to a much improved format
3402 which includes recognition of ranges. It now supports the ranges that
3403 are defined in UnicodeData.txt, and it also amalgamates other
3404 characters into ranges. This has reduced the number of entries in the
3405 table from around 16,000 to around 3,000, thus reducing its size
3406 considerably. I realized I did not need to use a tree structure after
3407 all - a binary chop search is just as efficient. Having reduced the
3408 number of entries, I extended their size from 6 bytes to 8 bytes to
3409 allow for more data.
3410
3411 (d) Added support for Unicode script names via properties such as \p{Han}.
3412
3413 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3414 matching that character.
3415
3416 20. When matching a repeated Unicode property with a minimum greater than zero,
3417 (for example \pL{2,}), PCRE could look past the end of the subject if it
3418 reached it while seeking the minimum number of characters. This could
3419 happen only if some of the characters were more than one byte long, because
3420 there is a check for at least the minimum number of bytes.
3421
3422 21. Refactored the implementation of \p and \P so as to be more general, to
3423 allow for more different types of property in future. This has changed the
3424 compiled form incompatibly. Anybody with saved compiled patterns that use
3425 \p or \P will have to recompile them.
3426
3427 22. Added "Any" and "L&" to the supported property types.
3428
3429 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3430 but give a compile time error if the value is greater than 0xff.
3431
3432 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3433 accidentally not being installed or uninstalled.
3434
3435 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3436 made were to insert the current release number. This seemed silly, because
3437 it made things harder for people building PCRE on systems that don't run
3438 "configure". I have turned pcre.h into a distributed file, no longer built
3439 by "configure", with the version identification directly included. There is
3440 no longer a pcre.h.in file.
3441
3442 However, this change necessitated a change to the pcre-config script as
3443 well. It is built from pcre-config.in, and one of the substitutions was the
3444 release number. I have updated configure.ac so that ./configure now finds
3445 the release number by grepping pcre.h.
3446
3447 26. Added the ability to run the tests under valgrind.
3448
3449
3450 Version 6.4 05-Sep-05
3451 ---------------------
3452
3453 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3454 "--" to be printed when multiple files were scanned, even when none of the
3455 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3456 consider it to be a bug, and have restored the previous behaviour.
3457
3458 2. A couple of code tidies to get rid of compiler warnings.
3459
3460 3. The pcretest program used to cheat by referring to symbols in the library
3461 whose names begin with _pcre_. These are internal symbols that are not
3462 really supposed to be visible externally, and in some environments it is
3463 possible to suppress them. The cheating is now confined to including
3464 certain files from the library's source, which is a bit cleaner.
3465
3466 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3467 file's purpose clearer.
3468
3469 5. Reorganized pcre_ucp_findchar().
3470
3471
3472 Version 6.3 15-Aug-05
3473 ---------------------
3474
3475 1. The file libpcre.pc.in did not have general read permission in the tarball.
3476
3477 2. There were some problems when building without C++ support:
3478
3479 (a) If C++ support was not built, "make install" and "make test" still
3480 tried to test it.
3481
3482 (b) There were problems when the value of CXX was explicitly set. Some
3483 changes have been made to try to fix these, and ...
3484
3485 (c) --disable-cpp can now be used to explicitly disable C++ support.
3486
3487 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3488 backslash in a target when C++ was disabled. This confuses some
3489 versions of "make", apparently. Using an intermediate variable solves
3490 this. (Same for CPP_LOBJ.)
3491
3492 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3493 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3494 necessary on certain architectures.
3495
3496 4. Added a setting of -export-symbols-regex to the link command to remove
3497 those symbols that are exported in the C sense, but actually are local
3498 within the library, and not documented. Their names all begin with
3499 "_pcre_". This is not a perfect job, because (a) we have to except some
3500 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3501 available (and never for static libraries). I have made a note to try to
3502 find a way round (a) in the future.
3503
3504
3505 Version 6.2 01-Aug-05
3506 ---------------------
3507
3508 1. There was no test for integer overflow of quantifier values. A construction
3509 such as {1111111111111111} would give undefined results. What is worse, if
3510 a minimum quantifier for a parenthesized subpattern overflowed and became
3511 negative, the calculation of the memory size went wrong. This could have
3512 led to memory overwriting.
3513
3514 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3515
3516 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3517 operating environments where this matters.
3518
3519 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3520 PCRE options from within the C++ wrapper.
3521
3522 5. Named capturing subpatterns were not being correctly counted when a pattern
3523 was compiled. This caused two problems: (a) If there were more than 100
3524 such subpatterns, the calculation of the memory needed for the whole
3525 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3526 back references of the form \12, where the number was greater than 9, were
3527 not recognized as back references, even though there were sufficient
3528 previous subpatterns.
3529
3530 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3531 versions of gcc, e.g. 2.95.4.
3532
3533
3534 Version 6.1 21-Jun-05
3535 ---------------------
3536
3537 1. There was one reference to the variable "posix" in pcretest.c that was not
3538 surrounded by "#if !defined NOPOSIX".
3539
3540 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3541 the cross-check on the old pcre_info() function, for the benefit of the
3542 cut-down version of PCRE that is currently imported into Exim.
3543
3544 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3545 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3546 patterns that start (?i) but I don't think that matters. The use of (?i) is
3547 just an example; this all applies to the other options as well.
3548
3549 4. Since libtool seems to echo the compile commands it is issuing, the output
3550 from "make" can be reduced a bit by putting "@" in front of each libtool
3551 compile command.
3552
3553 5. Patch from the folks at Google for configure.in to be a bit more thorough
3554 in checking for a suitable C++ installation before trying to compile the
3555 C++ stuff. This should fix a reported problem when a compiler was present,
3556 but no suitable headers.
3557
3558 6. The man pages all had just "PCRE" as their title. I have changed them to
3559 be the relevant file name. I have also arranged that these names are
3560 retained in the file doc/pcre.txt, which is a concatenation in text format
3561 of all the man pages except the little individual ones for each function.
3562
3563 7. The NON-UNIX-USE file had not been updated for the different set of source
3564 files that come with release 6. I also added a few comments about the C++
3565 wrapper.
3566
3567
3568 Version 6.0 07-Jun-05
3569 ---------------------
3570
3571 1. Some minor internal re-organization to help with my DFA experiments.
3572
3573 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3574 didn't matter for the library itself when fully configured, but did matter
3575 when compiling without UCP support, or within Exim, where the ucp files are
3576 not imported.
3577
3578 3. Refactoring of the library code to split up the various functions into
3579 different source modules. The addition of the new DFA matching code (see
3580 below) to a single monolithic source would have made it really too
3581 unwieldy, quite apart from causing all the code to be include in a
3582 statically linked application, when only some functions are used. This is
3583 relevant even without the DFA addition now that patterns can be compiled in
3584 one application and matched in another.
3585
3586 The downside of splitting up is that there have to be some external
3587 functions and data tables that are used internally in different modules of
3588 the library but which are not part of the API. These have all had their
3589 names changed to start with "_pcre_" so that they are unlikely to clash
3590 with other external names.
3591
3592 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3593 a different (DFA) algorithm. Although it is slower than the original
3594 function, it does have some advantages for certain types of matching
3595 problem.
3596
3597 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3598 including restarting after a partial match.
3599
3600 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3601 defined when compiling for Windows was sent to me. I have put it into the
3602 code, though I have no means of testing or verifying it.
3603
3604 7. Added the pcre_refcount() auxiliary function.
3605
3606 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3607 match before or at the first newline in the subject string. In pcretest,
3608 the /f option on a pattern can be used to set this.
3609
3610 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3611 would behave wrongly. This has been present in PCRE since release 4.0.
3612
3613 10. A number of changes to the pcregrep command:
3614
3615 (a) Refactored how -x works; insert ^(...)$ instead of setting
3616 PCRE_ANCHORED and checking the length, in preparation for adding
3617 something similar for -w.
3618
3619 (b) Added the -w (match as a word) option.
3620
3621 (c) Refactored the way lines are read and buffered so as to have more
3622 than one at a time available.
3623
3624 (d) Implemented a pcregrep test script.
