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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1387 - (show annotations)
Sat Nov 2 18:29:05 2013 UTC (6 years, 1 month ago) by ph10
File size: 255441 byte(s)
Update POSIX class handling in UCP mode.
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
156 Version 8.33 28-May-2013
157 ------------------------
158
159 1. Added 'U' to some constants that are compared to unsigned integers, to
160 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
161 variables that are added to signed variables, to ensure the result is
162 signed and can be negated.
163
164 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
165 CMake config files.
166
167 3. Revise the creation of config.h.generic so that all boolean macros are
168 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
169 overriding via -D on the command line possible.
170
171 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
172 to unsigned int is reported to make a quite noticeable speed difference in
173 a specific Windows environment. Testing on Linux did also appear to show
174 some benefit (and it is clearly not harmful). Also fixed the definition of
175 Xop which should be unsigned.
176
177 5. Related to (4), changing the definition of the intermediate variable cc
178 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
179 improvements.
180
181 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
182 unnecessary spaces.
183
184 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
185 and later.
186
187 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
188
189 9. Optimizing fast_forward_start_bits in JIT.
190
191 10. Adding support for callouts in JIT, and fixing some issues revealed
192 during this work. Namely:
193
194 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
195
196 (b) Minimum length was not checked before the matching is started.
197
198 11. The value of capture_last that is passed to callouts was incorrect in some
199 cases when there was a capture on one path that was subsequently abandoned
200 after a backtrack. Also, the capture_last value is now reset after a
201 recursion, since all captures are also reset in this case.
202
203 12. The interpreter no longer returns the "too many substrings" error in the
204 case when an overflowing capture is in a branch that is subsequently
205 abandoned after a backtrack.
206
207 13. In the pathological case when an offset vector of size 2 is used, pcretest
208 now prints out the matched string after a yield of 0 or 1.
209
210 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
211 Only supported by the JIT compiler at the moment.
212
213 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
214
215 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
216 offset of the starting point of the matching process, provided the offsets
217 vector is large enough.
218
219 17. The \A escape now records a lookbehind value of 1, though its execution
220 does not actually inspect the previous character. This is to ensure that,
221 in partial multi-segment matching, at least one character from the old
222 segment is retained when a new segment is processed. Otherwise, if there
223 are no lookbehinds in the pattern, \A might match incorrectly at the start
224 of a new segment.
225
226 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
227
228 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
229 gives some modest performance improvement in 8-bit mode.
230
231 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
232 be expressed in certain programming languages using Universal Character
233 Names.
234
235 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
236 which points out that "non characters" are not "characters that may not
237 appear in Unicode strings" but rather "characters that are reserved for
238 internal use and have only local meaning".
239
240 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
241 there was a conditional group that depended on an assertion, if the
242 assertion was false, the callout that immediately followed the alternation
243 in the condition was skipped when pcre_exec() was used for matching.
244
245 23. Allow an explicit callout to be inserted before an assertion that is the
246 condition for a conditional group, for compatibility with automatic
247 callouts, which always insert a callout at this point.
248
249 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
250 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
251
252 25. (*PRUNE) is now supported by the JIT compiler.
253
254 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
255
256 27. Fix the case where there are two or more SKIPs with arguments that may be
257 ignored.
258
259 28. (*SKIP) is now supported by the JIT compiler.
260
261 29. (*THEN) is now supported by the JIT compiler.
262
263 30. Update RunTest with additional test selector options.
264
265 31. The way PCRE handles backtracking verbs has been changed in two ways.
266
267 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
268 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
269 backtracking. In some cases this makes it more Perl-compatible, but Perl's
270 rather obscure rules do not always do the same thing.
271
272 (2) Previously, backtracking verbs were confined within assertions. This is
273 no longer the case for positive assertions, except for (*ACCEPT). Again,
274 this sometimes improves Perl compatibility, and sometimes does not.
275
276 32. A number of tests that were in test 2 because Perl did things differently
277 have been moved to test 1, because either Perl or PCRE has changed, and
278 these tests are now compatible.
279
280 32. Backtracking control verbs are now handled in the same way in JIT and
281 interpreter.
282
283 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
284 contained a forward subroutine reference caused a compile error.
285
286 34. Auto-detect and optimize limited repetitions in JIT.
287
288 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
289 blocking (*UTF) etc.
290
291 36. In the interpreter, maximizing pattern repetitions for characters and
292 character types now use tail recursion, which reduces stack usage.
293
294 37. The value of the max lookbehind was not correctly preserved if a compiled
295 and saved regex was reloaded on a host of different endianness.
296
297 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
298 of the compiled pattern block, expand the flags field from 16 to 32 bits
299 because it was almost full.
300
301 39. Try madvise first before posix_madvise.
302
303 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
304 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
305 original change locked out all matches of empty strings. This has been
306 changed so that one match of an empty string per line is recognized.
307 Subsequent searches on the same line (for colouring or for --only-matching,
308 for example) do not recognize empty strings.
309
310 41. Applied a user patch to fix a number of spelling mistakes in comments.
311
312 42. Data lines longer than 65536 caused pcretest to crash.
313
314 43. Clarified the data type for length and startoffset arguments for pcre_exec
315 and pcre_dfa_exec in the function-specific man pages, where they were
316 explicitly stated to be in bytes, never having been updated. I also added
317 some clarification to the pcreapi man page.
318
319 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
320 a segmentation fault.
321
322
323 Version 8.32 30-November-2012
324 -----------------------------
325
326 1. Improved JIT compiler optimizations for first character search and single
327 character iterators.
328
329 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
330 Patch by Daniel Richard G.
331
332 3. Single character iterator optimizations in the JIT compiler.
333
334 4. Improved JIT compiler optimizations for character ranges.
335
336 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
337 Reported by Giuseppe D'Angelo.
338
339 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
340 of a line, was being set incorrectly in cases where .* appeared inside
341 atomic brackets at the start of a pattern, or where there was a subsequent
342 *PRUNE or *SKIP.
343
344 7. Improved instruction cache flush for POWER/PowerPC.
345 Patch by Daniel Richard G.
346
347 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
348 grep:
349
350 (a) There is now no limit to the number of patterns to be matched.
351
352 (b) An error is given if a pattern is too long.
353
354 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
355 are now supported.
356
357 (d) --exclude-from and --include-from (multiple use) have been added.
358
359 (e) Exclusions and inclusions now apply to all files and directories, not
360 just to those obtained from scanning a directory recursively.
361
362 (f) Multiple uses of -f and --file-list are now supported.
363
364 (g) In a Windows environment, the default for -d has been changed from
365 "read" (the GNU grep default) to "skip", because otherwise the presence
366 of a directory in the file list provokes an error.
367
368 (h) The documentation has been revised and clarified in places.
369
370 9. Improve the matching speed of capturing brackets.
371
372 10. Changed the meaning of \X so that it now matches a Unicode extended
373 grapheme cluster.
374
375 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
376 out POSIX threads when JIT support is configured.
377
378 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
379
380 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
381 structure could go wrong in environments where size_t is not the same size
382 as int.
383
384 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
385 set.
386
387 15. The EBCDIC support had decayed; later updates to the code had included
388 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
389 general tidy up of EBCDIC-related issues, and the documentation was also
390 not quite right. There is now a test that can be run on ASCII systems to
391 check some of the EBCDIC-related things (but is it not a full test).
392
393 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
394 in a small tidy to the code.
395
396 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
397
398 18. If the --only-matching (-o) option in pcregrep is specified multiple
399 times, each one causes appropriate output. For example, -o1 -o2 outputs the
400 substrings matched by the 1st and 2nd capturing parentheses. A separating
401 string can be specified by --om-separator (default empty).
402
403 19. Improving the first n character searches.
404
405 20. Turn case lists for horizontal and vertical white space into macros so that
406 they are defined only once.
407
408 21. This set of changes together give more compatible Unicode case-folding
409 behaviour for characters that have more than one other case when UCP
410 support is available.
411
412 (a) The Unicode property table now has offsets into a new table of sets of
413 three or more characters that are case-equivalent. The MultiStage2.py
414 script that generates these tables (the pcre_ucd.c file) now scans
415 CaseFolding.txt instead of UnicodeData.txt for character case
416 information.
417
418 (b) The code for adding characters or ranges of characters to a character
419 class has been abstracted into a generalized function that also handles
420 case-independence. In UTF-mode with UCP support, this uses the new data
421 to handle characters with more than one other case.
422
423 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
424 whose other case is greater than 256 are now correctly matched
425 caselessly. Previously, the high codepoint matched the low one, but not
426 vice versa.
427
428 (d) The processing of \h, \H, \v, and \ in character classes now makes use
429 of the new class addition function, using character lists defined as
430 macros alongside the case definitions of 20 above.
431
432 (e) Caseless back references now work with characters that have more than
433 one other case.
434
435 (f) General caseless matching of characters with more than one other case
436 is supported.
437
438 22. Unicode character properties were updated from Unicode 6.2.0
439
440 23. Improved CMake support under Windows. Patch by Daniel Richard G.
441
442 24. Add support for 32-bit character strings, and UTF-32
443
444 25. Major JIT compiler update (code refactoring and bugfixing).
445 Experimental Sparc 32 support is added.
446
447 26. Applied a modified version of Daniel Richard G's patch to create
448 pcre.h.generic and config.h.generic by "make" instead of in the
449 PrepareRelease script.
450
451 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
452 pcre_compile.c when checking for a zero character.
453
454 28. Introducing a native interface for JIT. Through this interface, the compiled
455 machine code can be directly executed. The purpose of this interface is to
456 provide fast pattern matching, so several sanity checks are not performed.
457 However, feature tests are still performed. The new interface provides
458 1.4x speedup compared to the old one.
459
460 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
461 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
462 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
463
464 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
465 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
466 the "old" RFC 2279). Instead, it ended up passing a negative length to
467 pcre_exec().
468
469 31. Add support for GCC's visibility feature to hide internal functions.
470
471 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
472 "unknown -C option" after outputting 0 or 1.
473
474 33. There is now support for generating a code coverage report for the test
475 suite in environments where gcc is the compiler and lcov is installed. This
476 is mainly for the benefit of the developers.
477
478 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
479 unaddressable using valgrind annotations, allowing valgrind to detect
480 invalid memory accesses. This is mainly for the benefit of the developers.
481
482 25. (*UTF) can now be used to start a pattern in any of the three libraries.
483
484 26. Give configure error if --enable-cpp but no C++ compiler found.
485
486
487 Version 8.31 06-July-2012
488 -------------------------
489
490 1. Fixing a wrong JIT test case and some compiler warnings.
491
492 2. Removed a bashism from the RunTest script.
493
494 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
495 to unsigned type, result still unsigned" that was given by an MS compiler
496 on encountering the code "-sizeof(xxx)".
497
498 4. Partial matching support is added to the JIT compiler.
499
500 5. Fixed several bugs concerned with partial matching of items that consist
501 of more than one character:
502
503 (a) /^(..)\1/ did not partially match "aba" because checking references was
504 done on an "all or nothing" basis. This also applied to repeated
505 references.
506
507 (b) \R did not give a hard partial match if \r was found at the end of the
508 subject.
509
510 (c) \X did not give a hard partial match after matching one or more
511 characters at the end of the subject.
512
513 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
514 a partial match for the string "\r".
515
516 (e) When newline was set to CRLF, the metacharacter "." did not recognize
517 a partial match for a CR character at the end of the subject string.
518
519 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
520 running pcretest, the text "(JIT)" added to the output whenever JIT is
521 actually used to run the match.
522
523 7. Individual JIT compile options can be set in pcretest by following -s+[+]
524 or /S+[+] with a digit between 1 and 7.
525
526 8. OP_NOT now supports any UTF character not just single-byte ones.
527
528 9. (*MARK) control verb is now supported by the JIT compiler.
529
530 10. The command "./RunTest list" lists the available tests without actually
531 running any of them. (Because I keep forgetting what they all are.)
532
533 11. Add PCRE_INFO_MAXLOOKBEHIND.
534
535 12. Applied a (slightly modified) user-supplied patch that improves performance
536 when the heap is used for recursion (compiled with --disable-stack-for-
537 recursion). Instead of malloc and free for each heap frame each time a
538 logical recursion happens, frames are retained on a chain and re-used where
539 possible. This sometimes gives as much as 30% improvement.
540
541 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
542 call.
543
544 14. As documented, (*COMMIT) is now confined to within a positive assertion.
545
546 15. It is now possible to link pcretest with libedit as an alternative to
547 libreadline.
548
549 16. (*COMMIT) control verb is now supported by the JIT compiler.
550
551 17. The Unicode data tables have been updated to Unicode 6.1.0.
552
553 18. Added --file-list option to pcregrep.
554
555 19. Added binary file support to pcregrep, including the -a, --binary-files,
556 -I, and --text options.
557
558 20. The madvise function is renamed for posix_madvise for QNX compatibility
559 reasons. Fixed by Giuseppe D'Angelo.
560
561 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
562
563 22. Changed the option for creating symbolic links for 16-bit man pages from
564 -s to -sf so that re-installing does not cause issues.
565
566 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
567
568 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
569 matches in certain environments (the workspace was not being correctly
570 retained). Also added to pcre_dfa_exec() a simple plausibility check on
571 some of the workspace data at the beginning of a restart.
572
573 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
574 was not doing so when it should - probably a typo introduced by SVN 528
575 (change 8.10/14).
576
577 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
578 \w+ when the character tables indicated that \x{c4} was a word character.
579 There were several related cases, all because the tests for doing a table
580 lookup were testing for characters less than 127 instead of 255.
581
582 27. If a pattern contains capturing parentheses that are not used in a match,
583 their slots in the ovector are set to -1. For those that are higher than
584 any matched groups, this happens at the end of processing. In the case when
585 there were back references that the ovector was too small to contain
586 (causing temporary malloc'd memory to be used during matching), and the
587 highest capturing number was not used, memory off the end of the ovector
588 was incorrectly being set to -1. (It was using the size of the temporary
589 memory instead of the true size.)
590
591 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
592 ovector size, it uses memory at the end of the block that it has got.
593
594 29. Check for an overlong MARK name and give an error at compile time. The
595 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
596
597 30. JIT compiler update.
598
599 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
600 Rill for the patch.
601
602 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
603
604 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
605
606 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
607 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
608 was enabled.
609
610 35. Improve JIT code generation for greedy plus quantifier.
611
612 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
613 1 to "aa" instead of to an empty string. The bug affected repeated groups
614 that could potentially match an empty string.
615
616 37. Optimizing single character iterators in JIT.
617
618 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
619 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
620 codepoints that are too big for the mode are faulted, and in a UTF mode,
621 disallowed codepoints are also faulted.
622
623 39. If PCRE was compiled with UTF support, in three places in the DFA
624 matcher there was code that should only have been obeyed in UTF mode, but
625 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
626 processing when bytes with values greater than 127 were present. In 16-bit
627 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
628 both cases the values are those that cannot be the first data item in a UTF
629 character. The three items that might have provoked this were recursions,
630 possessively repeated groups, and atomic groups.
631
632 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
633 and pcregrep, because some OS require shared objects to be explicitly
634 passed to ld, causing the link step to fail if they are not.
635
636 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
637 mode, patterns that started with \h* or \R* might be incorrectly matched.
638
639
640 Version 8.30 04-February-2012
641 -----------------------------
642
643 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
644 name is defined in ctype.h.
645
646 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
647 only in quite long subpatterns.
648
649 3. Removed the function pcre_info(), which has been obsolete and deprecated
650 since it was replaced by pcre_fullinfo() in February 2000.
651
652 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
653 match a (*MARK), and the match failed at the start of the subject, a
654 reference to memory before the start of the subject could occur. This bug
655 was introduced by fix 17 of release 8.21.
656
657 5. A reference to an unset group with zero minimum repetition was giving
658 totally wrong answers (in non-JavaScript-compatibility mode). For example,
659 /(another)?(\1?)test/ matched against "hello world test". This bug was
660 introduced in release 8.13.
661
662 6. Add support for 16-bit character strings (a large amount of work involving
663 many changes and refactorings).
664
665 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
666 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
667 from a file.
668
669 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
670 rounding is not applied in this particular case).
671
672 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
673 if they appear, or are escaped, in patterns.
674
675 10. Get rid of a number of -Wunused-but-set-variable warnings.
676
677 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
678 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
679 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
680 also returns the mark "x". This bug applied to capturing parentheses,
681 non-capturing parentheses, and atomic parentheses. It also applied to some
682 assertions.
683
684 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
685 information out of configure.ac instead of relying on pcre.h.generic, which
686 is not stored in the repository.
687
688 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
689 -lreadline.
690
691 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
692
693 15. Applied Graycode's patch to put the top-level frame on the stack rather
694 than the heap when not using the stack for recursion. This gives a
695 performance improvement in many cases when recursion is not deep.
696
697 16. Experimental code added to "pcretest -C" to output the stack frame size.
698
699
700 Version 8.21 12-Dec-2011
701 ------------------------
702
703 1. Updating the JIT compiler.
704
705 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
706 are added as well.
707
708 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
709 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
710 calling _pcre_jit_exec. Some extra comments are added.
711
712 4. (*MARK) settings inside atomic groups that do not contain any capturing
713 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
714 was introduced by change 18 for 8.20.
715
716 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
717 ECMA-262 standard.
718
719 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
720 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
721 This bug was probably introduced by change 9 of 8.13.
722
723 7. While fixing 6 above, I noticed that a number of other items were being
724 incorrectly rejected as "not fixed length". This arose partly because newer
725 opcodes had not been added to the fixed-length checking code. I have (a)
726 corrected the bug and added tests for these items, and (b) arranged for an
727 error to occur if an unknown opcode is encountered while checking for fixed
728 length instead of just assuming "not fixed length". The items that were
729 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
730 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
731 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
732
733 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
734 being incorrectly compiled and would have given unpredicatble results.
735
736 9. A possessively repeated subpattern with minimum repeat count greater than
737 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
738 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
739 the first (A) could occur when it should not.
740
741 10. Add a cast and remove a redundant test from the code.
742
743 11. JIT should use pcre_malloc/pcre_free for allocation.
744
745 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
746 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
747 is anything other than /usr, -L is still shown).
748
749 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
750
751 14. Perl does not support \N without a following name in a [] class; PCRE now
752 also gives an error.
753
754 15. If a forward reference was repeated with an upper limit of around 2000,
755 it caused the error "internal error: overran compiling workspace". The
756 maximum number of forward references (including repeats) was limited by the
757 internal workspace, and dependent on the LINK_SIZE. The code has been
758 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
759 the default depends on LINK_SIZE. There is a new upper limit (for safety)
760 of around 200,000 forward references. While doing this, I also speeded up
761 the filling in of repeated forward references.
762
763 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
764 incorrectly expecting the subject to contain another "a" after the start.
765
766 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
767 in the match, the SKIP should be ignored. This was not happening; instead
768 the SKIP was being treated as NOMATCH. For patterns such as
769 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
770 tested.
771
772 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
773 now much more compatible with Perl, in particular in cases where the result
774 is a non-match for a non-anchored pattern. For example, if
775 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
776 "m", where previously it did not return a name. A side effect of this
777 change is that for partial matches, the last encountered mark name is
778 returned, as for non matches. A number of tests that were previously not
779 Perl-compatible have been moved into the Perl-compatible test files. The
780 refactoring has had the pleasing side effect of removing one argument from
781 the match() function, thus reducing its stack requirements.
782
783 19. If the /S+ option was used in pcretest to study a pattern using JIT,
784 subsequent uses of /S (without +) incorrectly behaved like /S+.
785
786 21. Retrieve executable code size support for the JIT compiler and fixing
787 some warnings.
788
789 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
790 not work when the shorter character appeared right at the end of the
791 subject string.
792
793 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
794 systems.
795
796 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
797 output it when the /M option is used in pcretest.
798
799 25. The CheckMan script was not being included in the distribution. Also, added
800 an explicit "perl" to run Perl scripts from the PrepareRelease script
801 because this is reportedly needed in Windows.
802
803 26. If study data was being save in a file and studying had not found a set of
804 "starts with" bytes for the pattern, the data written to the file (though
805 never used) was taken from uninitialized memory and so caused valgrind to
806 complain.
807
808 27. Updated RunTest.bat as provided by Sheri Pierce.
809
810 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
811
812 29. Computation of memory usage for the table of capturing group names was
813 giving an unnecessarily large value.
814
815
816 Version 8.20 21-Oct-2011
817 ------------------------
818
819 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
820 a POSIX class. After further experiments with Perl, which convinced me that
821 Perl has bugs and confusions, a closing square bracket is no longer allowed
822 in a POSIX name. This bug also affected patterns with classes that started
823 with full stops.
824
825 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
826 captured substring, but while checking the failing first alternative,
827 substring 1 is temporarily captured. If the output vector supplied to
828 pcre_exec() was not big enough for this capture, the yield of the function
829 was still zero ("insufficient space for captured substrings"). This cannot
830 be totally fixed without adding another stack variable, which seems a lot
831 of expense for a edge case. However, I have improved the situation in cases
832 such as /(a)(b)x|abc/ matched against "abc", where the return code
833 indicates that fewer than the maximum number of slots in the ovector have
834 been set.
835
836 3. Related to (2) above: when there are more back references in a pattern than
837 slots in the output vector, pcre_exec() uses temporary memory during
838 matching, and copies in the captures as far as possible afterwards. It was
839 using the entire output vector, but this conflicts with the specification
840 that only 2/3 is used for passing back captured substrings. Now it uses
841 only the first 2/3, for compatibility. This is, of course, another edge
842 case.
843
844 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
845 main code base, and can be used by building with --enable-jit. When this is
846 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
847 runtime --no-jit option is given.
