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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1393 - (show annotations)
Fri Nov 8 16:37:21 2013 UTC (5 years, 11 months ago) by ph10
File size: 256387 byte(s)
Error occurred while calculating annotation data.
Allow quantifiers on (?!) so as to be the same as other assertions.
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.
90
91 18. The character VT has been added to the set of characters that match \s and
92 are generally treated as white space, following this same change in Perl
93 5.18. There is now no difference between "Perl space" and "POSIX space".
94
95 19. The code for checking named groups as conditions, either for being set or
96 for being recursed, has been refactored (this is related to 14 and 15
97 above). Processing unduplicated named groups should now be as fast at
98 numerical groups, and processing duplicated groups should be faster than
99 before.
100
101 20. Two patches to the CMake build system, by Alexander Barkov:
102
103 (1) Replace the "source" command by "." in CMakeLists.txt because
104 "source" is a bash-ism.
105
106 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
107 without these the CMake build does not work on Solaris.
108
109 21. Perl has changed its handling of \8 and \9. If there is no previously
110 encountered capturing group of those numbers, they are treated as the
111 literal characters 8 and 9 instead of a binary zero followed by the
112 literals. PCRE now does the same.
113
114 22. Following Perl, added \o{} to specify codepoints in octal, making it
115 possible to specify values greater than 0777 and also making them
116 unambiguous.
117
118 23. Perl now gives an error for missing closing braces after \x{... instead of
119 treating the string as literal. PCRE now does the same.
120
121 24. RunTest used to grumble if an inappropriate test was selected explicitly,
122 but just skip it when running all tests. This make it awkward to run ranges
123 of tests when one of them was inappropriate. Now it just skips any
124 inappropriate tests, as it always did when running all tests.
125
126 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
127 character types such as \d or \w, too many callouts were inserted, and the
128 data that they returned was rubbish.
129
130 26. In UCP mode, \s was not matching two of the characters that Perl matches,
131 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
132 were matched by \h. The code has now been refactored so that the lists of
133 the horizontal and vertical whitespace characters used for \h and \v (which
134 are defined only in one place) are now also used for \s.
135
136 27. Add JIT support for the 64 bit TileGX architecture.
137 Patch by Jiong Wang (Tilera Corporation).
138
139 28. Possessive quantifiers for classes (both explicit and automatically
140 generated) now use special opcodes instead of wrapping in ONCE brackets.
141
142 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
143 (because it's meaningless), this was not happening when PCRE_CASELESS was
144 set. Not wrong, but inefficient.
145
146 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
147 properties for \w, \d, etc) is present in a test regex. Otherwise if the
148 test contains no characters greater than 255, Perl doesn't realise it
149 should be using Unicode semantics.
150
151 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
152 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
153 does in Unicode mode.
154
155 32. Added the "forbid" facility to pcretest so that putting tests into the
156 wrong test files can sometimes be quickly detected.
157
158 33. There is now a limit (default 250) on the depth of nesting of parentheses.
159 This limit is imposed to control the amount of system stack used at compile
160 time. It can be changed at build time by --with-parens-nest-limit=xxx or
161 the equivalent in CMake.
162
163 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
164 errors. Perl warns for these when in warning mode, but PCRE has no facility
165 for giving warnings.
166
167 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
168 However, this was not working for (?!) because it is optimized to (*FAIL),
169 for which PCRE does not allow quantifiers. The optimization is now disabled
170 when a quantifier follows (?!). I can't see any use for this, but it makes
171 things uniform.
172
173
174 Version 8.33 28-May-2013
175 ------------------------
176
177 1. Added 'U' to some constants that are compared to unsigned integers, to
178 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
179 variables that are added to signed variables, to ensure the result is
180 signed and can be negated.
181
182 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
183 CMake config files.
184
185 3. Revise the creation of config.h.generic so that all boolean macros are
186 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
187 overriding via -D on the command line possible.
188
189 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
190 to unsigned int is reported to make a quite noticeable speed difference in
191 a specific Windows environment. Testing on Linux did also appear to show
192 some benefit (and it is clearly not harmful). Also fixed the definition of
193 Xop which should be unsigned.
194
195 5. Related to (4), changing the definition of the intermediate variable cc
196 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
197 improvements.
198
199 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
200 unnecessary spaces.
201
202 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
203 and later.
204
205 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
206
207 9. Optimizing fast_forward_start_bits in JIT.
208
209 10. Adding support for callouts in JIT, and fixing some issues revealed
210 during this work. Namely:
211
212 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
213
214 (b) Minimum length was not checked before the matching is started.
215
216 11. The value of capture_last that is passed to callouts was incorrect in some
217 cases when there was a capture on one path that was subsequently abandoned
218 after a backtrack. Also, the capture_last value is now reset after a
219 recursion, since all captures are also reset in this case.
220
221 12. The interpreter no longer returns the "too many substrings" error in the
222 case when an overflowing capture is in a branch that is subsequently
223 abandoned after a backtrack.
224
225 13. In the pathological case when an offset vector of size 2 is used, pcretest
226 now prints out the matched string after a yield of 0 or 1.
227
228 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
229 Only supported by the JIT compiler at the moment.
230
231 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
232
233 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
234 offset of the starting point of the matching process, provided the offsets
235 vector is large enough.
236
237 17. The \A escape now records a lookbehind value of 1, though its execution
238 does not actually inspect the previous character. This is to ensure that,
239 in partial multi-segment matching, at least one character from the old
240 segment is retained when a new segment is processed. Otherwise, if there
241 are no lookbehinds in the pattern, \A might match incorrectly at the start
242 of a new segment.
243
244 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
245
246 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
247 gives some modest performance improvement in 8-bit mode.
248
249 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
250 be expressed in certain programming languages using Universal Character
251 Names.
252
253 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
254 which points out that "non characters" are not "characters that may not
255 appear in Unicode strings" but rather "characters that are reserved for
256 internal use and have only local meaning".
257
258 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
259 there was a conditional group that depended on an assertion, if the
260 assertion was false, the callout that immediately followed the alternation
261 in the condition was skipped when pcre_exec() was used for matching.
262
263 23. Allow an explicit callout to be inserted before an assertion that is the
264 condition for a conditional group, for compatibility with automatic
265 callouts, which always insert a callout at this point.
266
267 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
268 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
269
270 25. (*PRUNE) is now supported by the JIT compiler.
271
272 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
273
274 27. Fix the case where there are two or more SKIPs with arguments that may be
275 ignored.
276
277 28. (*SKIP) is now supported by the JIT compiler.
278
279 29. (*THEN) is now supported by the JIT compiler.
280
281 30. Update RunTest with additional test selector options.
282
283 31. The way PCRE handles backtracking verbs has been changed in two ways.
284
285 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
286 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
287 backtracking. In some cases this makes it more Perl-compatible, but Perl's
288 rather obscure rules do not always do the same thing.
289
290 (2) Previously, backtracking verbs were confined within assertions. This is
291 no longer the case for positive assertions, except for (*ACCEPT). Again,
292 this sometimes improves Perl compatibility, and sometimes does not.
293
294 32. A number of tests that were in test 2 because Perl did things differently
295 have been moved to test 1, because either Perl or PCRE has changed, and
296 these tests are now compatible.
297
298 32. Backtracking control verbs are now handled in the same way in JIT and
299 interpreter.
300
301 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
302 contained a forward subroutine reference caused a compile error.
303
304 34. Auto-detect and optimize limited repetitions in JIT.
305
306 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
307 blocking (*UTF) etc.
308
309 36. In the interpreter, maximizing pattern repetitions for characters and
310 character types now use tail recursion, which reduces stack usage.
311
312 37. The value of the max lookbehind was not correctly preserved if a compiled
313 and saved regex was reloaded on a host of different endianness.
314
315 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
316 of the compiled pattern block, expand the flags field from 16 to 32 bits
317 because it was almost full.
318
319 39. Try madvise first before posix_madvise.
320
321 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
322 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
323 original change locked out all matches of empty strings. This has been
324 changed so that one match of an empty string per line is recognized.
325 Subsequent searches on the same line (for colouring or for --only-matching,
326 for example) do not recognize empty strings.
327
328 41. Applied a user patch to fix a number of spelling mistakes in comments.
329
330 42. Data lines longer than 65536 caused pcretest to crash.
331
332 43. Clarified the data type for length and startoffset arguments for pcre_exec
333 and pcre_dfa_exec in the function-specific man pages, where they were
334 explicitly stated to be in bytes, never having been updated. I also added
335 some clarification to the pcreapi man page.
336
337 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
338 a segmentation fault.
339
340
341 Version 8.32 30-November-2012
342 -----------------------------
343
344 1. Improved JIT compiler optimizations for first character search and single
345 character iterators.
346
347 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
348 Patch by Daniel Richard G.
349
350 3. Single character iterator optimizations in the JIT compiler.
351
352 4. Improved JIT compiler optimizations for character ranges.
353
354 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
355 Reported by Giuseppe D'Angelo.
356
357 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
358 of a line, was being set incorrectly in cases where .* appeared inside
359 atomic brackets at the start of a pattern, or where there was a subsequent
360 *PRUNE or *SKIP.
361
362 7. Improved instruction cache flush for POWER/PowerPC.
363 Patch by Daniel Richard G.
364
365 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
366 grep:
367
368 (a) There is now no limit to the number of patterns to be matched.
369
370 (b) An error is given if a pattern is too long.
371
372 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
373 are now supported.
374
375 (d) --exclude-from and --include-from (multiple use) have been added.
376
377 (e) Exclusions and inclusions now apply to all files and directories, not
378 just to those obtained from scanning a directory recursively.
379
380 (f) Multiple uses of -f and --file-list are now supported.
381
382 (g) In a Windows environment, the default for -d has been changed from
383 "read" (the GNU grep default) to "skip", because otherwise the presence
384 of a directory in the file list provokes an error.
385
386 (h) The documentation has been revised and clarified in places.
387
388 9. Improve the matching speed of capturing brackets.
389
390 10. Changed the meaning of \X so that it now matches a Unicode extended
391 grapheme cluster.
392
393 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
394 out POSIX threads when JIT support is configured.
395
396 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
397
398 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
399 structure could go wrong in environments where size_t is not the same size
400 as int.
401
402 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
403 set.
404
405 15. The EBCDIC support had decayed; later updates to the code had included
406 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
407 general tidy up of EBCDIC-related issues, and the documentation was also
408 not quite right. There is now a test that can be run on ASCII systems to
409 check some of the EBCDIC-related things (but is it not a full test).
410
411 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
412 in a small tidy to the code.
413
414 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
415
416 18. If the --only-matching (-o) option in pcregrep is specified multiple
417 times, each one causes appropriate output. For example, -o1 -o2 outputs the
418 substrings matched by the 1st and 2nd capturing parentheses. A separating
419 string can be specified by --om-separator (default empty).
420
421 19. Improving the first n character searches.
422
423 20. Turn case lists for horizontal and vertical white space into macros so that
424 they are defined only once.
425
426 21. This set of changes together give more compatible Unicode case-folding
427 behaviour for characters that have more than one other case when UCP
428 support is available.
429
430 (a) The Unicode property table now has offsets into a new table of sets of
431 three or more characters that are case-equivalent. The MultiStage2.py
432 script that generates these tables (the pcre_ucd.c file) now scans
433 CaseFolding.txt instead of UnicodeData.txt for character case
434 information.
435
436 (b) The code for adding characters or ranges of characters to a character
437 class has been abstracted into a generalized function that also handles
438 case-independence. In UTF-mode with UCP support, this uses the new data
439 to handle characters with more than one other case.
440
441 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
442 whose other case is greater than 256 are now correctly matched
443 caselessly. Previously, the high codepoint matched the low one, but not
444 vice versa.
445
446 (d) The processing of \h, \H, \v, and \ in character classes now makes use
447 of the new class addition function, using character lists defined as
448 macros alongside the case definitions of 20 above.
449
450 (e) Caseless back references now work with characters that have more than
451 one other case.
452
453 (f) General caseless matching of characters with more than one other case
454 is supported.
455
456 22. Unicode character properties were updated from Unicode 6.2.0
457
458 23. Improved CMake support under Windows. Patch by Daniel Richard G.
459
460 24. Add support for 32-bit character strings, and UTF-32
461
462 25. Major JIT compiler update (code refactoring and bugfixing).
463 Experimental Sparc 32 support is added.
464
465 26. Applied a modified version of Daniel Richard G's patch to create
466 pcre.h.generic and config.h.generic by "make" instead of in the
467 PrepareRelease script.
468
469 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
470 pcre_compile.c when checking for a zero character.
471
472 28. Introducing a native interface for JIT. Through this interface, the compiled
473 machine code can be directly executed. The purpose of this interface is to
474 provide fast pattern matching, so several sanity checks are not performed.
475 However, feature tests are still performed. The new interface provides
476 1.4x speedup compared to the old one.
477
478 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
479 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
480 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
481
482 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
483 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
484 the "old" RFC 2279). Instead, it ended up passing a negative length to
485 pcre_exec().
486
487 31. Add support for GCC's visibility feature to hide internal functions.
488
489 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
490 "unknown -C option" after outputting 0 or 1.
491
492 33. There is now support for generating a code coverage report for the test
493 suite in environments where gcc is the compiler and lcov is installed. This
494 is mainly for the benefit of the developers.
495
496 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
497 unaddressable using valgrind annotations, allowing valgrind to detect
498 invalid memory accesses. This is mainly for the benefit of the developers.
499
500 25. (*UTF) can now be used to start a pattern in any of the three libraries.
501
502 26. Give configure error if --enable-cpp but no C++ compiler found.
503
504
505 Version 8.31 06-July-2012
506 -------------------------
507
508 1. Fixing a wrong JIT test case and some compiler warnings.
509
510 2. Removed a bashism from the RunTest script.
511
512 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
513 to unsigned type, result still unsigned" that was given by an MS compiler
514 on encountering the code "-sizeof(xxx)".
515
516 4. Partial matching support is added to the JIT compiler.
517
518 5. Fixed several bugs concerned with partial matching of items that consist
519 of more than one character:
520
521 (a) /^(..)\1/ did not partially match "aba" because checking references was
522 done on an "all or nothing" basis. This also applied to repeated
523 references.
524
525 (b) \R did not give a hard partial match if \r was found at the end of the
526 subject.
527
528 (c) \X did not give a hard partial match after matching one or more
529 characters at the end of the subject.
530
531 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
532 a partial match for the string "\r".
533
534 (e) When newline was set to CRLF, the metacharacter "." did not recognize
535 a partial match for a CR character at the end of the subject string.
536
537 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
538 running pcretest, the text "(JIT)" added to the output whenever JIT is
539 actually used to run the match.
540
541 7. Individual JIT compile options can be set in pcretest by following -s+[+]
542 or /S+[+] with a digit between 1 and 7.
543
544 8. OP_NOT now supports any UTF character not just single-byte ones.
545
546 9. (*MARK) control verb is now supported by the JIT compiler.
547
548 10. The command "./RunTest list" lists the available tests without actually
549 running any of them. (Because I keep forgetting what they all are.)
550
551 11. Add PCRE_INFO_MAXLOOKBEHIND.
552
553 12. Applied a (slightly modified) user-supplied patch that improves performance
554 when the heap is used for recursion (compiled with --disable-stack-for-
555 recursion). Instead of malloc and free for each heap frame each time a
556 logical recursion happens, frames are retained on a chain and re-used where
557 possible. This sometimes gives as much as 30% improvement.
558
559 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
560 call.
561
562 14. As documented, (*COMMIT) is now confined to within a positive assertion.
563
564 15. It is now possible to link pcretest with libedit as an alternative to
565 libreadline.
566
567 16. (*COMMIT) control verb is now supported by the JIT compiler.
568
569 17. The Unicode data tables have been updated to Unicode 6.1.0.
570
571 18. Added --file-list option to pcregrep.
572
573 19. Added binary file support to pcregrep, including the -a, --binary-files,
574 -I, and --text options.
575
576 20. The madvise function is renamed for posix_madvise for QNX compatibility
577 reasons. Fixed by Giuseppe D'Angelo.
578
579 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
580
581 22. Changed the option for creating symbolic links for 16-bit man pages from
582 -s to -sf so that re-installing does not cause issues.
583
584 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
585
586 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
587 matches in certain environments (the workspace was not being correctly
588 retained). Also added to pcre_dfa_exec() a simple plausibility check on
589 some of the workspace data at the beginning of a restart.
590
591 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
592 was not doing so when it should - probably a typo introduced by SVN 528
593 (change 8.10/14).
594
595 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
596 \w+ when the character tables indicated that \x{c4} was a word character.
597 There were several related cases, all because the tests for doing a table
598 lookup were testing for characters less than 127 instead of 255.
599
600 27. If a pattern contains capturing parentheses that are not used in a match,
601 their slots in the ovector are set to -1. For those that are higher than
602 any matched groups, this happens at the end of processing. In the case when
603 there were back references that the ovector was too small to contain
604 (causing temporary malloc'd memory to be used during matching), and the
605 highest capturing number was not used, memory off the end of the ovector
606 was incorrectly being set to -1. (It was using the size of the temporary
607 memory instead of the true size.)
608
609 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
610 ovector size, it uses memory at the end of the block that it has got.
611
612 29. Check for an overlong MARK name and give an error at compile time. The
613 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
614
615 30. JIT compiler update.
616
617 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
618 Rill for the patch.
619
620 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
621
622 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
623
624 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
625 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
626 was enabled.
627
628 35. Improve JIT code generation for greedy plus quantifier.
629
630 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
631 1 to "aa" instead of to an empty string. The bug affected repeated groups
632 that could potentially match an empty string.
633
634 37. Optimizing single character iterators in JIT.
635
636 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
637 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
638 codepoints that are too big for the mode are faulted, and in a UTF mode,
639 disallowed codepoints are also faulted.
640
641 39. If PCRE was compiled with UTF support, in three places in the DFA
642 matcher there was code that should only have been obeyed in UTF mode, but
643 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
644 processing when bytes with values greater than 127 were present. In 16-bit
645 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
646 both cases the values are those that cannot be the first data item in a UTF
647 character. The three items that might have provoked this were recursions,
648 possessively repeated groups, and atomic groups.
649
650 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
651 and pcregrep, because some OS require shared objects to be explicitly
652 passed to ld, causing the link step to fail if they are not.
653
654 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
655 mode, patterns that started with \h* or \R* might be incorrectly matched.
656
657
658 Version 8.30 04-February-2012
659 -----------------------------
660
661 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
662 name is defined in ctype.h.
663
664 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
665 only in quite long subpatterns.
666
667 3. Removed the function pcre_info(), which has been obsolete and deprecated
668 since it was replaced by pcre_fullinfo() in February 2000.
669
670 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
671 match a (*MARK), and the match failed at the start of the subject, a
672 reference to memory before the start of the subject could occur. This bug
673 was introduced by fix 17 of release 8.21.
674
675 5. A reference to an unset group with zero minimum repetition was giving
676 totally wrong answers (in non-JavaScript-compatibility mode). For example,
677 /(another)?(\1?)test/ matched against "hello world test". This bug was
678 introduced in release 8.13.
679
680 6. Add support for 16-bit character strings (a large amount of work involving
681 many changes and refactorings).
682
683 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
684 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
685 from a file.
686
687 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
688 rounding is not applied in this particular case).
689
690 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
691 if they appear, or are escaped, in patterns.
692
693 10. Get rid of a number of -Wunused-but-set-variable warnings.
694
695 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
696 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
697 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
698 also returns the mark "x". This bug applied to capturing parentheses,
699 non-capturing parentheses, and atomic parentheses. It also applied to some
700 assertions.
701
702 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
703 information out of configure.ac instead of relying on pcre.h.generic, which
704 is not stored in the repository.
705
706 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
707 -lreadline.
708
709 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
710
711 15. Applied Graycode's patch to put the top-level frame on the stack rather
712 than the heap when not using the stack for recursion. This gives a
713 performance improvement in many cases when recursion is not deep.
714
715 16. Experimental code added to "pcretest -C" to output the stack frame size.
716
717
718 Version 8.21 12-Dec-2011
719 ------------------------
720
721 1. Updating the JIT compiler.
722
723 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
724 are added as well.
725
726 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
727 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
728 calling _pcre_jit_exec. Some extra comments are added.
729
730 4. (*MARK) settings inside atomic groups that do not contain any capturing
731 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
732 was introduced by change 18 for 8.20.
733
734 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
735 ECMA-262 standard.
736
737 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
738 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
739 This bug was probably introduced by change 9 of 8.13.
740
741 7. While fixing 6 above, I noticed that a number of other items were being
742 incorrectly rejected as "not fixed length". This arose partly because newer
743 opcodes had not been added to the fixed-length checking code. I have (a)
744 corrected the bug and added tests for these items, and (b) arranged for an
745 error to occur if an unknown opcode is encountered while checking for fixed
746 length instead of just assuming "not fixed length". The items that were
747 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
748 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
749 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
750
751 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
752 being incorrectly compiled and would have given unpredicatble results.
753
754 9. A possessively repeated subpattern with minimum repeat count greater than
755 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
756 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
757 the first (A) could occur when it should not.
758
759 10. Add a cast and remove a redundant test from the code.
760
761 11. JIT should use pcre_malloc/pcre_free for allocation.
762
763 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
764 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
765 is anything other than /usr, -L is still shown).
766
767 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
768
769 14. Perl does not support \N without a following name in a [] class; PCRE now
770 also gives an error.
771
772 15. If a forward reference was repeated with an upper limit of around 2000,
773 it caused the error "internal error: overran compiling workspace". The
774 maximum number of forward references (including repeats) was limited by the
775 internal workspace, and dependent on the LINK_SIZE. The code has been
776 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
777 the default depends on LINK_SIZE. There is a new upper limit (for safety)
778 of around 200,000 forward references. While doing this, I also speeded up
779 the filling in of repeated forward references.
780
781 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
782 incorrectly expecting the subject to contain another "a" after the start.
783
784 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
785 in the match, the SKIP should be ignored. This was not happening; instead
786 the SKIP was being treated as NOMATCH. For patterns such as
787 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
788 tested.