3625
3626 (e) Added the -M (multiline match) option. This allows patterns to match
3627 over several lines of the subject. The buffering ensures that at least
3628 8K, or the rest of the document (whichever is the shorter) is available
3629 for matching (and similarly the previous 8K for lookbehind assertions).
3630
3631 (f) Changed the --help output so that it now says
3632
3633 -w, --word-regex(p)
3634
3635 instead of two lines, one with "regex" and the other with "regexp"
3636 because that confused at least one person since the short forms are the
3637 same. (This required a bit of code, as the output is generated
3638 automatically from a table. It wasn't just a text change.)
3639
3640 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3641 option but starts with a hyphen. Could be a pattern or a path name
3642 starting with a hyphen, for instance.
3643
3644 (h) "-" can be given as a file name to represent stdin.
3645
3646 (i) When file names are being printed, "(standard input)" is used for
3647 the standard input, for compatibility with GNU grep. Previously
3648 "<stdin>" was used.
3649
3650 (j) The option --label=xxx can be used to supply a name to be used for
3651 stdin when file names are being printed. There is no short form.
3652
3653 (k) Re-factored the options decoding logic because we are going to add
3654 two more options that take data. Such options can now be given in four
3655 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3656
3657 (l) Added the -A, -B, and -C options for requesting that lines of context
3658 around matches be printed.
3659
3660 (m) Added the -L option to print the names of files that do not contain
3661 any matching lines, that is, the complement of -l.
3662
3663 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3664 continue to scan other files.
3665
3666 (o) The -s option was incorrectly implemented. For compatibility with other
3667 greps, it now suppresses the error message for a non-existent or non-
3668 accessible file (but not the return code). There is a new option called
3669 -q that suppresses the output of matching lines, which was what -s was
3670 previously doing.
3671
3672 (p) Added --include and --exclude options to specify files for inclusion
3673 and exclusion when recursing.
3674
3675 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3676 Hopefully, it now does.
3677
3678 12. Missing cast in pcre_study().
3679
3680 13. Added an "uninstall" target to the makefile.
3681
3682 14. Replaced "extern" in the function prototypes in Makefile.in with
3683 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3684 world, but is set differently for Windows.
3685
3686 15. Added a second compiling function called pcre_compile2(). The only
3687 difference is that it has an extra argument, which is a pointer to an
3688 integer error code. When there is a compile-time failure, this is set
3689 non-zero, in addition to the error test pointer being set to point to an
3690 error message. The new argument may be NULL if no error number is required
3691 (but then you may as well call pcre_compile(), which is now just a
3692 wrapper). This facility is provided because some applications need a
3693 numeric error indication, but it has also enabled me to tidy up the way
3694 compile-time errors are handled in the POSIX wrapper.
3695
3696 16. Added VPATH=.libs to the makefile; this should help when building with one
3697 prefix path and installing with another. (Or so I'm told by someone who
3698 knows more about this stuff than I do.)
3699
3700 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3701 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3702 match everything, including newlines. This is not POSIX-compatible, but
3703 somebody wanted the feature. From pcretest it can be activated by using
3704 both the P and the s flags.
3705
3706 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3707
3708 19. libpcre.pc was being incorrectly installed as executable.
3709
3710 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3711 it now also looks for '\r' so that it will work unmodified on Windows.
3712
3713 21. Added Google's contributed C++ wrapper to the distribution.
3714
3715 22. Added some untidy missing memory free() calls in pcretest, to keep
3716 Electric Fence happy when testing.
3717
3718
3719
3720 Version 5.0 13-Sep-04
3721 ---------------------
3722
3723 1. Internal change: literal characters are no longer packed up into items
3724 containing multiple characters in a single byte-string. Each character
3725 is now matched using a separate opcode. However, there may be more than one
3726 byte in the character in UTF-8 mode.
3727
3728 2. The pcre_callout_block structure has two new fields: pattern_position and
3729 next_item_length. These contain the offset in the pattern to the next match
3730 item, and its length, respectively.
3731
3732 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3733 insertion of callouts before each pattern item. Added the /C option to
3734 pcretest to make use of this.
3735
3736 4. On the advice of a Windows user, the lines
3737
3738 #if defined(_WIN32) || defined(WIN32)
3739 _setmode( _fileno( stdout ), 0x8000 );
3740 #endif /* defined(_WIN32) || defined(WIN32) */
3741
3742 have been added to the source of pcretest. This apparently does useful
3743 magic in relation to line terminators.
3744
3745 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3746 for the benefit of those environments where the "b" makes a difference.
3747
3748 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3749 to know about it. I have put a hack into configure.in that adds in code
3750 to set GCC=yes if CC=icc. This seems to end up at a point in the
3751 generated configure script that is early enough to affect the setting of
3752 compiler options, which is what is needed, but I have no means of testing
3753 whether it really works. (The user who reported this had patched the
3754 generated configure script, which of course I cannot do.)
3755
3756 LATER: After change 22 below (new libtool files), the configure script
3757 seems to know about icc (and also ecc). Therefore, I have commented out
3758 this hack in configure.in.
3759
3760 7. Added support for pkg-config (2 patches were sent in).
3761
3762 8. Negated POSIX character classes that used a combination of internal tables
3763 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3764 [[:^ascii]]. Typically, they would match almost any characters. The other
3765 POSIX classes were not broken in this way.
3766
3767 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3768 to find the match, as PCRE was deluded into thinking that the match had to
3769 start at the start point or following a newline. The same bug applied to
3770 patterns with negative forward assertions or any backward assertions
3771 preceding ".*" at the start, unless the pattern required a fixed first
3772 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3773
3774 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3775 starting at the last subject character, bytes beyond the end of the subject
3776 string were read.
3777
3778 11. Renamed the variable "class" as "classbits" to make life easier for C++
3779 users. (Previously there was a macro definition, but it apparently wasn't
3780 enough.)
3781
3782 12. Added the new field "tables" to the extra data so that tables can be passed
3783 in at exec time, or the internal tables can be re-selected. This allows
3784 a compiled regex to be saved and re-used at a later time by a different
3785 program that might have everything at different addresses.
3786
3787 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3788 -R library as well as a -L library.
3789
3790 14. The debugging options of pcretest (-d on the command line or D on a
3791 pattern) showed incorrect output for anything following an extended class
3792 that contained multibyte characters and which was followed by a quantifier.
3793
3794 15. Added optional support for general category Unicode character properties
3795 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3796 support. It adds about 90K to the size of the library. The meanings of the
3797 inbuilt class escapes such as \d and \s have NOT been changed.
3798
3799 16. Updated pcredemo.c to include calls to free() to release the memory for the
3800 compiled pattern.
3801
3802 17. The generated file chartables.c was being created in the source directory
3803 instead of in the building directory. This caused the build to fail if the
3804 source directory was different from the building directory, and was
3805 read-only.
3806
3807 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3808 file. No doubt somebody will tell me if they don't make sense... Also added
3809 Dan Mooney's comments about building on OpenVMS.
3810
3811 19. Added support for partial matching via the PCRE_PARTIAL option for
3812 pcre_exec() and the \P data escape in pcretest.
3813
3814 20. Extended pcretest with 3 new pattern features:
3815
3816 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3817 write the compiled pattern to the file whose name is "rest-of-line".
3818 This is a straight binary dump of the data, with the saved pointer to
3819 the character tables forced to be NULL. The study data, if any, is
3820 written too. After writing, pcretest reads a new pattern.
3821
3822 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3823 compiled pattern from the given file. There must not be any
3824 occurrences of "<" in the file name (pretty unlikely); if there are,
3825 pcretest will instead treat the initial "<" as a pattern delimiter.
3826 After reading in the pattern, pcretest goes on to read data lines as
3827 usual.
3828
3829 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3830 and 16-bit fields in a compiled pattern, to simulate a pattern that
3831 was compiled on a host of opposite endianness.
3832
3833 21. The pcre-exec() function can now cope with patterns that were compiled on
3834 hosts of opposite endianness, with this restriction:
3835
3836 As for any compiled expression that is saved and used later, the tables
3837 pointer field cannot be preserved; the extra_data field in the arguments
3838 to pcre_exec() should be used to pass in a tables address if a value
3839 other than the default internal tables were used at compile time.