848
849 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
850 ovector, the return from the function was zero, implying that there were
851 other matches that did not fit. The correct "exactly full" value is now
852 returned.
853
854 6. If a subpattern that was called recursively or as a subroutine contained
855 (*PRUNE) or any other control that caused it to give a non-standard return,
856 invalid errors such as "Error -26 (nested recursion at the same subject
857 position)" or even infinite loops could occur.
858
859 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
860 computing the minimum length on reaching *ACCEPT, and so ended up with the
861 wrong value of 1 rather than 0. Further investigation indicates that
862 computing a minimum subject length in the presence of *ACCEPT is difficult
863 (think back references, subroutine calls), and so I have changed the code
864 so that no minimum is registered for a pattern that contains *ACCEPT.
865
866 8. If (*THEN) was present in the first (true) branch of a conditional group,
867 it was not handled as intended. [But see 16 below.]
868
869 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
870 Sheri Pierce.
871
872 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
873 the first byte in a match must be "a".
874
875 11. Change 17 for 8.13 increased the recursion depth for patterns like
876 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
877 pattern contains any instances of (*THEN). If it does not, the old
878 optimizations are restored. It would be nice to do this on a per-group
879 basis, but at the moment that is not feasible.
880
881 12. In some environments, the output of pcretest -C is CRLF terminated. This
882 broke RunTest's code that checks for the link size. A single white space
883 character after the value is now allowed for.
884
885 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
886 For "fr", it uses the Windows-specific input and output files.
887
888 14. If (*THEN) appeared in a group that was called recursively or as a
889 subroutine, it did not work as intended. [But see next item.]
890
891 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
892 pattern fragments (but not containing any | characters). If A and B are
893 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
894 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
895 D. In other words, Perl considers parentheses that do not contain any |
896 characters to be part of a surrounding alternative, whereas PCRE was
897 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
898 differently. PCRE now behaves in the same way as Perl, except in the case
899 of subroutine/recursion calls such as (?1) which have in any case always
900 been different (but PCRE had them first :-).
901
902 16. Related to 15 above: Perl does not treat the | in a conditional group as
903 creating alternatives. Such a group is treated in the same way as an
904 ordinary group without any | characters when processing (*THEN). PCRE has
905 been changed to match Perl's behaviour.
906
907 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
908 RunGrepTest script failed.
909
910 18. Change 22 for version 13 caused atomic groups to use more stack. This is
911 inevitable for groups that contain captures, but it can lead to a lot of
912 stack use in large patterns. The old behaviour has been restored for atomic
913 groups that do not contain any capturing parentheses.
914
915 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
916 suppress the check for a minimum subject length at run time. (If it was
917 given to pcre_exec() or pcre_dfa_exec() it did work.)
918
919 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
920 fail to work when decoding hex characters in data strings in EBCDIC
921 environments.
922
923 21. It appears that in at least one Mac OS environment, the isxdigit() function
924 is implemented as a macro that evaluates to its argument more than once,
925 contravening the C 90 Standard (I haven't checked a later standard). There
926 was an instance in pcretest which caused it to go wrong when processing
927 \x{...} escapes in subject strings. The has been rewritten to avoid using
928 things like p++ in the argument of isxdigit().
929
930
931 Version 8.13 16-Aug-2011
932 ------------------------
933
934 1. The Unicode data tables have been updated to Unicode 6.0.0.
935
936 2. Two minor typos in pcre_internal.h have been fixed.
937
938 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
939 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
940 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
941
942 4. There were a number of related bugs in the code for matching backrefences
943 caselessly in UTF-8 mode when codes for the characters concerned were
944 different numbers of bytes. For example, U+023A and U+2C65 are an upper
945 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
946 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
947 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
948 2-byte code at the end of the subject (it thought there wasn't enough data
949 left).
950
951 5. Comprehensive information about what went wrong is now returned by
952 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
953 as the output vector has at least 2 elements. The offset of the start of
954 the failing character and a reason code are placed in the vector.
955
956 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
957 now returned is for the first byte of the failing character, instead of the
958 last byte inspected. This is an incompatible change, but I hope it is small
959 enough not to be a problem. It makes the returned offset consistent with
960 pcre_exec() and pcre_dfa_exec().
961
962 7. pcretest now gives a text phrase as well as the error number when
963 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
964 failure, the offset and reason code are output.
965
966 8. When \R was used with a maximizing quantifier it failed to skip backwards
967 over a \r\n pair if the subsequent match failed. Instead, it just skipped
968 back over a single character (\n). This seems wrong (because it treated the
969 two characters as a single entity when going forwards), conflicts with the
970 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
971 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
972 has been changed.
973
974 9. Some internal refactoring has changed the processing so that the handling
975 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
976 time (the PCRE_DOTALL option was changed this way some time ago: version
977 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
978 which was always a bit of a fudge. It also means that there is one less
979 argument for the match() function, which reduces its stack requirements
980 slightly. This change also fixes an incompatibility with Perl: the pattern
981 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
982
983 10. More internal refactoring has drastically reduced the number of recursive
984 calls to match() for possessively repeated groups such as (abc)++ when
985 using pcre_exec().
986
987 11. While implementing 10, a number of bugs in the handling of groups were
988 discovered and fixed:
989
990 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
991 (a|)*(?1) gave a compile-time internal error.
992 ((a|)+)+ did not notice that the outer group could match an empty string.
993 (^a|^)+ was not marked as anchored.
994 (.*a|.*)+ was not marked as matching at start or after a newline.
995
996 12. Yet more internal refactoring has removed another argument from the match()
997 function. Special calls to this function are now indicated by setting a
998 value in a variable in the "match data" data block.
999
1000 13. Be more explicit in pcre_study() instead of relying on "default" for
1001 opcodes that mean there is no starting character; this means that when new
1002 ones are added and accidentally left out of pcre_study(), testing should
1003 pick them up.
1004
1005 14. The -s option of pcretest has been documented for ages as being an old
1006 synonym of -m (show memory usage). I have changed it to mean "force study
1007 for every regex", that is, assume /S for every regex. This is similar to -i
1008 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1009 using it. It makes it easier to run collections of tests with and without
1010 study enabled, and thereby test pcre_study() more easily. All the standard
1011 tests are now run with and without -s (but some patterns can be marked as
1012 "never study" - see 20 below).
1013
1014 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1015 restoration of the capturing data to the outer values was not happening
1016 correctly.
1017
1018 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1019 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1020 pattern had matched an empty string, and so incorrectly returned a no
1021 match.
1022
1023 17. There was optimizing code for the last branch of non-capturing parentheses,
1024 and also for the obeyed branch of a conditional subexpression, which used
1025 tail recursion to cut down on stack usage. Unfortunately, now that there is
1026 the possibility of (*THEN) occurring in these branches, tail recursion is
1027 no longer possible because the return has to be checked for (*THEN). These
1028 two optimizations have therefore been removed. [But see 8.20/11 above.]
1029
1030 18. If a pattern containing \R was studied, it was assumed that \R always
1031 matched two bytes, thus causing the minimum subject length to be
1032 incorrectly computed because \R can also match just one byte.
1033
1034 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1035 was incorrectly computed.
1036
1037 20. If /S is present twice on a test pattern in pcretest input, it now
1038 *disables* studying, thereby overriding the use of -s on the command line
1039 (see 14 above). This is necessary for one or two tests to keep the output
1040 identical in both cases.
1041
1042 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1043 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1044
1045 22. When an atomic group that contained a capturing parenthesis was
1046 successfully matched, but the branch in which it appeared failed, the
1047 capturing was not being forgotten if a higher numbered group was later
1048 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1049 group 1 to "a", when in fact it should be unset. This applied to multi-
1050 branched capturing and non-capturing groups, repeated or not, and also to
1051 positive assertions (capturing in negative assertions does not happen
1052 in PCRE) and also to nested atomic groups.
1053
1054 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1055 subject after a captured substring, to make it easier to tell which of a
1056 number of identical substrings has been captured.
1057
1058 24. The way atomic groups are processed by pcre_exec() has been changed so that
1059 if they are repeated, backtracking one repetition now resets captured
1060 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1061 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1062 "aaa". Previously, it would have been "a". As part of this code
1063 refactoring, the way recursive calls are handled has also been changed.
1064
1065 25. If an assertion condition captured any substrings, they were not passed
1066 back unless some other capturing happened later. For example, if
1067 (?(?=(a))a) was matched against "a", no capturing was returned.
1068
1069 26. When studying a pattern that contained subroutine calls or assertions,
1070 the code for finding the minimum length of a possible match was handling
1071 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1072 group 1 called group 2 while simultaneously a separate group 2 called group
1073 1). A stack overflow occurred in this case. I have fixed this by limiting
1074 the recursion depth to 10.
1075
1076 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1077 Fortmann. This supports explicit test numbers on the command line, and has
1078 argument validation and error reporting.
1079
1080 28. An instance of \X with an unlimited repeat could fail if at any point the
1081 first character it looked at was a mark character.
1082
1083 29. Some minor code refactoring concerning Unicode properties and scripts
1084 should reduce the stack requirement of match() slightly.
1085
1086 30. Added the '=' option to pcretest to check the setting of unused capturing
1087 slots at the end of the pattern, which are documented as being -1, but are
1088 not included in the return count.
1089
1090 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1091 compiled something random. Now it gives a compile-time error (as does
1092 Perl).
1093
1094 32. A *MARK encountered during the processing of a positive assertion is now
1095 recorded and passed back (compatible with Perl).
1096
1097 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1098 had alternative anchored branches, the search for a second match in a line
1099 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1100 matched the line "0102" twice. The same bug affected patterns that started
1101 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1102 twice.
1103
1104 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1105 does, and because of capturing effects, quantifying parenthesized
1106 assertions may at times be useful. Quantifiers are now allowed for
1107 parenthesized assertions.
1108
1109 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1110
1111 36. \g was being checked for fancy things in a character class, when it should
1112 just be a literal "g".
1113
1114 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1115 appearance of a nested POSIX class supersedes an apparent external class.
1116 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1117 unescaped square brackets may also appear as part of class names. For
1118 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1119 more like Perl. (But see 8.20/1 above.)
1120
1121 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1122 was because it thought it was \N{name}, which is not supported).
1123
1124 39. Add minix to OS list not supporting the -S option in pcretest.
1125
1126 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1127 cannot analyze patterns in sufficient detail to catch mutual recursions
1128 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1129 subgroup is called recursively as a subpattern for a second time at the
1130 same position in the subject string. In previous releases this might have
1131 been caught by the recursion limit, or it might have run out of stack.
1132
1133 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1134 happen only once. PCRE was, however incorrectly giving a compile time error
1135 "recursive call could loop indefinitely" because it cannot analyze the
1136 pattern in sufficient detail. The compile time test no longer happens when
1137 PCRE is compiling a conditional subpattern, but actual runaway loops are
1138 now caught at runtime (see 40 above).
1139
1140 42. It seems that Perl allows any characters other than a closing parenthesis
1141 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1142 has been changed to be the same.
1143
1144 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1145 as not to get warnings when autogen.sh is called. Also changed
1146 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1147
1148 44. To help people who use pcregrep to scan files containing exceedingly long
1149 lines, the following changes have been made:
1150
1151 (a) The default value of the buffer size parameter has been increased from
1152 8K to 20K. (The actual buffer used is three times this size.)
1153
1154 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1155 PCRE is built.
1156
1157 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1158 to be set at run time.
1159
1160 (d) Numerical values in pcregrep options can be followed by K or M, for
1161 example --buffer-size=50K.
1162
1163 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1164 given and the return code is set to 2.
1165
1166 45. Add a pointer to the latest mark to the callout data block.
1167
1168 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1169 partial match of an empty string instead of no match. This was specific to
1170 the use of ".".
1171
1172 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1173 complete match instead of a partial match. This bug was dependent on both
1174 the PCRE_UTF8 and PCRE_DOTALL options being set.
1175
1176 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1177 starting byte set, because \b was not being ignored.
1178
1179
1180 Version 8.12 15-Jan-2011
1181 ------------------------
1182
1183 1. Fixed some typos in the markup of the man pages, and wrote a script that
1184 checks for such things as part of the documentation building process.
1185
1186 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1187 --match-limit and --recursion-limit options (added for 8.11). In
1188 particular, this made one of the standard tests fail. (The integer value
1189 went into the wrong half of a long int.)
1190
1191 3. If the --colour option was given to pcregrep with -v (invert match), it
1192 did strange things, either producing crazy output, or crashing. It should,
1193 of course, ignore a request for colour when reporting lines that do not
1194 match.
1195
1196 4. Another pcregrep bug caused similar problems if --colour was specified with
1197 -M (multiline) and the pattern match finished with a line ending.
1198
1199 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1200 matched in multiline mode, the following line was shown as part of the
1201 match. This seems wrong, so I have changed it.
1202
1203 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1204 the check for further matches in the same line (so they could be coloured)
1205 to overrun the end of the current line. If another match was found, it was
1206 incorrectly shown (and then shown again when found in the next line).
1207
1208 7. If pcregrep was compiled under Windows, there was a reference to the
1209 function pcregrep_exit() before it was defined. I am assuming this was
1210 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1211 reported by a user. I've moved the definition above the reference.
1212
1213
1214 Version 8.11 10-Dec-2010
1215 ------------------------
1216
1217 1. (*THEN) was not working properly if there were untried alternatives prior
1218 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1219 backtracked to try for "b" instead of moving to the next alternative branch
1220 at the same level (in this case, to look for "c"). The Perl documentation
1221 is clear that when (*THEN) is backtracked onto, it goes to the "next
1222 alternative in the innermost enclosing group".
1223
1224 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1225 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1226 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1227 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1228 (*THEN).
1229
1230 3. If \s appeared in a character class, it removed the VT character from
1231 the class, even if it had been included by some previous item, for example
1232 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1233 of \s, but is part of the POSIX "space" class.)
1234
1235 4. A partial match never returns an empty string (because you can always
1236 match an empty string at the end of the subject); however the checking for
1237 an empty string was starting at the "start of match" point. This has been
1238 changed to the "earliest inspected character" point, because the returned
1239 data for a partial match starts at this character. This means that, for
1240 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1241 (previously it gave "no match").
1242
1243 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1244 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1245 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1246 has an implication that the given string is incomplete (because a partial
1247 match is preferred over a full match). For this reason, these items now
1248 give a partial match in this situation. [Aside: previously, the one case
1249 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1250 match rather than a full match, which was wrong by the old rules, but is
1251 now correct.]
1252
1253 6. There was a bug in the handling of #-introduced comments, recognized when
1254 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1255 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1256 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1257 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1258 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1259 places in pcre_compile().
1260
1261 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1262 comments when looking ahead for named forward references to subpatterns,
1263 the only newline sequence it recognized was NL. It now handles newlines
1264 according to the set newline convention.
1265
1266 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1267 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1268 cater for a lack of strerror(). These oversights have been fixed.
1269
1270 9. Added --match-limit and --recursion-limit to pcregrep.
1271
1272 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1273
1274 11. When the -o option was used, pcregrep was setting a return code of 1, even
1275 when matches were found, and --line-buffered was not being honoured.
1276
1277 12. Added an optional parentheses number to the -o and --only-matching options
1278 of pcregrep.
1279
1280 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1281 can match an empty string. The code to do it in pcretest and pcredemo
1282 needed fixing:
1283
1284 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1285 only one byte after an empty string match just before CRLF (this case
1286 just got forgotten; "any" and "anycrlf" were OK).
1287
1288 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1289 mode when an empty string match preceded an ASCII character followed by
1290 a non-ASCII character. (The code for advancing by one character rather
1291 than one byte was nonsense.)
1292
1293 (c) The pcredemo.c sample program did not have any code at all to handle
1294 the cases when CRLF is a valid newline sequence.
1295
1296 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1297 as a starting offset was within the subject string. There is now a new
1298 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1299 negative or greater than the length of the string. In order to test this,
1300 pcretest is extended to allow the setting of negative starting offsets.
1301
1302 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1303 starting offset points to the beginning of a UTF-8 character was
1304 unnecessarily clumsy. I tidied it up.
1305
1306 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1307 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1308
1309 17. Nobody had reported that the --include_dir option, which was added in
1310 release 7.7 should have been called --include-dir (hyphen, not underscore)
1311 for compatibility with GNU grep. I have changed it to --include-dir, but
1312 left --include_dir as an undocumented synonym, and the same for
1313 --exclude-dir, though that is not available in GNU grep, at least as of
1314 release 2.5.4.
1315
1316 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1317 characters from a string of bytes) have been redefined so as not to use
1318 loops, in order to improve performance in some environments. At the same
1319 time, I abstracted some of the common code into auxiliary macros to save
1320 repetition (this should not affect the compiled code).
1321
1322 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1323 compile-time error is now given if \c is not followed by an ASCII
1324 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1325 different, and any byte value is allowed.)
1326
1327 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1328 START_OPTIMIZE option, which is now allowed at compile time - but just
1329 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1330 to pcregrep and other applications that have no direct access to PCRE
1331 options. The new /Y option in pcretest sets this option when calling
1332 pcre_compile().
1333
1334 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1335 back references. Groups containing recursive back references were forced to
1336 be atomic by that change, but in the case of named groups, the amount of
1337 memory required was incorrectly computed, leading to "Failed: internal
1338 error: code overflow". This has been fixed.
1339
1340 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1341 pcretest.c, to avoid build problems in some Borland environments.
1342
1343
1344 Version 8.10 25-Jun-2010
1345 ------------------------
1346
1347 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1348 THEN.
1349
1350 2. (*ACCEPT) was not working when inside an atomic group.
1351
1352 3. Inside a character class, \B is treated as a literal by default, but
1353 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1354 causes the error). The code is unchanged, but I tidied the documentation.
1355
1356 4. Inside a character class, PCRE always treated \R and \X as literals,
1357 whereas Perl faults them if its -w option is set. I have changed PCRE so
1358 that it faults them when PCRE_EXTRA is set.
1359
1360 5. Added support for \N, which always matches any character other than
1361 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1362
1363 6. When compiling pcregrep with newer versions of gcc which may have
1364 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1365 declared with attribute warn_unused_result" were given. Just casting the
1366 result to (void) does not stop the warnings; a more elaborate fudge is
1367 needed. I've used a macro to implement this.
1368
1369 7. Minor change to pcretest.c to avoid a compiler warning.
1370
1371 8. Added four artifical Unicode properties to help with an option to make
1372 \s etc use properties (see next item). The new properties are: Xan
1373 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1374
1375 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1376 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1377 this option. Modified pcretest to add /W to test this facility. Added
1378 REG_UCP to make it available via the POSIX interface.
1379
1380 10. Added --line-buffered to pcregrep.
1381
1382 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1383 studied, and the match started with a letter with a code point greater than
1384 127 whose first byte was different to the first byte of the other case of
1385 the letter, the other case of this starting letter was not recognized
1386 (#976).
1387
1388 12. If a pattern that was studied started with a repeated Unicode property
1389 test, for example, \p{Nd}+, there was the theoretical possibility of
1390 setting up an incorrect bitmap of starting bytes, but fortunately it could
1391 not have actually happened in practice until change 8 above was made (it
1392 added property types that matched character-matching opcodes).
1393
1394 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1395 possible starting bytes for non-anchored patterns.
1396
1397 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1398 \R, and also a number of cases that involve Unicode properties, both
1399 explicit and implicit when PCRE_UCP is set.
1400
1401 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1402 input, it could crash or give wrong results if characters with values
1403 greater than 0xc0 were present in the subject string. (Detail: it assumed
1404 UTF-8 input when processing these items.)
1405
1406 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1407 size_t is 64-bit (#991).
1408
1409 17. Added a check for running out of memory when PCRE is compiled with
1410 --disable-stack-for-recursion (#990).
1411
1412 18. If the last data line in a file for pcretest does not have a newline on
1413 the end, a newline was missing in the output.
1414
1415 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1416 less than 128) in its various bitmaps. However, there is a facility for
1417 generating tables according to the current locale when PCRE is compiled. It
1418 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1419 characters, are recognized by isspace() and therefore were getting set in
1420 these tables, and indeed these tables seem to approximate to ISO 8859. This
1421 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1422 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1423 which of course cannot start UTF-8 characters. I have changed the code so
1424 that only real ASCII characters (less than 128) and the correct starting
1425 bytes for UTF-8 encodings are set for characters greater than 127 when in
1426 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1427 altogether.)
1428
1429 20. Added the /T option to pcretest so as to be able to run tests with non-
1430 standard character tables, thus making it possible to include the tests
1431 used for 19 above in the standard set of tests.
1432
1433 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1434 reference to a subpattern the other side of a comment that contains an
1435 opening parenthesis caused either an internal compiling error, or a
1436 reference to the wrong subpattern.
1437
1438
1439 Version 8.02 19-Mar-2010
1440 ------------------------
1441
1442 1. The Unicode data tables have been updated to Unicode 5.2.0.
1443
1444 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1445 configured.
1446
1447 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1448 original author of that file, following a query about its status.
1449
1450 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1451 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1452
1453 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1454 quantifier applied to a forward-referencing subroutine call, could compile
1455 incorrect code or give the error "internal error: previously-checked
1456 referenced subpattern not found".
1457
1458 6. Both MS Visual Studio and Symbian OS have problems with initializing
1459 variables to point to external functions. For these systems, therefore,
1460 pcre_malloc etc. are now initialized to local functions that call the
1461 relevant global functions.
1462
1463 7. There were two entries missing in the vectors called coptable and poptable
1464 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1465 I've fixed the data, and added a kludgy way of testing at compile time that
1466 the lengths are correct (equal to the number of opcodes).
