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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1322 - (show annotations)
Wed May 1 19:42:31 2013 UTC (6 years, 8 months ago) by zherczeg
File size: 246381 byte(s)
Try madvise first before posix_madvise.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 28-April-2013
5 --------------------------
6
7 1. Added 'U' to some constants that are compared to unsigned integers, to
8 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9 variables that are added to signed variables, to ensure the result is
10 signed and can be negated.
11
12 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
13 CMake config files.
14
15 3. Revise the creation of config.h.generic so that all boolean macros are
16 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17 overriding via -D on the command line possible.
18
19 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20 to unsigned int is reported to make a quite noticeable speed difference in
21 a specific Windows environment. Testing on Linux did also appear to show
22 some benefit (and it is clearly not harmful). Also fixed the definition of
23 Xop which should be unsigned.
24
25 5. Related to (4), changing the definition of the intermediate variable cc
26 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27 improvements.
28
29 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
30 unnecessary spaces.
31
32 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33 and later.
34
35 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36
37 9. Optimizing fast_forward_start_bits in JIT.
38
39 10. Adding experimental support for callouts in JIT, and fixing some
40 issues revealed during this work. Namely:
41
42 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43
44 (b) Minimum length was not checked before the matching is started.
45
46 11. The value of capture_last that is passed to callouts was incorrect in some
47 cases when there was a capture on one path that was subsequently abandoned
48 after a backtrack. Also, the capture_last value is now reset after a
49 recursion, since all captures are also reset in this case.
50
51 12. The interpreter no longer returns the "too many substrings" error in the
52 case when an overflowing capture is in a branch that is subsequently
53 abandoned after a backtrack.
54
55 13. In the pathological case when an offset vector of size 2 is used, pcretest
56 now prints out the matched string after a yield of 0 or 1.
57
58 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59 Only supported by the JIT compiler at the moment.
60
61 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62
63 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
64 offset of the starting point of the matching process, provided the offsets
65 vector is large enough.
66
67 17. The \A escape now records a lookbehind value of 1, though its execution
68 does not actually inspect the previous character. This is to ensure that,
69 in partial multi-segment matching, at least one character from the old
70 segment is retained when a new segment is processed. Otherwise, if there
71 are no lookbehinds in the pattern, \A might match incorrectly at the start
72 of a new segment.
73
74 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75
76 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
77 gives some modest performance improvement in 8-bit mode.
78
79 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
80 be expressed in certain programming languages using Universal Character
81 Names.
82
83 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
84 which points out that "non characters" are not "characters that may not
85 appear in Unicode strings" but rather "characters that are reserved for
86 internal use and have only local meaning".
87
88 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
89 there was a conditional group that depended on an assertion, if the
90 assertion was false, the callout that immediately followed the alternation
91 in the condition was skipped when pcre_exec() was used for matching.
92
93 23. Allow an explicit callout to be inserted before an assertion that is the
94 condition for a conditional group, for compatibility with automatic
95 callouts, which always insert a callout at this point.
96
97 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
98 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
99
100 25. (*PRUNE) is now supported by the JIT compiler.
101
102 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
103
104 27. Fix the case where there are two or more SKIPs with arguments that may be
105 ignored.
106
107 28. (*SKIP) is now supported by the JIT compiler.
108
109 29. (*THEN) is now supported by the JIT compiler.
110
111 30. Update RunTest with additional test selector options.
112
113 31. The way PCRE handles backtracking verbs has been changed in two ways.
114
115 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
116 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
117 backtracking. In some cases this makes it more Perl-compatible, but Perl's
118 rather obscure rules do not always do the same thing.
119
120 (2) Previously, backtracking verbs were confined within assertions. This is
121 no longer the case for positive assertions, except for (*ACCEPT). Again,
122 this sometimes improves Perl compatibility, and sometimes does not.
123
124 32. A number of tests that were in test 2 because Perl did things differently
125 have been moved to test 1, because either Perl or PCRE has changed, and
126 these tests are now compatible.
127
128 32. Control verbs are handled in the same way in JIT and interpreter.
129
130 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
131 contained a forward subroutine reference caused a compile error.
132
133 34. Auto-detect and optimize limited repetitions in JIT.
134
135 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
136 blocking (*UTF) etc.
137
138 36. In the interpreter, maximizing pattern repetitions for characters and
139 character types now use tail recursion, which reduces stack usage.
140
141 37. The value of the max lookbehind was not correctly preserved if a compiled
142 and saved regex was reloaded on a host of different endianness.
143
144 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
145 of the compiled pattern block, expand the flags field from 16 to 32 bits
146 because it was almost full.
147
148 39. Try madvise first before posix_madvise.
149
150
151 Version 8.32 30-November-2012
152 -----------------------------
153
154 1. Improved JIT compiler optimizations for first character search and single
155 character iterators.
156
157 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
158 Patch by Daniel Richard G.
159
160 3. Single character iterator optimizations in the JIT compiler.
161
162 4. Improved JIT compiler optimizations for character ranges.
163
164 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
165 Reported by Giuseppe D'Angelo.
166
167 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
168 of a line, was being set incorrectly in cases where .* appeared inside
169 atomic brackets at the start of a pattern, or where there was a subsequent
170 *PRUNE or *SKIP.
171
172 7. Improved instruction cache flush for POWER/PowerPC.
173 Patch by Daniel Richard G.
174
175 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
176 grep:
177
178 (a) There is now no limit to the number of patterns to be matched.
179
180 (b) An error is given if a pattern is too long.
181
182 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
183 are now supported.
184
185 (d) --exclude-from and --include-from (multiple use) have been added.
186
187 (e) Exclusions and inclusions now apply to all files and directories, not
188 just to those obtained from scanning a directory recursively.
189
190 (f) Multiple uses of -f and --file-list are now supported.
191
192 (g) In a Windows environment, the default for -d has been changed from
193 "read" (the GNU grep default) to "skip", because otherwise the presence
194 of a directory in the file list provokes an error.
195
196 (h) The documentation has been revised and clarified in places.
197
198 9. Improve the matching speed of capturing brackets.
199
200 10. Changed the meaning of \X so that it now matches a Unicode extended
201 grapheme cluster.
202
203 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
204 out POSIX threads when JIT support is configured.
205
206 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
207
208 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
209 structure could go wrong in environments where size_t is not the same size
210 as int.
211
212 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
213 set.
214
215 15. The EBCDIC support had decayed; later updates to the code had included
216 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
217 general tidy up of EBCDIC-related issues, and the documentation was also
218 not quite right. There is now a test that can be run on ASCII systems to
219 check some of the EBCDIC-related things (but is it not a full test).
220
221 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
222 in a small tidy to the code.
223
224 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
225
226 18. If the --only-matching (-o) option in pcregrep is specified multiple
227 times, each one causes appropriate output. For example, -o1 -o2 outputs the
228 substrings matched by the 1st and 2nd capturing parentheses. A separating
229 string can be specified by --om-separator (default empty).
230
231 19. Improving the first n character searches.
232
233 20. Turn case lists for horizontal and vertical white space into macros so that
234 they are defined only once.
235
236 21. This set of changes together give more compatible Unicode case-folding
237 behaviour for characters that have more than one other case when UCP
238 support is available.
239
240 (a) The Unicode property table now has offsets into a new table of sets of
241 three or more characters that are case-equivalent. The MultiStage2.py
242 script that generates these tables (the pcre_ucd.c file) now scans
243 CaseFolding.txt instead of UnicodeData.txt for character case
244 information.
245
246 (b) The code for adding characters or ranges of characters to a character
247 class has been abstracted into a generalized function that also handles
248 case-independence. In UTF-mode with UCP support, this uses the new data
249 to handle characters with more than one other case.
250
251 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
252 whose other case is greater than 256 are now correctly matched
253 caselessly. Previously, the high codepoint matched the low one, but not
254 vice versa.
255
256 (d) The processing of \h, \H, \v, and \ in character classes now makes use
257 of the new class addition function, using character lists defined as
258 macros alongside the case definitions of 20 above.
259
260 (e) Caseless back references now work with characters that have more than
261 one other case.
262
263 (f) General caseless matching of characters with more than one other case
264 is supported.
265
266 22. Unicode character properties were updated from Unicode 6.2.0
267
268 23. Improved CMake support under Windows. Patch by Daniel Richard G.
269
270 24. Add support for 32-bit character strings, and UTF-32
271
272 25. Major JIT compiler update (code refactoring and bugfixing).
273 Experimental Sparc 32 support is added.
274
275 26. Applied a modified version of Daniel Richard G's patch to create
276 pcre.h.generic and config.h.generic by "make" instead of in the
277 PrepareRelease script.
278
279 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
280 pcre_compile.c when checking for a zero character.
281
282 28. Introducing a native interface for JIT. Through this interface, the compiled
283 machine code can be directly executed. The purpose of this interface is to
284 provide fast pattern matching, so several sanity checks are not performed.
285 However, feature tests are still performed. The new interface provides
286 1.4x speedup compared to the old one.
287
288 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
289 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
290 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
291
292 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
293 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
294 the "old" RFC 2279). Instead, it ended up passing a negative length to
295 pcre_exec().
296
297 31. Add support for GCC's visibility feature to hide internal functions.
298
299 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
300 "unknown -C option" after outputting 0 or 1.
301
302 33. There is now support for generating a code coverage report for the test
303 suite in environments where gcc is the compiler and lcov is installed. This
304 is mainly for the benefit of the developers.
305
306 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
307 unaddressable using valgrind annotations, allowing valgrind to detect
308 invalid memory accesses. This is mainly for the benefit of the developers.
309
310 25. (*UTF) can now be used to start a pattern in any of the three libraries.
311
312 26. Give configure error if --enable-cpp but no C++ compiler found.
313
314
315 Version 8.31 06-July-2012
316 -------------------------
317
318 1. Fixing a wrong JIT test case and some compiler warnings.
319
320 2. Removed a bashism from the RunTest script.
321
322 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
323 to unsigned type, result still unsigned" that was given by an MS compiler
324 on encountering the code "-sizeof(xxx)".
325
326 4. Partial matching support is added to the JIT compiler.
327
328 5. Fixed several bugs concerned with partial matching of items that consist
329 of more than one character:
330
331 (a) /^(..)\1/ did not partially match "aba" because checking references was
332 done on an "all or nothing" basis. This also applied to repeated
333 references.
334
335 (b) \R did not give a hard partial match if \r was found at the end of the
336 subject.
337
338 (c) \X did not give a hard partial match after matching one or more
339 characters at the end of the subject.
340
341 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
342 a partial match for the string "\r".
343
344 (e) When newline was set to CRLF, the metacharacter "." did not recognize
345 a partial match for a CR character at the end of the subject string.