3840
3841 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3842 now diagnosed as an error. Previously, most of the time, a negative number
3843 would have been treated as zero, but if in addition "ovector" was passed as
3844 NULL, a crash could occur.
3845
3846 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3847 new versions from the libtool 1.5 distribution (the last one is a copy of
3848 a file called libtool.m4). This seems to have fixed the need to patch
3849 "configure" to support Darwin 1.3 (which I used to do). However, I still
3850 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3851 workstation).
3852
3853 24. Changed the PCRE licence to be the more standard "BSD" licence.
3854
3855
3856 Version 4.5 01-Dec-03
3857 ---------------------
3858
3859 1. There has been some re-arrangement of the code for the match() function so
3860 that it can be compiled in a version that does not call itself recursively.
3861 Instead, it keeps those local variables that need separate instances for
3862 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3863 needs to "recurse". Keeping track of where control must go is done by means
3864 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3865 hide most of the details from the main code, and operates only if
3866 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3867 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3868 operating.
3869
3870 To make it easier for callers to provide specially tailored get/free
3871 functions for this usage, two new functions, pcre_stack_malloc, and
3872 pcre_stack_free, are used. They are always called in strict stacking order,
3873 and the size of block requested is always the same.
3874
3875 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3876 PCRE has been compiled to use the stack or the heap for recursion. The
3877 -C option of pcretest uses this to show which version is compiled.
3878
3879 A new data escape \S, is added to pcretest; it causes the amounts of store
3880 obtained and freed by both kinds of malloc/free at match time to be added
3881 to the output.
3882
3883 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3884 what's available on my current Linux desktop machine.
3885
3886 3. When matching a UTF-8 string, the test for a valid string at the start has
3887 been extended. If start_offset is not zero, PCRE now checks that it points
3888 to a byte that is the start of a UTF-8 character. If not, it returns
3889 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3890 this is necessary because there may be backward assertions in the pattern.
3891 When matching the same subject several times, it may save resources to use
3892 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3893
3894 4. The code for checking the validity of UTF-8 strings has been tightened so
3895 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3896 containing "overlong sequences".
3897
3898 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3899 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3900 should have been "&", but it just so happened that all the cases this let
3901 through by mistake were picked up later in the function.
3902
3903 6. I had used a variable called "isblank" - this is a C99 function, causing
3904 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3905
3906 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3907 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3908 so the version is shown; (c) stop "make test" echoing "./RunTest".
3909
3910 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3911
3912 9. The prototype for memmove() for systems that don't have it was using
3913 size_t, but the inclusion of the header that defines size_t was later. I've
3914 moved the #includes for the C headers earlier to avoid this.
3915
3916 10. Added some adjustments to the code to make it easier to compiler on certain
3917 special systems:
3918
3919 (a) Some "const" qualifiers were missing.
3920 (b) Added the macro EXPORT before all exported functions; by default this
3921 is defined to be empty.
3922 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3923 that it reads its output file name as an argument instead of writing
3924 to the standard output and assuming this can be redirected.
3925
3926 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3927 class containing characters with values greater than 255, PCRE compilation
3928 went into a loop.
3929
3930 12. A recursive reference to a subpattern that was within another subpattern
3931 that had a minimum quantifier of zero caused PCRE to crash. For example,
3932 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3933 recursion. If the recursively-called subpattern itself had a zero repeat,
3934 that was OK.
3935
3936 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3937 buffer into which it was copied (for escape processing) was still set at
3938 1024, so long lines caused crashes.
3939
3940 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3941 "internal error: code overflow...". This applied to any character class
3942 that was followed by a possessive quantifier.
3943
3944 15. Modified the Makefile to add libpcre.la as a prerequisite for
3945 libpcreposix.la because I was told this is needed for a parallel build to
3946 work.
3947
3948 16. If a pattern that contained .* following optional items at the start was
3949 studied, the wrong optimizing data was generated, leading to matching
3950 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3951 matching string must start with a or b or c. The correct conclusion for
3952 this pattern is that a match can start with any character.
3953
3954
3955 Version 4.4 13-Aug-03
3956 ---------------------
3957
3958 1. In UTF-8 mode, a character class containing characters with values between
3959 127 and 255 was not handled correctly if the compiled pattern was studied.
3960 In fixing this, I have also improved the studying algorithm for such
3961 classes (slightly).
3962
3963 2. Three internal functions had redundant arguments passed to them. Removal
3964 might give a very teeny performance improvement.
3965
3966 3. Documentation bug: the value of the capture_top field in a callout is *one
3967 more than* the number of the hightest numbered captured substring.
3968
3969 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3970 in incorrectly linking with a previously installed version. They now link
3971 explicitly with libpcre.la.
3972
3973 5. configure.in no longer needs to recognize Cygwin specially.
3974
3975 6. A problem in pcre.in for Windows platforms is fixed.
3976
3977 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3978 pcretest, it used to include the size of the study block as part of its
3979 output. Unfortunately, the structure contains a field that has a different
3980 size on different hardware architectures. This meant that the tests that
3981 showed this size failed. As the block is currently always of a fixed size,
3982 this information isn't actually particularly useful in pcretest output, so
3983 I have just removed it.
3984
3985 8. Three pre-processor statements accidentally did not start in column 1.
3986 Sadly, there are *still* compilers around that complain, even though
3987 standard C has not required this for well over a decade. Sigh.
3988
3989 9. In pcretest, the code for checking callouts passed small integers in the
3990 callout_data field, which is a void * field. However, some picky compilers
3991 complained about the casts involved for this on 64-bit systems. Now
3992 pcretest passes the address of the small integer instead, which should get
3993 rid of the warnings.
3994
3995 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3996 both compile and run time, and gives an error if an invalid UTF-8 sequence
3997 is found. There is a option for disabling this check in cases where the
3998 string is known to be correct and/or the maximum performance is wanted.
3999
4000 11. In response to a bug report, I changed one line in Makefile.in from
4001
4002 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4003 to
4004 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4005
4006 to look similar to other lines, but I have no way of telling whether this
4007 is the right thing to do, as I do not use Windows. No doubt I'll get told
4008 if it's wrong...
4009
4010
4011 Version 4.3 21-May-03
4012 ---------------------
4013
4014 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4015 Makefile.
4016
4017 2. Some refactoring to improve the quality of the code:
4018
4019 (i) The utf8_table... variables are now declared "const".
4020
4021 (ii) The code for \cx, which used the "case flipping" table to upper case
4022 lower case letters, now just substracts 32. This is ASCII-specific,
4023 but the whole concept of \cx is ASCII-specific, so it seems
4024 reasonable.
4025
4026 (iii) PCRE was using its character types table to recognize decimal and
4027 hexadecimal digits in the pattern. This is silly, because it handles
4028 only 0-9, a-f, and A-F, but the character types table is locale-
4029 specific, which means strange things might happen. A private
4030 table is now used for this - though it costs 256 bytes, a table is
4031 much faster than multiple explicit tests. Of course, the standard
4032 character types table is still used for matching digits in subject
4033 strings against \d.
4034
4035 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4036 ending in _t are). So I've renamed it as ESC_tee.
4037
4038 3. The first argument for regexec() in the POSIX wrapper should have been
4039 defined as "const".
4040
4041 4. Changed pcretest to use malloc() for its buffers so that they can be
4042 Electric Fenced for debugging.
4043
4044 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4045 to read one or more bytes before the start of the subject string. Often this
4046 had no effect on PCRE's behaviour, but in some circumstances it could
4047 provoke a segmentation fault.
4048
4049 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4050 to try to read one or more bytes before the start of the subject string.
4051
4052 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4053 UTF-8 support could misbehave in various ways if the subject string
4054 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4055 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4056 back over UTF-8 characters.)