1467
1468 8. Following on from 7, I added a similar kludge to check the length of the
1469 eint vector in pcreposix.c.
1470
1471 9. Error texts for pcre_compile() are held as one long string to avoid too
1472 much relocation at load time. To find a text, the string is searched,
1473 counting zeros. There was no check for running off the end of the string,
1474 which could happen if a new error number was added without updating the
1475 string.
1476
1477 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1478
1479 11. \K was not working if it appeared in an atomic group or in a group that
1480 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1481 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1482 the assertion is positive, but not if it is negative.
1483
1484 12. Change 11 fortuitously reduced the size of the stack frame used in the
1485 "match()" function of pcre_exec.c by one pointer. Forthcoming
1486 implementation of support for (*MARK) will need an extra pointer on the
1487 stack; I have reserved it now, so that the stack frame size does not
1488 decrease.
1489
1490 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1491 item in branch that calls a recursion is a subroutine call - as in the
1492 second branch in the above example - was incorrectly given the compile-
1493 time error "recursive call could loop indefinitely" because pcre_compile()
1494 was not correctly checking the subroutine for matching a non-empty string.
1495
1496 14. The checks for overrunning compiling workspace could trigger after an
1497 overrun had occurred. This is a "should never occur" error, but it can be
1498 triggered by pathological patterns such as hundreds of nested parentheses.
1499 The checks now trigger 100 bytes before the end of the workspace.
1500
1501 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1502
1503
1504 Version 8.01 19-Jan-2010
1505 ------------------------
1506
1507 1. If a pattern contained a conditional subpattern with only one branch (in
1508 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1509 computed the wrong minimum data length (which is of course zero for such
1510 subpatterns). This could cause incorrect "no match" results.
1511
1512 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1513 the pattern is reset in the first branch, pcre_compile() failed with
1514 "internal error: code overflow at offset...". This happened only when
1515 the reset was to the original external option setting. (An optimization
1516 abstracts leading options settings into an external setting, which was the
1517 cause of this.)
1518
1519 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1520 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1521 assertion pattern did not match (meaning that the assertion was true), it
1522 was incorrectly treated as false if the SKIP had been reached during the
1523 matching. This also applied to assertions used as conditions.
1524
1525 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1526 assertion subpattern, including such a pattern used as a condition,
1527 unpredictable results occurred, instead of the error return
1528 PCRE_ERROR_DFA_UITEM.
1529
1530 5. The C++ GlobalReplace function was not working like Perl for the special
1531 situation when an empty string is matched. It now does the fancy magic
1532 stuff that is necessary.
1533
1534 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1535 removed. (These were left over from very, very early versions of PCRE.)
1536
1537 7. Some cosmetic changes to the code to make life easier when compiling it
1538 as part of something else:
1539
1540 (a) Change DEBUG to PCRE_DEBUG.
1541
1542 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1543 called "current" as "current_branch", to prevent a collision with the
1544 Linux macro when compiled as a kernel module.
1545
1546 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1547 prevent a collision with the Linux macro when compiled as a kernel
1548 module.
1549
1550 8. In pcre_compile() there are some checks for integer overflows that used to
1551 cast potentially large values to (double). This has been changed to that
1552 when building, a check for int64_t is made, and if it is found, it is used
1553 instead, thus avoiding the use of floating point arithmetic. (There is no
1554 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1555 double.
1556
1557 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1558 2005 (difference between two addresses compared to an unsigned value).
1559
1560 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1561 custom one, because of the following reported problem in Windows:
1562
1563 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1564 under Win32.
1565 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1566 therefore missing the function definition.
1567 - The compiler thus generates a "C" signature for the test function.
1568 - The linker fails to find the "C" function.
1569 - PCRE fails to configure if asked to do so against libbz2.
1570
1571 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1572 messages were output:
1573
1574 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1575 rerunning libtoolize, to keep the correct libtool macros in-tree.
1576 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1577
1578 I have done both of these things.
1579
1580 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1581 most of the time, it *can* run out if it is given a pattern that contains a
1582 runaway infinite recursion. I updated the discussion in the pcrestack man
1583 page.
1584
1585 13. Now that we have gone to the x.xx style of version numbers, the minor
1586 version may start with zero. Using 08 or 09 is a bad idea because users
1587 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1588 interpreted as invalid octal numbers. I've updated the previous comment in
1589 configure.ac, and also added a check that gives an error if 08 or 09 are
1590 used.
1591
1592 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1593 causing partial matching to fail when the end of the subject matched \W
1594 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1595
1596 15. There were some discrepancies between the declarations in pcre_internal.h
1597 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1598 their definitions. The declarations used "const uschar *" and the
1599 definitions used USPTR. Even though USPTR is normally defined as "const
1600 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1601 reported that: "This difference in casting confuses some C++ compilers, for
1602 example, SunCC recognizes above declarations as different functions and
1603 generates broken code for hbpcre." I have changed the declarations to use
1604 USPTR.
1605
1606 16. GNU libtool is named differently on some systems. The autogen.sh script now
1607 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1608 (FreeBSD).
1609
1610 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1611 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1612 comment: "Figure out how to create a longlong from a string: strtoll and
1613 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1614 instance, but it only takes 2 args instead of 3!"
1615
1616 18. A subtle bug concerned with back references has been fixed by a change of
1617 specification, with a corresponding code fix. A pattern such as
1618 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1619 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1620 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1621 same bug. Such groups have to be quantified to be useful, or contained
1622 inside another quantified group. (If there's no repetition, the reference
1623 can never match.) The problem arises because, having left the group and
1624 moved on to the rest of the pattern, a later failure that backtracks into
1625 the group uses the captured value from the final iteration of the group
1626 rather than the correct earlier one. I have fixed this in PCRE by forcing
1627 any group that contains a reference to itself to be an atomic group; that
1628 is, there cannot be any backtracking into it once it has completed. This is
1629 similar to recursive and subroutine calls.
1630
1631
1632 Version 8.00 19-Oct-09
1633 ----------------------
1634
1635 1. The table for translating pcre_compile() error codes into POSIX error codes
1636 was out-of-date, and there was no check on the pcre_compile() error code
1637 being within the table. This could lead to an OK return being given in
1638 error.
1639
1640 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1641 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1642 in a Windows environment.
1643
1644 3. The pcregrep --count option prints the count for each file even when it is
1645 zero, as does GNU grep. However, pcregrep was also printing all files when
1646 --files-with-matches was added. Now, when both options are given, it prints
1647 counts only for those files that have at least one match. (GNU grep just
1648 prints the file name in this circumstance, but including the count seems
1649 more useful - otherwise, why use --count?) Also ensured that the
1650 combination -clh just lists non-zero counts, with no names.
1651
1652 4. The long form of the pcregrep -F option was incorrectly implemented as
1653 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1654 but it seems right to fix it, and I didn't think it was worth preserving
1655 the old behaviour.
1656
1657 5. The command line items --regex=pattern and --regexp=pattern were not
1658 recognized by pcregrep, which required --regex pattern or --regexp pattern
1659 (with a space rather than an '='). The man page documented the '=' forms,
1660 which are compatible with GNU grep; these now work.
1661
1662 6. No libpcreposix.pc file was created for pkg-config; there was just
1663 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1664
1665 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1666 when UCP support is not needed, by modifying the Python script that
1667 generates it from Unicode data files. This should not matter if the module
1668 is correctly used as a library, but I received one complaint about 50K of
1669 unwanted data. My guess is that the person linked everything into his
1670 program rather than using a library. Anyway, it does no harm.
1671
1672 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1673 was a minimum greater than 1 for a wide character in a possessive
1674 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1675 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1676 character. Chaos in the form of incorrect output or a compiling loop could
1677 result.
1678
1679 9. The restrictions on what a pattern can contain when partial matching is
1680 requested for pcre_exec() have been removed. All patterns can now be
1681 partially matched by this function. In addition, if there are at least two
1682 slots in the offset vector, the offset of the earliest inspected character
1683 for the match and the offset of the end of the subject are set in them when
1684 PCRE_ERROR_PARTIAL is returned.
1685
1686 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1687 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1688 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1689 and may be more useful for multi-segment matching.
1690
1691 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1692 used to be given if ever the end of the subject was reached; now it is
1693 given only if matching could not proceed because another character was
1694 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1695 string "Z", which now yields "no match" instead of "partial match". In the
1696 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1697 final character ended with (*FAIL).
1698
1699 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1700 if the pattern had a "must contain" character that was already found in the
1701 earlier partial match, unless partial matching was again requested. For
1702 example, with the pattern /dog.(body)?/, the "must contain" character is
1703 "g". If the first part-match was for the string "dog", restarting with
1704 "sbody" failed. This bug has been fixed.
1705
1706 13. The string returned by pcre_dfa_exec() after a partial match has been
1707 changed so that it starts at the first inspected character rather than the
1708 first character of the match. This makes a difference only if the pattern
1709 starts with a lookbehind assertion or \b or \B (\K is not supported by
1710 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1711 matching functions compatible, and I think it's the right thing to do.
1712
1713 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1714 so that the demonstration program is easily available in environments where
1715 PCRE has not been installed from source.
1716
1717 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1718 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1719 library.
1720
1721 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1722 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1723 is not the first non-POSIX option to be added. Clearly some people find
1724 these options useful.
1725
1726 17. If a caller to the POSIX matching function regexec() passes a non-zero
1727 value for nmatch with a NULL value for pmatch, the value of
1728 nmatch is forced to zero.
1729
1730 18. RunGrepTest did not have a test for the availability of the -u option of
1731 the diff command, as RunTest does. It now checks in the same way as
1732 RunTest, and also checks for the -b option.
1733
1734 19. If an odd number of negated classes containing just a single character
1735 interposed, within parentheses, between a forward reference to a named
1736 subpattern and the definition of the subpattern, compilation crashed with
1737 an internal error, complaining that it could not find the referenced
1738 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1739 [The bug was that it was starting one character too far in when skipping
1740 over the character class, thus treating the ] as data rather than
1741 terminating the class. This meant it could skip too much.]
1742
1743 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1744 /g option in pcretest when the pattern contains \K, which makes it possible
1745 to have an empty string match not at the start, even when the pattern is
1746 anchored. Updated pcretest and pcredemo to use this option.
1747
1748 21. If the maximum number of capturing subpatterns in a recursion was greater
1749 than the maximum at the outer level, the higher number was returned, but
1750 with unset values at the outer level. The correct (outer level) value is
1751 now given.
1752
1753 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1754 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1755 make it do so. The string so far is captured, making this feature
1756 compatible with Perl.
1757
1758 23. The tests have been re-organized, adding tests 11 and 12, to make it
1759 possible to check the Perl 5.10 features against Perl 5.10.
1760
1761 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1762 pattern matches a fixed length string. PCRE did not allow this; now it
1763 does. Neither allows recursion.
1764
1765 25. I finally figured out how to implement a request to provide the minimum
1766 length of subject string that was needed in order to match a given pattern.
1767 (It was back references and recursion that I had previously got hung up
1768 on.) This code has now been added to pcre_study(); it finds a lower bound
1769 to the length of subject needed. It is not necessarily the greatest lower
1770 bound, but using it to avoid searching strings that are too short does give
1771 some useful speed-ups. The value is available to calling programs via
1772 pcre_fullinfo().
1773
1774 26. While implementing 25, I discovered to my embarrassment that pcretest had
1775 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1776 study optimizations had never been tested with that matching function.
1777 Oops. What is worse, even when it was passed study data, there was a bug in
1778 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1779 were also very few tests of studied patterns with pcre_dfa_exec().
1780
1781 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1782 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1783 on the other side of the coin, they are no longer allowed to have different
1784 names, because these cannot be distinguished in PCRE, and this has caused
1785 confusion. (This is a difference from Perl.)
1786
1787 28. When duplicate subpattern names are present (necessarily with different
1788 numbers, as required by 27 above), and a test is made by name in a
1789 conditional pattern, either for a subpattern having been matched, or for
1790 recursion in such a pattern, all the associated numbered subpatterns are
1791 tested, and the overall condition is true if the condition is true for any
1792 one of them. This is the way Perl works, and is also more like the way
1793 testing by number works.
1794
1795
1796 Version 7.9 11-Apr-09
1797 ---------------------
1798
1799 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1800 (pcretest), all targets were linked against these libraries. This included
1801 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1802 libraries. This caused unwanted dependencies to be created. This problem
1803 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1804 pcretest is linked with readline.
1805
1806 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1807 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1808 moved outside it again, because FALSE and TRUE are already defined in AIX,
1809 but BOOL is not.
1810
1811 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1812 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1813
1814 4. The pcregrep documentation said spaces were inserted as well as colons (or
1815 hyphens) following file names and line numbers when outputting matching
1816 lines. This is not true; no spaces are inserted. I have also clarified the
1817 wording for the --colour (or --color) option.
1818
1819 5. In pcregrep, when --colour was used with -o, the list of matching strings
1820 was not coloured; this is different to GNU grep, so I have changed it to be
1821 the same.
1822
1823 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1824 each matching line was coloured. Now it goes on to look for further matches
1825 of any of the test patterns, which is the same behaviour as GNU grep.
1826
1827 7. A pattern that could match an empty string could cause pcregrep to loop; it
1828 doesn't make sense to accept an empty string match in pcregrep, so I have
1829 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1830 seems to be how GNU grep behaves. [But see later change 40 for release
1831 8.33.]
1832
1833 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1834 start or after a newline", because the conditional assertion was not being
1835 correctly handled. The rule now is that both the assertion and what follows
1836 in the first alternative must satisfy the test.
1837
1838 9. If auto-callout was enabled in a pattern with a conditional group whose
1839 condition was an assertion, PCRE could crash during matching, both with
1840 pcre_exec() and pcre_dfa_exec().
1841
1842 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1843 used for matching.
1844
1845 11. Unicode property support in character classes was not working for
1846 characters (bytes) greater than 127 when not in UTF-8 mode.
1847
1848 12. Added the -M command line option to pcretest.
1849
1850 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1851
1852 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1853
1854 16. Added comments and documentation about mis-use of no_arg in the C++
1855 wrapper.
1856
1857 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1858 from Martin Jerabek that uses macro names for all relevant character and
1859 string constants.
1860
1861 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1862 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1863 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1864 these, but not everybody uses configure.
1865
1866 19. A conditional group that had only one branch was not being correctly
1867 recognized as an item that could match an empty string. This meant that an
1868 enclosing group might also not be so recognized, causing infinite looping
1869 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1870 with the subject "ab", where knowledge that the repeated group can match
1871 nothing is needed in order to break the loop.
1872
1873 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1874 exec(), but without supplying a callout function, matching went wrong.
1875
1876 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1877 leak if the size of the offset vector was greater than 30. When the vector
1878 is smaller, the saved offsets during recursion go onto a local stack
1879 vector, but for larger vectors malloc() is used. It was failing to free
1880 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1881 error, in fact).
1882
1883 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1884 heapframe that is used only when UTF-8 support is enabled. This caused no
1885 problem, but was untidy.
1886
1887 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1888 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1889 included within another project.
1890
1891 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1892 slightly modified by me:
1893
1894 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1895 not building pcregrep.
1896
1897 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1898 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1899
1900 25. Forward references, both numeric and by name, in patterns that made use of
1901 duplicate group numbers, could behave incorrectly or give incorrect errors,
1902 because when scanning forward to find the reference group, PCRE was not
1903 taking into account the duplicate group numbers. A pattern such as
1904 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1905
1906 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1907 the feature of a custom pointer more persuasive (as requested by a user).
1908
1909 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1910 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1911 pre-defined.
1912
1913 28. Added support for (*UTF8) at the start of a pattern.
1914
1915 29. Arrange for flags added by the "release type" setting in CMake to be shown
1916 in the configuration summary.
1917
1918
1919 Version 7.8 05-Sep-08
1920 ---------------------
1921
1922 1. Replaced UCP searching code with optimized version as implemented for Ad
1923 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1924 stage table and inline lookup instead of a function, giving speed ups of 2
1925 to 5 times on some simple patterns that I tested. Permission was given to
1926 distribute the MultiStage2.py script that generates the tables (it's not in
1927 the tarball, but is in the Subversion repository).
1928
1929 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1930 scripts.
1931
1932 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1933 a group with a zero qualifier. The result of the study could be incorrect,
1934 or the function might crash, depending on the pattern.
1935
1936 4. Caseless matching was not working for non-ASCII characters in back
1937 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1938 It now works when Unicode Property Support is available.
1939
1940 5. In pcretest, an escape such as \x{de} in the data was always generating
1941 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1942 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1943 truncation.
1944
1945 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1946
1947 7. Added two (int) casts to pcregrep when printing the difference of two
1948 pointers, in case they are 64-bit values.
1949
1950 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1951 test 2 if it fails.
1952
1953 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1954 and a #define of that name to empty if it is not externally set. This is to
1955 allow users of MSVC to set it if necessary.
1956
1957 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1958 the convenience functions in the pcre_get.c source file.
1959
1960 11. An option change at the start of a pattern that had top-level alternatives
1961 could cause overwriting and/or a crash. This command provoked a crash in
1962 some environments:
1963
1964 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1965
1966 This potential security problem was recorded as CVE-2008-2371.
1967
1968 12. For a pattern where the match had to start at the beginning or immediately
1969 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1970 pcre_dfa_exec() could read past the end of the passed subject if there was
1971 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1972 pcretest so that it places the subject at the end of its malloc-ed buffer.
1973
1974 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1975 exec() might read past the end of the data buffer in UTF-8 mode.
1976
1977 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1978 the data contained the byte 0x85 as part of a UTF-8 character within its
1979 first line. This applied both to normal and DFA matching.
1980
1981 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1982 /^[^d]*?$/8 failed to match "abc".
1983
1984 16. Added a missing copyright notice to pcrecpp_internal.h.
1985
1986 17. Make it more clear in the documentation that values returned from
1987 pcre_exec() in ovector are byte offsets, not character counts.
1988
1989 18. Tidied a few places to stop certain compilers from issuing warnings.
1990
1991 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1992 supplied by Stefan Weber. I made a further small update for 7.8 because
1993 there is a change of source arrangements: the pcre_searchfuncs.c module is
1994 replaced by pcre_ucd.c.
1995
1996
1997 Version 7.7 07-May-08
1998 ---------------------
1999
2000 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2001 a string to a long long, pretend we don't even have a long long." This is
2002 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2003
2004 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2005 pre-7.6 versions, which defined a global no_arg variable instead of putting
2006 it in the RE class. (See also #8 below.)
2007
2008 3. Remove a line of dead code, identified by coverity and reported by Nuno
2009 Lopes.
2010
2011 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2012
2013 (1) The include/exclude patterns were being applied to the whole pathnames
2014 of files, instead of just to the final components.
2015
2016 (2) If there was more than one level of directory, the subdirectories were
2017 skipped unless they satisfied the include/exclude conditions. This is
2018 inconsistent with GNU grep (and could even be seen as contrary to the
2019 pcregrep specification - which I improved to make it absolutely clear).
2020 The action now is always to scan all levels of directory, and just
2021 apply the include/exclude patterns to regular files.
2022
2023 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2024 --exclude_dir in the tests to avoid scanning .svn directories.
2025
2026 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2027 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2028 doesn't support NULs in patterns.
2029
2030 7. Added some missing "const"s to declarations of static tables in
2031 pcre_compile.c and pcre_dfa_exec.c.
2032
2033 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2034 caused by fix #2 above. (Subsequently also a second patch to fix the
2035 first patch. And a third patch - this was a messy problem.)
2036
2037 9. Applied Craig's patch to remove the use of push_back().
2038
2039 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2040 matching function regexec().
2041
2042 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2043 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2044 references. PCRE supports relative numbers with this syntax (I don't think
2045 Oniguruma does).
2046
2047 12. Previously, a group with a zero repeat such as (...){0} was completely
2048 omitted from the compiled regex. However, this means that if the group
2049 was called as a subroutine from elsewhere in the pattern, things went wrong
2050 (an internal error was given). Such groups are now left in the compiled
2051 pattern, with a new opcode that causes them to be skipped at execution
2052 time.
2053
2054 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2055 to the way PCRE behaves:
2056
2057 (a) A lone ] character is dis-allowed (Perl treats it as data).
2058
2059 (b) A back reference to an unmatched subpattern matches an empty string
2060 (Perl fails the current match path).
2061
2062 (c) A data ] in a character class must be notated as \] because if the
2063 first data character in a class is ], it defines an empty class. (In
2064 Perl it is not possible to have an empty class.) The empty class []
2065 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2066 The negative empty class [^] matches any one character, independently
2067 of the DOTALL setting.
2068
2069 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2070 non-existent subpattern following a character class starting with ']' and
2071 containing () gave an internal compiling error instead of "reference to
2072 non-existent subpattern". Fortunately, when the pattern did exist, the
2073 compiled code was correct. (When scanning forwards to check for the
2074 existence of the subpattern, it was treating the data ']' as terminating
2075 the class, so got the count wrong. When actually compiling, the reference
2076 was subsequently set up correctly.)
2077
2078 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2079 it was being rejected as not supported by pcre_dfa_exec(), even though
2080 other assertions are supported. I have made pcre_dfa_exec() support
2081 (*FAIL).
2082
2083 16. The implementation of 13c above involved the invention of a new opcode,
2084 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2085 cannot be changed at match time, I realized I could make a small
2086 improvement to matching performance by compiling OP_ALLANY instead of
2087 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2088 on the OP_ANY path.
2089
2090 17. Compiling pcretest on Windows with readline support failed without the
2091 following two fixes: (1) Make the unistd.h include conditional on
2092 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2093
2094 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2095 ncurses library to be included for pcretest when ReadLine support is
2096 requested, but also to allow for it to be overridden. This patch came from
2097 Daniel Bergström.
2098
2099 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2100 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2101 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2102 spotting this.
