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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1352 - (show annotations)
Mon Jul 29 15:49:21 2013 UTC (6 years, 4 months ago) by ph10
File size: 250474 byte(s)
Error occurred while calculating annotation data.
Fix incorrect "first data item" recorded for an assertion condition.
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 firse 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
62 Version 8.33 28-May-2013
63 --------------------------
64
65 1. Added 'U' to some constants that are compared to unsigned integers, to
66 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
67 variables that are added to signed variables, to ensure the result is
68 signed and can be negated.
69
70 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
71 CMake config files.
72
73 3. Revise the creation of config.h.generic so that all boolean macros are
74 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
75 overriding via -D on the command line possible.
76
77 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
78 to unsigned int is reported to make a quite noticeable speed difference in
79 a specific Windows environment. Testing on Linux did also appear to show
80 some benefit (and it is clearly not harmful). Also fixed the definition of
81 Xop which should be unsigned.
82
83 5. Related to (4), changing the definition of the intermediate variable cc
84 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
85 improvements.
86
87 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
88 unnecessary spaces.
89
90 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
91 and later.
92
93 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
94
95 9. Optimizing fast_forward_start_bits in JIT.
96
97 10. Adding support for callouts in JIT, and fixing some issues revealed
98 during this work. Namely:
99
100 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
101
102 (b) Minimum length was not checked before the matching is started.
103
104 11. The value of capture_last that is passed to callouts was incorrect in some
105 cases when there was a capture on one path that was subsequently abandoned
106 after a backtrack. Also, the capture_last value is now reset after a
107 recursion, since all captures are also reset in this case.
108
109 12. The interpreter no longer returns the "too many substrings" error in the
110 case when an overflowing capture is in a branch that is subsequently
111 abandoned after a backtrack.
112
113 13. In the pathological case when an offset vector of size 2 is used, pcretest
114 now prints out the matched string after a yield of 0 or 1.
115
116 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
117 Only supported by the JIT compiler at the moment.
118
119 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
120
121 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
122 offset of the starting point of the matching process, provided the offsets
123 vector is large enough.
124
125 17. The \A escape now records a lookbehind value of 1, though its execution
126 does not actually inspect the previous character. This is to ensure that,
127 in partial multi-segment matching, at least one character from the old
128 segment is retained when a new segment is processed. Otherwise, if there
129 are no lookbehinds in the pattern, \A might match incorrectly at the start
130 of a new segment.
131
132 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
133
134 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
135 gives some modest performance improvement in 8-bit mode.
136
137 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
138 be expressed in certain programming languages using Universal Character
139 Names.
140
141 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
142 which points out that "non characters" are not "characters that may not
143 appear in Unicode strings" but rather "characters that are reserved for
144 internal use and have only local meaning".
145
146 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
147 there was a conditional group that depended on an assertion, if the
148 assertion was false, the callout that immediately followed the alternation
149 in the condition was skipped when pcre_exec() was used for matching.
150
151 23. Allow an explicit callout to be inserted before an assertion that is the
152 condition for a conditional group, for compatibility with automatic
153 callouts, which always insert a callout at this point.
154
155 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
156 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
157
158 25. (*PRUNE) is now supported by the JIT compiler.
159
160 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
161
162 27. Fix the case where there are two or more SKIPs with arguments that may be
163 ignored.
164
165 28. (*SKIP) is now supported by the JIT compiler.
166
167 29. (*THEN) is now supported by the JIT compiler.
168
169 30. Update RunTest with additional test selector options.
170
171 31. The way PCRE handles backtracking verbs has been changed in two ways.
172
173 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
174 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
175 backtracking. In some cases this makes it more Perl-compatible, but Perl's
176 rather obscure rules do not always do the same thing.
177
178 (2) Previously, backtracking verbs were confined within assertions. This is
179 no longer the case for positive assertions, except for (*ACCEPT). Again,
180 this sometimes improves Perl compatibility, and sometimes does not.
181
182 32. A number of tests that were in test 2 because Perl did things differently
183 have been moved to test 1, because either Perl or PCRE has changed, and
184 these tests are now compatible.
185
186 32. Backtracking control verbs are now handled in the same way in JIT and
187 interpreter.
188
189 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
190 contained a forward subroutine reference caused a compile error.
191
192 34. Auto-detect and optimize limited repetitions in JIT.
193
194 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
195 blocking (*UTF) etc.
196
197 36. In the interpreter, maximizing pattern repetitions for characters and
198 character types now use tail recursion, which reduces stack usage.
199
200 37. The value of the max lookbehind was not correctly preserved if a compiled
201 and saved regex was reloaded on a host of different endianness.
202
203 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
204 of the compiled pattern block, expand the flags field from 16 to 32 bits
205 because it was almost full.
206
207 39. Try madvise first before posix_madvise.
208
209 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
210 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
211 original change locked out all matches of empty strings. This has been
212 changed so that one match of an empty string per line is recognized.
213 Subsequent searches on the same line (for colouring or for --only-matching,
214 for example) do not recognize empty strings.
215
216 41. Applied a user patch to fix a number of spelling mistakes in comments.
217
218 42. Data lines longer than 65536 caused pcretest to crash.
219
220 43. Clarified the data type for length and startoffset arguments for pcre_exec
221 and pcre_dfa_exec in the function-specific man pages, where they were
222 explicitly stated to be in bytes, never having been updated. I also added
223 some clarification to the pcreapi man page.
224
225 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
226 a segmentation fault.
227
228
229 Version 8.32 30-November-2012
230 -----------------------------
231
232 1. Improved JIT compiler optimizations for first character search and single
233 character iterators.
234
235 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
236 Patch by Daniel Richard G.
237
238 3. Single character iterator optimizations in the JIT compiler.
239
240 4. Improved JIT compiler optimizations for character ranges.
241
242 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
243 Reported by Giuseppe D'Angelo.
244
245 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
246 of a line, was being set incorrectly in cases where .* appeared inside
247 atomic brackets at the start of a pattern, or where there was a subsequent
248 *PRUNE or *SKIP.
249
250 7. Improved instruction cache flush for POWER/PowerPC.
251 Patch by Daniel Richard G.
252
253 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
254 grep:
255
256 (a) There is now no limit to the number of patterns to be matched.
257
258 (b) An error is given if a pattern is too long.
259
260 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
261 are now supported.
262
263 (d) --exclude-from and --include-from (multiple use) have been added.
264
265 (e) Exclusions and inclusions now apply to all files and directories, not
266 just to those obtained from scanning a directory recursively.
267
268 (f) Multiple uses of -f and --file-list are now supported.
269
270 (g) In a Windows environment, the default for -d has been changed from
271 "read" (the GNU grep default) to "skip", because otherwise the presence
272 of a directory in the file list provokes an error.
273
274 (h) The documentation has been revised and clarified in places.
275
276 9. Improve the matching speed of capturing brackets.
277
278 10. Changed the meaning of \X so that it now matches a Unicode extended
279 grapheme cluster.
280
281 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
282 out POSIX threads when JIT support is configured.
283
284 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
285
286 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
287 structure could go wrong in environments where size_t is not the same size
288 as int.
289
290 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
291 set.
292
293 15. The EBCDIC support had decayed; later updates to the code had included
294 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
295 general tidy up of EBCDIC-related issues, and the documentation was also
296 not quite right. There is now a test that can be run on ASCII systems to
297 check some of the EBCDIC-related things (but is it not a full test).
298
299 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
300 in a small tidy to the code.
301
302 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
303
304 18. If the --only-matching (-o) option in pcregrep is specified multiple
305 times, each one causes appropriate output. For example, -o1 -o2 outputs the
306 substrings matched by the 1st and 2nd capturing parentheses. A separating
307 string can be specified by --om-separator (default empty).
308
309 19. Improving the first n character searches.
310
311 20. Turn case lists for horizontal and vertical white space into macros so that
312 they are defined only once.
313
314 21. This set of changes together give more compatible Unicode case-folding
315 behaviour for characters that have more than one other case when UCP
316 support is available.
317
318 (a) The Unicode property table now has offsets into a new table of sets of
319 three or more characters that are case-equivalent. The MultiStage2.py
320 script that generates these tables (the pcre_ucd.c file) now scans
321 CaseFolding.txt instead of UnicodeData.txt for character case
322 information.
323
324 (b) The code for adding characters or ranges of characters to a character
325 class has been abstracted into a generalized function that also handles
326 case-independence. In UTF-mode with UCP support, this uses the new data
327 to handle characters with more than one other case.
328
329 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
330 whose other case is greater than 256 are now correctly matched
331 caselessly. Previously, the high codepoint matched the low one, but not
332 vice versa.
333
334 (d) The processing of \h, \H, \v, and \ in character classes now makes use
335 of the new class addition function, using character lists defined as
336 macros alongside the case definitions of 20 above.
337
338 (e) Caseless back references now work with characters that have more than
339 one other case.
340
341 (f) General caseless matching of characters with more than one other case
342 is supported.
343
344 22. Unicode character properties were updated from Unicode 6.2.0
345
346 23. Improved CMake support under Windows. Patch by Daniel Richard G.
347
348 24. Add support for 32-bit character strings, and UTF-32
349
350 25. Major JIT compiler update (code refactoring and bugfixing).
351 Experimental Sparc 32 support is added.
352
353 26. Applied a modified version of Daniel Richard G's patch to create
354 pcre.h.generic and config.h.generic by "make" instead of in the
355 PrepareRelease script.
356
357 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
358 pcre_compile.c when checking for a zero character.
359
360 28. Introducing a native interface for JIT. Through this interface, the compiled
361 machine code can be directly executed. The purpose of this interface is to
362 provide fast pattern matching, so several sanity checks are not performed.
363 However, feature tests are still performed. The new interface provides
364 1.4x speedup compared to the old one.
365
366 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
367 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
368 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
369
370 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
371 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
372 the "old" RFC 2279). Instead, it ended up passing a negative length to
373 pcre_exec().
374
375 31. Add support for GCC's visibility feature to hide internal functions.
376
377 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
378 "unknown -C option" after outputting 0 or 1.
379
380 33. There is now support for generating a code coverage report for the test
381 suite in environments where gcc is the compiler and lcov is installed. This
382 is mainly for the benefit of the developers.
383
384 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
385 unaddressable using valgrind annotations, allowing valgrind to detect
386 invalid memory accesses. This is mainly for the benefit of the developers.
387
388 25. (*UTF) can now be used to start a pattern in any of the three libraries.
389
390 26. Give configure error if --enable-cpp but no C++ compiler found.
391
392
393 Version 8.31 06-July-2012
394 -------------------------
395
396 1. Fixing a wrong JIT test case and some compiler warnings.
397
398 2. Removed a bashism from the RunTest script.
399
400 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
401 to unsigned type, result still unsigned" that was given by an MS compiler
402 on encountering the code "-sizeof(xxx)".
403
404 4. Partial matching support is added to the JIT compiler.
405
406 5. Fixed several bugs concerned with partial matching of items that consist
407 of more than one character:
408
409 (a) /^(..)\1/ did not partially match "aba" because checking references was
410 done on an "all or nothing" basis. This also applied to repeated
411 references.
412
413 (b) \R did not give a hard partial match if \r was found at the end of the
414 subject.
415
416 (c) \X did not give a hard partial match after matching one or more
417 characters at the end of the subject.
418
419 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
420 a partial match for the string "\r".
421
422 (e) When newline was set to CRLF, the metacharacter "." did not recognize
423 a partial match for a CR character at the end of the subject string.
424
425 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
426 running pcretest, the text "(JIT)" added to the output whenever JIT is
427 actually used to run the match.
428
429 7. Individual JIT compile options can be set in pcretest by following -s+[+]
430 or /S+[+] with a digit between 1 and 7.
431
432 8. OP_NOT now supports any UTF character not just single-byte ones.
433
434 9. (*MARK) control verb is now supported by the JIT compiler.
435
436 10. The command "./RunTest list" lists the available tests without actually
437 running any of them. (Because I keep forgetting what they all are.)
438
439 11. Add PCRE_INFO_MAXLOOKBEHIND.
440
441 12. Applied a (slightly modified) user-supplied patch that improves performance
442 when the heap is used for recursion (compiled with --disable-stack-for-
443 recursion). Instead of malloc and free for each heap frame each time a
444 logical recursion happens, frames are retained on a chain and re-used where
445 possible. This sometimes gives as much as 30% improvement.
446
447 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
448 call.
449
450 14. As documented, (*COMMIT) is now confined to within a positive assertion.
451
452 15. It is now possible to link pcretest with libedit as an alternative to
453 libreadline.
454
455 16. (*COMMIT) control verb is now supported by the JIT compiler.
456
457 17. The Unicode data tables have been updated to Unicode 6.1.0.
458
459 18. Added --file-list option to pcregrep.
460
461 19. Added binary file support to pcregrep, including the -a, --binary-files,
462 -I, and --text options.
463
464 20. The madvise function is renamed for posix_madvise for QNX compatibility
465 reasons. Fixed by Giuseppe D'Angelo.
466
467 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
468
469 22. Changed the option for creating symbolic links for 16-bit man pages from
470 -s to -sf so that re-installing does not cause issues.
471
472 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
473
474 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
475 matches in certain environments (the workspace was not being correctly
476 retained). Also added to pcre_dfa_exec() a simple plausibility check on
477 some of the workspace data at the beginning of a restart.
478
479 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
480 was not doing so when it should - probably a typo introduced by SVN 528
481 (change 8.10/14).
482
483 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
484 \w+ when the character tables indicated that \x{c4} was a word character.
485 There were several related cases, all because the tests for doing a table
486 lookup were testing for characters less than 127 instead of 255.
487
488 27. If a pattern contains capturing parentheses that are not used in a match,
489 their slots in the ovector are set to -1. For those that are higher than
490 any matched groups, this happens at the end of processing. In the case when
491 there were back references that the ovector was too small to contain
492 (causing temporary malloc'd memory to be used during matching), and the
493 highest capturing number was not used, memory off the end of the ovector
494 was incorrectly being set to -1. (It was using the size of the temporary
495 memory instead of the true size.)
496
497 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
498 ovector size, it uses memory at the end of the block that it has got.
499
500 29. Check for an overlong MARK name and give an error at compile time. The
501 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
502
503 30. JIT compiler update.
504
505 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
506 Rill for the patch.
507
508 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
509
510 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
511
512 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
513 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
514 was enabled.
515
516 35. Improve JIT code generation for greedy plus quantifier.
517
518 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
519 1 to "aa" instead of to an empty string. The bug affected repeated groups
520 that could potentially match an empty string.
521
522 37. Optimizing single character iterators in JIT.
523
524 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
525 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
526 codepoints that are too big for the mode are faulted, and in a UTF mode,
527 disallowed codepoints are also faulted.
528
529 39. If PCRE was compiled with UTF support, in three places in the DFA
530 matcher there was code that should only have been obeyed in UTF mode, but
531 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
532 processing when bytes with values greater than 127 were present. In 16-bit
533 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
534 both cases the values are those that cannot be the first data item in a UTF
535 character. The three items that might have provoked this were recursions,
536 possessively repeated groups, and atomic groups.
537
538 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
539 and pcregrep, because some OS require shared objects to be explicitly
540 passed to ld, causing the link step to fail if they are not.
541
542 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
543 mode, patterns that started with \h* or \R* might be incorrectly matched.
544
545
546 Version 8.30 04-February-2012
547 -----------------------------
548
549 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
550 name is defined in ctype.h.
551
552 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
553 only in quite long subpatterns.
554
555 3. Removed the function pcre_info(), which has been obsolete and deprecated
556 since it was replaced by pcre_fullinfo() in February 2000.
557
558 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
559 match a (*MARK), and the match failed at the start of the subject, a
560 reference to memory before the start of the subject could occur. This bug
561 was introduced by fix 17 of release 8.21.
562
563 5. A reference to an unset group with zero minimum repetition was giving
564 totally wrong answers (in non-JavaScript-compatibility mode). For example,
565 /(another)?(\1?)test/ matched against "hello world test". This bug was
566 introduced in release 8.13.
567
568 6. Add support for 16-bit character strings (a large amount of work involving
569 many changes and refactorings).
570
571 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
572 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
573 from a file.
574
575 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
576 rounding is not applied in this particular case).
577
578 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
579 if they appear, or are escaped, in patterns.
580
581 10. Get rid of a number of -Wunused-but-set-variable warnings.
582
583 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
584 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
585 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
586 also returns the mark "x". This bug applied to capturing parentheses,
587 non-capturing parentheses, and atomic parentheses. It also applied to some
588 assertions.
589
590 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
591 information out of configure.ac instead of relying on pcre.h.generic, which
592 is not stored in the repository.
593
594 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
595 -lreadline.
596
597 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
598
599 15. Applied Graycode's patch to put the top-level frame on the stack rather
600 than the heap when not using the stack for recursion. This gives a
601 performance improvement in many cases when recursion is not deep.
602
603 16. Experimental code added to "pcretest -C" to output the stack frame size.
604
605
606 Version 8.21 12-Dec-2011
607 ------------------------
608
609 1. Updating the JIT compiler.
610
611 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
612 are added as well.
613
614 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
615 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
616 calling _pcre_jit_exec. Some extra comments are added.
617
618 4. (*MARK) settings inside atomic groups that do not contain any capturing
619 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
620 was introduced by change 18 for 8.20.
621
622 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
623 ECMA-262 standard.
624
625 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
626 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
627 This bug was probably introduced by change 9 of 8.13.
628
629 7. While fixing 6 above, I noticed that a number of other items were being
630 incorrectly rejected as "not fixed length". This arose partly because newer
631 opcodes had not been added to the fixed-length checking code. I have (a)
632 corrected the bug and added tests for these items, and (b) arranged for an
633 error to occur if an unknown opcode is encountered while checking for fixed
634 length instead of just assuming "not fixed length". The items that were
635 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
636 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
637 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
638
639 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
640 being incorrectly compiled and would have given unpredicatble results.
641
642 9. A possessively repeated subpattern with minimum repeat count greater than
643 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
644 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
645 the first (A) could occur when it should not.
646
647 10. Add a cast and remove a redundant test from the code.
648
649 11. JIT should use pcre_malloc/pcre_free for allocation.
650
651 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
652 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
653 is anything other than /usr, -L is still shown).
654
655 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
656
657 14. Perl does not support \N without a following name in a [] class; PCRE now
658 also gives an error.
659
660 15. If a forward reference was repeated with an upper limit of around 2000,
661 it caused the error "internal error: overran compiling workspace". The
662 maximum number of forward references (including repeats) was limited by the
663 internal workspace, and dependent on the LINK_SIZE. The code has been
664 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
665 the default depends on LINK_SIZE. There is a new upper limit (for safety)
666 of around 200,000 forward references. While doing this, I also speeded up
667 the filling in of repeated forward references.
668
669 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
670 incorrectly expecting the subject to contain another "a" after the start.
671
672 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
673 in the match, the SKIP should be ignored. This was not happening; instead
674 the SKIP was being treated as NOMATCH. For patterns such as
675 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
676 tested.
677
678 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
679 now much more compatible with Perl, in particular in cases where the result
680 is a non-match for a non-anchored pattern. For example, if
681 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
682 "m", where previously it did not return a name. A side effect of this
683 change is that for partial matches, the last encountered mark name is
684 returned, as for non matches. A number of tests that were previously not
685 Perl-compatible have been moved into the Perl-compatible test files. The
686 refactoring has had the pleasing side effect of removing one argument from
687 the match() function, thus reducing its stack requirements.
688
689 19. If the /S+ option was used in pcretest to study a pattern using JIT,
690 subsequent uses of /S (without +) incorrectly behaved like /S+.
691
692 21. Retrieve executable code size support for the JIT compiler and fixing
693 some warnings.
694
695 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
696 not work when the shorter character appeared right at the end of the
697 subject string.
698
699 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
700 systems.
701
702 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
703 output it when the /M option is used in pcretest.
704
705 25. The CheckMan script was not being included in the distribution. Also, added
706 an explicit "perl" to run Perl scripts from the PrepareRelease script
707 because this is reportedly needed in Windows.
708
709 26. If study data was being save in a file and studying had not found a set of
710 "starts with" bytes for the pattern, the data written to the file (though
711 never used) was taken from uninitialized memory and so caused valgrind to
712 complain.
713
714 27. Updated RunTest.bat as provided by Sheri Pierce.
715
716 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
717
718 29. Computation of memory usage for the table of capturing group names was
719 giving an unnecessarily large value.
720
721
722 Version 8.20 21-Oct-2011
723 ------------------------
724
725 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
726 a POSIX class. After further experiments with Perl, which convinced me that
727 Perl has bugs and confusions, a closing square bracket is no longer allowed
728 in a POSIX name. This bug also affected patterns with classes that started
729 with full stops.
730
731 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
732 captured substring, but while checking the failing first alternative,
733 substring 1 is temporarily captured. If the output vector supplied to
734 pcre_exec() was not big enough for this capture, the yield of the function
735 was still zero ("insufficient space for captured substrings"). This cannot
736 be totally fixed without adding another stack variable, which seems a lot
737 of expense for a edge case. However, I have improved the situation in cases
738 such as /(a)(b)x|abc/ matched against "abc", where the return code
739 indicates that fewer than the maximum number of slots in the ovector have
740 been set.
741
742 3. Related to (2) above: when there are more back references in a pattern than
743 slots in the output vector, pcre_exec() uses temporary memory during
744 matching, and copies in the captures as far as possible afterwards. It was
745 using the entire output vector, but this conflicts with the specification
746 that only 2/3 is used for passing back captured substrings. Now it uses
747 only the first 2/3, for compatibility. This is, of course, another edge
748 case.