4057
4058
4059 Version 4.2 14-Apr-03
4060 ---------------------
4061
4062 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4063
4064 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4065 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4066 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4067 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4068 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4069 and BUILD_EXEEXT
4070 Note: automatic setting of the BUILD variables is not yet working
4071 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4072 compile-time but not at link-time
4073 [LINK]: use for linking executables only
4074 make different versions for Windows and non-Windows
4075 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4076 libraries
4077 [LINK_FOR_BUILD]: new variable
4078 [OBJEXT]: use throughout
4079 [EXEEXT]: use throughout
4080 <winshared>: new target
4081 <wininstall>: new target
4082 <dftables.o>: use native compiler
4083 <dftables>: use native linker
4084 <install>: handle Windows platform correctly
4085 <clean>: ditto
4086 <check>: ditto
4087 copy DLL to top builddir before testing
4088
4089 As part of these changes, -no-undefined was removed again. This was reported
4090 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4091 in any case.
4092
4093 3. Some tidies to get rid of compiler warnings:
4094
4095 . In the match_data structure, match_limit was an unsigned long int, whereas
4096 match_call_count was an int. I've made them both unsigned long ints.
4097
4098 . In pcretest the fact that a const uschar * doesn't automatically cast to
4099 a void * provoked a warning.
4100
4101 . Turning on some more compiler warnings threw up some "shadow" variables
4102 and a few more missing casts.
4103
4104 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4105 option, a class that contained a single character with a value between 128
4106 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4107
4108 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4109 option, a class that contained several characters, but with at least one
4110 whose value was between 128 and 255 caused PCRE to crash.
4111
4112
4113 Version 4.1 12-Mar-03
4114 ---------------------
4115
4116 1. Compiling with gcc -pedantic found a couple of places where casts were
4117 needed, and a string in dftables.c that was longer than standard compilers are
4118 required to support.
4119
4120 2. Compiling with Sun's compiler found a few more places where the code could
4121 be tidied up in order to avoid warnings.
4122
4123 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4124 first of these names is deprecated in the latest Autoconf in favour of the name
4125 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4126 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4127 analogy I have changed it to CFLAGS_FOR_BUILD.
4128
4129 4. Added -no-undefined to the linking command in the Makefile, because this is
4130 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4131 linking step for the pcreposix library.
4132
4133 5. PCRE was failing to diagnose the case of two named groups with the same
4134 name.
4135
4136 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4137 literal character that is needed in the subject for a match, and scans along to
4138 ensure that it is present before embarking on the full matching process. This
4139 saves time in cases of nested unlimited repeats that are never going to match.
4140 Problem: the scan can take a lot of time if the subject is very long (e.g.
4141 megabytes), thus penalizing straightforward matches. It is now done only if the
4142 amount of subject to be scanned is less than 1000 bytes.
4143
4144 7. A lesser problem with the same optimization is that it was recording the
4145 first character of an anchored pattern as "needed", thus provoking a search
4146 right along the subject, even when the first match of the pattern was going to
4147 fail. The "needed" character is now not set for anchored patterns, unless it
4148 follows something in the pattern that is of non-fixed length. Thus, it still
4149 fulfils its original purpose of finding quick non-matches in cases of nested
4150 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4151
4152
4153 Version 4.0 17-Feb-03
4154 ---------------------
4155
4156 1. If a comment in an extended regex that started immediately after a meta-item
4157 extended to the end of string, PCRE compiled incorrect data. This could lead to
4158 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4159
4160 2. Moved to autoconf 2.53 and libtool 1.4.2.
4161
4162 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4163 the special perltest8 script is no longer needed - all the tests can be run
4164 from a single perltest script.
4165
4166 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4167 by \s. It has now been removed in PCRE. This means it isn't recognized as
4168 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4169 class [:space:] *does* include VT, thereby creating a mess.
4170
4171 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4172 space and tab.
4173
4174 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4175 its new features into the main test script, reducing the number of scripts.
4176
4177 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4178 were backward compatible, and made the (?i) apply to the whole pattern, as if
4179 /i were given. Now it behaves more logically, and applies the option setting
4180 only to what follows. PCRE has been changed to follow suit. However, if it
4181 finds options settings right at the start of the pattern, it extracts them into
4182 the global options, as before. Thus, they show up in the info data.
4183
4184 8. Added support for the \Q...\E escape sequence. Characters in between are
4185 treated as literals. This is slightly different from Perl in that $ and @ are
4186 also handled as literals inside the quotes. In Perl, they will cause variable
4187 interpolation. Note the following examples:
4188
4189 Pattern PCRE matches Perl matches
4190
4191 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4192 \Qabc\$xyz\E abc\$xyz abc\$xyz
4193 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4194
4195 For compatibility with Perl, \Q...\E sequences are recognized inside character
4196 classes as well as outside them.
4197
4198 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4199 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4200 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4201 signed/unsigned warnings.
4202
4203 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4204 option for pcretest, so I've replaced it by a simple function that does just
4205 that job.
4206
4207 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4208 "pcregrep -".
4209
4210 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4211 Java package. This provides some syntactic sugar for simple cases of what my
4212 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4213 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4214 item, you can use this simplified notation. Note that only makes sense with
4215 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4216 greediness, whatever the setting of the PCRE_UNGREEDY option.
4217
4218 13. A change of greediness default within a pattern was not taking effect at
4219 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4220 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4221 was abstracted outside.
4222
4223 14. PCRE now supports the \G assertion. It is true when the current matching
4224 position is at the start point of the match. This differs from \A when the
4225 starting offset is non-zero. Used with the /g option of pcretest (or similar
4226 code), it works in the same way as it does for Perl's /g option. If all
4227 alternatives of a regex begin with \G, the expression is anchored to the start
4228 match position, and the "anchored" flag is set in the compiled expression.
4229
4230 15. Some bugs concerning the handling of certain option changes within patterns
4231 have been fixed. These applied to options other than (?ims). For example,
4232 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4233 been the other way round. Some of this was related to change 7 above.
4234
4235 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4236 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4237 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4238 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4239
4240 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4241 mode. Unlike ".", it always matches newline, whatever the setting of
4242 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4243 assertions. Perl allows it, but it doesn't (in general) work because it can't
4244 calculate the length of the lookbehind. At least, that's the case for Perl
4245 5.8.0 - I've been told they are going to document that it doesn't work in
4246 future.
4247
4248 18. Added an error diagnosis for escapes that PCRE does not support: these are
4249 \L, \l, \N, \P, \p, \U, \u, and \X.
4250
4251 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4252 reading past the end of the pattern in cases such as /[abcd/.
4253
4254 20. PCRE was getting more memory than necessary for patterns with classes that
4255 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4256
4257 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4258 compiling PCRE for use with Virtual Pascal.
4259
4260 22. Small fix to the Makefile to make it work properly if the build is done
4261 outside the source tree.
4262
4263 23. Added a new extension: a condition to go with recursion. If a conditional
4264 subpattern starts with (?(R) the "true" branch is used if recursion has
4265 happened, whereas the "false" branch is used only at the top level.
4266
4267 24. When there was a very long string of literal characters (over 255 bytes
4268 without UTF support, over 250 bytes with UTF support), the computation of how
4269 much memory was required could be incorrect, leading to segfaults or other
4270 strange effects.
4271
4272 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4273 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4274 there was a subsequent back reference to those brackets. This meant that, for
4275 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4276 possible to check for precisely this case. All we can do is abandon the
4277 optimization if .* occurs inside capturing brackets when there are any back
4278 references whatsoever. (See below for a better fix that came later.)
4279
4280 26. The handling of the optimization for finding the first character of a
4281 non-anchored pattern, and for finding a character that is required later in the
4282 match were failing in some cases. This didn't break the matching; it just
4283 failed to optimize when it could. The way this is done has been re-implemented.
4284
4285 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4286
4287 28. Added a new feature that provides some of the functionality that Perl
4288 provides with (?{...}). The facility is termed a "callout". The way it is done
4289 in PCRE is for the caller to provide an optional function, by setting
4290 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4291 global variable. By default it is unset, which disables all calling out. To get
4292 the function called, the regex must include (?C) at appropriate points. This
4293 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4294 This provides a means of identifying different callout points. When PCRE
4295 reaches such a point in the regex, if pcre_callout has been set, the external
4296 function is called. It is provided with data in a structure called
4297 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4298 matching continues; if it returns a non-zero value, the match at the current
4299 point fails. However, backtracking will occur if possible. [This was changed
4300 later and other features added - see item 49 below.]