346
347 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
348 running pcretest, the text "(JIT)" added to the output whenever JIT is
349 actually used to run the match.
350
351 7. Individual JIT compile options can be set in pcretest by following -s+[+]
352 or /S+[+] with a digit between 1 and 7.
353
354 8. OP_NOT now supports any UTF character not just single-byte ones.
355
356 9. (*MARK) control verb is now supported by the JIT compiler.
357
358 10. The command "./RunTest list" lists the available tests without actually
359 running any of them. (Because I keep forgetting what they all are.)
360
361 11. Add PCRE_INFO_MAXLOOKBEHIND.
362
363 12. Applied a (slightly modified) user-supplied patch that improves performance
364 when the heap is used for recursion (compiled with --disable-stack-for-
365 recursion). Instead of malloc and free for each heap frame each time a
366 logical recursion happens, frames are retained on a chain and re-used where
367 possible. This sometimes gives as much as 30% improvement.
368
369 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
370 call.
371
372 14. As documented, (*COMMIT) is now confined to within a positive assertion.
373
374 15. It is now possible to link pcretest with libedit as an alternative to
375 libreadline.
376
377 16. (*COMMIT) control verb is now supported by the JIT compiler.
378
379 17. The Unicode data tables have been updated to Unicode 6.1.0.
380
381 18. Added --file-list option to pcregrep.
382
383 19. Added binary file support to pcregrep, including the -a, --binary-files,
384 -I, and --text options.
385
386 20. The madvise function is renamed for posix_madvise for QNX compatibility
387 reasons. Fixed by Giuseppe D'Angelo.
388
389 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
390
391 22. Changed the option for creating symbolic links for 16-bit man pages from
392 -s to -sf so that re-installing does not cause issues.
393
394 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
395
396 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
397 matches in certain environments (the workspace was not being correctly
398 retained). Also added to pcre_dfa_exec() a simple plausibility check on
399 some of the workspace data at the beginning of a restart.
400
401 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
402 was not doing so when it should - probably a typo introduced by SVN 528
403 (change 8.10/14).
404
405 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
406 \w+ when the character tables indicated that \x{c4} was a word character.
407 There were several related cases, all because the tests for doing a table
408 lookup were testing for characters less than 127 instead of 255.
409
410 27. If a pattern contains capturing parentheses that are not used in a match,
411 their slots in the ovector are set to -1. For those that are higher than
412 any matched groups, this happens at the end of processing. In the case when
413 there were back references that the ovector was too small to contain
414 (causing temporary malloc'd memory to be used during matching), and the
415 highest capturing number was not used, memory off the end of the ovector
416 was incorrectly being set to -1. (It was using the size of the temporary
417 memory instead of the true size.)
418
419 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
420 ovector size, it uses memory at the end of the block that it has got.
421
422 29. Check for an overlong MARK name and give an error at compile time. The
423 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
424
425 30. JIT compiler update.
426
427 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
428 Rill for the patch.
429
430 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
431
432 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
433
434 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
435 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
436 was enabled.
437
438 35. Improve JIT code generation for greedy plus quantifier.
439
440 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
441 1 to "aa" instead of to an empty string. The bug affected repeated groups
442 that could potentially match an empty string.
443
444 37. Optimizing single character iterators in JIT.
445
446 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
447 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
448 codepoints that are too big for the mode are faulted, and in a UTF mode,
449 disallowed codepoints are also faulted.
450
451 39. If PCRE was compiled with UTF support, in three places in the DFA
452 matcher there was code that should only have been obeyed in UTF mode, but
453 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
454 processing when bytes with values greater than 127 were present. In 16-bit
455 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
456 both cases the values are those that cannot be the first data item in a UTF
457 character. The three items that might have provoked this were recursions,
458 possessively repeated groups, and atomic groups.
459
460 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
461 and pcregrep, because some OS require shared objects to be explicitly
462 passed to ld, causing the link step to fail if they are not.
463
464 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
465 mode, patterns that started with \h* or \R* might be incorrectly matched.
466
467
468 Version 8.30 04-February-2012
469 -----------------------------
470
471 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
472 name is defined in ctype.h.
473
474 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
475 only in quite long subpatterns.
476
477 3. Removed the function pcre_info(), which has been obsolete and deprecated
478 since it was replaced by pcre_fullinfo() in February 2000.
479
480 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
481 match a (*MARK), and the match failed at the start of the subject, a
482 reference to memory before the start of the subject could occur. This bug
483 was introduced by fix 17 of release 8.21.
484
485 5. A reference to an unset group with zero minimum repetition was giving
486 totally wrong answers (in non-JavaScript-compatibility mode). For example,
487 /(another)?(\1?)test/ matched against "hello world test". This bug was
488 introduced in release 8.13.
489
490 6. Add support for 16-bit character strings (a large amount of work involving
491 many changes and refactorings).
492
493 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
494 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
495 from a file.
496
497 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
498 rounding is not applied in this particular case).
499
500 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
501 if they appear, or are escaped, in patterns.
502
503 10. Get rid of a number of -Wunused-but-set-variable warnings.
504
505 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
506 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
507 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
508 also returns the mark "x". This bug applied to capturing parentheses,
509 non-capturing parentheses, and atomic parentheses. It also applied to some
510 assertions.
511
512 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
513 information out of configure.ac instead of relying on pcre.h.generic, which
514 is not stored in the repository.
515
516 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
517 -lreadline.
518
519 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
520
521 15. Applied Graycode's patch to put the top-level frame on the stack rather
522 than the heap when not using the stack for recursion. This gives a
523 performance improvement in many cases when recursion is not deep.
524
525 16. Experimental code added to "pcretest -C" to output the stack frame size.
526
527
528 Version 8.21 12-Dec-2011
529 ------------------------
530
531 1. Updating the JIT compiler.
532
533 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
534 are added as well.
535
536 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
537 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
538 calling _pcre_jit_exec. Some extra comments are added.
539
540 4. (*MARK) settings inside atomic groups that do not contain any capturing
541 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
542 was introduced by change 18 for 8.20.
543
544 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
545 ECMA-262 standard.
546
547 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
548 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
549 This bug was probably introduced by change 9 of 8.13.
550
551 7. While fixing 6 above, I noticed that a number of other items were being
552 incorrectly rejected as "not fixed length". This arose partly because newer
553 opcodes had not been added to the fixed-length checking code. I have (a)
554 corrected the bug and added tests for these items, and (b) arranged for an
555 error to occur if an unknown opcode is encountered while checking for fixed
556 length instead of just assuming "not fixed length". The items that were
557 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
558 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
559 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
560
561 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
562 being incorrectly compiled and would have given unpredicatble results.
563
564 9. A possessively repeated subpattern with minimum repeat count greater than
565 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
566 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
567 the first (A) could occur when it should not.
568
569 10. Add a cast and remove a redundant test from the code.
570
571 11. JIT should use pcre_malloc/pcre_free for allocation.
572
573 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
574 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
575 is anything other than /usr, -L is still shown).
576
577 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
578
579 14. Perl does not support \N without a following name in a [] class; PCRE now
580 also gives an error.
581
582 15. If a forward reference was repeated with an upper limit of around 2000,
583 it caused the error "internal error: overran compiling workspace". The
584 maximum number of forward references (including repeats) was limited by the
585 internal workspace, and dependent on the LINK_SIZE. The code has been
586 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
587 the default depends on LINK_SIZE. There is a new upper limit (for safety)
588 of around 200,000 forward references. While doing this, I also speeded up
589 the filling in of repeated forward references.
590
591 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
592 incorrectly expecting the subject to contain another "a" after the start.
593
594 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
595 in the match, the SKIP should be ignored. This was not happening; instead
596 the SKIP was being treated as NOMATCH. For patterns such as
597 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
598 tested.
599
600 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
601 now much more compatible with Perl, in particular in cases where the result
602 is a non-match for a non-anchored pattern. For example, if
603 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
604 "m", where previously it did not return a name. A side effect of this
605 change is that for partial matches, the last encountered mark name is
606 returned, as for non matches. A number of tests that were previously not
607 Perl-compatible have been moved into the Perl-compatible test files. The
608 refactoring has had the pleasing side effect of removing one argument from
609 the match() function, thus reducing its stack requirements.
610
611 19. If the /S+ option was used in pcretest to study a pattern using JIT,
612 subsequent uses of /S (without +) incorrectly behaved like /S+.
613
614 21. Retrieve executable code size support for the JIT compiler and fixing
615 some warnings.
616
617 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
618 not work when the shorter character appeared right at the end of the
619 subject string.
620
621 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
622 systems.
623
624 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
625 output it when the /M option is used in pcretest.
626
627 25. The CheckMan script was not being included in the distribution. Also, added
628 an explicit "perl" to run Perl scripts from the PrepareRelease script
629 because this is reportedly needed in Windows.
630
631 26. If study data was being save in a file and studying had not found a set of
632 "starts with" bytes for the pattern, the data written to the file (though
633 never used) was taken from uninitialized memory and so caused valgrind to
634 complain.
635
636 27. Updated RunTest.bat as provided by Sheri Pierce.
637
638 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
639
640 29. Computation of memory usage for the table of capturing group names was
641 giving an unnecessarily large value.
642
643
644 Version 8.20 21-Oct-2011
645 ------------------------
646
647 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
648 a POSIX class. After further experiments with Perl, which convinced me that
649 Perl has bugs and confusions, a closing square bracket is no longer allowed
650 in a POSIX name. This bug also affected patterns with classes that started
651 with full stops.
652
653 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
654 captured substring, but while checking the failing first alternative,
655 substring 1 is temporarily captured. If the output vector supplied to
656 pcre_exec() was not big enough for this capture, the yield of the function
657 was still zero ("insufficient space for captured substrings"). This cannot
658 be totally fixed without adding another stack variable, which seems a lot
659 of expense for a edge case. However, I have improved the situation in cases
660 such as /(a)(b)x|abc/ matched against "abc", where the return code
661 indicates that fewer than the maximum number of slots in the ovector have
662 been set.
663
664 3. Related to (2) above: when there are more back references in a pattern than
665 slots in the output vector, pcre_exec() uses temporary memory during
666 matching, and copies in the captures as far as possible afterwards. It was
667 using the entire output vector, but this conflicts with the specification
668 that only 2/3 is used for passing back captured substrings. Now it uses
669 only the first 2/3, for compatibility. This is, of course, another edge
670 case.
671
672 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
673 main code base, and can be used by building with --enable-jit. When this is
674 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
675 runtime --no-jit option is given.
676
677 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
678 ovector, the return from the function was zero, implying that there were
679 other matches that did not fit. The correct "exactly full" value is now
680 returned.