749
750 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
751 main code base, and can be used by building with --enable-jit. When this is
752 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
753 runtime --no-jit option is given.
754
755 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
756 ovector, the return from the function was zero, implying that there were
757 other matches that did not fit. The correct "exactly full" value is now
758 returned.
759
760 6. If a subpattern that was called recursively or as a subroutine contained
761 (*PRUNE) or any other control that caused it to give a non-standard return,
762 invalid errors such as "Error -26 (nested recursion at the same subject
763 position)" or even infinite loops could occur.
764
765 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
766 computing the minimum length on reaching *ACCEPT, and so ended up with the
767 wrong value of 1 rather than 0. Further investigation indicates that
768 computing a minimum subject length in the presence of *ACCEPT is difficult
769 (think back references, subroutine calls), and so I have changed the code
770 so that no minimum is registered for a pattern that contains *ACCEPT.
771
772 8. If (*THEN) was present in the first (true) branch of a conditional group,
773 it was not handled as intended. [But see 16 below.]
774
775 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
776 Sheri Pierce.
777
778 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
779 the first byte in a match must be "a".
780
781 11. Change 17 for 8.13 increased the recursion depth for patterns like
782 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
783 pattern contains any instances of (*THEN). If it does not, the old
784 optimizations are restored. It would be nice to do this on a per-group
785 basis, but at the moment that is not feasible.
786
787 12. In some environments, the output of pcretest -C is CRLF terminated. This
788 broke RunTest's code that checks for the link size. A single white space
789 character after the value is now allowed for.
790
791 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
792 For "fr", it uses the Windows-specific input and output files.
793
794 14. If (*THEN) appeared in a group that was called recursively or as a
795 subroutine, it did not work as intended. [But see next item.]
796
797 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
798 pattern fragments (but not containing any | characters). If A and B are
799 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
800 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
801 D. In other words, Perl considers parentheses that do not contain any |
802 characters to be part of a surrounding alternative, whereas PCRE was
803 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
804 differently. PCRE now behaves in the same way as Perl, except in the case
805 of subroutine/recursion calls such as (?1) which have in any case always
806 been different (but PCRE had them first :-).
807
808 16. Related to 15 above: Perl does not treat the | in a conditional group as
809 creating alternatives. Such a group is treated in the same way as an
810 ordinary group without any | characters when processing (*THEN). PCRE has
811 been changed to match Perl's behaviour.
812
813 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
814 RunGrepTest script failed.
815
816 18. Change 22 for version 13 caused atomic groups to use more stack. This is
817 inevitable for groups that contain captures, but it can lead to a lot of
818 stack use in large patterns. The old behaviour has been restored for atomic
819 groups that do not contain any capturing parentheses.
820
821 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
822 suppress the check for a minimum subject length at run time. (If it was
823 given to pcre_exec() or pcre_dfa_exec() it did work.)
824
825 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
826 fail to work when decoding hex characters in data strings in EBCDIC
827 environments.
828
829 21. It appears that in at least one Mac OS environment, the isxdigit() function
830 is implemented as a macro that evaluates to its argument more than once,
831 contravening the C 90 Standard (I haven't checked a later standard). There
832 was an instance in pcretest which caused it to go wrong when processing
833 \x{...} escapes in subject strings. The has been rewritten to avoid using
834 things like p++ in the argument of isxdigit().
835
836
837 Version 8.13 16-Aug-2011
838 ------------------------
839
840 1. The Unicode data tables have been updated to Unicode 6.0.0.
841
842 2. Two minor typos in pcre_internal.h have been fixed.
843
844 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
845 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
846 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
847
848 4. There were a number of related bugs in the code for matching backrefences
849 caselessly in UTF-8 mode when codes for the characters concerned were
850 different numbers of bytes. For example, U+023A and U+2C65 are an upper
851 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
852 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
853 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
854 2-byte code at the end of the subject (it thought there wasn't enough data
855 left).
856
857 5. Comprehensive information about what went wrong is now returned by
858 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
859 as the output vector has at least 2 elements. The offset of the start of
860 the failing character and a reason code are placed in the vector.
861
862 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
863 now returned is for the first byte of the failing character, instead of the
864 last byte inspected. This is an incompatible change, but I hope it is small
865 enough not to be a problem. It makes the returned offset consistent with
866 pcre_exec() and pcre_dfa_exec().
867
868 7. pcretest now gives a text phrase as well as the error number when
869 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
870 failure, the offset and reason code are output.
871
872 8. When \R was used with a maximizing quantifier it failed to skip backwards
873 over a \r\n pair if the subsequent match failed. Instead, it just skipped
874 back over a single character (\n). This seems wrong (because it treated the
875 two characters as a single entity when going forwards), conflicts with the
876 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
877 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
878 has been changed.
879
880 9. Some internal refactoring has changed the processing so that the handling
881 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
882 time (the PCRE_DOTALL option was changed this way some time ago: version
883 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
884 which was always a bit of a fudge. It also means that there is one less
885 argument for the match() function, which reduces its stack requirements
886 slightly. This change also fixes an incompatibility with Perl: the pattern
887 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
888
889 10. More internal refactoring has drastically reduced the number of recursive
890 calls to match() for possessively repeated groups such as (abc)++ when
891 using pcre_exec().
892
893 11. While implementing 10, a number of bugs in the handling of groups were
894 discovered and fixed:
895
896 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
897 (a|)*(?1) gave a compile-time internal error.
898 ((a|)+)+ did not notice that the outer group could match an empty string.
899 (^a|^)+ was not marked as anchored.
900 (.*a|.*)+ was not marked as matching at start or after a newline.
901
902 12. Yet more internal refactoring has removed another argument from the match()
903 function. Special calls to this function are now indicated by setting a
904 value in a variable in the "match data" data block.
905
906 13. Be more explicit in pcre_study() instead of relying on "default" for
907 opcodes that mean there is no starting character; this means that when new
908 ones are added and accidentally left out of pcre_study(), testing should
909 pick them up.
910
911 14. The -s option of pcretest has been documented for ages as being an old
912 synonym of -m (show memory usage). I have changed it to mean "force study
913 for every regex", that is, assume /S for every regex. This is similar to -i
914 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
915 using it. It makes it easier to run collections of tests with and without
916 study enabled, and thereby test pcre_study() more easily. All the standard
917 tests are now run with and without -s (but some patterns can be marked as
918 "never study" - see 20 below).
919
920 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
921 restoration of the capturing data to the outer values was not happening
922 correctly.
923
924 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
925 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
926 pattern had matched an empty string, and so incorrectly returned a no
927 match.
928
929 17. There was optimizing code for the last branch of non-capturing parentheses,
930 and also for the obeyed branch of a conditional subexpression, which used
931 tail recursion to cut down on stack usage. Unfortunately, now that there is
932 the possibility of (*THEN) occurring in these branches, tail recursion is
933 no longer possible because the return has to be checked for (*THEN). These
934 two optimizations have therefore been removed. [But see 8.20/11 above.]
935
936 18. If a pattern containing \R was studied, it was assumed that \R always
937 matched two bytes, thus causing the minimum subject length to be
938 incorrectly computed because \R can also match just one byte.
939
940 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
941 was incorrectly computed.
942
943 20. If /S is present twice on a test pattern in pcretest input, it now
944 *disables* studying, thereby overriding the use of -s on the command line
945 (see 14 above). This is necessary for one or two tests to keep the output
946 identical in both cases.
947
948 21. When (*ACCEPT) was used in an assertion that matched an empty string and
949 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
950
951 22. When an atomic group that contained a capturing parenthesis was
952 successfully matched, but the branch in which it appeared failed, the
953 capturing was not being forgotten if a higher numbered group was later
954 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
955 group 1 to "a", when in fact it should be unset. This applied to multi-
956 branched capturing and non-capturing groups, repeated or not, and also to
957 positive assertions (capturing in negative assertions does not happen
958 in PCRE) and also to nested atomic groups.
959
960 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
961 subject after a captured substring, to make it easier to tell which of a
962 number of identical substrings has been captured.
963
964 24. The way atomic groups are processed by pcre_exec() has been changed so that
965 if they are repeated, backtracking one repetition now resets captured
966 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
967 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
968 "aaa". Previously, it would have been "a". As part of this code
969 refactoring, the way recursive calls are handled has also been changed.
970
971 25. If an assertion condition captured any substrings, they were not passed
972 back unless some other capturing happened later. For example, if
973 (?(?=(a))a) was matched against "a", no capturing was returned.
974
975 26. When studying a pattern that contained subroutine calls or assertions,
976 the code for finding the minimum length of a possible match was handling
977 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
978 group 1 called group 2 while simultaneously a separate group 2 called group
979 1). A stack overflow occurred in this case. I have fixed this by limiting
980 the recursion depth to 10.
981
982 27. Updated RunTest.bat in the distribution to the version supplied by Tom
983 Fortmann. This supports explicit test numbers on the command line, and has
984 argument validation and error reporting.
985
986 28. An instance of \X with an unlimited repeat could fail if at any point the
987 first character it looked at was a mark character.
988
989 29. Some minor code refactoring concerning Unicode properties and scripts
990 should reduce the stack requirement of match() slightly.
991
992 30. Added the '=' option to pcretest to check the setting of unused capturing
993 slots at the end of the pattern, which are documented as being -1, but are
994 not included in the return count.
995
996 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
997 compiled something random. Now it gives a compile-time error (as does
998 Perl).
999
1000 32. A *MARK encountered during the processing of a positive assertion is now
1001 recorded and passed back (compatible with Perl).
1002
1003 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1004 had alternative anchored branches, the search for a second match in a line
1005 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1006 matched the line "0102" twice. The same bug affected patterns that started
1007 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1008 twice.
1009
1010 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1011 does, and because of capturing effects, quantifying parenthesized
1012 assertions may at times be useful. Quantifiers are now allowed for
1013 parenthesized assertions.
1014
1015 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1016
1017 36. \g was being checked for fancy things in a character class, when it should
1018 just be a literal "g".
1019
1020 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1021 appearance of a nested POSIX class supersedes an apparent external class.
1022 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1023 unescaped square brackets may also appear as part of class names. For
1024 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1025 more like Perl. (But see 8.20/1 above.)
1026
1027 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1028 was because it thought it was \N{name}, which is not supported).
1029
1030 39. Add minix to OS list not supporting the -S option in pcretest.
1031
1032 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1033 cannot analyze patterns in sufficient detail to catch mutual recursions
1034 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1035 subgroup is called recursively as a subpattern for a second time at the
1036 same position in the subject string. In previous releases this might have
1037 been caught by the recursion limit, or it might have run out of stack.
1038
1039 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1040 happen only once. PCRE was, however incorrectly giving a compile time error
1041 "recursive call could loop indefinitely" because it cannot analyze the
1042 pattern in sufficient detail. The compile time test no longer happens when
1043 PCRE is compiling a conditional subpattern, but actual runaway loops are
1044 now caught at runtime (see 40 above).
1045
1046 42. It seems that Perl allows any characters other than a closing parenthesis
1047 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1048 has been changed to be the same.
1049
1050 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1051 as not to get warnings when autogen.sh is called. Also changed
1052 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1053
1054 44. To help people who use pcregrep to scan files containing exceedingly long
1055 lines, the following changes have been made:
1056
1057 (a) The default value of the buffer size parameter has been increased from
1058 8K to 20K. (The actual buffer used is three times this size.)
1059
1060 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1061 PCRE is built.
1062
1063 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1064 to be set at run time.
1065
1066 (d) Numerical values in pcregrep options can be followed by K or M, for
1067 example --buffer-size=50K.
1068
1069 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1070 given and the return code is set to 2.
1071
1072 45. Add a pointer to the latest mark to the callout data block.
1073
1074 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1075 partial match of an empty string instead of no match. This was specific to
1076 the use of ".".
1077
1078 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1079 complete match instead of a partial match. This bug was dependent on both
1080 the PCRE_UTF8 and PCRE_DOTALL options being set.
1081
1082 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1083 starting byte set, because \b was not being ignored.
1084
1085
1086 Version 8.12 15-Jan-2011
1087 ------------------------
1088
1089 1. Fixed some typos in the markup of the man pages, and wrote a script that
1090 checks for such things as part of the documentation building process.
1091
1092 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1093 --match-limit and --recursion-limit options (added for 8.11). In
1094 particular, this made one of the standard tests fail. (The integer value
1095 went into the wrong half of a long int.)
1096
1097 3. If the --colour option was given to pcregrep with -v (invert match), it
1098 did strange things, either producing crazy output, or crashing. It should,
1099 of course, ignore a request for colour when reporting lines that do not
1100 match.
1101
1102 4. Another pcregrep bug caused similar problems if --colour was specified with
1103 -M (multiline) and the pattern match finished with a line ending.
1104
1105 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1106 matched in multiline mode, the following line was shown as part of the
1107 match. This seems wrong, so I have changed it.
1108
1109 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1110 the check for further matches in the same line (so they could be coloured)
1111 to overrun the end of the current line. If another match was found, it was
1112 incorrectly shown (and then shown again when found in the next line).
1113
1114 7. If pcregrep was compiled under Windows, there was a reference to the
1115 function pcregrep_exit() before it was defined. I am assuming this was
1116 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1117 reported by a user. I've moved the definition above the reference.
1118
1119
1120 Version 8.11 10-Dec-2010
1121 ------------------------
1122
1123 1. (*THEN) was not working properly if there were untried alternatives prior
1124 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1125 backtracked to try for "b" instead of moving to the next alternative branch
1126 at the same level (in this case, to look for "c"). The Perl documentation
1127 is clear that when (*THEN) is backtracked onto, it goes to the "next
1128 alternative in the innermost enclosing group".
1129
1130 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1131 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1132 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1133 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1134 (*THEN).
1135
1136 3. If \s appeared in a character class, it removed the VT character from
1137 the class, even if it had been included by some previous item, for example
1138 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1139 of \s, but is part of the POSIX "space" class.)
1140
1141 4. A partial match never returns an empty string (because you can always
1142 match an empty string at the end of the subject); however the checking for
1143 an empty string was starting at the "start of match" point. This has been
1144 changed to the "earliest inspected character" point, because the returned
1145 data for a partial match starts at this character. This means that, for
1146 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1147 (previously it gave "no match").
1148
1149 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1150 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1151 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1152 has an implication that the given string is incomplete (because a partial
1153 match is preferred over a full match). For this reason, these items now
1154 give a partial match in this situation. [Aside: previously, the one case
1155 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1156 match rather than a full match, which was wrong by the old rules, but is
1157 now correct.]
1158
1159 6. There was a bug in the handling of #-introduced comments, recognized when
1160 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1161 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1162 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1163 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1164 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1165 places in pcre_compile().
1166
1167 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1168 comments when looking ahead for named forward references to subpatterns,
1169 the only newline sequence it recognized was NL. It now handles newlines
1170 according to the set newline convention.
1171
1172 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1173 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1174 cater for a lack of strerror(). These oversights have been fixed.
1175
1176 9. Added --match-limit and --recursion-limit to pcregrep.
1177
1178 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1179
1180 11. When the -o option was used, pcregrep was setting a return code of 1, even
1181 when matches were found, and --line-buffered was not being honoured.
1182
1183 12. Added an optional parentheses number to the -o and --only-matching options
1184 of pcregrep.
1185
1186 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1187 can match an empty string. The code to do it in pcretest and pcredemo
1188 needed fixing:
1189
1190 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1191 only one byte after an empty string match just before CRLF (this case
1192 just got forgotten; "any" and "anycrlf" were OK).
1193
1194 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1195 mode when an empty string match preceded an ASCII character followed by
1196 a non-ASCII character. (The code for advancing by one character rather
1197 than one byte was nonsense.)
1198
1199 (c) The pcredemo.c sample program did not have any code at all to handle
1200 the cases when CRLF is a valid newline sequence.
1201
1202 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1203 as a starting offset was within the subject string. There is now a new
1204 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1205 negative or greater than the length of the string. In order to test this,
1206 pcretest is extended to allow the setting of negative starting offsets.
1207
1208 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1209 starting offset points to the beginning of a UTF-8 character was
1210 unnecessarily clumsy. I tidied it up.
1211
1212 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1213 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1214
1215 17. Nobody had reported that the --include_dir option, which was added in
1216 release 7.7 should have been called --include-dir (hyphen, not underscore)
1217 for compatibility with GNU grep. I have changed it to --include-dir, but
1218 left --include_dir as an undocumented synonym, and the same for
1219 --exclude-dir, though that is not available in GNU grep, at least as of
1220 release 2.5.4.
1221
1222 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1223 characters from a string of bytes) have been redefined so as not to use
1224 loops, in order to improve performance in some environments. At the same
1225 time, I abstracted some of the common code into auxiliary macros to save
1226 repetition (this should not affect the compiled code).
1227
1228 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1229 compile-time error is now given if \c is not followed by an ASCII
1230 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1231 different, and any byte value is allowed.)
1232
1233 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1234 START_OPTIMIZE option, which is now allowed at compile time - but just
1235 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1236 to pcregrep and other applications that have no direct access to PCRE
1237 options. The new /Y option in pcretest sets this option when calling
1238 pcre_compile().
1239
1240 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1241 back references. Groups containing recursive back references were forced to
1242 be atomic by that change, but in the case of named groups, the amount of
1243 memory required was incorrectly computed, leading to "Failed: internal
1244 error: code overflow". This has been fixed.
1245
1246 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1247 pcretest.c, to avoid build problems in some Borland environments.
1248
1249
1250 Version 8.10 25-Jun-2010
1251 ------------------------
1252
1253 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1254 THEN.
1255
1256 2. (*ACCEPT) was not working when inside an atomic group.
1257
1258 3. Inside a character class, \B is treated as a literal by default, but
1259 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1260 causes the error). The code is unchanged, but I tidied the documentation.
1261
1262 4. Inside a character class, PCRE always treated \R and \X as literals,
1263 whereas Perl faults them if its -w option is set. I have changed PCRE so
1264 that it faults them when PCRE_EXTRA is set.
1265
1266 5. Added support for \N, which always matches any character other than
1267 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1268
1269 6. When compiling pcregrep with newer versions of gcc which may have
1270 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1271 declared with attribute warn_unused_result" were given. Just casting the
1272 result to (void) does not stop the warnings; a more elaborate fudge is
1273 needed. I've used a macro to implement this.
1274
1275 7. Minor change to pcretest.c to avoid a compiler warning.
1276
1277 8. Added four artifical Unicode properties to help with an option to make
1278 \s etc use properties (see next item). The new properties are: Xan
1279 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1280
1281 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1282 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1283 this option. Modified pcretest to add /W to test this facility. Added
1284 REG_UCP to make it available via the POSIX interface.
1285
1286 10. Added --line-buffered to pcregrep.
1287
1288 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1289 studied, and the match started with a letter with a code point greater than
1290 127 whose first byte was different to the first byte of the other case of
1291 the letter, the other case of this starting letter was not recognized
1292 (#976).
1293
1294 12. If a pattern that was studied started with a repeated Unicode property
1295 test, for example, \p{Nd}+, there was the theoretical possibility of
1296 setting up an incorrect bitmap of starting bytes, but fortunately it could
1297 not have actually happened in practice until change 8 above was made (it
1298 added property types that matched character-matching opcodes).
1299
1300 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1301 possible starting bytes for non-anchored patterns.
1302
1303 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1304 \R, and also a number of cases that involve Unicode properties, both
1305 explicit and implicit when PCRE_UCP is set.
1306
1307 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1308 input, it could crash or give wrong results if characters with values
1309 greater than 0xc0 were present in the subject string. (Detail: it assumed
1310 UTF-8 input when processing these items.)
1311
1312 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1313 size_t is 64-bit (#991).
1314
1315 17. Added a check for running out of memory when PCRE is compiled with
1316 --disable-stack-for-recursion (#990).
1317
1318 18. If the last data line in a file for pcretest does not have a newline on
1319 the end, a newline was missing in the output.
1320
1321 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1322 less than 128) in its various bitmaps. However, there is a facility for
1323 generating tables according to the current locale when PCRE is compiled. It
1324 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1325 characters, are recognized by isspace() and therefore were getting set in
1326 these tables, and indeed these tables seem to approximate to ISO 8859. This
1327 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1328 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1329 which of course cannot start UTF-8 characters. I have changed the code so
1330 that only real ASCII characters (less than 128) and the correct starting
1331 bytes for UTF-8 encodings are set for characters greater than 127 when in
1332 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1333 altogether.)
1334
1335 20. Added the /T option to pcretest so as to be able to run tests with non-
1336 standard character tables, thus making it possible to include the tests
1337 used for 19 above in the standard set of tests.
1338
1339 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1340 reference to a subpattern the other side of a comment that contains an
1341 opening parenthesis caused either an internal compiling error, or a
1342 reference to the wrong subpattern.
1343
1344
1345 Version 8.02 19-Mar-2010
1346 ------------------------
1347
1348 1. The Unicode data tables have been updated to Unicode 5.2.0.
1349
1350 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1351 configured.
1352
1353 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1354 original author of that file, following a query about its status.
1355
1356 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1357 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1358
1359 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1360 quantifier applied to a forward-referencing subroutine call, could compile
1361 incorrect code or give the error "internal error: previously-checked
1362 referenced subpattern not found".
1363
1364 6. Both MS Visual Studio and Symbian OS have problems with initializing
1365 variables to point to external functions. For these systems, therefore,
1366 pcre_malloc etc. are now initialized to local functions that call the
1367 relevant global functions.
1368
1369 7. There were two entries missing in the vectors called coptable and poptable
1370 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1371 I've fixed the data, and added a kludgy way of testing at compile time that
1372 the lengths are correct (equal to the number of opcodes).