4301
4302 29. pcretest is upgraded to test the callout functionality. It provides a
4303 callout function that displays information. By default, it shows the start of
4304 the match and the current position in the text. There are some new data escapes
4305 to vary what happens:
4306
4307 \C+ in addition, show current contents of captured substrings
4308 \C- do not supply a callout function
4309 \C!n return 1 when callout number n is reached
4310 \C!n!m return 1 when callout number n is reached for the mth time
4311
4312 30. If pcregrep was called with the -l option and just a single file name, it
4313 output "<stdin>" if a match was found, instead of the file name.
4314
4315 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4316 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4317 pcre_exec(). This saves a malloc/free per call. The default value of
4318 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4319 when configuring.
4320
4321 32. The default maximum size of a compiled pattern is 64K. There have been a
4322 few cases of people hitting this limit. The code now uses macros to handle the
4323 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4324 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4325 configuring. Tests 2 and 5 work only with 2-byte links because they output
4326 debugging information about compiled patterns.
4327
4328 33. Internal code re-arrangements:
4329
4330 (a) Moved the debugging function for printing out a compiled regex into
4331 its own source file (printint.c) and used #include to pull it into
4332 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4333 separate copies.
4334
4335 (b) Defined the list of op-code names for debugging as a macro in
4336 internal.h so that it is next to the definition of the opcodes.
4337
4338 (c) Defined a table of op-code lengths for simpler skipping along compiled
4339 code. This is again a macro in internal.h so that it is next to the
4340 definition of the opcodes.
4341
4342 34. Added support for recursive calls to individual subpatterns, along the
4343 lines of Robin Houston's patch (but implemented somewhat differently).
4344
4345 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4346 allow it to read and process whole directories in Win32. This code was
4347 contributed by Lionel Fourquaux; it has not been tested by me.
4348
4349 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4350 used to name a group. Names consist of alphanumerics and underscores, and must
4351 be unique. Back references use the syntax (?P=name) and recursive calls use
4352 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4353 numbers. The function pcre_fullinfo() can be used after compilation to extract
4354 a name/number map. There are three relevant calls:
4355
4356 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4357 PCRE_INFO_NAMECOUNT yields the number of entries
4358 PCRE_INFO_NAMETABLE yields a pointer to the map.
4359
4360 The map is a vector of fixed-size entries. The size of each entry depends on
4361 the length of the longest name used. The first two bytes of each entry are the
4362 group number, most significant byte first. There follows the corresponding
4363 name, zero terminated. The names are in alphabetical order.
4364
4365 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4366 case instead of 255. Making it the same both with and without UTF-8 support
4367 means that the same test output works with both.
4368
4369 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4370 calling malloc() with a zero argument.
4371
4372 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4373 optimization. I've improved things by keeping a bitmap of backreferences with
4374 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4375 fact referenced, the optimization can be applied. It is unlikely that a
4376 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4377 the match to follow \n) will appear inside brackets with a number greater than
4378 31, but if it does, any back reference > 31 suppresses the optimization.
4379
4380 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4381 of disabling numbered capturing parentheses. Any opening parenthesis that is
4382 not followed by ? behaves as if it were followed by ?: but named parentheses
4383 can still be used for capturing (and they will acquire numbers in the usual
4384 way).
4385
4386 41. Redesigned the return codes from the match() function into yes/no/error so
4387 that errors can be passed back from deep inside the nested calls. A malloc
4388 failure while inside a recursive subpattern call now causes the
4389 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4390
4391 42. It is now possible to set a limit on the number of times the match()
4392 function is called in a call to pcre_exec(). This facility makes it possible to
4393 limit the amount of recursion and backtracking, though not in a directly
4394 obvious way, because the match() function is used in a number of different
4395 circumstances. The count starts from zero for each position in the subject
4396 string (for non-anchored patterns). The default limit is, for compatibility, a
4397 large number, namely 10 000 000. You can change this in two ways:
4398
4399 (a) When configuring PCRE before making, you can use --with-match-limit=n
4400 to set a default value for the compiled library.
4401
4402 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4403 a different value is set. See 45 below.
4404
4405 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4406
4407 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4408 of things that can be changed at compile time. The first argument specifies
4409 what is wanted and the second points to where the information is to be placed.
4410 The current list of available information is:
4411
4412 PCRE_CONFIG_UTF8
4413
4414 The output is an integer that is set to one if UTF-8 support is available;
4415 otherwise it is set to zero.
4416
4417 PCRE_CONFIG_NEWLINE
4418
4419 The output is an integer that it set to the value of the code that is used for
4420 newline. It is either LF (10) or CR (13).
4421
4422 PCRE_CONFIG_LINK_SIZE
4423
4424 The output is an integer that contains the number of bytes used for internal
4425 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4426
4427 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4428
4429 The output is an integer that contains the threshold above which the POSIX
4430 interface uses malloc() for output vectors. See item 31 above.
4431
4432 PCRE_CONFIG_MATCH_LIMIT
4433
4434 The output is an unsigned integer that contains the default limit of the number
4435 of match() calls in a pcre_exec() execution. See 42 above.
4436
4437 44. pcretest has been upgraded by the addition of the -C option. This causes it
4438 to extract all the available output from the new pcre_config() function, and to
4439 output it. The program then exits immediately.
4440
4441 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4442 order to support additional features. One way would have been to define
4443 pcre_exec2() (for example) with extra arguments, but this would not have been
4444 extensible, and would also have required all calls to the original function to
4445 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4446 is used for passing in "extra" data from pcre_study().
4447
4448 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4449 contains the following fields:
4450
4451 flags a bitmap indicating which of the following fields are set
4452 study_data opaque data from pcre_study()
4453 match_limit a way of specifying a limit on match() calls for a specific
4454 call to pcre_exec()
4455 callout_data data for callouts (see 49 below)
4456
4457 The flag bits are also defined in pcre.h, and are
4458
4459 PCRE_EXTRA_STUDY_DATA
4460 PCRE_EXTRA_MATCH_LIMIT
4461 PCRE_EXTRA_CALLOUT_DATA
4462
4463 The pcre_study() function now returns one of these new pcre_extra blocks, with
4464 the actual study data pointed to by the study_data field, and the
4465 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4466 before. That is, this change is entirely upwards-compatible and requires no
4467 change to existing code.
4468
4469 If you want to pass in additional data to pcre_exec(), you can either place it
4470 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4471 block.
4472
4473 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4474 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4475 times with different match limits, until it finds the minimum value needed for
4476 pcre_exec() to complete. The value is then output. This can be instructive; for
4477 most simple matches the number is quite small, but for pathological cases it
4478 gets very large very quickly.
4479
4480 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4481 returns the size of the data block pointed to by the study_data field in a
4482 pcre_extra block, that is, the value that was passed as the argument to
4483 pcre_malloc() when PCRE was getting memory in which to place the information
4484 created by pcre_study(). The fourth argument should point to a size_t variable.
4485 pcretest has been extended so that this information is shown after a successful
4486 pcre_study() call when information about the compiled regex is being displayed.
4487
4488 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4489 because what follows is always an absolute path. (Later: it turns out that this
4490 is more than cosmetic for MinGW, because it doesn't like empty path
4491 components.)
4492
4493 49. Some changes have been made to the callout feature (see 28 above):
4494
4495 (i) A callout function now has three choices for what it returns:
4496
4497 0 => success, carry on matching
4498 > 0 => failure at this point, but backtrack if possible
4499 < 0 => serious error, return this value from pcre_exec()
4500
4501 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4502 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4503 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4504 use by callout functions. It will never be used by PCRE itself.
4505
4506 (ii) The pcre_extra structure (see 45 above) has a void * field called
4507 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4508 pcre_callout_block structure has a field of the same name. The contents of
4509 the field passed in the pcre_extra structure are passed to the callout
4510 function in the corresponding field in the callout block. This makes it
4511 easier to use the same callout-containing regex from multiple threads. For
4512 testing, the pcretest program has a new data escape
4513
4514 \C*n pass the number n (may be negative) as callout_data
4515
4516 If the callout function in pcretest receives a non-zero value as
4517 callout_data, it returns that value.