789
790 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
791 now much more compatible with Perl, in particular in cases where the result
792 is a non-match for a non-anchored pattern. For example, if
793 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
794 "m", where previously it did not return a name. A side effect of this
795 change is that for partial matches, the last encountered mark name is
796 returned, as for non matches. A number of tests that were previously not
797 Perl-compatible have been moved into the Perl-compatible test files. The
798 refactoring has had the pleasing side effect of removing one argument from
799 the match() function, thus reducing its stack requirements.
800
801 19. If the /S+ option was used in pcretest to study a pattern using JIT,
802 subsequent uses of /S (without +) incorrectly behaved like /S+.
803
804 21. Retrieve executable code size support for the JIT compiler and fixing
805 some warnings.
806
807 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
808 not work when the shorter character appeared right at the end of the
809 subject string.
810
811 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
812 systems.
813
814 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
815 output it when the /M option is used in pcretest.
816
817 25. The CheckMan script was not being included in the distribution. Also, added
818 an explicit "perl" to run Perl scripts from the PrepareRelease script
819 because this is reportedly needed in Windows.
820
821 26. If study data was being save in a file and studying had not found a set of
822 "starts with" bytes for the pattern, the data written to the file (though
823 never used) was taken from uninitialized memory and so caused valgrind to
824 complain.
825
826 27. Updated RunTest.bat as provided by Sheri Pierce.
827
828 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
829
830 29. Computation of memory usage for the table of capturing group names was
831 giving an unnecessarily large value.
832
833
834 Version 8.20 21-Oct-2011
835 ------------------------
836
837 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
838 a POSIX class. After further experiments with Perl, which convinced me that
839 Perl has bugs and confusions, a closing square bracket is no longer allowed
840 in a POSIX name. This bug also affected patterns with classes that started
841 with full stops.
842
843 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
844 captured substring, but while checking the failing first alternative,
845 substring 1 is temporarily captured. If the output vector supplied to
846 pcre_exec() was not big enough for this capture, the yield of the function
847 was still zero ("insufficient space for captured substrings"). This cannot
848 be totally fixed without adding another stack variable, which seems a lot
849 of expense for a edge case. However, I have improved the situation in cases
850 such as /(a)(b)x|abc/ matched against "abc", where the return code
851 indicates that fewer than the maximum number of slots in the ovector have
852 been set.
853
854 3. Related to (2) above: when there are more back references in a pattern than
855 slots in the output vector, pcre_exec() uses temporary memory during
856 matching, and copies in the captures as far as possible afterwards. It was
857 using the entire output vector, but this conflicts with the specification
858 that only 2/3 is used for passing back captured substrings. Now it uses
859 only the first 2/3, for compatibility. This is, of course, another edge
860 case.
861
862 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
863 main code base, and can be used by building with --enable-jit. When this is
864 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
865 runtime --no-jit option is given.
866
867 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
868 ovector, the return from the function was zero, implying that there were
869 other matches that did not fit. The correct "exactly full" value is now
870 returned.
871
872 6. If a subpattern that was called recursively or as a subroutine contained
873 (*PRUNE) or any other control that caused it to give a non-standard return,
874 invalid errors such as "Error -26 (nested recursion at the same subject
875 position)" or even infinite loops could occur.
876
877 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
878 computing the minimum length on reaching *ACCEPT, and so ended up with the
879 wrong value of 1 rather than 0. Further investigation indicates that
880 computing a minimum subject length in the presence of *ACCEPT is difficult
881 (think back references, subroutine calls), and so I have changed the code
882 so that no minimum is registered for a pattern that contains *ACCEPT.
883
884 8. If (*THEN) was present in the first (true) branch of a conditional group,
885 it was not handled as intended. [But see 16 below.]
886
887 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
888 Sheri Pierce.
889
890 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
891 the first byte in a match must be "a".
892
893 11. Change 17 for 8.13 increased the recursion depth for patterns like
894 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
895 pattern contains any instances of (*THEN). If it does not, the old
896 optimizations are restored. It would be nice to do this on a per-group
897 basis, but at the moment that is not feasible.
898
899 12. In some environments, the output of pcretest -C is CRLF terminated. This
900 broke RunTest's code that checks for the link size. A single white space
901 character after the value is now allowed for.
902
903 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
904 For "fr", it uses the Windows-specific input and output files.
905
906 14. If (*THEN) appeared in a group that was called recursively or as a
907 subroutine, it did not work as intended. [But see next item.]
908
909 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
910 pattern fragments (but not containing any | characters). If A and B are
911 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
912 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
913 D. In other words, Perl considers parentheses that do not contain any |
914 characters to be part of a surrounding alternative, whereas PCRE was
915 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
916 differently. PCRE now behaves in the same way as Perl, except in the case
917 of subroutine/recursion calls such as (?1) which have in any case always
918 been different (but PCRE had them first :-).
919
920 16. Related to 15 above: Perl does not treat the | in a conditional group as
921 creating alternatives. Such a group is treated in the same way as an
922 ordinary group without any | characters when processing (*THEN). PCRE has
923 been changed to match Perl's behaviour.
924
925 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
926 RunGrepTest script failed.
927
928 18. Change 22 for version 13 caused atomic groups to use more stack. This is
929 inevitable for groups that contain captures, but it can lead to a lot of
930 stack use in large patterns. The old behaviour has been restored for atomic
931 groups that do not contain any capturing parentheses.
932
933 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
934 suppress the check for a minimum subject length at run time. (If it was
935 given to pcre_exec() or pcre_dfa_exec() it did work.)
936
937 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
938 fail to work when decoding hex characters in data strings in EBCDIC
939 environments.
940
941 21. It appears that in at least one Mac OS environment, the isxdigit() function
942 is implemented as a macro that evaluates to its argument more than once,
943 contravening the C 90 Standard (I haven't checked a later standard). There
944 was an instance in pcretest which caused it to go wrong when processing
945 \x{...} escapes in subject strings. The has been rewritten to avoid using
946 things like p++ in the argument of isxdigit().
947
948
949 Version 8.13 16-Aug-2011
950 ------------------------
951
952 1. The Unicode data tables have been updated to Unicode 6.0.0.
953
954 2. Two minor typos in pcre_internal.h have been fixed.
955
956 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
957 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
958 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
959
960 4. There were a number of related bugs in the code for matching backrefences
961 caselessly in UTF-8 mode when codes for the characters concerned were
962 different numbers of bytes. For example, U+023A and U+2C65 are an upper
963 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
964 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
965 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
966 2-byte code at the end of the subject (it thought there wasn't enough data
967 left).
968
969 5. Comprehensive information about what went wrong is now returned by
970 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
971 as the output vector has at least 2 elements. The offset of the start of
972 the failing character and a reason code are placed in the vector.
973
974 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
975 now returned is for the first byte of the failing character, instead of the
976 last byte inspected. This is an incompatible change, but I hope it is small
977 enough not to be a problem. It makes the returned offset consistent with
978 pcre_exec() and pcre_dfa_exec().
979
980 7. pcretest now gives a text phrase as well as the error number when
981 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
982 failure, the offset and reason code are output.
983
984 8. When \R was used with a maximizing quantifier it failed to skip backwards
985 over a \r\n pair if the subsequent match failed. Instead, it just skipped
986 back over a single character (\n). This seems wrong (because it treated the
987 two characters as a single entity when going forwards), conflicts with the
988 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
989 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
990 has been changed.
991
992 9. Some internal refactoring has changed the processing so that the handling
993 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
994 time (the PCRE_DOTALL option was changed this way some time ago: version
995 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
996 which was always a bit of a fudge. It also means that there is one less
997 argument for the match() function, which reduces its stack requirements
998 slightly. This change also fixes an incompatibility with Perl: the pattern
999 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1000
1001 10. More internal refactoring has drastically reduced the number of recursive
1002 calls to match() for possessively repeated groups such as (abc)++ when
1003 using pcre_exec().
1004
1005 11. While implementing 10, a number of bugs in the handling of groups were
1006 discovered and fixed:
1007
1008 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1009 (a|)*(?1) gave a compile-time internal error.
1010 ((a|)+)+ did not notice that the outer group could match an empty string.
1011 (^a|^)+ was not marked as anchored.
1012 (.*a|.*)+ was not marked as matching at start or after a newline.
1013
1014 12. Yet more internal refactoring has removed another argument from the match()
1015 function. Special calls to this function are now indicated by setting a
1016 value in a variable in the "match data" data block.
1017
1018 13. Be more explicit in pcre_study() instead of relying on "default" for
1019 opcodes that mean there is no starting character; this means that when new
1020 ones are added and accidentally left out of pcre_study(), testing should
1021 pick them up.
1022
1023 14. The -s option of pcretest has been documented for ages as being an old
1024 synonym of -m (show memory usage). I have changed it to mean "force study
1025 for every regex", that is, assume /S for every regex. This is similar to -i
1026 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1027 using it. It makes it easier to run collections of tests with and without
1028 study enabled, and thereby test pcre_study() more easily. All the standard
1029 tests are now run with and without -s (but some patterns can be marked as
1030 "never study" - see 20 below).
1031
1032 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1033 restoration of the capturing data to the outer values was not happening
1034 correctly.
1035
1036 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1037 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1038 pattern had matched an empty string, and so incorrectly returned a no
1039 match.
1040
1041 17. There was optimizing code for the last branch of non-capturing parentheses,
1042 and also for the obeyed branch of a conditional subexpression, which used
1043 tail recursion to cut down on stack usage. Unfortunately, now that there is
1044 the possibility of (*THEN) occurring in these branches, tail recursion is
1045 no longer possible because the return has to be checked for (*THEN). These
1046 two optimizations have therefore been removed. [But see 8.20/11 above.]
1047
1048 18. If a pattern containing \R was studied, it was assumed that \R always
1049 matched two bytes, thus causing the minimum subject length to be
1050 incorrectly computed because \R can also match just one byte.
1051
1052 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1053 was incorrectly computed.
1054
1055 20. If /S is present twice on a test pattern in pcretest input, it now
1056 *disables* studying, thereby overriding the use of -s on the command line
1057 (see 14 above). This is necessary for one or two tests to keep the output
1058 identical in both cases.
1059
1060 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1061 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1062
1063 22. When an atomic group that contained a capturing parenthesis was
1064 successfully matched, but the branch in which it appeared failed, the
1065 capturing was not being forgotten if a higher numbered group was later
1066 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1067 group 1 to "a", when in fact it should be unset. This applied to multi-
1068 branched capturing and non-capturing groups, repeated or not, and also to
1069 positive assertions (capturing in negative assertions does not happen
1070 in PCRE) and also to nested atomic groups.
1071
1072 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1073 subject after a captured substring, to make it easier to tell which of a
1074 number of identical substrings has been captured.
1075
1076 24. The way atomic groups are processed by pcre_exec() has been changed so that
1077 if they are repeated, backtracking one repetition now resets captured
1078 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1079 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1080 "aaa". Previously, it would have been "a". As part of this code
1081 refactoring, the way recursive calls are handled has also been changed.
1082
1083 25. If an assertion condition captured any substrings, they were not passed
1084 back unless some other capturing happened later. For example, if
1085 (?(?=(a))a) was matched against "a", no capturing was returned.
1086
1087 26. When studying a pattern that contained subroutine calls or assertions,
1088 the code for finding the minimum length of a possible match was handling
1089 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1090 group 1 called group 2 while simultaneously a separate group 2 called group
1091 1). A stack overflow occurred in this case. I have fixed this by limiting
1092 the recursion depth to 10.
1093
1094 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1095 Fortmann. This supports explicit test numbers on the command line, and has
1096 argument validation and error reporting.
1097
1098 28. An instance of \X with an unlimited repeat could fail if at any point the
1099 first character it looked at was a mark character.
1100
1101 29. Some minor code refactoring concerning Unicode properties and scripts
1102 should reduce the stack requirement of match() slightly.
1103
1104 30. Added the '=' option to pcretest to check the setting of unused capturing
1105 slots at the end of the pattern, which are documented as being -1, but are
1106 not included in the return count.
1107
1108 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1109 compiled something random. Now it gives a compile-time error (as does
1110 Perl).
1111
1112 32. A *MARK encountered during the processing of a positive assertion is now
1113 recorded and passed back (compatible with Perl).
1114
1115 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1116 had alternative anchored branches, the search for a second match in a line
1117 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1118 matched the line "0102" twice. The same bug affected patterns that started
1119 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1120 twice.
1121
1122 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1123 does, and because of capturing effects, quantifying parenthesized
1124 assertions may at times be useful. Quantifiers are now allowed for
1125 parenthesized assertions.
1126
1127 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1128
1129 36. \g was being checked for fancy things in a character class, when it should
1130 just be a literal "g".
1131
1132 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1133 appearance of a nested POSIX class supersedes an apparent external class.
1134 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1135 unescaped square brackets may also appear as part of class names. For
1136 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1137 more like Perl. (But see 8.20/1 above.)
1138
1139 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1140 was because it thought it was \N{name}, which is not supported).
1141
1142 39. Add minix to OS list not supporting the -S option in pcretest.
1143
1144 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1145 cannot analyze patterns in sufficient detail to catch mutual recursions
1146 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1147 subgroup is called recursively as a subpattern for a second time at the
1148 same position in the subject string. In previous releases this might have
1149 been caught by the recursion limit, or it might have run out of stack.
1150
1151 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1152 happen only once. PCRE was, however incorrectly giving a compile time error
1153 "recursive call could loop indefinitely" because it cannot analyze the
1154 pattern in sufficient detail. The compile time test no longer happens when
1155 PCRE is compiling a conditional subpattern, but actual runaway loops are
1156 now caught at runtime (see 40 above).
1157
1158 42. It seems that Perl allows any characters other than a closing parenthesis
1159 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1160 has been changed to be the same.
1161
1162 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1163 as not to get warnings when autogen.sh is called. Also changed
1164 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1165
1166 44. To help people who use pcregrep to scan files containing exceedingly long
1167 lines, the following changes have been made:
1168
1169 (a) The default value of the buffer size parameter has been increased from
1170 8K to 20K. (The actual buffer used is three times this size.)
1171
1172 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1173 PCRE is built.
1174
1175 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1176 to be set at run time.
1177
1178 (d) Numerical values in pcregrep options can be followed by K or M, for
1179 example --buffer-size=50K.
1180
1181 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1182 given and the return code is set to 2.
1183
1184 45. Add a pointer to the latest mark to the callout data block.
1185
1186 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1187 partial match of an empty string instead of no match. This was specific to
1188 the use of ".".
1189
1190 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1191 complete match instead of a partial match. This bug was dependent on both
1192 the PCRE_UTF8 and PCRE_DOTALL options being set.
1193
1194 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1195 starting byte set, because \b was not being ignored.
1196
1197
1198 Version 8.12 15-Jan-2011
1199 ------------------------
1200
1201 1. Fixed some typos in the markup of the man pages, and wrote a script that
1202 checks for such things as part of the documentation building process.
1203
1204 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1205 --match-limit and --recursion-limit options (added for 8.11). In
1206 particular, this made one of the standard tests fail. (The integer value
1207 went into the wrong half of a long int.)
1208
1209 3. If the --colour option was given to pcregrep with -v (invert match), it
1210 did strange things, either producing crazy output, or crashing. It should,
1211 of course, ignore a request for colour when reporting lines that do not
1212 match.
1213
1214 4. Another pcregrep bug caused similar problems if --colour was specified with
1215 -M (multiline) and the pattern match finished with a line ending.
1216
1217 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1218 matched in multiline mode, the following line was shown as part of the
1219 match. This seems wrong, so I have changed it.
1220
1221 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1222 the check for further matches in the same line (so they could be coloured)
1223 to overrun the end of the current line. If another match was found, it was
1224 incorrectly shown (and then shown again when found in the next line).
1225
1226 7. If pcregrep was compiled under Windows, there was a reference to the
1227 function pcregrep_exit() before it was defined. I am assuming this was
1228 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1229 reported by a user. I've moved the definition above the reference.
1230
1231
1232 Version 8.11 10-Dec-2010
1233 ------------------------
1234
1235 1. (*THEN) was not working properly if there were untried alternatives prior
1236 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1237 backtracked to try for "b" instead of moving to the next alternative branch
1238 at the same level (in this case, to look for "c"). The Perl documentation
1239 is clear that when (*THEN) is backtracked onto, it goes to the "next
1240 alternative in the innermost enclosing group".
1241
1242 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1243 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1244 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1245 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1246 (*THEN).
1247
1248 3. If \s appeared in a character class, it removed the VT character from
1249 the class, even if it had been included by some previous item, for example
1250 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1251 of \s, but is part of the POSIX "space" class.)
1252
1253 4. A partial match never returns an empty string (because you can always
1254 match an empty string at the end of the subject); however the checking for
1255 an empty string was starting at the "start of match" point. This has been
1256 changed to the "earliest inspected character" point, because the returned
1257 data for a partial match starts at this character. This means that, for
1258 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1259 (previously it gave "no match").
1260
1261 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1262 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1263 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1264 has an implication that the given string is incomplete (because a partial
1265 match is preferred over a full match). For this reason, these items now
1266 give a partial match in this situation. [Aside: previously, the one case
1267 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1268 match rather than a full match, which was wrong by the old rules, but is
1269 now correct.]
1270
1271 6. There was a bug in the handling of #-introduced comments, recognized when
1272 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1273 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1274 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1275 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1276 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1277 places in pcre_compile().
1278
1279 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1280 comments when looking ahead for named forward references to subpatterns,
1281 the only newline sequence it recognized was NL. It now handles newlines
1282 according to the set newline convention.
1283
1284 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1285 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1286 cater for a lack of strerror(). These oversights have been fixed.
1287
1288 9. Added --match-limit and --recursion-limit to pcregrep.
1289
1290 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1291
1292 11. When the -o option was used, pcregrep was setting a return code of 1, even
1293 when matches were found, and --line-buffered was not being honoured.
1294
1295 12. Added an optional parentheses number to the -o and --only-matching options
1296 of pcregrep.
1297
1298 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1299 can match an empty string. The code to do it in pcretest and pcredemo
1300 needed fixing:
1301
1302 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1303 only one byte after an empty string match just before CRLF (this case
1304 just got forgotten; "any" and "anycrlf" were OK).
1305
1306 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1307 mode when an empty string match preceded an ASCII character followed by
1308 a non-ASCII character. (The code for advancing by one character rather
1309 than one byte was nonsense.)
1310
1311 (c) The pcredemo.c sample program did not have any code at all to handle
1312 the cases when CRLF is a valid newline sequence.
1313
1314 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1315 as a starting offset was within the subject string. There is now a new
1316 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1317 negative or greater than the length of the string. In order to test this,
1318 pcretest is extended to allow the setting of negative starting offsets.
1319
1320 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1321 starting offset points to the beginning of a UTF-8 character was
1322 unnecessarily clumsy. I tidied it up.
1323
1324 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1325 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1326
1327 17. Nobody had reported that the --include_dir option, which was added in
1328 release 7.7 should have been called --include-dir (hyphen, not underscore)
1329 for compatibility with GNU grep. I have changed it to --include-dir, but
1330 left --include_dir as an undocumented synonym, and the same for
1331 --exclude-dir, though that is not available in GNU grep, at least as of
1332 release 2.5.4.
1333
1334 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1335 characters from a string of bytes) have been redefined so as not to use
1336 loops, in order to improve performance in some environments. At the same
1337 time, I abstracted some of the common code into auxiliary macros to save
1338 repetition (this should not affect the compiled code).
1339
1340 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1341 compile-time error is now given if \c is not followed by an ASCII
1342 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1343 different, and any byte value is allowed.)
1344
1345 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1346 START_OPTIMIZE option, which is now allowed at compile time - but just
1347 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1348 to pcregrep and other applications that have no direct access to PCRE
1349 options. The new /Y option in pcretest sets this option when calling
1350 pcre_compile().
1351
1352 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1353 back references. Groups containing recursive back references were forced to
1354 be atomic by that change, but in the case of named groups, the amount of
1355 memory required was incorrectly computed, leading to "Failed: internal
1356 error: code overflow". This has been fixed.
1357
1358 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1359 pcretest.c, to avoid build problems in some Borland environments.
1360
1361
1362 Version 8.10 25-Jun-2010
1363 ------------------------
1364
1365 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1366 THEN.
1367
1368 2. (*ACCEPT) was not working when inside an atomic group.
1369
1370 3. Inside a character class, \B is treated as a literal by default, but
1371 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1372 causes the error). The code is unchanged, but I tidied the documentation.
1373
1374 4. Inside a character class, PCRE always treated \R and \X as literals,
1375 whereas Perl faults them if its -w option is set. I have changed PCRE so
1376 that it faults them when PCRE_EXTRA is set.
1377
1378 5. Added support for \N, which always matches any character other than
1379 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1380
1381 6. When compiling pcregrep with newer versions of gcc which may have
1382 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1383 declared with attribute warn_unused_result" were given. Just casting the
1384 result to (void) does not stop the warnings; a more elaborate fudge is
1385 needed. I've used a macro to implement this.
1386
1387 7. Minor change to pcretest.c to avoid a compiler warning.
1388
1389 8. Added four artifical Unicode properties to help with an option to make
1390 \s etc use properties (see next item). The new properties are: Xan
1391 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1392
1393 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1394 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1395 this option. Modified pcretest to add /W to test this facility. Added
1396 REG_UCP to make it available via the POSIX interface.
1397
1398 10. Added --line-buffered to pcregrep.
1399
1400 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1401 studied, and the match started with a letter with a code point greater than
1402 127 whose first byte was different to the first byte of the other case of
1403 the letter, the other case of this starting letter was not recognized
1404 (#976).
1405
1406 12. If a pattern that was studied started with a repeated Unicode property
1407 test, for example, \p{Nd}+, there was the theoretical possibility of
1408 setting up an incorrect bitmap of starting bytes, but fortunately it could
1409 not have actually happened in practice until change 8 above was made (it
1410 added property types that matched character-matching opcodes).
1411
1412 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1413 possible starting bytes for non-anchored patterns.