1373
1374 8. Following on from 7, I added a similar kludge to check the length of the
1375 eint vector in pcreposix.c.
1376
1377 9. Error texts for pcre_compile() are held as one long string to avoid too
1378 much relocation at load time. To find a text, the string is searched,
1379 counting zeros. There was no check for running off the end of the string,
1380 which could happen if a new error number was added without updating the
1381 string.
1382
1383 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1384
1385 11. \K was not working if it appeared in an atomic group or in a group that
1386 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1387 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1388 the assertion is positive, but not if it is negative.
1389
1390 12. Change 11 fortuitously reduced the size of the stack frame used in the
1391 "match()" function of pcre_exec.c by one pointer. Forthcoming
1392 implementation of support for (*MARK) will need an extra pointer on the
1393 stack; I have reserved it now, so that the stack frame size does not
1394 decrease.
1395
1396 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1397 item in branch that calls a recursion is a subroutine call - as in the
1398 second branch in the above example - was incorrectly given the compile-
1399 time error "recursive call could loop indefinitely" because pcre_compile()
1400 was not correctly checking the subroutine for matching a non-empty string.
1401
1402 14. The checks for overrunning compiling workspace could trigger after an
1403 overrun had occurred. This is a "should never occur" error, but it can be
1404 triggered by pathological patterns such as hundreds of nested parentheses.
1405 The checks now trigger 100 bytes before the end of the workspace.
1406
1407 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1408
1409
1410 Version 8.01 19-Jan-2010
1411 ------------------------
1412
1413 1. If a pattern contained a conditional subpattern with only one branch (in
1414 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1415 computed the wrong minimum data length (which is of course zero for such
1416 subpatterns). This could cause incorrect "no match" results.
1417
1418 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1419 the pattern is reset in the first branch, pcre_compile() failed with
1420 "internal error: code overflow at offset...". This happened only when
1421 the reset was to the original external option setting. (An optimization
1422 abstracts leading options settings into an external setting, which was the
1423 cause of this.)
1424
1425 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1426 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1427 assertion pattern did not match (meaning that the assertion was true), it
1428 was incorrectly treated as false if the SKIP had been reached during the
1429 matching. This also applied to assertions used as conditions.
1430
1431 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1432 assertion subpattern, including such a pattern used as a condition,
1433 unpredictable results occurred, instead of the error return
1434 PCRE_ERROR_DFA_UITEM.
1435
1436 5. The C++ GlobalReplace function was not working like Perl for the special
1437 situation when an empty string is matched. It now does the fancy magic
1438 stuff that is necessary.
1439
1440 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1441 removed. (These were left over from very, very early versions of PCRE.)
1442
1443 7. Some cosmetic changes to the code to make life easier when compiling it
1444 as part of something else:
1445
1446 (a) Change DEBUG to PCRE_DEBUG.
1447
1448 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1449 called "current" as "current_branch", to prevent a collision with the
1450 Linux macro when compiled as a kernel module.
1451
1452 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1453 prevent a collision with the Linux macro when compiled as a kernel
1454 module.
1455
1456 8. In pcre_compile() there are some checks for integer overflows that used to
1457 cast potentially large values to (double). This has been changed to that
1458 when building, a check for int64_t is made, and if it is found, it is used
1459 instead, thus avoiding the use of floating point arithmetic. (There is no
1460 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1461 double.
1462
1463 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1464 2005 (difference between two addresses compared to an unsigned value).
1465
1466 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1467 custom one, because of the following reported problem in Windows:
1468
1469 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1470 under Win32.
1471 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1472 therefore missing the function definition.
1473 - The compiler thus generates a "C" signature for the test function.
1474 - The linker fails to find the "C" function.
1475 - PCRE fails to configure if asked to do so against libbz2.
1476
1477 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1478 messages were output:
1479
1480 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1481 rerunning libtoolize, to keep the correct libtool macros in-tree.
1482 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1483
1484 I have done both of these things.
1485
1486 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1487 most of the time, it *can* run out if it is given a pattern that contains a
1488 runaway infinite recursion. I updated the discussion in the pcrestack man
1489 page.
1490
1491 13. Now that we have gone to the x.xx style of version numbers, the minor
1492 version may start with zero. Using 08 or 09 is a bad idea because users
1493 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1494 interpreted as invalid octal numbers. I've updated the previous comment in
1495 configure.ac, and also added a check that gives an error if 08 or 09 are
1496 used.
1497
1498 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1499 causing partial matching to fail when the end of the subject matched \W
1500 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1501
1502 15. There were some discrepancies between the declarations in pcre_internal.h
1503 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1504 their definitions. The declarations used "const uschar *" and the
1505 definitions used USPTR. Even though USPTR is normally defined as "const
1506 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1507 reported that: "This difference in casting confuses some C++ compilers, for
1508 example, SunCC recognizes above declarations as different functions and
1509 generates broken code for hbpcre." I have changed the declarations to use
1510 USPTR.
1511
1512 16. GNU libtool is named differently on some systems. The autogen.sh script now
1513 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1514 (FreeBSD).
1515
1516 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1517 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1518 comment: "Figure out how to create a longlong from a string: strtoll and
1519 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1520 instance, but it only takes 2 args instead of 3!"
1521
1522 18. A subtle bug concerned with back references has been fixed by a change of
1523 specification, with a corresponding code fix. A pattern such as
1524 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1525 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1526 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1527 same bug. Such groups have to be quantified to be useful, or contained
1528 inside another quantified group. (If there's no repetition, the reference
1529 can never match.) The problem arises because, having left the group and
1530 moved on to the rest of the pattern, a later failure that backtracks into
1531 the group uses the captured value from the final iteration of the group
1532 rather than the correct earlier one. I have fixed this in PCRE by forcing
1533 any group that contains a reference to itself to be an atomic group; that
1534 is, there cannot be any backtracking into it once it has completed. This is
1535 similar to recursive and subroutine calls.
1536
1537
1538 Version 8.00 19-Oct-09
1539 ----------------------
1540
1541 1. The table for translating pcre_compile() error codes into POSIX error codes
1542 was out-of-date, and there was no check on the pcre_compile() error code
1543 being within the table. This could lead to an OK return being given in
1544 error.
1545
1546 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1547 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1548 in a Windows environment.
1549
1550 3. The pcregrep --count option prints the count for each file even when it is
1551 zero, as does GNU grep. However, pcregrep was also printing all files when
1552 --files-with-matches was added. Now, when both options are given, it prints
1553 counts only for those files that have at least one match. (GNU grep just
1554 prints the file name in this circumstance, but including the count seems
1555 more useful - otherwise, why use --count?) Also ensured that the
1556 combination -clh just lists non-zero counts, with no names.
1557
1558 4. The long form of the pcregrep -F option was incorrectly implemented as
1559 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1560 but it seems right to fix it, and I didn't think it was worth preserving
1561 the old behaviour.
1562
1563 5. The command line items --regex=pattern and --regexp=pattern were not
1564 recognized by pcregrep, which required --regex pattern or --regexp pattern
1565 (with a space rather than an '='). The man page documented the '=' forms,
1566 which are compatible with GNU grep; these now work.
1567
1568 6. No libpcreposix.pc file was created for pkg-config; there was just
1569 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1570
1571 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1572 when UCP support is not needed, by modifying the Python script that
1573 generates it from Unicode data files. This should not matter if the module
1574 is correctly used as a library, but I received one complaint about 50K of
1575 unwanted data. My guess is that the person linked everything into his
1576 program rather than using a library. Anyway, it does no harm.
1577
1578 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1579 was a minimum greater than 1 for a wide character in a possessive
1580 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1581 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1582 character. Chaos in the form of incorrect output or a compiling loop could
1583 result.
1584
1585 9. The restrictions on what a pattern can contain when partial matching is
1586 requested for pcre_exec() have been removed. All patterns can now be
1587 partially matched by this function. In addition, if there are at least two
1588 slots in the offset vector, the offset of the earliest inspected character
1589 for the match and the offset of the end of the subject are set in them when
1590 PCRE_ERROR_PARTIAL is returned.
1591
1592 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1593 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1594 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1595 and may be more useful for multi-segment matching.
1596
1597 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1598 used to be given if ever the end of the subject was reached; now it is
1599 given only if matching could not proceed because another character was
1600 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1601 string "Z", which now yields "no match" instead of "partial match". In the
1602 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1603 final character ended with (*FAIL).
1604
1605 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1606 if the pattern had a "must contain" character that was already found in the
1607 earlier partial match, unless partial matching was again requested. For
1608 example, with the pattern /dog.(body)?/, the "must contain" character is
1609 "g". If the first part-match was for the string "dog", restarting with
1610 "sbody" failed. This bug has been fixed.
1611
1612 13. The string returned by pcre_dfa_exec() after a partial match has been
1613 changed so that it starts at the first inspected character rather than the
1614 first character of the match. This makes a difference only if the pattern
1615 starts with a lookbehind assertion or \b or \B (\K is not supported by
1616 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1617 matching functions compatible, and I think it's the right thing to do.
1618
1619 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1620 so that the demonstration program is easily available in environments where
1621 PCRE has not been installed from source.
1622
1623 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1624 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1625 library.
1626
1627 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1628 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1629 is not the first non-POSIX option to be added. Clearly some people find
1630 these options useful.
1631
1632 17. If a caller to the POSIX matching function regexec() passes a non-zero
1633 value for nmatch with a NULL value for pmatch, the value of
1634 nmatch is forced to zero.
1635
1636 18. RunGrepTest did not have a test for the availability of the -u option of
1637 the diff command, as RunTest does. It now checks in the same way as
1638 RunTest, and also checks for the -b option.
1639
1640 19. If an odd number of negated classes containing just a single character
1641 interposed, within parentheses, between a forward reference to a named
1642 subpattern and the definition of the subpattern, compilation crashed with
1643 an internal error, complaining that it could not find the referenced
1644 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1645 [The bug was that it was starting one character too far in when skipping
1646 over the character class, thus treating the ] as data rather than
1647 terminating the class. This meant it could skip too much.]
1648
1649 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1650 /g option in pcretest when the pattern contains \K, which makes it possible
1651 to have an empty string match not at the start, even when the pattern is
1652 anchored. Updated pcretest and pcredemo to use this option.
1653
1654 21. If the maximum number of capturing subpatterns in a recursion was greater
1655 than the maximum at the outer level, the higher number was returned, but
1656 with unset values at the outer level. The correct (outer level) value is
1657 now given.
1658
1659 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1660 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1661 make it do so. The string so far is captured, making this feature
1662 compatible with Perl.
1663
1664 23. The tests have been re-organized, adding tests 11 and 12, to make it
1665 possible to check the Perl 5.10 features against Perl 5.10.
1666
1667 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1668 pattern matches a fixed length string. PCRE did not allow this; now it
1669 does. Neither allows recursion.
1670
1671 25. I finally figured out how to implement a request to provide the minimum
1672 length of subject string that was needed in order to match a given pattern.
1673 (It was back references and recursion that I had previously got hung up
1674 on.) This code has now been added to pcre_study(); it finds a lower bound
1675 to the length of subject needed. It is not necessarily the greatest lower
1676 bound, but using it to avoid searching strings that are too short does give
1677 some useful speed-ups. The value is available to calling programs via
1678 pcre_fullinfo().
1679
1680 26. While implementing 25, I discovered to my embarrassment that pcretest had
1681 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1682 study optimizations had never been tested with that matching function.
1683 Oops. What is worse, even when it was passed study data, there was a bug in
1684 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1685 were also very few tests of studied patterns with pcre_dfa_exec().
1686
1687 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1688 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1689 on the other side of the coin, they are no longer allowed to have different
1690 names, because these cannot be distinguished in PCRE, and this has caused
1691 confusion. (This is a difference from Perl.)
1692
1693 28. When duplicate subpattern names are present (necessarily with different
1694 numbers, as required by 27 above), and a test is made by name in a
1695 conditional pattern, either for a subpattern having been matched, or for
1696 recursion in such a pattern, all the associated numbered subpatterns are
1697 tested, and the overall condition is true if the condition is true for any
1698 one of them. This is the way Perl works, and is also more like the way
1699 testing by number works.
1700
1701
1702 Version 7.9 11-Apr-09
1703 ---------------------
1704
1705 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1706 (pcretest), all targets were linked against these libraries. This included
1707 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1708 libraries. This caused unwanted dependencies to be created. This problem
1709 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1710 pcretest is linked with readline.
1711
1712 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1713 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1714 moved outside it again, because FALSE and TRUE are already defined in AIX,
1715 but BOOL is not.
1716
1717 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1718 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1719
1720 4. The pcregrep documentation said spaces were inserted as well as colons (or
1721 hyphens) following file names and line numbers when outputting matching
1722 lines. This is not true; no spaces are inserted. I have also clarified the
1723 wording for the --colour (or --color) option.
1724
1725 5. In pcregrep, when --colour was used with -o, the list of matching strings
1726 was not coloured; this is different to GNU grep, so I have changed it to be
1727 the same.
1728
1729 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1730 each matching line was coloured. Now it goes on to look for further matches
1731 of any of the test patterns, which is the same behaviour as GNU grep.
1732
1733 7. A pattern that could match an empty string could cause pcregrep to loop; it
1734 doesn't make sense to accept an empty string match in pcregrep, so I have
1735 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1736 seems to be how GNU grep behaves. [But see later change 40 for release
1737 8.33.]
1738
1739 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1740 start or after a newline", because the conditional assertion was not being
1741 correctly handled. The rule now is that both the assertion and what follows
1742 in the first alternative must satisfy the test.
1743
1744 9. If auto-callout was enabled in a pattern with a conditional group whose
1745 condition was an assertion, PCRE could crash during matching, both with
1746 pcre_exec() and pcre_dfa_exec().
1747
1748 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1749 used for matching.
1750
1751 11. Unicode property support in character classes was not working for
1752 characters (bytes) greater than 127 when not in UTF-8 mode.
1753
1754 12. Added the -M command line option to pcretest.
1755
1756 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1757
1758 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1759
1760 16. Added comments and documentation about mis-use of no_arg in the C++
1761 wrapper.
1762
1763 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1764 from Martin Jerabek that uses macro names for all relevant character and
1765 string constants.
1766
1767 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1768 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1769 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1770 these, but not everybody uses configure.
1771
1772 19. A conditional group that had only one branch was not being correctly
1773 recognized as an item that could match an empty string. This meant that an
1774 enclosing group might also not be so recognized, causing infinite looping
1775 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1776 with the subject "ab", where knowledge that the repeated group can match
1777 nothing is needed in order to break the loop.
1778
1779 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1780 exec(), but without supplying a callout function, matching went wrong.
1781
1782 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1783 leak if the size of the offset vector was greater than 30. When the vector
1784 is smaller, the saved offsets during recursion go onto a local stack
1785 vector, but for larger vectors malloc() is used. It was failing to free
1786 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1787 error, in fact).
1788
1789 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1790 heapframe that is used only when UTF-8 support is enabled. This caused no
1791 problem, but was untidy.
1792
1793 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1794 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1795 included within another project.
1796
1797 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1798 slightly modified by me:
1799
1800 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1801 not building pcregrep.
1802
1803 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1804 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1805
1806 25. Forward references, both numeric and by name, in patterns that made use of
1807 duplicate group numbers, could behave incorrectly or give incorrect errors,
1808 because when scanning forward to find the reference group, PCRE was not
1809 taking into account the duplicate group numbers. A pattern such as
1810 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1811
1812 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1813 the feature of a custom pointer more persuasive (as requested by a user).
1814
1815 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1816 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1817 pre-defined.
1818
1819 28. Added support for (*UTF8) at the start of a pattern.
1820
1821 29. Arrange for flags added by the "release type" setting in CMake to be shown
1822 in the configuration summary.
1823
1824
1825 Version 7.8 05-Sep-08
1826 ---------------------
1827
1828 1. Replaced UCP searching code with optimized version as implemented for Ad
1829 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1830 stage table and inline lookup instead of a function, giving speed ups of 2
1831 to 5 times on some simple patterns that I tested. Permission was given to
1832 distribute the MultiStage2.py script that generates the tables (it's not in
1833 the tarball, but is in the Subversion repository).
1834
1835 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1836 scripts.
1837
1838 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1839 a group with a zero qualifier. The result of the study could be incorrect,
1840 or the function might crash, depending on the pattern.
1841
1842 4. Caseless matching was not working for non-ASCII characters in back
1843 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1844 It now works when Unicode Property Support is available.
1845
1846 5. In pcretest, an escape such as \x{de} in the data was always generating
1847 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1848 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1849 truncation.
1850
1851 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1852
1853 7. Added two (int) casts to pcregrep when printing the difference of two
1854 pointers, in case they are 64-bit values.
1855
1856 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1857 test 2 if it fails.
1858
1859 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1860 and a #define of that name to empty if it is not externally set. This is to
1861 allow users of MSVC to set it if necessary.
1862
1863 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1864 the convenience functions in the pcre_get.c source file.
1865
1866 11. An option change at the start of a pattern that had top-level alternatives
1867 could cause overwriting and/or a crash. This command provoked a crash in
1868 some environments:
1869
1870 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1871
1872 This potential security problem was recorded as CVE-2008-2371.
1873
1874 12. For a pattern where the match had to start at the beginning or immediately
1875 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1876 pcre_dfa_exec() could read past the end of the passed subject if there was
1877 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1878 pcretest so that it places the subject at the end of its malloc-ed buffer.
1879
1880 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1881 exec() might read past the end of the data buffer in UTF-8 mode.
1882
1883 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1884 the data contained the byte 0x85 as part of a UTF-8 character within its
1885 first line. This applied both to normal and DFA matching.
1886
1887 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1888 /^[^d]*?$/8 failed to match "abc".
1889
1890 16. Added a missing copyright notice to pcrecpp_internal.h.
1891
1892 17. Make it more clear in the documentation that values returned from
1893 pcre_exec() in ovector are byte offsets, not character counts.
1894
1895 18. Tidied a few places to stop certain compilers from issuing warnings.
1896
1897 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1898 supplied by Stefan Weber. I made a further small update for 7.8 because
1899 there is a change of source arrangements: the pcre_searchfuncs.c module is
1900 replaced by pcre_ucd.c.
1901
1902
1903 Version 7.7 07-May-08
1904 ---------------------
1905
1906 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1907 a string to a long long, pretend we don't even have a long long." This is
1908 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1909
1910 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1911 pre-7.6 versions, which defined a global no_arg variable instead of putting
1912 it in the RE class. (See also #8 below.)
1913
1914 3. Remove a line of dead code, identified by coverity and reported by Nuno
1915 Lopes.
1916
1917 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1918
1919 (1) The include/exclude patterns were being applied to the whole pathnames
1920 of files, instead of just to the final components.
1921
1922 (2) If there was more than one level of directory, the subdirectories were
1923 skipped unless they satisfied the include/exclude conditions. This is
1924 inconsistent with GNU grep (and could even be seen as contrary to the
1925 pcregrep specification - which I improved to make it absolutely clear).
1926 The action now is always to scan all levels of directory, and just
1927 apply the include/exclude patterns to regular files.
1928
1929 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1930 --exclude_dir in the tests to avoid scanning .svn directories.
1931
1932 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1933 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1934 doesn't support NULs in patterns.
1935
1936 7. Added some missing "const"s to declarations of static tables in
1937 pcre_compile.c and pcre_dfa_exec.c.
1938
1939 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1940 caused by fix #2 above. (Subsequently also a second patch to fix the
1941 first patch. And a third patch - this was a messy problem.)
1942
1943 9. Applied Craig's patch to remove the use of push_back().
1944
1945 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1946 matching function regexec().
1947
1948 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1949 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1950 references. PCRE supports relative numbers with this syntax (I don't think
1951 Oniguruma does).
1952
1953 12. Previously, a group with a zero repeat such as (...){0} was completely
1954 omitted from the compiled regex. However, this means that if the group
1955 was called as a subroutine from elsewhere in the pattern, things went wrong
1956 (an internal error was given). Such groups are now left in the compiled
1957 pattern, with a new opcode that causes them to be skipped at execution
1958 time.
1959
1960 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1961 to the way PCRE behaves:
1962
1963 (a) A lone ] character is dis-allowed (Perl treats it as data).
1964
1965 (b) A back reference to an unmatched subpattern matches an empty string
1966 (Perl fails the current match path).
1967
1968 (c) A data ] in a character class must be notated as \] because if the
1969 first data character in a class is ], it defines an empty class. (In
1970 Perl it is not possible to have an empty class.) The empty class []
1971 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1972 The negative empty class [^] matches any one character, independently
1973 of the DOTALL setting.
1974
1975 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1976 non-existent subpattern following a character class starting with ']' and
1977 containing () gave an internal compiling error instead of "reference to
1978 non-existent subpattern". Fortunately, when the pattern did exist, the
1979 compiled code was correct. (When scanning forwards to check for the
1980 existence of the subpattern, it was treating the data ']' as terminating
1981 the class, so got the count wrong. When actually compiling, the reference
1982 was subsequently set up correctly.)
1983
1984 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1985 it was being rejected as not supported by pcre_dfa_exec(), even though
1986 other assertions are supported. I have made pcre_dfa_exec() support
1987 (*FAIL).
1988
1989 16. The implementation of 13c above involved the invention of a new opcode,
1990 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1991 cannot be changed at match time, I realized I could make a small
1992 improvement to matching performance by compiling OP_ALLANY instead of
1993 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1994 on the OP_ANY path.
1995
1996 17. Compiling pcretest on Windows with readline support failed without the
1997 following two fixes: (1) Make the unistd.h include conditional on
1998 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1999
2000 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2001 ncurses library to be included for pcretest when ReadLine support is
2002 requested, but also to allow for it to be overridden. This patch came from
2003 Daniel Bergström.
2004
2005 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2006 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2007 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2008 spotting this.