4518
4519 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4520 there were some redundant $(CFLAGS) in commands that are now specified as
4521 $(LINK), which already includes $(CFLAGS).
4522
4523 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4524 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4525 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4526 one-byte characters throughout. Note that case-insensitive matching applies
4527 only to characters whose values are less than 256. PCRE doesn't support the
4528 notion of cases for higher-valued characters.
4529
4530 (i) A character class whose characters are all within 0-255 is handled as
4531 a bit map, and the map is inverted for negative classes. Previously, a
4532 character > 255 always failed to match such a class; however it should
4533 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4534
4535 (ii) A negated character class with a single character < 255 is coded as
4536 "not this character" (OP_NOT). This wasn't working properly when the test
4537 character was multibyte, either singly or repeated.
4538
4539 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4540 mode, for example: \x{100}{2,3}.
4541
4542 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4543 singly or repeated) now correctly test multibyte characters. However,
4544 PCRE doesn't recognize any characters with values greater than 255 as
4545 digits, spaces, or word characters. Such characters always match \D, \S,
4546 and \W, and never match \d, \s, or \w.
4547
4548 (v) Classes may now contain characters and character ranges with values
4549 greater than 255. For example: [ab\x{100}-\x{400}].
4550
4551 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4552 PCRE in UTF-8 mode.
4553
4554 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4555 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4556 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4557 value.)
4558
4559 53. The single man page has become too large. I have therefore split it up into
4560 a number of separate man pages. These also give rise to individual HTML pages;
4561 these are now put in a separate directory, and there is an index.html page that
4562 lists them all. Some hyperlinking between the pages has been installed.
4563
4564 54. Added convenience functions for handling named capturing parentheses.
4565
4566 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4567 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4568 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4569 are faulted.
4570
4571 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4572 calling configure. These values are used when compiling the dftables.c program
4573 which is run to generate the source of the default character tables. They
4574 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4575 you will need to set these values.
4576
4577 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4578
4579
4580 Version 3.9 02-Jan-02
4581 ---------------------
4582
4583 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4584
4585 2. If --disable-static was given, the building process failed when trying to
4586 build pcretest and pcregrep. (For some reason it was using libtool to compile
4587 them, which is not right, as they aren't part of the library.)
4588
4589
4590 Version 3.8 18-Dec-01
4591 ---------------------
4592
4593 1. The experimental UTF-8 code was completely screwed up. It was packing the
4594 bytes in the wrong order. How dumb can you get?
4595
4596
4597 Version 3.7 29-Oct-01
4598 ---------------------
4599
4600 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4601 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4602 this didn't happen under Solaris 8, where I normally test things.
4603
4604 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4605 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4606 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4607 this fix an hour or so after the initial 3.7 release.)
4608
4609
4610 Version 3.6 23-Oct-01
4611 ---------------------
4612
4613 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4614 offsets passed as NULL with zero offset count.
4615
4616 2. The config.guess and config.sub files had not been updated when I moved to
4617 the latest autoconf.
4618
4619
4620 Version 3.5 15-Aug-01
4621 ---------------------
4622
4623 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4624 had been forgotten.
4625
4626 2. By using declared but undefined structures, we can avoid using "void"
4627 definitions in pcre.h while keeping the internal definitions of the structures
4628 private.
4629
4630 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4631 user point of view, this means that both static and shared libraries are built
4632 by default, but this can be individually controlled. More of the work of
4633 handling this static/shared cases is now inside libtool instead of PCRE's make
4634 file.
4635
4636 4. The pcretest utility is now installed along with pcregrep because it is
4637 useful for users (to test regexs) and by doing this, it automatically gets
4638 relinked by libtool. The documentation has been turned into a man page, so
4639 there are now .1, .txt, and .html versions in /doc.
4640
4641 5. Upgrades to pcregrep:
4642 (i) Added long-form option names like gnu grep.
4643 (ii) Added --help to list all options with an explanatory phrase.
4644 (iii) Added -r, --recursive to recurse into sub-directories.
4645 (iv) Added -f, --file to read patterns from a file.
4646
4647 6. pcre_exec() was referring to its "code" argument before testing that
4648 argument for NULL (and giving an error if it was NULL).
4649
4650 7. Upgraded Makefile.in to allow for compiling in a different directory from
4651 the source directory.
4652
4653 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4654 options bits, the pointer it was passed was to an int instead of to an unsigned
4655 long int. This mattered only on 64-bit systems.
4656
4657 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4658 generated) instead of pcre.in, which it its source. Also made the same change
4659 in several of the .c files.
4660
4661 10. A new release of gcc defines printf() as a macro, which broke pcretest
4662 because it had an ifdef in the middle of a string argument for printf(). Fixed
4663 by using separate calls to printf().
4664
4665 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4666 script, to force use of CR or LF instead of \n in the source. On non-Unix
4667 systems, the value can be set in config.h.
4668
4669 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4670 absolute limit. Changed the text of the error message to make this clear, and
4671 likewise updated the man page.
4672
4673 13. The limit of 99 on the number of capturing subpatterns has been removed.
4674 The new limit is 65535, which I hope will not be a "real" limit.
4675
4676
4677 Version 3.4 22-Aug-00
4678 ---------------------
4679
4680 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4681
4682 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4683
4684
4685 Version 3.3 01-Aug-00
4686 ---------------------
4687
4688 1. If an octal character was given, but the value was greater than \377, it
4689 was not getting masked to the least significant bits, as documented. This could
4690 lead to crashes in some systems.
4691
4692 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4693 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4694
4695 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4696 These just pass their arguments on to (pcre_free)(), but they are provided
4697 because some uses of PCRE bind it to non-C systems that can call its functions,
4698 but cannot call free() or pcre_free() directly.
4699
4700 4. Add "make test" as a synonym for "make check". Corrected some comments in
4701 the Makefile.
4702
4703 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4704 Makefile.
4705
4706 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4707 command called pgrep for grepping around the active processes.
4708
4709 7. Added the beginnings of support for UTF-8 character strings.
4710
4711 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4712 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4713 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4714 out for the ar command.)
4715
4716
4717 Version 3.2 12-May-00
4718 ---------------------
4719
4720 This is purely a bug fixing release.
4721
4722 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4723 of ZA. This was just one example of several cases that could provoke this bug,
4724 which was introduced by change 9 of version 2.00. The code for breaking
4725 infinite loops after an iteration that matches an empty string was't working
4726 correctly.
4727
4728 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4729 when matched against abbab (for example). After matching an empty string, it
4730 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4731 caused it to match further down the string than it should.
4732
4733 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4734 was there because it doesn't seem to be needed, and it causes trouble on some
4735 systems, as it is not a Standard C header. It has been removed.
4736
4737 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4738 were reported on the Macintosh. The changes were from
4739
4740 while ((c = *(++ptr)) != 0 && c != '\n');
4741 to
4742 while ((c = *(++ptr)) != 0 && c != '\n') ;
4743
4744 Totally extraordinary, but if that's what it takes...
4745
4746 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4747 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4748 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4749 assumes the way PCRE uses memmove() (always moving upwards).
4750
4751 6. PCRE is being used in one environment where strchr() is not available. There
4752 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4753 faster code anyway.
4754
4755
4756 Version 3.1 09-Feb-00
4757 ---------------------
4758
4759 The only change in this release is the fixing of some bugs in Makefile.in for
4760 the "install" target:
4761
4762 (1) It was failing to install pcreposix.h.
4763
4764 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4765
4766
4767 Version 3.0 01-Feb-00
4768 ---------------------
4769
4770 1. Add support for the /+ modifier to perltest (to output $` like it does in
4771 pcretest).
4772
4773 2. Add support for the /g modifier to perltest.
4774
4775 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4776 matches null strings.
4777
4778 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4779 pattern. Perl treats empty patterns specially - it reuses the most recent
4780 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4781 effect.
4782
4783 5. The POSIX interface was broken in that it was just handing over the POSIX
4784 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4785 required a bigger vector, with some working space on the end. This means that
4786 the POSIX wrapper now has to get and free some memory, and copy the results.