681
682 6. If a subpattern that was called recursively or as a subroutine contained
683 (*PRUNE) or any other control that caused it to give a non-standard return,
684 invalid errors such as "Error -26 (nested recursion at the same subject
685 position)" or even infinite loops could occur.
686
687 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
688 computing the minimum length on reaching *ACCEPT, and so ended up with the
689 wrong value of 1 rather than 0. Further investigation indicates that
690 computing a minimum subject length in the presence of *ACCEPT is difficult
691 (think back references, subroutine calls), and so I have changed the code
692 so that no minimum is registered for a pattern that contains *ACCEPT.
693
694 8. If (*THEN) was present in the first (true) branch of a conditional group,
695 it was not handled as intended. [But see 16 below.]
696
697 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
698 Sheri Pierce.
699
700 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
701 the first byte in a match must be "a".
702
703 11. Change 17 for 8.13 increased the recursion depth for patterns like
704 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
705 pattern contains any instances of (*THEN). If it does not, the old
706 optimizations are restored. It would be nice to do this on a per-group
707 basis, but at the moment that is not feasible.
708
709 12. In some environments, the output of pcretest -C is CRLF terminated. This
710 broke RunTest's code that checks for the link size. A single white space
711 character after the value is now allowed for.
712
713 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
714 For "fr", it uses the Windows-specific input and output files.
715
716 14. If (*THEN) appeared in a group that was called recursively or as a
717 subroutine, it did not work as intended. [But see next item.]
718
719 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
720 pattern fragments (but not containing any | characters). If A and B are
721 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
722 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
723 D. In other words, Perl considers parentheses that do not contain any |
724 characters to be part of a surrounding alternative, whereas PCRE was
725 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
726 differently. PCRE now behaves in the same way as Perl, except in the case
727 of subroutine/recursion calls such as (?1) which have in any case always
728 been different (but PCRE had them first :-).
729
730 16. Related to 15 above: Perl does not treat the | in a conditional group as
731 creating alternatives. Such a group is treated in the same way as an
732 ordinary group without any | characters when processing (*THEN). PCRE has
733 been changed to match Perl's behaviour.
734
735 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
736 RunGrepTest script failed.
737
738 18. Change 22 for version 13 caused atomic groups to use more stack. This is
739 inevitable for groups that contain captures, but it can lead to a lot of
740 stack use in large patterns. The old behaviour has been restored for atomic
741 groups that do not contain any capturing parentheses.
742
743 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
744 suppress the check for a minimum subject length at run time. (If it was
745 given to pcre_exec() or pcre_dfa_exec() it did work.)
746
747 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
748 fail to work when decoding hex characters in data strings in EBCDIC
749 environments.
750
751 21. It appears that in at least one Mac OS environment, the isxdigit() function
752 is implemented as a macro that evaluates to its argument more than once,
753 contravening the C 90 Standard (I haven't checked a later standard). There
754 was an instance in pcretest which caused it to go wrong when processing
755 \x{...} escapes in subject strings. The has been rewritten to avoid using
756 things like p++ in the argument of isxdigit().
757
758
759 Version 8.13 16-Aug-2011
760 ------------------------
761
762 1. The Unicode data tables have been updated to Unicode 6.0.0.
763
764 2. Two minor typos in pcre_internal.h have been fixed.
765
766 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
767 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
768 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
769
770 4. There were a number of related bugs in the code for matching backrefences
771 caselessly in UTF-8 mode when codes for the characters concerned were
772 different numbers of bytes. For example, U+023A and U+2C65 are an upper
773 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
774 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
775 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
776 2-byte code at the end of the subject (it thought there wasn't enough data
777 left).
778
779 5. Comprehensive information about what went wrong is now returned by
780 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
781 as the output vector has at least 2 elements. The offset of the start of
782 the failing character and a reason code are placed in the vector.
783
784 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
785 now returned is for the first byte of the failing character, instead of the
786 last byte inspected. This is an incompatible change, but I hope it is small
787 enough not to be a problem. It makes the returned offset consistent with
788 pcre_exec() and pcre_dfa_exec().
789
790 7. pcretest now gives a text phrase as well as the error number when
791 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
792 failure, the offset and reason code are output.
793
794 8. When \R was used with a maximizing quantifier it failed to skip backwards
795 over a \r\n pair if the subsequent match failed. Instead, it just skipped
796 back over a single character (\n). This seems wrong (because it treated the
797 two characters as a single entity when going forwards), conflicts with the
798 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
799 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
800 has been changed.
801
802 9. Some internal refactoring has changed the processing so that the handling
803 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
804 time (the PCRE_DOTALL option was changed this way some time ago: version
805 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
806 which was always a bit of a fudge. It also means that there is one less
807 argument for the match() function, which reduces its stack requirements
808 slightly. This change also fixes an incompatibility with Perl: the pattern
809 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
810
811 10. More internal refactoring has drastically reduced the number of recursive
812 calls to match() for possessively repeated groups such as (abc)++ when
813 using pcre_exec().
814
815 11. While implementing 10, a number of bugs in the handling of groups were
816 discovered and fixed:
817
818 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
819 (a|)*(?1) gave a compile-time internal error.
820 ((a|)+)+ did not notice that the outer group could match an empty string.
821 (^a|^)+ was not marked as anchored.
822 (.*a|.*)+ was not marked as matching at start or after a newline.
823
824 12. Yet more internal refactoring has removed another argument from the match()
825 function. Special calls to this function are now indicated by setting a
826 value in a variable in the "match data" data block.
827
828 13. Be more explicit in pcre_study() instead of relying on "default" for
829 opcodes that mean there is no starting character; this means that when new
830 ones are added and accidentally left out of pcre_study(), testing should
831 pick them up.
832
833 14. The -s option of pcretest has been documented for ages as being an old
834 synonym of -m (show memory usage). I have changed it to mean "force study
835 for every regex", that is, assume /S for every regex. This is similar to -i
836 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
837 using it. It makes it easier to run collections of tests with and without
838 study enabled, and thereby test pcre_study() more easily. All the standard
839 tests are now run with and without -s (but some patterns can be marked as
840 "never study" - see 20 below).
841
842 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
843 restoration of the capturing data to the outer values was not happening
844 correctly.
845
846 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
847 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
848 pattern had matched an empty string, and so incorrectly returned a no
849 match.
850
851 17. There was optimizing code for the last branch of non-capturing parentheses,
852 and also for the obeyed branch of a conditional subexpression, which used
853 tail recursion to cut down on stack usage. Unfortunately, now that there is
854 the possibility of (*THEN) occurring in these branches, tail recursion is
855 no longer possible because the return has to be checked for (*THEN). These
856 two optimizations have therefore been removed. [But see 8.20/11 above.]
857
858 18. If a pattern containing \R was studied, it was assumed that \R always
859 matched two bytes, thus causing the minimum subject length to be
860 incorrectly computed because \R can also match just one byte.
861
862 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
863 was incorrectly computed.
864
865 20. If /S is present twice on a test pattern in pcretest input, it now
866 *disables* studying, thereby overriding the use of -s on the command line
867 (see 14 above). This is necessary for one or two tests to keep the output
868 identical in both cases.
869
870 21. When (*ACCEPT) was used in an assertion that matched an empty string and
871 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
872
873 22. When an atomic group that contained a capturing parenthesis was
874 successfully matched, but the branch in which it appeared failed, the
875 capturing was not being forgotten if a higher numbered group was later
876 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
877 group 1 to "a", when in fact it should be unset. This applied to multi-
878 branched capturing and non-capturing groups, repeated or not, and also to
879 positive assertions (capturing in negative assertions does not happen
880 in PCRE) and also to nested atomic groups.
881
882 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
883 subject after a captured substring, to make it easier to tell which of a
884 number of identical substrings has been captured.
885
886 24. The way atomic groups are processed by pcre_exec() has been changed so that
887 if they are repeated, backtracking one repetition now resets captured
888 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
889 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
890 "aaa". Previously, it would have been "a". As part of this code
891 refactoring, the way recursive calls are handled has also been changed.
892
893 25. If an assertion condition captured any substrings, they were not passed
894 back unless some other capturing happened later. For example, if
895 (?(?=(a))a) was matched against "a", no capturing was returned.
896
897 26. When studying a pattern that contained subroutine calls or assertions,
898 the code for finding the minimum length of a possible match was handling
899 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
900 group 1 called group 2 while simultaneously a separate group 2 called group
901 1). A stack overflow occurred in this case. I have fixed this by limiting
902 the recursion depth to 10.
903
904 27. Updated RunTest.bat in the distribution to the version supplied by Tom
905 Fortmann. This supports explicit test numbers on the command line, and has
906 argument validation and error reporting.
907
908 28. An instance of \X with an unlimited repeat could fail if at any point the
909 first character it looked at was a mark character.
910
911 29. Some minor code refactoring concerning Unicode properties and scripts
912 should reduce the stack requirement of match() slightly.
913
914 30. Added the '=' option to pcretest to check the setting of unused capturing
915 slots at the end of the pattern, which are documented as being -1, but are
916 not included in the return count.
917
918 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
919 compiled something random. Now it gives a compile-time error (as does
920 Perl).
921
922 32. A *MARK encountered during the processing of a positive assertion is now
923 recorded and passed back (compatible with Perl).
924
925 33. If --only-matching or --colour was set on a pcregrep call whose pattern
926 had alternative anchored branches, the search for a second match in a line
927 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
928 matched the line "0102" twice. The same bug affected patterns that started
929 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
930 twice.
931
932 34. Previously, PCRE did not allow quantification of assertions. However, Perl
933 does, and because of capturing effects, quantifying parenthesized
934 assertions may at times be useful. Quantifiers are now allowed for
935 parenthesized assertions.
936
937 35. A minor code tidy in pcre_compile() when checking options for \R usage.
938
939 36. \g was being checked for fancy things in a character class, when it should
940 just be a literal "g".
941
942 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
943 appearance of a nested POSIX class supersedes an apparent external class.
944 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
945 unescaped square brackets may also appear as part of class names. For
946 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
947 more like Perl. (But see 8.20/1 above.)
948
949 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
950 was because it thought it was \N{name}, which is not supported).
951
952 39. Add minix to OS list not supporting the -S option in pcretest.
953
954 40. PCRE tries to detect cases of infinite recursion at compile time, but it
955 cannot analyze patterns in sufficient detail to catch mutual recursions
956 such as ((?1))((?2)). There is now a runtime test that gives an error if a
957 subgroup is called recursively as a subpattern for a second time at the
958 same position in the subject string. In previous releases this might have
959 been caught by the recursion limit, or it might have run out of stack.