1414
1415 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1416 \R, and also a number of cases that involve Unicode properties, both
1417 explicit and implicit when PCRE_UCP is set.
1418
1419 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1420 input, it could crash or give wrong results if characters with values
1421 greater than 0xc0 were present in the subject string. (Detail: it assumed
1422 UTF-8 input when processing these items.)
1423
1424 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1425 size_t is 64-bit (#991).
1426
1427 17. Added a check for running out of memory when PCRE is compiled with
1428 --disable-stack-for-recursion (#990).
1429
1430 18. If the last data line in a file for pcretest does not have a newline on
1431 the end, a newline was missing in the output.
1432
1433 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1434 less than 128) in its various bitmaps. However, there is a facility for
1435 generating tables according to the current locale when PCRE is compiled. It
1436 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1437 characters, are recognized by isspace() and therefore were getting set in
1438 these tables, and indeed these tables seem to approximate to ISO 8859. This
1439 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1440 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1441 which of course cannot start UTF-8 characters. I have changed the code so
1442 that only real ASCII characters (less than 128) and the correct starting
1443 bytes for UTF-8 encodings are set for characters greater than 127 when in
1444 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1445 altogether.)
1446
1447 20. Added the /T option to pcretest so as to be able to run tests with non-
1448 standard character tables, thus making it possible to include the tests
1449 used for 19 above in the standard set of tests.
1450
1451 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1452 reference to a subpattern the other side of a comment that contains an
1453 opening parenthesis caused either an internal compiling error, or a
1454 reference to the wrong subpattern.
1455
1456
1457 Version 8.02 19-Mar-2010
1458 ------------------------
1459
1460 1. The Unicode data tables have been updated to Unicode 5.2.0.
1461
1462 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1463 configured.
1464
1465 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1466 original author of that file, following a query about its status.
1467
1468 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1469 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1470
1471 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1472 quantifier applied to a forward-referencing subroutine call, could compile
1473 incorrect code or give the error "internal error: previously-checked
1474 referenced subpattern not found".
1475
1476 6. Both MS Visual Studio and Symbian OS have problems with initializing
1477 variables to point to external functions. For these systems, therefore,
1478 pcre_malloc etc. are now initialized to local functions that call the
1479 relevant global functions.
1480
1481 7. There were two entries missing in the vectors called coptable and poptable
1482 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1483 I've fixed the data, and added a kludgy way of testing at compile time that
1484 the lengths are correct (equal to the number of opcodes).
1485
1486 8. Following on from 7, I added a similar kludge to check the length of the
1487 eint vector in pcreposix.c.
1488
1489 9. Error texts for pcre_compile() are held as one long string to avoid too
1490 much relocation at load time. To find a text, the string is searched,
1491 counting zeros. There was no check for running off the end of the string,
1492 which could happen if a new error number was added without updating the
1493 string.
1494
1495 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1496
1497 11. \K was not working if it appeared in an atomic group or in a group that
1498 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1499 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1500 the assertion is positive, but not if it is negative.
1501
1502 12. Change 11 fortuitously reduced the size of the stack frame used in the
1503 "match()" function of pcre_exec.c by one pointer. Forthcoming
1504 implementation of support for (*MARK) will need an extra pointer on the
1505 stack; I have reserved it now, so that the stack frame size does not
1506 decrease.
1507
1508 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1509 item in branch that calls a recursion is a subroutine call - as in the
1510 second branch in the above example - was incorrectly given the compile-
1511 time error "recursive call could loop indefinitely" because pcre_compile()
1512 was not correctly checking the subroutine for matching a non-empty string.
1513
1514 14. The checks for overrunning compiling workspace could trigger after an
1515 overrun had occurred. This is a "should never occur" error, but it can be
1516 triggered by pathological patterns such as hundreds of nested parentheses.
1517 The checks now trigger 100 bytes before the end of the workspace.
1518
1519 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1520
1521
1522 Version 8.01 19-Jan-2010
1523 ------------------------
1524
1525 1. If a pattern contained a conditional subpattern with only one branch (in
1526 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1527 computed the wrong minimum data length (which is of course zero for such
1528 subpatterns). This could cause incorrect "no match" results.
1529
1530 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1531 the pattern is reset in the first branch, pcre_compile() failed with
1532 "internal error: code overflow at offset...". This happened only when
1533 the reset was to the original external option setting. (An optimization
1534 abstracts leading options settings into an external setting, which was the
1535 cause of this.)
1536
1537 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1538 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1539 assertion pattern did not match (meaning that the assertion was true), it
1540 was incorrectly treated as false if the SKIP had been reached during the
1541 matching. This also applied to assertions used as conditions.
1542
1543 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1544 assertion subpattern, including such a pattern used as a condition,
1545 unpredictable results occurred, instead of the error return
1546 PCRE_ERROR_DFA_UITEM.
1547
1548 5. The C++ GlobalReplace function was not working like Perl for the special
1549 situation when an empty string is matched. It now does the fancy magic
1550 stuff that is necessary.
1551
1552 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1553 removed. (These were left over from very, very early versions of PCRE.)
1554
1555 7. Some cosmetic changes to the code to make life easier when compiling it
1556 as part of something else:
1557
1558 (a) Change DEBUG to PCRE_DEBUG.
1559
1560 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1561 called "current" as "current_branch", to prevent a collision with the
1562 Linux macro when compiled as a kernel module.
1563
1564 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1565 prevent a collision with the Linux macro when compiled as a kernel
1566 module.
1567
1568 8. In pcre_compile() there are some checks for integer overflows that used to
1569 cast potentially large values to (double). This has been changed to that
1570 when building, a check for int64_t is made, and if it is found, it is used
1571 instead, thus avoiding the use of floating point arithmetic. (There is no
1572 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1573 double.
1574
1575 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1576 2005 (difference between two addresses compared to an unsigned value).
1577
1578 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1579 custom one, because of the following reported problem in Windows:
1580
1581 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1582 under Win32.
1583 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1584 therefore missing the function definition.
1585 - The compiler thus generates a "C" signature for the test function.
1586 - The linker fails to find the "C" function.
1587 - PCRE fails to configure if asked to do so against libbz2.
1588
1589 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1590 messages were output:
1591
1592 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1593 rerunning libtoolize, to keep the correct libtool macros in-tree.
1594 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1595
1596 I have done both of these things.
1597
1598 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1599 most of the time, it *can* run out if it is given a pattern that contains a
1600 runaway infinite recursion. I updated the discussion in the pcrestack man
1601 page.
1602
1603 13. Now that we have gone to the x.xx style of version numbers, the minor
1604 version may start with zero. Using 08 or 09 is a bad idea because users
1605 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1606 interpreted as invalid octal numbers. I've updated the previous comment in
1607 configure.ac, and also added a check that gives an error if 08 or 09 are
1608 used.
1609
1610 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1611 causing partial matching to fail when the end of the subject matched \W
1612 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1613
1614 15. There were some discrepancies between the declarations in pcre_internal.h
1615 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1616 their definitions. The declarations used "const uschar *" and the
1617 definitions used USPTR. Even though USPTR is normally defined as "const
1618 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1619 reported that: "This difference in casting confuses some C++ compilers, for
1620 example, SunCC recognizes above declarations as different functions and
1621 generates broken code for hbpcre." I have changed the declarations to use
1622 USPTR.
1623
1624 16. GNU libtool is named differently on some systems. The autogen.sh script now
1625 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1626 (FreeBSD).
1627
1628 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1629 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1630 comment: "Figure out how to create a longlong from a string: strtoll and
1631 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1632 instance, but it only takes 2 args instead of 3!"
1633
1634 18. A subtle bug concerned with back references has been fixed by a change of
1635 specification, with a corresponding code fix. A pattern such as
1636 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1637 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1638 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1639 same bug. Such groups have to be quantified to be useful, or contained
1640 inside another quantified group. (If there's no repetition, the reference
1641 can never match.) The problem arises because, having left the group and
1642 moved on to the rest of the pattern, a later failure that backtracks into
1643 the group uses the captured value from the final iteration of the group
1644 rather than the correct earlier one. I have fixed this in PCRE by forcing
1645 any group that contains a reference to itself to be an atomic group; that
1646 is, there cannot be any backtracking into it once it has completed. This is
1647 similar to recursive and subroutine calls.
1648
1649
1650 Version 8.00 19-Oct-09
1651 ----------------------
1652
1653 1. The table for translating pcre_compile() error codes into POSIX error codes
1654 was out-of-date, and there was no check on the pcre_compile() error code
1655 being within the table. This could lead to an OK return being given in
1656 error.
1657
1658 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1659 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1660 in a Windows environment.
1661
1662 3. The pcregrep --count option prints the count for each file even when it is
1663 zero, as does GNU grep. However, pcregrep was also printing all files when
1664 --files-with-matches was added. Now, when both options are given, it prints
1665 counts only for those files that have at least one match. (GNU grep just
1666 prints the file name in this circumstance, but including the count seems
1667 more useful - otherwise, why use --count?) Also ensured that the
1668 combination -clh just lists non-zero counts, with no names.
1669
1670 4. The long form of the pcregrep -F option was incorrectly implemented as
1671 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1672 but it seems right to fix it, and I didn't think it was worth preserving
1673 the old behaviour.
1674
1675 5. The command line items --regex=pattern and --regexp=pattern were not
1676 recognized by pcregrep, which required --regex pattern or --regexp pattern
1677 (with a space rather than an '='). The man page documented the '=' forms,
1678 which are compatible with GNU grep; these now work.
1679
1680 6. No libpcreposix.pc file was created for pkg-config; there was just
1681 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1682
1683 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1684 when UCP support is not needed, by modifying the Python script that
1685 generates it from Unicode data files. This should not matter if the module
1686 is correctly used as a library, but I received one complaint about 50K of
1687 unwanted data. My guess is that the person linked everything into his
1688 program rather than using a library. Anyway, it does no harm.
1689
1690 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1691 was a minimum greater than 1 for a wide character in a possessive
1692 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1693 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1694 character. Chaos in the form of incorrect output or a compiling loop could
1695 result.
1696
1697 9. The restrictions on what a pattern can contain when partial matching is
1698 requested for pcre_exec() have been removed. All patterns can now be
1699 partially matched by this function. In addition, if there are at least two
1700 slots in the offset vector, the offset of the earliest inspected character
1701 for the match and the offset of the end of the subject are set in them when
1702 PCRE_ERROR_PARTIAL is returned.
1703
1704 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1705 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1706 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1707 and may be more useful for multi-segment matching.
1708
1709 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1710 used to be given if ever the end of the subject was reached; now it is
1711 given only if matching could not proceed because another character was
1712 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1713 string "Z", which now yields "no match" instead of "partial match". In the
1714 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1715 final character ended with (*FAIL).
1716
1717 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1718 if the pattern had a "must contain" character that was already found in the
1719 earlier partial match, unless partial matching was again requested. For
1720 example, with the pattern /dog.(body)?/, the "must contain" character is
1721 "g". If the first part-match was for the string "dog", restarting with
1722 "sbody" failed. This bug has been fixed.
1723
1724 13. The string returned by pcre_dfa_exec() after a partial match has been
1725 changed so that it starts at the first inspected character rather than the
1726 first character of the match. This makes a difference only if the pattern
1727 starts with a lookbehind assertion or \b or \B (\K is not supported by
1728 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1729 matching functions compatible, and I think it's the right thing to do.
1730
1731 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1732 so that the demonstration program is easily available in environments where
1733 PCRE has not been installed from source.
1734
1735 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1736 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1737 library.
1738
1739 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1740 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1741 is not the first non-POSIX option to be added. Clearly some people find
1742 these options useful.
1743
1744 17. If a caller to the POSIX matching function regexec() passes a non-zero
1745 value for nmatch with a NULL value for pmatch, the value of
1746 nmatch is forced to zero.
1747
1748 18. RunGrepTest did not have a test for the availability of the -u option of
1749 the diff command, as RunTest does. It now checks in the same way as
1750 RunTest, and also checks for the -b option.
1751
1752 19. If an odd number of negated classes containing just a single character
1753 interposed, within parentheses, between a forward reference to a named
1754 subpattern and the definition of the subpattern, compilation crashed with
1755 an internal error, complaining that it could not find the referenced
1756 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1757 [The bug was that it was starting one character too far in when skipping
1758 over the character class, thus treating the ] as data rather than
1759 terminating the class. This meant it could skip too much.]
1760
1761 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1762 /g option in pcretest when the pattern contains \K, which makes it possible
1763 to have an empty string match not at the start, even when the pattern is
1764 anchored. Updated pcretest and pcredemo to use this option.
1765
1766 21. If the maximum number of capturing subpatterns in a recursion was greater
1767 than the maximum at the outer level, the higher number was returned, but
1768 with unset values at the outer level. The correct (outer level) value is
1769 now given.
1770
1771 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1772 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1773 make it do so. The string so far is captured, making this feature
1774 compatible with Perl.
1775
1776 23. The tests have been re-organized, adding tests 11 and 12, to make it
1777 possible to check the Perl 5.10 features against Perl 5.10.
1778
1779 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1780 pattern matches a fixed length string. PCRE did not allow this; now it
1781 does. Neither allows recursion.
1782
1783 25. I finally figured out how to implement a request to provide the minimum
1784 length of subject string that was needed in order to match a given pattern.
1785 (It was back references and recursion that I had previously got hung up
1786 on.) This code has now been added to pcre_study(); it finds a lower bound
1787 to the length of subject needed. It is not necessarily the greatest lower
1788 bound, but using it to avoid searching strings that are too short does give
1789 some useful speed-ups. The value is available to calling programs via
1790 pcre_fullinfo().
1791
1792 26. While implementing 25, I discovered to my embarrassment that pcretest had
1793 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1794 study optimizations had never been tested with that matching function.
1795 Oops. What is worse, even when it was passed study data, there was a bug in
1796 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1797 were also very few tests of studied patterns with pcre_dfa_exec().
1798
1799 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1800 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1801 on the other side of the coin, they are no longer allowed to have different
1802 names, because these cannot be distinguished in PCRE, and this has caused
1803 confusion. (This is a difference from Perl.)
1804
1805 28. When duplicate subpattern names are present (necessarily with different
1806 numbers, as required by 27 above), and a test is made by name in a
1807 conditional pattern, either for a subpattern having been matched, or for
1808 recursion in such a pattern, all the associated numbered subpatterns are
1809 tested, and the overall condition is true if the condition is true for any
1810 one of them. This is the way Perl works, and is also more like the way
1811 testing by number works.
1812
1813
1814 Version 7.9 11-Apr-09
1815 ---------------------
1816
1817 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1818 (pcretest), all targets were linked against these libraries. This included
1819 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1820 libraries. This caused unwanted dependencies to be created. This problem
1821 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1822 pcretest is linked with readline.
1823
1824 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1825 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1826 moved outside it again, because FALSE and TRUE are already defined in AIX,
1827 but BOOL is not.
1828
1829 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1830 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1831
1832 4. The pcregrep documentation said spaces were inserted as well as colons (or
1833 hyphens) following file names and line numbers when outputting matching
1834 lines. This is not true; no spaces are inserted. I have also clarified the
1835 wording for the --colour (or --color) option.
1836
1837 5. In pcregrep, when --colour was used with -o, the list of matching strings
1838 was not coloured; this is different to GNU grep, so I have changed it to be
1839 the same.
1840
1841 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1842 each matching line was coloured. Now it goes on to look for further matches
1843 of any of the test patterns, which is the same behaviour as GNU grep.
1844
1845 7. A pattern that could match an empty string could cause pcregrep to loop; it
1846 doesn't make sense to accept an empty string match in pcregrep, so I have
1847 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1848 seems to be how GNU grep behaves. [But see later change 40 for release
1849 8.33.]
1850
1851 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1852 start or after a newline", because the conditional assertion was not being
1853 correctly handled. The rule now is that both the assertion and what follows
1854 in the first alternative must satisfy the test.
1855
1856 9. If auto-callout was enabled in a pattern with a conditional group whose
1857 condition was an assertion, PCRE could crash during matching, both with
1858 pcre_exec() and pcre_dfa_exec().
1859
1860 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1861 used for matching.
1862
1863 11. Unicode property support in character classes was not working for
1864 characters (bytes) greater than 127 when not in UTF-8 mode.
1865
1866 12. Added the -M command line option to pcretest.
1867
1868 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1869
1870 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1871
1872 16. Added comments and documentation about mis-use of no_arg in the C++
1873 wrapper.
1874
1875 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1876 from Martin Jerabek that uses macro names for all relevant character and
1877 string constants.
1878
1879 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1880 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1881 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1882 these, but not everybody uses configure.
1883
1884 19. A conditional group that had only one branch was not being correctly
1885 recognized as an item that could match an empty string. This meant that an
1886 enclosing group might also not be so recognized, causing infinite looping
1887 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1888 with the subject "ab", where knowledge that the repeated group can match
1889 nothing is needed in order to break the loop.
1890
1891 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1892 exec(), but without supplying a callout function, matching went wrong.
1893
1894 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1895 leak if the size of the offset vector was greater than 30. When the vector
1896 is smaller, the saved offsets during recursion go onto a local stack
1897 vector, but for larger vectors malloc() is used. It was failing to free
1898 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1899 error, in fact).
1900
1901 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1902 heapframe that is used only when UTF-8 support is enabled. This caused no
1903 problem, but was untidy.
1904
1905 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1906 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1907 included within another project.
1908
1909 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1910 slightly modified by me:
1911
1912 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1913 not building pcregrep.
1914
1915 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1916 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1917
1918 25. Forward references, both numeric and by name, in patterns that made use of
1919 duplicate group numbers, could behave incorrectly or give incorrect errors,
1920 because when scanning forward to find the reference group, PCRE was not
1921 taking into account the duplicate group numbers. A pattern such as
1922 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1923
1924 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1925 the feature of a custom pointer more persuasive (as requested by a user).
1926
1927 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1928 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1929 pre-defined.
1930
1931 28. Added support for (*UTF8) at the start of a pattern.
1932
1933 29. Arrange for flags added by the "release type" setting in CMake to be shown
1934 in the configuration summary.
1935
1936
1937 Version 7.8 05-Sep-08
1938 ---------------------
1939
1940 1. Replaced UCP searching code with optimized version as implemented for Ad
1941 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1942 stage table and inline lookup instead of a function, giving speed ups of 2
1943 to 5 times on some simple patterns that I tested. Permission was given to
1944 distribute the MultiStage2.py script that generates the tables (it's not in
1945 the tarball, but is in the Subversion repository).
1946
1947 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1948 scripts.
1949
1950 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1951 a group with a zero qualifier. The result of the study could be incorrect,
1952 or the function might crash, depending on the pattern.
1953
1954 4. Caseless matching was not working for non-ASCII characters in back
1955 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1956 It now works when Unicode Property Support is available.
1957
1958 5. In pcretest, an escape such as \x{de} in the data was always generating
1959 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1960 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1961 truncation.
1962
1963 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1964
1965 7. Added two (int) casts to pcregrep when printing the difference of two
1966 pointers, in case they are 64-bit values.
1967
1968 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1969 test 2 if it fails.
1970
1971 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1972 and a #define of that name to empty if it is not externally set. This is to
1973 allow users of MSVC to set it if necessary.
1974
1975 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1976 the convenience functions in the pcre_get.c source file.
1977
1978 11. An option change at the start of a pattern that had top-level alternatives
1979 could cause overwriting and/or a crash. This command provoked a crash in
1980 some environments:
1981
1982 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1983
1984 This potential security problem was recorded as CVE-2008-2371.
1985
1986 12. For a pattern where the match had to start at the beginning or immediately
1987 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1988 pcre_dfa_exec() could read past the end of the passed subject if there was
1989 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1990 pcretest so that it places the subject at the end of its malloc-ed buffer.
1991
1992 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1993 exec() might read past the end of the data buffer in UTF-8 mode.
1994
1995 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1996 the data contained the byte 0x85 as part of a UTF-8 character within its
1997 first line. This applied both to normal and DFA matching.
1998
1999 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2000 /^[^d]*?$/8 failed to match "abc".
2001
2002 16. Added a missing copyright notice to pcrecpp_internal.h.
2003
2004 17. Make it more clear in the documentation that values returned from
2005 pcre_exec() in ovector are byte offsets, not character counts.
2006
2007 18. Tidied a few places to stop certain compilers from issuing warnings.
2008
2009 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2010 supplied by Stefan Weber. I made a further small update for 7.8 because
2011 there is a change of source arrangements: the pcre_searchfuncs.c module is
2012 replaced by pcre_ucd.c.
2013
2014
2015 Version 7.7 07-May-08
2016 ---------------------
2017
2018 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2019 a string to a long long, pretend we don't even have a long long." This is
2020 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2021
2022 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2023 pre-7.6 versions, which defined a global no_arg variable instead of putting
2024 it in the RE class. (See also #8 below.)
2025
2026 3. Remove a line of dead code, identified by coverity and reported by Nuno
2027 Lopes.
2028
2029 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2030
2031 (1) The include/exclude patterns were being applied to the whole pathnames
2032 of files, instead of just to the final components.
2033
2034 (2) If there was more than one level of directory, the subdirectories were
2035 skipped unless they satisfied the include/exclude conditions. This is
2036 inconsistent with GNU grep (and could even be seen as contrary to the
2037 pcregrep specification - which I improved to make it absolutely clear).
2038 The action now is always to scan all levels of directory, and just
2039 apply the include/exclude patterns to regular files.
2040
2041 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2042 --exclude_dir in the tests to avoid scanning .svn directories.
2043
2044 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2045 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2046 doesn't support NULs in patterns.
2047
2048 7. Added some missing "const"s to declarations of static tables in
2049 pcre_compile.c and pcre_dfa_exec.c.
2050
2051 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2052 caused by fix #2 above. (Subsequently also a second patch to fix the
2053 first patch. And a third patch - this was a messy problem.)
2054
2055 9. Applied Craig's patch to remove the use of push_back().
2056
2057 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2058 matching function regexec().