2103
2104
2105 Version 7.6 28-Jan-08
2106 ---------------------
2107
2108 1. A character class containing a very large number of characters with
2109 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2110 overflow.
2111
2112 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2113 HAVE_LONG_LONG is not defined.
2114
2115 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2116 bring them up to date and include new features. This patch includes:
2117
2118 - Fixed PH's badly added libz and libbz2 support.
2119 - Fixed a problem with static linking.
2120 - Added pcredemo. [But later removed - see 7 below.]
2121 - Fixed dftables problem and added an option.
2122 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2123 HAVE_LONG_LONG.
2124 - Added readline support for pcretest.
2125 - Added an listing of the option settings after cmake has run.
2126
2127 4. A user submitted a patch to Makefile that makes it easy to create
2128 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2129 Makefile.am that cause it to include this special target, without
2130 affecting anything else. Note that the same mingw target plus all
2131 the other distribution libraries and programs are now supported
2132 when configuring with CMake (see 6 below) instead of with
2133 Configure/Make.
2134
2135 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2136 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2137 exported in the Windows port". It has not yet been confirmed that the patch
2138 solves the problem, but it does no harm.
2139
2140 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2141 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2142 with CMake, and also correct the comment about stack recursion.
2143
2144 7. Remove the automatic building of pcredemo from the ./configure system and
2145 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2146 of a program that users should build themselves after PCRE is installed, so
2147 building it automatically is not really right. What is more, it gave
2148 trouble in some build environments.
2149
2150 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2151
2152
2153 Version 7.5 10-Jan-08
2154 ---------------------
2155
2156 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2157 values in parens when parsing an RE using the C++ wrapper."
2158
2159 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2160 Characters greater than 255 were excluded from the class instead of being
2161 included.
2162
2163 3. The same bug as (2) above applied to negated POSIX classes such as
2164 [:^space:].
2165
2166 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2167 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2168 I have changed it.
2169
2170 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2171 first named subpattern) and a construct such as (?&a) would reference the
2172 first named subpattern whose name started with "a" (in other words, the
2173 length check was missing). Both these problems are fixed. "Subpattern name
2174 expected" is now given for (?&) (a zero-length name), and this patch also
2175 makes it give the same error for \k'' (previously it complained that that
2176 was a reference to a non-existent subpattern).
2177
2178 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2179 this is right because (?- can be followed by option settings as well as by
2180 digits. I have, however, made the messages clearer.
2181
2182 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2183 than the number used in the conditional) now cause a compile-time error.
2184 This is actually not compatible with Perl, which accepts such patterns, but
2185 treats the conditional as always being FALSE (as PCRE used to), but it
2186 seems to me that giving a diagnostic is better.
2187
2188 8. Change "alphameric" to the more common word "alphanumeric" in comments
2189 and messages.
2190
2191 9. Fix two occurrences of "backslash" in comments that should have been
2192 "backspace".
2193
2194 10. Remove two redundant lines of code that can never be obeyed (their function
2195 was moved elsewhere).
2196
2197 11. The program that makes PCRE's Unicode character property table had a bug
2198 which caused it to generate incorrect table entries for sequences of
2199 characters that have the same character type, but are in different scripts.
2200 It amalgamated them into a single range, with the script of the first of
2201 them. In other words, some characters were in the wrong script. There were
2202 thirteen such cases, affecting characters in the following ranges:
2203
2204 U+002b0 - U+002c1
2205 U+0060c - U+0060d
2206 U+0061e - U+00612
2207 U+0064b - U+0065e
2208 U+0074d - U+0076d
2209 U+01800 - U+01805
2210 U+01d00 - U+01d77
2211 U+01d9b - U+01dbf
2212 U+0200b - U+0200f
2213 U+030fc - U+030fe
2214 U+03260 - U+0327f
2215 U+0fb46 - U+0fbb1
2216 U+10450 - U+1049d
2217
2218 12. The -o option (show only the matching part of a line) for pcregrep was not
2219 compatible with GNU grep in that, if there was more than one match in a
2220 line, it showed only the first of them. It now behaves in the same way as
2221 GNU grep.
2222
2223 13. If the -o and -v options were combined for pcregrep, it printed a blank
2224 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2225 does the same. The return code can be used to tell if there were any
2226 non-matching lines.
2227
2228 14. Added --file-offsets and --line-offsets to pcregrep.
2229
2230 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2231 infinitely looping recursion. The bug was that positive lookaheads were not
2232 being skipped when checking for a possible empty match (negative lookaheads
2233 and both kinds of lookbehind were skipped).
2234
2235 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2236 inclusion of <windows.h> to before rather than after the definition of
2237 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2238
2239 17. Specifying a possessive quantifier with a specific limit for a Unicode
2240 character property caused pcre_compile() to compile bad code, which led at
2241 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2242 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2243 caused the error; without that there was no problem.
2244
2245 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2246
2247 19. Added --enable-pcretest-libreadline.
2248
2249 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2250 RE::GlobalReplace(). As a result, the number of replacements returned was
2251 double what it should be. I removed one of the increments, but Craig sent a
2252 later patch that removed the other one (the right fix) and added unit tests
2253 that check the return values (which was not done before).
2254
2255 21. Several CMake things:
2256
2257 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2258 the names libpcre and libpcreposix, not just pcre and pcreposix.
2259
2260 (2) The above change means that pcretest and pcregrep are now correctly
2261 linked with the newly-built libraries, not previously installed ones.
2262
2263 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2264
2265 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2266 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2267 UTF-8 newline character). The key issue is that the pattern starts .*;
2268 this means that the match must be either at the beginning, or after a
2269 newline. The bug was in the code for advancing after a failed match and
2270 checking that the new position followed a newline. It was not taking
2271 account of UTF-8 characters correctly.
2272
2273 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2274 character classes. PCRE was not treating the sequence [:...:] as a
2275 character class unless the ... were all letters. Perl, however, seems to
2276 allow any characters between [: and :], though of course it rejects as
2277 unknown any "names" that contain non-letters, because all the known class
2278 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2279 for example, whereas PCRE did not - it did not recognize a POSIX character
2280 class. This seemed a bit dangerous, so the code has been changed to be
2281 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2282 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2283 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2284 Perl does, and where it didn't before.
2285
2286 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2287 Windows environments %n is disabled by default.
2288
2289
2290 Version 7.4 21-Sep-07
2291 ---------------------
2292
2293 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2294 means that a class such as [\s] counted as "explicit reference to CR or
2295 LF". That isn't really right - the whole point of the change was to try to
2296 help when there was an actual mention of one of the two characters. So now
2297 the change happens only if \r or \n (or a literal CR or LF) character is
2298 encountered.
2299
2300 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2301 of both had grown to the point where there were only 3 bits left.
2302 Fortunately, there was spare space in the data structure, and so I have
2303 moved the internal flags into a new 16-bit field to free up more option
2304 bits.
2305
2306 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2307 but did not set the internal JCHANGED flag - either of these is enough to
2308 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2309 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2310 start sets both bits.
2311
2312 4. Added options (at build time, compile time, exec time) to change \R from
2313 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2314
2315 5. doc/pcresyntax.html was missing from the distribution.
2316
2317 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2318 compatibility, even though it is no longer used.
2319
2320 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2321 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2322 windows.h file is present (where different names are used). [This was
2323 reversed later after testing - see 16 below.]
2324
2325 8. Changed all #include <config.h> to #include "config.h". There were also
2326 some further <pcre.h> cases that I changed to "pcre.h".
2327
2328 9. When pcregrep was used with the --colour option, it missed the line ending
2329 sequence off the lines that it output.
2330
2331 10. It was pointed out to me that arrays of string pointers cause lots of
2332 relocations when a shared library is dynamically loaded. A technique of
2333 using a single long string with a table of offsets can drastically reduce
2334 these. I have refactored PCRE in four places to do this. The result is
2335 dramatic:
2336
2337 Originally: 290
2338 After changing UCP table: 187
2339 After changing error message table: 43
2340 After changing table of "verbs" 36
2341 After changing table of Posix names 22
2342
2343 Thanks to the folks working on Gregex for glib for this insight.
2344
2345 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2346 unicode-properties was also set.
2347
2348 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2349
2350 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2351 checked only for CRLF.
2352
2353 14. Added casts to pcretest.c to avoid compiler warnings.
2354
2355 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2356
2357 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2358 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2359 entirely. This removes changes made in 7 above.
2360
2361 17. The CMake files have been updated, and there is now more information about
2362 building with CMake in the NON-UNIX-USE document.
2363
2364
2365 Version 7.3 28-Aug-07
2366 ---------------------
2367
2368 1. In the rejigging of the build system that eventually resulted in 7.1, the
2369 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2370 brackets there is not right, since it causes compilers to look for an
2371 installed pcre.h, not the version that is in the source that is being
2372 compiled (which of course may be different). I have changed it back to:
2373
2374 #include "pcre.h"
2375
2376 I have a vague recollection that the change was concerned with compiling in
2377 different directories, but in the new build system, that is taken care of
2378 by the VPATH setting the Makefile.
2379
2380 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2381 when the subject happened to end in the byte 0x85 (e.g. if the last
2382 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2383 characters but of course it shouldn't be taken as a newline when it is part
2384 of another character. The bug was that, for an unlimited repeat of . in
2385 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2386 characters when looking for a newline.
2387
2388 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2389
2390 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2391 in debug output.
2392
2393 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2394 long printing in the pcrecpp unittest when running under MinGW.
2395
2396 6. ESC_K was left out of the EBCDIC table.
2397
2398 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2399 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2400 limit also applies to "virtual nesting" when a pattern is recursive, and in
2401 this case 1000 isn't so big. I have been able to remove this limit at the
2402 expense of backing off one optimization in certain circumstances. Normally,
2403 when pcre_exec() would call its internal match() function recursively and
2404 immediately return the result unconditionally, it uses a "tail recursion"
2405 feature to save stack. However, when a subpattern that can match an empty
2406 string has an unlimited repetition quantifier, it no longer makes this
2407 optimization. That gives it a stack frame in which to save the data for
2408 checking that an empty string has been matched. Previously this was taken
2409 from the 1000-entry workspace that had been reserved. So now there is no
2410 explicit limit, but more stack is used.
2411
2412 8. Applied Daniel's patches to solve problems with the import/export magic
2413 syntax that is required for Windows, and which was going wrong for the
2414 pcreposix and pcrecpp parts of the library. These were overlooked when this
2415 problem was solved for the main library.
2416
2417 9. There were some crude static tests to avoid integer overflow when computing
2418 the size of patterns that contain repeated groups with explicit upper
2419 limits. As the maximum quantifier is 65535, the maximum group length was
2420 set at 30,000 so that the product of these two numbers did not overflow a
2421 32-bit integer. However, it turns out that people want to use groups that
2422 are longer than 30,000 bytes (though not repeat them that many times).
2423 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2424 made it possible to implement the integer overflow checks in a much more
2425 dynamic way, which I have now done. The artificial limitation on group
2426 length has been removed - we now have only the limit on the total length of
2427 the compiled pattern, which depends on the LINK_SIZE setting.
2428
2429 10. Fixed a bug in the documentation for get/copy named substring when
2430 duplicate names are permitted. If none of the named substrings are set, the
2431 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2432 empty string.
2433
2434 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2435 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2436 because the ] is interpreted as the first data character and the
2437 terminating ] is not found. PCRE has been made compatible with Perl in this
2438 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2439 cause memory overwriting.
2440
2441 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2442 string has been matched (to stop an infinite loop). It was not recognizing
2443 a conditional subpattern that could match an empty string if that
2444 subpattern was within another subpattern. For example, it looped when
2445 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2446 condition was not nested. This bug has been fixed.
2447
2448 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2449 past the start of the subject in the presence of bytes with the top bit
2450 set, for example "\x8aBCD".
2451
2452 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2453 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2454
2455 14. Optimized (?!) to (*FAIL).
2456
2457 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2458 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2459 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2460 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2461 does: it's just the validity check that is more restrictive.
2462
2463 16. Inserted checks for integer overflows during escape sequence (backslash)
2464 processing, and also fixed erroneous offset values for syntax errors during
2465 backslash processing.
2466
2467 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2468 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2469
2470 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2471 caused an overrun.
2472
2473 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2474 something other than just ASCII characters) inside a group that had an
2475 unlimited repeat caused a loop at compile time (while checking to see
2476 whether the group could match an empty string).
2477
2478 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2479 [\P{Any}] did so. (Error in the code for printing property names.)
2480
2481 21. An orphan \E inside a character class could cause a crash.
2482
2483 22. A repeated capturing bracket such as (A)? could cause a wild memory
2484 reference during compilation.
2485
2486 23. There are several functions in pcre_compile() that scan along a compiled
2487 expression for various reasons (e.g. to see if it's fixed length for look
2488 behind). There were bugs in these functions when a repeated \p or \P was
2489 present in the pattern. These operators have additional parameters compared
2490 with \d, etc, and these were not being taken into account when moving along
2491 the compiled data. Specifically:
2492
2493 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2494 length.
2495
2496 (b) An item such as \pL+ within a repeated group could cause crashes or
2497 loops.
2498
2499 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2500 "reference to non-existent subpattern" error.
2501
2502 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2503
2504 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2505 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2506
2507 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2508
2509 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2510 character were causing crashes (broken optimization).
2511
2512 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2513 \p or \P) caused a compile-time loop.
2514
2515 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2516 break. For example, the unstudied pattern [\r\n]A does not match the string
2517 "\r\nA" because change 7.0/46 below moves the current point on by two
2518 characters after failing to match at the start. However, the pattern \nA
2519 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2520 the same is true. There doesn't seem any very clean way out of this, but
2521 what I have chosen to do makes the common cases work: PCRE now takes note
2522 of whether there can be an explicit match for \r or \n anywhere in the
2523 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2524 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2525 pattern has explicit CR or LF references.
2526
2527 29. Added (*CR) etc for changing newline setting at start of pattern.
2528
2529
2530 Version 7.2 19-Jun-07
2531 ---------------------
2532
2533 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2534 which is apparently normally available under Windows.
2535
2536 2. Re-jig the pcregrep tests with different newline settings in an attempt
2537 to make them independent of the local environment's newline setting.
2538
2539 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2540
2541 4. Some of the "internals" tests were previously cut out when the link size
2542 was not 2, because the output contained actual offsets. The recent new
2543 "Z" feature of pcretest means that these can be cut out, making the tests
2544 usable with all link sizes.
2545
2546 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2547 stack recursion. This gives a massive performance boost under BSD, but just
2548 a small improvement under Linux. However, it saves one field in the frame
2549 in all cases.
2550
2551 6. Added more features from the forthcoming Perl 5.10:
2552
2553 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2554 recursion call. It refers to the nth most recently opened parentheses.
2555
2556 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2557 to be opened parentheses.
2558
2559 (c) Conditions that refer to capturing parentheses can be specified
2560 relatively, for example, (?(-2)... or (?(+3)...
2561
2562 (d) \K resets the start of the current match so that everything before
2563 is not part of it.
2564
2565 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2566
2567 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2568 reference syntax.
2569
2570 (g) (?| introduces a group in which the numbering of parentheses in each
2571 alternative starts with the same number.
2572
2573 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2574
2575 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2576 PCRE_INFO_JCHANGED.
2577
2578 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2579 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2580 for detecting groups that can match an empty string.
2581
2582 9. A pattern with a very large number of alternatives (more than several
2583 hundred) was running out of internal workspace during the pre-compile
2584 phase, where pcre_compile() figures out how much memory will be needed. A
2585 bit of new cunning has reduced the workspace needed for groups with
2586 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2587 workspace instead of running out of the 4096 that are available.
2588
2589 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2590
2591 11. Applied patch from Google to remove an optimization that didn't quite work.
2592 The report of the bug said:
2593
2594 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2595 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2596 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2597
2598 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2599 it matched the wrong number of bytes.
2600
2601
2602 Version 7.1 24-Apr-07
2603 ---------------------
2604
2605 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2606 that is more "standard", making use of automake and other Autotools. There
2607 is some re-arrangement of the files and adjustment of comments consequent
2608 on this.
2609
2610 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2611 for recursive directory scanning broke on some systems because the files
2612 are not scanned in any specific order and on different systems the order
2613 was different. A call to "sort" has been inserted into RunGrepTest for the
2614 approprate test as a short-term fix. In the longer term there may be an
2615 alternative.
2616
2617 3. I had an email from Eric Raymond about problems translating some of PCRE's
2618 man pages to HTML (despite the fact that I distribute HTML pages, some
2619 people do their own conversions for various reasons). The problems
2620 concerned the use of low-level troff macros .br and .in. I have therefore
2621 removed all such uses from the man pages (some were redundant, some could
2622 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2623 HTML has been updated to handle .nf/.fi and to complain if it encounters
2624 .br or .in.
2625
2626 4. Updated comments in configure.ac that get placed in config.h.in and also
2627 arranged for config.h to be included in the distribution, with the name
2628 config.h.generic, for the benefit of those who have to compile without
2629 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2630
2631 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2632 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2633 makevp.bat for the current PCRE, using the additional files
2634 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2635
2636 6. A Windows user reported a minor discrepancy with test 2, which turned out
2637 to be caused by a trailing space on an input line that had got lost in his
2638 copy. The trailing space was an accident, so I've just removed it.
2639
2640 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2641 that is needed.
2642
2643 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2644 as "const" (a) because they are and (b) because it helps the PHP
2645 maintainers who have recently made a script to detect big data structures
2646 in the php code that should be moved to the .rodata section. I remembered
2647 to update Builducptable as well, so it won't revert if ucptable.h is ever
2648 re-created.
2649
2650 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2651 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2652 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2653 support is not required. This saves 1.5-2K of code, which is important in
2654 some applications.
2655
2656 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2657 so as not to refer to the tables, even though these functions will never be
2658 called when UTF-8 support is disabled. Otherwise there are problems with a
2659 shared library.
2660
2661 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2662
2663 (a) It was defining its arguments as char * instead of void *.
2664
2665 (b) It was assuming that all moves were upwards in memory; this was true
2666 a long time ago when I wrote it, but is no longer the case.
2667
2668 The emulated memove() is provided for those environments that have neither
2669 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2670 is clearly not the case, as these two bugs were recently reported.
2671
2672 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2673 and Detrail to create the HTML documentation, the .txt form of the man
2674 pages, and it removes trailing spaces from listed files. It also creates
2675 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2676 case, it wraps all the #defines with #ifndefs. This script should be run
2677 before "make dist".
2678
2679 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2680 with Unicode property support.
2681
2682 (a) For a maximizing quantifier, if the two different cases of the
2683 character were of different lengths in their UTF-8 codings (there are
2684 some cases like this - I found 11), and the matching function had to
2685 back up over a mixture of the two cases, it incorrectly assumed they
2686 were both the same length.
2687
2688 (b) When PCRE was configured to use the heap rather than the stack for
2689 recursion during matching, it was not correctly preserving the data for
2690 the other case of a UTF-8 character when checking ahead for a match
2691 while processing a minimizing repeat. If the check also involved
2692 matching a wide character, but failed, corruption could cause an
2693 erroneous result when trying to check for a repeat of the original
2694 character.
2695
2696 13. Some tidying changes to the testing mechanism:
2697
2698 (a) The RunTest script now detects the internal link size and whether there
2699 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2700 values substituted by "configure". (The RunGrepTest script already did
2701 this for UTF-8.) The configure.ac script no longer substitutes the
2702 relevant variables.
2703
2704 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2705 with length and offset values. This means that the output is different
2706 for different internal link sizes. Test 2 is skipped for link sizes
2707 other than 2 because of this, bypassing the problem. Unfortunately,
2708 there was also a test in test 3 (the locale tests) that used /B and
2709 failed for link sizes other than 2. Rather than cut the whole test out,
2710 I have added a new /Z option to pcretest that replaces the length and
2711 offset values with spaces. This is now used to make test 3 independent
2712 of link size. (Test 2 will be tidied up later.)
2713
2714 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2715 segmentation fault instead of returning the appropriate error message.
2716
2717 15. In multiline mode when the newline sequence was set to "any", the pattern
2718 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2719 This doesn't seem right; it now treats the CRLF combination as the line
2720 ending, and so does not match in that case. It's only a pattern such as ^$
2721 that would hit this one: something like ^ABC$ would have failed after \r
2722 and then tried again after \r\n.
2723
2724 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2725 in an attempt to make files that differ only in their line terminators
2726 compare equal. This works on Linux.
2727
2728 17. Under certain error circumstances pcregrep might try to free random memory
2729 as it exited. This is now fixed, thanks to valgrind.
2730
2731 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2732 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2733 was because its rules for how to advance for /g after matching an empty
2734 string at the end of a line did not allow for this case. They now check for
2735 it specially.
2736
2737 20. pcretest is supposed to handle patterns and data of any length, by
2738 extending its buffers when necessary. It was getting this wrong when the
2739 buffer for a data line had to be extended.
2740
2741 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2742 CRLF as a newline sequence.
2743
2744 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2745 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2746 I have nevertheless tidied it up.
2747
2748 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2749
2750 24. Added a man page for pcre-config.
2751
2752
2753 Version 7.0 19-Dec-06
2754 ---------------------
2755
2756 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2757 moving to gcc 4.1.1.
2758
2759 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2760 sys/time.h, which is documented as needed for this function. It doesn't
2761 seem to matter on Linux, but it showed up on some releases of OS X.