2009
2010
2011 Version 7.6 28-Jan-08
2012 ---------------------
2013
2014 1. A character class containing a very large number of characters with
2015 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2016 overflow.
2017
2018 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2019 HAVE_LONG_LONG is not defined.
2020
2021 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2022 bring them up to date and include new features. This patch includes:
2023
2024 - Fixed PH's badly added libz and libbz2 support.
2025 - Fixed a problem with static linking.
2026 - Added pcredemo. [But later removed - see 7 below.]
2027 - Fixed dftables problem and added an option.
2028 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2029 HAVE_LONG_LONG.
2030 - Added readline support for pcretest.
2031 - Added an listing of the option settings after cmake has run.
2032
2033 4. A user submitted a patch to Makefile that makes it easy to create
2034 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2035 Makefile.am that cause it to include this special target, without
2036 affecting anything else. Note that the same mingw target plus all
2037 the other distribution libraries and programs are now supported
2038 when configuring with CMake (see 6 below) instead of with
2039 Configure/Make.
2040
2041 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2042 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2043 exported in the Windows port". It has not yet been confirmed that the patch
2044 solves the problem, but it does no harm.
2045
2046 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2047 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2048 with CMake, and also correct the comment about stack recursion.
2049
2050 7. Remove the automatic building of pcredemo from the ./configure system and
2051 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2052 of a program that users should build themselves after PCRE is installed, so
2053 building it automatically is not really right. What is more, it gave
2054 trouble in some build environments.
2055
2056 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2057
2058
2059 Version 7.5 10-Jan-08
2060 ---------------------
2061
2062 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2063 values in parens when parsing an RE using the C++ wrapper."
2064
2065 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2066 Characters greater than 255 were excluded from the class instead of being
2067 included.
2068
2069 3. The same bug as (2) above applied to negated POSIX classes such as
2070 [:^space:].
2071
2072 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2073 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2074 I have changed it.
2075
2076 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2077 first named subpattern) and a construct such as (?&a) would reference the
2078 first named subpattern whose name started with "a" (in other words, the
2079 length check was missing). Both these problems are fixed. "Subpattern name
2080 expected" is now given for (?&) (a zero-length name), and this patch also
2081 makes it give the same error for \k'' (previously it complained that that
2082 was a reference to a non-existent subpattern).
2083
2084 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2085 this is right because (?- can be followed by option settings as well as by
2086 digits. I have, however, made the messages clearer.
2087
2088 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2089 than the number used in the conditional) now cause a compile-time error.
2090 This is actually not compatible with Perl, which accepts such patterns, but
2091 treats the conditional as always being FALSE (as PCRE used to), but it
2092 seems to me that giving a diagnostic is better.
2093
2094 8. Change "alphameric" to the more common word "alphanumeric" in comments
2095 and messages.
2096
2097 9. Fix two occurrences of "backslash" in comments that should have been
2098 "backspace".
2099
2100 10. Remove two redundant lines of code that can never be obeyed (their function
2101 was moved elsewhere).
2102
2103 11. The program that makes PCRE's Unicode character property table had a bug
2104 which caused it to generate incorrect table entries for sequences of
2105 characters that have the same character type, but are in different scripts.
2106 It amalgamated them into a single range, with the script of the first of
2107 them. In other words, some characters were in the wrong script. There were
2108 thirteen such cases, affecting characters in the following ranges:
2109
2110 U+002b0 - U+002c1
2111 U+0060c - U+0060d
2112 U+0061e - U+00612
2113 U+0064b - U+0065e
2114 U+0074d - U+0076d
2115 U+01800 - U+01805
2116 U+01d00 - U+01d77
2117 U+01d9b - U+01dbf
2118 U+0200b - U+0200f
2119 U+030fc - U+030fe
2120 U+03260 - U+0327f
2121 U+0fb46 - U+0fbb1
2122 U+10450 - U+1049d
2123
2124 12. The -o option (show only the matching part of a line) for pcregrep was not
2125 compatible with GNU grep in that, if there was more than one match in a
2126 line, it showed only the first of them. It now behaves in the same way as
2127 GNU grep.
2128
2129 13. If the -o and -v options were combined for pcregrep, it printed a blank
2130 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2131 does the same. The return code can be used to tell if there were any
2132 non-matching lines.
2133
2134 14. Added --file-offsets and --line-offsets to pcregrep.
2135
2136 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2137 infinitely looping recursion. The bug was that positive lookaheads were not
2138 being skipped when checking for a possible empty match (negative lookaheads
2139 and both kinds of lookbehind were skipped).
2140
2141 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2142 inclusion of <windows.h> to before rather than after the definition of
2143 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2144
2145 17. Specifying a possessive quantifier with a specific limit for a Unicode
2146 character property caused pcre_compile() to compile bad code, which led at
2147 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2148 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2149 caused the error; without that there was no problem.
2150
2151 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2152
2153 19. Added --enable-pcretest-libreadline.
2154
2155 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2156 RE::GlobalReplace(). As a result, the number of replacements returned was
2157 double what it should be. I removed one of the increments, but Craig sent a
2158 later patch that removed the other one (the right fix) and added unit tests
2159 that check the return values (which was not done before).
2160
2161 21. Several CMake things:
2162
2163 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2164 the names libpcre and libpcreposix, not just pcre and pcreposix.
2165
2166 (2) The above change means that pcretest and pcregrep are now correctly
2167 linked with the newly-built libraries, not previously installed ones.
2168
2169 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2170
2171 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2172 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2173 UTF-8 newline character). The key issue is that the pattern starts .*;
2174 this means that the match must be either at the beginning, or after a
2175 newline. The bug was in the code for advancing after a failed match and
2176 checking that the new position followed a newline. It was not taking
2177 account of UTF-8 characters correctly.
2178
2179 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2180 character classes. PCRE was not treating the sequence [:...:] as a
2181 character class unless the ... were all letters. Perl, however, seems to
2182 allow any characters between [: and :], though of course it rejects as
2183 unknown any "names" that contain non-letters, because all the known class
2184 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2185 for example, whereas PCRE did not - it did not recognize a POSIX character
2186 class. This seemed a bit dangerous, so the code has been changed to be
2187 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2188 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2189 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2190 Perl does, and where it didn't before.
2191
2192 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2193 Windows environments %n is disabled by default.
2194
2195
2196 Version 7.4 21-Sep-07
2197 ---------------------
2198
2199 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2200 means that a class such as [\s] counted as "explicit reference to CR or
2201 LF". That isn't really right - the whole point of the change was to try to
2202 help when there was an actual mention of one of the two characters. So now
2203 the change happens only if \r or \n (or a literal CR or LF) character is
2204 encountered.
2205
2206 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2207 of both had grown to the point where there were only 3 bits left.
2208 Fortunately, there was spare space in the data structure, and so I have
2209 moved the internal flags into a new 16-bit field to free up more option
2210 bits.
2211
2212 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2213 but did not set the internal JCHANGED flag - either of these is enough to
2214 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2215 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2216 start sets both bits.
2217
2218 4. Added options (at build time, compile time, exec time) to change \R from
2219 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2220
2221 5. doc/pcresyntax.html was missing from the distribution.
2222
2223 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2224 compatibility, even though it is no longer used.
2225
2226 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2227 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2228 windows.h file is present (where different names are used). [This was
2229 reversed later after testing - see 16 below.]
2230
2231 8. Changed all #include <config.h> to #include "config.h". There were also
2232 some further <pcre.h> cases that I changed to "pcre.h".
2233
2234 9. When pcregrep was used with the --colour option, it missed the line ending
2235 sequence off the lines that it output.
2236
2237 10. It was pointed out to me that arrays of string pointers cause lots of
2238 relocations when a shared library is dynamically loaded. A technique of
2239 using a single long string with a table of offsets can drastically reduce
2240 these. I have refactored PCRE in four places to do this. The result is
2241 dramatic:
2242
2243 Originally: 290
2244 After changing UCP table: 187
2245 After changing error message table: 43
2246 After changing table of "verbs" 36
2247 After changing table of Posix names 22
2248
2249 Thanks to the folks working on Gregex for glib for this insight.
2250
2251 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2252 unicode-properties was also set.
2253
2254 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2255
2256 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2257 checked only for CRLF.
2258
2259 14. Added casts to pcretest.c to avoid compiler warnings.
2260
2261 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2262
2263 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2264 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2265 entirely. This removes changes made in 7 above.
2266
2267 17. The CMake files have been updated, and there is now more information about
2268 building with CMake in the NON-UNIX-USE document.
2269
2270
2271 Version 7.3 28-Aug-07
2272 ---------------------
2273
2274 1. In the rejigging of the build system that eventually resulted in 7.1, the
2275 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2276 brackets there is not right, since it causes compilers to look for an
2277 installed pcre.h, not the version that is in the source that is being
2278 compiled (which of course may be different). I have changed it back to:
2279
2280 #include "pcre.h"
2281
2282 I have a vague recollection that the change was concerned with compiling in
2283 different directories, but in the new build system, that is taken care of
2284 by the VPATH setting the Makefile.
2285
2286 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2287 when the subject happened to end in the byte 0x85 (e.g. if the last
2288 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2289 characters but of course it shouldn't be taken as a newline when it is part
2290 of another character. The bug was that, for an unlimited repeat of . in
2291 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2292 characters when looking for a newline.
2293
2294 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2295
2296 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2297 in debug output.
2298
2299 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2300 long printing in the pcrecpp unittest when running under MinGW.
2301
2302 6. ESC_K was left out of the EBCDIC table.
2303
2304 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2305 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2306 limit also applies to "virtual nesting" when a pattern is recursive, and in
2307 this case 1000 isn't so big. I have been able to remove this limit at the
2308 expense of backing off one optimization in certain circumstances. Normally,
2309 when pcre_exec() would call its internal match() function recursively and
2310 immediately return the result unconditionally, it uses a "tail recursion"
2311 feature to save stack. However, when a subpattern that can match an empty
2312 string has an unlimited repetition quantifier, it no longer makes this
2313 optimization. That gives it a stack frame in which to save the data for
2314 checking that an empty string has been matched. Previously this was taken
2315 from the 1000-entry workspace that had been reserved. So now there is no
2316 explicit limit, but more stack is used.
2317
2318 8. Applied Daniel's patches to solve problems with the import/export magic
2319 syntax that is required for Windows, and which was going wrong for the
2320 pcreposix and pcrecpp parts of the library. These were overlooked when this
2321 problem was solved for the main library.
2322
2323 9. There were some crude static tests to avoid integer overflow when computing
2324 the size of patterns that contain repeated groups with explicit upper
2325 limits. As the maximum quantifier is 65535, the maximum group length was
2326 set at 30,000 so that the product of these two numbers did not overflow a
2327 32-bit integer. However, it turns out that people want to use groups that
2328 are longer than 30,000 bytes (though not repeat them that many times).
2329 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2330 made it possible to implement the integer overflow checks in a much more
2331 dynamic way, which I have now done. The artificial limitation on group
2332 length has been removed - we now have only the limit on the total length of
2333 the compiled pattern, which depends on the LINK_SIZE setting.
2334
2335 10. Fixed a bug in the documentation for get/copy named substring when
2336 duplicate names are permitted. If none of the named substrings are set, the
2337 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2338 empty string.
2339
2340 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2341 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2342 because the ] is interpreted as the first data character and the
2343 terminating ] is not found. PCRE has been made compatible with Perl in this
2344 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2345 cause memory overwriting.
2346
2347 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2348 string has been matched (to stop an infinite loop). It was not recognizing
2349 a conditional subpattern that could match an empty string if that
2350 subpattern was within another subpattern. For example, it looped when
2351 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2352 condition was not nested. This bug has been fixed.
2353
2354 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2355 past the start of the subject in the presence of bytes with the top bit
2356 set, for example "\x8aBCD".
2357
2358 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2359 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2360
2361 14. Optimized (?!) to (*FAIL).
2362
2363 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2364 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2365 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2366 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2367 does: it's just the validity check that is more restrictive.
2368
2369 16. Inserted checks for integer overflows during escape sequence (backslash)
2370 processing, and also fixed erroneous offset values for syntax errors during
2371 backslash processing.
2372
2373 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2374 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2375
2376 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2377 caused an overrun.
2378
2379 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2380 something other than just ASCII characters) inside a group that had an
2381 unlimited repeat caused a loop at compile time (while checking to see
2382 whether the group could match an empty string).
2383
2384 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2385 [\P{Any}] did so. (Error in the code for printing property names.)
2386
2387 21. An orphan \E inside a character class could cause a crash.
2388
2389 22. A repeated capturing bracket such as (A)? could cause a wild memory
2390 reference during compilation.
2391
2392 23. There are several functions in pcre_compile() that scan along a compiled
2393 expression for various reasons (e.g. to see if it's fixed length for look
2394 behind). There were bugs in these functions when a repeated \p or \P was
2395 present in the pattern. These operators have additional parameters compared
2396 with \d, etc, and these were not being taken into account when moving along
2397 the compiled data. Specifically:
2398
2399 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2400 length.
2401
2402 (b) An item such as \pL+ within a repeated group could cause crashes or
2403 loops.
2404
2405 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2406 "reference to non-existent subpattern" error.
2407
2408 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2409
2410 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2411 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2412
2413 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2414
2415 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2416 character were causing crashes (broken optimization).
2417
2418 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2419 \p or \P) caused a compile-time loop.
2420
2421 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2422 break. For example, the unstudied pattern [\r\n]A does not match the string
2423 "\r\nA" because change 7.0/46 below moves the current point on by two
2424 characters after failing to match at the start. However, the pattern \nA
2425 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2426 the same is true. There doesn't seem any very clean way out of this, but
2427 what I have chosen to do makes the common cases work: PCRE now takes note
2428 of whether there can be an explicit match for \r or \n anywhere in the
2429 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2430 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2431 pattern has explicit CR or LF references.
2432
2433 29. Added (*CR) etc for changing newline setting at start of pattern.
2434
2435
2436 Version 7.2 19-Jun-07
2437 ---------------------
2438
2439 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2440 which is apparently normally available under Windows.
2441
2442 2. Re-jig the pcregrep tests with different newline settings in an attempt
2443 to make them independent of the local environment's newline setting.
2444
2445 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2446
2447 4. Some of the "internals" tests were previously cut out when the link size
2448 was not 2, because the output contained actual offsets. The recent new
2449 "Z" feature of pcretest means that these can be cut out, making the tests
2450 usable with all link sizes.
2451
2452 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2453 stack recursion. This gives a massive performance boost under BSD, but just
2454 a small improvement under Linux. However, it saves one field in the frame
2455 in all cases.
2456
2457 6. Added more features from the forthcoming Perl 5.10:
2458
2459 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2460 recursion call. It refers to the nth most recently opened parentheses.
2461
2462 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2463 to be opened parentheses.
2464
2465 (c) Conditions that refer to capturing parentheses can be specified
2466 relatively, for example, (?(-2)... or (?(+3)...
2467
2468 (d) \K resets the start of the current match so that everything before
2469 is not part of it.
2470
2471 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2472
2473 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2474 reference syntax.
2475
2476 (g) (?| introduces a group in which the numbering of parentheses in each
2477 alternative starts with the same number.
2478
2479 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2480
2481 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2482 PCRE_INFO_JCHANGED.
2483
2484 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2485 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2486 for detecting groups that can match an empty string.
2487
2488 9. A pattern with a very large number of alternatives (more than several
2489 hundred) was running out of internal workspace during the pre-compile
2490 phase, where pcre_compile() figures out how much memory will be needed. A
2491 bit of new cunning has reduced the workspace needed for groups with
2492 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2493 workspace instead of running out of the 4096 that are available.
2494
2495 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2496
2497 11. Applied patch from Google to remove an optimization that didn't quite work.
2498 The report of the bug said:
2499
2500 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2501 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2502 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2503
2504 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2505 it matched the wrong number of bytes.
2506
2507
2508 Version 7.1 24-Apr-07
2509 ---------------------
2510
2511 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2512 that is more "standard", making use of automake and other Autotools. There
2513 is some re-arrangement of the files and adjustment of comments consequent
2514 on this.
2515
2516 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2517 for recursive directory scanning broke on some systems because the files
2518 are not scanned in any specific order and on different systems the order
2519 was different. A call to "sort" has been inserted into RunGrepTest for the
2520 approprate test as a short-term fix. In the longer term there may be an
2521 alternative.
2522
2523 3. I had an email from Eric Raymond about problems translating some of PCRE's
2524 man pages to HTML (despite the fact that I distribute HTML pages, some
2525 people do their own conversions for various reasons). The problems
2526 concerned the use of low-level troff macros .br and .in. I have therefore
2527 removed all such uses from the man pages (some were redundant, some could
2528 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2529 HTML has been updated to handle .nf/.fi and to complain if it encounters
2530 .br or .in.
2531
2532 4. Updated comments in configure.ac that get placed in config.h.in and also
2533 arranged for config.h to be included in the distribution, with the name
2534 config.h.generic, for the benefit of those who have to compile without
2535 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2536
2537 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2538 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2539 makevp.bat for the current PCRE, using the additional files
2540 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2541
2542 6. A Windows user reported a minor discrepancy with test 2, which turned out
2543 to be caused by a trailing space on an input line that had got lost in his
2544 copy. The trailing space was an accident, so I've just removed it.
2545
2546 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2547 that is needed.
2548
2549 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2550 as "const" (a) because they are and (b) because it helps the PHP
2551 maintainers who have recently made a script to detect big data structures
2552 in the php code that should be moved to the .rodata section. I remembered
2553 to update Builducptable as well, so it won't revert if ucptable.h is ever
2554 re-created.
2555
2556 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2557 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2558 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2559 support is not required. This saves 1.5-2K of code, which is important in
2560 some applications.
2561
2562 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2563 so as not to refer to the tables, even though these functions will never be
2564 called when UTF-8 support is disabled. Otherwise there are problems with a
2565 shared library.
2566
2567 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2568
2569 (a) It was defining its arguments as char * instead of void *.
2570
2571 (b) It was assuming that all moves were upwards in memory; this was true
2572 a long time ago when I wrote it, but is no longer the case.
2573
2574 The emulated memove() is provided for those environments that have neither
2575 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2576 is clearly not the case, as these two bugs were recently reported.
2577
2578 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2579 and Detrail to create the HTML documentation, the .txt form of the man
2580 pages, and it removes trailing spaces from listed files. It also creates
2581 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2582 case, it wraps all the #defines with #ifndefs. This script should be run
2583 before "make dist".
2584
2585 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2586 with Unicode property support.
2587
2588 (a) For a maximizing quantifier, if the two different cases of the
2589 character were of different lengths in their UTF-8 codings (there are
2590 some cases like this - I found 11), and the matching function had to
2591 back up over a mixture of the two cases, it incorrectly assumed they
2592 were both the same length.
2593
2594 (b) When PCRE was configured to use the heap rather than the stack for
2595 recursion during matching, it was not correctly preserving the data for
2596 the other case of a UTF-8 character when checking ahead for a match
2597 while processing a minimizing repeat. If the check also involved
2598 matching a wide character, but failed, corruption could cause an
2599 erroneous result when trying to check for a repeat of the original
2600 character.
2601
2602 13. Some tidying changes to the testing mechanism:
2603
2604 (a) The RunTest script now detects the internal link size and whether there
2605 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2606 values substituted by "configure". (The RunGrepTest script already did
2607 this for UTF-8.) The configure.ac script no longer substitutes the
2608 relevant variables.
2609
2610 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2611 with length and offset values. This means that the output is different
2612 for different internal link sizes. Test 2 is skipped for link sizes
2613 other than 2 because of this, bypassing the problem. Unfortunately,
2614 there was also a test in test 3 (the locale tests) that used /B and
2615 failed for link sizes other than 2. Rather than cut the whole test out,
2616 I have added a new /Z option to pcretest that replaces the length and
2617 offset values with spaces. This is now used to make test 3 independent
2618 of link size. (Test 2 will be tidied up later.)
2619
2620 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2621 segmentation fault instead of returning the appropriate error message.
2622
2623 15. In multiline mode when the newline sequence was set to "any", the pattern
2624 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2625 This doesn't seem right; it now treats the CRLF combination as the line
2626 ending, and so does not match in that case. It's only a pattern such as ^$
2627 that would hit this one: something like ^ABC$ would have failed after \r
2628 and then tried again after \r\n.
2629
2630 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2631 in an attempt to make files that differ only in their line terminators
2632 compare equal. This works on Linux.
2633
2634 17. Under certain error circumstances pcregrep might try to free random memory
2635 as it exited. This is now fixed, thanks to valgrind.
2636
2637 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2638 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2639 was because its rules for how to advance for /g after matching an empty
2640 string at the end of a line did not allow for this case. They now check for
2641 it specially.
2642
2643 20. pcretest is supposed to handle patterns and data of any length, by
2644 extending its buffers when necessary. It was getting this wrong when the
2645 buffer for a data line had to be extended.
2646
2647 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2648 CRLF as a newline sequence.
2649
2650 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2651 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2652 I have nevertheless tidied it up.
2653
2654 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2655
2656 24. Added a man page for pcre-config.
2657
2658
2659 Version 7.0 19-Dec-06
2660 ---------------------
2661
2662 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2663 moving to gcc 4.1.1.
2664
2665 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2666 sys/time.h, which is documented as needed for this function. It doesn't
2667 seem to matter on Linux, but it showed up on some releases of OS X.