2059
2060 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2061 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2062 references. PCRE supports relative numbers with this syntax (I don't think
2063 Oniguruma does).
2064
2065 12. Previously, a group with a zero repeat such as (...){0} was completely
2066 omitted from the compiled regex. However, this means that if the group
2067 was called as a subroutine from elsewhere in the pattern, things went wrong
2068 (an internal error was given). Such groups are now left in the compiled
2069 pattern, with a new opcode that causes them to be skipped at execution
2070 time.
2071
2072 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2073 to the way PCRE behaves:
2074
2075 (a) A lone ] character is dis-allowed (Perl treats it as data).
2076
2077 (b) A back reference to an unmatched subpattern matches an empty string
2078 (Perl fails the current match path).
2079
2080 (c) A data ] in a character class must be notated as \] because if the
2081 first data character in a class is ], it defines an empty class. (In
2082 Perl it is not possible to have an empty class.) The empty class []
2083 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2084 The negative empty class [^] matches any one character, independently
2085 of the DOTALL setting.
2086
2087 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2088 non-existent subpattern following a character class starting with ']' and
2089 containing () gave an internal compiling error instead of "reference to
2090 non-existent subpattern". Fortunately, when the pattern did exist, the
2091 compiled code was correct. (When scanning forwards to check for the
2092 existence of the subpattern, it was treating the data ']' as terminating
2093 the class, so got the count wrong. When actually compiling, the reference
2094 was subsequently set up correctly.)
2095
2096 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2097 it was being rejected as not supported by pcre_dfa_exec(), even though
2098 other assertions are supported. I have made pcre_dfa_exec() support
2099 (*FAIL).
2100
2101 16. The implementation of 13c above involved the invention of a new opcode,
2102 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2103 cannot be changed at match time, I realized I could make a small
2104 improvement to matching performance by compiling OP_ALLANY instead of
2105 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2106 on the OP_ANY path.
2107
2108 17. Compiling pcretest on Windows with readline support failed without the
2109 following two fixes: (1) Make the unistd.h include conditional on
2110 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2111
2112 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2113 ncurses library to be included for pcretest when ReadLine support is
2114 requested, but also to allow for it to be overridden. This patch came from
2115 Daniel Bergström.
2116
2117 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2118 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2119 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2120 spotting this.
2121
2122
2123 Version 7.6 28-Jan-08
2124 ---------------------
2125
2126 1. A character class containing a very large number of characters with
2127 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2128 overflow.
2129
2130 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2131 HAVE_LONG_LONG is not defined.
2132
2133 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2134 bring them up to date and include new features. This patch includes:
2135
2136 - Fixed PH's badly added libz and libbz2 support.
2137 - Fixed a problem with static linking.
2138 - Added pcredemo. [But later removed - see 7 below.]
2139 - Fixed dftables problem and added an option.
2140 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2141 HAVE_LONG_LONG.
2142 - Added readline support for pcretest.
2143 - Added an listing of the option settings after cmake has run.
2144
2145 4. A user submitted a patch to Makefile that makes it easy to create
2146 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2147 Makefile.am that cause it to include this special target, without
2148 affecting anything else. Note that the same mingw target plus all
2149 the other distribution libraries and programs are now supported
2150 when configuring with CMake (see 6 below) instead of with
2151 Configure/Make.
2152
2153 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2154 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2155 exported in the Windows port". It has not yet been confirmed that the patch
2156 solves the problem, but it does no harm.
2157
2158 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2159 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2160 with CMake, and also correct the comment about stack recursion.
2161
2162 7. Remove the automatic building of pcredemo from the ./configure system and
2163 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2164 of a program that users should build themselves after PCRE is installed, so
2165 building it automatically is not really right. What is more, it gave
2166 trouble in some build environments.
2167
2168 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2169
2170
2171 Version 7.5 10-Jan-08
2172 ---------------------
2173
2174 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2175 values in parens when parsing an RE using the C++ wrapper."
2176
2177 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2178 Characters greater than 255 were excluded from the class instead of being
2179 included.
2180
2181 3. The same bug as (2) above applied to negated POSIX classes such as
2182 [:^space:].
2183
2184 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2185 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2186 I have changed it.
2187
2188 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2189 first named subpattern) and a construct such as (?&a) would reference the
2190 first named subpattern whose name started with "a" (in other words, the
2191 length check was missing). Both these problems are fixed. "Subpattern name
2192 expected" is now given for (?&) (a zero-length name), and this patch also
2193 makes it give the same error for \k'' (previously it complained that that
2194 was a reference to a non-existent subpattern).
2195
2196 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2197 this is right because (?- can be followed by option settings as well as by
2198 digits. I have, however, made the messages clearer.
2199
2200 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2201 than the number used in the conditional) now cause a compile-time error.
2202 This is actually not compatible with Perl, which accepts such patterns, but
2203 treats the conditional as always being FALSE (as PCRE used to), but it
2204 seems to me that giving a diagnostic is better.
2205
2206 8. Change "alphameric" to the more common word "alphanumeric" in comments
2207 and messages.
2208
2209 9. Fix two occurrences of "backslash" in comments that should have been
2210 "backspace".
2211
2212 10. Remove two redundant lines of code that can never be obeyed (their function
2213 was moved elsewhere).
2214
2215 11. The program that makes PCRE's Unicode character property table had a bug
2216 which caused it to generate incorrect table entries for sequences of
2217 characters that have the same character type, but are in different scripts.
2218 It amalgamated them into a single range, with the script of the first of
2219 them. In other words, some characters were in the wrong script. There were
2220 thirteen such cases, affecting characters in the following ranges:
2221
2222 U+002b0 - U+002c1
2223 U+0060c - U+0060d
2224 U+0061e - U+00612
2225 U+0064b - U+0065e
2226 U+0074d - U+0076d
2227 U+01800 - U+01805
2228 U+01d00 - U+01d77
2229 U+01d9b - U+01dbf
2230 U+0200b - U+0200f
2231 U+030fc - U+030fe
2232 U+03260 - U+0327f
2233 U+0fb46 - U+0fbb1
2234 U+10450 - U+1049d
2235
2236 12. The -o option (show only the matching part of a line) for pcregrep was not
2237 compatible with GNU grep in that, if there was more than one match in a
2238 line, it showed only the first of them. It now behaves in the same way as
2239 GNU grep.
2240
2241 13. If the -o and -v options were combined for pcregrep, it printed a blank
2242 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2243 does the same. The return code can be used to tell if there were any
2244 non-matching lines.
2245
2246 14. Added --file-offsets and --line-offsets to pcregrep.
2247
2248 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2249 infinitely looping recursion. The bug was that positive lookaheads were not
2250 being skipped when checking for a possible empty match (negative lookaheads
2251 and both kinds of lookbehind were skipped).
2252
2253 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2254 inclusion of <windows.h> to before rather than after the definition of
2255 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2256
2257 17. Specifying a possessive quantifier with a specific limit for a Unicode
2258 character property caused pcre_compile() to compile bad code, which led at
2259 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2260 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2261 caused the error; without that there was no problem.
2262
2263 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2264
2265 19. Added --enable-pcretest-libreadline.
2266
2267 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2268 RE::GlobalReplace(). As a result, the number of replacements returned was
2269 double what it should be. I removed one of the increments, but Craig sent a
2270 later patch that removed the other one (the right fix) and added unit tests
2271 that check the return values (which was not done before).
2272
2273 21. Several CMake things:
2274
2275 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2276 the names libpcre and libpcreposix, not just pcre and pcreposix.
2277
2278 (2) The above change means that pcretest and pcregrep are now correctly
2279 linked with the newly-built libraries, not previously installed ones.
2280
2281 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2282
2283 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2284 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2285 UTF-8 newline character). The key issue is that the pattern starts .*;
2286 this means that the match must be either at the beginning, or after a
2287 newline. The bug was in the code for advancing after a failed match and
2288 checking that the new position followed a newline. It was not taking
2289 account of UTF-8 characters correctly.
2290
2291 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2292 character classes. PCRE was not treating the sequence [:...:] as a
2293 character class unless the ... were all letters. Perl, however, seems to
2294 allow any characters between [: and :], though of course it rejects as
2295 unknown any "names" that contain non-letters, because all the known class
2296 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2297 for example, whereas PCRE did not - it did not recognize a POSIX character
2298 class. This seemed a bit dangerous, so the code has been changed to be
2299 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2300 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2301 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2302 Perl does, and where it didn't before.
2303
2304 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2305 Windows environments %n is disabled by default.
2306
2307
2308 Version 7.4 21-Sep-07
2309 ---------------------
2310
2311 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2312 means that a class such as [\s] counted as "explicit reference to CR or
2313 LF". That isn't really right - the whole point of the change was to try to
2314 help when there was an actual mention of one of the two characters. So now
2315 the change happens only if \r or \n (or a literal CR or LF) character is
2316 encountered.
2317
2318 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2319 of both had grown to the point where there were only 3 bits left.
2320 Fortunately, there was spare space in the data structure, and so I have
2321 moved the internal flags into a new 16-bit field to free up more option
2322 bits.
2323
2324 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2325 but did not set the internal JCHANGED flag - either of these is enough to
2326 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2327 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2328 start sets both bits.
2329
2330 4. Added options (at build time, compile time, exec time) to change \R from
2331 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2332
2333 5. doc/pcresyntax.html was missing from the distribution.
2334
2335 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2336 compatibility, even though it is no longer used.
2337
2338 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2339 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2340 windows.h file is present (where different names are used). [This was
2341 reversed later after testing - see 16 below.]
2342
2343 8. Changed all #include <config.h> to #include "config.h". There were also
2344 some further <pcre.h> cases that I changed to "pcre.h".
2345
2346 9. When pcregrep was used with the --colour option, it missed the line ending
2347 sequence off the lines that it output.
2348
2349 10. It was pointed out to me that arrays of string pointers cause lots of
2350 relocations when a shared library is dynamically loaded. A technique of
2351 using a single long string with a table of offsets can drastically reduce
2352 these. I have refactored PCRE in four places to do this. The result is
2353 dramatic:
2354
2355 Originally: 290
2356 After changing UCP table: 187
2357 After changing error message table: 43
2358 After changing table of "verbs" 36
2359 After changing table of Posix names 22
2360
2361 Thanks to the folks working on Gregex for glib for this insight.
2362
2363 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2364 unicode-properties was also set.
2365
2366 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2367
2368 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2369 checked only for CRLF.
2370
2371 14. Added casts to pcretest.c to avoid compiler warnings.
2372
2373 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2374
2375 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2376 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2377 entirely. This removes changes made in 7 above.
2378
2379 17. The CMake files have been updated, and there is now more information about
2380 building with CMake in the NON-UNIX-USE document.
2381
2382
2383 Version 7.3 28-Aug-07
2384 ---------------------
2385
2386 1. In the rejigging of the build system that eventually resulted in 7.1, the
2387 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2388 brackets there is not right, since it causes compilers to look for an
2389 installed pcre.h, not the version that is in the source that is being
2390 compiled (which of course may be different). I have changed it back to:
2391
2392 #include "pcre.h"
2393
2394 I have a vague recollection that the change was concerned with compiling in
2395 different directories, but in the new build system, that is taken care of
2396 by the VPATH setting the Makefile.
2397
2398 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2399 when the subject happened to end in the byte 0x85 (e.g. if the last
2400 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2401 characters but of course it shouldn't be taken as a newline when it is part
2402 of another character. The bug was that, for an unlimited repeat of . in
2403 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2404 characters when looking for a newline.
2405
2406 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2407
2408 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2409 in debug output.
2410
2411 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2412 long printing in the pcrecpp unittest when running under MinGW.
2413
2414 6. ESC_K was left out of the EBCDIC table.
2415
2416 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2417 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2418 limit also applies to "virtual nesting" when a pattern is recursive, and in
2419 this case 1000 isn't so big. I have been able to remove this limit at the
2420 expense of backing off one optimization in certain circumstances. Normally,
2421 when pcre_exec() would call its internal match() function recursively and
2422 immediately return the result unconditionally, it uses a "tail recursion"
2423 feature to save stack. However, when a subpattern that can match an empty
2424 string has an unlimited repetition quantifier, it no longer makes this
2425 optimization. That gives it a stack frame in which to save the data for
2426 checking that an empty string has been matched. Previously this was taken
2427 from the 1000-entry workspace that had been reserved. So now there is no
2428 explicit limit, but more stack is used.
2429
2430 8. Applied Daniel's patches to solve problems with the import/export magic
2431 syntax that is required for Windows, and which was going wrong for the
2432 pcreposix and pcrecpp parts of the library. These were overlooked when this
2433 problem was solved for the main library.
2434
2435 9. There were some crude static tests to avoid integer overflow when computing
2436 the size of patterns that contain repeated groups with explicit upper
2437 limits. As the maximum quantifier is 65535, the maximum group length was
2438 set at 30,000 so that the product of these two numbers did not overflow a
2439 32-bit integer. However, it turns out that people want to use groups that
2440 are longer than 30,000 bytes (though not repeat them that many times).
2441 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2442 made it possible to implement the integer overflow checks in a much more
2443 dynamic way, which I have now done. The artificial limitation on group
2444 length has been removed - we now have only the limit on the total length of
2445 the compiled pattern, which depends on the LINK_SIZE setting.
2446
2447 10. Fixed a bug in the documentation for get/copy named substring when
2448 duplicate names are permitted. If none of the named substrings are set, the
2449 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2450 empty string.
2451
2452 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2453 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2454 because the ] is interpreted as the first data character and the
2455 terminating ] is not found. PCRE has been made compatible with Perl in this
2456 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2457 cause memory overwriting.
2458
2459 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2460 string has been matched (to stop an infinite loop). It was not recognizing
2461 a conditional subpattern that could match an empty string if that
2462 subpattern was within another subpattern. For example, it looped when
2463 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2464 condition was not nested. This bug has been fixed.
2465
2466 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2467 past the start of the subject in the presence of bytes with the top bit
2468 set, for example "\x8aBCD".
2469
2470 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2471 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2472
2473 14. Optimized (?!) to (*FAIL).
2474
2475 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2476 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2477 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2478 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2479 does: it's just the validity check that is more restrictive.
2480
2481 16. Inserted checks for integer overflows during escape sequence (backslash)
2482 processing, and also fixed erroneous offset values for syntax errors during
2483 backslash processing.
2484
2485 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2486 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2487
2488 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2489 caused an overrun.
2490
2491 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2492 something other than just ASCII characters) inside a group that had an
2493 unlimited repeat caused a loop at compile time (while checking to see
2494 whether the group could match an empty string).
2495
2496 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2497 [\P{Any}] did so. (Error in the code for printing property names.)
2498
2499 21. An orphan \E inside a character class could cause a crash.
2500
2501 22. A repeated capturing bracket such as (A)? could cause a wild memory
2502 reference during compilation.
2503
2504 23. There are several functions in pcre_compile() that scan along a compiled
2505 expression for various reasons (e.g. to see if it's fixed length for look
2506 behind). There were bugs in these functions when a repeated \p or \P was
2507 present in the pattern. These operators have additional parameters compared
2508 with \d, etc, and these were not being taken into account when moving along
2509 the compiled data. Specifically:
2510
2511 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2512 length.
2513
2514 (b) An item such as \pL+ within a repeated group could cause crashes or
2515 loops.
2516
2517 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2518 "reference to non-existent subpattern" error.
2519
2520 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2521
2522 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2523 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2524
2525 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2526
2527 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2528 character were causing crashes (broken optimization).
2529
2530 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2531 \p or \P) caused a compile-time loop.
2532
2533 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2534 break. For example, the unstudied pattern [\r\n]A does not match the string
2535 "\r\nA" because change 7.0/46 below moves the current point on by two
2536 characters after failing to match at the start. However, the pattern \nA
2537 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2538 the same is true. There doesn't seem any very clean way out of this, but
2539 what I have chosen to do makes the common cases work: PCRE now takes note
2540 of whether there can be an explicit match for \r or \n anywhere in the
2541 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2542 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2543 pattern has explicit CR or LF references.
2544
2545 29. Added (*CR) etc for changing newline setting at start of pattern.
2546
2547
2548 Version 7.2 19-Jun-07
2549 ---------------------
2550
2551 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2552 which is apparently normally available under Windows.
2553
2554 2. Re-jig the pcregrep tests with different newline settings in an attempt
2555 to make them independent of the local environment's newline setting.
2556
2557 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2558
2559 4. Some of the "internals" tests were previously cut out when the link size
2560 was not 2, because the output contained actual offsets. The recent new
2561 "Z" feature of pcretest means that these can be cut out, making the tests
2562 usable with all link sizes.
2563
2564 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2565 stack recursion. This gives a massive performance boost under BSD, but just
2566 a small improvement under Linux. However, it saves one field in the frame
2567 in all cases.
2568
2569 6. Added more features from the forthcoming Perl 5.10:
2570
2571 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2572 recursion call. It refers to the nth most recently opened parentheses.
2573
2574 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2575 to be opened parentheses.
2576
2577 (c) Conditions that refer to capturing parentheses can be specified
2578 relatively, for example, (?(-2)... or (?(+3)...
2579
2580 (d) \K resets the start of the current match so that everything before
2581 is not part of it.
2582
2583 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2584
2585 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2586 reference syntax.
2587
2588 (g) (?| introduces a group in which the numbering of parentheses in each
2589 alternative starts with the same number.
2590
2591 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2592
2593 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2594 PCRE_INFO_JCHANGED.
2595
2596 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2597 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2598 for detecting groups that can match an empty string.
2599
2600 9. A pattern with a very large number of alternatives (more than several
2601 hundred) was running out of internal workspace during the pre-compile
2602 phase, where pcre_compile() figures out how much memory will be needed. A
2603 bit of new cunning has reduced the workspace needed for groups with
2604 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2605 workspace instead of running out of the 4096 that are available.
2606
2607 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2608
2609 11. Applied patch from Google to remove an optimization that didn't quite work.
2610 The report of the bug said:
2611
2612 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2613 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2614 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2615
2616 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2617 it matched the wrong number of bytes.
2618
2619
2620 Version 7.1 24-Apr-07
2621 ---------------------
2622
2623 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2624 that is more "standard", making use of automake and other Autotools. There
2625 is some re-arrangement of the files and adjustment of comments consequent
2626 on this.
2627
2628 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2629 for recursive directory scanning broke on some systems because the files
2630 are not scanned in any specific order and on different systems the order
2631 was different. A call to "sort" has been inserted into RunGrepTest for the
2632 approprate test as a short-term fix. In the longer term there may be an
2633 alternative.
2634
2635 3. I had an email from Eric Raymond about problems translating some of PCRE's
2636 man pages to HTML (despite the fact that I distribute HTML pages, some
2637 people do their own conversions for various reasons). The problems
2638 concerned the use of low-level troff macros .br and .in. I have therefore
2639 removed all such uses from the man pages (some were redundant, some could
2640 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2641 HTML has been updated to handle .nf/.fi and to complain if it encounters
2642 .br or .in.
2643
2644 4. Updated comments in configure.ac that get placed in config.h.in and also
2645 arranged for config.h to be included in the distribution, with the name
2646 config.h.generic, for the benefit of those who have to compile without
2647 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2648
2649 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2650 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2651 makevp.bat for the current PCRE, using the additional files
2652 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2653
2654 6. A Windows user reported a minor discrepancy with test 2, which turned out
2655 to be caused by a trailing space on an input line that had got lost in his
2656 copy. The trailing space was an accident, so I've just removed it.
2657
2658 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2659 that is needed.
2660
2661 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2662 as "const" (a) because they are and (b) because it helps the PHP
2663 maintainers who have recently made a script to detect big data structures
2664 in the php code that should be moved to the .rodata section. I remembered
2665 to update Builducptable as well, so it won't revert if ucptable.h is ever
2666 re-created.
2667
2668 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2669 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2670 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2671 support is not required. This saves 1.5-2K of code, which is important in
2672 some applications.
2673
2674 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2675 so as not to refer to the tables, even though these functions will never be
2676 called when UTF-8 support is disabled. Otherwise there are problems with a
2677 shared library.
2678
2679 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2680
2681 (a) It was defining its arguments as char * instead of void *.
2682
2683 (b) It was assuming that all moves were upwards in memory; this was true
2684 a long time ago when I wrote it, but is no longer the case.
2685
2686 The emulated memove() is provided for those environments that have neither
2687 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2688 is clearly not the case, as these two bugs were recently reported.
2689
2690 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2691 and Detrail to create the HTML documentation, the .txt form of the man
2692 pages, and it removes trailing spaces from listed files. It also creates
2693 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2694 case, it wraps all the #defines with #ifndefs. This script should be run
2695 before "make dist".
2696
2697 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2698 with Unicode property support.
2699
2700 (a) For a maximizing quantifier, if the two different cases of the
2701 character were of different lengths in their UTF-8 codings (there are
2702 some cases like this - I found 11), and the matching function had to
2703 back up over a mixture of the two cases, it incorrectly assumed they
2704 were both the same length.
2705
2706 (b) When PCRE was configured to use the heap rather than the stack for
2707 recursion during matching, it was not correctly preserving the data for
2708 the other case of a UTF-8 character when checking ahead for a match
2709 while processing a minimizing repeat. If the check also involved
2710 matching a wide character, but failed, corruption could cause an
2711 erroneous result when trying to check for a repeat of the original
2712 character.
2713
2714 13. Some tidying changes to the testing mechanism:
2715
2716 (a) The RunTest script now detects the internal link size and whether there
2717 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2718 values substituted by "configure". (The RunGrepTest script already did
2719 this for UTF-8.) The configure.ac script no longer substitutes the
2720 relevant variables.
2721
2722 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2723 with length and offset values. This means that the output is different
2724 for different internal link sizes. Test 2 is skipped for link sizes
2725 other than 2 because of this, bypassing the problem. Unfortunately,
2726 there was also a test in test 3 (the locale tests) that used /B and
2727 failed for link sizes other than 2. Rather than cut the whole test out,
2728 I have added a new /Z option to pcretest that replaces the length and
2729 offset values with spaces. This is now used to make test 3 independent
2730 of link size. (Test 2 will be tidied up later.)