960
961 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
962 happen only once. PCRE was, however incorrectly giving a compile time error
963 "recursive call could loop indefinitely" because it cannot analyze the
964 pattern in sufficient detail. The compile time test no longer happens when
965 PCRE is compiling a conditional subpattern, but actual runaway loops are
966 now caught at runtime (see 40 above).
967
968 42. It seems that Perl allows any characters other than a closing parenthesis
969 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
970 has been changed to be the same.
971
972 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
973 as not to get warnings when autogen.sh is called. Also changed
974 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
975
976 44. To help people who use pcregrep to scan files containing exceedingly long
977 lines, the following changes have been made:
978
979 (a) The default value of the buffer size parameter has been increased from
980 8K to 20K. (The actual buffer used is three times this size.)
981
982 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
983 PCRE is built.
984
985 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
986 to be set at run time.
987
988 (d) Numerical values in pcregrep options can be followed by K or M, for
989 example --buffer-size=50K.
990
991 (e) If a line being scanned overflows pcregrep's buffer, an error is now
992 given and the return code is set to 2.
993
994 45. Add a pointer to the latest mark to the callout data block.
995
996 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
997 partial match of an empty string instead of no match. This was specific to
998 the use of ".".
999
1000 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1001 complete match instead of a partial match. This bug was dependent on both
1002 the PCRE_UTF8 and PCRE_DOTALL options being set.
1003
1004 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1005 starting byte set, because \b was not being ignored.
1006
1007
1008 Version 8.12 15-Jan-2011
1009 ------------------------
1010
1011 1. Fixed some typos in the markup of the man pages, and wrote a script that
1012 checks for such things as part of the documentation building process.
1013
1014 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1015 --match-limit and --recursion-limit options (added for 8.11). In
1016 particular, this made one of the standard tests fail. (The integer value
1017 went into the wrong half of a long int.)
1018
1019 3. If the --colour option was given to pcregrep with -v (invert match), it
1020 did strange things, either producing crazy output, or crashing. It should,
1021 of course, ignore a request for colour when reporting lines that do not
1022 match.
1023
1024 4. Another pcregrep bug caused similar problems if --colour was specified with
1025 -M (multiline) and the pattern match finished with a line ending.
1026
1027 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1028 matched in multiline mode, the following line was shown as part of the
1029 match. This seems wrong, so I have changed it.
1030
1031 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1032 the check for further matches in the same line (so they could be coloured)
1033 to overrun the end of the current line. If another match was found, it was
1034 incorrectly shown (and then shown again when found in the next line).
1035
1036 7. If pcregrep was compiled under Windows, there was a reference to the
1037 function pcregrep_exit() before it was defined. I am assuming this was
1038 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1039 reported by a user. I've moved the definition above the reference.
1040
1041
1042 Version 8.11 10-Dec-2010
1043 ------------------------
1044
1045 1. (*THEN) was not working properly if there were untried alternatives prior
1046 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1047 backtracked to try for "b" instead of moving to the next alternative branch
1048 at the same level (in this case, to look for "c"). The Perl documentation
1049 is clear that when (*THEN) is backtracked onto, it goes to the "next
1050 alternative in the innermost enclosing group".
1051
1052 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1053 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1054 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1055 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1056 (*THEN).
1057
1058 3. If \s appeared in a character class, it removed the VT character from
1059 the class, even if it had been included by some previous item, for example
1060 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1061 of \s, but is part of the POSIX "space" class.)
1062
1063 4. A partial match never returns an empty string (because you can always
1064 match an empty string at the end of the subject); however the checking for
1065 an empty string was starting at the "start of match" point. This has been
1066 changed to the "earliest inspected character" point, because the returned
1067 data for a partial match starts at this character. This means that, for
1068 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1069 (previously it gave "no match").
1070
1071 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1072 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1073 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1074 has an implication that the given string is incomplete (because a partial
1075 match is preferred over a full match). For this reason, these items now
1076 give a partial match in this situation. [Aside: previously, the one case
1077 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1078 match rather than a full match, which was wrong by the old rules, but is
1079 now correct.]
1080
1081 6. There was a bug in the handling of #-introduced comments, recognized when
1082 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1083 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1084 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1085 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1086 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1087 places in pcre_compile().
1088
1089 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1090 comments when looking ahead for named forward references to subpatterns,
1091 the only newline sequence it recognized was NL. It now handles newlines
1092 according to the set newline convention.
1093
1094 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1095 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1096 cater for a lack of strerror(). These oversights have been fixed.
1097
1098 9. Added --match-limit and --recursion-limit to pcregrep.
1099
1100 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1101
1102 11. When the -o option was used, pcregrep was setting a return code of 1, even
1103 when matches were found, and --line-buffered was not being honoured.
1104
1105 12. Added an optional parentheses number to the -o and --only-matching options
1106 of pcregrep.
1107
1108 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1109 can match an empty string. The code to do it in pcretest and pcredemo
1110 needed fixing:
1111
1112 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1113 only one byte after an empty string match just before CRLF (this case
1114 just got forgotten; "any" and "anycrlf" were OK).
1115
1116 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1117 mode when an empty string match preceded an ASCII character followed by
1118 a non-ASCII character. (The code for advancing by one character rather
1119 than one byte was nonsense.)
1120
1121 (c) The pcredemo.c sample program did not have any code at all to handle
1122 the cases when CRLF is a valid newline sequence.
1123
1124 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1125 as a starting offset was within the subject string. There is now a new
1126 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1127 negative or greater than the length of the string. In order to test this,
1128 pcretest is extended to allow the setting of negative starting offsets.
1129
1130 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1131 starting offset points to the beginning of a UTF-8 character was
1132 unnecessarily clumsy. I tidied it up.
1133
1134 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1135 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1136
1137 17. Nobody had reported that the --include_dir option, which was added in
1138 release 7.7 should have been called --include-dir (hyphen, not underscore)
1139 for compatibility with GNU grep. I have changed it to --include-dir, but
1140 left --include_dir as an undocumented synonym, and the same for
1141 --exclude-dir, though that is not available in GNU grep, at least as of
1142 release 2.5.4.
1143
1144 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1145 characters from a string of bytes) have been redefined so as not to use
1146 loops, in order to improve performance in some environments. At the same
1147 time, I abstracted some of the common code into auxiliary macros to save
1148 repetition (this should not affect the compiled code).
1149
1150 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1151 compile-time error is now given if \c is not followed by an ASCII
1152 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1153 different, and any byte value is allowed.)
1154
1155 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1156 START_OPTIMIZE option, which is now allowed at compile time - but just
1157 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1158 to pcregrep and other applications that have no direct access to PCRE
1159 options. The new /Y option in pcretest sets this option when calling
1160 pcre_compile().
1161
1162 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1163 back references. Groups containing recursive back references were forced to
1164 be atomic by that change, but in the case of named groups, the amount of
1165 memory required was incorrectly computed, leading to "Failed: internal
1166 error: code overflow". This has been fixed.
1167
1168 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1169 pcretest.c, to avoid build problems in some Borland environments.
1170
1171
1172 Version 8.10 25-Jun-2010
1173 ------------------------
1174
1175 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1176 THEN.
1177
1178 2. (*ACCEPT) was not working when inside an atomic group.
1179
1180 3. Inside a character class, \B is treated as a literal by default, but
1181 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1182 causes the error). The code is unchanged, but I tidied the documentation.
1183
1184 4. Inside a character class, PCRE always treated \R and \X as literals,
1185 whereas Perl faults them if its -w option is set. I have changed PCRE so
1186 that it faults them when PCRE_EXTRA is set.
1187
1188 5. Added support for \N, which always matches any character other than
1189 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1190
1191 6. When compiling pcregrep with newer versions of gcc which may have
1192 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1193 declared with attribute warn_unused_result" were given. Just casting the
1194 result to (void) does not stop the warnings; a more elaborate fudge is
1195 needed. I've used a macro to implement this.
1196
1197 7. Minor change to pcretest.c to avoid a compiler warning.
1198
1199 8. Added four artifical Unicode properties to help with an option to make
1200 \s etc use properties (see next item). The new properties are: Xan
1201 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1202
1203 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1204 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1205 this option. Modified pcretest to add /W to test this facility. Added
1206 REG_UCP to make it available via the POSIX interface.
1207
1208 10. Added --line-buffered to pcregrep.
1209
1210 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1211 studied, and the match started with a letter with a code point greater than
1212 127 whose first byte was different to the first byte of the other case of
1213 the letter, the other case of this starting letter was not recognized
1214 (#976).
1215
1216 12. If a pattern that was studied started with a repeated Unicode property
1217 test, for example, \p{Nd}+, there was the theoretical possibility of
1218 setting up an incorrect bitmap of starting bytes, but fortunately it could
1219 not have actually happened in practice until change 8 above was made (it
1220 added property types that matched character-matching opcodes).
1221
1222 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1223 possible starting bytes for non-anchored patterns.
1224
1225 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1226 \R, and also a number of cases that involve Unicode properties, both
1227 explicit and implicit when PCRE_UCP is set.
1228
1229 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1230 input, it could crash or give wrong results if characters with values
1231 greater than 0xc0 were present in the subject string. (Detail: it assumed
1232 UTF-8 input when processing these items.)
1233
1234 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1235 size_t is 64-bit (#991).
1236
1237 17. Added a check for running out of memory when PCRE is compiled with
1238 --disable-stack-for-recursion (#990).
1239
1240 18. If the last data line in a file for pcretest does not have a newline on
1241 the end, a newline was missing in the output.
1242
1243 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1244 less than 128) in its various bitmaps. However, there is a facility for
1245 generating tables according to the current locale when PCRE is compiled. It
1246 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1247 characters, are recognized by isspace() and therefore were getting set in
1248 these tables, and indeed these tables seem to approximate to ISO 8859. This
1249 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1250 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1251 which of course cannot start UTF-8 characters. I have changed the code so
1252 that only real ASCII characters (less than 128) and the correct starting
1253 bytes for UTF-8 encodings are set for characters greater than 127 when in
1254 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1255 altogether.)
1256
1257 20. Added the /T option to pcretest so as to be able to run tests with non-
1258 standard character tables, thus making it possible to include the tests
1259 used for 19 above in the standard set of tests.
1260
1261 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1262 reference to a subpattern the other side of a comment that contains an
1263 opening parenthesis caused either an internal compiling error, or a
1264 reference to the wrong subpattern.
1265
1266
1267 Version 8.02 19-Mar-2010
1268 ------------------------
1269
1270 1. The Unicode data tables have been updated to Unicode 5.2.0.
1271
1272 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1273 configured.
1274
1275 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1276 original author of that file, following a query about its status.