2668
2669 3. It seems that there are systems where bytes whose values are greater than
2670 127 match isprint() in the "C" locale. The "C" locale should be the
2671 default when a C program starts up. In most systems, only ASCII printing
2672 characters match isprint(). This difference caused the output from pcretest
2673 to vary, making some of the tests fail. I have changed pcretest so that:
2674
2675 (a) When it is outputting text in the compiled version of a pattern, bytes
2676 other than 32-126 are always shown as hex escapes.
2677
2678 (b) When it is outputting text that is a matched part of a subject string,
2679 it does the same, unless a different locale has been set for the match
2680 (using the /L modifier). In this case, it uses isprint() to decide.
2681
2682 4. Fixed a major bug that caused incorrect computation of the amount of memory
2683 required for a compiled pattern when options that changed within the
2684 pattern affected the logic of the preliminary scan that determines the
2685 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2686 that the computed length was too small. The symptoms of this bug were
2687 either the PCRE error "internal error: code overflow" from pcre_compile(),
2688 or a glibc crash with a message such as "pcretest: free(): invalid next
2689 size (fast)". Examples of patterns that provoked this bug (shown in
2690 pcretest format) are:
2691
2692 /(?-x: )/x
2693 /(?x)(?-x: \s*#\s*)/
2694 /((?i)[\x{c0}])/8
2695 /(?i:[\x{c0}])/8
2696
2697 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2698 is now done differently.
2699
2700 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2701 wrapper classes; (b) implement a new function in the C++ scanner that is
2702 more efficient than the old way of doing things because it avoids levels of
2703 recursion in the regex matching; (c) add a paragraph to the documentation
2704 for the FullMatch() function.
2705
2706 6. The escape sequence \n was being treated as whatever was defined as
2707 "newline". Not only was this contrary to the documentation, which states
2708 that \n is character 10 (hex 0A), but it also went horribly wrong when
2709 "newline" was defined as CRLF. This has been fixed.
2710
2711 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2712 was being set to -1 for the "end of line" case (supposedly a value that no
2713 character can have). Though this value is never used (the check for end of
2714 line is "zero bytes in current character"), it caused compiler complaints.
2715 I've changed it to 0xffffffff.
2716
2717 8. In pcre_version.c, the version string was being built by a sequence of
2718 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2719 string (as it is for production releases) called a macro with an empty
2720 argument. The C standard says the result of this is undefined. The gcc
2721 compiler treats it as an empty string (which was what was wanted) but it is
2722 reported that Visual C gives an error. The source has been hacked around to
2723 avoid this problem.
2724
2725 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2726 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2727 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2728 of them did).
2729
2730 10. Originally, pcretest opened its input and output without "b"; then I was
2731 told that "b" was needed in some environments, so it was added for release
2732 5.0 to both the input and output. (It makes no difference on Unix-like
2733 systems.) Later I was told that it is wrong for the input on Windows. I've
2734 now abstracted the modes into two macros, to make it easier to fiddle with
2735 them, and removed "b" from the input mode under Windows.
2736
2737 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2738
2739 12. Added -help and --help to pcretest as an official way of being reminded
2740 of the options.
2741
2742 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2743 and pcrecpp.cc because they annoy compilers at high warning levels.
2744
2745 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2746
2747 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2748 scripts are not C programs :-) and which was not noticed because it works
2749 on Linux.
2750
2751 16. pcretest is supposed to handle any length of pattern and data line (as one
2752 line or as a continued sequence of lines) by extending its input buffer if
2753 necessary. This feature was broken for very long pattern lines, leading to
2754 a string of junk being passed to pcre_compile() if the pattern was longer
2755 than about 50K.
2756
2757 17. I have done a major re-factoring of the way pcre_compile() computes the
2758 amount of memory needed for a compiled pattern. Previously, there was code
2759 that made a preliminary scan of the pattern in order to do this. That was
2760 OK when PCRE was new, but as the facilities have expanded, it has become
2761 harder and harder to keep it in step with the real compile phase, and there
2762 have been a number of bugs (see for example, 4 above). I have now found a
2763 cunning way of running the real compile function in a "fake" mode that
2764 enables it to compute how much memory it would need, while actually only
2765 ever using a few hundred bytes of working memory and without too many
2766 tests of the mode. This should make future maintenance and development
2767 easier. A side effect of this work is that the limit of 200 on the nesting
2768 depth of parentheses has been removed (though this was never a serious
2769 limitation, I suspect). However, there is a downside: pcre_compile() now
2770 runs more slowly than before (30% or more, depending on the pattern). I
2771 hope this isn't a big issue. There is no effect on runtime performance.
2772
2773 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2774 newline (only possible for the last line of a file) and it was a
2775 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2776
2777 19. Added additional timing features to pcretest. (1) The -tm option now times
2778 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2779 separate command line item, by a number that specifies the number of
2780 repeats to use when timing. The default is 50000; this gives better
2781 precision, but takes uncomfortably long for very large patterns.
2782
2783 20. Extended pcre_study() to be more clever in cases where a branch of a
2784 subpattern has no definite first character. For example, (a*|b*)[cd] would
2785 previously give no result from pcre_study(). Now it recognizes that the
2786 first character must be a, b, c, or d.
2787
2788 21. There was an incorrect error "recursive call could loop indefinitely" if
2789 a subpattern (or the entire pattern) that was being tested for matching an
2790 empty string contained only one non-empty item after a nested subpattern.
2791 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2792 incorrectly, because the \d was being skipped in the check.
2793
2794 22. The pcretest program now has a new pattern option /B and a command line
2795 option -b, which is equivalent to adding /B to every pattern. This causes
2796 it to show the compiled bytecode, without the additional information that
2797 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2798 is the same as /B/I).
2799
2800 23. A new optimization is now able automatically to treat some sequences such
2801 as a*b as a*+b. More specifically, if something simple (such as a character
2802 or a simple class like \d) has an unlimited quantifier, and is followed by
2803 something that cannot possibly match the quantified thing, the quantifier
2804 is automatically "possessified".
2805
2806 24. A recursive reference to a subpattern whose number was greater than 39
2807 went wrong under certain circumstances in UTF-8 mode. This bug could also
2808 have affected the operation of pcre_study().
2809
2810 25. Realized that a little bit of performance could be had by replacing
2811 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2812
2813 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2814
2815 27. Possessive quantifiers such as a++ were previously implemented by turning
2816 them into atomic groups such as ($>a+). Now they have their own opcodes,
2817 which improves performance. This includes the automatically created ones
2818 from 23 above.
2819
2820 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2821 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2822 the first matched character to be a colon. This applied both to named and
2823 numbered groups.
2824
2825 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2826
2827 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2828 building PCRE on Windows easier, so I have included it in the distribution.
2829
2830 31. There is now a check in pcretest against a ridiculously large number being
2831 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2832 loop, the loop is abandoned.
2833
2834 32. Forward references to subpatterns in conditions such as (?(2)...) where
2835 subpattern 2 is defined later cause pcre_compile() to search forwards in
2836 the pattern for the relevant set of parentheses. This search went wrong
2837 when there were unescaped parentheses in a character class, parentheses
2838 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2839
2840 33. "Subroutine" calls and backreferences were previously restricted to
2841 referencing subpatterns earlier in the regex. This restriction has now
2842 been removed.
2843
2844 34. Added a number of extra features that are going to be in Perl 5.10. On the
2845 whole, these are just syntactic alternatives for features that PCRE had
2846 previously implemented using the Python syntax or my own invention. The
2847 other formats are all retained for compatibility.
2848
2849 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2850 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2851 also .NET compatible.
2852
2853 (b) A recursion or subroutine call to a named group can now be defined as
2854 (?&name) as well as (?P>name).
2855
2856 (c) A backreference to a named group can now be defined as \k<name> or
2857 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2858 5.10, are also .NET compatible.
2859
2860 (d) A conditional reference to a named group can now use the syntax
2861 (?(<name>) or (?('name') as well as (?(name).
2862
2863 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2864 groups (named and numbered) that are never evaluated inline, but can be
2865 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2866 is always false. There may be only one alternative in such a group.
2867
2868 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2869 as the simple (?(R). The condition is true only if the most recent
2870 recursion is that of the given number or name. It does not search out
2871 through the entire recursion stack.
2872
2873 (g) The escape \gN or \g{N} has been added, where N is a positive or
2874 negative number, specifying an absolute or relative reference.
2875
2876 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2877 some "unreachable code" warnings.
2878
2879 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2880 things, this adds five new scripts.
2881
2882 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2883 There were also incompatibilities regarding the handling of \Q..\E inside
2884 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2885 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2886
2887 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2888 matches an empty string, and forcibly breaks the loop. There were bugs in
2889 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2890 against aaaa the result was just "a" rather than "aaaa", for example. Two
2891 separate and independent bugs (that affected different cases) have been
2892 fixed.
2893
2894 39. Refactored the code to abolish the use of different opcodes for small
2895 capturing bracket numbers. This is a tidy that I avoided doing when I
2896 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2897 The new approach is not only tidier, it makes it possible to reduce the
2898 memory needed to fix the previous bug (38).
2899
2900 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2901 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2902 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2903 mode.
2904
2905 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2906 report.
2907
2908 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2909 copy construction and assignment in the C++ wrapper.
2910
2911 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2912 couple of bugs that could have given wrong results in the "--newline=crlf"
2913 case.
2914
2915 44. Added a number of casts and did some reorganization of signed/unsigned int
2916 variables following suggestions from Dair Grant. Also renamed the variable
2917 "this" as "item" because it is a C++ keyword.
2918
2919 45. Arranged for dftables to add
2920
2921 #include "pcre_internal.h"
2922
2923 to pcre_chartables.c because without it, gcc 4.x may remove the array
2924 definition from the final binary if PCRE is built into a static library and
2925 dead code stripping is activated.
2926
2927 46. For an unanchored pattern, if a match attempt fails at the start of a
2928 newline sequence, and the newline setting is CRLF or ANY, and the next two
2929 characters are CRLF, advance by two characters instead of one.
2930
2931
2932 Version 6.7 04-Jul-06
2933 ---------------------
2934
2935 1. In order to handle tests when input lines are enormously long, pcretest has
2936 been re-factored so that it automatically extends its buffers when
2937 necessary. The code is crude, but this _is_ just a test program. The
2938 default size has been increased from 32K to 50K.
2939
2940 2. The code in pcre_study() was using the value of the re argument before
2941 testing it for NULL. (Of course, in any sensible call of the function, it
2942 won't be NULL.)
2943
2944 3. The memmove() emulation function in pcre_internal.h, which is used on
2945 systems that lack both memmove() and bcopy() - that is, hardly ever -
2946 was missing a "static" storage class specifier.
2947
2948 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2949 containing an extended class (one that cannot be represented by a bitmap
2950 because it contains high-valued characters or Unicode property items, e.g.
2951 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2952 pattern, but PCRE should not loop if you do not (it no longer does).
2953 [Detail: two cases were found: (a) a repeated subpattern containing an
2954 extended class; (b) a recursive reference to a subpattern that followed a
2955 previous extended class. It wasn't skipping over the extended class
2956 correctly when UTF-8 mode was not set.]
2957
2958 5. A negated single-character class was not being recognized as fixed-length
2959 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2960 compile error "lookbehind assertion is not fixed length".
2961
2962 6. The RunPerlTest auxiliary script was showing an unexpected difference
2963 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2964 write a Perl script that can interpret lines of an input file either as
2965 byte characters or as UTF-8, which is what "perltest" was being required to
2966 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2967 can't do is switch easily at run time between having the "use utf8;" pragma
2968 or not. In the end, I fudged it by using the RunPerlTest script to insert
2969 "use utf8;" explicitly for the UTF-8 tests.
2970
2971 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2972 the end of the subject string, contrary to the documentation and to what
2973 Perl does. This was true of both matching functions. Now it matches only at
2974 the start of the subject and immediately after *internal* newlines.
2975
2976 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2977 a pointer to an int instead of a pointer to an unsigned long int. This
2978 caused problems on 64-bit systems.
2979
2980 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2981 instance of the 'standard' template library not being so standard".
2982
2983 10. There was no check on the number of named subpatterns nor the maximum
2984 length of a subpattern name. The product of these values is used to compute
2985 the size of the memory block for a compiled pattern. By supplying a very
2986 long subpattern name and a large number of named subpatterns, the size
2987 computation could be caused to overflow. This is now prevented by limiting
2988 the length of names to 32 characters, and the number of named subpatterns
2989 to 10,000.
2990
2991 11. Subpatterns that are repeated with specific counts have to be replicated in
2992 the compiled pattern. The size of memory for this was computed from the
2993 length of the subpattern and the repeat count. The latter is limited to
2994 65535, but there was no limit on the former, meaning that integer overflow
2995 could in principle occur. The compiled length of a repeated subpattern is
2996 now limited to 30,000 bytes in order to prevent this.
2997
2998 12. Added the optional facility to have named substrings with the same name.
2999
3000 13. Added the ability to use a named substring as a condition, using the
3001 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3002 are numbers (not recommended). Forward references are permitted.
3003
3004 14. Added forward references in named backreferences (if you see what I mean).
3005
3006 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3007 pattern could run off the end of the subject. For example, the pattern
3008 "(?s)(.{1,5})"8 did this with the subject "ab".
3009
3010 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3011 PCRE_CASELESS was set when matching characters that were quantified with ?
3012 or *.
3013
3014 17. A character class other than a single negated character that had a minimum
3015 but no maximum quantifier - for example [ab]{6,} - was not handled
3016 correctly by pce_dfa_exec(). It would match only one character.
3017
3018 18. A valid (though odd) pattern that looked like a POSIX character
3019 class but used an invalid character after [ (for example [[,abc,]]) caused
3020 pcre_compile() to give the error "Failed: internal error: code overflow" or
3021 in some cases to crash with a glibc free() error. This could even happen if
3022 the pattern terminated after [[ but there just happened to be a sequence of
3023 letters, a binary zero, and a closing ] in the memory that followed.
3024
3025 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3026 over the years. Originally (before any Unicode support), just the bottom 8
3027 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3028 output from "man perlunicode" includes this:
3029
3030 The regular expression compiler produces polymorphic opcodes. That
3031 is, the pattern adapts to the data and automatically switches to
3032 the Unicode character scheme when presented with Unicode data--or
3033 instead uses a traditional byte scheme when presented with byte
3034 data.
3035
3036 Sadly, a wide octal escape does not cause a switch, and in a string with
3037 no other multibyte characters, these octal escapes are treated as before.
3038 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3039 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3040 Unicode string.
3041
3042 I have not perpetrated such confusion in PCRE. Up till now, it took just
3043 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3044 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3045 translate to the appropriate multibyte character.
3046
3047 29. Applied some refactoring to reduce the number of warnings from Microsoft
3048 and Borland compilers. This has included removing the fudge introduced
3049 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3050 a warning about an unused variable.
3051
3052 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3053 characters since release 4.0, because Perl (from release 5.004) does not.
3054 [Or at least, is documented not to: some releases seem to be in conflict
3055 with the documentation.] However, when a pattern was studied with
3056 pcre_study() and all its branches started with \s, PCRE still included VT
3057 as a possible starting character. Of course, this did no harm; it just
3058 caused an unnecessary match attempt.
3059
3060 22. Removed a now-redundant internal flag bit that recorded the fact that case
3061 dependency changed within the pattern. This was once needed for "required
3062 byte" processing, but is no longer used. This recovers a now-scarce options
3063 bit. Also moved the least significant internal flag bit to the most-
3064 significant bit of the word, which was not previously used (hangover from
3065 the days when it was an int rather than a uint) to free up another bit for
3066 the future.
3067
3068 23. Added support for CRLF line endings as well as CR and LF. As well as the
3069 default being selectable at build time, it can now be changed at runtime
3070 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3071 specify that it is scanning data with non-default line endings.
3072
3073 24. Changed the definition of CXXLINK to make it agree with the definition of
3074 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3075
3076 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3077 recursions. This makes a big different to stack usage for some patterns.
3078
3079 26. If a subpattern containing a named recursion or subroutine reference such
3080 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3081 the space required for the compiled pattern went wrong and gave too small a
3082 value. Depending on the environment, this could lead to "Failed: internal
3083 error: code overflow at offset 49" or "glibc detected double free or
3084 corruption" errors.
3085
3086 27. Applied patches from Google (a) to support the new newline modes and (b) to
3087 advance over multibyte UTF-8 characters in GlobalReplace.
3088
3089 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3090 difference for some implementation of PCRE in some Windows version.
3091
3092 29. Added some extra testing facilities to pcretest:
3093
3094 \q<number> in a data line sets the "match limit" value
3095 \Q<number> in a data line sets the "match recursion limt" value
3096 -S <number> sets the stack size, where <number> is in megabytes
3097
3098 The -S option isn't available for Windows.
3099
3100
3101 Version 6.6 06-Feb-06
3102 ---------------------
3103
3104 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3105 in pcreposix.h. I have copied the definition from pcre.h.
3106
3107 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3108 because pcre.h is no longer a built file.
3109
3110 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3111 not normally included in the compiled code.
3112
3113
3114 Version 6.5 01-Feb-06
3115 ---------------------
3116
3117 1. When using the partial match feature with pcre_dfa_exec(), it was not
3118 anchoring the second and subsequent partial matches at the new starting
3119 point. This could lead to incorrect results. For example, with the pattern
3120 /1234/, partially matching against "123" and then "a4" gave a match.
3121
3122 2. Changes to pcregrep:
3123
3124 (a) All non-match returns from pcre_exec() were being treated as failures
3125 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3126 error message is output. Some extra information is given for the
3127 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3128 probably the only errors that are likely to be caused by users (by
3129 specifying a regex that has nested indefinite repeats, for instance).
3130 If there are more than 20 of these errors, pcregrep is abandoned.
3131
3132 (b) A binary zero was treated as data while matching, but terminated the
3133 output line if it was written out. This has been fixed: binary zeroes
3134 are now no different to any other data bytes.
3135
3136 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3137 used to set a locale for matching. The --locale=xxxx long option has
3138 been added (no short equivalent) to specify a locale explicitly on the
3139 pcregrep command, overriding the environment variables.
3140
3141 (d) When -B was used with -n, some line numbers in the output were one less
3142 than they should have been.
3143
3144 (e) Added the -o (--only-matching) option.
3145
3146 (f) If -A or -C was used with -c (count only), some lines of context were
3147 accidentally printed for the final match.
3148
3149 (g) Added the -H (--with-filename) option.
3150
3151 (h) The combination of options -rh failed to suppress file names for files
3152 that were found from directory arguments.
3153
3154 (i) Added the -D (--devices) and -d (--directories) options.
3155
3156 (j) Added the -F (--fixed-strings) option.
3157
3158 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3159
3160 (l) Added the --colo(u)r option.
3161
3162 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3163 is not present by default.
3164
3165 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3166 items such as (?R) or (?1), when the recursion could match a number of
3167 alternatives. If it matched one of the alternatives, but subsequently,
3168 outside the recursion, there was a failure, the code tried to back up into
3169 the recursion. However, because of the way PCRE is implemented, this is not
3170 possible, and the result was an incorrect result from the match.
3171
3172 In order to prevent this happening, the specification of recursion has
3173 been changed so that all such subpatterns are automatically treated as
3174 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3175
3176 4. I had overlooked the fact that, in some locales, there are characters for
3177 which isalpha() is true but neither isupper() nor islower() are true. In
3178 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3179 and ordfeminine) are like this. This affected the treatment of \w and \W
3180 when they appeared in character classes, but not when they appeared outside
3181 a character class. The bit map for "word" characters is now created
3182 separately from the results of isalnum() instead of just taking it from the
3183 upper, lower, and digit maps. (Plus the underscore character, of course.)
3184
3185 5. The above bug also affected the handling of POSIX character classes such as
3186 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3187 permanent tables. Instead, the bit maps for such a class were previously
3188 created as the appropriate unions of the upper, lower, and digit bitmaps.
3189 Now they are created by subtraction from the [[:word:]] class, which has
3190 its own bitmap.
3191
3192 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3193 It is created by subtracting the vertical space characters (\x09, \x0a,
3194 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3195 subtraction was done in the overall bitmap for a character class, meaning
3196 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3197 be recognized. This bug has been fixed.
3198
3199 7. Patches from the folks at Google:
3200
3201 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3202 real life, but is still worth protecting against".
3203
3204 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3205 regular expressions".
3206
3207 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3208 have it.
3209
3210 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3211 "configure" and the latter not, in order to fix a problem somebody had
3212 with compiling the Arg class on HP-UX.
3213
3214 (e) Improve the error-handling of the C++ wrapper a little bit.
3215
3216 (f) New tests for checking recursion limiting.
3217
3218 8. The pcre_memmove() function, which is used only if the environment does not
3219 have a standard memmove() function (and is therefore rarely compiled),
3220 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3221 returning a result (though PCRE never actually uses the result).
3222
3223 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3224 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3225 returned instead of calling malloc() with an overflowing number that would
3226 most likely cause subsequent chaos.
3227
3228 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3229
3230 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3231 with this option is matched, the nmatch and pmatch options of regexec() are
3232 ignored.
3233
3234 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3235 provided in case anyone wants to the the POSIX interface with UTF-8
3236 strings.
3237
3238 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3239 C++ linking (needed for some HP-UX environments).
3240
3241 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3242 (unused parameter) and in the pcre_printint() function (omitted "default"
3243 switch label when the default is to do nothing).
3244
3245 15. Added some code to make it possible, when PCRE is compiled as a C++
3246 library, to replace subject pointers for pcre_exec() with a smart pointer
3247 class, thus making it possible to process discontinuous strings.