2731
2732 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2733 segmentation fault instead of returning the appropriate error message.
2734
2735 15. In multiline mode when the newline sequence was set to "any", the pattern
2736 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2737 This doesn't seem right; it now treats the CRLF combination as the line
2738 ending, and so does not match in that case. It's only a pattern such as ^$
2739 that would hit this one: something like ^ABC$ would have failed after \r
2740 and then tried again after \r\n.
2741
2742 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2743 in an attempt to make files that differ only in their line terminators
2744 compare equal. This works on Linux.
2745
2746 17. Under certain error circumstances pcregrep might try to free random memory
2747 as it exited. This is now fixed, thanks to valgrind.
2748
2749 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2750 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2751 was because its rules for how to advance for /g after matching an empty
2752 string at the end of a line did not allow for this case. They now check for
2753 it specially.
2754
2755 20. pcretest is supposed to handle patterns and data of any length, by
2756 extending its buffers when necessary. It was getting this wrong when the
2757 buffer for a data line had to be extended.
2758
2759 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2760 CRLF as a newline sequence.
2761
2762 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2763 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2764 I have nevertheless tidied it up.
2765
2766 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2767
2768 24. Added a man page for pcre-config.
2769
2770
2771 Version 7.0 19-Dec-06
2772 ---------------------
2773
2774 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2775 moving to gcc 4.1.1.
2776
2777 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2778 sys/time.h, which is documented as needed for this function. It doesn't
2779 seem to matter on Linux, but it showed up on some releases of OS X.
2780
2781 3. It seems that there are systems where bytes whose values are greater than
2782 127 match isprint() in the "C" locale. The "C" locale should be the
2783 default when a C program starts up. In most systems, only ASCII printing
2784 characters match isprint(). This difference caused the output from pcretest
2785 to vary, making some of the tests fail. I have changed pcretest so that:
2786
2787 (a) When it is outputting text in the compiled version of a pattern, bytes
2788 other than 32-126 are always shown as hex escapes.
2789
2790 (b) When it is outputting text that is a matched part of a subject string,
2791 it does the same, unless a different locale has been set for the match
2792 (using the /L modifier). In this case, it uses isprint() to decide.
2793
2794 4. Fixed a major bug that caused incorrect computation of the amount of memory
2795 required for a compiled pattern when options that changed within the
2796 pattern affected the logic of the preliminary scan that determines the
2797 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2798 that the computed length was too small. The symptoms of this bug were
2799 either the PCRE error "internal error: code overflow" from pcre_compile(),
2800 or a glibc crash with a message such as "pcretest: free(): invalid next
2801 size (fast)". Examples of patterns that provoked this bug (shown in
2802 pcretest format) are:
2803
2804 /(?-x: )/x
2805 /(?x)(?-x: \s*#\s*)/
2806 /((?i)[\x{c0}])/8
2807 /(?i:[\x{c0}])/8
2808
2809 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2810 is now done differently.
2811
2812 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2813 wrapper classes; (b) implement a new function in the C++ scanner that is
2814 more efficient than the old way of doing things because it avoids levels of
2815 recursion in the regex matching; (c) add a paragraph to the documentation
2816 for the FullMatch() function.
2817
2818 6. The escape sequence \n was being treated as whatever was defined as
2819 "newline". Not only was this contrary to the documentation, which states
2820 that \n is character 10 (hex 0A), but it also went horribly wrong when
2821 "newline" was defined as CRLF. This has been fixed.
2822
2823 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2824 was being set to -1 for the "end of line" case (supposedly a value that no
2825 character can have). Though this value is never used (the check for end of
2826 line is "zero bytes in current character"), it caused compiler complaints.
2827 I've changed it to 0xffffffff.
2828
2829 8. In pcre_version.c, the version string was being built by a sequence of
2830 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2831 string (as it is for production releases) called a macro with an empty
2832 argument. The C standard says the result of this is undefined. The gcc
2833 compiler treats it as an empty string (which was what was wanted) but it is
2834 reported that Visual C gives an error. The source has been hacked around to
2835 avoid this problem.
2836
2837 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2838 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2839 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2840 of them did).
2841
2842 10. Originally, pcretest opened its input and output without "b"; then I was
2843 told that "b" was needed in some environments, so it was added for release
2844 5.0 to both the input and output. (It makes no difference on Unix-like
2845 systems.) Later I was told that it is wrong for the input on Windows. I've
2846 now abstracted the modes into two macros, to make it easier to fiddle with
2847 them, and removed "b" from the input mode under Windows.
2848
2849 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2850
2851 12. Added -help and --help to pcretest as an official way of being reminded
2852 of the options.
2853
2854 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2855 and pcrecpp.cc because they annoy compilers at high warning levels.
2856
2857 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2858
2859 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2860 scripts are not C programs :-) and which was not noticed because it works
2861 on Linux.
2862
2863 16. pcretest is supposed to handle any length of pattern and data line (as one
2864 line or as a continued sequence of lines) by extending its input buffer if
2865 necessary. This feature was broken for very long pattern lines, leading to
2866 a string of junk being passed to pcre_compile() if the pattern was longer
2867 than about 50K.
2868
2869 17. I have done a major re-factoring of the way pcre_compile() computes the
2870 amount of memory needed for a compiled pattern. Previously, there was code
2871 that made a preliminary scan of the pattern in order to do this. That was
2872 OK when PCRE was new, but as the facilities have expanded, it has become
2873 harder and harder to keep it in step with the real compile phase, and there
2874 have been a number of bugs (see for example, 4 above). I have now found a
2875 cunning way of running the real compile function in a "fake" mode that
2876 enables it to compute how much memory it would need, while actually only
2877 ever using a few hundred bytes of working memory and without too many
2878 tests of the mode. This should make future maintenance and development
2879 easier. A side effect of this work is that the limit of 200 on the nesting
2880 depth of parentheses has been removed (though this was never a serious
2881 limitation, I suspect). However, there is a downside: pcre_compile() now
2882 runs more slowly than before (30% or more, depending on the pattern). I
2883 hope this isn't a big issue. There is no effect on runtime performance.
2884
2885 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2886 newline (only possible for the last line of a file) and it was a
2887 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2888
2889 19. Added additional timing features to pcretest. (1) The -tm option now times
2890 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2891 separate command line item, by a number that specifies the number of
2892 repeats to use when timing. The default is 50000; this gives better
2893 precision, but takes uncomfortably long for very large patterns.
2894
2895 20. Extended pcre_study() to be more clever in cases where a branch of a
2896 subpattern has no definite first character. For example, (a*|b*)[cd] would
2897 previously give no result from pcre_study(). Now it recognizes that the
2898 first character must be a, b, c, or d.
2899
2900 21. There was an incorrect error "recursive call could loop indefinitely" if
2901 a subpattern (or the entire pattern) that was being tested for matching an
2902 empty string contained only one non-empty item after a nested subpattern.
2903 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2904 incorrectly, because the \d was being skipped in the check.
2905
2906 22. The pcretest program now has a new pattern option /B and a command line
2907 option -b, which is equivalent to adding /B to every pattern. This causes
2908 it to show the compiled bytecode, without the additional information that
2909 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2910 is the same as /B/I).
2911
2912 23. A new optimization is now able automatically to treat some sequences such
2913 as a*b as a*+b. More specifically, if something simple (such as a character
2914 or a simple class like \d) has an unlimited quantifier, and is followed by
2915 something that cannot possibly match the quantified thing, the quantifier
2916 is automatically "possessified".
2917
2918 24. A recursive reference to a subpattern whose number was greater than 39
2919 went wrong under certain circumstances in UTF-8 mode. This bug could also
2920 have affected the operation of pcre_study().
2921
2922 25. Realized that a little bit of performance could be had by replacing
2923 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2924
2925 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2926
2927 27. Possessive quantifiers such as a++ were previously implemented by turning
2928 them into atomic groups such as ($>a+). Now they have their own opcodes,
2929 which improves performance. This includes the automatically created ones
2930 from 23 above.
2931
2932 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2933 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2934 the first matched character to be a colon. This applied both to named and
2935 numbered groups.
2936
2937 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2938
2939 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2940 building PCRE on Windows easier, so I have included it in the distribution.
2941
2942 31. There is now a check in pcretest against a ridiculously large number being
2943 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2944 loop, the loop is abandoned.
2945
2946 32. Forward references to subpatterns in conditions such as (?(2)...) where
2947 subpattern 2 is defined later cause pcre_compile() to search forwards in
2948 the pattern for the relevant set of parentheses. This search went wrong
2949 when there were unescaped parentheses in a character class, parentheses
2950 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2951
2952 33. "Subroutine" calls and backreferences were previously restricted to
2953 referencing subpatterns earlier in the regex. This restriction has now
2954 been removed.
2955
2956 34. Added a number of extra features that are going to be in Perl 5.10. On the
2957 whole, these are just syntactic alternatives for features that PCRE had
2958 previously implemented using the Python syntax or my own invention. The
2959 other formats are all retained for compatibility.
2960
2961 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2962 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2963 also .NET compatible.
2964
2965 (b) A recursion or subroutine call to a named group can now be defined as
2966 (?&name) as well as (?P>name).
2967
2968 (c) A backreference to a named group can now be defined as \k<name> or
2969 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2970 5.10, are also .NET compatible.
2971
2972 (d) A conditional reference to a named group can now use the syntax
2973 (?(<name>) or (?('name') as well as (?(name).
2974
2975 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2976 groups (named and numbered) that are never evaluated inline, but can be
2977 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2978 is always false. There may be only one alternative in such a group.
2979
2980 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2981 as the simple (?(R). The condition is true only if the most recent
2982 recursion is that of the given number or name. It does not search out
2983 through the entire recursion stack.
2984
2985 (g) The escape \gN or \g{N} has been added, where N is a positive or
2986 negative number, specifying an absolute or relative reference.
2987
2988 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2989 some "unreachable code" warnings.
2990
2991 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2992 things, this adds five new scripts.
2993
2994 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2995 There were also incompatibilities regarding the handling of \Q..\E inside
2996 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2997 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2998
2999 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3000 matches an empty string, and forcibly breaks the loop. There were bugs in
3001 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3002 against aaaa the result was just "a" rather than "aaaa", for example. Two
3003 separate and independent bugs (that affected different cases) have been
3004 fixed.
3005
3006 39. Refactored the code to abolish the use of different opcodes for small
3007 capturing bracket numbers. This is a tidy that I avoided doing when I
3008 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3009 The new approach is not only tidier, it makes it possible to reduce the
3010 memory needed to fix the previous bug (38).
3011
3012 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3013 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3014 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3015 mode.
3016
3017 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3018 report.
3019
3020 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3021 copy construction and assignment in the C++ wrapper.
3022
3023 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3024 couple of bugs that could have given wrong results in the "--newline=crlf"
3025 case.
3026
3027 44. Added a number of casts and did some reorganization of signed/unsigned int
3028 variables following suggestions from Dair Grant. Also renamed the variable
3029 "this" as "item" because it is a C++ keyword.
3030
3031 45. Arranged for dftables to add
3032
3033 #include "pcre_internal.h"
3034
3035 to pcre_chartables.c because without it, gcc 4.x may remove the array
3036 definition from the final binary if PCRE is built into a static library and
3037 dead code stripping is activated.
3038
3039 46. For an unanchored pattern, if a match attempt fails at the start of a
3040 newline sequence, and the newline setting is CRLF or ANY, and the next two
3041 characters are CRLF, advance by two characters instead of one.
3042
3043
3044 Version 6.7 04-Jul-06
3045 ---------------------
3046
3047 1. In order to handle tests when input lines are enormously long, pcretest has
3048 been re-factored so that it automatically extends its buffers when
3049 necessary. The code is crude, but this _is_ just a test program. The
3050 default size has been increased from 32K to 50K.
3051
3052 2. The code in pcre_study() was using the value of the re argument before
3053 testing it for NULL. (Of course, in any sensible call of the function, it
3054 won't be NULL.)
3055
3056 3. The memmove() emulation function in pcre_internal.h, which is used on
3057 systems that lack both memmove() and bcopy() - that is, hardly ever -
3058 was missing a "static" storage class specifier.
3059
3060 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3061 containing an extended class (one that cannot be represented by a bitmap
3062 because it contains high-valued characters or Unicode property items, e.g.
3063 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3064 pattern, but PCRE should not loop if you do not (it no longer does).
3065 [Detail: two cases were found: (a) a repeated subpattern containing an
3066 extended class; (b) a recursive reference to a subpattern that followed a
3067 previous extended class. It wasn't skipping over the extended class
3068 correctly when UTF-8 mode was not set.]
3069
3070 5. A negated single-character class was not being recognized as fixed-length
3071 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3072 compile error "lookbehind assertion is not fixed length".
3073
3074 6. The RunPerlTest auxiliary script was showing an unexpected difference
3075 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3076 write a Perl script that can interpret lines of an input file either as
3077 byte characters or as UTF-8, which is what "perltest" was being required to
3078 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3079 can't do is switch easily at run time between having the "use utf8;" pragma
3080 or not. In the end, I fudged it by using the RunPerlTest script to insert
3081 "use utf8;" explicitly for the UTF-8 tests.
3082
3083 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3084 the end of the subject string, contrary to the documentation and to what
3085 Perl does. This was true of both matching functions. Now it matches only at
3086 the start of the subject and immediately after *internal* newlines.
3087
3088 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3089 a pointer to an int instead of a pointer to an unsigned long int. This
3090 caused problems on 64-bit systems.
3091
3092 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3093 instance of the 'standard' template library not being so standard".
3094
3095 10. There was no check on the number of named subpatterns nor the maximum
3096 length of a subpattern name. The product of these values is used to compute
3097 the size of the memory block for a compiled pattern. By supplying a very
3098 long subpattern name and a large number of named subpatterns, the size
3099 computation could be caused to overflow. This is now prevented by limiting
3100 the length of names to 32 characters, and the number of named subpatterns
3101 to 10,000.
3102
3103 11. Subpatterns that are repeated with specific counts have to be replicated in
3104 the compiled pattern. The size of memory for this was computed from the
3105 length of the subpattern and the repeat count. The latter is limited to
3106 65535, but there was no limit on the former, meaning that integer overflow
3107 could in principle occur. The compiled length of a repeated subpattern is
3108 now limited to 30,000 bytes in order to prevent this.
3109
3110 12. Added the optional facility to have named substrings with the same name.
3111
3112 13. Added the ability to use a named substring as a condition, using the
3113 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3114 are numbers (not recommended). Forward references are permitted.
3115
3116 14. Added forward references in named backreferences (if you see what I mean).
3117
3118 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3119 pattern could run off the end of the subject. For example, the pattern
3120 "(?s)(.{1,5})"8 did this with the subject "ab".
3121
3122 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3123 PCRE_CASELESS was set when matching characters that were quantified with ?
3124 or *.
3125
3126 17. A character class other than a single negated character that had a minimum
3127 but no maximum quantifier - for example [ab]{6,} - was not handled
3128 correctly by pce_dfa_exec(). It would match only one character.
3129
3130 18. A valid (though odd) pattern that looked like a POSIX character
3131 class but used an invalid character after [ (for example [[,abc,]]) caused
3132 pcre_compile() to give the error "Failed: internal error: code overflow" or
3133 in some cases to crash with a glibc free() error. This could even happen if
3134 the pattern terminated after [[ but there just happened to be a sequence of
3135 letters, a binary zero, and a closing ] in the memory that followed.
3136
3137 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3138 over the years. Originally (before any Unicode support), just the bottom 8
3139 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3140 output from "man perlunicode" includes this:
3141
3142 The regular expression compiler produces polymorphic opcodes. That
3143 is, the pattern adapts to the data and automatically switches to
3144 the Unicode character scheme when presented with Unicode data--or
3145 instead uses a traditional byte scheme when presented with byte
3146 data.
3147
3148 Sadly, a wide octal escape does not cause a switch, and in a string with
3149 no other multibyte characters, these octal escapes are treated as before.
3150 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3151 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3152 Unicode string.
3153
3154 I have not perpetrated such confusion in PCRE. Up till now, it took just
3155 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3156 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3157 translate to the appropriate multibyte character.
3158
3159 29. Applied some refactoring to reduce the number of warnings from Microsoft
3160 and Borland compilers. This has included removing the fudge introduced
3161 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3162 a warning about an unused variable.
3163
3164 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3165 characters since release 4.0, because Perl (from release 5.004) does not.
3166 [Or at least, is documented not to: some releases seem to be in conflict
3167 with the documentation.] However, when a pattern was studied with
3168 pcre_study() and all its branches started with \s, PCRE still included VT
3169 as a possible starting character. Of course, this did no harm; it just
3170 caused an unnecessary match attempt.
3171
3172 22. Removed a now-redundant internal flag bit that recorded the fact that case
3173 dependency changed within the pattern. This was once needed for "required
3174 byte" processing, but is no longer used. This recovers a now-scarce options
3175 bit. Also moved the least significant internal flag bit to the most-
3176 significant bit of the word, which was not previously used (hangover from
3177 the days when it was an int rather than a uint) to free up another bit for
3178 the future.
3179
3180 23. Added support for CRLF line endings as well as CR and LF. As well as the
3181 default being selectable at build time, it can now be changed at runtime
3182 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3183 specify that it is scanning data with non-default line endings.
3184
3185 24. Changed the definition of CXXLINK to make it agree with the definition of
3186 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3187
3188 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3189 recursions. This makes a big different to stack usage for some patterns.
3190
3191 26. If a subpattern containing a named recursion or subroutine reference such
3192 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3193 the space required for the compiled pattern went wrong and gave too small a
3194 value. Depending on the environment, this could lead to "Failed: internal
3195 error: code overflow at offset 49" or "glibc detected double free or
3196 corruption" errors.
3197
3198 27. Applied patches from Google (a) to support the new newline modes and (b) to
3199 advance over multibyte UTF-8 characters in GlobalReplace.
3200
3201 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3202 difference for some implementation of PCRE in some Windows version.
3203
3204 29. Added some extra testing facilities to pcretest:
3205
3206 \q<number> in a data line sets the "match limit" value
3207 \Q<number> in a data line sets the "match recursion limt" value
3208 -S <number> sets the stack size, where <number> is in megabytes
3209
3210 The -S option isn't available for Windows.
3211
3212
3213 Version 6.6 06-Feb-06
3214 ---------------------
3215
3216 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3217 in pcreposix.h. I have copied the definition from pcre.h.
3218
3219 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3220 because pcre.h is no longer a built file.
3221
3222 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3223 not normally included in the compiled code.
3224
3225
3226 Version 6.5 01-Feb-06
3227 ---------------------
3228
3229 1. When using the partial match feature with pcre_dfa_exec(), it was not
3230 anchoring the second and subsequent partial matches at the new starting
3231 point. This could lead to incorrect results. For example, with the pattern
3232 /1234/, partially matching against "123" and then "a4" gave a match.
3233
3234 2. Changes to pcregrep:
3235
3236 (a) All non-match returns from pcre_exec() were being treated as failures
3237 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3238 error message is output. Some extra information is given for the
3239 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3240 probably the only errors that are likely to be caused by users (by
3241 specifying a regex that has nested indefinite repeats, for instance).
3242 If there are more than 20 of these errors, pcregrep is abandoned.
3243
3244 (b) A binary zero was treated as data while matching, but terminated the
3245 output line if it was written out. This has been fixed: binary zeroes
3246 are now no different to any other data bytes.
3247
3248 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3249 used to set a locale for matching. The --locale=xxxx long option has
3250 been added (no short equivalent) to specify a locale explicitly on the
3251 pcregrep command, overriding the environment variables.
3252
3253 (d) When -B was used with -n, some line numbers in the output were one less
3254 than they should have been.
3255
3256 (e) Added the -o (--only-matching) option.
3257
3258 (f) If -A or -C was used with -c (count only), some lines of context were
3259 accidentally printed for the final match.
3260
3261 (g) Added the -H (--with-filename) option.
3262
3263 (h) The combination of options -rh failed to suppress file names for files
3264 that were found from directory arguments.
3265
3266 (i) Added the -D (--devices) and -d (--directories) options.
3267
3268 (j) Added the -F (--fixed-strings) option.
3269
3270 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3271
3272 (l) Added the --colo(u)r option.
3273
3274 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3275 is not present by default.
3276
3277 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3278 items such as (?R) or (?1), when the recursion could match a number of
3279 alternatives. If it matched one of the alternatives, but subsequently,
3280 outside the recursion, there was a failure, the code tried to back up into
3281 the recursion. However, because of the way PCRE is implemented, this is not
3282 possible, and the result was an incorrect result from the match.
3283
3284 In order to prevent this happening, the specification of recursion has
3285 been changed so that all such subpatterns are automatically treated as
3286 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3287
3288 4. I had overlooked the fact that, in some locales, there are characters for
3289 which isalpha() is true but neither isupper() nor islower() are true. In
3290 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3291 and ordfeminine) are like this. This affected the treatment of \w and \W
3292 when they appeared in character classes, but not when they appeared outside
3293 a character class. The bit map for "word" characters is now created
3294 separately from the results of isalnum() instead of just taking it from the
3295 upper, lower, and digit maps. (Plus the underscore character, of course.)
3296
3297 5. The above bug also affected the handling of POSIX character classes such as
3298 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3299 permanent tables. Instead, the bit maps for such a class were previously
3300 created as the appropriate unions of the upper, lower, and digit bitmaps.
3301 Now they are created by subtraction from the [[:word:]] class, which has
3302 its own bitmap.
3303
3304 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3305 It is created by subtracting the vertical space characters (\x09, \x0a,
3306 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3307 subtraction was done in the overall bitmap for a character class, meaning
3308 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3309 be recognized. This bug has been fixed.
3310
3311 7. Patches from the folks at Google:
3312
3313 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3314 real life, but is still worth protecting against".
3315
3316 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3317 regular expressions".
3318
3319 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3320 have it.
3321
3322 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3323 "configure" and the latter not, in order to fix a problem somebody had
3324 with compiling the Arg class on HP-UX.
3325
3326 (e) Improve the error-handling of the C++ wrapper a little bit.