1277
1278 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1279 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1280
1281 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1282 quantifier applied to a forward-referencing subroutine call, could compile
1283 incorrect code or give the error "internal error: previously-checked
1284 referenced subpattern not found".
1285
1286 6. Both MS Visual Studio and Symbian OS have problems with initializing
1287 variables to point to external functions. For these systems, therefore,
1288 pcre_malloc etc. are now initialized to local functions that call the
1289 relevant global functions.
1290
1291 7. There were two entries missing in the vectors called coptable and poptable
1292 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1293 I've fixed the data, and added a kludgy way of testing at compile time that
1294 the lengths are correct (equal to the number of opcodes).
1295
1296 8. Following on from 7, I added a similar kludge to check the length of the
1297 eint vector in pcreposix.c.
1298
1299 9. Error texts for pcre_compile() are held as one long string to avoid too
1300 much relocation at load time. To find a text, the string is searched,
1301 counting zeros. There was no check for running off the end of the string,
1302 which could happen if a new error number was added without updating the
1303 string.
1304
1305 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1306
1307 11. \K was not working if it appeared in an atomic group or in a group that
1308 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1309 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1310 the assertion is positive, but not if it is negative.
1311
1312 12. Change 11 fortuitously reduced the size of the stack frame used in the
1313 "match()" function of pcre_exec.c by one pointer. Forthcoming
1314 implementation of support for (*MARK) will need an extra pointer on the
1315 stack; I have reserved it now, so that the stack frame size does not
1316 decrease.
1317
1318 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1319 item in branch that calls a recursion is a subroutine call - as in the
1320 second branch in the above example - was incorrectly given the compile-
1321 time error "recursive call could loop indefinitely" because pcre_compile()
1322 was not correctly checking the subroutine for matching a non-empty string.
1323
1324 14. The checks for overrunning compiling workspace could trigger after an
1325 overrun had occurred. This is a "should never occur" error, but it can be
1326 triggered by pathological patterns such as hundreds of nested parentheses.
1327 The checks now trigger 100 bytes before the end of the workspace.
1328
1329 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1330
1331
1332 Version 8.01 19-Jan-2010
1333 ------------------------
1334
1335 1. If a pattern contained a conditional subpattern with only one branch (in
1336 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1337 computed the wrong minimum data length (which is of course zero for such
1338 subpatterns). This could cause incorrect "no match" results.
1339
1340 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1341 the pattern is reset in the first branch, pcre_compile() failed with
1342 "internal error: code overflow at offset...". This happened only when
1343 the reset was to the original external option setting. (An optimization
1344 abstracts leading options settings into an external setting, which was the
1345 cause of this.)
1346
1347 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1348 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1349 assertion pattern did not match (meaning that the assertion was true), it
1350 was incorrectly treated as false if the SKIP had been reached during the
1351 matching. This also applied to assertions used as conditions.
1352
1353 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1354 assertion subpattern, including such a pattern used as a condition,
1355 unpredictable results occurred, instead of the error return
1356 PCRE_ERROR_DFA_UITEM.
1357
1358 5. The C++ GlobalReplace function was not working like Perl for the special
1359 situation when an empty string is matched. It now does the fancy magic
1360 stuff that is necessary.
1361
1362 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1363 removed. (These were left over from very, very early versions of PCRE.)
1364
1365 7. Some cosmetic changes to the code to make life easier when compiling it
1366 as part of something else:
1367
1368 (a) Change DEBUG to PCRE_DEBUG.
1369
1370 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1371 called "current" as "current_branch", to prevent a collision with the
1372 Linux macro when compiled as a kernel module.
1373
1374 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1375 prevent a collision with the Linux macro when compiled as a kernel
1376 module.
1377
1378 8. In pcre_compile() there are some checks for integer overflows that used to
1379 cast potentially large values to (double). This has been changed to that
1380 when building, a check for int64_t is made, and if it is found, it is used
1381 instead, thus avoiding the use of floating point arithmetic. (There is no
1382 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1383 double.
1384
1385 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1386 2005 (difference between two addresses compared to an unsigned value).
1387
1388 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1389 custom one, because of the following reported problem in Windows:
1390
1391 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1392 under Win32.
1393 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1394 therefore missing the function definition.
1395 - The compiler thus generates a "C" signature for the test function.
1396 - The linker fails to find the "C" function.
1397 - PCRE fails to configure if asked to do so against libbz2.
1398
1399 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1400 messages were output:
1401
1402 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1403 rerunning libtoolize, to keep the correct libtool macros in-tree.
1404 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1405
1406 I have done both of these things.
1407
1408 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1409 most of the time, it *can* run out if it is given a pattern that contains a
1410 runaway infinite recursion. I updated the discussion in the pcrestack man
1411 page.
1412
1413 13. Now that we have gone to the x.xx style of version numbers, the minor
1414 version may start with zero. Using 08 or 09 is a bad idea because users
1415 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1416 interpreted as invalid octal numbers. I've updated the previous comment in
1417 configure.ac, and also added a check that gives an error if 08 or 09 are
1418 used.
1419
1420 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1421 causing partial matching to fail when the end of the subject matched \W
1422 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1423
1424 15. There were some discrepancies between the declarations in pcre_internal.h
1425 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1426 their definitions. The declarations used "const uschar *" and the
1427 definitions used USPTR. Even though USPTR is normally defined as "const
1428 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1429 reported that: "This difference in casting confuses some C++ compilers, for
1430 example, SunCC recognizes above declarations as different functions and
1431 generates broken code for hbpcre." I have changed the declarations to use
1432 USPTR.
1433
1434 16. GNU libtool is named differently on some systems. The autogen.sh script now
1435 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1436 (FreeBSD).
1437
1438 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1439 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1440 comment: "Figure out how to create a longlong from a string: strtoll and
1441 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1442 instance, but it only takes 2 args instead of 3!"
1443
1444 18. A subtle bug concerned with back references has been fixed by a change of
1445 specification, with a corresponding code fix. A pattern such as
1446 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1447 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1448 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1449 same bug. Such groups have to be quantified to be useful, or contained
1450 inside another quantified group. (If there's no repetition, the reference
1451 can never match.) The problem arises because, having left the group and
1452 moved on to the rest of the pattern, a later failure that backtracks into
1453 the group uses the captured value from the final iteration of the group
1454 rather than the correct earlier one. I have fixed this in PCRE by forcing
1455 any group that contains a reference to itself to be an atomic group; that
1456 is, there cannot be any backtracking into it once it has completed. This is
1457 similar to recursive and subroutine calls.
1458
1459
1460 Version 8.00 19-Oct-09
1461 ----------------------
1462
1463 1. The table for translating pcre_compile() error codes into POSIX error codes
1464 was out-of-date, and there was no check on the pcre_compile() error code
1465 being within the table. This could lead to an OK return being given in
1466 error.
1467
1468 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1469 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1470 in a Windows environment.
1471
1472 3. The pcregrep --count option prints the count for each file even when it is
1473 zero, as does GNU grep. However, pcregrep was also printing all files when
1474 --files-with-matches was added. Now, when both options are given, it prints
1475 counts only for those files that have at least one match. (GNU grep just
1476 prints the file name in this circumstance, but including the count seems
1477 more useful - otherwise, why use --count?) Also ensured that the
1478 combination -clh just lists non-zero counts, with no names.
1479
1480 4. The long form of the pcregrep -F option was incorrectly implemented as
1481 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1482 but it seems right to fix it, and I didn't think it was worth preserving
1483 the old behaviour.
1484
1485 5. The command line items --regex=pattern and --regexp=pattern were not
1486 recognized by pcregrep, which required --regex pattern or --regexp pattern
1487 (with a space rather than an '='). The man page documented the '=' forms,
1488 which are compatible with GNU grep; these now work.
1489
1490 6. No libpcreposix.pc file was created for pkg-config; there was just
1491 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1492
1493 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1494 when UCP support is not needed, by modifying the Python script that
1495 generates it from Unicode data files. This should not matter if the module
1496 is correctly used as a library, but I received one complaint about 50K of
1497 unwanted data. My guess is that the person linked everything into his
1498 program rather than using a library. Anyway, it does no harm.
1499
1500 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1501 was a minimum greater than 1 for a wide character in a possessive
1502 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1503 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1504 character. Chaos in the form of incorrect output or a compiling loop could
1505 result.
1506
1507 9. The restrictions on what a pattern can contain when partial matching is
1508 requested for pcre_exec() have been removed. All patterns can now be
1509 partially matched by this function. In addition, if there are at least two
1510 slots in the offset vector, the offset of the earliest inspected character
1511 for the match and the offset of the end of the subject are set in them when
1512 PCRE_ERROR_PARTIAL is returned.
1513
1514 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1515 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1516 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1517 and may be more useful for multi-segment matching.
1518
1519 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1520 used to be given if ever the end of the subject was reached; now it is
1521 given only if matching could not proceed because another character was
1522 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1523 string "Z", which now yields "no match" instead of "partial match". In the
1524 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1525 final character ended with (*FAIL).
1526
1527 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1528 if the pattern had a "must contain" character that was already found in the
1529 earlier partial match, unless partial matching was again requested. For
1530 example, with the pattern /dog.(body)?/, the "must contain" character is
1531 "g". If the first part-match was for the string "dog", restarting with
1532 "sbody" failed. This bug has been fixed.
1533
1534 13. The string returned by pcre_dfa_exec() after a partial match has been
1535 changed so that it starts at the first inspected character rather than the
1536 first character of the match. This makes a difference only if the pattern
1537 starts with a lookbehind assertion or \b or \B (\K is not supported by
1538 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1539 matching functions compatible, and I think it's the right thing to do.
1540
1541 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1542 so that the demonstration program is easily available in environments where
1543 PCRE has not been installed from source.
1544
1545 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1546 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1547 library.
1548
1549 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1550 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1551 is not the first non-POSIX option to be added. Clearly some people find
1552 these options useful.
1553
1554 17. If a caller to the POSIX matching function regexec() passes a non-zero
1555 value for nmatch with a NULL value for pmatch, the value of
1556 nmatch is forced to zero.
1557
1558 18. RunGrepTest did not have a test for the availability of the -u option of
1559 the diff command, as RunTest does. It now checks in the same way as
1560 RunTest, and also checks for the -b option.
1561
1562 19. If an odd number of negated classes containing just a single character
1563 interposed, within parentheses, between a forward reference to a named
1564 subpattern and the definition of the subpattern, compilation crashed with
1565 an internal error, complaining that it could not find the referenced
1566 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1567 [The bug was that it was starting one character too far in when skipping
1568 over the character class, thus treating the ] as data rather than
1569 terminating the class. This meant it could skip too much.]