3248
3249 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3250 much the same function. They were added by different people who were trying
3251 to make PCRE easy to compile on non-Unix systems. It has been suggested
3252 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3253 for compiling on Windows systems. I have therefore replaced it with
3254 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3255 defaults to "extern" for C or "extern C" for C++, which works fine on
3256 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3257 SCOPE with something explicit in config.h. In addition:
3258
3259 (a) pcreposix.h still had just "extern" instead of either of these macros;
3260 I have replaced it with PCRE_DATA_SCOPE.
3261
3262 (b) Functions such as _pcre_xclass(), which are internal to the library,
3263 but external in the C sense, all had PCRE_EXPORT in their definitions.
3264 This is apparently wrong for the Windows case, so I have removed it.
3265 (It makes no difference on Unix-like systems.)
3266
3267 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3268 of recursive calls to match(). This is different to MATCH_LIMIT because
3269 that limits the total number of calls to match(), not all of which increase
3270 the depth of recursion. Limiting the recursion depth limits the amount of
3271 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3272 when PCRE is compiled, and changed at run time. A patch from Google adds
3273 this functionality to the C++ interface.
3274
3275 18. Changes to the handling of Unicode character properties:
3276
3277 (a) Updated the table to Unicode 4.1.0.
3278
3279 (b) Recognize characters that are not in the table as "Cn" (undefined).
3280
3281 (c) I revised the way the table is implemented to a much improved format
3282 which includes recognition of ranges. It now supports the ranges that
3283 are defined in UnicodeData.txt, and it also amalgamates other
3284 characters into ranges. This has reduced the number of entries in the
3285 table from around 16,000 to around 3,000, thus reducing its size
3286 considerably. I realized I did not need to use a tree structure after
3287 all - a binary chop search is just as efficient. Having reduced the
3288 number of entries, I extended their size from 6 bytes to 8 bytes to
3289 allow for more data.
3290
3291 (d) Added support for Unicode script names via properties such as \p{Han}.
3292
3293 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3294 matching that character.
3295
3296 20. When matching a repeated Unicode property with a minimum greater than zero,
3297 (for example \pL{2,}), PCRE could look past the end of the subject if it
3298 reached it while seeking the minimum number of characters. This could
3299 happen only if some of the characters were more than one byte long, because
3300 there is a check for at least the minimum number of bytes.
3301
3302 21. Refactored the implementation of \p and \P so as to be more general, to
3303 allow for more different types of property in future. This has changed the
3304 compiled form incompatibly. Anybody with saved compiled patterns that use
3305 \p or \P will have to recompile them.
3306
3307 22. Added "Any" and "L&" to the supported property types.
3308
3309 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3310 but give a compile time error if the value is greater than 0xff.
3311
3312 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3313 accidentally not being installed or uninstalled.
3314
3315 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3316 made were to insert the current release number. This seemed silly, because
3317 it made things harder for people building PCRE on systems that don't run
3318 "configure". I have turned pcre.h into a distributed file, no longer built
3319 by "configure", with the version identification directly included. There is
3320 no longer a pcre.h.in file.
3321
3322 However, this change necessitated a change to the pcre-config script as
3323 well. It is built from pcre-config.in, and one of the substitutions was the
3324 release number. I have updated configure.ac so that ./configure now finds
3325 the release number by grepping pcre.h.
3326
3327 26. Added the ability to run the tests under valgrind.
3328
3329
3330 Version 6.4 05-Sep-05
3331 ---------------------
3332
3333 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3334 "--" to be printed when multiple files were scanned, even when none of the
3335 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3336 consider it to be a bug, and have restored the previous behaviour.
3337
3338 2. A couple of code tidies to get rid of compiler warnings.
3339
3340 3. The pcretest program used to cheat by referring to symbols in the library
3341 whose names begin with _pcre_. These are internal symbols that are not
3342 really supposed to be visible externally, and in some environments it is
3343 possible to suppress them. The cheating is now confined to including
3344 certain files from the library's source, which is a bit cleaner.
3345
3346 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3347 file's purpose clearer.
3348
3349 5. Reorganized pcre_ucp_findchar().
3350
3351
3352 Version 6.3 15-Aug-05
3353 ---------------------
3354
3355 1. The file libpcre.pc.in did not have general read permission in the tarball.
3356
3357 2. There were some problems when building without C++ support:
3358
3359 (a) If C++ support was not built, "make install" and "make test" still
3360 tried to test it.
3361
3362 (b) There were problems when the value of CXX was explicitly set. Some
3363 changes have been made to try to fix these, and ...
3364
3365 (c) --disable-cpp can now be used to explicitly disable C++ support.
3366
3367 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3368 backslash in a target when C++ was disabled. This confuses some
3369 versions of "make", apparently. Using an intermediate variable solves
3370 this. (Same for CPP_LOBJ.)
3371
3372 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3373 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3374 necessary on certain architectures.
3375
3376 4. Added a setting of -export-symbols-regex to the link command to remove
3377 those symbols that are exported in the C sense, but actually are local
3378 within the library, and not documented. Their names all begin with
3379 "_pcre_". This is not a perfect job, because (a) we have to except some
3380 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3381 available (and never for static libraries). I have made a note to try to
3382 find a way round (a) in the future.
3383
3384
3385 Version 6.2 01-Aug-05
3386 ---------------------
3387
3388 1. There was no test for integer overflow of quantifier values. A construction
3389 such as {1111111111111111} would give undefined results. What is worse, if
3390 a minimum quantifier for a parenthesized subpattern overflowed and became
3391 negative, the calculation of the memory size went wrong. This could have
3392 led to memory overwriting.
3393
3394 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3395
3396 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3397 operating environments where this matters.
3398
3399 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3400 PCRE options from within the C++ wrapper.
3401
3402 5. Named capturing subpatterns were not being correctly counted when a pattern
3403 was compiled. This caused two problems: (a) If there were more than 100
3404 such subpatterns, the calculation of the memory needed for the whole
3405 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3406 back references of the form \12, where the number was greater than 9, were
3407 not recognized as back references, even though there were sufficient
3408 previous subpatterns.
3409
3410 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3411 versions of gcc, e.g. 2.95.4.
3412
3413
3414 Version 6.1 21-Jun-05
3415 ---------------------
3416
3417 1. There was one reference to the variable "posix" in pcretest.c that was not
3418 surrounded by "#if !defined NOPOSIX".
3419
3420 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3421 the cross-check on the old pcre_info() function, for the benefit of the
3422 cut-down version of PCRE that is currently imported into Exim.
3423
3424 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3425 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3426 patterns that start (?i) but I don't think that matters. The use of (?i) is
3427 just an example; this all applies to the other options as well.
3428
3429 4. Since libtool seems to echo the compile commands it is issuing, the output
3430 from "make" can be reduced a bit by putting "@" in front of each libtool
3431 compile command.
3432
3433 5. Patch from the folks at Google for configure.in to be a bit more thorough
3434 in checking for a suitable C++ installation before trying to compile the
3435 C++ stuff. This should fix a reported problem when a compiler was present,
3436 but no suitable headers.
3437
3438 6. The man pages all had just "PCRE" as their title. I have changed them to
3439 be the relevant file name. I have also arranged that these names are
3440 retained in the file doc/pcre.txt, which is a concatenation in text format
3441 of all the man pages except the little individual ones for each function.
3442
3443 7. The NON-UNIX-USE file had not been updated for the different set of source
3444 files that come with release 6. I also added a few comments about the C++
3445 wrapper.
3446
3447
3448 Version 6.0 07-Jun-05
3449 ---------------------
3450
3451 1. Some minor internal re-organization to help with my DFA experiments.
3452
3453 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3454 didn't matter for the library itself when fully configured, but did matter
3455 when compiling without UCP support, or within Exim, where the ucp files are
3456 not imported.
3457
3458 3. Refactoring of the library code to split up the various functions into
3459 different source modules. The addition of the new DFA matching code (see
3460 below) to a single monolithic source would have made it really too
3461 unwieldy, quite apart from causing all the code to be include in a
3462 statically linked application, when only some functions are used. This is
3463 relevant even without the DFA addition now that patterns can be compiled in
3464 one application and matched in another.
3465
3466 The downside of splitting up is that there have to be some external
3467 functions and data tables that are used internally in different modules of
3468 the library but which are not part of the API. These have all had their
3469 names changed to start with "_pcre_" so that they are unlikely to clash
3470 with other external names.
3471
3472 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3473 a different (DFA) algorithm. Although it is slower than the original
3474 function, it does have some advantages for certain types of matching
3475 problem.
3476
3477 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3478 including restarting after a partial match.
3479
3480 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3481 defined when compiling for Windows was sent to me. I have put it into the
3482 code, though I have no means of testing or verifying it.
3483
3484 7. Added the pcre_refcount() auxiliary function.
3485
3486 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3487 match before or at the first newline in the subject string. In pcretest,
3488 the /f option on a pattern can be used to set this.
3489
3490 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3491 would behave wrongly. This has been present in PCRE since release 4.0.
3492
3493 10. A number of changes to the pcregrep command:
3494
3495 (a) Refactored how -x works; insert ^(...)$ instead of setting
3496 PCRE_ANCHORED and checking the length, in preparation for adding
3497 something similar for -w.
3498
3499 (b) Added the -w (match as a word) option.
3500
3501 (c) Refactored the way lines are read and buffered so as to have more
3502 than one at a time available.
3503
3504 (d) Implemented a pcregrep test script.
3505
3506 (e) Added the -M (multiline match) option. This allows patterns to match
3507 over several lines of the subject. The buffering ensures that at least
3508 8K, or the rest of the document (whichever is the shorter) is available
3509 for matching (and similarly the previous 8K for lookbehind assertions).
3510
3511 (f) Changed the --help output so that it now says
3512
3513 -w, --word-regex(p)
3514
3515 instead of two lines, one with "regex" and the other with "regexp"
3516 because that confused at least one person since the short forms are the
3517 same. (This required a bit of code, as the output is generated
3518 automatically from a table. It wasn't just a text change.)
3519
3520 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3521 option but starts with a hyphen. Could be a pattern or a path name
3522 starting with a hyphen, for instance.
3523
3524 (h) "-" can be given as a file name to represent stdin.
3525
3526 (i) When file names are being printed, "(standard input)" is used for
3527 the standard input, for compatibility with GNU grep. Previously
3528 "<stdin>" was used.
3529
3530 (j) The option --label=xxx can be used to supply a name to be used for
3531 stdin when file names are being printed. There is no short form.
3532
3533 (k) Re-factored the options decoding logic because we are going to add
3534 two more options that take data. Such options can now be given in four
3535 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3536
3537 (l) Added the -A, -B, and -C options for requesting that lines of context
3538 around matches be printed.
3539
3540 (m) Added the -L option to print the names of files that do not contain
3541 any matching lines, that is, the complement of -l.
3542
3543 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3544 continue to scan other files.
3545
3546 (o) The -s option was incorrectly implemented. For compatibility with other
3547 greps, it now suppresses the error message for a non-existent or non-
3548 accessible file (but not the return code). There is a new option called
3549 -q that suppresses the output of matching lines, which was what -s was
3550 previously doing.
3551
3552 (p) Added --include and --exclude options to specify files for inclusion
3553 and exclusion when recursing.
3554
3555 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3556 Hopefully, it now does.
3557
3558 12. Missing cast in pcre_study().
3559
3560 13. Added an "uninstall" target to the makefile.
3561
3562 14. Replaced "extern" in the function prototypes in Makefile.in with
3563 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3564 world, but is set differently for Windows.
3565
3566 15. Added a second compiling function called pcre_compile2(). The only
3567 difference is that it has an extra argument, which is a pointer to an
3568 integer error code. When there is a compile-time failure, this is set
3569 non-zero, in addition to the error test pointer being set to point to an
3570 error message. The new argument may be NULL if no error number is required
3571 (but then you may as well call pcre_compile(), which is now just a
3572 wrapper). This facility is provided because some applications need a
3573 numeric error indication, but it has also enabled me to tidy up the way
3574 compile-time errors are handled in the POSIX wrapper.
3575
3576 16. Added VPATH=.libs to the makefile; this should help when building with one
3577 prefix path and installing with another. (Or so I'm told by someone who
3578 knows more about this stuff than I do.)
3579
3580 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3581 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3582 match everything, including newlines. This is not POSIX-compatible, but
3583 somebody wanted the feature. From pcretest it can be activated by using
3584 both the P and the s flags.
3585
3586 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3587
3588 19. libpcre.pc was being incorrectly installed as executable.
3589
3590 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3591 it now also looks for '\r' so that it will work unmodified on Windows.
3592
3593 21. Added Google's contributed C++ wrapper to the distribution.
3594
3595 22. Added some untidy missing memory free() calls in pcretest, to keep
3596 Electric Fence happy when testing.
3597
3598
3599
3600 Version 5.0 13-Sep-04
3601 ---------------------
3602
3603 1. Internal change: literal characters are no longer packed up into items
3604 containing multiple characters in a single byte-string. Each character
3605 is now matched using a separate opcode. However, there may be more than one
3606 byte in the character in UTF-8 mode.
3607
3608 2. The pcre_callout_block structure has two new fields: pattern_position and
3609 next_item_length. These contain the offset in the pattern to the next match
3610 item, and its length, respectively.
3611
3612 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3613 insertion of callouts before each pattern item. Added the /C option to
3614 pcretest to make use of this.
3615
3616 4. On the advice of a Windows user, the lines
3617
3618 #if defined(_WIN32) || defined(WIN32)
3619 _setmode( _fileno( stdout ), 0x8000 );
3620 #endif /* defined(_WIN32) || defined(WIN32) */
3621
3622 have been added to the source of pcretest. This apparently does useful
3623 magic in relation to line terminators.
3624
3625 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3626 for the benefit of those environments where the "b" makes a difference.
3627
3628 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3629 to know about it. I have put a hack into configure.in that adds in code
3630 to set GCC=yes if CC=icc. This seems to end up at a point in the
3631 generated configure script that is early enough to affect the setting of
3632 compiler options, which is what is needed, but I have no means of testing
3633 whether it really works. (The user who reported this had patched the
3634 generated configure script, which of course I cannot do.)
3635
3636 LATER: After change 22 below (new libtool files), the configure script
3637 seems to know about icc (and also ecc). Therefore, I have commented out
3638 this hack in configure.in.
3639
3640 7. Added support for pkg-config (2 patches were sent in).
3641
3642 8. Negated POSIX character classes that used a combination of internal tables
3643 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3644 [[:^ascii]]. Typically, they would match almost any characters. The other
3645 POSIX classes were not broken in this way.
3646
3647 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3648 to find the match, as PCRE was deluded into thinking that the match had to
3649 start at the start point or following a newline. The same bug applied to
3650 patterns with negative forward assertions or any backward assertions
3651 preceding ".*" at the start, unless the pattern required a fixed first
3652 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3653
3654 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3655 starting at the last subject character, bytes beyond the end of the subject
3656 string were read.
3657
3658 11. Renamed the variable "class" as "classbits" to make life easier for C++
3659 users. (Previously there was a macro definition, but it apparently wasn't
3660 enough.)
3661
3662 12. Added the new field "tables" to the extra data so that tables can be passed
3663 in at exec time, or the internal tables can be re-selected. This allows
3664 a compiled regex to be saved and re-used at a later time by a different
3665 program that might have everything at different addresses.
3666
3667 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3668 -R library as well as a -L library.
3669
3670 14. The debugging options of pcretest (-d on the command line or D on a
3671 pattern) showed incorrect output for anything following an extended class
3672 that contained multibyte characters and which was followed by a quantifier.
3673
3674 15. Added optional support for general category Unicode character properties
3675 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3676 support. It adds about 90K to the size of the library. The meanings of the
3677 inbuilt class escapes such as \d and \s have NOT been changed.
3678
3679 16. Updated pcredemo.c to include calls to free() to release the memory for the
3680 compiled pattern.
3681
3682 17. The generated file chartables.c was being created in the source directory
3683 instead of in the building directory. This caused the build to fail if the
3684 source directory was different from the building directory, and was
3685 read-only.
3686
3687 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3688 file. No doubt somebody will tell me if they don't make sense... Also added
3689 Dan Mooney's comments about building on OpenVMS.
3690
3691 19. Added support for partial matching via the PCRE_PARTIAL option for
3692 pcre_exec() and the \P data escape in pcretest.
3693
3694 20. Extended pcretest with 3 new pattern features:
3695
3696 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3697 write the compiled pattern to the file whose name is "rest-of-line".
3698 This is a straight binary dump of the data, with the saved pointer to
3699 the character tables forced to be NULL. The study data, if any, is
3700 written too. After writing, pcretest reads a new pattern.
3701
3702 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3703 compiled pattern from the given file. There must not be any
3704 occurrences of "<" in the file name (pretty unlikely); if there are,
3705 pcretest will instead treat the initial "<" as a pattern delimiter.
3706 After reading in the pattern, pcretest goes on to read data lines as
3707 usual.
3708
3709 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3710 and 16-bit fields in a compiled pattern, to simulate a pattern that
3711 was compiled on a host of opposite endianness.
3712
3713 21. The pcre-exec() function can now cope with patterns that were compiled on
3714 hosts of opposite endianness, with this restriction:
3715
3716 As for any compiled expression that is saved and used later, the tables
3717 pointer field cannot be preserved; the extra_data field in the arguments
3718 to pcre_exec() should be used to pass in a tables address if a value
3719 other than the default internal tables were used at compile time.
3720
3721 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3722 now diagnosed as an error. Previously, most of the time, a negative number
3723 would have been treated as zero, but if in addition "ovector" was passed as
3724 NULL, a crash could occur.
3725
3726 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3727 new versions from the libtool 1.5 distribution (the last one is a copy of
3728 a file called libtool.m4). This seems to have fixed the need to patch
3729 "configure" to support Darwin 1.3 (which I used to do). However, I still
3730 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3731 workstation).
3732
3733 24. Changed the PCRE licence to be the more standard "BSD" licence.
3734
3735
3736 Version 4.5 01-Dec-03
3737 ---------------------
3738
3739 1. There has been some re-arrangement of the code for the match() function so
3740 that it can be compiled in a version that does not call itself recursively.
3741 Instead, it keeps those local variables that need separate instances for
3742 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3743 needs to "recurse". Keeping track of where control must go is done by means
3744 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3745 hide most of the details from the main code, and operates only if
3746 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3747 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3748 operating.
3749
3750 To make it easier for callers to provide specially tailored get/free
3751 functions for this usage, two new functions, pcre_stack_malloc, and
3752 pcre_stack_free, are used. They are always called in strict stacking order,
3753 and the size of block requested is always the same.
3754
3755 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3756 PCRE has been compiled to use the stack or the heap for recursion. The
3757 -C option of pcretest uses this to show which version is compiled.
3758
3759 A new data escape \S, is added to pcretest; it causes the amounts of store
3760 obtained and freed by both kinds of malloc/free at match time to be added
3761 to the output.
3762
3763 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3764 what's available on my current Linux desktop machine.
3765
3766 3. When matching a UTF-8 string, the test for a valid string at the start has
3767 been extended. If start_offset is not zero, PCRE now checks that it points
3768 to a byte that is the start of a UTF-8 character. If not, it returns
3769 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3770 this is necessary because there may be backward assertions in the pattern.
3771 When matching the same subject several times, it may save resources to use
3772 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3773
3774 4. The code for checking the validity of UTF-8 strings has been tightened so
3775 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3776 containing "overlong sequences".
3777
3778 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3779 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3780 should have been "&", but it just so happened that all the cases this let
3781 through by mistake were picked up later in the function.
3782
3783 6. I had used a variable called "isblank" - this is a C99 function, causing
3784 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3785
3786 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3787 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3788 so the version is shown; (c) stop "make test" echoing "./RunTest".
3789
3790 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3791
3792 9. The prototype for memmove() for systems that don't have it was using
3793 size_t, but the inclusion of the header that defines size_t was later. I've
3794 moved the #includes for the C headers earlier to avoid this.
3795
3796 10. Added some adjustments to the code to make it easier to compiler on certain
3797 special systems:
3798
3799 (a) Some "const" qualifiers were missing.
3800 (b) Added the macro EXPORT before all exported functions; by default this
3801 is defined to be empty.
3802 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3803 that it reads its output file name as an argument instead of writing
3804 to the standard output and assuming this can be redirected.
3805
3806 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3807 class containing characters with values greater than 255, PCRE compilation
3808 went into a loop.
3809
3810 12. A recursive reference to a subpattern that was within another subpattern
3811 that had a minimum quantifier of zero caused PCRE to crash. For example,
3812 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3813 recursion. If the recursively-called subpattern itself had a zero repeat,
3814 that was OK.
3815
3816 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3817 buffer into which it was copied (for escape processing) was still set at
3818 1024, so long lines caused crashes.
3819
3820 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3821 "internal error: code overflow...". This applied to any character class
3822 that was followed by a possessive quantifier.
3823
3824 15. Modified the Makefile to add libpcre.la as a prerequisite for
3825 libpcreposix.la because I was told this is needed for a parallel build to
3826 work.
3827
3828 16. If a pattern that contained .* following optional items at the start was
3829 studied, the wrong optimizing data was generated, leading to matching
3830 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3831 matching string must start with a or b or c. The correct conclusion for
3832 this pattern is that a match can start with any character.
3833
3834
3835 Version 4.4 13-Aug-03
3836 ---------------------
3837
3838 1. In UTF-8 mode, a character class containing characters with values between
3839 127 and 255 was not handled correctly if the compiled pattern was studied.
3840 In fixing this, I have also improved the studying algorithm for such
3841 classes (slightly).
3842
3843 2. Three internal functions had redundant arguments passed to them. Removal
3844 might give a very teeny performance improvement.
3845
3846 3. Documentation bug: the value of the capture_top field in a callout is *one
3847 more than* the number of the hightest numbered captured substring.