3327
3328 (f) New tests for checking recursion limiting.
3329
3330 8. The pcre_memmove() function, which is used only if the environment does not
3331 have a standard memmove() function (and is therefore rarely compiled),
3332 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3333 returning a result (though PCRE never actually uses the result).
3334
3335 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3336 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3337 returned instead of calling malloc() with an overflowing number that would
3338 most likely cause subsequent chaos.
3339
3340 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3341
3342 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3343 with this option is matched, the nmatch and pmatch options of regexec() are
3344 ignored.
3345
3346 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3347 provided in case anyone wants to the the POSIX interface with UTF-8
3348 strings.
3349
3350 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3351 C++ linking (needed for some HP-UX environments).
3352
3353 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3354 (unused parameter) and in the pcre_printint() function (omitted "default"
3355 switch label when the default is to do nothing).
3356
3357 15. Added some code to make it possible, when PCRE is compiled as a C++
3358 library, to replace subject pointers for pcre_exec() with a smart pointer
3359 class, thus making it possible to process discontinuous strings.
3360
3361 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3362 much the same function. They were added by different people who were trying
3363 to make PCRE easy to compile on non-Unix systems. It has been suggested
3364 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3365 for compiling on Windows systems. I have therefore replaced it with
3366 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3367 defaults to "extern" for C or "extern C" for C++, which works fine on
3368 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3369 SCOPE with something explicit in config.h. In addition:
3370
3371 (a) pcreposix.h still had just "extern" instead of either of these macros;
3372 I have replaced it with PCRE_DATA_SCOPE.
3373
3374 (b) Functions such as _pcre_xclass(), which are internal to the library,
3375 but external in the C sense, all had PCRE_EXPORT in their definitions.
3376 This is apparently wrong for the Windows case, so I have removed it.
3377 (It makes no difference on Unix-like systems.)
3378
3379 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3380 of recursive calls to match(). This is different to MATCH_LIMIT because
3381 that limits the total number of calls to match(), not all of which increase
3382 the depth of recursion. Limiting the recursion depth limits the amount of
3383 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3384 when PCRE is compiled, and changed at run time. A patch from Google adds
3385 this functionality to the C++ interface.
3386
3387 18. Changes to the handling of Unicode character properties:
3388
3389 (a) Updated the table to Unicode 4.1.0.
3390
3391 (b) Recognize characters that are not in the table as "Cn" (undefined).
3392
3393 (c) I revised the way the table is implemented to a much improved format
3394 which includes recognition of ranges. It now supports the ranges that
3395 are defined in UnicodeData.txt, and it also amalgamates other
3396 characters into ranges. This has reduced the number of entries in the
3397 table from around 16,000 to around 3,000, thus reducing its size
3398 considerably. I realized I did not need to use a tree structure after
3399 all - a binary chop search is just as efficient. Having reduced the
3400 number of entries, I extended their size from 6 bytes to 8 bytes to
3401 allow for more data.
3402
3403 (d) Added support for Unicode script names via properties such as \p{Han}.
3404
3405 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3406 matching that character.
3407
3408 20. When matching a repeated Unicode property with a minimum greater than zero,
3409 (for example \pL{2,}), PCRE could look past the end of the subject if it
3410 reached it while seeking the minimum number of characters. This could
3411 happen only if some of the characters were more than one byte long, because
3412 there is a check for at least the minimum number of bytes.
3413
3414 21. Refactored the implementation of \p and \P so as to be more general, to
3415 allow for more different types of property in future. This has changed the
3416 compiled form incompatibly. Anybody with saved compiled patterns that use
3417 \p or \P will have to recompile them.
3418
3419 22. Added "Any" and "L&" to the supported property types.
3420
3421 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3422 but give a compile time error if the value is greater than 0xff.
3423
3424 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3425 accidentally not being installed or uninstalled.
3426
3427 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3428 made were to insert the current release number. This seemed silly, because
3429 it made things harder for people building PCRE on systems that don't run
3430 "configure". I have turned pcre.h into a distributed file, no longer built
3431 by "configure", with the version identification directly included. There is
3432 no longer a pcre.h.in file.
3433
3434 However, this change necessitated a change to the pcre-config script as
3435 well. It is built from pcre-config.in, and one of the substitutions was the
3436 release number. I have updated configure.ac so that ./configure now finds
3437 the release number by grepping pcre.h.
3438
3439 26. Added the ability to run the tests under valgrind.
3440
3441
3442 Version 6.4 05-Sep-05
3443 ---------------------
3444
3445 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3446 "--" to be printed when multiple files were scanned, even when none of the
3447 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3448 consider it to be a bug, and have restored the previous behaviour.
3449
3450 2. A couple of code tidies to get rid of compiler warnings.
3451
3452 3. The pcretest program used to cheat by referring to symbols in the library
3453 whose names begin with _pcre_. These are internal symbols that are not
3454 really supposed to be visible externally, and in some environments it is
3455 possible to suppress them. The cheating is now confined to including
3456 certain files from the library's source, which is a bit cleaner.
3457
3458 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3459 file's purpose clearer.
3460
3461 5. Reorganized pcre_ucp_findchar().
3462
3463
3464 Version 6.3 15-Aug-05
3465 ---------------------
3466
3467 1. The file libpcre.pc.in did not have general read permission in the tarball.
3468
3469 2. There were some problems when building without C++ support:
3470
3471 (a) If C++ support was not built, "make install" and "make test" still
3472 tried to test it.
3473
3474 (b) There were problems when the value of CXX was explicitly set. Some
3475 changes have been made to try to fix these, and ...
3476
3477 (c) --disable-cpp can now be used to explicitly disable C++ support.
3478
3479 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3480 backslash in a target when C++ was disabled. This confuses some
3481 versions of "make", apparently. Using an intermediate variable solves
3482 this. (Same for CPP_LOBJ.)
3483
3484 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3485 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3486 necessary on certain architectures.
3487
3488 4. Added a setting of -export-symbols-regex to the link command to remove
3489 those symbols that are exported in the C sense, but actually are local
3490 within the library, and not documented. Their names all begin with
3491 "_pcre_". This is not a perfect job, because (a) we have to except some
3492 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3493 available (and never for static libraries). I have made a note to try to
3494 find a way round (a) in the future.
3495
3496
3497 Version 6.2 01-Aug-05
3498 ---------------------
3499
3500 1. There was no test for integer overflow of quantifier values. A construction
3501 such as {1111111111111111} would give undefined results. What is worse, if
3502 a minimum quantifier for a parenthesized subpattern overflowed and became
3503 negative, the calculation of the memory size went wrong. This could have
3504 led to memory overwriting.
3505
3506 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3507
3508 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3509 operating environments where this matters.
3510
3511 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3512 PCRE options from within the C++ wrapper.
3513
3514 5. Named capturing subpatterns were not being correctly counted when a pattern
3515 was compiled. This caused two problems: (a) If there were more than 100
3516 such subpatterns, the calculation of the memory needed for the whole
3517 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3518 back references of the form \12, where the number was greater than 9, were
3519 not recognized as back references, even though there were sufficient
3520 previous subpatterns.
3521
3522 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3523 versions of gcc, e.g. 2.95.4.
3524
3525
3526 Version 6.1 21-Jun-05
3527 ---------------------
3528
3529 1. There was one reference to the variable "posix" in pcretest.c that was not
3530 surrounded by "#if !defined NOPOSIX".
3531
3532 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3533 the cross-check on the old pcre_info() function, for the benefit of the
3534 cut-down version of PCRE that is currently imported into Exim.
3535
3536 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3537 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3538 patterns that start (?i) but I don't think that matters. The use of (?i) is
3539 just an example; this all applies to the other options as well.
3540
3541 4. Since libtool seems to echo the compile commands it is issuing, the output
3542 from "make" can be reduced a bit by putting "@" in front of each libtool
3543 compile command.
3544
3545 5. Patch from the folks at Google for configure.in to be a bit more thorough
3546 in checking for a suitable C++ installation before trying to compile the
3547 C++ stuff. This should fix a reported problem when a compiler was present,
3548 but no suitable headers.
3549
3550 6. The man pages all had just "PCRE" as their title. I have changed them to
3551 be the relevant file name. I have also arranged that these names are
3552 retained in the file doc/pcre.txt, which is a concatenation in text format
3553 of all the man pages except the little individual ones for each function.
3554
3555 7. The NON-UNIX-USE file had not been updated for the different set of source
3556 files that come with release 6. I also added a few comments about the C++
3557 wrapper.
3558
3559
3560 Version 6.0 07-Jun-05
3561 ---------------------
3562
3563 1. Some minor internal re-organization to help with my DFA experiments.
3564
3565 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3566 didn't matter for the library itself when fully configured, but did matter
3567 when compiling without UCP support, or within Exim, where the ucp files are
3568 not imported.
3569
3570 3. Refactoring of the library code to split up the various functions into
3571 different source modules. The addition of the new DFA matching code (see
3572 below) to a single monolithic source would have made it really too
3573 unwieldy, quite apart from causing all the code to be include in a
3574 statically linked application, when only some functions are used. This is
3575 relevant even without the DFA addition now that patterns can be compiled in
3576 one application and matched in another.
3577
3578 The downside of splitting up is that there have to be some external
3579 functions and data tables that are used internally in different modules of
3580 the library but which are not part of the API. These have all had their
3581 names changed to start with "_pcre_" so that they are unlikely to clash
3582 with other external names.
3583
3584 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3585 a different (DFA) algorithm. Although it is slower than the original
3586 function, it does have some advantages for certain types of matching
3587 problem.
3588
3589 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3590 including restarting after a partial match.
3591
3592 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3593 defined when compiling for Windows was sent to me. I have put it into the
3594 code, though I have no means of testing or verifying it.
3595
3596 7. Added the pcre_refcount() auxiliary function.
3597
3598 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3599 match before or at the first newline in the subject string. In pcretest,
3600 the /f option on a pattern can be used to set this.
3601
3602 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3603 would behave wrongly. This has been present in PCRE since release 4.0.
3604
3605 10. A number of changes to the pcregrep command:
3606
3607 (a) Refactored how -x works; insert ^(...)$ instead of setting
3608 PCRE_ANCHORED and checking the length, in preparation for adding
3609 something similar for -w.
3610
3611 (b) Added the -w (match as a word) option.
3612
3613 (c) Refactored the way lines are read and buffered so as to have more
3614 than one at a time available.
3615
3616 (d) Implemented a pcregrep test script.
3617
3618 (e) Added the -M (multiline match) option. This allows patterns to match
3619 over several lines of the subject. The buffering ensures that at least
3620 8K, or the rest of the document (whichever is the shorter) is available
3621 for matching (and similarly the previous 8K for lookbehind assertions).
3622
3623 (f) Changed the --help output so that it now says
3624
3625 -w, --word-regex(p)
3626
3627 instead of two lines, one with "regex" and the other with "regexp"
3628 because that confused at least one person since the short forms are the
3629 same. (This required a bit of code, as the output is generated
3630 automatically from a table. It wasn't just a text change.)
3631
3632 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3633 option but starts with a hyphen. Could be a pattern or a path name
3634 starting with a hyphen, for instance.
3635
3636 (h) "-" can be given as a file name to represent stdin.
3637
3638 (i) When file names are being printed, "(standard input)" is used for
3639 the standard input, for compatibility with GNU grep. Previously
3640 "<stdin>" was used.
3641
3642 (j) The option --label=xxx can be used to supply a name to be used for
3643 stdin when file names are being printed. There is no short form.
3644
3645 (k) Re-factored the options decoding logic because we are going to add
3646 two more options that take data. Such options can now be given in four
3647 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3648
3649 (l) Added the -A, -B, and -C options for requesting that lines of context
3650 around matches be printed.
3651
3652 (m) Added the -L option to print the names of files that do not contain
3653 any matching lines, that is, the complement of -l.
3654
3655 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3656 continue to scan other files.
3657
3658 (o) The -s option was incorrectly implemented. For compatibility with other
3659 greps, it now suppresses the error message for a non-existent or non-
3660 accessible file (but not the return code). There is a new option called
3661 -q that suppresses the output of matching lines, which was what -s was
3662 previously doing.
3663
3664 (p) Added --include and --exclude options to specify files for inclusion
3665 and exclusion when recursing.
3666
3667 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3668 Hopefully, it now does.
3669
3670 12. Missing cast in pcre_study().
3671
3672 13. Added an "uninstall" target to the makefile.
3673
3674 14. Replaced "extern" in the function prototypes in Makefile.in with
3675 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3676 world, but is set differently for Windows.
3677
3678 15. Added a second compiling function called pcre_compile2(). The only
3679 difference is that it has an extra argument, which is a pointer to an
3680 integer error code. When there is a compile-time failure, this is set
3681 non-zero, in addition to the error test pointer being set to point to an
3682 error message. The new argument may be NULL if no error number is required
3683 (but then you may as well call pcre_compile(), which is now just a
3684 wrapper). This facility is provided because some applications need a
3685 numeric error indication, but it has also enabled me to tidy up the way
3686 compile-time errors are handled in the POSIX wrapper.
3687
3688 16. Added VPATH=.libs to the makefile; this should help when building with one
3689 prefix path and installing with another. (Or so I'm told by someone who
3690 knows more about this stuff than I do.)
3691
3692 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3693 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3694 match everything, including newlines. This is not POSIX-compatible, but
3695 somebody wanted the feature. From pcretest it can be activated by using
3696 both the P and the s flags.
3697
3698 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3699
3700 19. libpcre.pc was being incorrectly installed as executable.
3701
3702 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3703 it now also looks for '\r' so that it will work unmodified on Windows.
3704
3705 21. Added Google's contributed C++ wrapper to the distribution.
3706
3707 22. Added some untidy missing memory free() calls in pcretest, to keep
3708 Electric Fence happy when testing.
3709
3710
3711
3712 Version 5.0 13-Sep-04
3713 ---------------------
3714
3715 1. Internal change: literal characters are no longer packed up into items
3716 containing multiple characters in a single byte-string. Each character
3717 is now matched using a separate opcode. However, there may be more than one
3718 byte in the character in UTF-8 mode.
3719
3720 2. The pcre_callout_block structure has two new fields: pattern_position and
3721 next_item_length. These contain the offset in the pattern to the next match
3722 item, and its length, respectively.
3723
3724 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3725 insertion of callouts before each pattern item. Added the /C option to
3726 pcretest to make use of this.
3727
3728 4. On the advice of a Windows user, the lines
3729
3730 #if defined(_WIN32) || defined(WIN32)
3731 _setmode( _fileno( stdout ), 0x8000 );
3732 #endif /* defined(_WIN32) || defined(WIN32) */
3733
3734 have been added to the source of pcretest. This apparently does useful
3735 magic in relation to line terminators.
3736
3737 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3738 for the benefit of those environments where the "b" makes a difference.
3739
3740 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3741 to know about it. I have put a hack into configure.in that adds in code
3742 to set GCC=yes if CC=icc. This seems to end up at a point in the
3743 generated configure script that is early enough to affect the setting of
3744 compiler options, which is what is needed, but I have no means of testing
3745 whether it really works. (The user who reported this had patched the
3746 generated configure script, which of course I cannot do.)
3747
3748 LATER: After change 22 below (new libtool files), the configure script
3749 seems to know about icc (and also ecc). Therefore, I have commented out
3750 this hack in configure.in.
3751
3752 7. Added support for pkg-config (2 patches were sent in).
3753
3754 8. Negated POSIX character classes that used a combination of internal tables
3755 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3756 [[:^ascii]]. Typically, they would match almost any characters. The other
3757 POSIX classes were not broken in this way.
3758
3759 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3760 to find the match, as PCRE was deluded into thinking that the match had to
3761 start at the start point or following a newline. The same bug applied to
3762 patterns with negative forward assertions or any backward assertions
3763 preceding ".*" at the start, unless the pattern required a fixed first
3764 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3765
3766 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3767 starting at the last subject character, bytes beyond the end of the subject
3768 string were read.
3769
3770 11. Renamed the variable "class" as "classbits" to make life easier for C++
3771 users. (Previously there was a macro definition, but it apparently wasn't
3772 enough.)
3773
3774 12. Added the new field "tables" to the extra data so that tables can be passed
3775 in at exec time, or the internal tables can be re-selected. This allows
3776 a compiled regex to be saved and re-used at a later time by a different
3777 program that might have everything at different addresses.
3778
3779 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3780 -R library as well as a -L library.
3781
3782 14. The debugging options of pcretest (-d on the command line or D on a
3783 pattern) showed incorrect output for anything following an extended class
3784 that contained multibyte characters and which was followed by a quantifier.
3785
3786 15. Added optional support for general category Unicode character properties
3787 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3788 support. It adds about 90K to the size of the library. The meanings of the
3789 inbuilt class escapes such as \d and \s have NOT been changed.
3790
3791 16. Updated pcredemo.c to include calls to free() to release the memory for the
3792 compiled pattern.
3793
3794 17. The generated file chartables.c was being created in the source directory
3795 instead of in the building directory. This caused the build to fail if the
3796 source directory was different from the building directory, and was
3797 read-only.
3798
3799 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3800 file. No doubt somebody will tell me if they don't make sense... Also added
3801 Dan Mooney's comments about building on OpenVMS.
3802
3803 19. Added support for partial matching via the PCRE_PARTIAL option for
3804 pcre_exec() and the \P data escape in pcretest.
3805
3806 20. Extended pcretest with 3 new pattern features:
3807
3808 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3809 write the compiled pattern to the file whose name is "rest-of-line".
3810 This is a straight binary dump of the data, with the saved pointer to
3811 the character tables forced to be NULL. The study data, if any, is
3812 written too. After writing, pcretest reads a new pattern.
3813
3814 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3815 compiled pattern from the given file. There must not be any
3816 occurrences of "<" in the file name (pretty unlikely); if there are,
3817 pcretest will instead treat the initial "<" as a pattern delimiter.
3818 After reading in the pattern, pcretest goes on to read data lines as
3819 usual.
3820
3821 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3822 and 16-bit fields in a compiled pattern, to simulate a pattern that
3823 was compiled on a host of opposite endianness.
3824
3825 21. The pcre-exec() function can now cope with patterns that were compiled on
3826 hosts of opposite endianness, with this restriction:
3827
3828 As for any compiled expression that is saved and used later, the tables
3829 pointer field cannot be preserved; the extra_data field in the arguments
3830 to pcre_exec() should be used to pass in a tables address if a value
3831 other than the default internal tables were used at compile time.
3832
3833 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3834 now diagnosed as an error. Previously, most of the time, a negative number
3835 would have been treated as zero, but if in addition "ovector" was passed as
3836 NULL, a crash could occur.
3837
3838 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3839 new versions from the libtool 1.5 distribution (the last one is a copy of
3840 a file called libtool.m4). This seems to have fixed the need to patch
3841 "configure" to support Darwin 1.3 (which I used to do). However, I still
3842 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3843 workstation).
3844
3845 24. Changed the PCRE licence to be the more standard "BSD" licence.
3846
3847
3848 Version 4.5 01-Dec-03
3849 ---------------------
3850
3851 1. There has been some re-arrangement of the code for the match() function so
3852 that it can be compiled in a version that does not call itself recursively.
3853 Instead, it keeps those local variables that need separate instances for
3854 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3855 needs to "recurse". Keeping track of where control must go is done by means
3856 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3857 hide most of the details from the main code, and operates only if
3858 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3859 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3860 operating.
3861
3862 To make it easier for callers to provide specially tailored get/free
3863 functions for this usage, two new functions, pcre_stack_malloc, and
3864 pcre_stack_free, are used. They are always called in strict stacking order,
3865 and the size of block requested is always the same.
3866
3867 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3868 PCRE has been compiled to use the stack or the heap for recursion. The
3869 -C option of pcretest uses this to show which version is compiled.
3870
3871 A new data escape \S, is added to pcretest; it causes the amounts of store
3872 obtained and freed by both kinds of malloc/free at match time to be added
3873 to the output.
3874
3875 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3876 what's available on my current Linux desktop machine.
3877
3878 3. When matching a UTF-8 string, the test for a valid string at the start has
3879 been extended. If start_offset is not zero, PCRE now checks that it points
3880 to a byte that is the start of a UTF-8 character. If not, it returns
3881 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3882 this is necessary because there may be backward assertions in the pattern.
3883 When matching the same subject several times, it may save resources to use
3884 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3885
3886 4. The code for checking the validity of UTF-8 strings has been tightened so
3887 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3888 containing "overlong sequences".
3889
3890 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3891 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3892 should have been "&", but it just so happened that all the cases this let
3893 through by mistake were picked up later in the function.
3894
3895 6. I had used a variable called "isblank" - this is a C99 function, causing
3896 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3897
3898 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3899 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3900 so the version is shown; (c) stop "make test" echoing "./RunTest".
3901
3902 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3903
3904 9. The prototype for memmove() for systems that don't have it was using
3905 size_t, but the inclusion of the header that defines size_t was later. I've
3906 moved the #includes for the C headers earlier to avoid this.
3907
3908 10. Added some adjustments to the code to make it easier to compiler on certain
3909 special systems:
3910
3911 (a) Some "const" qualifiers were missing.
3912 (b) Added the macro EXPORT before all exported functions; by default this
3913 is defined to be empty.
3914 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3915 that it reads its output file name as an argument instead of writing
3916 to the standard output and assuming this can be redirected.
3917
3918 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3919 class containing characters with values greater than 255, PCRE compilation
3920 went into a loop.
3921
3922 12. A recursive reference to a subpattern that was within another subpattern
3923 that had a minimum quantifier of zero caused PCRE to crash. For example,
3924 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3925 recursion. If the recursively-called subpattern itself had a zero repeat,
3926 that was OK.
3927
3928 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3929 buffer into which it was copied (for escape processing) was still set at
3930 1024, so long lines caused crashes.
3931
3932 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3933 "internal error: code overflow...". This applied to any character class
3934 that was followed by a possessive quantifier.
3935
3936 15. Modified the Makefile to add libpcre.la as a prerequisite for
3937 libpcreposix.la because I was told this is needed for a parallel build to
3938 work.