4787
4788 6. Added some simple autoconf support, placing the test data and the
4789 documentation in separate directories, re-organizing some of the
4790 information files, and making it build pcre-config (a GNU standard). Also added
4791 libtool support for building PCRE as a shared library, which is now the
4792 default.
4793
4794 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4795 09 are not valid octal constants. Single digits will be used for minor values
4796 less than 10.
4797
4798 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4799 existing programs that set these in the POSIX interface can use PCRE without
4800 modification.
4801
4802 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4803 return all that pcre_info() returns, plus additional data. The pcre_info()
4804 function is retained for compatibility, but is considered to be obsolete.
4805
4806 10. Added experimental recursion feature (?R) to handle one common case that
4807 Perl 5.6 will be able to do with (?p{...}).
4808
4809 11. Added support for POSIX character classes like [:alpha:], which Perl is
4810 adopting.
4811
4812
4813 Version 2.08 31-Aug-99
4814 ----------------------
4815
4816 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4817 trying to match at the startoffset position, but instead was moving forward to
4818 the next newline as if a previous match had failed.
4819
4820 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4821 and could get into a loop if a null string was matched other than at the start
4822 of the subject.
4823
4824 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4825 be distinguished at compile time, and for completeness also added PCRE_DATE.
4826
4827 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4828 in GnuWin32 environments.
4829
4830
4831 Version 2.07 29-Jul-99
4832 ----------------------
4833
4834 1. The documentation is now supplied in plain text form and HTML as well as in
4835 the form of man page sources.
4836
4837 2. C++ compilers don't like assigning (void *) values to other pointer types.
4838 In particular this affects malloc(). Although there is no problem in Standard
4839 C, I've put in casts to keep C++ compilers happy.
4840
4841 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4842 should be (const char *).
4843
4844 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4845 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4846 However, I haven't made this a standard facility. The documentation doesn't
4847 mention it, and the Makefile doesn't support it.
4848
4849 5. The Makefile now contains an "install" target, with editable destinations at
4850 the top of the file. The pcretest program is not installed.
4851
4852 6. pgrep -V now gives the PCRE version number and date.
4853
4854 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4855 causing the entire string to be ignored, instead of just the last character.
4856
4857 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4858 non-matching string, it can take a very, very long time, even for strings of
4859 quite modest length, because of the nested recursion. PCRE now does better in
4860 some of these cases. It does this by remembering the last required literal
4861 character in the pattern, and pre-searching the subject to ensure it is present
4862 before running the real match. In other words, it applies a heuristic to detect
4863 some types of certain failure quickly, and in the above example, if presented
4864 with a string that has no trailing " it gives "no match" very quickly.
4865
4866 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4867 other alternatives are tried instead.
4868
4869
4870 Version 2.06 09-Jun-99
4871 ----------------------
4872
4873 1. Change pcretest's output for amount of store used to show just the code
4874 space, because the remainder (the data block) varies in size between 32-bit and
4875 64-bit systems.
4876
4877 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4878 start matching at. This allows lookbehinds to work when searching for multiple
4879 occurrences in a string.
4880
4881 3. Added additional options to pcretest for testing multiple occurrences:
4882
4883 /+ outputs the rest of the string that follows a match
4884 /g loops for multiple occurrences, using the new startoffset argument
4885 /G loops for multiple occurrences by passing an incremented pointer
4886
4887 4. PCRE wasn't doing the "first character" optimization for patterns starting
4888 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4889 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4890 the letter 'x'. On long subject strings, this gives a significant speed-up.
4891
4892
4893 Version 2.05 21-Apr-99
4894 ----------------------
4895
4896 1. Changed the type of magic_number from int to long int so that it works
4897 properly on 16-bit systems.
4898
4899 2. Fixed a bug which caused patterns starting with .* not to work correctly
4900 when the subject string contained newline characters. PCRE was assuming
4901 anchoring for such patterns in all cases, which is not correct because .* will
4902 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4903 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4904 must be retried after every newline in the subject.
4905
4906
4907 Version 2.04 18-Feb-99
4908 ----------------------
4909
4910 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4911 computation of the store needed to hold the pattern was incorrect (too large).
4912 If such patterns were nested a few deep, this could multiply and become a real
4913 problem.
4914
4915 2. Added /M option to pcretest to show the memory requirement of a specific
4916 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4917
4918 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4919 compiled in such a way that the backtracking after subsequent failure was
4920 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4921 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4922
4923
4924 Version 2.03 02-Feb-99
4925 ----------------------
4926
4927 1. Fixed typo and small mistake in man page.
4928
4929 2. Added 4th condition (GPL supersedes if conflict) and created separate
4930 LICENCE file containing the conditions.
4931
4932 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4933 Perl, that is the internal \ allows the delimiter to be included in the
4934 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4935 the final delimiter, add \ to the end of the pattern (to test the error).
4936
4937 4. Added the convenience functions for extracting substrings after a successful
4938 match. Updated pcretest to make it able to test these functions.
4939
4940
4941 Version 2.02 14-Jan-99
4942 ----------------------
4943
4944 1. Initialized the working variables associated with each extraction so that
4945 their saving and restoring doesn't refer to uninitialized store.
4946
4947 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4948 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4949 fix the problem.
4950
4951 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4952 calls, and wasn't printing the correct value for compiling calls. Increased the
4953 default value of LOOPREPEAT, and the number of significant figures in the
4954 times.
4955
4956 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4957
4958 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4959 a building problem on Windows NT with a FAT file system.
4960
4961
4962 Version 2.01 21-Oct-98
4963 ----------------------
4964
4965 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4966 to character tables built by pcre_maketables() in the current locale. If NULL
4967 is passed, the default tables are used.
4968
4969
4970 Version 2.00 24-Sep-98
4971 ----------------------
4972
4973 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4974 it any more.
4975
4976 2. Allow quantification of (?>) groups, and make it work correctly.
4977
4978 3. The first character computation wasn't working for (?>) groups.
4979
4980 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4981 end of the subject) and add 5.005's \z, which really does match only at the
4982 very end of the subject.
4983
4984 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4985
4986 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4987 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4988 localized options. All options to pcre_study() were also removed.
4989
4990 7. Add other new features from 5.005:
4991
4992 $(?<= positive lookbehind
4993 $(?<! negative lookbehind
4994 (?imsx-imsx) added the unsetting capability
4995 such a setting is global if at outer level; local otherwise
4996 (?imsx-imsx:) non-capturing groups with option setting
4997 (?(cond)re|re) conditional pattern matching
4998
4999 A backreference to itself in a repeated group matches the previous
5000 captured string.
5001
5002 8. General tidying up of studying (both automatic and via "study")
5003 consequential on the addition of new assertions.
5004
5005 9. As in 5.005, unlimited repeated groups that could match an empty substring
5006 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5007 runtime if any iteration does actually match an empty substring.
5008
5009 10. Include the RunTest script in the distribution.
5010
5011 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5012 discrepancies, some of which were old and were also with respect to 5.004. They
5013 have now been fixed.
5014
5015
5016 Version 1.09 28-Apr-98
5017 ----------------------
5018
5019 1. A negated single character class followed by a quantifier with a minimum
5020 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5021 program crashes, or just wrong answers. This did not apply to negated classes
5022 containing more than one character, or to minima other than one.
5023
5024
5025 Version 1.08 27-Mar-98
5026 ----------------------
5027
5028 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5029
5030 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5031 latter must appear before anything that relies on it in the pattern.
5032
5033
5034 Version 1.07 16-Feb-98
5035 ----------------------
5036
5037 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5038 repeat of a potentially empty string).
5039
5040
5041 Version 1.06 23-Jan-98
5042 ----------------------
5043
5044 1. Added Markus Oberhumer's little patches for C++.
5045
5046 2. Literal strings longer than 255 characters were broken.
5047
5048
5049 Version 1.05 23-Dec-97
5050 ----------------------
5051
5052 1. Negated character classes containing more than one character were failing if
5053 PCRE_CASELESS was set at run time.
5054
5055
5056 Version 1.04 19-Dec-97
5057 ----------------------
5058
5059 1. Corrected the man page, where some "const" qualifiers had been omitted.