3848
3849 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3850 in incorrectly linking with a previously installed version. They now link
3851 explicitly with libpcre.la.
3852
3853 5. configure.in no longer needs to recognize Cygwin specially.
3854
3855 6. A problem in pcre.in for Windows platforms is fixed.
3856
3857 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3858 pcretest, it used to include the size of the study block as part of its
3859 output. Unfortunately, the structure contains a field that has a different
3860 size on different hardware architectures. This meant that the tests that
3861 showed this size failed. As the block is currently always of a fixed size,
3862 this information isn't actually particularly useful in pcretest output, so
3863 I have just removed it.
3864
3865 8. Three pre-processor statements accidentally did not start in column 1.
3866 Sadly, there are *still* compilers around that complain, even though
3867 standard C has not required this for well over a decade. Sigh.
3868
3869 9. In pcretest, the code for checking callouts passed small integers in the
3870 callout_data field, which is a void * field. However, some picky compilers
3871 complained about the casts involved for this on 64-bit systems. Now
3872 pcretest passes the address of the small integer instead, which should get
3873 rid of the warnings.
3874
3875 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3876 both compile and run time, and gives an error if an invalid UTF-8 sequence
3877 is found. There is a option for disabling this check in cases where the
3878 string is known to be correct and/or the maximum performance is wanted.
3879
3880 11. In response to a bug report, I changed one line in Makefile.in from
3881
3882 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3883 to
3884 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3885
3886 to look similar to other lines, but I have no way of telling whether this
3887 is the right thing to do, as I do not use Windows. No doubt I'll get told
3888 if it's wrong...
3889
3890
3891 Version 4.3 21-May-03
3892 ---------------------
3893
3894 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3895 Makefile.
3896
3897 2. Some refactoring to improve the quality of the code:
3898
3899 (i) The utf8_table... variables are now declared "const".
3900
3901 (ii) The code for \cx, which used the "case flipping" table to upper case
3902 lower case letters, now just substracts 32. This is ASCII-specific,
3903 but the whole concept of \cx is ASCII-specific, so it seems
3904 reasonable.
3905
3906 (iii) PCRE was using its character types table to recognize decimal and
3907 hexadecimal digits in the pattern. This is silly, because it handles
3908 only 0-9, a-f, and A-F, but the character types table is locale-
3909 specific, which means strange things might happen. A private
3910 table is now used for this - though it costs 256 bytes, a table is
3911 much faster than multiple explicit tests. Of course, the standard
3912 character types table is still used for matching digits in subject
3913 strings against \d.
3914
3915 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3916 ending in _t are). So I've renamed it as ESC_tee.
3917
3918 3. The first argument for regexec() in the POSIX wrapper should have been
3919 defined as "const".
3920
3921 4. Changed pcretest to use malloc() for its buffers so that they can be
3922 Electric Fenced for debugging.
3923
3924 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3925 to read one or more bytes before the start of the subject string. Often this
3926 had no effect on PCRE's behaviour, but in some circumstances it could
3927 provoke a segmentation fault.
3928
3929 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3930 to try to read one or more bytes before the start of the subject string.
3931
3932 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3933 UTF-8 support could misbehave in various ways if the subject string
3934 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3935 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3936 back over UTF-8 characters.)
3937
3938
3939 Version 4.2 14-Apr-03
3940 ---------------------
3941
3942 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3943
3944 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3945 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3946 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3947 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3948 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3949 and BUILD_EXEEXT
3950 Note: automatic setting of the BUILD variables is not yet working
3951 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3952 compile-time but not at link-time
3953 [LINK]: use for linking executables only
3954 make different versions for Windows and non-Windows
3955 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3956 libraries
3957 [LINK_FOR_BUILD]: new variable
3958 [OBJEXT]: use throughout
3959 [EXEEXT]: use throughout
3960 <winshared>: new target
3961 <wininstall>: new target
3962 <dftables.o>: use native compiler
3963 <dftables>: use native linker
3964 <install>: handle Windows platform correctly
3965 <clean>: ditto
3966 <check>: ditto
3967 copy DLL to top builddir before testing
3968
3969 As part of these changes, -no-undefined was removed again. This was reported
3970 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3971 in any case.
3972
3973 3. Some tidies to get rid of compiler warnings:
3974
3975 . In the match_data structure, match_limit was an unsigned long int, whereas
3976 match_call_count was an int. I've made them both unsigned long ints.
3977
3978 . In pcretest the fact that a const uschar * doesn't automatically cast to
3979 a void * provoked a warning.
3980
3981 . Turning on some more compiler warnings threw up some "shadow" variables
3982 and a few more missing casts.
3983
3984 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3985 option, a class that contained a single character with a value between 128
3986 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3987
3988 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3989 option, a class that contained several characters, but with at least one
3990 whose value was between 128 and 255 caused PCRE to crash.
3991
3992
3993 Version 4.1 12-Mar-03
3994 ---------------------
3995
3996 1. Compiling with gcc -pedantic found a couple of places where casts were
3997 needed, and a string in dftables.c that was longer than standard compilers are
3998 required to support.
3999
4000 2. Compiling with Sun's compiler found a few more places where the code could
4001 be tidied up in order to avoid warnings.
4002
4003 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4004 first of these names is deprecated in the latest Autoconf in favour of the name
4005 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4006 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4007 analogy I have changed it to CFLAGS_FOR_BUILD.
4008
4009 4. Added -no-undefined to the linking command in the Makefile, because this is
4010 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4011 linking step for the pcreposix library.
4012
4013 5. PCRE was failing to diagnose the case of two named groups with the same
4014 name.
4015
4016 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4017 literal character that is needed in the subject for a match, and scans along to
4018 ensure that it is present before embarking on the full matching process. This
4019 saves time in cases of nested unlimited repeats that are never going to match.
4020 Problem: the scan can take a lot of time if the subject is very long (e.g.
4021 megabytes), thus penalizing straightforward matches. It is now done only if the
4022 amount of subject to be scanned is less than 1000 bytes.
4023
4024 7. A lesser problem with the same optimization is that it was recording the
4025 first character of an anchored pattern as "needed", thus provoking a search
4026 right along the subject, even when the first match of the pattern was going to
4027 fail. The "needed" character is now not set for anchored patterns, unless it
4028 follows something in the pattern that is of non-fixed length. Thus, it still
4029 fulfils its original purpose of finding quick non-matches in cases of nested
4030 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4031
4032
4033 Version 4.0 17-Feb-03
4034 ---------------------
4035
4036 1. If a comment in an extended regex that started immediately after a meta-item
4037 extended to the end of string, PCRE compiled incorrect data. This could lead to
4038 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4039
4040 2. Moved to autoconf 2.53 and libtool 1.4.2.
4041
4042 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4043 the special perltest8 script is no longer needed - all the tests can be run
4044 from a single perltest script.
4045
4046 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4047 by \s. It has now been removed in PCRE. This means it isn't recognized as
4048 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4049 class [:space:] *does* include VT, thereby creating a mess.
4050
4051 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4052 space and tab.
4053
4054 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4055 its new features into the main test script, reducing the number of scripts.
4056
4057 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4058 were backward compatible, and made the (?i) apply to the whole pattern, as if
4059 /i were given. Now it behaves more logically, and applies the option setting
4060 only to what follows. PCRE has been changed to follow suit. However, if it
4061 finds options settings right at the start of the pattern, it extracts them into
4062 the global options, as before. Thus, they show up in the info data.
4063
4064 8. Added support for the \Q...\E escape sequence. Characters in between are
4065 treated as literals. This is slightly different from Perl in that $ and @ are
4066 also handled as literals inside the quotes. In Perl, they will cause variable
4067 interpolation. Note the following examples:
4068
4069 Pattern PCRE matches Perl matches
4070
4071 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4072 \Qabc\$xyz\E abc\$xyz abc\$xyz
4073 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4074
4075 For compatibility with Perl, \Q...\E sequences are recognized inside character
4076 classes as well as outside them.
4077
4078 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4079 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4080 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4081 signed/unsigned warnings.
4082
4083 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4084 option for pcretest, so I've replaced it by a simple function that does just
4085 that job.
4086
4087 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4088 "pcregrep -".
4089
4090 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4091 Java package. This provides some syntactic sugar for simple cases of what my
4092 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4093 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4094 item, you can use this simplified notation. Note that only makes sense with
4095 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4096 greediness, whatever the setting of the PCRE_UNGREEDY option.
4097
4098 13. A change of greediness default within a pattern was not taking effect at
4099 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4100 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4101 was abstracted outside.
4102
4103 14. PCRE now supports the \G assertion. It is true when the current matching
4104 position is at the start point of the match. This differs from \A when the
4105 starting offset is non-zero. Used with the /g option of pcretest (or similar
4106 code), it works in the same way as it does for Perl's /g option. If all
4107 alternatives of a regex begin with \G, the expression is anchored to the start
4108 match position, and the "anchored" flag is set in the compiled expression.
4109
4110 15. Some bugs concerning the handling of certain option changes within patterns
4111 have been fixed. These applied to options other than (?ims). For example,
4112 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4113 been the other way round. Some of this was related to change 7 above.
4114
4115 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4116 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4117 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4118 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4119
4120 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4121 mode. Unlike ".", it always matches newline, whatever the setting of
4122 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4123 assertions. Perl allows it, but it doesn't (in general) work because it can't
4124 calculate the length of the lookbehind. At least, that's the case for Perl
4125 5.8.0 - I've been told they are going to document that it doesn't work in
4126 future.
4127
4128 18. Added an error diagnosis for escapes that PCRE does not support: these are
4129 \L, \l, \N, \P, \p, \U, \u, and \X.
4130
4131 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4132 reading past the end of the pattern in cases such as /[abcd/.
4133
4134 20. PCRE was getting more memory than necessary for patterns with classes that
4135 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4136
4137 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4138 compiling PCRE for use with Virtual Pascal.
4139
4140 22. Small fix to the Makefile to make it work properly if the build is done
4141 outside the source tree.
4142
4143 23. Added a new extension: a condition to go with recursion. If a conditional
4144 subpattern starts with (?(R) the "true" branch is used if recursion has
4145 happened, whereas the "false" branch is used only at the top level.
4146
4147 24. When there was a very long string of literal characters (over 255 bytes
4148 without UTF support, over 250 bytes with UTF support), the computation of how
4149 much memory was required could be incorrect, leading to segfaults or other
4150 strange effects.
4151
4152 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4153 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4154 there was a subsequent back reference to those brackets. This meant that, for
4155 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4156 possible to check for precisely this case. All we can do is abandon the
4157 optimization if .* occurs inside capturing brackets when there are any back
4158 references whatsoever. (See below for a better fix that came later.)
4159
4160 26. The handling of the optimization for finding the first character of a
4161 non-anchored pattern, and for finding a character that is required later in the
4162 match were failing in some cases. This didn't break the matching; it just
4163 failed to optimize when it could. The way this is done has been re-implemented.
4164
4165 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4166
4167 28. Added a new feature that provides some of the functionality that Perl
4168 provides with (?{...}). The facility is termed a "callout". The way it is done
4169 in PCRE is for the caller to provide an optional function, by setting
4170 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4171 global variable. By default it is unset, which disables all calling out. To get
4172 the function called, the regex must include (?C) at appropriate points. This
4173 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4174 This provides a means of identifying different callout points. When PCRE
4175 reaches such a point in the regex, if pcre_callout has been set, the external
4176 function is called. It is provided with data in a structure called
4177 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4178 matching continues; if it returns a non-zero value, the match at the current
4179 point fails. However, backtracking will occur if possible. [This was changed
4180 later and other features added - see item 49 below.]
4181
4182 29. pcretest is upgraded to test the callout functionality. It provides a
4183 callout function that displays information. By default, it shows the start of
4184 the match and the current position in the text. There are some new data escapes
4185 to vary what happens:
4186
4187 \C+ in addition, show current contents of captured substrings
4188 \C- do not supply a callout function
4189 \C!n return 1 when callout number n is reached
4190 \C!n!m return 1 when callout number n is reached for the mth time
4191
4192 30. If pcregrep was called with the -l option and just a single file name, it
4193 output "<stdin>" if a match was found, instead of the file name.
4194
4195 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4196 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4197 pcre_exec(). This saves a malloc/free per call. The default value of
4198 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4199 when configuring.
4200
4201 32. The default maximum size of a compiled pattern is 64K. There have been a
4202 few cases of people hitting this limit. The code now uses macros to handle the
4203 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4204 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4205 configuring. Tests 2 and 5 work only with 2-byte links because they output
4206 debugging information about compiled patterns.
4207
4208 33. Internal code re-arrangements:
4209
4210 (a) Moved the debugging function for printing out a compiled regex into
4211 its own source file (printint.c) and used #include to pull it into
4212 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4213 separate copies.
4214
4215 (b) Defined the list of op-code names for debugging as a macro in
4216 internal.h so that it is next to the definition of the opcodes.
4217
4218 (c) Defined a table of op-code lengths for simpler skipping along compiled
4219 code. This is again a macro in internal.h so that it is next to the
4220 definition of the opcodes.
4221
4222 34. Added support for recursive calls to individual subpatterns, along the
4223 lines of Robin Houston's patch (but implemented somewhat differently).
4224
4225 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4226 allow it to read and process whole directories in Win32. This code was
4227 contributed by Lionel Fourquaux; it has not been tested by me.
4228
4229 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4230 used to name a group. Names consist of alphanumerics and underscores, and must
4231 be unique. Back references use the syntax (?P=name) and recursive calls use
4232 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4233 numbers. The function pcre_fullinfo() can be used after compilation to extract
4234 a name/number map. There are three relevant calls:
4235
4236 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4237 PCRE_INFO_NAMECOUNT yields the number of entries
4238 PCRE_INFO_NAMETABLE yields a pointer to the map.
4239
4240 The map is a vector of fixed-size entries. The size of each entry depends on
4241 the length of the longest name used. The first two bytes of each entry are the
4242 group number, most significant byte first. There follows the corresponding
4243 name, zero terminated. The names are in alphabetical order.
4244
4245 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4246 case instead of 255. Making it the same both with and without UTF-8 support
4247 means that the same test output works with both.
4248
4249 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4250 calling malloc() with a zero argument.
4251
4252 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4253 optimization. I've improved things by keeping a bitmap of backreferences with
4254 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4255 fact referenced, the optimization can be applied. It is unlikely that a
4256 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4257 the match to follow \n) will appear inside brackets with a number greater than
4258 31, but if it does, any back reference > 31 suppresses the optimization.
4259
4260 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4261 of disabling numbered capturing parentheses. Any opening parenthesis that is
4262 not followed by ? behaves as if it were followed by ?: but named parentheses
4263 can still be used for capturing (and they will acquire numbers in the usual
4264 way).
4265
4266 41. Redesigned the return codes from the match() function into yes/no/error so
4267 that errors can be passed back from deep inside the nested calls. A malloc
4268 failure while inside a recursive subpattern call now causes the
4269 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4270
4271 42. It is now possible to set a limit on the number of times the match()
4272 function is called in a call to pcre_exec(). This facility makes it possible to
4273 limit the amount of recursion and backtracking, though not in a directly
4274 obvious way, because the match() function is used in a number of different
4275 circumstances. The count starts from zero for each position in the subject
4276 string (for non-anchored patterns). The default limit is, for compatibility, a
4277 large number, namely 10 000 000. You can change this in two ways:
4278
4279 (a) When configuring PCRE before making, you can use --with-match-limit=n
4280 to set a default value for the compiled library.
4281
4282 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4283 a different value is set. See 45 below.
4284
4285 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4286
4287 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4288 of things that can be changed at compile time. The first argument specifies
4289 what is wanted and the second points to where the information is to be placed.
4290 The current list of available information is:
4291
4292 PCRE_CONFIG_UTF8
4293
4294 The output is an integer that is set to one if UTF-8 support is available;
4295 otherwise it is set to zero.
4296
4297 PCRE_CONFIG_NEWLINE
4298
4299 The output is an integer that it set to the value of the code that is used for
4300 newline. It is either LF (10) or CR (13).
4301
4302 PCRE_CONFIG_LINK_SIZE
4303
4304 The output is an integer that contains the number of bytes used for internal
4305 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4306
4307 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4308
4309 The output is an integer that contains the threshold above which the POSIX
4310 interface uses malloc() for output vectors. See item 31 above.
4311
4312 PCRE_CONFIG_MATCH_LIMIT
4313
4314 The output is an unsigned integer that contains the default limit of the number
4315 of match() calls in a pcre_exec() execution. See 42 above.
4316
4317 44. pcretest has been upgraded by the addition of the -C option. This causes it
4318 to extract all the available output from the new pcre_config() function, and to
4319 output it. The program then exits immediately.
4320
4321 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4322 order to support additional features. One way would have been to define
4323 pcre_exec2() (for example) with extra arguments, but this would not have been
4324 extensible, and would also have required all calls to the original function to
4325 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4326 is used for passing in "extra" data from pcre_study().
4327
4328 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4329 contains the following fields:
4330
4331 flags a bitmap indicating which of the following fields are set
4332 study_data opaque data from pcre_study()
4333 match_limit a way of specifying a limit on match() calls for a specific
4334 call to pcre_exec()
4335 callout_data data for callouts (see 49 below)
4336
4337 The flag bits are also defined in pcre.h, and are
4338
4339 PCRE_EXTRA_STUDY_DATA
4340 PCRE_EXTRA_MATCH_LIMIT
4341 PCRE_EXTRA_CALLOUT_DATA
4342
4343 The pcre_study() function now returns one of these new pcre_extra blocks, with
4344 the actual study data pointed to by the study_data field, and the
4345 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4346 before. That is, this change is entirely upwards-compatible and requires no
4347 change to existing code.
4348
4349 If you want to pass in additional data to pcre_exec(), you can either place it
4350 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4351 block.
4352
4353 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4354 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4355 times with different match limits, until it finds the minimum value needed for
4356 pcre_exec() to complete. The value is then output. This can be instructive; for
4357 most simple matches the number is quite small, but for pathological cases it
4358 gets very large very quickly.
4359
4360 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4361 returns the size of the data block pointed to by the study_data field in a
4362 pcre_extra block, that is, the value that was passed as the argument to
4363 pcre_malloc() when PCRE was getting memory in which to place the information
4364 created by pcre_study(). The fourth argument should point to a size_t variable.
4365 pcretest has been extended so that this information is shown after a successful
4366 pcre_study() call when information about the compiled regex is being displayed.
4367
4368 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4369 because what follows is always an absolute path. (Later: it turns out that this
4370 is more than cosmetic for MinGW, because it doesn't like empty path
4371 components.)
4372
4373 49. Some changes have been made to the callout feature (see 28 above):
4374
4375 (i) A callout function now has three choices for what it returns:
4376
4377 0 => success, carry on matching
4378 > 0 => failure at this point, but backtrack if possible
4379 < 0 => serious error, return this value from pcre_exec()
4380
4381 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4382 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4383 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4384 use by callout functions. It will never be used by PCRE itself.
4385
4386 (ii) The pcre_extra structure (see 45 above) has a void * field called
4387 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4388 pcre_callout_block structure has a field of the same name. The contents of
4389 the field passed in the pcre_extra structure are passed to the callout
4390 function in the corresponding field in the callout block. This makes it
4391 easier to use the same callout-containing regex from multiple threads. For
4392 testing, the pcretest program has a new data escape
4393
4394 \C*n pass the number n (may be negative) as callout_data
4395
4396 If the callout function in pcretest receives a non-zero value as
4397 callout_data, it returns that value.
4398
4399 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4400 there were some redundant $(CFLAGS) in commands that are now specified as
4401 $(LINK), which already includes $(CFLAGS).
4402
4403 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4404 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4405 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4406 one-byte characters throughout. Note that case-insensitive matching applies
4407 only to characters whose values are less than 256. PCRE doesn't support the
4408 notion of cases for higher-valued characters.
4409
4410 (i) A character class whose characters are all within 0-255 is handled as
4411 a bit map, and the map is inverted for negative classes. Previously, a
4412 character > 255 always failed to match such a class; however it should
4413 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4414
4415 (ii) A negated character class with a single character < 255 is coded as
4416 "not this character" (OP_NOT). This wasn't working properly when the test
4417 character was multibyte, either singly or repeated.
4418
4419 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4420 mode, for example: \x{100}{2,3}.
4421
4422 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4423 singly or repeated) now correctly test multibyte characters. However,
4424 PCRE doesn't recognize any characters with values greater than 255 as
4425 digits, spaces, or word characters. Such characters always match \D, \S,
4426 and \W, and never match \d, \s, or \w.
4427
4428 (v) Classes may now contain characters and character ranges with values
4429 greater than 255. For example: [ab\x{100}-\x{400}].
4430
4431 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4432 PCRE in UTF-8 mode.
4433
4434 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4435 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4436 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4437 value.)
4438
4439 53. The single man page has become too large. I have therefore split it up into
4440 a number of separate man pages. These also give rise to individual HTML pages;
4441 these are now put in a separate directory, and there is an index.html page that
4442 lists them all. Some hyperlinking between the pages has been installed.
4443
4444 54. Added convenience functions for handling named capturing parentheses.
4445
4446 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4447 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4448 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4449 are faulted.
4450
4451 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4452 calling configure. These values are used when compiling the dftables.c program
4453 which is run to generate the source of the default character tables. They
4454 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4455 you will need to set these values.
4456
4457 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4458
4459
4460 Version 3.9 02-Jan-02
4461 ---------------------
4462
4463 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4464
4465 2. If --disable-static was given, the building process failed when trying to
4466 build pcretest and pcregrep. (For some reason it was using libtool to compile
4467 them, which is not right, as they aren't part of the library.)