1570
1571 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1572 /g option in pcretest when the pattern contains \K, which makes it possible
1573 to have an empty string match not at the start, even when the pattern is
1574 anchored. Updated pcretest and pcredemo to use this option.
1575
1576 21. If the maximum number of capturing subpatterns in a recursion was greater
1577 than the maximum at the outer level, the higher number was returned, but
1578 with unset values at the outer level. The correct (outer level) value is
1579 now given.
1580
1581 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1582 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1583 make it do so. The string so far is captured, making this feature
1584 compatible with Perl.
1585
1586 23. The tests have been re-organized, adding tests 11 and 12, to make it
1587 possible to check the Perl 5.10 features against Perl 5.10.
1588
1589 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1590 pattern matches a fixed length string. PCRE did not allow this; now it
1591 does. Neither allows recursion.
1592
1593 25. I finally figured out how to implement a request to provide the minimum
1594 length of subject string that was needed in order to match a given pattern.
1595 (It was back references and recursion that I had previously got hung up
1596 on.) This code has now been added to pcre_study(); it finds a lower bound
1597 to the length of subject needed. It is not necessarily the greatest lower
1598 bound, but using it to avoid searching strings that are too short does give
1599 some useful speed-ups. The value is available to calling programs via
1600 pcre_fullinfo().
1601
1602 26. While implementing 25, I discovered to my embarrassment that pcretest had
1603 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1604 study optimizations had never been tested with that matching function.
1605 Oops. What is worse, even when it was passed study data, there was a bug in
1606 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1607 were also very few tests of studied patterns with pcre_dfa_exec().
1608
1609 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1610 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1611 on the other side of the coin, they are no longer allowed to have different
1612 names, because these cannot be distinguished in PCRE, and this has caused
1613 confusion. (This is a difference from Perl.)
1614
1615 28. When duplicate subpattern names are present (necessarily with different
1616 numbers, as required by 27 above), and a test is made by name in a
1617 conditional pattern, either for a subpattern having been matched, or for
1618 recursion in such a pattern, all the associated numbered subpatterns are
1619 tested, and the overall condition is true if the condition is true for any
1620 one of them. This is the way Perl works, and is also more like the way
1621 testing by number works.
1622
1623
1624 Version 7.9 11-Apr-09
1625 ---------------------
1626
1627 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1628 (pcretest), all targets were linked against these libraries. This included
1629 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1630 libraries. This caused unwanted dependencies to be created. This problem
1631 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1632 pcretest is linked with readline.
1633
1634 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1635 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1636 moved outside it again, because FALSE and TRUE are already defined in AIX,
1637 but BOOL is not.
1638
1639 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1640 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1641
1642 4. The pcregrep documentation said spaces were inserted as well as colons (or
1643 hyphens) following file names and line numbers when outputting matching
1644 lines. This is not true; no spaces are inserted. I have also clarified the
1645 wording for the --colour (or --color) option.
1646
1647 5. In pcregrep, when --colour was used with -o, the list of matching strings
1648 was not coloured; this is different to GNU grep, so I have changed it to be
1649 the same.
1650
1651 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1652 each matching line was coloured. Now it goes on to look for further matches
1653 of any of the test patterns, which is the same behaviour as GNU grep.
1654
1655 7. A pattern that could match an empty string could cause pcregrep to loop; it
1656 doesn't make sense to accept an empty string match in pcregrep, so I have
1657 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1658 seems to be how GNU grep behaves.
1659
1660 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1661 start or after a newline", because the conditional assertion was not being
1662 correctly handled. The rule now is that both the assertion and what follows
1663 in the first alternative must satisfy the test.
1664
1665 9. If auto-callout was enabled in a pattern with a conditional group whose
1666 condition was an assertion, PCRE could crash during matching, both with
1667 pcre_exec() and pcre_dfa_exec().
1668
1669 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1670 used for matching.
1671
1672 11. Unicode property support in character classes was not working for
1673 characters (bytes) greater than 127 when not in UTF-8 mode.
1674
1675 12. Added the -M command line option to pcretest.
1676
1677 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1678
1679 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1680
1681 16. Added comments and documentation about mis-use of no_arg in the C++
1682 wrapper.
1683
1684 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1685 from Martin Jerabek that uses macro names for all relevant character and
1686 string constants.
1687
1688 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1689 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1690 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1691 these, but not everybody uses configure.
1692
1693 19. A conditional group that had only one branch was not being correctly
1694 recognized as an item that could match an empty string. This meant that an
1695 enclosing group might also not be so recognized, causing infinite looping
1696 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1697 with the subject "ab", where knowledge that the repeated group can match
1698 nothing is needed in order to break the loop.
1699
1700 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1701 exec(), but without supplying a callout function, matching went wrong.
1702
1703 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1704 leak if the size of the offset vector was greater than 30. When the vector
1705 is smaller, the saved offsets during recursion go onto a local stack
1706 vector, but for larger vectors malloc() is used. It was failing to free
1707 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1708 error, in fact).
1709
1710 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1711 heapframe that is used only when UTF-8 support is enabled. This caused no
1712 problem, but was untidy.
1713
1714 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1715 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1716 included within another project.
1717
1718 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1719 slightly modified by me:
1720
1721 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1722 not building pcregrep.
1723
1724 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1725 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1726
1727 25. Forward references, both numeric and by name, in patterns that made use of
1728 duplicate group numbers, could behave incorrectly or give incorrect errors,
1729 because when scanning forward to find the reference group, PCRE was not
1730 taking into account the duplicate group numbers. A pattern such as
1731 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1732
1733 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1734 the feature of a custom pointer more persuasive (as requested by a user).
1735
1736 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1737 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1738 pre-defined.
1739
1740 28. Added support for (*UTF8) at the start of a pattern.
1741
1742 29. Arrange for flags added by the "release type" setting in CMake to be shown
1743 in the configuration summary.
1744
1745
1746 Version 7.8 05-Sep-08
1747 ---------------------
1748
1749 1. Replaced UCP searching code with optimized version as implemented for Ad
1750 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1751 stage table and inline lookup instead of a function, giving speed ups of 2
1752 to 5 times on some simple patterns that I tested. Permission was given to
1753 distribute the MultiStage2.py script that generates the tables (it's not in
1754 the tarball, but is in the Subversion repository).
1755
1756 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1757 scripts.
1758
1759 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1760 a group with a zero qualifier. The result of the study could be incorrect,
1761 or the function might crash, depending on the pattern.
1762
1763 4. Caseless matching was not working for non-ASCII characters in back
1764 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1765 It now works when Unicode Property Support is available.
1766
1767 5. In pcretest, an escape such as \x{de} in the data was always generating
1768 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1769 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1770 truncation.
1771
1772 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1773
1774 7. Added two (int) casts to pcregrep when printing the difference of two
1775 pointers, in case they are 64-bit values.
1776
1777 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1778 test 2 if it fails.
1779
1780 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1781 and a #define of that name to empty if it is not externally set. This is to
1782 allow users of MSVC to set it if necessary.
1783
1784 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1785 the convenience functions in the pcre_get.c source file.
1786
1787 11. An option change at the start of a pattern that had top-level alternatives
1788 could cause overwriting and/or a crash. This command provoked a crash in
1789 some environments:
1790
1791 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1792
1793 This potential security problem was recorded as CVE-2008-2371.
1794
1795 12. For a pattern where the match had to start at the beginning or immediately
1796 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1797 pcre_dfa_exec() could read past the end of the passed subject if there was
1798 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1799 pcretest so that it places the subject at the end of its malloc-ed buffer.
1800
1801 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1802 exec() might read past the end of the data buffer in UTF-8 mode.
1803
1804 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1805 the data contained the byte 0x85 as part of a UTF-8 character within its
1806 first line. This applied both to normal and DFA matching.
1807
1808 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1809 /^[^d]*?$/8 failed to match "abc".
1810
1811 16. Added a missing copyright notice to pcrecpp_internal.h.
1812
1813 17. Make it more clear in the documentation that values returned from
1814 pcre_exec() in ovector are byte offsets, not character counts.
1815
1816 18. Tidied a few places to stop certain compilers from issuing warnings.
1817
1818 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1819 supplied by Stefan Weber. I made a further small update for 7.8 because
1820 there is a change of source arrangements: the pcre_searchfuncs.c module is
1821 replaced by pcre_ucd.c.
1822
1823
1824 Version 7.7 07-May-08
1825 ---------------------
1826
1827 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1828 a string to a long long, pretend we don't even have a long long." This is
1829 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1830
1831 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1832 pre-7.6 versions, which defined a global no_arg variable instead of putting
1833 it in the RE class. (See also #8 below.)
1834
1835 3. Remove a line of dead code, identified by coverity and reported by Nuno
1836 Lopes.
1837
1838 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1839
1840 (1) The include/exclude patterns were being applied to the whole pathnames
1841 of files, instead of just to the final components.
1842
1843 (2) If there was more than one level of directory, the subdirectories were
1844 skipped unless they satisfied the include/exclude conditions. This is
1845 inconsistent with GNU grep (and could even be seen as contrary to the
1846 pcregrep specification - which I improved to make it absolutely clear).
1847 The action now is always to scan all levels of directory, and just
1848 apply the include/exclude patterns to regular files.
1849
1850 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1851 --exclude_dir in the tests to avoid scanning .svn directories.
1852
1853 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1854 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1855 doesn't support NULs in patterns.
1856
1857 7. Added some missing "const"s to declarations of static tables in
1858 pcre_compile.c and pcre_dfa_exec.c.
1859
1860 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1861 caused by fix #2 above. (Subsequently also a second patch to fix the
1862 first patch. And a third patch - this was a messy problem.)
1863
1864 9. Applied Craig's patch to remove the use of push_back().
1865
1866 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1867 matching function regexec().
1868
1869 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1870 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1871 references. PCRE supports relative numbers with this syntax (I don't think
1872 Oniguruma does).
1873
1874 12. Previously, a group with a zero repeat such as (...){0} was completely
1875 omitted from the compiled regex. However, this means that if the group
1876 was called as a subroutine from elsewhere in the pattern, things went wrong
1877 (an internal error was given). Such groups are now left in the compiled
1878 pattern, with a new opcode that causes them to be skipped at execution
1879 time.
1880
1881 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1882 to the way PCRE behaves:
1883
1884 (a) A lone ] character is dis-allowed (Perl treats it as data).
1885
1886 (b) A back reference to an unmatched subpattern matches an empty string
1887 (Perl fails the current match path).
1888
1889 (c) A data ] in a character class must be notated as \] because if the
1890 first data character in a class is ], it defines an empty class. (In
1891 Perl it is not possible to have an empty class.) The empty class []
1892 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1893 The negative empty class [^] matches any one character, independently
1894 of the DOTALL setting.