2762
2763 3. It seems that there are systems where bytes whose values are greater than
2764 127 match isprint() in the "C" locale. The "C" locale should be the
2765 default when a C program starts up. In most systems, only ASCII printing
2766 characters match isprint(). This difference caused the output from pcretest
2767 to vary, making some of the tests fail. I have changed pcretest so that:
2768
2769 (a) When it is outputting text in the compiled version of a pattern, bytes
2770 other than 32-126 are always shown as hex escapes.
2771
2772 (b) When it is outputting text that is a matched part of a subject string,
2773 it does the same, unless a different locale has been set for the match
2774 (using the /L modifier). In this case, it uses isprint() to decide.
2775
2776 4. Fixed a major bug that caused incorrect computation of the amount of memory
2777 required for a compiled pattern when options that changed within the
2778 pattern affected the logic of the preliminary scan that determines the
2779 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2780 that the computed length was too small. The symptoms of this bug were
2781 either the PCRE error "internal error: code overflow" from pcre_compile(),
2782 or a glibc crash with a message such as "pcretest: free(): invalid next
2783 size (fast)". Examples of patterns that provoked this bug (shown in
2784 pcretest format) are:
2785
2786 /(?-x: )/x
2787 /(?x)(?-x: \s*#\s*)/
2788 /((?i)[\x{c0}])/8
2789 /(?i:[\x{c0}])/8
2790
2791 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2792 is now done differently.
2793
2794 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2795 wrapper classes; (b) implement a new function in the C++ scanner that is
2796 more efficient than the old way of doing things because it avoids levels of
2797 recursion in the regex matching; (c) add a paragraph to the documentation
2798 for the FullMatch() function.
2799
2800 6. The escape sequence \n was being treated as whatever was defined as
2801 "newline". Not only was this contrary to the documentation, which states
2802 that \n is character 10 (hex 0A), but it also went horribly wrong when
2803 "newline" was defined as CRLF. This has been fixed.
2804
2805 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2806 was being set to -1 for the "end of line" case (supposedly a value that no
2807 character can have). Though this value is never used (the check for end of
2808 line is "zero bytes in current character"), it caused compiler complaints.
2809 I've changed it to 0xffffffff.
2810
2811 8. In pcre_version.c, the version string was being built by a sequence of
2812 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2813 string (as it is for production releases) called a macro with an empty
2814 argument. The C standard says the result of this is undefined. The gcc
2815 compiler treats it as an empty string (which was what was wanted) but it is
2816 reported that Visual C gives an error. The source has been hacked around to
2817 avoid this problem.
2818
2819 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2820 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2821 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2822 of them did).
2823
2824 10. Originally, pcretest opened its input and output without "b"; then I was
2825 told that "b" was needed in some environments, so it was added for release
2826 5.0 to both the input and output. (It makes no difference on Unix-like
2827 systems.) Later I was told that it is wrong for the input on Windows. I've
2828 now abstracted the modes into two macros, to make it easier to fiddle with
2829 them, and removed "b" from the input mode under Windows.
2830
2831 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2832
2833 12. Added -help and --help to pcretest as an official way of being reminded
2834 of the options.
2835
2836 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2837 and pcrecpp.cc because they annoy compilers at high warning levels.
2838
2839 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2840
2841 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2842 scripts are not C programs :-) and which was not noticed because it works
2843 on Linux.
2844
2845 16. pcretest is supposed to handle any length of pattern and data line (as one
2846 line or as a continued sequence of lines) by extending its input buffer if
2847 necessary. This feature was broken for very long pattern lines, leading to
2848 a string of junk being passed to pcre_compile() if the pattern was longer
2849 than about 50K.
2850
2851 17. I have done a major re-factoring of the way pcre_compile() computes the
2852 amount of memory needed for a compiled pattern. Previously, there was code
2853 that made a preliminary scan of the pattern in order to do this. That was
2854 OK when PCRE was new, but as the facilities have expanded, it has become
2855 harder and harder to keep it in step with the real compile phase, and there
2856 have been a number of bugs (see for example, 4 above). I have now found a
2857 cunning way of running the real compile function in a "fake" mode that
2858 enables it to compute how much memory it would need, while actually only
2859 ever using a few hundred bytes of working memory and without too many
2860 tests of the mode. This should make future maintenance and development
2861 easier. A side effect of this work is that the limit of 200 on the nesting
2862 depth of parentheses has been removed (though this was never a serious
2863 limitation, I suspect). However, there is a downside: pcre_compile() now
2864 runs more slowly than before (30% or more, depending on the pattern). I
2865 hope this isn't a big issue. There is no effect on runtime performance.
2866
2867 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2868 newline (only possible for the last line of a file) and it was a
2869 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2870
2871 19. Added additional timing features to pcretest. (1) The -tm option now times
2872 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2873 separate command line item, by a number that specifies the number of
2874 repeats to use when timing. The default is 50000; this gives better
2875 precision, but takes uncomfortably long for very large patterns.
2876
2877 20. Extended pcre_study() to be more clever in cases where a branch of a
2878 subpattern has no definite first character. For example, (a*|b*)[cd] would
2879 previously give no result from pcre_study(). Now it recognizes that the
2880 first character must be a, b, c, or d.
2881
2882 21. There was an incorrect error "recursive call could loop indefinitely" if
2883 a subpattern (or the entire pattern) that was being tested for matching an
2884 empty string contained only one non-empty item after a nested subpattern.
2885 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2886 incorrectly, because the \d was being skipped in the check.
2887
2888 22. The pcretest program now has a new pattern option /B and a command line
2889 option -b, which is equivalent to adding /B to every pattern. This causes
2890 it to show the compiled bytecode, without the additional information that
2891 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2892 is the same as /B/I).
2893
2894 23. A new optimization is now able automatically to treat some sequences such
2895 as a*b as a*+b. More specifically, if something simple (such as a character
2896 or a simple class like \d) has an unlimited quantifier, and is followed by
2897 something that cannot possibly match the quantified thing, the quantifier
2898 is automatically "possessified".
2899
2900 24. A recursive reference to a subpattern whose number was greater than 39
2901 went wrong under certain circumstances in UTF-8 mode. This bug could also
2902 have affected the operation of pcre_study().
2903
2904 25. Realized that a little bit of performance could be had by replacing
2905 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2906
2907 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2908
2909 27. Possessive quantifiers such as a++ were previously implemented by turning
2910 them into atomic groups such as ($>a+). Now they have their own opcodes,
2911 which improves performance. This includes the automatically created ones
2912 from 23 above.
2913
2914 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2915 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2916 the first matched character to be a colon. This applied both to named and
2917 numbered groups.
2918
2919 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2920
2921 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2922 building PCRE on Windows easier, so I have included it in the distribution.
2923
2924 31. There is now a check in pcretest against a ridiculously large number being
2925 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2926 loop, the loop is abandoned.
2927
2928 32. Forward references to subpatterns in conditions such as (?(2)...) where
2929 subpattern 2 is defined later cause pcre_compile() to search forwards in
2930 the pattern for the relevant set of parentheses. This search went wrong
2931 when there were unescaped parentheses in a character class, parentheses
2932 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2933
2934 33. "Subroutine" calls and backreferences were previously restricted to
2935 referencing subpatterns earlier in the regex. This restriction has now
2936 been removed.
2937
2938 34. Added a number of extra features that are going to be in Perl 5.10. On the
2939 whole, these are just syntactic alternatives for features that PCRE had
2940 previously implemented using the Python syntax or my own invention. The
2941 other formats are all retained for compatibility.
2942
2943 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2944 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2945 also .NET compatible.
2946
2947 (b) A recursion or subroutine call to a named group can now be defined as
2948 (?&name) as well as (?P>name).
2949
2950 (c) A backreference to a named group can now be defined as \k<name> or
2951 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2952 5.10, are also .NET compatible.
2953
2954 (d) A conditional reference to a named group can now use the syntax
2955 (?(<name>) or (?('name') as well as (?(name).
2956
2957 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2958 groups (named and numbered) that are never evaluated inline, but can be
2959 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2960 is always false. There may be only one alternative in such a group.
2961
2962 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2963 as the simple (?(R). The condition is true only if the most recent
2964 recursion is that of the given number or name. It does not search out
2965 through the entire recursion stack.
2966
2967 (g) The escape \gN or \g{N} has been added, where N is a positive or
2968 negative number, specifying an absolute or relative reference.
2969
2970 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2971 some "unreachable code" warnings.
2972
2973 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2974 things, this adds five new scripts.
2975
2976 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2977 There were also incompatibilities regarding the handling of \Q..\E inside
2978 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2979 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2980
2981 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2982 matches an empty string, and forcibly breaks the loop. There were bugs in
2983 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2984 against aaaa the result was just "a" rather than "aaaa", for example. Two
2985 separate and independent bugs (that affected different cases) have been
2986 fixed.
2987
2988 39. Refactored the code to abolish the use of different opcodes for small
2989 capturing bracket numbers. This is a tidy that I avoided doing when I
2990 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2991 The new approach is not only tidier, it makes it possible to reduce the
2992 memory needed to fix the previous bug (38).
2993
2994 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2995 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2996 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2997 mode.
2998
2999 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3000 report.
3001
3002 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3003 copy construction and assignment in the C++ wrapper.
3004
3005 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3006 couple of bugs that could have given wrong results in the "--newline=crlf"
3007 case.
3008
3009 44. Added a number of casts and did some reorganization of signed/unsigned int
3010 variables following suggestions from Dair Grant. Also renamed the variable
3011 "this" as "item" because it is a C++ keyword.
3012
3013 45. Arranged for dftables to add
3014
3015 #include "pcre_internal.h"
3016
3017 to pcre_chartables.c because without it, gcc 4.x may remove the array
3018 definition from the final binary if PCRE is built into a static library and
3019 dead code stripping is activated.
3020
3021 46. For an unanchored pattern, if a match attempt fails at the start of a
3022 newline sequence, and the newline setting is CRLF or ANY, and the next two
3023 characters are CRLF, advance by two characters instead of one.
3024
3025
3026 Version 6.7 04-Jul-06
3027 ---------------------
3028
3029 1. In order to handle tests when input lines are enormously long, pcretest has
3030 been re-factored so that it automatically extends its buffers when
3031 necessary. The code is crude, but this _is_ just a test program. The
3032 default size has been increased from 32K to 50K.
3033
3034 2. The code in pcre_study() was using the value of the re argument before
3035 testing it for NULL. (Of course, in any sensible call of the function, it
3036 won't be NULL.)
3037
3038 3. The memmove() emulation function in pcre_internal.h, which is used on
3039 systems that lack both memmove() and bcopy() - that is, hardly ever -
3040 was missing a "static" storage class specifier.
3041
3042 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3043 containing an extended class (one that cannot be represented by a bitmap
3044 because it contains high-valued characters or Unicode property items, e.g.
3045 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3046 pattern, but PCRE should not loop if you do not (it no longer does).
3047 [Detail: two cases were found: (a) a repeated subpattern containing an
3048 extended class; (b) a recursive reference to a subpattern that followed a
3049 previous extended class. It wasn't skipping over the extended class
3050 correctly when UTF-8 mode was not set.]
3051
3052 5. A negated single-character class was not being recognized as fixed-length
3053 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3054 compile error "lookbehind assertion is not fixed length".
3055
3056 6. The RunPerlTest auxiliary script was showing an unexpected difference
3057 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3058 write a Perl script that can interpret lines of an input file either as
3059 byte characters or as UTF-8, which is what "perltest" was being required to
3060 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3061 can't do is switch easily at run time between having the "use utf8;" pragma
3062 or not. In the end, I fudged it by using the RunPerlTest script to insert
3063 "use utf8;" explicitly for the UTF-8 tests.
3064
3065 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3066 the end of the subject string, contrary to the documentation and to what
3067 Perl does. This was true of both matching functions. Now it matches only at
3068 the start of the subject and immediately after *internal* newlines.
3069
3070 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3071 a pointer to an int instead of a pointer to an unsigned long int. This
3072 caused problems on 64-bit systems.
3073
3074 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3075 instance of the 'standard' template library not being so standard".
3076
3077 10. There was no check on the number of named subpatterns nor the maximum
3078 length of a subpattern name. The product of these values is used to compute
3079 the size of the memory block for a compiled pattern. By supplying a very
3080 long subpattern name and a large number of named subpatterns, the size
3081 computation could be caused to overflow. This is now prevented by limiting
3082 the length of names to 32 characters, and the number of named subpatterns
3083 to 10,000.
3084
3085 11. Subpatterns that are repeated with specific counts have to be replicated in
3086 the compiled pattern. The size of memory for this was computed from the
3087 length of the subpattern and the repeat count. The latter is limited to
3088 65535, but there was no limit on the former, meaning that integer overflow
3089 could in principle occur. The compiled length of a repeated subpattern is
3090 now limited to 30,000 bytes in order to prevent this.
3091
3092 12. Added the optional facility to have named substrings with the same name.
3093
3094 13. Added the ability to use a named substring as a condition, using the
3095 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3096 are numbers (not recommended). Forward references are permitted.
3097
3098 14. Added forward references in named backreferences (if you see what I mean).
3099
3100 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3101 pattern could run off the end of the subject. For example, the pattern
3102 "(?s)(.{1,5})"8 did this with the subject "ab".
3103
3104 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3105 PCRE_CASELESS was set when matching characters that were quantified with ?
3106 or *.
3107
3108 17. A character class other than a single negated character that had a minimum
3109 but no maximum quantifier - for example [ab]{6,} - was not handled
3110 correctly by pce_dfa_exec(). It would match only one character.
3111
3112 18. A valid (though odd) pattern that looked like a POSIX character
3113 class but used an invalid character after [ (for example [[,abc,]]) caused
3114 pcre_compile() to give the error "Failed: internal error: code overflow" or
3115 in some cases to crash with a glibc free() error. This could even happen if
3116 the pattern terminated after [[ but there just happened to be a sequence of
3117 letters, a binary zero, and a closing ] in the memory that followed.
3118
3119 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3120 over the years. Originally (before any Unicode support), just the bottom 8
3121 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3122 output from "man perlunicode" includes this:
3123
3124 The regular expression compiler produces polymorphic opcodes. That
3125 is, the pattern adapts to the data and automatically switches to
3126 the Unicode character scheme when presented with Unicode data--or
3127 instead uses a traditional byte scheme when presented with byte
3128 data.
3129
3130 Sadly, a wide octal escape does not cause a switch, and in a string with
3131 no other multibyte characters, these octal escapes are treated as before.
3132 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3133 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3134 Unicode string.
3135
3136 I have not perpetrated such confusion in PCRE. Up till now, it took just
3137 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3138 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3139 translate to the appropriate multibyte character.
3140
3141 29. Applied some refactoring to reduce the number of warnings from Microsoft
3142 and Borland compilers. This has included removing the fudge introduced
3143 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3144 a warning about an unused variable.
3145
3146 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3147 characters since release 4.0, because Perl (from release 5.004) does not.
3148 [Or at least, is documented not to: some releases seem to be in conflict
3149 with the documentation.] However, when a pattern was studied with
3150 pcre_study() and all its branches started with \s, PCRE still included VT
3151 as a possible starting character. Of course, this did no harm; it just
3152 caused an unnecessary match attempt.
3153
3154 22. Removed a now-redundant internal flag bit that recorded the fact that case
3155 dependency changed within the pattern. This was once needed for "required
3156 byte" processing, but is no longer used. This recovers a now-scarce options
3157 bit. Also moved the least significant internal flag bit to the most-
3158 significant bit of the word, which was not previously used (hangover from
3159 the days when it was an int rather than a uint) to free up another bit for
3160 the future.
3161
3162 23. Added support for CRLF line endings as well as CR and LF. As well as the
3163 default being selectable at build time, it can now be changed at runtime
3164 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3165 specify that it is scanning data with non-default line endings.
3166
3167 24. Changed the definition of CXXLINK to make it agree with the definition of
3168 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3169
3170 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3171 recursions. This makes a big different to stack usage for some patterns.
3172
3173 26. If a subpattern containing a named recursion or subroutine reference such
3174 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3175 the space required for the compiled pattern went wrong and gave too small a
3176 value. Depending on the environment, this could lead to "Failed: internal
3177 error: code overflow at offset 49" or "glibc detected double free or
3178 corruption" errors.
3179
3180 27. Applied patches from Google (a) to support the new newline modes and (b) to
3181 advance over multibyte UTF-8 characters in GlobalReplace.
3182
3183 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3184 difference for some implementation of PCRE in some Windows version.
3185
3186 29. Added some extra testing facilities to pcretest:
3187
3188 \q<number> in a data line sets the "match limit" value
3189 \Q<number> in a data line sets the "match recursion limt" value
3190 -S <number> sets the stack size, where <number> is in megabytes
3191
3192 The -S option isn't available for Windows.
3193
3194
3195 Version 6.6 06-Feb-06
3196 ---------------------
3197
3198 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3199 in pcreposix.h. I have copied the definition from pcre.h.
3200
3201 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3202 because pcre.h is no longer a built file.
3203
3204 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3205 not normally included in the compiled code.
3206
3207
3208 Version 6.5 01-Feb-06
3209 ---------------------
3210
3211 1. When using the partial match feature with pcre_dfa_exec(), it was not
3212 anchoring the second and subsequent partial matches at the new starting
3213 point. This could lead to incorrect results. For example, with the pattern
3214 /1234/, partially matching against "123" and then "a4" gave a match.
3215
3216 2. Changes to pcregrep:
3217
3218 (a) All non-match returns from pcre_exec() were being treated as failures
3219 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3220 error message is output. Some extra information is given for the
3221 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3222 probably the only errors that are likely to be caused by users (by
3223 specifying a regex that has nested indefinite repeats, for instance).
3224 If there are more than 20 of these errors, pcregrep is abandoned.
3225
3226 (b) A binary zero was treated as data while matching, but terminated the
3227 output line if it was written out. This has been fixed: binary zeroes
3228 are now no different to any other data bytes.
3229
3230 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3231 used to set a locale for matching. The --locale=xxxx long option has
3232 been added (no short equivalent) to specify a locale explicitly on the
3233 pcregrep command, overriding the environment variables.
3234
3235 (d) When -B was used with -n, some line numbers in the output were one less
3236 than they should have been.
3237
3238 (e) Added the -o (--only-matching) option.
3239
3240 (f) If -A or -C was used with -c (count only), some lines of context were
3241 accidentally printed for the final match.
3242
3243 (g) Added the -H (--with-filename) option.
3244
3245 (h) The combination of options -rh failed to suppress file names for files
3246 that were found from directory arguments.
3247
3248 (i) Added the -D (--devices) and -d (--directories) options.
3249
3250 (j) Added the -F (--fixed-strings) option.
3251
3252 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3253
3254 (l) Added the --colo(u)r option.
3255
3256 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3257 is not present by default.
3258
3259 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3260 items such as (?R) or (?1), when the recursion could match a number of
3261 alternatives. If it matched one of the alternatives, but subsequently,
3262 outside the recursion, there was a failure, the code tried to back up into
3263 the recursion. However, because of the way PCRE is implemented, this is not
3264 possible, and the result was an incorrect result from the match.
3265
3266 In order to prevent this happening, the specification of recursion has
3267 been changed so that all such subpatterns are automatically treated as
3268 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3269
3270 4. I had overlooked the fact that, in some locales, there are characters for
3271 which isalpha() is true but neither isupper() nor islower() are true. In
3272 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3273 and ordfeminine) are like this. This affected the treatment of \w and \W
3274 when they appeared in character classes, but not when they appeared outside
3275 a character class. The bit map for "word" characters is now created
3276 separately from the results of isalnum() instead of just taking it from the
3277 upper, lower, and digit maps. (Plus the underscore character, of course.)
3278
3279 5. The above bug also affected the handling of POSIX character classes such as
3280 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3281 permanent tables. Instead, the bit maps for such a class were previously
3282 created as the appropriate unions of the upper, lower, and digit bitmaps.
3283 Now they are created by subtraction from the [[:word:]] class, which has
3284 its own bitmap.
3285
3286 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3287 It is created by subtracting the vertical space characters (\x09, \x0a,
3288 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3289 subtraction was done in the overall bitmap for a character class, meaning
3290 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3291 be recognized. This bug has been fixed.
3292
3293 7. Patches from the folks at Google:
3294
3295 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3296 real life, but is still worth protecting against".
3297
3298 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3299 regular expressions".
3300
3301 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3302 have it.
3303
3304 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3305 "configure" and the latter not, in order to fix a problem somebody had
3306 with compiling the Arg class on HP-UX.
3307
3308 (e) Improve the error-handling of the C++ wrapper a little bit.
3309
3310 (f) New tests for checking recursion limiting.
3311
3312 8. The pcre_memmove() function, which is used only if the environment does not
3313 have a standard memmove() function (and is therefore rarely compiled),
3314 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3315 returning a result (though PCRE never actually uses the result).
3316
3317 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3318 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3319 returned instead of calling malloc() with an overflowing number that would
3320 most likely cause subsequent chaos.
3321
3322 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3323
3324 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3325 with this option is matched, the nmatch and pmatch options of regexec() are
3326 ignored.
3327
3328 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3329 provided in case anyone wants to the the POSIX interface with UTF-8
3330 strings.
3331
3332 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3333 C++ linking (needed for some HP-UX environments).
3334
3335 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3336 (unused parameter) and in the pcre_printint() function (omitted "default"
3337 switch label when the default is to do nothing).
3338
3339 15. Added some code to make it possible, when PCRE is compiled as a C++
3340 library, to replace subject pointers for pcre_exec() with a smart pointer
3341 class, thus making it possible to process discontinuous strings.