4468
4469
4470 Version 3.8 18-Dec-01
4471 ---------------------
4472
4473 1. The experimental UTF-8 code was completely screwed up. It was packing the
4474 bytes in the wrong order. How dumb can you get?
4475
4476
4477 Version 3.7 29-Oct-01
4478 ---------------------
4479
4480 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4481 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4482 this didn't happen under Solaris 8, where I normally test things.
4483
4484 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4485 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4486 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4487 this fix an hour or so after the initial 3.7 release.)
4488
4489
4490 Version 3.6 23-Oct-01
4491 ---------------------
4492
4493 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4494 offsets passed as NULL with zero offset count.
4495
4496 2. The config.guess and config.sub files had not been updated when I moved to
4497 the latest autoconf.
4498
4499
4500 Version 3.5 15-Aug-01
4501 ---------------------
4502
4503 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4504 had been forgotten.
4505
4506 2. By using declared but undefined structures, we can avoid using "void"
4507 definitions in pcre.h while keeping the internal definitions of the structures
4508 private.
4509
4510 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4511 user point of view, this means that both static and shared libraries are built
4512 by default, but this can be individually controlled. More of the work of
4513 handling this static/shared cases is now inside libtool instead of PCRE's make
4514 file.
4515
4516 4. The pcretest utility is now installed along with pcregrep because it is
4517 useful for users (to test regexs) and by doing this, it automatically gets
4518 relinked by libtool. The documentation has been turned into a man page, so
4519 there are now .1, .txt, and .html versions in /doc.
4520
4521 5. Upgrades to pcregrep:
4522 (i) Added long-form option names like gnu grep.
4523 (ii) Added --help to list all options with an explanatory phrase.
4524 (iii) Added -r, --recursive to recurse into sub-directories.
4525 (iv) Added -f, --file to read patterns from a file.
4526
4527 6. pcre_exec() was referring to its "code" argument before testing that
4528 argument for NULL (and giving an error if it was NULL).
4529
4530 7. Upgraded Makefile.in to allow for compiling in a different directory from
4531 the source directory.
4532
4533 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4534 options bits, the pointer it was passed was to an int instead of to an unsigned
4535 long int. This mattered only on 64-bit systems.
4536
4537 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4538 generated) instead of pcre.in, which it its source. Also made the same change
4539 in several of the .c files.
4540
4541 10. A new release of gcc defines printf() as a macro, which broke pcretest
4542 because it had an ifdef in the middle of a string argument for printf(). Fixed
4543 by using separate calls to printf().
4544
4545 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4546 script, to force use of CR or LF instead of \n in the source. On non-Unix
4547 systems, the value can be set in config.h.
4548
4549 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4550 absolute limit. Changed the text of the error message to make this clear, and
4551 likewise updated the man page.
4552
4553 13. The limit of 99 on the number of capturing subpatterns has been removed.
4554 The new limit is 65535, which I hope will not be a "real" limit.
4555
4556
4557 Version 3.4 22-Aug-00
4558 ---------------------
4559
4560 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4561
4562 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4563
4564
4565 Version 3.3 01-Aug-00
4566 ---------------------
4567
4568 1. If an octal character was given, but the value was greater than \377, it
4569 was not getting masked to the least significant bits, as documented. This could
4570 lead to crashes in some systems.
4571
4572 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4573 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4574
4575 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4576 These just pass their arguments on to (pcre_free)(), but they are provided
4577 because some uses of PCRE bind it to non-C systems that can call its functions,
4578 but cannot call free() or pcre_free() directly.
4579
4580 4. Add "make test" as a synonym for "make check". Corrected some comments in
4581 the Makefile.
4582
4583 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4584 Makefile.
4585
4586 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4587 command called pgrep for grepping around the active processes.
4588
4589 7. Added the beginnings of support for UTF-8 character strings.
4590
4591 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4592 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4593 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4594 out for the ar command.)
4595
4596
4597 Version 3.2 12-May-00
4598 ---------------------
4599
4600 This is purely a bug fixing release.
4601
4602 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4603 of ZA. This was just one example of several cases that could provoke this bug,
4604 which was introduced by change 9 of version 2.00. The code for breaking
4605 infinite loops after an iteration that matches an empty string was't working
4606 correctly.
4607
4608 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4609 when matched against abbab (for example). After matching an empty string, it
4610 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4611 caused it to match further down the string than it should.
4612
4613 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4614 was there because it doesn't seem to be needed, and it causes trouble on some
4615 systems, as it is not a Standard C header. It has been removed.
4616
4617 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4618 were reported on the Macintosh. The changes were from
4619
4620 while ((c = *(++ptr)) != 0 && c != '\n');
4621 to
4622 while ((c = *(++ptr)) != 0 && c != '\n') ;
4623
4624 Totally extraordinary, but if that's what it takes...
4625
4626 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4627 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4628 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4629 assumes the way PCRE uses memmove() (always moving upwards).
4630
4631 6. PCRE is being used in one environment where strchr() is not available. There
4632 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4633 faster code anyway.
4634
4635
4636 Version 3.1 09-Feb-00
4637 ---------------------
4638
4639 The only change in this release is the fixing of some bugs in Makefile.in for
4640 the "install" target:
4641
4642 (1) It was failing to install pcreposix.h.
4643
4644 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4645
4646
4647 Version 3.0 01-Feb-00
4648 ---------------------
4649
4650 1. Add support for the /+ modifier to perltest (to output $` like it does in
4651 pcretest).
4652
4653 2. Add support for the /g modifier to perltest.
4654
4655 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4656 matches null strings.
4657
4658 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4659 pattern. Perl treats empty patterns specially - it reuses the most recent
4660 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4661 effect.
4662
4663 5. The POSIX interface was broken in that it was just handing over the POSIX
4664 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4665 required a bigger vector, with some working space on the end. This means that
4666 the POSIX wrapper now has to get and free some memory, and copy the results.
4667
4668 6. Added some simple autoconf support, placing the test data and the
4669 documentation in separate directories, re-organizing some of the
4670 information files, and making it build pcre-config (a GNU standard). Also added
4671 libtool support for building PCRE as a shared library, which is now the
4672 default.
4673
4674 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4675 09 are not valid octal constants. Single digits will be used for minor values
4676 less than 10.
4677
4678 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4679 existing programs that set these in the POSIX interface can use PCRE without
4680 modification.
4681
4682 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4683 return all that pcre_info() returns, plus additional data. The pcre_info()
4684 function is retained for compatibility, but is considered to be obsolete.
4685
4686 10. Added experimental recursion feature (?R) to handle one common case that
4687 Perl 5.6 will be able to do with (?p{...}).
4688
4689 11. Added support for POSIX character classes like [:alpha:], which Perl is
4690 adopting.
4691
4692
4693 Version 2.08 31-Aug-99
4694 ----------------------
4695
4696 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4697 trying to match at the startoffset position, but instead was moving forward to
4698 the next newline as if a previous match had failed.
4699
4700 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4701 and could get into a loop if a null string was matched other than at the start
4702 of the subject.
4703
4704 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4705 be distinguished at compile time, and for completeness also added PCRE_DATE.
4706
4707 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4708 in GnuWin32 environments.
4709
4710
4711 Version 2.07 29-Jul-99
4712 ----------------------
4713
4714 1. The documentation is now supplied in plain text form and HTML as well as in
4715 the form of man page sources.
4716
4717 2. C++ compilers don't like assigning (void *) values to other pointer types.
4718 In particular this affects malloc(). Although there is no problem in Standard
4719 C, I've put in casts to keep C++ compilers happy.
4720
4721 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4722 should be (const char *).
4723
4724 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4725 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4726 However, I haven't made this a standard facility. The documentation doesn't
4727 mention it, and the Makefile doesn't support it.
4728
4729 5. The Makefile now contains an "install" target, with editable destinations at
4730 the top of the file. The pcretest program is not installed.
4731
4732 6. pgrep -V now gives the PCRE version number and date.
4733
4734 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4735 causing the entire string to be ignored, instead of just the last character.
4736
4737 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4738 non-matching string, it can take a very, very long time, even for strings of
4739 quite modest length, because of the nested recursion. PCRE now does better in
4740 some of these cases. It does this by remembering the last required literal
4741 character in the pattern, and pre-searching the subject to ensure it is present
4742 before running the real match. In other words, it applies a heuristic to detect
4743 some types of certain failure quickly, and in the above example, if presented
4744 with a string that has no trailing " it gives "no match" very quickly.
4745
4746 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4747 other alternatives are tried instead.
4748
4749
4750 Version 2.06 09-Jun-99
4751 ----------------------
4752
4753 1. Change pcretest's output for amount of store used to show just the code
4754 space, because the remainder (the data block) varies in size between 32-bit and
4755 64-bit systems.
4756
4757 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4758 start matching at. This allows lookbehinds to work when searching for multiple
4759 occurrences in a string.
4760
4761 3. Added additional options to pcretest for testing multiple occurrences:
4762
4763 /+ outputs the rest of the string that follows a match
4764 /g loops for multiple occurrences, using the new startoffset argument
4765 /G loops for multiple occurrences by passing an incremented pointer
4766
4767 4. PCRE wasn't doing the "first character" optimization for patterns starting
4768 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4769 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4770 the letter 'x'. On long subject strings, this gives a significant speed-up.
4771
4772
4773 Version 2.05 21-Apr-99
4774 ----------------------
4775
4776 1. Changed the type of magic_number from int to long int so that it works
4777 properly on 16-bit systems.
4778
4779 2. Fixed a bug which caused patterns starting with .* not to work correctly
4780 when the subject string contained newline characters. PCRE was assuming
4781 anchoring for such patterns in all cases, which is not correct because .* will
4782 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4783 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4784 must be retried after every newline in the subject.
4785
4786
4787 Version 2.04 18-Feb-99
4788 ----------------------
4789
4790 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4791 computation of the store needed to hold the pattern was incorrect (too large).
4792 If such patterns were nested a few deep, this could multiply and become a real
4793 problem.
4794
4795 2. Added /M option to pcretest to show the memory requirement of a specific
4796 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4797
4798 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4799 compiled in such a way that the backtracking after subsequent failure was
4800 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4801 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4802
4803
4804 Version 2.03 02-Feb-99
4805 ----------------------
4806
4807 1. Fixed typo and small mistake in man page.
4808
4809 2. Added 4th condition (GPL supersedes if conflict) and created separate
4810 LICENCE file containing the conditions.
4811
4812 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4813 Perl, that is the internal \ allows the delimiter to be included in the
4814 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4815 the final delimiter, add \ to the end of the pattern (to test the error).
4816
4817 4. Added the convenience functions for extracting substrings after a successful
4818 match. Updated pcretest to make it able to test these functions.
4819
4820
4821 Version 2.02 14-Jan-99
4822 ----------------------
4823
4824 1. Initialized the working variables associated with each extraction so that
4825 their saving and restoring doesn't refer to uninitialized store.
4826
4827 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4828 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4829 fix the problem.
4830
4831 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4832 calls, and wasn't printing the correct value for compiling calls. Increased the
4833 default value of LOOPREPEAT, and the number of significant figures in the
4834 times.
4835
4836 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4837
4838 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4839 a building problem on Windows NT with a FAT file system.
4840
4841
4842 Version 2.01 21-Oct-98
4843 ----------------------
4844
4845 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4846 to character tables built by pcre_maketables() in the current locale. If NULL
4847 is passed, the default tables are used.
4848
4849
4850 Version 2.00 24-Sep-98
4851 ----------------------
4852
4853 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4854 it any more.
4855
4856 2. Allow quantification of (?>) groups, and make it work correctly.
4857
4858 3. The first character computation wasn't working for (?>) groups.
4859
4860 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4861 end of the subject) and add 5.005's \z, which really does match only at the
4862 very end of the subject.
4863
4864 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4865
4866 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4867 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4868 localized options. All options to pcre_study() were also removed.
4869
4870 7. Add other new features from 5.005:
4871
4872 $(?<= positive lookbehind
4873 $(?<! negative lookbehind
4874 (?imsx-imsx) added the unsetting capability
4875 such a setting is global if at outer level; local otherwise
4876 (?imsx-imsx:) non-capturing groups with option setting
4877 (?(cond)re|re) conditional pattern matching
4878
4879 A backreference to itself in a repeated group matches the previous
4880 captured string.
4881
4882 8. General tidying up of studying (both automatic and via "study")
4883 consequential on the addition of new assertions.
4884
4885 9. As in 5.005, unlimited repeated groups that could match an empty substring
4886 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4887 runtime if any iteration does actually match an empty substring.
4888
4889 10. Include the RunTest script in the distribution.
4890
4891 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4892 discrepancies, some of which were old and were also with respect to 5.004. They
4893 have now been fixed.
4894
4895
4896 Version 1.09 28-Apr-98
4897 ----------------------
4898
4899 1. A negated single character class followed by a quantifier with a minimum
4900 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4901 program crashes, or just wrong answers. This did not apply to negated classes
4902 containing more than one character, or to minima other than one.
4903
4904
4905 Version 1.08 27-Mar-98
4906 ----------------------
4907
4908 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4909
4910 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4911 latter must appear before anything that relies on it in the pattern.
4912
4913
4914 Version 1.07 16-Feb-98
4915 ----------------------
4916
4917 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4918 repeat of a potentially empty string).
4919
4920
4921 Version 1.06 23-Jan-98
4922 ----------------------
4923
4924 1. Added Markus Oberhumer's little patches for C++.
4925
4926 2. Literal strings longer than 255 characters were broken.
4927
4928
4929 Version 1.05 23-Dec-97
4930 ----------------------
4931
4932 1. Negated character classes containing more than one character were failing if
4933 PCRE_CASELESS was set at run time.
4934
4935
4936 Version 1.04 19-Dec-97
4937 ----------------------
4938
4939 1. Corrected the man page, where some "const" qualifiers had been omitted.
4940
4941 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4942 input syntax.
4943
4944 3. Fixed memory leak which occurred when a regex with back references was
4945 matched with an offsets vector that wasn't big enough. The temporary memory
4946 that is used in this case wasn't being freed if the match failed.
4947
4948 4. Tidied pcretest to ensure it frees memory that it gets.
4949
4950 5. Temporary memory was being obtained in the case where the passed offsets
4951 vector was exactly big enough.
4952
4953 6. Corrected definition of offsetof() from change 5 below.
4954
4955 7. I had screwed up change 6 below and broken the rules for the use of
4956 setjmp(). Now fixed.
4957
4958
4959 Version 1.03 18-Dec-97
4960 ----------------------
4961
4962 1. A erroneous regex with a missing opening parenthesis was correctly
4963 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4964 on some systems.
4965
4966 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4967 it was reported that one broken compiler failed on the former because "code" is
4968 also an independent variable.
4969
4970 3. The erroneous regex a[]b caused an array overrun reference.
4971
4972 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4973 fail on data ending with that character. (It was going on too far, and checking
4974 the next character, typically a binary zero.) This was specific to the
4975 optimized code for single-character negative classes.
4976
4977 5. Added a contributed patch from the TIN world which does the following:
4978
4979 + Add an undef for memmove, in case the the system defines a macro for it.
4980
4981 + Add a definition of offsetof(), in case there isn't one. (I don't know
4982 the reason behind this - offsetof() is part of the ANSI standard - but
4983 it does no harm).
4984
4985 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4986 most of the places where whitespace preceded '#'. I have given up and
4987 allowed the remaining 2 cases to be at the margin.
4988
4989 + Rename some variables in pcre to eliminate shadowing. This seems very
4990 pedantic, but does no harm, of course.
4991
4992 6. Moved the call to setjmp() into its own function, to get rid of warnings
4993 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4994
4995 7. Constructs such as \d{8,} were compiling into the equivalent of
4996 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4997 outcome, but in this particular case used more store than had been allocated,
4998 which caused the bug to be discovered because it threw up an internal error.
4999
5000 8. The debugging code in both pcre and pcretest for outputting the compiled
5001 form of a regex was going wrong in the case of back references followed by
5002 curly-bracketed repeats.
5003
5004
5005 Version 1.02 12-Dec-97
5006 ----------------------
5007
5008 1. Typos in pcre.3 and comments in the source fixed.
5009
5010 2. Applied a contributed patch to get rid of places where it used to remove
5011 'const' from variables, and fixed some signed/unsigned and uninitialized
5012 variable warnings.
5013
5014 3. Added the "runtest" target to Makefile.
5015
5016 4. Set default compiler flag to -O2 rather than just -O.
5017
5018
5019 Version 1.01 19-Nov-97
5020 ----------------------
5021
5022 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5023 like /([ab]*)*/, that is, for classes with more than one character in them.
5024
5025 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5026 as /((?>a*))*/ (a PCRE_EXTRA facility).
5027
5028
5029 Version 1.00 18-Nov-97
5030 ----------------------
5031
5032 1. Added compile-time macros to support systems such as SunOS4 which don't have
5033 memmove() or strerror() but have other things that can be used instead.
5034
5035 2. Arranged that "make clean" removes the executables.
5036
5037
5038 Version 0.99 27-Oct-97
5039 ----------------------
5040
5041 1. Fixed bug in code for optimizing classes with only one character. It was
5042 initializing a 32-byte map regardless, which could cause it to run off the end
5043 of the memory it had got.
5044
5045 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5046
5047
5048 Version 0.98 22-Oct-97
5049 ----------------------
5050
5051 1. Fixed bug in code for handling temporary memory usage when there are more
5052 back references than supplied space in the ovector. This could cause segfaults.
5053
5054
5055 Version 0.97 21-Oct-97
5056 ----------------------
5057
5058 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5059
5060 2. Optimized negated single characters not to use a bit map.
5061
5062 3. Brought error texts together as macro definitions; clarified some of them;
5063 fixed one that was wrong - it said "range out of order" when it meant "invalid
5064 escape sequence".
5065
5066 4. Changed some char * arguments to const char *.
5067
5068 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5069
5070 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5071 pcretest.
5072
5073
5074 Version 0.96 16-Oct-97
5075 ----------------------
5076
5077 1. Added a simple "pgrep" utility to the distribution.
5078
5079 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5080 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5081 where "ddd" means "one or more decimal digits".
5082
5083 3. Fixed serious bug. If a pattern had a back reference, but the call to
5084 pcre_exec() didn't supply a large enough ovector to record the related
5085 identifying subpattern, the match always failed. PCRE now remembers the number
5086 of the largest back reference, and gets some temporary memory in which to save
5087 the offsets during matching if necessary, in order to ensure that
5088 backreferences always work.
5089
5090 4. Increased the compatibility with Perl in a number of ways:
5091
5092 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5093 to request this handling. The option can be set at compile or exec time.
5094
5095 (b) $ matches before a terminating newline by default; an option
5096 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5097 mode). The option can be set at compile or exec time.
5098
5099 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5100 the same as Perl's. If the decimal number it represents is less than 10
5101 or there aren't that many previous left capturing parentheses, an octal
5102 escape is read. Inside a character class, it's always an octal escape,
5103 even if it is a single digit.
5104
5105 (d) An escaped but undefined alphabetic character is taken as a literal,
5106 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5107 escapes.
5108
5109 (e) {0} is now permitted. (The previous item is removed from the compiled
5110 pattern).
5111
5112 5. Changed all the names of code files so that the basic parts are no longer
5113 than 10 characters, and abolished the teeny "globals.c" file.
5114
5115 6. Changed the handling of character classes; they are now done with a 32-byte
5116 bit map always.
5117
5118 7. Added the -d and /D options to pcretest to make it possible to look at the
5119 internals of compilation without having to recompile pcre.
5120
5121
5122 Version 0.95 23-Sep-97
5123 ----------------------
5124
5125 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5126 \x20 at the start of a run of normal characters. These were being treated as
5127 real characters, instead of the source characters being re-checked.
5128
5129
5130 Version 0.94 18-Sep-97
5131 ----------------------
5132
5133 1. The functions are now thread-safe, with the caveat that the global variables
5134 containing pointers to malloc() and free() or alternative functions are the
5135 same for all threads.
5136
5137 2. Get pcre_study() to generate a bitmap of initial characters for non-
5138 anchored patterns when this is possible, and use it if passed to pcre_exec().
5139
5140
5141 Version 0.93 15-Sep-97
5142 ----------------------
5143
5144 1. /(b)|(:+)/ was computing an incorrect first character.
5145
5146 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5147 but not actually doing anything yet.
5148
5149 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5150 as Perl does (e.g. [-az] or [az-]).
5151
5152 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5153 all possible positions.
5154
5155 5. Split up into different modules to avoid including unneeded functions in a
5156 compiled binary. However, compile and exec are still in one module. The "study"
5157 function is split off.
5158
5159 6. The character tables are now in a separate module whose source is generated
5160 by an auxiliary program - but can then be edited by hand if required. There are
5161 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5162 toupper() in the code.
5163
5164 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5165 make them global. Abolish the function for setting them, as the caller can now
5166 set them directly.
5167
5168
5169 Version 0.92 11-Sep-97
5170 ----------------------
5171
5172 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5173 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5174
5175 2. Caseless matching was not working in character classes if the characters in
5176 the pattern were in upper case.
5177
5178 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5179
5180 4. Make PCRE_ANCHORED public and accept as a compile option.
5181
5182 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5183 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5184 pass them.
5185
5186 6. Give an error if bad option bits passed at compile or run time.
5187
5188 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5189 pcretest to cause it to pass that flag.
5190
5191 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5192 options, and the first character, if set.
5193
5194 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5195
5196
5197 Version 0.91 10-Sep-97
5198 ----------------------
5199
5200 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5201 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5202
5203 2. PCRE was looping on encountering an indefinitely repeated back reference to
5204 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5205 Perl does - treats the match as successful.
5206
5207 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5