1895
1896 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1897 non-existent subpattern following a character class starting with ']' and
1898 containing () gave an internal compiling error instead of "reference to
1899 non-existent subpattern". Fortunately, when the pattern did exist, the
1900 compiled code was correct. (When scanning forwards to check for the
1901 existencd of the subpattern, it was treating the data ']' as terminating
1902 the class, so got the count wrong. When actually compiling, the reference
1903 was subsequently set up correctly.)
1904
1905 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1906 it was being rejected as not supported by pcre_dfa_exec(), even though
1907 other assertions are supported. I have made pcre_dfa_exec() support
1908 (*FAIL).
1909
1910 16. The implementation of 13c above involved the invention of a new opcode,
1911 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1912 cannot be changed at match time, I realized I could make a small
1913 improvement to matching performance by compiling OP_ALLANY instead of
1914 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1915 on the OP_ANY path.
1916
1917 17. Compiling pcretest on Windows with readline support failed without the
1918 following two fixes: (1) Make the unistd.h include conditional on
1919 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1920
1921 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1922 ncurses library to be included for pcretest when ReadLine support is
1923 requested, but also to allow for it to be overridden. This patch came from
1924 Daniel Bergström.
1925
1926 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1927 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1928 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1929 spotting this.
1930
1931
1932 Version 7.6 28-Jan-08
1933 ---------------------
1934
1935 1. A character class containing a very large number of characters with
1936 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1937 overflow.
1938
1939 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1940 HAVE_LONG_LONG is not defined.
1941
1942 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1943 bring them up to date and include new features. This patch includes:
1944
1945 - Fixed PH's badly added libz and libbz2 support.
1946 - Fixed a problem with static linking.
1947 - Added pcredemo. [But later removed - see 7 below.]
1948 - Fixed dftables problem and added an option.
1949 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1950 HAVE_LONG_LONG.
1951 - Added readline support for pcretest.
1952 - Added an listing of the option settings after cmake has run.
1953
1954 4. A user submitted a patch to Makefile that makes it easy to create
1955 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1956 Makefile.am that cause it to include this special target, without
1957 affecting anything else. Note that the same mingw target plus all
1958 the other distribution libraries and programs are now supported
1959 when configuring with CMake (see 6 below) instead of with
1960 Configure/Make.
1961
1962 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1963 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1964 exported in the Windows port". It has not yet been confirmed that the patch
1965 solves the problem, but it does no harm.
1966
1967 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1968 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1969 with CMake, and also correct the comment about stack recursion.
1970
1971 7. Remove the automatic building of pcredemo from the ./configure system and
1972 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1973 of a program that users should build themselves after PCRE is installed, so
1974 building it automatically is not really right. What is more, it gave
1975 trouble in some build environments.
1976
1977 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1978
1979
1980 Version 7.5 10-Jan-08
1981 ---------------------
1982
1983 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1984 values in parens when parsing an RE using the C++ wrapper."
1985
1986 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1987 Characters greater than 255 were excluded from the class instead of being
1988 included.
1989
1990 3. The same bug as (2) above applied to negated POSIX classes such as
1991 [:^space:].
1992
1993 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1994 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1995 I have changed it.
1996
1997 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1998 first named subpattern) and a construct such as (?&a) would reference the
1999 first named subpattern whose name started with "a" (in other words, the
2000 length check was missing). Both these problems are fixed. "Subpattern name
2001 expected" is now given for (?&) (a zero-length name), and this patch also
2002 makes it give the same error for \k'' (previously it complained that that
2003 was a reference to a non-existent subpattern).
2004
2005 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2006 this is right because (?- can be followed by option settings as well as by
2007 digits. I have, however, made the messages clearer.
2008
2009 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2010 than the number used in the conditional) now cause a compile-time error.
2011 This is actually not compatible with Perl, which accepts such patterns, but
2012 treats the conditional as always being FALSE (as PCRE used to), but it
2013 seems to me that giving a diagnostic is better.
2014
2015 8. Change "alphameric" to the more common word "alphanumeric" in comments
2016 and messages.
2017
2018 9. Fix two occurrences of "backslash" in comments that should have been
2019 "backspace".
2020
2021 10. Remove two redundant lines of code that can never be obeyed (their function
2022 was moved elsewhere).
2023
2024 11. The program that makes PCRE's Unicode character property table had a bug
2025 which caused it to generate incorrect table entries for sequences of
2026 characters that have the same character type, but are in different scripts.
2027 It amalgamated them into a single range, with the script of the first of
2028 them. In other words, some characters were in the wrong script. There were
2029 thirteen such cases, affecting characters in the following ranges:
2030
2031 U+002b0 - U+002c1
2032 U+0060c - U+0060d
2033 U+0061e - U+00612
2034 U+0064b - U+0065e
2035 U+0074d - U+0076d
2036 U+01800 - U+01805
2037 U+01d00 - U+01d77
2038 U+01d9b - U+01dbf
2039 U+0200b - U+0200f
2040 U+030fc - U+030fe
2041 U+03260 - U+0327f
2042 U+0fb46 - U+0fbb1
2043 U+10450 - U+1049d
2044
2045 12. The -o option (show only the matching part of a line) for pcregrep was not
2046 compatible with GNU grep in that, if there was more than one match in a
2047 line, it showed only the first of them. It now behaves in the same way as
2048 GNU grep.
2049
2050 13. If the -o and -v options were combined for pcregrep, it printed a blank
2051 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2052 does the same. The return code can be used to tell if there were any
2053 non-matching lines.
2054
2055 14. Added --file-offsets and --line-offsets to pcregrep.
2056
2057 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2058 infinitely looping recursion. The bug was that positive lookaheads were not
2059 being skipped when checking for a possible empty match (negative lookaheads
2060 and both kinds of lookbehind were skipped).
2061
2062 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2063 inclusion of <windows.h> to before rather than after the definition of
2064 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2065
2066 17. Specifying a possessive quantifier with a specific limit for a Unicode
2067 character property caused pcre_compile() to compile bad code, which led at
2068 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2069 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2070 caused the error; without that there was no problem.
2071
2072 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2073
2074 19. Added --enable-pcretest-libreadline.
2075
2076 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2077 RE::GlobalReplace(). As a result, the number of replacements returned was
2078 double what it should be. I removed one of the increments, but Craig sent a
2079 later patch that removed the other one (the right fix) and added unit tests
2080 that check the return values (which was not done before).
2081
2082 21. Several CMake things:
2083
2084 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2085 the names libpcre and libpcreposix, not just pcre and pcreposix.
2086
2087 (2) The above change means that pcretest and pcregrep are now correctly
2088 linked with the newly-built libraries, not previously installed ones.
2089
2090 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2091
2092 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2093 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2094 UTF-8 newline character). The key issue is that the pattern starts .*;
2095 this means that the match must be either at the beginning, or after a
2096 newline. The bug was in the code for advancing after a failed match and
2097 checking that the new position followed a newline. It was not taking
2098 account of UTF-8 characters correctly.
2099
2100 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2101 character classes. PCRE was not treating the sequence [:...:] as a
2102 character class unless the ... were all letters. Perl, however, seems to
2103 allow any characters between [: and :], though of course it rejects as
2104 unknown any "names" that contain non-letters, because all the known class
2105 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2106 for example, whereas PCRE did not - it did not recognize a POSIX character
2107 class. This seemed a bit dangerous, so the code has been changed to be
2108 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2109 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2110 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2111 Perl does, and where it didn't before.
2112
2113 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2114 Windows environments %n is disabled by default.
2115
2116
2117 Version 7.4 21-Sep-07
2118 ---------------------
2119
2120 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2121 means that a class such as [\s] counted as "explicit reference to CR or
2122 LF". That isn't really right - the whole point of the change was to try to
2123 help when there was an actual mention of one of the two characters. So now
2124 the change happens only if \r or \n (or a literal CR or LF) character is
2125 encountered.
2126
2127 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2128 of both had grown to the point where there were only 3 bits left.
2129 Fortunately, there was spare space in the data structure, and so I have
2130 moved the internal flags into a new 16-bit field to free up more option
2131 bits.
2132
2133 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2134 but did not set the internal JCHANGED flag - either of these is enough to
2135 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2136 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2137 start sets both bits.
2138
2139 4. Added options (at build time, compile time, exec time) to change \R from
2140 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2141
2142 5. doc/pcresyntax.html was missing from the distribution.
2143
2144 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2145 compatibility, even though it is no longer used.
2146
2147 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2148 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2149 windows.h file is present (where different names are used). [This was
2150 reversed later after testing - see 16 below.]
2151
2152 8. Changed all #include <config.h> to #include "config.h". There were also
2153 some further <pcre.h> cases that I changed to "pcre.h".
2154
2155 9. When pcregrep was used with the --colour option, it missed the line ending
2156 sequence off the lines that it output.
2157
2158 10. It was pointed out to me that arrays of string pointers cause lots of
2159 relocations when a shared library is dynamically loaded. A technique of
2160 using a single long string with a table of offsets can drastically reduce
2161 these. I have refactored PCRE in four places to do this. The result is
2162 dramatic:
2163
2164 Originally: 290
2165 After changing UCP table: 187
2166 After changing error message table: 43
2167 After changing table of "verbs" 36
2168 After changing table of Posix names 22
2169
2170 Thanks to the folks working on Gregex for glib for this insight.
2171
2172 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2173 unicode-properties was also set.
2174
2175 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2176
2177 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2178 checked only for CRLF.
2179
2180 14. Added casts to pcretest.c to avoid compiler warnings.
2181
2182 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2183
2184 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2185 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2186 entirely. This removes changes made in 7 above.
2187
2188 17. The CMake files have been updated, and there is now more information about
2189 building with CMake in the NON-UNIX-USE document.
2190
2191
2192 Version 7.3 28-Aug-07
2193 ---------------------
2194
2195 1. In the rejigging of the build system that eventually resulted in 7.1, the
2196 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2197 brackets there is not right, since it causes compilers to look for an
2198 installed pcre.h, not the version that is in the source that is being
2199 compiled (which of course may be different). I have changed it back to:
2200
2201 #include "pcre.h"
2202
2203 I have a vague recollection that the change was concerned with compiling in
2204 different directories, but in the new build system, that is taken care of
2205 by the VPATH setting the Makefile.
2206
2207 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2208 when the subject happened to end in the byte 0x85 (e.g. if the last
2209 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2210 characters but of course it shouldn't be taken as a newline when it is part
2211 of another character. The bug was that, for an unlimited repeat of . in
2212 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2213 characters when looking for a newline.
2214
2215 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2216
2217 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2218 in debug output.