3342
3343 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3344 much the same function. They were added by different people who were trying
3345 to make PCRE easy to compile on non-Unix systems. It has been suggested
3346 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3347 for compiling on Windows systems. I have therefore replaced it with
3348 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3349 defaults to "extern" for C or "extern C" for C++, which works fine on
3350 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3351 SCOPE with something explicit in config.h. In addition:
3352
3353 (a) pcreposix.h still had just "extern" instead of either of these macros;
3354 I have replaced it with PCRE_DATA_SCOPE.
3355
3356 (b) Functions such as _pcre_xclass(), which are internal to the library,
3357 but external in the C sense, all had PCRE_EXPORT in their definitions.
3358 This is apparently wrong for the Windows case, so I have removed it.
3359 (It makes no difference on Unix-like systems.)
3360
3361 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3362 of recursive calls to match(). This is different to MATCH_LIMIT because
3363 that limits the total number of calls to match(), not all of which increase
3364 the depth of recursion. Limiting the recursion depth limits the amount of
3365 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3366 when PCRE is compiled, and changed at run time. A patch from Google adds
3367 this functionality to the C++ interface.
3368
3369 18. Changes to the handling of Unicode character properties:
3370
3371 (a) Updated the table to Unicode 4.1.0.
3372
3373 (b) Recognize characters that are not in the table as "Cn" (undefined).
3374
3375 (c) I revised the way the table is implemented to a much improved format
3376 which includes recognition of ranges. It now supports the ranges that
3377 are defined in UnicodeData.txt, and it also amalgamates other
3378 characters into ranges. This has reduced the number of entries in the
3379 table from around 16,000 to around 3,000, thus reducing its size
3380 considerably. I realized I did not need to use a tree structure after
3381 all - a binary chop search is just as efficient. Having reduced the
3382 number of entries, I extended their size from 6 bytes to 8 bytes to
3383 allow for more data.
3384
3385 (d) Added support for Unicode script names via properties such as \p{Han}.
3386
3387 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3388 matching that character.
3389
3390 20. When matching a repeated Unicode property with a minimum greater than zero,
3391 (for example \pL{2,}), PCRE could look past the end of the subject if it
3392 reached it while seeking the minimum number of characters. This could
3393 happen only if some of the characters were more than one byte long, because
3394 there is a check for at least the minimum number of bytes.
3395
3396 21. Refactored the implementation of \p and \P so as to be more general, to
3397 allow for more different types of property in future. This has changed the
3398 compiled form incompatibly. Anybody with saved compiled patterns that use
3399 \p or \P will have to recompile them.
3400
3401 22. Added "Any" and "L&" to the supported property types.
3402
3403 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3404 but give a compile time error if the value is greater than 0xff.
3405
3406 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3407 accidentally not being installed or uninstalled.
3408
3409 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3410 made were to insert the current release number. This seemed silly, because
3411 it made things harder for people building PCRE on systems that don't run
3412 "configure". I have turned pcre.h into a distributed file, no longer built
3413 by "configure", with the version identification directly included. There is
3414 no longer a pcre.h.in file.
3415
3416 However, this change necessitated a change to the pcre-config script as
3417 well. It is built from pcre-config.in, and one of the substitutions was the
3418 release number. I have updated configure.ac so that ./configure now finds
3419 the release number by grepping pcre.h.
3420
3421 26. Added the ability to run the tests under valgrind.
3422
3423
3424 Version 6.4 05-Sep-05
3425 ---------------------
3426
3427 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3428 "--" to be printed when multiple files were scanned, even when none of the
3429 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3430 consider it to be a bug, and have restored the previous behaviour.
3431
3432 2. A couple of code tidies to get rid of compiler warnings.
3433
3434 3. The pcretest program used to cheat by referring to symbols in the library
3435 whose names begin with _pcre_. These are internal symbols that are not
3436 really supposed to be visible externally, and in some environments it is
3437 possible to suppress them. The cheating is now confined to including
3438 certain files from the library's source, which is a bit cleaner.
3439
3440 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3441 file's purpose clearer.
3442
3443 5. Reorganized pcre_ucp_findchar().
3444
3445
3446 Version 6.3 15-Aug-05
3447 ---------------------
3448
3449 1. The file libpcre.pc.in did not have general read permission in the tarball.
3450
3451 2. There were some problems when building without C++ support:
3452
3453 (a) If C++ support was not built, "make install" and "make test" still
3454 tried to test it.
3455
3456 (b) There were problems when the value of CXX was explicitly set. Some
3457 changes have been made to try to fix these, and ...
3458
3459 (c) --disable-cpp can now be used to explicitly disable C++ support.
3460
3461 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3462 backslash in a target when C++ was disabled. This confuses some
3463 versions of "make", apparently. Using an intermediate variable solves
3464 this. (Same for CPP_LOBJ.)
3465
3466 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3467 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3468 necessary on certain architectures.
3469
3470 4. Added a setting of -export-symbols-regex to the link command to remove
3471 those symbols that are exported in the C sense, but actually are local
3472 within the library, and not documented. Their names all begin with
3473 "_pcre_". This is not a perfect job, because (a) we have to except some
3474 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3475 available (and never for static libraries). I have made a note to try to
3476 find a way round (a) in the future.
3477
3478
3479 Version 6.2 01-Aug-05
3480 ---------------------
3481
3482 1. There was no test for integer overflow of quantifier values. A construction
3483 such as {1111111111111111} would give undefined results. What is worse, if
3484 a minimum quantifier for a parenthesized subpattern overflowed and became
3485 negative, the calculation of the memory size went wrong. This could have
3486 led to memory overwriting.
3487
3488 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3489
3490 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3491 operating environments where this matters.
3492
3493 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3494 PCRE options from within the C++ wrapper.
3495
3496 5. Named capturing subpatterns were not being correctly counted when a pattern
3497 was compiled. This caused two problems: (a) If there were more than 100
3498 such subpatterns, the calculation of the memory needed for the whole
3499 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3500 back references of the form \12, where the number was greater than 9, were
3501 not recognized as back references, even though there were sufficient
3502 previous subpatterns.
3503
3504 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3505 versions of gcc, e.g. 2.95.4.
3506
3507
3508 Version 6.1 21-Jun-05
3509 ---------------------
3510
3511 1. There was one reference to the variable "posix" in pcretest.c that was not
3512 surrounded by "#if !defined NOPOSIX".
3513
3514 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3515 the cross-check on the old pcre_info() function, for the benefit of the
3516 cut-down version of PCRE that is currently imported into Exim.
3517
3518 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3519 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3520 patterns that start (?i) but I don't think that matters. The use of (?i) is
3521 just an example; this all applies to the other options as well.
3522
3523 4. Since libtool seems to echo the compile commands it is issuing, the output
3524 from "make" can be reduced a bit by putting "@" in front of each libtool
3525 compile command.
3526
3527 5. Patch from the folks at Google for configure.in to be a bit more thorough
3528 in checking for a suitable C++ installation before trying to compile the
3529 C++ stuff. This should fix a reported problem when a compiler was present,
3530 but no suitable headers.
3531
3532 6. The man pages all had just "PCRE" as their title. I have changed them to
3533 be the relevant file name. I have also arranged that these names are
3534 retained in the file doc/pcre.txt, which is a concatenation in text format
3535 of all the man pages except the little individual ones for each function.
3536
3537 7. The NON-UNIX-USE file had not been updated for the different set of source
3538 files that come with release 6. I also added a few comments about the C++
3539 wrapper.
3540
3541
3542 Version 6.0 07-Jun-05
3543 ---------------------
3544
3545 1. Some minor internal re-organization to help with my DFA experiments.
3546
3547 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3548 didn't matter for the library itself when fully configured, but did matter
3549 when compiling without UCP support, or within Exim, where the ucp files are
3550 not imported.
3551
3552 3. Refactoring of the library code to split up the various functions into
3553 different source modules. The addition of the new DFA matching code (see
3554 below) to a single monolithic source would have made it really too
3555 unwieldy, quite apart from causing all the code to be include in a
3556 statically linked application, when only some functions are used. This is
3557 relevant even without the DFA addition now that patterns can be compiled in
3558 one application and matched in another.
3559
3560 The downside of splitting up is that there have to be some external
3561 functions and data tables that are used internally in different modules of
3562 the library but which are not part of the API. These have all had their
3563 names changed to start with "_pcre_" so that they are unlikely to clash
3564 with other external names.
3565
3566 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3567 a different (DFA) algorithm. Although it is slower than the original
3568 function, it does have some advantages for certain types of matching
3569 problem.
3570
3571 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3572 including restarting after a partial match.
3573
3574 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3575 defined when compiling for Windows was sent to me. I have put it into the
3576 code, though I have no means of testing or verifying it.
3577
3578 7. Added the pcre_refcount() auxiliary function.
3579
3580 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3581 match before or at the first newline in the subject string. In pcretest,
3582 the /f option on a pattern can be used to set this.
3583
3584 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3585 would behave wrongly. This has been present in PCRE since release 4.0.
3586
3587 10. A number of changes to the pcregrep command:
3588
3589 (a) Refactored how -x works; insert ^(...)$ instead of setting
3590 PCRE_ANCHORED and checking the length, in preparation for adding
3591 something similar for -w.
3592
3593 (b) Added the -w (match as a word) option.
3594
3595 (c) Refactored the way lines are read and buffered so as to have more
3596 than one at a time available.
3597
3598 (d) Implemented a pcregrep test script.
3599
3600 (e) Added the -M (multiline match) option. This allows patterns to match
3601 over several lines of the subject. The buffering ensures that at least
3602 8K, or the rest of the document (whichever is the shorter) is available
3603 for matching (and similarly the previous 8K for lookbehind assertions).
3604
3605 (f) Changed the --help output so that it now says
3606
3607 -w, --word-regex(p)
3608
3609 instead of two lines, one with "regex" and the other with "regexp"
3610 because that confused at least one person since the short forms are the
3611 same. (This required a bit of code, as the output is generated
3612 automatically from a table. It wasn't just a text change.)
3613
3614 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3615 option but starts with a hyphen. Could be a pattern or a path name
3616 starting with a hyphen, for instance.
3617
3618 (h) "-" can be given as a file name to represent stdin.
3619
3620 (i) When file names are being printed, "(standard input)" is used for
3621 the standard input, for compatibility with GNU grep. Previously
3622 "<stdin>" was used.
3623
3624 (j) The option --label=xxx can be used to supply a name to be used for
3625 stdin when file names are being printed. There is no short form.
3626
3627 (k) Re-factored the options decoding logic because we are going to add
3628 two more options that take data. Such options can now be given in four
3629 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3630
3631 (l) Added the -A, -B, and -C options for requesting that lines of context
3632 around matches be printed.
3633
3634 (m) Added the -L option to print the names of files that do not contain
3635 any matching lines, that is, the complement of -l.
3636
3637 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3638 continue to scan other files.
3639
3640 (o) The -s option was incorrectly implemented. For compatibility with other
3641 greps, it now suppresses the error message for a non-existent or non-
3642 accessible file (but not the return code). There is a new option called
3643 -q that suppresses the output of matching lines, which was what -s was
3644 previously doing.
3645
3646 (p) Added --include and --exclude options to specify files for inclusion
3647 and exclusion when recursing.
3648
3649 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3650 Hopefully, it now does.
3651
3652 12. Missing cast in pcre_study().
3653
3654 13. Added an "uninstall" target to the makefile.
3655
3656 14. Replaced "extern" in the function prototypes in Makefile.in with
3657 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3658 world, but is set differently for Windows.
3659
3660 15. Added a second compiling function called pcre_compile2(). The only
3661 difference is that it has an extra argument, which is a pointer to an
3662 integer error code. When there is a compile-time failure, this is set
3663 non-zero, in addition to the error test pointer being set to point to an
3664 error message. The new argument may be NULL if no error number is required
3665 (but then you may as well call pcre_compile(), which is now just a
3666 wrapper). This facility is provided because some applications need a
3667 numeric error indication, but it has also enabled me to tidy up the way
3668 compile-time errors are handled in the POSIX wrapper.
3669
3670 16. Added VPATH=.libs to the makefile; this should help when building with one
3671 prefix path and installing with another. (Or so I'm told by someone who
3672 knows more about this stuff than I do.)
3673
3674 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3675 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3676 match everything, including newlines. This is not POSIX-compatible, but
3677 somebody wanted the feature. From pcretest it can be activated by using
3678 both the P and the s flags.
3679
3680 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3681
3682 19. libpcre.pc was being incorrectly installed as executable.
3683
3684 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3685 it now also looks for '\r' so that it will work unmodified on Windows.
3686
3687 21. Added Google's contributed C++ wrapper to the distribution.
3688
3689 22. Added some untidy missing memory free() calls in pcretest, to keep
3690 Electric Fence happy when testing.
3691
3692
3693
3694 Version 5.0 13-Sep-04
3695 ---------------------
3696
3697 1. Internal change: literal characters are no longer packed up into items
3698 containing multiple characters in a single byte-string. Each character
3699 is now matched using a separate opcode. However, there may be more than one
3700 byte in the character in UTF-8 mode.
3701
3702 2. The pcre_callout_block structure has two new fields: pattern_position and
3703 next_item_length. These contain the offset in the pattern to the next match
3704 item, and its length, respectively.
3705
3706 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3707 insertion of callouts before each pattern item. Added the /C option to
3708 pcretest to make use of this.
3709
3710 4. On the advice of a Windows user, the lines
3711
3712 #if defined(_WIN32) || defined(WIN32)
3713 _setmode( _fileno( stdout ), 0x8000 );
3714 #endif /* defined(_WIN32) || defined(WIN32) */
3715
3716 have been added to the source of pcretest. This apparently does useful
3717 magic in relation to line terminators.
3718
3719 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3720 for the benefit of those environments where the "b" makes a difference.
3721
3722 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3723 to know about it. I have put a hack into configure.in that adds in code
3724 to set GCC=yes if CC=icc. This seems to end up at a point in the
3725 generated configure script that is early enough to affect the setting of
3726 compiler options, which is what is needed, but I have no means of testing
3727 whether it really works. (The user who reported this had patched the
3728 generated configure script, which of course I cannot do.)
3729
3730 LATER: After change 22 below (new libtool files), the configure script
3731 seems to know about icc (and also ecc). Therefore, I have commented out
3732 this hack in configure.in.
3733
3734 7. Added support for pkg-config (2 patches were sent in).
3735
3736 8. Negated POSIX character classes that used a combination of internal tables
3737 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3738 [[:^ascii]]. Typically, they would match almost any characters. The other
3739 POSIX classes were not broken in this way.
3740
3741 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3742 to find the match, as PCRE was deluded into thinking that the match had to
3743 start at the start point or following a newline. The same bug applied to
3744 patterns with negative forward assertions or any backward assertions
3745 preceding ".*" at the start, unless the pattern required a fixed first
3746 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3747
3748 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3749 starting at the last subject character, bytes beyond the end of the subject
3750 string were read.
3751
3752 11. Renamed the variable "class" as "classbits" to make life easier for C++
3753 users. (Previously there was a macro definition, but it apparently wasn't
3754 enough.)
3755
3756 12. Added the new field "tables" to the extra data so that tables can be passed
3757 in at exec time, or the internal tables can be re-selected. This allows
3758 a compiled regex to be saved and re-used at a later time by a different
3759 program that might have everything at different addresses.
3760
3761 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3762 -R library as well as a -L library.
3763
3764 14. The debugging options of pcretest (-d on the command line or D on a
3765 pattern) showed incorrect output for anything following an extended class
3766 that contained multibyte characters and which was followed by a quantifier.
3767
3768 15. Added optional support for general category Unicode character properties
3769 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3770 support. It adds about 90K to the size of the library. The meanings of the
3771 inbuilt class escapes such as \d and \s have NOT been changed.
3772
3773 16. Updated pcredemo.c to include calls to free() to release the memory for the
3774 compiled pattern.
3775
3776 17. The generated file chartables.c was being created in the source directory
3777 instead of in the building directory. This caused the build to fail if the
3778 source directory was different from the building directory, and was
3779 read-only.
3780
3781 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3782 file. No doubt somebody will tell me if they don't make sense... Also added
3783 Dan Mooney's comments about building on OpenVMS.
3784
3785 19. Added support for partial matching via the PCRE_PARTIAL option for
3786 pcre_exec() and the \P data escape in pcretest.
3787
3788 20. Extended pcretest with 3 new pattern features:
3789
3790 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3791 write the compiled pattern to the file whose name is "rest-of-line".
3792 This is a straight binary dump of the data, with the saved pointer to
3793 the character tables forced to be NULL. The study data, if any, is
3794 written too. After writing, pcretest reads a new pattern.
3795
3796 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3797 compiled pattern from the given file. There must not be any
3798 occurrences of "<" in the file name (pretty unlikely); if there are,
3799 pcretest will instead treat the initial "<" as a pattern delimiter.
3800 After reading in the pattern, pcretest goes on to read data lines as
3801 usual.
3802
3803 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3804 and 16-bit fields in a compiled pattern, to simulate a pattern that
3805 was compiled on a host of opposite endianness.
3806
3807 21. The pcre-exec() function can now cope with patterns that were compiled on
3808 hosts of opposite endianness, with this restriction:
3809
3810 As for any compiled expression that is saved and used later, the tables
3811 pointer field cannot be preserved; the extra_data field in the arguments
3812 to pcre_exec() should be used to pass in a tables address if a value
3813 other than the default internal tables were used at compile time.
3814
3815 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3816 now diagnosed as an error. Previously, most of the time, a negative number
3817 would have been treated as zero, but if in addition "ovector" was passed as
3818 NULL, a crash could occur.
3819
3820 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3821 new versions from the libtool 1.5 distribution (the last one is a copy of
3822 a file called libtool.m4). This seems to have fixed the need to patch
3823 "configure" to support Darwin 1.3 (which I used to do). However, I still
3824 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3825 workstation).
3826
3827 24. Changed the PCRE licence to be the more standard "BSD" licence.
3828
3829
3830 Version 4.5 01-Dec-03
3831 ---------------------
3832
3833 1. There has been some re-arrangement of the code for the match() function so
3834 that it can be compiled in a version that does not call itself recursively.
3835 Instead, it keeps those local variables that need separate instances for
3836 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3837 needs to "recurse". Keeping track of where control must go is done by means
3838 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3839 hide most of the details from the main code, and operates only if
3840 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3841 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3842 operating.
3843
3844 To make it easier for callers to provide specially tailored get/free
3845 functions for this usage, two new functions, pcre_stack_malloc, and
3846 pcre_stack_free, are used. They are always called in strict stacking order,
3847 and the size of block requested is always the same.
3848
3849 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3850 PCRE has been compiled to use the stack or the heap for recursion. The
3851 -C option of pcretest uses this to show which version is compiled.
3852
3853 A new data escape \S, is added to pcretest; it causes the amounts of store
3854 obtained and freed by both kinds of malloc/free at match time to be added
3855 to the output.
3856
3857 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3858 what's available on my current Linux desktop machine.
3859
3860 3. When matching a UTF-8 string, the test for a valid string at the start has
3861 been extended. If start_offset is not zero, PCRE now checks that it points
3862 to a byte that is the start of a UTF-8 character. If not, it returns
3863 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3864 this is necessary because there may be backward assertions in the pattern.
3865 When matching the same subject several times, it may save resources to use
3866 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3867
3868 4. The code for checking the validity of UTF-8 strings has been tightened so
3869 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3870 containing "overlong sequences".
3871
3872 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3873 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3874 should have been "&", but it just so happened that all the cases this let
3875 through by mistake were picked up later in the function.
3876
3877 6. I had used a variable called "isblank" - this is a C99 function, causing
3878 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3879
3880 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3881 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3882 so the version is shown; (c) stop "make test" echoing "./RunTest".
3883
3884 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3885
3886 9. The prototype for memmove() for systems that don't have it was using
3887 size_t, but the inclusion of the header that defines size_t was later. I've
3888 moved the #includes for the C headers earlier to avoid this.
3889
3890 10. Added some adjustments to the code to make it easier to compiler on certain
3891 special systems:
3892
3893 (a) Some "const" qualifiers were missing.
3894 (b) Added the macro EXPORT before all exported functions; by default this
3895 is defined to be empty.
3896 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3897 that it reads its output file name as an argument instead of writing
3898 to the standard output and assuming this can be redirected.
3899
3900 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3901 class containing characters with values greater than 255, PCRE compilation
3902 went into a loop.
3903
3904 12. A recursive reference to a subpattern that was within another subpattern
3905 that had a minimum quantifier of zero caused PCRE to crash. For example,
3906 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3907 recursion. If the recursively-called subpattern itself had a zero repeat,
3908 that was OK.
3909
3910 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3911 buffer into which it was copied (for escape processing) was still set at
3912 1024, so long lines caused crashes.
3913
3914 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3915 "internal error: code overflow...". This applied to any character class
3916 that was followed by a possessive quantifier.
3917
3918 15. Modified the Makefile to add libpcre.la as a prerequisite for
3919 libpcreposix.la because I was told this is needed for a parallel build to
3920 work.
3921
3922 16. If a pattern that contained .* following optional items at the start was
3923 studied, the wrong optimizing data was generated, leading to matching
3924 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3925 matching string must start with a or b or c. The correct conclusion for
3926 this pattern is that a match can start with any character.
3927
3928
3929 Version 4.4 13-Aug-03
3930 ---------------------
3931
3932 1. In UTF-8 mode, a character class containing characters with values between
3933 127 and 255 was not handled correctly if the compiled pattern was studied.
3934 In fixing this, I have also improved the studying algorithm for such
3935 classes (slightly).
3936
3937 2. Three internal functions had redundant arguments passed to them. Removal
3938 might give a very teeny performance improvement.
3939
3940 3. Documentation bug: the value of the capture_top field in a callout is *one
3941 more than* the number of the hightest numbered captured substring.