3939
3940 16. If a pattern that contained .* following optional items at the start was
3941 studied, the wrong optimizing data was generated, leading to matching
3942 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3943 matching string must start with a or b or c. The correct conclusion for
3944 this pattern is that a match can start with any character.
3945
3946
3947 Version 4.4 13-Aug-03
3948 ---------------------
3949
3950 1. In UTF-8 mode, a character class containing characters with values between
3951 127 and 255 was not handled correctly if the compiled pattern was studied.
3952 In fixing this, I have also improved the studying algorithm for such
3953 classes (slightly).
3954
3955 2. Three internal functions had redundant arguments passed to them. Removal
3956 might give a very teeny performance improvement.
3957
3958 3. Documentation bug: the value of the capture_top field in a callout is *one
3959 more than* the number of the hightest numbered captured substring.
3960
3961 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3962 in incorrectly linking with a previously installed version. They now link
3963 explicitly with libpcre.la.
3964
3965 5. configure.in no longer needs to recognize Cygwin specially.
3966
3967 6. A problem in pcre.in for Windows platforms is fixed.
3968
3969 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3970 pcretest, it used to include the size of the study block as part of its
3971 output. Unfortunately, the structure contains a field that has a different
3972 size on different hardware architectures. This meant that the tests that
3973 showed this size failed. As the block is currently always of a fixed size,
3974 this information isn't actually particularly useful in pcretest output, so
3975 I have just removed it.
3976
3977 8. Three pre-processor statements accidentally did not start in column 1.
3978 Sadly, there are *still* compilers around that complain, even though
3979 standard C has not required this for well over a decade. Sigh.
3980
3981 9. In pcretest, the code for checking callouts passed small integers in the
3982 callout_data field, which is a void * field. However, some picky compilers
3983 complained about the casts involved for this on 64-bit systems. Now
3984 pcretest passes the address of the small integer instead, which should get
3985 rid of the warnings.
3986
3987 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3988 both compile and run time, and gives an error if an invalid UTF-8 sequence
3989 is found. There is a option for disabling this check in cases where the
3990 string is known to be correct and/or the maximum performance is wanted.
3991
3992 11. In response to a bug report, I changed one line in Makefile.in from
3993
3994 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3995 to
3996 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3997
3998 to look similar to other lines, but I have no way of telling whether this
3999 is the right thing to do, as I do not use Windows. No doubt I'll get told
4000 if it's wrong...
4001
4002
4003 Version 4.3 21-May-03
4004 ---------------------
4005
4006 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4007 Makefile.
4008
4009 2. Some refactoring to improve the quality of the code:
4010
4011 (i) The utf8_table... variables are now declared "const".
4012
4013 (ii) The code for \cx, which used the "case flipping" table to upper case
4014 lower case letters, now just substracts 32. This is ASCII-specific,
4015 but the whole concept of \cx is ASCII-specific, so it seems
4016 reasonable.
4017
4018 (iii) PCRE was using its character types table to recognize decimal and
4019 hexadecimal digits in the pattern. This is silly, because it handles
4020 only 0-9, a-f, and A-F, but the character types table is locale-
4021 specific, which means strange things might happen. A private
4022 table is now used for this - though it costs 256 bytes, a table is
4023 much faster than multiple explicit tests. Of course, the standard
4024 character types table is still used for matching digits in subject
4025 strings against \d.
4026
4027 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4028 ending in _t are). So I've renamed it as ESC_tee.
4029
4030 3. The first argument for regexec() in the POSIX wrapper should have been
4031 defined as "const".
4032
4033 4. Changed pcretest to use malloc() for its buffers so that they can be
4034 Electric Fenced for debugging.
4035
4036 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4037 to read one or more bytes before the start of the subject string. Often this
4038 had no effect on PCRE's behaviour, but in some circumstances it could
4039 provoke a segmentation fault.
4040
4041 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4042 to try to read one or more bytes before the start of the subject string.
4043
4044 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4045 UTF-8 support could misbehave in various ways if the subject string
4046 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4047 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4048 back over UTF-8 characters.)
4049
4050
4051 Version 4.2 14-Apr-03
4052 ---------------------
4053
4054 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4055
4056 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4057 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4058 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4059 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4060 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4061 and BUILD_EXEEXT
4062 Note: automatic setting of the BUILD variables is not yet working
4063 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4064 compile-time but not at link-time
4065 [LINK]: use for linking executables only
4066 make different versions for Windows and non-Windows
4067 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4068 libraries
4069 [LINK_FOR_BUILD]: new variable
4070 [OBJEXT]: use throughout
4071 [EXEEXT]: use throughout
4072 <winshared>: new target
4073 <wininstall>: new target
4074 <dftables.o>: use native compiler
4075 <dftables>: use native linker
4076 <install>: handle Windows platform correctly
4077 <clean>: ditto
4078 <check>: ditto
4079 copy DLL to top builddir before testing
4080
4081 As part of these changes, -no-undefined was removed again. This was reported
4082 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4083 in any case.
4084
4085 3. Some tidies to get rid of compiler warnings:
4086
4087 . In the match_data structure, match_limit was an unsigned long int, whereas
4088 match_call_count was an int. I've made them both unsigned long ints.
4089
4090 . In pcretest the fact that a const uschar * doesn't automatically cast to
4091 a void * provoked a warning.
4092
4093 . Turning on some more compiler warnings threw up some "shadow" variables
4094 and a few more missing casts.
4095
4096 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4097 option, a class that contained a single character with a value between 128
4098 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4099
4100 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4101 option, a class that contained several characters, but with at least one
4102 whose value was between 128 and 255 caused PCRE to crash.
4103
4104
4105 Version 4.1 12-Mar-03
4106 ---------------------
4107
4108 1. Compiling with gcc -pedantic found a couple of places where casts were
4109 needed, and a string in dftables.c that was longer than standard compilers are
4110 required to support.
4111
4112 2. Compiling with Sun's compiler found a few more places where the code could
4113 be tidied up in order to avoid warnings.
4114
4115 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4116 first of these names is deprecated in the latest Autoconf in favour of the name
4117 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4118 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4119 analogy I have changed it to CFLAGS_FOR_BUILD.
4120
4121 4. Added -no-undefined to the linking command in the Makefile, because this is
4122 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4123 linking step for the pcreposix library.
4124
4125 5. PCRE was failing to diagnose the case of two named groups with the same
4126 name.
4127
4128 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4129 literal character that is needed in the subject for a match, and scans along to
4130 ensure that it is present before embarking on the full matching process. This
4131 saves time in cases of nested unlimited repeats that are never going to match.
4132 Problem: the scan can take a lot of time if the subject is very long (e.g.
4133 megabytes), thus penalizing straightforward matches. It is now done only if the
4134 amount of subject to be scanned is less than 1000 bytes.
4135
4136 7. A lesser problem with the same optimization is that it was recording the
4137 first character of an anchored pattern as "needed", thus provoking a search
4138 right along the subject, even when the first match of the pattern was going to
4139 fail. The "needed" character is now not set for anchored patterns, unless it
4140 follows something in the pattern that is of non-fixed length. Thus, it still
4141 fulfils its original purpose of finding quick non-matches in cases of nested
4142 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4143
4144
4145 Version 4.0 17-Feb-03
4146 ---------------------
4147
4148 1. If a comment in an extended regex that started immediately after a meta-item
4149 extended to the end of string, PCRE compiled incorrect data. This could lead to
4150 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4151
4152 2. Moved to autoconf 2.53 and libtool 1.4.2.
4153
4154 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4155 the special perltest8 script is no longer needed - all the tests can be run
4156 from a single perltest script.
4157
4158 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4159 by \s. It has now been removed in PCRE. This means it isn't recognized as
4160 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4161 class [:space:] *does* include VT, thereby creating a mess.
4162
4163 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4164 space and tab.
4165
4166 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4167 its new features into the main test script, reducing the number of scripts.
4168
4169 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4170 were backward compatible, and made the (?i) apply to the whole pattern, as if
4171 /i were given. Now it behaves more logically, and applies the option setting
4172 only to what follows. PCRE has been changed to follow suit. However, if it
4173 finds options settings right at the start of the pattern, it extracts them into
4174 the global options, as before. Thus, they show up in the info data.
4175
4176 8. Added support for the \Q...\E escape sequence. Characters in between are
4177 treated as literals. This is slightly different from Perl in that $ and @ are
4178 also handled as literals inside the quotes. In Perl, they will cause variable
4179 interpolation. Note the following examples:
4180
4181 Pattern PCRE matches Perl matches
4182
4183 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4184 \Qabc\$xyz\E abc\$xyz abc\$xyz
4185 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4186
4187 For compatibility with Perl, \Q...\E sequences are recognized inside character
4188 classes as well as outside them.
4189
4190 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4191 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4192 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4193 signed/unsigned warnings.
4194
4195 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4196 option for pcretest, so I've replaced it by a simple function that does just
4197 that job.
4198
4199 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4200 "pcregrep -".
4201
4202 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4203 Java package. This provides some syntactic sugar for simple cases of what my
4204 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4205 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4206 item, you can use this simplified notation. Note that only makes sense with
4207 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4208 greediness, whatever the setting of the PCRE_UNGREEDY option.
4209
4210 13. A change of greediness default within a pattern was not taking effect at
4211 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4212 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4213 was abstracted outside.
4214
4215 14. PCRE now supports the \G assertion. It is true when the current matching
4216 position is at the start point of the match. This differs from \A when the
4217 starting offset is non-zero. Used with the /g option of pcretest (or similar
4218 code), it works in the same way as it does for Perl's /g option. If all
4219 alternatives of a regex begin with \G, the expression is anchored to the start
4220 match position, and the "anchored" flag is set in the compiled expression.
4221
4222 15. Some bugs concerning the handling of certain option changes within patterns
4223 have been fixed. These applied to options other than (?ims). For example,
4224 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4225 been the other way round. Some of this was related to change 7 above.
4226
4227 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4228 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4229 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4230 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4231
4232 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4233 mode. Unlike ".", it always matches newline, whatever the setting of
4234 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4235 assertions. Perl allows it, but it doesn't (in general) work because it can't
4236 calculate the length of the lookbehind. At least, that's the case for Perl
4237 5.8.0 - I've been told they are going to document that it doesn't work in
4238 future.
4239
4240 18. Added an error diagnosis for escapes that PCRE does not support: these are
4241 \L, \l, \N, \P, \p, \U, \u, and \X.
4242
4243 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4244 reading past the end of the pattern in cases such as /[abcd/.
4245
4246 20. PCRE was getting more memory than necessary for patterns with classes that
4247 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4248
4249 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4250 compiling PCRE for use with Virtual Pascal.
4251
4252 22. Small fix to the Makefile to make it work properly if the build is done
4253 outside the source tree.
4254
4255 23. Added a new extension: a condition to go with recursion. If a conditional
4256 subpattern starts with (?(R) the "true" branch is used if recursion has
4257 happened, whereas the "false" branch is used only at the top level.
4258
4259 24. When there was a very long string of literal characters (over 255 bytes
4260 without UTF support, over 250 bytes with UTF support), the computation of how
4261 much memory was required could be incorrect, leading to segfaults or other
4262 strange effects.
4263
4264 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4265 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4266 there was a subsequent back reference to those brackets. This meant that, for
4267 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4268 possible to check for precisely this case. All we can do is abandon the
4269 optimization if .* occurs inside capturing brackets when there are any back
4270 references whatsoever. (See below for a better fix that came later.)
4271
4272 26. The handling of the optimization for finding the first character of a
4273 non-anchored pattern, and for finding a character that is required later in the
4274 match were failing in some cases. This didn't break the matching; it just
4275 failed to optimize when it could. The way this is done has been re-implemented.
4276
4277 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4278
4279 28. Added a new feature that provides some of the functionality that Perl
4280 provides with (?{...}). The facility is termed a "callout". The way it is done
4281 in PCRE is for the caller to provide an optional function, by setting
4282 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4283 global variable. By default it is unset, which disables all calling out. To get
4284 the function called, the regex must include (?C) at appropriate points. This
4285 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4286 This provides a means of identifying different callout points. When PCRE
4287 reaches such a point in the regex, if pcre_callout has been set, the external
4288 function is called. It is provided with data in a structure called
4289 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4290 matching continues; if it returns a non-zero value, the match at the current
4291 point fails. However, backtracking will occur if possible. [This was changed
4292 later and other features added - see item 49 below.]
4293
4294 29. pcretest is upgraded to test the callout functionality. It provides a
4295 callout function that displays information. By default, it shows the start of
4296 the match and the current position in the text. There are some new data escapes
4297 to vary what happens:
4298
4299 \C+ in addition, show current contents of captured substrings
4300 \C- do not supply a callout function
4301 \C!n return 1 when callout number n is reached
4302 \C!n!m return 1 when callout number n is reached for the mth time
4303
4304 30. If pcregrep was called with the -l option and just a single file name, it
4305 output "<stdin>" if a match was found, instead of the file name.
4306
4307 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4308 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4309 pcre_exec(). This saves a malloc/free per call. The default value of
4310 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4311 when configuring.
4312
4313 32. The default maximum size of a compiled pattern is 64K. There have been a
4314 few cases of people hitting this limit. The code now uses macros to handle the
4315 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4316 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4317 configuring. Tests 2 and 5 work only with 2-byte links because they output
4318 debugging information about compiled patterns.
4319
4320 33. Internal code re-arrangements:
4321
4322 (a) Moved the debugging function for printing out a compiled regex into
4323 its own source file (printint.c) and used #include to pull it into
4324 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4325 separate copies.
4326
4327 (b) Defined the list of op-code names for debugging as a macro in
4328 internal.h so that it is next to the definition of the opcodes.
4329
4330 (c) Defined a table of op-code lengths for simpler skipping along compiled
4331 code. This is again a macro in internal.h so that it is next to the
4332 definition of the opcodes.
4333
4334 34. Added support for recursive calls to individual subpatterns, along the
4335 lines of Robin Houston's patch (but implemented somewhat differently).
4336
4337 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4338 allow it to read and process whole directories in Win32. This code was
4339 contributed by Lionel Fourquaux; it has not been tested by me.
4340
4341 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4342 used to name a group. Names consist of alphanumerics and underscores, and must
4343 be unique. Back references use the syntax (?P=name) and recursive calls use
4344 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4345 numbers. The function pcre_fullinfo() can be used after compilation to extract
4346 a name/number map. There are three relevant calls:
4347
4348 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4349 PCRE_INFO_NAMECOUNT yields the number of entries
4350 PCRE_INFO_NAMETABLE yields a pointer to the map.
4351
4352 The map is a vector of fixed-size entries. The size of each entry depends on
4353 the length of the longest name used. The first two bytes of each entry are the
4354 group number, most significant byte first. There follows the corresponding
4355 name, zero terminated. The names are in alphabetical order.
4356
4357 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4358 case instead of 255. Making it the same both with and without UTF-8 support
4359 means that the same test output works with both.
4360
4361 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4362 calling malloc() with a zero argument.
4363
4364 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4365 optimization. I've improved things by keeping a bitmap of backreferences with
4366 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4367 fact referenced, the optimization can be applied. It is unlikely that a
4368 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4369 the match to follow \n) will appear inside brackets with a number greater than
4370 31, but if it does, any back reference > 31 suppresses the optimization.
4371
4372 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4373 of disabling numbered capturing parentheses. Any opening parenthesis that is
4374 not followed by ? behaves as if it were followed by ?: but named parentheses
4375 can still be used for capturing (and they will acquire numbers in the usual
4376 way).
4377
4378 41. Redesigned the return codes from the match() function into yes/no/error so
4379 that errors can be passed back from deep inside the nested calls. A malloc
4380 failure while inside a recursive subpattern call now causes the
4381 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4382
4383 42. It is now possible to set a limit on the number of times the match()
4384 function is called in a call to pcre_exec(). This facility makes it possible to
4385 limit the amount of recursion and backtracking, though not in a directly
4386 obvious way, because the match() function is used in a number of different
4387 circumstances. The count starts from zero for each position in the subject
4388 string (for non-anchored patterns). The default limit is, for compatibility, a
4389 large number, namely 10 000 000. You can change this in two ways:
4390
4391 (a) When configuring PCRE before making, you can use --with-match-limit=n
4392 to set a default value for the compiled library.
4393
4394 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4395 a different value is set. See 45 below.
4396
4397 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4398
4399 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4400 of things that can be changed at compile time. The first argument specifies
4401 what is wanted and the second points to where the information is to be placed.
4402 The current list of available information is:
4403
4404 PCRE_CONFIG_UTF8
4405
4406 The output is an integer that is set to one if UTF-8 support is available;
4407 otherwise it is set to zero.
4408
4409 PCRE_CONFIG_NEWLINE
4410
4411 The output is an integer that it set to the value of the code that is used for
4412 newline. It is either LF (10) or CR (13).
4413
4414 PCRE_CONFIG_LINK_SIZE
4415
4416 The output is an integer that contains the number of bytes used for internal
4417 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4418
4419 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4420
4421 The output is an integer that contains the threshold above which the POSIX
4422 interface uses malloc() for output vectors. See item 31 above.
4423
4424 PCRE_CONFIG_MATCH_LIMIT
4425
4426 The output is an unsigned integer that contains the default limit of the number
4427 of match() calls in a pcre_exec() execution. See 42 above.
4428
4429 44. pcretest has been upgraded by the addition of the -C option. This causes it
4430 to extract all the available output from the new pcre_config() function, and to
4431 output it. The program then exits immediately.
4432
4433 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4434 order to support additional features. One way would have been to define
4435 pcre_exec2() (for example) with extra arguments, but this would not have been
4436 extensible, and would also have required all calls to the original function to
4437 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4438 is used for passing in "extra" data from pcre_study().
4439
4440 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4441 contains the following fields:
4442
4443 flags a bitmap indicating which of the following fields are set
4444 study_data opaque data from pcre_study()
4445 match_limit a way of specifying a limit on match() calls for a specific
4446 call to pcre_exec()
4447 callout_data data for callouts (see 49 below)
4448
4449 The flag bits are also defined in pcre.h, and are
4450
4451 PCRE_EXTRA_STUDY_DATA
4452 PCRE_EXTRA_MATCH_LIMIT
4453 PCRE_EXTRA_CALLOUT_DATA
4454
4455 The pcre_study() function now returns one of these new pcre_extra blocks, with
4456 the actual study data pointed to by the study_data field, and the
4457 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4458 before. That is, this change is entirely upwards-compatible and requires no
4459 change to existing code.
4460
4461 If you want to pass in additional data to pcre_exec(), you can either place it
4462 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4463 block.
4464
4465 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4466 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4467 times with different match limits, until it finds the minimum value needed for
4468 pcre_exec() to complete. The value is then output. This can be instructive; for
4469 most simple matches the number is quite small, but for pathological cases it
4470 gets very large very quickly.
4471
4472 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4473 returns the size of the data block pointed to by the study_data field in a
4474 pcre_extra block, that is, the value that was passed as the argument to
4475 pcre_malloc() when PCRE was getting memory in which to place the information
4476 created by pcre_study(). The fourth argument should point to a size_t variable.
4477 pcretest has been extended so that this information is shown after a successful
4478 pcre_study() call when information about the compiled regex is being displayed.
4479
4480 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4481 because what follows is always an absolute path. (Later: it turns out that this
4482 is more than cosmetic for MinGW, because it doesn't like empty path
4483 components.)
4484
4485 49. Some changes have been made to the callout feature (see 28 above):
4486
4487 (i) A callout function now has three choices for what it returns:
4488
4489 0 => success, carry on matching
4490 > 0 => failure at this point, but backtrack if possible
4491 < 0 => serious error, return this value from pcre_exec()
4492
4493 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4494 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4495 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4496 use by callout functions. It will never be used by PCRE itself.
4497
4498 (ii) The pcre_extra structure (see 45 above) has a void * field called
4499 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4500 pcre_callout_block structure has a field of the same name. The contents of
4501 the field passed in the pcre_extra structure are passed to the callout
4502 function in the corresponding field in the callout block. This makes it
4503 easier to use the same callout-containing regex from multiple threads. For
4504 testing, the pcretest program has a new data escape
4505
4506 \C*n pass the number n (may be negative) as callout_data
4507
4508 If the callout function in pcretest receives a non-zero value as
4509 callout_data, it returns that value.
4510
4511 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4512 there were some redundant $(CFLAGS) in commands that are now specified as
4513 $(LINK), which already includes $(CFLAGS).
4514
4515 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4516 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4517 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4518 one-byte characters throughout. Note that case-insensitive matching applies
4519 only to characters whose values are less than 256. PCRE doesn't support the
4520 notion of cases for higher-valued characters.
4521
4522 (i) A character class whose characters are all within 0-255 is handled as
4523 a bit map, and the map is inverted for negative classes. Previously, a
4524 character > 255 always failed to match such a class; however it should
4525 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4526
4527 (ii) A negated character class with a single character < 255 is coded as
4528 "not this character" (OP_NOT). This wasn't working properly when the test
4529 character was multibyte, either singly or repeated.
4530
4531 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4532 mode, for example: \x{100}{2,3}.
4533
4534 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4535 singly or repeated) now correctly test multibyte characters. However,
4536 PCRE doesn't recognize any characters with values greater than 255 as
4537 digits, spaces, or word characters. Such characters always match \D, \S,
4538 and \W, and never match \d, \s, or \w.
4539
4540 (v) Classes may now contain characters and character ranges with values
4541 greater than 255. For example: [ab\x{100}-\x{400}].
4542
4543 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4544 PCRE in UTF-8 mode.
4545
4546 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4547 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4548 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4549 value.)
4550
4551 53. The single man page has become too large. I have therefore split it up into
4552 a number of separate man pages. These also give rise to individual HTML pages;
4553 these are now put in a separate directory, and there is an index.html page that
4554 lists them all. Some hyperlinking between the pages has been installed.
4555
4556 54. Added convenience functions for handling named capturing parentheses.
4557
4558 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4559 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4560 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4561 are faulted.
4562
4563 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4564 calling configure. These values are used when compiling the dftables.c program
4565 which is run to generate the source of the default character tables. They
4566 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4567 you will need to set these values.