5060
5061 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5062 input syntax.
5063
5064 3. Fixed memory leak which occurred when a regex with back references was
5065 matched with an offsets vector that wasn't big enough. The temporary memory
5066 that is used in this case wasn't being freed if the match failed.
5067
5068 4. Tidied pcretest to ensure it frees memory that it gets.
5069
5070 5. Temporary memory was being obtained in the case where the passed offsets
5071 vector was exactly big enough.
5072
5073 6. Corrected definition of offsetof() from change 5 below.
5074
5075 7. I had screwed up change 6 below and broken the rules for the use of
5076 setjmp(). Now fixed.
5077
5078
5079 Version 1.03 18-Dec-97
5080 ----------------------
5081
5082 1. A erroneous regex with a missing opening parenthesis was correctly
5083 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5084 on some systems.
5085
5086 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5087 it was reported that one broken compiler failed on the former because "code" is
5088 also an independent variable.
5089
5090 3. The erroneous regex a[]b caused an array overrun reference.
5091
5092 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5093 fail on data ending with that character. (It was going on too far, and checking
5094 the next character, typically a binary zero.) This was specific to the
5095 optimized code for single-character negative classes.
5096
5097 5. Added a contributed patch from the TIN world which does the following:
5098
5099 + Add an undef for memmove, in case the the system defines a macro for it.
5100
5101 + Add a definition of offsetof(), in case there isn't one. (I don't know
5102 the reason behind this - offsetof() is part of the ANSI standard - but
5103 it does no harm).
5104
5105 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5106 most of the places where whitespace preceded '#'. I have given up and
5107 allowed the remaining 2 cases to be at the margin.
5108
5109 + Rename some variables in pcre to eliminate shadowing. This seems very
5110 pedantic, but does no harm, of course.
5111
5112 6. Moved the call to setjmp() into its own function, to get rid of warnings
5113 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5114
5115 7. Constructs such as \d{8,} were compiling into the equivalent of
5116 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5117 outcome, but in this particular case used more store than had been allocated,
5118 which caused the bug to be discovered because it threw up an internal error.
5119
5120 8. The debugging code in both pcre and pcretest for outputting the compiled
5121 form of a regex was going wrong in the case of back references followed by
5122 curly-bracketed repeats.
5123
5124
5125 Version 1.02 12-Dec-97
5126 ----------------------
5127
5128 1. Typos in pcre.3 and comments in the source fixed.
5129
5130 2. Applied a contributed patch to get rid of places where it used to remove
5131 'const' from variables, and fixed some signed/unsigned and uninitialized
5132 variable warnings.
5133
5134 3. Added the "runtest" target to Makefile.
5135
5136 4. Set default compiler flag to -O2 rather than just -O.
5137
5138
5139 Version 1.01 19-Nov-97
5140 ----------------------
5141
5142 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5143 like /([ab]*)*/, that is, for classes with more than one character in them.
5144
5145 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5146 as /((?>a*))*/ (a PCRE_EXTRA facility).
5147
5148
5149 Version 1.00 18-Nov-97
5150 ----------------------
5151
5152 1. Added compile-time macros to support systems such as SunOS4 which don't have
5153 memmove() or strerror() but have other things that can be used instead.
5154
5155 2. Arranged that "make clean" removes the executables.
5156
5157
5158 Version 0.99 27-Oct-97
5159 ----------------------
5160
5161 1. Fixed bug in code for optimizing classes with only one character. It was
5162 initializing a 32-byte map regardless, which could cause it to run off the end
5163 of the memory it had got.
5164
5165 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5166
5167
5168 Version 0.98 22-Oct-97
5169 ----------------------
5170
5171 1. Fixed bug in code for handling temporary memory usage when there are more
5172 back references than supplied space in the ovector. This could cause segfaults.
5173
5174
5175 Version 0.97 21-Oct-97
5176 ----------------------
5177
5178 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5179
5180 2. Optimized negated single characters not to use a bit map.
5181
5182 3. Brought error texts together as macro definitions; clarified some of them;
5183 fixed one that was wrong - it said "range out of order" when it meant "invalid
5184 escape sequence".
5185
5186 4. Changed some char * arguments to const char *.
5187
5188 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5189
5190 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5191 pcretest.
5192
5193
5194 Version 0.96 16-Oct-97
5195 ----------------------
5196
5197 1. Added a simple "pgrep" utility to the distribution.
5198
5199 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5200 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5201 where "ddd" means "one or more decimal digits".
5202
5203 3. Fixed serious bug. If a pattern had a back reference, but the call to
5204 pcre_exec() didn't supply a large enough ovector to record the related
5205 identifying subpattern, the match always failed. PCRE now remembers the number
5206 of the largest back reference, and gets some temporary memory in which to save
5207 the offsets during matching if necessary, in order to ensure that
5208 backreferences always work.
5209
5210 4. Increased the compatibility with Perl in a number of ways:
5211
5212 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5213 to request this handling. The option can be set at compile or exec time.
5214
5215 (b) $ matches before a terminating newline by default; an option
5216 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5217 mode). The option can be set at compile or exec time.
5218
5219 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5220 the same as Perl's. If the decimal number it represents is less than 10
5221 or there aren't that many previous left capturing parentheses, an octal
5222 escape is read. Inside a character class, it's always an octal escape,
5223 even if it is a single digit.
5224
5225 (d) An escaped but undefined alphabetic character is taken as a literal,
5226 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5227 escapes.
5228
5229 (e) {0} is now permitted. (The previous item is removed from the compiled
5230 pattern).
5231
5232 5. Changed all the names of code files so that the basic parts are no longer
5233 than 10 characters, and abolished the teeny "globals.c" file.
5234
5235 6. Changed the handling of character classes; they are now done with a 32-byte
5236 bit map always.
5237
5238 7. Added the -d and /D options to pcretest to make it possible to look at the
5239 internals of compilation without having to recompile pcre.
5240
5241
5242 Version 0.95 23-Sep-97
5243 ----------------------
5244
5245 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5246 \x20 at the start of a run of normal characters. These were being treated as
5247 real characters, instead of the source characters being re-checked.
5248
5249
5250 Version 0.94 18-Sep-97
5251 ----------------------
5252
5253 1. The functions are now thread-safe, with the caveat that the global variables
5254 containing pointers to malloc() and free() or alternative functions are the
5255 same for all threads.
5256
5257 2. Get pcre_study() to generate a bitmap of initial characters for non-
5258 anchored patterns when this is possible, and use it if passed to pcre_exec().
5259
5260
5261 Version 0.93 15-Sep-97
5262 ----------------------
5263
5264 1. /(b)|(:+)/ was computing an incorrect first character.
5265
5266 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5267 but not actually doing anything yet.
5268
5269 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5270 as Perl does (e.g. [-az] or [az-]).
5271
5272 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5273 all possible positions.
5274
5275 5. Split up into different modules to avoid including unneeded functions in a
5276 compiled binary. However, compile and exec are still in one module. The "study"
5277 function is split off.
5278
5279 6. The character tables are now in a separate module whose source is generated
5280 by an auxiliary program - but can then be edited by hand if required. There are
5281 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5282 toupper() in the code.
5283
5284 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5285 make them global. Abolish the function for setting them, as the caller can now
5286 set them directly.
5287
5288
5289 Version 0.92 11-Sep-97
5290 ----------------------
5291
5292 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5293 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5294
5295 2. Caseless matching was not working in character classes if the characters in
5296 the pattern were in upper case.
5297
5298 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5299
5300 4. Make PCRE_ANCHORED public and accept as a compile option.
5301
5302 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5303 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5304 pass them.
5305
5306 6. Give an error if bad option bits passed at compile or run time.
5307
5308 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5309 pcretest to cause it to pass that flag.
5310
5311 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5312 options, and the first character, if set.
5313
5314 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5315
5316
5317 Version 0.91 10-Sep-97
5318 ----------------------
5319
5320 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5321 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5322
5323 2. PCRE was looping on encountering an indefinitely repeated back reference to
5324 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5325 Perl does - treats the match as successful.
5326
5327 ****

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

  ViewVC Help
Powered by ViewVC 1.1.5