3942
3943 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3944 in incorrectly linking with a previously installed version. They now link
3945 explicitly with libpcre.la.
3946
3947 5. configure.in no longer needs to recognize Cygwin specially.
3948
3949 6. A problem in pcre.in for Windows platforms is fixed.
3950
3951 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3952 pcretest, it used to include the size of the study block as part of its
3953 output. Unfortunately, the structure contains a field that has a different
3954 size on different hardware architectures. This meant that the tests that
3955 showed this size failed. As the block is currently always of a fixed size,
3956 this information isn't actually particularly useful in pcretest output, so
3957 I have just removed it.
3958
3959 8. Three pre-processor statements accidentally did not start in column 1.
3960 Sadly, there are *still* compilers around that complain, even though
3961 standard C has not required this for well over a decade. Sigh.
3962
3963 9. In pcretest, the code for checking callouts passed small integers in the
3964 callout_data field, which is a void * field. However, some picky compilers
3965 complained about the casts involved for this on 64-bit systems. Now
3966 pcretest passes the address of the small integer instead, which should get
3967 rid of the warnings.
3968
3969 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3970 both compile and run time, and gives an error if an invalid UTF-8 sequence
3971 is found. There is a option for disabling this check in cases where the
3972 string is known to be correct and/or the maximum performance is wanted.
3973
3974 11. In response to a bug report, I changed one line in Makefile.in from
3975
3976 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3977 to
3978 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3979
3980 to look similar to other lines, but I have no way of telling whether this
3981 is the right thing to do, as I do not use Windows. No doubt I'll get told
3982 if it's wrong...
3983
3984
3985 Version 4.3 21-May-03
3986 ---------------------
3987
3988 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3989 Makefile.
3990
3991 2. Some refactoring to improve the quality of the code:
3992
3993 (i) The utf8_table... variables are now declared "const".
3994
3995 (ii) The code for \cx, which used the "case flipping" table to upper case
3996 lower case letters, now just substracts 32. This is ASCII-specific,
3997 but the whole concept of \cx is ASCII-specific, so it seems
3998 reasonable.
3999
4000 (iii) PCRE was using its character types table to recognize decimal and
4001 hexadecimal digits in the pattern. This is silly, because it handles
4002 only 0-9, a-f, and A-F, but the character types table is locale-
4003 specific, which means strange things might happen. A private
4004 table is now used for this - though it costs 256 bytes, a table is
4005 much faster than multiple explicit tests. Of course, the standard
4006 character types table is still used for matching digits in subject
4007 strings against \d.
4008
4009 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4010 ending in _t are). So I've renamed it as ESC_tee.
4011
4012 3. The first argument for regexec() in the POSIX wrapper should have been
4013 defined as "const".
4014
4015 4. Changed pcretest to use malloc() for its buffers so that they can be
4016 Electric Fenced for debugging.
4017
4018 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4019 to read one or more bytes before the start of the subject string. Often this
4020 had no effect on PCRE's behaviour, but in some circumstances it could
4021 provoke a segmentation fault.
4022
4023 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4024 to try to read one or more bytes before the start of the subject string.
4025
4026 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4027 UTF-8 support could misbehave in various ways if the subject string
4028 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4029 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4030 back over UTF-8 characters.)
4031
4032
4033 Version 4.2 14-Apr-03
4034 ---------------------
4035
4036 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4037
4038 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4039 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4040 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4041 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4042 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4043 and BUILD_EXEEXT
4044 Note: automatic setting of the BUILD variables is not yet working
4045 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4046 compile-time but not at link-time
4047 [LINK]: use for linking executables only
4048 make different versions for Windows and non-Windows
4049 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4050 libraries
4051 [LINK_FOR_BUILD]: new variable
4052 [OBJEXT]: use throughout
4053 [EXEEXT]: use throughout
4054 <winshared>: new target
4055 <wininstall>: new target
4056 <dftables.o>: use native compiler
4057 <dftables>: use native linker
4058 <install>: handle Windows platform correctly
4059 <clean>: ditto
4060 <check>: ditto
4061 copy DLL to top builddir before testing
4062
4063 As part of these changes, -no-undefined was removed again. This was reported
4064 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4065 in any case.
4066
4067 3. Some tidies to get rid of compiler warnings:
4068
4069 . In the match_data structure, match_limit was an unsigned long int, whereas
4070 match_call_count was an int. I've made them both unsigned long ints.
4071
4072 . In pcretest the fact that a const uschar * doesn't automatically cast to
4073 a void * provoked a warning.
4074
4075 . Turning on some more compiler warnings threw up some "shadow" variables
4076 and a few more missing casts.
4077
4078 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4079 option, a class that contained a single character with a value between 128
4080 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4081
4082 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4083 option, a class that contained several characters, but with at least one
4084 whose value was between 128 and 255 caused PCRE to crash.
4085
4086
4087 Version 4.1 12-Mar-03
4088 ---------------------
4089
4090 1. Compiling with gcc -pedantic found a couple of places where casts were
4091 needed, and a string in dftables.c that was longer than standard compilers are
4092 required to support.
4093
4094 2. Compiling with Sun's compiler found a few more places where the code could
4095 be tidied up in order to avoid warnings.
4096
4097 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4098 first of these names is deprecated in the latest Autoconf in favour of the name
4099 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4100 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4101 analogy I have changed it to CFLAGS_FOR_BUILD.
4102
4103 4. Added -no-undefined to the linking command in the Makefile, because this is
4104 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4105 linking step for the pcreposix library.
4106
4107 5. PCRE was failing to diagnose the case of two named groups with the same
4108 name.
4109
4110 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4111 literal character that is needed in the subject for a match, and scans along to
4112 ensure that it is present before embarking on the full matching process. This
4113 saves time in cases of nested unlimited repeats that are never going to match.
4114 Problem: the scan can take a lot of time if the subject is very long (e.g.
4115 megabytes), thus penalizing straightforward matches. It is now done only if the
4116 amount of subject to be scanned is less than 1000 bytes.
4117
4118 7. A lesser problem with the same optimization is that it was recording the
4119 first character of an anchored pattern as "needed", thus provoking a search
4120 right along the subject, even when the first match of the pattern was going to
4121 fail. The "needed" character is now not set for anchored patterns, unless it
4122 follows something in the pattern that is of non-fixed length. Thus, it still
4123 fulfils its original purpose of finding quick non-matches in cases of nested
4124 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4125
4126
4127 Version 4.0 17-Feb-03
4128 ---------------------
4129
4130 1. If a comment in an extended regex that started immediately after a meta-item
4131 extended to the end of string, PCRE compiled incorrect data. This could lead to
4132 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4133
4134 2. Moved to autoconf 2.53 and libtool 1.4.2.
4135
4136 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4137 the special perltest8 script is no longer needed - all the tests can be run
4138 from a single perltest script.
4139
4140 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4141 by \s. It has now been removed in PCRE. This means it isn't recognized as
4142 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4143 class [:space:] *does* include VT, thereby creating a mess.
4144
4145 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4146 space and tab.
4147
4148 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4149 its new features into the main test script, reducing the number of scripts.
4150
4151 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4152 were backward compatible, and made the (?i) apply to the whole pattern, as if
4153 /i were given. Now it behaves more logically, and applies the option setting
4154 only to what follows. PCRE has been changed to follow suit. However, if it
4155 finds options settings right at the start of the pattern, it extracts them into
4156 the global options, as before. Thus, they show up in the info data.
4157
4158 8. Added support for the \Q...\E escape sequence. Characters in between are
4159 treated as literals. This is slightly different from Perl in that $ and @ are
4160 also handled as literals inside the quotes. In Perl, they will cause variable
4161 interpolation. Note the following examples:
4162
4163 Pattern PCRE matches Perl matches
4164
4165 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4166 \Qabc\$xyz\E abc\$xyz abc\$xyz
4167 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4168
4169 For compatibility with Perl, \Q...\E sequences are recognized inside character
4170 classes as well as outside them.
4171
4172 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4173 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4174 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4175 signed/unsigned warnings.
4176
4177 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4178 option for pcretest, so I've replaced it by a simple function that does just
4179 that job.
4180
4181 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4182 "pcregrep -".
4183
4184 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4185 Java package. This provides some syntactic sugar for simple cases of what my
4186 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4187 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4188 item, you can use this simplified notation. Note that only makes sense with
4189 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4190 greediness, whatever the setting of the PCRE_UNGREEDY option.
4191
4192 13. A change of greediness default within a pattern was not taking effect at
4193 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4194 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4195 was abstracted outside.
4196
4197 14. PCRE now supports the \G assertion. It is true when the current matching
4198 position is at the start point of the match. This differs from \A when the
4199 starting offset is non-zero. Used with the /g option of pcretest (or similar
4200 code), it works in the same way as it does for Perl's /g option. If all
4201 alternatives of a regex begin with \G, the expression is anchored to the start
4202 match position, and the "anchored" flag is set in the compiled expression.
4203
4204 15. Some bugs concerning the handling of certain option changes within patterns
4205 have been fixed. These applied to options other than (?ims). For example,
4206 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4207 been the other way round. Some of this was related to change 7 above.
4208
4209 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4210 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4211 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4212 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4213
4214 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4215 mode. Unlike ".", it always matches newline, whatever the setting of
4216 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4217 assertions. Perl allows it, but it doesn't (in general) work because it can't
4218 calculate the length of the lookbehind. At least, that's the case for Perl
4219 5.8.0 - I've been told they are going to document that it doesn't work in
4220 future.
4221
4222 18. Added an error diagnosis for escapes that PCRE does not support: these are
4223 \L, \l, \N, \P, \p, \U, \u, and \X.
4224
4225 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4226 reading past the end of the pattern in cases such as /[abcd/.
4227
4228 20. PCRE was getting more memory than necessary for patterns with classes that
4229 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4230
4231 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4232 compiling PCRE for use with Virtual Pascal.
4233
4234 22. Small fix to the Makefile to make it work properly if the build is done
4235 outside the source tree.
4236
4237 23. Added a new extension: a condition to go with recursion. If a conditional
4238 subpattern starts with (?(R) the "true" branch is used if recursion has
4239 happened, whereas the "false" branch is used only at the top level.
4240
4241 24. When there was a very long string of literal characters (over 255 bytes
4242 without UTF support, over 250 bytes with UTF support), the computation of how
4243 much memory was required could be incorrect, leading to segfaults or other
4244 strange effects.
4245
4246 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4247 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4248 there was a subsequent back reference to those brackets. This meant that, for
4249 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4250 possible to check for precisely this case. All we can do is abandon the
4251 optimization if .* occurs inside capturing brackets when there are any back
4252 references whatsoever. (See below for a better fix that came later.)
4253
4254 26. The handling of the optimization for finding the first character of a
4255 non-anchored pattern, and for finding a character that is required later in the
4256 match were failing in some cases. This didn't break the matching; it just
4257 failed to optimize when it could. The way this is done has been re-implemented.
4258
4259 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4260
4261 28. Added a new feature that provides some of the functionality that Perl
4262 provides with (?{...}). The facility is termed a "callout". The way it is done
4263 in PCRE is for the caller to provide an optional function, by setting
4264 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4265 global variable. By default it is unset, which disables all calling out. To get
4266 the function called, the regex must include (?C) at appropriate points. This
4267 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4268 This provides a means of identifying different callout points. When PCRE
4269 reaches such a point in the regex, if pcre_callout has been set, the external
4270 function is called. It is provided with data in a structure called
4271 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4272 matching continues; if it returns a non-zero value, the match at the current
4273 point fails. However, backtracking will occur if possible. [This was changed
4274 later and other features added - see item 49 below.]
4275
4276 29. pcretest is upgraded to test the callout functionality. It provides a
4277 callout function that displays information. By default, it shows the start of
4278 the match and the current position in the text. There are some new data escapes
4279 to vary what happens:
4280
4281 \C+ in addition, show current contents of captured substrings
4282 \C- do not supply a callout function
4283 \C!n return 1 when callout number n is reached
4284 \C!n!m return 1 when callout number n is reached for the mth time
4285
4286 30. If pcregrep was called with the -l option and just a single file name, it
4287 output "<stdin>" if a match was found, instead of the file name.
4288
4289 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4290 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4291 pcre_exec(). This saves a malloc/free per call. The default value of
4292 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4293 when configuring.
4294
4295 32. The default maximum size of a compiled pattern is 64K. There have been a
4296 few cases of people hitting this limit. The code now uses macros to handle the
4297 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4298 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4299 configuring. Tests 2 and 5 work only with 2-byte links because they output
4300 debugging information about compiled patterns.
4301
4302 33. Internal code re-arrangements:
4303
4304 (a) Moved the debugging function for printing out a compiled regex into
4305 its own source file (printint.c) and used #include to pull it into
4306 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4307 separate copies.
4308
4309 (b) Defined the list of op-code names for debugging as a macro in
4310 internal.h so that it is next to the definition of the opcodes.
4311
4312 (c) Defined a table of op-code lengths for simpler skipping along compiled
4313 code. This is again a macro in internal.h so that it is next to the
4314 definition of the opcodes.
4315
4316 34. Added support for recursive calls to individual subpatterns, along the
4317 lines of Robin Houston's patch (but implemented somewhat differently).
4318
4319 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4320 allow it to read and process whole directories in Win32. This code was
4321 contributed by Lionel Fourquaux; it has not been tested by me.
4322
4323 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4324 used to name a group. Names consist of alphanumerics and underscores, and must
4325 be unique. Back references use the syntax (?P=name) and recursive calls use
4326 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4327 numbers. The function pcre_fullinfo() can be used after compilation to extract
4328 a name/number map. There are three relevant calls:
4329
4330 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4331 PCRE_INFO_NAMECOUNT yields the number of entries
4332 PCRE_INFO_NAMETABLE yields a pointer to the map.
4333
4334 The map is a vector of fixed-size entries. The size of each entry depends on
4335 the length of the longest name used. The first two bytes of each entry are the
4336 group number, most significant byte first. There follows the corresponding
4337 name, zero terminated. The names are in alphabetical order.
4338
4339 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4340 case instead of 255. Making it the same both with and without UTF-8 support
4341 means that the same test output works with both.
4342
4343 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4344 calling malloc() with a zero argument.
4345
4346 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4347 optimization. I've improved things by keeping a bitmap of backreferences with
4348 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4349 fact referenced, the optimization can be applied. It is unlikely that a
4350 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4351 the match to follow \n) will appear inside brackets with a number greater than
4352 31, but if it does, any back reference > 31 suppresses the optimization.
4353
4354 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4355 of disabling numbered capturing parentheses. Any opening parenthesis that is
4356 not followed by ? behaves as if it were followed by ?: but named parentheses
4357 can still be used for capturing (and they will acquire numbers in the usual
4358 way).
4359
4360 41. Redesigned the return codes from the match() function into yes/no/error so
4361 that errors can be passed back from deep inside the nested calls. A malloc
4362 failure while inside a recursive subpattern call now causes the
4363 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4364
4365 42. It is now possible to set a limit on the number of times the match()
4366 function is called in a call to pcre_exec(). This facility makes it possible to
4367 limit the amount of recursion and backtracking, though not in a directly
4368 obvious way, because the match() function is used in a number of different
4369 circumstances. The count starts from zero for each position in the subject
4370 string (for non-anchored patterns). The default limit is, for compatibility, a
4371 large number, namely 10 000 000. You can change this in two ways:
4372
4373 (a) When configuring PCRE before making, you can use --with-match-limit=n
4374 to set a default value for the compiled library.
4375
4376 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4377 a different value is set. See 45 below.
4378
4379 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4380
4381 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4382 of things that can be changed at compile time. The first argument specifies
4383 what is wanted and the second points to where the information is to be placed.
4384 The current list of available information is:
4385
4386 PCRE_CONFIG_UTF8
4387
4388 The output is an integer that is set to one if UTF-8 support is available;
4389 otherwise it is set to zero.
4390
4391 PCRE_CONFIG_NEWLINE
4392
4393 The output is an integer that it set to the value of the code that is used for
4394 newline. It is either LF (10) or CR (13).
4395
4396 PCRE_CONFIG_LINK_SIZE
4397
4398 The output is an integer that contains the number of bytes used for internal
4399 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4400
4401 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4402
4403 The output is an integer that contains the threshold above which the POSIX
4404 interface uses malloc() for output vectors. See item 31 above.
4405
4406 PCRE_CONFIG_MATCH_LIMIT
4407
4408 The output is an unsigned integer that contains the default limit of the number
4409 of match() calls in a pcre_exec() execution. See 42 above.
4410
4411 44. pcretest has been upgraded by the addition of the -C option. This causes it
4412 to extract all the available output from the new pcre_config() function, and to
4413 output it. The program then exits immediately.
4414
4415 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4416 order to support additional features. One way would have been to define
4417 pcre_exec2() (for example) with extra arguments, but this would not have been
4418 extensible, and would also have required all calls to the original function to
4419 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4420 is used for passing in "extra" data from pcre_study().
4421
4422 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4423 contains the following fields:
4424
4425 flags a bitmap indicating which of the following fields are set
4426 study_data opaque data from pcre_study()
4427 match_limit a way of specifying a limit on match() calls for a specific
4428 call to pcre_exec()
4429 callout_data data for callouts (see 49 below)
4430
4431 The flag bits are also defined in pcre.h, and are
4432
4433 PCRE_EXTRA_STUDY_DATA
4434 PCRE_EXTRA_MATCH_LIMIT
4435 PCRE_EXTRA_CALLOUT_DATA
4436
4437 The pcre_study() function now returns one of these new pcre_extra blocks, with
4438 the actual study data pointed to by the study_data field, and the
4439 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4440 before. That is, this change is entirely upwards-compatible and requires no
4441 change to existing code.
4442
4443 If you want to pass in additional data to pcre_exec(), you can either place it
4444 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4445 block.
4446
4447 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4448 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4449 times with different match limits, until it finds the minimum value needed for
4450 pcre_exec() to complete. The value is then output. This can be instructive; for
4451 most simple matches the number is quite small, but for pathological cases it
4452 gets very large very quickly.
4453
4454 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4455 returns the size of the data block pointed to by the study_data field in a
4456 pcre_extra block, that is, the value that was passed as the argument to
4457 pcre_malloc() when PCRE was getting memory in which to place the information
4458 created by pcre_study(). The fourth argument should point to a size_t variable.
4459 pcretest has been extended so that this information is shown after a successful
4460 pcre_study() call when information about the compiled regex is being displayed.
4461
4462 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4463 because what follows is always an absolute path. (Later: it turns out that this
4464 is more than cosmetic for MinGW, because it doesn't like empty path
4465 components.)
4466
4467 49. Some changes have been made to the callout feature (see 28 above):
4468
4469 (i) A callout function now has three choices for what it returns:
4470
4471 0 => success, carry on matching
4472 > 0 => failure at this point, but backtrack if possible
4473 < 0 => serious error, return this value from pcre_exec()
4474
4475 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4476 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4477 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4478 use by callout functions. It will never be used by PCRE itself.
4479
4480 (ii) The pcre_extra structure (see 45 above) has a void * field called
4481 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4482 pcre_callout_block structure has a field of the same name. The contents of
4483 the field passed in the pcre_extra structure are passed to the callout
4484 function in the corresponding field in the callout block. This makes it
4485 easier to use the same callout-containing regex from multiple threads. For
4486 testing, the pcretest program has a new data escape
4487
4488 \C*n pass the number n (may be negative) as callout_data
4489
4490 If the callout function in pcretest receives a non-zero value as
4491 callout_data, it returns that value.
4492
4493 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4494 there were some redundant $(CFLAGS) in commands that are now specified as
4495 $(LINK), which already includes $(CFLAGS).
4496
4497 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4498 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4499 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4500 one-byte characters throughout. Note that case-insensitive matching applies
4501 only to characters whose values are less than 256. PCRE doesn't support the
4502 notion of cases for higher-valued characters.
4503
4504 (i) A character class whose characters are all within 0-255 is handled as
4505 a bit map, and the map is inverted for negative classes. Previously, a
4506 character > 255 always failed to match such a class; however it should
4507 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4508
4509 (ii) A negated character class with a single character < 255 is coded as
4510 "not this character" (OP_NOT). This wasn't working properly when the test
4511 character was multibyte, either singly or repeated.
4512
4513 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4514 mode, for example: \x{100}{2,3}.
4515
4516 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4517 singly or repeated) now correctly test multibyte characters. However,
4518 PCRE doesn't recognize any characters with values greater than 255 as
4519 digits, spaces, or word characters. Such characters always match \D, \S,
4520 and \W, and never match \d, \s, or \w.
4521
4522 (v) Classes may now contain characters and character ranges with values
4523 greater than 255. For example: [ab\x{100}-\x{400}].
4524
4525 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4526 PCRE in UTF-8 mode.
4527
4528 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4529 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4530 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4531 value.)
4532
4533 53. The single man page has become too large. I have therefore split it up into
4534 a number of separate man pages. These also give rise to individual HTML pages;
4535 these are now put in a separate directory, and there is an index.html page that
4536 lists them all. Some hyperlinking between the pages has been installed.
4537
4538 54. Added convenience functions for handling named capturing parentheses.
4539
4540 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4541 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4542 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4543 are faulted.
4544
4545 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4546 calling configure. These values are used when compiling the dftables.c program
4547 which is run to generate the source of the default character tables. They
4548 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4549 you will need to set these values.
4550
4551 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4552
4553
4554 Version 3.9 02-Jan-02
4555 ---------------------
4556
4557 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4558
4559 2. If --disable-static was given, the building process failed when trying to
4560 build pcretest and pcregrep. (For some reason it was using libtool to compile
4561 them, which is not right, as they aren't part of the library.)