4568
4569 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4570
4571
4572 Version 3.9 02-Jan-02
4573 ---------------------
4574
4575 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4576
4577 2. If --disable-static was given, the building process failed when trying to
4578 build pcretest and pcregrep. (For some reason it was using libtool to compile
4579 them, which is not right, as they aren't part of the library.)
4580
4581
4582 Version 3.8 18-Dec-01
4583 ---------------------
4584
4585 1. The experimental UTF-8 code was completely screwed up. It was packing the
4586 bytes in the wrong order. How dumb can you get?
4587
4588
4589 Version 3.7 29-Oct-01
4590 ---------------------
4591
4592 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4593 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4594 this didn't happen under Solaris 8, where I normally test things.
4595
4596 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4597 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4598 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4599 this fix an hour or so after the initial 3.7 release.)
4600
4601
4602 Version 3.6 23-Oct-01
4603 ---------------------
4604
4605 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4606 offsets passed as NULL with zero offset count.
4607
4608 2. The config.guess and config.sub files had not been updated when I moved to
4609 the latest autoconf.
4610
4611
4612 Version 3.5 15-Aug-01
4613 ---------------------
4614
4615 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4616 had been forgotten.
4617
4618 2. By using declared but undefined structures, we can avoid using "void"
4619 definitions in pcre.h while keeping the internal definitions of the structures
4620 private.
4621
4622 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4623 user point of view, this means that both static and shared libraries are built
4624 by default, but this can be individually controlled. More of the work of
4625 handling this static/shared cases is now inside libtool instead of PCRE's make
4626 file.
4627
4628 4. The pcretest utility is now installed along with pcregrep because it is
4629 useful for users (to test regexs) and by doing this, it automatically gets
4630 relinked by libtool. The documentation has been turned into a man page, so
4631 there are now .1, .txt, and .html versions in /doc.
4632
4633 5. Upgrades to pcregrep:
4634 (i) Added long-form option names like gnu grep.
4635 (ii) Added --help to list all options with an explanatory phrase.
4636 (iii) Added -r, --recursive to recurse into sub-directories.
4637 (iv) Added -f, --file to read patterns from a file.
4638
4639 6. pcre_exec() was referring to its "code" argument before testing that
4640 argument for NULL (and giving an error if it was NULL).
4641
4642 7. Upgraded Makefile.in to allow for compiling in a different directory from
4643 the source directory.
4644
4645 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4646 options bits, the pointer it was passed was to an int instead of to an unsigned
4647 long int. This mattered only on 64-bit systems.
4648
4649 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4650 generated) instead of pcre.in, which it its source. Also made the same change
4651 in several of the .c files.
4652
4653 10. A new release of gcc defines printf() as a macro, which broke pcretest
4654 because it had an ifdef in the middle of a string argument for printf(). Fixed
4655 by using separate calls to printf().
4656
4657 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4658 script, to force use of CR or LF instead of \n in the source. On non-Unix
4659 systems, the value can be set in config.h.
4660
4661 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4662 absolute limit. Changed the text of the error message to make this clear, and
4663 likewise updated the man page.
4664
4665 13. The limit of 99 on the number of capturing subpatterns has been removed.
4666 The new limit is 65535, which I hope will not be a "real" limit.
4667
4668
4669 Version 3.4 22-Aug-00
4670 ---------------------
4671
4672 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4673
4674 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4675
4676
4677 Version 3.3 01-Aug-00
4678 ---------------------
4679
4680 1. If an octal character was given, but the value was greater than \377, it
4681 was not getting masked to the least significant bits, as documented. This could
4682 lead to crashes in some systems.
4683
4684 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4685 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4686
4687 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4688 These just pass their arguments on to (pcre_free)(), but they are provided
4689 because some uses of PCRE bind it to non-C systems that can call its functions,
4690 but cannot call free() or pcre_free() directly.
4691
4692 4. Add "make test" as a synonym for "make check". Corrected some comments in
4693 the Makefile.
4694
4695 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4696 Makefile.
4697
4698 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4699 command called pgrep for grepping around the active processes.
4700
4701 7. Added the beginnings of support for UTF-8 character strings.
4702
4703 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4704 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4705 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4706 out for the ar command.)
4707
4708
4709 Version 3.2 12-May-00
4710 ---------------------
4711
4712 This is purely a bug fixing release.
4713
4714 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4715 of ZA. This was just one example of several cases that could provoke this bug,
4716 which was introduced by change 9 of version 2.00. The code for breaking
4717 infinite loops after an iteration that matches an empty string was't working
4718 correctly.
4719
4720 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4721 when matched against abbab (for example). After matching an empty string, it
4722 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4723 caused it to match further down the string than it should.
4724
4725 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4726 was there because it doesn't seem to be needed, and it causes trouble on some
4727 systems, as it is not a Standard C header. It has been removed.
4728
4729 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4730 were reported on the Macintosh. The changes were from
4731
4732 while ((c = *(++ptr)) != 0 && c != '\n');
4733 to
4734 while ((c = *(++ptr)) != 0 && c != '\n') ;
4735
4736 Totally extraordinary, but if that's what it takes...
4737
4738 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4739 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4740 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4741 assumes the way PCRE uses memmove() (always moving upwards).
4742
4743 6. PCRE is being used in one environment where strchr() is not available. There
4744 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4745 faster code anyway.
4746
4747
4748 Version 3.1 09-Feb-00
4749 ---------------------
4750
4751 The only change in this release is the fixing of some bugs in Makefile.in for
4752 the "install" target:
4753
4754 (1) It was failing to install pcreposix.h.
4755
4756 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4757
4758
4759 Version 3.0 01-Feb-00
4760 ---------------------
4761
4762 1. Add support for the /+ modifier to perltest (to output $` like it does in
4763 pcretest).
4764
4765 2. Add support for the /g modifier to perltest.
4766
4767 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4768 matches null strings.
4769
4770 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4771 pattern. Perl treats empty patterns specially - it reuses the most recent
4772 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4773 effect.
4774
4775 5. The POSIX interface was broken in that it was just handing over the POSIX
4776 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4777 required a bigger vector, with some working space on the end. This means that
4778 the POSIX wrapper now has to get and free some memory, and copy the results.
4779
4780 6. Added some simple autoconf support, placing the test data and the
4781 documentation in separate directories, re-organizing some of the
4782 information files, and making it build pcre-config (a GNU standard). Also added
4783 libtool support for building PCRE as a shared library, which is now the
4784 default.
4785
4786 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4787 09 are not valid octal constants. Single digits will be used for minor values
4788 less than 10.
4789
4790 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4791 existing programs that set these in the POSIX interface can use PCRE without
4792 modification.
4793
4794 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4795 return all that pcre_info() returns, plus additional data. The pcre_info()
4796 function is retained for compatibility, but is considered to be obsolete.
4797
4798 10. Added experimental recursion feature (?R) to handle one common case that
4799 Perl 5.6 will be able to do with (?p{...}).
4800
4801 11. Added support for POSIX character classes like [:alpha:], which Perl is
4802 adopting.
4803
4804
4805 Version 2.08 31-Aug-99
4806 ----------------------
4807
4808 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4809 trying to match at the startoffset position, but instead was moving forward to
4810 the next newline as if a previous match had failed.
4811
4812 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4813 and could get into a loop if a null string was matched other than at the start
4814 of the subject.
4815
4816 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4817 be distinguished at compile time, and for completeness also added PCRE_DATE.
4818
4819 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4820 in GnuWin32 environments.
4821
4822
4823 Version 2.07 29-Jul-99
4824 ----------------------
4825
4826 1. The documentation is now supplied in plain text form and HTML as well as in
4827 the form of man page sources.
4828
4829 2. C++ compilers don't like assigning (void *) values to other pointer types.
4830 In particular this affects malloc(). Although there is no problem in Standard
4831 C, I've put in casts to keep C++ compilers happy.
4832
4833 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4834 should be (const char *).
4835
4836 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4837 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4838 However, I haven't made this a standard facility. The documentation doesn't
4839 mention it, and the Makefile doesn't support it.
4840
4841 5. The Makefile now contains an "install" target, with editable destinations at
4842 the top of the file. The pcretest program is not installed.
4843
4844 6. pgrep -V now gives the PCRE version number and date.
4845
4846 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4847 causing the entire string to be ignored, instead of just the last character.
4848
4849 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4850 non-matching string, it can take a very, very long time, even for strings of
4851 quite modest length, because of the nested recursion. PCRE now does better in
4852 some of these cases. It does this by remembering the last required literal
4853 character in the pattern, and pre-searching the subject to ensure it is present
4854 before running the real match. In other words, it applies a heuristic to detect
4855 some types of certain failure quickly, and in the above example, if presented
4856 with a string that has no trailing " it gives "no match" very quickly.
4857
4858 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4859 other alternatives are tried instead.
4860
4861
4862 Version 2.06 09-Jun-99
4863 ----------------------
4864
4865 1. Change pcretest's output for amount of store used to show just the code
4866 space, because the remainder (the data block) varies in size between 32-bit and
4867 64-bit systems.
4868
4869 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4870 start matching at. This allows lookbehinds to work when searching for multiple
4871 occurrences in a string.
4872
4873 3. Added additional options to pcretest for testing multiple occurrences:
4874
4875 /+ outputs the rest of the string that follows a match
4876 /g loops for multiple occurrences, using the new startoffset argument
4877 /G loops for multiple occurrences by passing an incremented pointer
4878
4879 4. PCRE wasn't doing the "first character" optimization for patterns starting
4880 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4881 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4882 the letter 'x'. On long subject strings, this gives a significant speed-up.
4883
4884
4885 Version 2.05 21-Apr-99
4886 ----------------------
4887
4888 1. Changed the type of magic_number from int to long int so that it works
4889 properly on 16-bit systems.
4890
4891 2. Fixed a bug which caused patterns starting with .* not to work correctly
4892 when the subject string contained newline characters. PCRE was assuming
4893 anchoring for such patterns in all cases, which is not correct because .* will
4894 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4895 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4896 must be retried after every newline in the subject.
4897
4898
4899 Version 2.04 18-Feb-99
4900 ----------------------
4901
4902 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4903 computation of the store needed to hold the pattern was incorrect (too large).
4904 If such patterns were nested a few deep, this could multiply and become a real
4905 problem.
4906
4907 2. Added /M option to pcretest to show the memory requirement of a specific
4908 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4909
4910 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4911 compiled in such a way that the backtracking after subsequent failure was
4912 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4913 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4914
4915
4916 Version 2.03 02-Feb-99
4917 ----------------------
4918
4919 1. Fixed typo and small mistake in man page.
4920
4921 2. Added 4th condition (GPL supersedes if conflict) and created separate
4922 LICENCE file containing the conditions.
4923
4924 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4925 Perl, that is the internal \ allows the delimiter to be included in the
4926 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4927 the final delimiter, add \ to the end of the pattern (to test the error).
4928
4929 4. Added the convenience functions for extracting substrings after a successful
4930 match. Updated pcretest to make it able to test these functions.
4931
4932
4933 Version 2.02 14-Jan-99
4934 ----------------------
4935
4936 1. Initialized the working variables associated with each extraction so that
4937 their saving and restoring doesn't refer to uninitialized store.
4938
4939 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4940 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4941 fix the problem.
4942
4943 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4944 calls, and wasn't printing the correct value for compiling calls. Increased the
4945 default value of LOOPREPEAT, and the number of significant figures in the
4946 times.
4947
4948 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4949
4950 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4951 a building problem on Windows NT with a FAT file system.
4952
4953
4954 Version 2.01 21-Oct-98
4955 ----------------------
4956
4957 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4958 to character tables built by pcre_maketables() in the current locale. If NULL
4959 is passed, the default tables are used.
4960
4961
4962 Version 2.00 24-Sep-98
4963 ----------------------
4964
4965 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4966 it any more.
4967
4968 2. Allow quantification of (?>) groups, and make it work correctly.
4969
4970 3. The first character computation wasn't working for (?>) groups.
4971
4972 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4973 end of the subject) and add 5.005's \z, which really does match only at the
4974 very end of the subject.
4975
4976 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4977
4978 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4979 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4980 localized options. All options to pcre_study() were also removed.
4981
4982 7. Add other new features from 5.005:
4983
4984 $(?<= positive lookbehind
4985 $(?<! negative lookbehind
4986 (?imsx-imsx) added the unsetting capability
4987 such a setting is global if at outer level; local otherwise
4988 (?imsx-imsx:) non-capturing groups with option setting
4989 (?(cond)re|re) conditional pattern matching
4990
4991 A backreference to itself in a repeated group matches the previous
4992 captured string.
4993
4994 8. General tidying up of studying (both automatic and via "study")
4995 consequential on the addition of new assertions.
4996
4997 9. As in 5.005, unlimited repeated groups that could match an empty substring
4998 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4999 runtime if any iteration does actually match an empty substring.
5000
5001 10. Include the RunTest script in the distribution.
5002
5003 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5004 discrepancies, some of which were old and were also with respect to 5.004. They
5005 have now been fixed.
5006
5007
5008 Version 1.09 28-Apr-98
5009 ----------------------
5010
5011 1. A negated single character class followed by a quantifier with a minimum
5012 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5013 program crashes, or just wrong answers. This did not apply to negated classes
5014 containing more than one character, or to minima other than one.
5015
5016
5017 Version 1.08 27-Mar-98
5018 ----------------------
5019
5020 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5021
5022 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5023 latter must appear before anything that relies on it in the pattern.
5024
5025
5026 Version 1.07 16-Feb-98
5027 ----------------------
5028
5029 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5030 repeat of a potentially empty string).
5031
5032
5033 Version 1.06 23-Jan-98
5034 ----------------------
5035
5036 1. Added Markus Oberhumer's little patches for C++.
5037
5038 2. Literal strings longer than 255 characters were broken.
5039
5040
5041 Version 1.05 23-Dec-97
5042 ----------------------
5043
5044 1. Negated character classes containing more than one character were failing if
5045 PCRE_CASELESS was set at run time.
5046
5047
5048 Version 1.04 19-Dec-97
5049 ----------------------
5050
5051 1. Corrected the man page, where some "const" qualifiers had been omitted.
5052
5053 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5054 input syntax.
5055
5056 3. Fixed memory leak which occurred when a regex with back references was
5057 matched with an offsets vector that wasn't big enough. The temporary memory
5058 that is used in this case wasn't being freed if the match failed.
5059
5060 4. Tidied pcretest to ensure it frees memory that it gets.
5061
5062 5. Temporary memory was being obtained in the case where the passed offsets
5063 vector was exactly big enough.
5064
5065 6. Corrected definition of offsetof() from change 5 below.
5066
5067 7. I had screwed up change 6 below and broken the rules for the use of
5068 setjmp(). Now fixed.
5069
5070
5071 Version 1.03 18-Dec-97
5072 ----------------------
5073
5074 1. A erroneous regex with a missing opening parenthesis was correctly
5075 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5076 on some systems.
5077
5078 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5079 it was reported that one broken compiler failed on the former because "code" is
5080 also an independent variable.
5081
5082 3. The erroneous regex a[]b caused an array overrun reference.
5083
5084 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5085 fail on data ending with that character. (It was going on too far, and checking
5086 the next character, typically a binary zero.) This was specific to the
5087 optimized code for single-character negative classes.
5088
5089 5. Added a contributed patch from the TIN world which does the following:
5090
5091 + Add an undef for memmove, in case the the system defines a macro for it.
5092
5093 + Add a definition of offsetof(), in case there isn't one. (I don't know
5094 the reason behind this - offsetof() is part of the ANSI standard - but
5095 it does no harm).
5096
5097 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5098 most of the places where whitespace preceded '#'. I have given up and
5099 allowed the remaining 2 cases to be at the margin.
5100
5101 + Rename some variables in pcre to eliminate shadowing. This seems very
5102 pedantic, but does no harm, of course.
5103
5104 6. Moved the call to setjmp() into its own function, to get rid of warnings
5105 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5106
5107 7. Constructs such as \d{8,} were compiling into the equivalent of
5108 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5109 outcome, but in this particular case used more store than had been allocated,
5110 which caused the bug to be discovered because it threw up an internal error.
5111
5112 8. The debugging code in both pcre and pcretest for outputting the compiled
5113 form of a regex was going wrong in the case of back references followed by
5114 curly-bracketed repeats.
5115
5116
5117 Version 1.02 12-Dec-97
5118 ----------------------
5119
5120 1. Typos in pcre.3 and comments in the source fixed.
5121
5122 2. Applied a contributed patch to get rid of places where it used to remove
5123 'const' from variables, and fixed some signed/unsigned and uninitialized
5124 variable warnings.
5125
5126 3. Added the "runtest" target to Makefile.
5127
5128 4. Set default compiler flag to -O2 rather than just -O.
5129
5130
5131 Version 1.01 19-Nov-97
5132 ----------------------
5133
5134 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5135 like /([ab]*)*/, that is, for classes with more than one character in them.
5136
5137 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5138 as /((?>a*))*/ (a PCRE_EXTRA facility).
5139
5140
5141 Version 1.00 18-Nov-97
5142 ----------------------
5143
5144 1. Added compile-time macros to support systems such as SunOS4 which don't have
5145 memmove() or strerror() but have other things that can be used instead.
5146
5147 2. Arranged that "make clean" removes the executables.
5148
5149
5150 Version 0.99 27-Oct-97
5151 ----------------------
5152
5153 1. Fixed bug in code for optimizing classes with only one character. It was
5154 initializing a 32-byte map regardless, which could cause it to run off the end
5155 of the memory it had got.
5156
5157 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5158
5159
5160 Version 0.98 22-Oct-97
5161 ----------------------
5162
5163 1. Fixed bug in code for handling temporary memory usage when there are more
5164 back references than supplied space in the ovector. This could cause segfaults.
5165
5166
5167 Version 0.97 21-Oct-97
5168 ----------------------
5169
5170 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5171
5172 2. Optimized negated single characters not to use a bit map.
5173
5174 3. Brought error texts together as macro definitions; clarified some of them;
5175 fixed one that was wrong - it said "range out of order" when it meant "invalid
5176 escape sequence".
5177
5178 4. Changed some char * arguments to const char *.
5179
5180 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5181
5182 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5183 pcretest.
5184
5185
5186 Version 0.96 16-Oct-97
5187 ----------------------
5188
5189 1. Added a simple "pgrep" utility to the distribution.
5190
5191 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5192 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5193 where "ddd" means "one or more decimal digits".
5194
5195 3. Fixed serious bug. If a pattern had a back reference, but the call to
5196 pcre_exec() didn't supply a large enough ovector to record the related
5197 identifying subpattern, the match always failed. PCRE now remembers the number
5198 of the largest back reference, and gets some temporary memory in which to save
5199 the offsets during matching if necessary, in order to ensure that
5200 backreferences always work.
5201
5202 4. Increased the compatibility with Perl in a number of ways:
5203
5204 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5205 to request this handling. The option can be set at compile or exec time.
5206
5207 (b) $ matches before a terminating newline by default; an option
5208 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5209 mode). The option can be set at compile or exec time.
5210
5211 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5212 the same as Perl's. If the decimal number it represents is less than 10
5213 or there aren't that many previous left capturing parentheses, an octal
5214 escape is read. Inside a character class, it's always an octal escape,
5215 even if it is a single digit.
5216
5217 (d) An escaped but undefined alphabetic character is taken as a literal,
5218 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5219 escapes.
5220
5221 (e) {0} is now permitted. (The previous item is removed from the compiled
5222 pattern).
5223
5224 5. Changed all the names of code files so that the basic parts are no longer
5225 than 10 characters, and abolished the teeny "globals.c" file.
5226
5227 6. Changed the handling of character classes; they are now done with a 32-byte
5228 bit map always.
5229
5230 7. Added the -d and /D options to pcretest to make it possible to look at the
5231 internals of compilation without having to recompile pcre.
5232
5233
5234 Version 0.95 23-Sep-97
5235 ----------------------
5236
5237 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5238 \x20 at the start of a run of normal characters. These were being treated as
5239 real characters, instead of the source characters being re-checked.
5240
5241
5242 Version 0.94 18-Sep-97
5243 ----------------------
5244
5245 1. The functions are now thread-safe, with the caveat that the global variables
5246 containing pointers to malloc() and free() or alternative functions are the
5247 same for all threads.
5248
5249 2. Get pcre_study() to generate a bitmap of initial characters for non-
5250 anchored patterns when this is possible, and use it if passed to pcre_exec().
5251
5252
5253 Version 0.93 15-Sep-97
5254 ----------------------
5255
5256 1. /(b)|(:+)/ was computing an incorrect first character.
5257
5258 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5259 but not actually doing anything yet.
5260
5261 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5262 as Perl does (e.g. [-az] or [az-]).
5263
5264 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5265 all possible positions.
5266
5267 5. Split up into different modules to avoid including unneeded functions in a
5268 compiled binary. However, compile and exec are still in one module. The "study"
5269 function is split off.
5270
5271 6. The character tables are now in a separate module whose source is generated
5272 by an auxiliary program - but can then be edited by hand if required. There are
5273 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5274 toupper() in the code.
5275
5276 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5277 make them global. Abolish the function for setting them, as the caller can now
5278 set them directly.
5279
5280
5281 Version 0.92 11-Sep-97
5282 ----------------------
5283
5284 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5285 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5286
5287 2. Caseless matching was not working in character classes if the characters in
5288 the pattern were in upper case.
5289
5290 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5291
5292 4. Make PCRE_ANCHORED public and accept as a compile option.
5293
5294 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5295 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5296 pass them.
5297
5298 6. Give an error if bad option bits passed at compile or run time.
5299
5300 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5301 pcretest to cause it to pass that flag.
5302
5303 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5304 options, and the first character, if set.
5305
5306 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5307
5308
5309 Version 0.91 10-Sep-97
5310 ----------------------
5311
5312 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5313 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5314
5315 2. PCRE was looping on encountering an indefinitely repeated back reference to
5316 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5317 Perl does - treats the match as successful.
5318
5319 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5