4562
4563
4564 Version 3.8 18-Dec-01
4565 ---------------------
4566
4567 1. The experimental UTF-8 code was completely screwed up. It was packing the
4568 bytes in the wrong order. How dumb can you get?
4569
4570
4571 Version 3.7 29-Oct-01
4572 ---------------------
4573
4574 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4575 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4576 this didn't happen under Solaris 8, where I normally test things.
4577
4578 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4579 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4580 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4581 this fix an hour or so after the initial 3.7 release.)
4582
4583
4584 Version 3.6 23-Oct-01
4585 ---------------------
4586
4587 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4588 offsets passed as NULL with zero offset count.
4589
4590 2. The config.guess and config.sub files had not been updated when I moved to
4591 the latest autoconf.
4592
4593
4594 Version 3.5 15-Aug-01
4595 ---------------------
4596
4597 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4598 had been forgotten.
4599
4600 2. By using declared but undefined structures, we can avoid using "void"
4601 definitions in pcre.h while keeping the internal definitions of the structures
4602 private.
4603
4604 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4605 user point of view, this means that both static and shared libraries are built
4606 by default, but this can be individually controlled. More of the work of
4607 handling this static/shared cases is now inside libtool instead of PCRE's make
4608 file.
4609
4610 4. The pcretest utility is now installed along with pcregrep because it is
4611 useful for users (to test regexs) and by doing this, it automatically gets
4612 relinked by libtool. The documentation has been turned into a man page, so
4613 there are now .1, .txt, and .html versions in /doc.
4614
4615 5. Upgrades to pcregrep:
4616 (i) Added long-form option names like gnu grep.
4617 (ii) Added --help to list all options with an explanatory phrase.
4618 (iii) Added -r, --recursive to recurse into sub-directories.
4619 (iv) Added -f, --file to read patterns from a file.
4620
4621 6. pcre_exec() was referring to its "code" argument before testing that
4622 argument for NULL (and giving an error if it was NULL).
4623
4624 7. Upgraded Makefile.in to allow for compiling in a different directory from
4625 the source directory.
4626
4627 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4628 options bits, the pointer it was passed was to an int instead of to an unsigned
4629 long int. This mattered only on 64-bit systems.
4630
4631 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4632 generated) instead of pcre.in, which it its source. Also made the same change
4633 in several of the .c files.
4634
4635 10. A new release of gcc defines printf() as a macro, which broke pcretest
4636 because it had an ifdef in the middle of a string argument for printf(). Fixed
4637 by using separate calls to printf().
4638
4639 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4640 script, to force use of CR or LF instead of \n in the source. On non-Unix
4641 systems, the value can be set in config.h.
4642
4643 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4644 absolute limit. Changed the text of the error message to make this clear, and
4645 likewise updated the man page.
4646
4647 13. The limit of 99 on the number of capturing subpatterns has been removed.
4648 The new limit is 65535, which I hope will not be a "real" limit.
4649
4650
4651 Version 3.4 22-Aug-00
4652 ---------------------
4653
4654 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4655
4656 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4657
4658
4659 Version 3.3 01-Aug-00
4660 ---------------------
4661
4662 1. If an octal character was given, but the value was greater than \377, it
4663 was not getting masked to the least significant bits, as documented. This could
4664 lead to crashes in some systems.
4665
4666 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4667 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4668
4669 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4670 These just pass their arguments on to (pcre_free)(), but they are provided
4671 because some uses of PCRE bind it to non-C systems that can call its functions,
4672 but cannot call free() or pcre_free() directly.
4673
4674 4. Add "make test" as a synonym for "make check". Corrected some comments in
4675 the Makefile.
4676
4677 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4678 Makefile.
4679
4680 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4681 command called pgrep for grepping around the active processes.
4682
4683 7. Added the beginnings of support for UTF-8 character strings.
4684
4685 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4686 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4687 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4688 out for the ar command.)
4689
4690
4691 Version 3.2 12-May-00
4692 ---------------------
4693
4694 This is purely a bug fixing release.
4695
4696 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4697 of ZA. This was just one example of several cases that could provoke this bug,
4698 which was introduced by change 9 of version 2.00. The code for breaking
4699 infinite loops after an iteration that matches an empty string was't working
4700 correctly.
4701
4702 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4703 when matched against abbab (for example). After matching an empty string, it
4704 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4705 caused it to match further down the string than it should.
4706
4707 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4708 was there because it doesn't seem to be needed, and it causes trouble on some
4709 systems, as it is not a Standard C header. It has been removed.
4710
4711 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4712 were reported on the Macintosh. The changes were from
4713
4714 while ((c = *(++ptr)) != 0 && c != '\n');
4715 to
4716 while ((c = *(++ptr)) != 0 && c != '\n') ;
4717
4718 Totally extraordinary, but if that's what it takes...
4719
4720 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4721 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4722 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4723 assumes the way PCRE uses memmove() (always moving upwards).
4724
4725 6. PCRE is being used in one environment where strchr() is not available. There
4726 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4727 faster code anyway.
4728
4729
4730 Version 3.1 09-Feb-00
4731 ---------------------
4732
4733 The only change in this release is the fixing of some bugs in Makefile.in for
4734 the "install" target:
4735
4736 (1) It was failing to install pcreposix.h.
4737
4738 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4739
4740
4741 Version 3.0 01-Feb-00
4742 ---------------------
4743
4744 1. Add support for the /+ modifier to perltest (to output $` like it does in
4745 pcretest).
4746
4747 2. Add support for the /g modifier to perltest.
4748
4749 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4750 matches null strings.
4751
4752 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4753 pattern. Perl treats empty patterns specially - it reuses the most recent
4754 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4755 effect.
4756
4757 5. The POSIX interface was broken in that it was just handing over the POSIX
4758 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4759 required a bigger vector, with some working space on the end. This means that
4760 the POSIX wrapper now has to get and free some memory, and copy the results.
4761
4762 6. Added some simple autoconf support, placing the test data and the
4763 documentation in separate directories, re-organizing some of the
4764 information files, and making it build pcre-config (a GNU standard). Also added
4765 libtool support for building PCRE as a shared library, which is now the
4766 default.
4767
4768 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4769 09 are not valid octal constants. Single digits will be used for minor values
4770 less than 10.
4771
4772 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4773 existing programs that set these in the POSIX interface can use PCRE without
4774 modification.
4775
4776 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4777 return all that pcre_info() returns, plus additional data. The pcre_info()
4778 function is retained for compatibility, but is considered to be obsolete.
4779
4780 10. Added experimental recursion feature (?R) to handle one common case that
4781 Perl 5.6 will be able to do with (?p{...}).
4782
4783 11. Added support for POSIX character classes like [:alpha:], which Perl is
4784 adopting.
4785
4786
4787 Version 2.08 31-Aug-99
4788 ----------------------
4789
4790 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4791 trying to match at the startoffset position, but instead was moving forward to
4792 the next newline as if a previous match had failed.
4793
4794 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4795 and could get into a loop if a null string was matched other than at the start
4796 of the subject.
4797
4798 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4799 be distinguished at compile time, and for completeness also added PCRE_DATE.
4800
4801 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4802 in GnuWin32 environments.
4803
4804
4805 Version 2.07 29-Jul-99
4806 ----------------------
4807
4808 1. The documentation is now supplied in plain text form and HTML as well as in
4809 the form of man page sources.
4810
4811 2. C++ compilers don't like assigning (void *) values to other pointer types.
4812 In particular this affects malloc(). Although there is no problem in Standard
4813 C, I've put in casts to keep C++ compilers happy.
4814
4815 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4816 should be (const char *).
4817
4818 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4819 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4820 However, I haven't made this a standard facility. The documentation doesn't
4821 mention it, and the Makefile doesn't support it.
4822
4823 5. The Makefile now contains an "install" target, with editable destinations at
4824 the top of the file. The pcretest program is not installed.
4825
4826 6. pgrep -V now gives the PCRE version number and date.
4827
4828 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4829 causing the entire string to be ignored, instead of just the last character.
4830
4831 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4832 non-matching string, it can take a very, very long time, even for strings of
4833 quite modest length, because of the nested recursion. PCRE now does better in
4834 some of these cases. It does this by remembering the last required literal
4835 character in the pattern, and pre-searching the subject to ensure it is present
4836 before running the real match. In other words, it applies a heuristic to detect
4837 some types of certain failure quickly, and in the above example, if presented
4838 with a string that has no trailing " it gives "no match" very quickly.
4839
4840 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4841 other alternatives are tried instead.
4842
4843
4844 Version 2.06 09-Jun-99
4845 ----------------------
4846
4847 1. Change pcretest's output for amount of store used to show just the code
4848 space, because the remainder (the data block) varies in size between 32-bit and
4849 64-bit systems.
4850
4851 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4852 start matching at. This allows lookbehinds to work when searching for multiple
4853 occurrences in a string.
4854
4855 3. Added additional options to pcretest for testing multiple occurrences:
4856
4857 /+ outputs the rest of the string that follows a match
4858 /g loops for multiple occurrences, using the new startoffset argument
4859 /G loops for multiple occurrences by passing an incremented pointer
4860
4861 4. PCRE wasn't doing the "first character" optimization for patterns starting
4862 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4863 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4864 the letter 'x'. On long subject strings, this gives a significant speed-up.
4865
4866
4867 Version 2.05 21-Apr-99
4868 ----------------------
4869
4870 1. Changed the type of magic_number from int to long int so that it works
4871 properly on 16-bit systems.
4872
4873 2. Fixed a bug which caused patterns starting with .* not to work correctly
4874 when the subject string contained newline characters. PCRE was assuming
4875 anchoring for such patterns in all cases, which is not correct because .* will
4876 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4877 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4878 must be retried after every newline in the subject.
4879
4880
4881 Version 2.04 18-Feb-99
4882 ----------------------
4883
4884 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4885 computation of the store needed to hold the pattern was incorrect (too large).
4886 If such patterns were nested a few deep, this could multiply and become a real
4887 problem.
4888
4889 2. Added /M option to pcretest to show the memory requirement of a specific
4890 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4891
4892 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4893 compiled in such a way that the backtracking after subsequent failure was
4894 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4895 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4896
4897
4898 Version 2.03 02-Feb-99
4899 ----------------------
4900
4901 1. Fixed typo and small mistake in man page.
4902
4903 2. Added 4th condition (GPL supersedes if conflict) and created separate
4904 LICENCE file containing the conditions.
4905
4906 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4907 Perl, that is the internal \ allows the delimiter to be included in the
4908 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4909 the final delimiter, add \ to the end of the pattern (to test the error).
4910
4911 4. Added the convenience functions for extracting substrings after a successful
4912 match. Updated pcretest to make it able to test these functions.
4913
4914
4915 Version 2.02 14-Jan-99
4916 ----------------------
4917
4918 1. Initialized the working variables associated with each extraction so that
4919 their saving and restoring doesn't refer to uninitialized store.
4920
4921 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4922 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4923 fix the problem.
4924
4925 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4926 calls, and wasn't printing the correct value for compiling calls. Increased the
4927 default value of LOOPREPEAT, and the number of significant figures in the
4928 times.
4929
4930 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4931
4932 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4933 a building problem on Windows NT with a FAT file system.
4934
4935
4936 Version 2.01 21-Oct-98
4937 ----------------------
4938
4939 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4940 to character tables built by pcre_maketables() in the current locale. If NULL
4941 is passed, the default tables are used.
4942
4943
4944 Version 2.00 24-Sep-98
4945 ----------------------
4946
4947 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4948 it any more.
4949
4950 2. Allow quantification of (?>) groups, and make it work correctly.
4951
4952 3. The first character computation wasn't working for (?>) groups.
4953
4954 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4955 end of the subject) and add 5.005's \z, which really does match only at the
4956 very end of the subject.
4957
4958 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4959
4960 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4961 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4962 localized options. All options to pcre_study() were also removed.
4963
4964 7. Add other new features from 5.005:
4965
4966 $(?<= positive lookbehind
4967 $(?<! negative lookbehind
4968 (?imsx-imsx) added the unsetting capability
4969 such a setting is global if at outer level; local otherwise
4970 (?imsx-imsx:) non-capturing groups with option setting
4971 (?(cond)re|re) conditional pattern matching
4972
4973 A backreference to itself in a repeated group matches the previous
4974 captured string.
4975
4976 8. General tidying up of studying (both automatic and via "study")
4977 consequential on the addition of new assertions.
4978
4979 9. As in 5.005, unlimited repeated groups that could match an empty substring
4980 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4981 runtime if any iteration does actually match an empty substring.
4982
4983 10. Include the RunTest script in the distribution.
4984
4985 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4986 discrepancies, some of which were old and were also with respect to 5.004. They
4987 have now been fixed.
4988
4989
4990 Version 1.09 28-Apr-98
4991 ----------------------
4992
4993 1. A negated single character class followed by a quantifier with a minimum
4994 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4995 program crashes, or just wrong answers. This did not apply to negated classes
4996 containing more than one character, or to minima other than one.
4997
4998
4999 Version 1.08 27-Mar-98
5000 ----------------------
5001
5002 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5003
5004 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5005 latter must appear before anything that relies on it in the pattern.
5006
5007
5008 Version 1.07 16-Feb-98
5009 ----------------------
5010
5011 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5012 repeat of a potentially empty string).
5013
5014
5015 Version 1.06 23-Jan-98
5016 ----------------------
5017
5018 1. Added Markus Oberhumer's little patches for C++.
5019
5020 2. Literal strings longer than 255 characters were broken.
5021
5022
5023 Version 1.05 23-Dec-97
5024 ----------------------
5025
5026 1. Negated character classes containing more than one character were failing if
5027 PCRE_CASELESS was set at run time.
5028
5029
5030 Version 1.04 19-Dec-97
5031 ----------------------
5032
5033 1. Corrected the man page, where some "const" qualifiers had been omitted.
5034
5035 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5036 input syntax.
5037
5038 3. Fixed memory leak which occurred when a regex with back references was
5039 matched with an offsets vector that wasn't big enough. The temporary memory
5040 that is used in this case wasn't being freed if the match failed.
5041
5042 4. Tidied pcretest to ensure it frees memory that it gets.
5043
5044 5. Temporary memory was being obtained in the case where the passed offsets
5045 vector was exactly big enough.
5046
5047 6. Corrected definition of offsetof() from change 5 below.
5048
5049 7. I had screwed up change 6 below and broken the rules for the use of
5050 setjmp(). Now fixed.
5051
5052
5053 Version 1.03 18-Dec-97
5054 ----------------------
5055
5056 1. A erroneous regex with a missing opening parenthesis was correctly
5057 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5058 on some systems.
5059
5060 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5061 it was reported that one broken compiler failed on the former because "code" is
5062 also an independent variable.
5063
5064 3. The erroneous regex a[]b caused an array overrun reference.
5065
5066 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5067 fail on data ending with that character. (It was going on too far, and checking
5068 the next character, typically a binary zero.) This was specific to the
5069 optimized code for single-character negative classes.
5070
5071 5. Added a contributed patch from the TIN world which does the following:
5072
5073 + Add an undef for memmove, in case the the system defines a macro for it.
5074
5075 + Add a definition of offsetof(), in case there isn't one. (I don't know
5076 the reason behind this - offsetof() is part of the ANSI standard - but
5077 it does no harm).
5078
5079 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5080 most of the places where whitespace preceded '#'. I have given up and
5081 allowed the remaining 2 cases to be at the margin.
5082
5083 + Rename some variables in pcre to eliminate shadowing. This seems very
5084 pedantic, but does no harm, of course.
5085
5086 6. Moved the call to setjmp() into its own function, to get rid of warnings
5087 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5088
5089 7. Constructs such as \d{8,} were compiling into the equivalent of
5090 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5091 outcome, but in this particular case used more store than had been allocated,
5092 which caused the bug to be discovered because it threw up an internal error.
5093
5094 8. The debugging code in both pcre and pcretest for outputting the compiled
5095 form of a regex was going wrong in the case of back references followed by
5096 curly-bracketed repeats.
5097
5098
5099 Version 1.02 12-Dec-97
5100 ----------------------
5101
5102 1. Typos in pcre.3 and comments in the source fixed.
5103
5104 2. Applied a contributed patch to get rid of places where it used to remove
5105 'const' from variables, and fixed some signed/unsigned and uninitialized
5106 variable warnings.
5107
5108 3. Added the "runtest" target to Makefile.
5109
5110 4. Set default compiler flag to -O2 rather than just -O.
5111
5112
5113 Version 1.01 19-Nov-97
5114 ----------------------
5115
5116 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5117 like /([ab]*)*/, that is, for classes with more than one character in them.
5118
5119 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5120 as /((?>a*))*/ (a PCRE_EXTRA facility).
5121
5122
5123 Version 1.00 18-Nov-97
5124 ----------------------
5125
5126 1. Added compile-time macros to support systems such as SunOS4 which don't have
5127 memmove() or strerror() but have other things that can be used instead.
5128
5129 2. Arranged that "make clean" removes the executables.
5130
5131
5132 Version 0.99 27-Oct-97
5133 ----------------------
5134
5135 1. Fixed bug in code for optimizing classes with only one character. It was
5136 initializing a 32-byte map regardless, which could cause it to run off the end
5137 of the memory it had got.
5138
5139 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5140
5141
5142 Version 0.98 22-Oct-97
5143 ----------------------
5144
5145 1. Fixed bug in code for handling temporary memory usage when there are more
5146 back references than supplied space in the ovector. This could cause segfaults.
5147
5148
5149 Version 0.97 21-Oct-97
5150 ----------------------
5151
5152 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5153
5154 2. Optimized negated single characters not to use a bit map.
5155
5156 3. Brought error texts together as macro definitions; clarified some of them;
5157 fixed one that was wrong - it said "range out of order" when it meant "invalid
5158 escape sequence".
5159
5160 4. Changed some char * arguments to const char *.
5161
5162 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5163
5164 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5165 pcretest.
5166
5167
5168 Version 0.96 16-Oct-97
5169 ----------------------
5170
5171 1. Added a simple "pgrep" utility to the distribution.
5172
5173 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5174 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5175 where "ddd" means "one or more decimal digits".
5176
5177 3. Fixed serious bug. If a pattern had a back reference, but the call to
5178 pcre_exec() didn't supply a large enough ovector to record the related
5179 identifying subpattern, the match always failed. PCRE now remembers the number
5180 of the largest back reference, and gets some temporary memory in which to save
5181 the offsets during matching if necessary, in order to ensure that
5182 backreferences always work.
5183
5184 4. Increased the compatibility with Perl in a number of ways:
5185
5186 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5187 to request this handling. The option can be set at compile or exec time.
5188
5189 (b) $ matches before a terminating newline by default; an option
5190 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5191 mode). The option can be set at compile or exec time.
5192
5193 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5194 the same as Perl's. If the decimal number it represents is less than 10
5195 or there aren't that many previous left capturing parentheses, an octal
5196 escape is read. Inside a character class, it's always an octal escape,
5197 even if it is a single digit.
5198
5199 (d) An escaped but undefined alphabetic character is taken as a literal,
5200 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5201 escapes.
5202
5203 (e) {0} is now permitted. (The previous item is removed from the compiled
5204 pattern).
5205
5206 5. Changed all the names of code files so that the basic parts are no longer
5207 than 10 characters, and abolished the teeny "globals.c" file.
5208
5209 6. Changed the handling of character classes; they are now done with a 32-byte
5210 bit map always.
5211
5212 7. Added the -d and /D options to pcretest to make it possible to look at the
5213 internals of compilation without having to recompile pcre.
5214
5215
5216 Version 0.95 23-Sep-97
5217 ----------------------
5218
5219 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5220 \x20 at the start of a run of normal characters. These were being treated as
5221 real characters, instead of the source characters being re-checked.
5222
5223
5224 Version 0.94 18-Sep-97
5225 ----------------------
5226
5227 1. The functions are now thread-safe, with the caveat that the global variables
5228 containing pointers to malloc() and free() or alternative functions are the
5229 same for all threads.
5230
5231 2. Get pcre_study() to generate a bitmap of initial characters for non-
5232 anchored patterns when this is possible, and use it if passed to pcre_exec().
5233
5234
5235 Version 0.93 15-Sep-97
5236 ----------------------
5237
5238 1. /(b)|(:+)/ was computing an incorrect first character.
5239
5240 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5241 but not actually doing anything yet.
5242
5243 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5244 as Perl does (e.g. [-az] or [az-]).
5245
5246 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5247 all possible positions.
5248
5249 5. Split up into different modules to avoid including unneeded functions in a
5250 compiled binary. However, compile and exec are still in one module. The "study"
5251 function is split off.
5252
5253 6. The character tables are now in a separate module whose source is generated
5254 by an auxiliary program - but can then be edited by hand if required. There are
5255 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5256 toupper() in the code.
5257
5258 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5259 make them global. Abolish the function for setting them, as the caller can now
5260 set them directly.
5261
5262
5263 Version 0.92 11-Sep-97
5264 ----------------------
5265
5266 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5267 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5268
5269 2. Caseless matching was not working in character classes if the characters in
5270 the pattern were in upper case.
5271
5272 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5273
5274 4. Make PCRE_ANCHORED public and accept as a compile option.
5275
5276 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5277 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5278 pass them.
5279
5280 6. Give an error if bad option bits passed at compile or run time.
5281
5282 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5283 pcretest to cause it to pass that flag.
5284
5285 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5286 options, and the first character, if set.
5287
5288 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5289
5290
5291 Version 0.91 10-Sep-97
5292 ----------------------
5293
5294 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5295 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5296
5297 2. PCRE was looping on encountering an indefinitely repeated back reference to
5298 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5299 Perl does - treats the match as successful.
5300
5301 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5