2219
2220 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2221 long printing in the pcrecpp unittest when running under MinGW.
2222
2223 6. ESC_K was left out of the EBCDIC table.
2224
2225 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2226 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2227 limit also applies to "virtual nesting" when a pattern is recursive, and in
2228 this case 1000 isn't so big. I have been able to remove this limit at the
2229 expense of backing off one optimization in certain circumstances. Normally,
2230 when pcre_exec() would call its internal match() function recursively and
2231 immediately return the result unconditionally, it uses a "tail recursion"
2232 feature to save stack. However, when a subpattern that can match an empty
2233 string has an unlimited repetition quantifier, it no longer makes this
2234 optimization. That gives it a stack frame in which to save the data for
2235 checking that an empty string has been matched. Previously this was taken
2236 from the 1000-entry workspace that had been reserved. So now there is no
2237 explicit limit, but more stack is used.
2238
2239 8. Applied Daniel's patches to solve problems with the import/export magic
2240 syntax that is required for Windows, and which was going wrong for the
2241 pcreposix and pcrecpp parts of the library. These were overlooked when this
2242 problem was solved for the main library.
2243
2244 9. There were some crude static tests to avoid integer overflow when computing
2245 the size of patterns that contain repeated groups with explicit upper
2246 limits. As the maximum quantifier is 65535, the maximum group length was
2247 set at 30,000 so that the product of these two numbers did not overflow a
2248 32-bit integer. However, it turns out that people want to use groups that
2249 are longer than 30,000 bytes (though not repeat them that many times).
2250 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2251 made it possible to implement the integer overflow checks in a much more
2252 dynamic way, which I have now done. The artificial limitation on group
2253 length has been removed - we now have only the limit on the total length of
2254 the compiled pattern, which depends on the LINK_SIZE setting.
2255
2256 10. Fixed a bug in the documentation for get/copy named substring when
2257 duplicate names are permitted. If none of the named substrings are set, the
2258 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2259 empty string.
2260
2261 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2262 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2263 because the ] is interpreted as the first data character and the
2264 terminating ] is not found. PCRE has been made compatible with Perl in this
2265 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2266 cause memory overwriting.
2267
2268 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2269 string has been matched (to stop an infinite loop). It was not recognizing
2270 a conditional subpattern that could match an empty string if that
2271 subpattern was within another subpattern. For example, it looped when
2272 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2273 condition was not nested. This bug has been fixed.
2274
2275 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2276 past the start of the subject in the presence of bytes with the top bit
2277 set, for example "\x8aBCD".
2278
2279 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2280 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2281
2282 14. Optimized (?!) to (*FAIL).
2283
2284 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2285 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2286 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2287 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2288 does: it's just the validity check that is more restrictive.
2289
2290 16. Inserted checks for integer overflows during escape sequence (backslash)
2291 processing, and also fixed erroneous offset values for syntax errors during
2292 backslash processing.
2293
2294 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2295 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2296
2297 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2298 caused an overrun.
2299
2300 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2301 something other than just ASCII characters) inside a group that had an
2302 unlimited repeat caused a loop at compile time (while checking to see
2303 whether the group could match an empty string).
2304
2305 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2306 [\P{Any}] did so. (Error in the code for printing property names.)
2307
2308 21. An orphan \E inside a character class could cause a crash.
2309
2310 22. A repeated capturing bracket such as (A)? could cause a wild memory
2311 reference during compilation.
2312
2313 23. There are several functions in pcre_compile() that scan along a compiled
2314 expression for various reasons (e.g. to see if it's fixed length for look
2315 behind). There were bugs in these functions when a repeated \p or \P was
2316 present in the pattern. These operators have additional parameters compared
2317 with \d, etc, and these were not being taken into account when moving along
2318 the compiled data. Specifically:
2319
2320 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2321 length.
2322
2323 (b) An item such as \pL+ within a repeated group could cause crashes or
2324 loops.
2325
2326 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2327 "reference to non-existent subpattern" error.
2328
2329 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2330
2331 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2332 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2333
2334 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2335
2336 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2337 character were causing crashes (broken optimization).
2338
2339 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2340 \p or \P) caused a compile-time loop.
2341
2342 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2343 break. For example, the unstudied pattern [\r\n]A does not match the string
2344 "\r\nA" because change 7.0/46 below moves the current point on by two
2345 characters after failing to match at the start. However, the pattern \nA
2346 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2347 the same is true. There doesn't seem any very clean way out of this, but
2348 what I have chosen to do makes the common cases work: PCRE now takes note
2349 of whether there can be an explicit match for \r or \n anywhere in the
2350 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2351 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2352 pattern has explicit CR or LF references.
2353
2354 29. Added (*CR) etc for changing newline setting at start of pattern.
2355
2356
2357 Version 7.2 19-Jun-07
2358 ---------------------
2359
2360 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2361 which is apparently normally available under Windows.
2362
2363 2. Re-jig the pcregrep tests with different newline settings in an attempt
2364 to make them independent of the local environment's newline setting.
2365
2366 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2367
2368 4. Some of the "internals" tests were previously cut out when the link size
2369 was not 2, because the output contained actual offsets. The recent new
2370 "Z" feature of pcretest means that these can be cut out, making the tests
2371 usable with all link sizes.
2372
2373 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2374 stack recursion. This gives a massive performance boost under BSD, but just
2375 a small improvement under Linux. However, it saves one field in the frame
2376 in all cases.
2377
2378 6. Added more features from the forthcoming Perl 5.10:
2379
2380 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2381 recursion call. It refers to the nth most recently opened parentheses.
2382
2383 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2384 to be opened parentheses.
2385
2386 (c) Conditions that refer to capturing parentheses can be specified
2387 relatively, for example, (?(-2)... or (?(+3)...
2388
2389 (d) \K resets the start of the current match so that everything before
2390 is not part of it.
2391
2392 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2393
2394 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2395 reference syntax.
2396
2397 (g) (?| introduces a group in which the numbering of parentheses in each
2398 alternative starts with the same number.
2399
2400 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2401
2402 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2403 PCRE_INFO_JCHANGED.
2404
2405 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2406 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2407 for detecting groups that can match an empty string.
2408
2409 9. A pattern with a very large number of alternatives (more than several
2410 hundred) was running out of internal workspace during the pre-compile
2411 phase, where pcre_compile() figures out how much memory will be needed. A
2412 bit of new cunning has reduced the workspace needed for groups with
2413 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2414 workspace instead of running out of the 4096 that are available.
2415
2416 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2417
2418 11. Applied patch from Google to remove an optimization that didn't quite work.
2419 The report of the bug said:
2420
2421 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2422 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2423 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2424
2425 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2426 it matched the wrong number of bytes.
2427
2428
2429 Version 7.1 24-Apr-07
2430 ---------------------
2431
2432 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2433 that is more "standard", making use of automake and other Autotools. There
2434 is some re-arrangement of the files and adjustment of comments consequent
2435 on this.
2436
2437 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2438 for recursive directory scanning broke on some systems because the files
2439 are not scanned in any specific order and on different systems the order
2440 was different. A call to "sort" has been inserted into RunGrepTest for the
2441 approprate test as a short-term fix. In the longer term there may be an
2442 alternative.
2443
2444 3. I had an email from Eric Raymond about problems translating some of PCRE's
2445 man pages to HTML (despite the fact that I distribute HTML pages, some
2446 people do their own conversions for various reasons). The problems
2447 concerned the use of low-level troff macros .br and .in. I have therefore
2448 removed all such uses from the man pages (some were redundant, some could
2449 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2450 HTML has been updated to handle .nf/.fi and to complain if it encounters
2451 .br or .in.
2452
2453 4. Updated comments in configure.ac that get placed in config.h.in and also
2454 arranged for config.h to be included in the distribution, with the name
2455 config.h.generic, for the benefit of those who have to compile without
2456 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2457
2458 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2459 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2460 makevp.bat for the current PCRE, using the additional files
2461 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2462
2463 6. A Windows user reported a minor discrepancy with test 2, which turned out
2464 to be caused by a trailing space on an input line that had got lost in his
2465 copy. The trailing space was an accident, so I've just removed it.
2466
2467 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2468 that is needed.
2469
2470 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2471 as "const" (a) because they are and (b) because it helps the PHP
2472 maintainers who have recently made a script to detect big data structures
2473 in the php code that should be moved to the .rodata section. I remembered
2474 to update Builducptable as well, so it won't revert if ucptable.h is ever
2475 re-created.
2476
2477 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2478 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2479 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2480 support is not required. This saves 1.5-2K of code, which is important in
2481 some applications.
2482
2483 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2484 so as not to refer to the tables, even though these functions will never be
2485 called when UTF-8 support is disabled. Otherwise there are problems with a
2486 shared library.
2487
2488 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2489
2490 (a) It was defining its arguments as char * instead of void *.
2491
2492 (b) It was assuming that all moves were upwards in memory; this was true
2493 a long time ago when I wrote it, but is no longer the case.
2494
2495 The emulated memove() is provided for those environments that have neither
2496 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2497 is clearly not the case, as these two bugs were recently reported.
2498
2499 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2500 and Detrail to create the HTML documentation, the .txt form of the man
2501 pages, and it removes trailing spaces from listed files. It also creates
2502 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2503 case, it wraps all the #defines with #ifndefs. This script should be run
2504 before "make dist".
2505
2506 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2507 with Unicode property support.
2508
2509 (a) For a maximizing quantifier, if the two different cases of the
2510 character were of different lengths in their UTF-8 codings (there are
2511 some cases like this - I found 11), and the matching function had to
2512 back up over a mixture of the two cases, it incorrectly assumed they
2513 were both the same length.
2514
2515 (b) When PCRE was configured to use the heap rather than the stack for
2516 recursion during matching, it was not correctly preserving the data for
2517 the other case of a UTF-8 character when checking ahead for a match
2518 while processing a minimizing repeat. If the check also involved
2519 matching a wide character, but failed, corruption could cause an
2520 erroneous result when trying to check for a repeat of the original
2521 character.
2522
2523 13. Some tidying changes to the testing mechanism:
2524
2525 (a) The RunTest script now detects the internal link size and whether there
2526 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2527 values substituted by "configure". (The RunGrepTest script already did
2528 this for UTF-8.) The configure.ac script no longer substitutes the
2529 relevant variables.
2530
2531 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2532 with length and offset values. This means that the output is different
2533 for different internal link sizes. Test 2 is skipped for link sizes
2534 other than 2 because of this, bypassing the problem. Unfortunately,
2535 there was also a test in test 3 (the locale tests) that used /B and
2536 failed for link sizes other than 2. Rather than cut the whole test out,
2537 I have added a new /Z option to pcretest that replaces the length and
2538 offset values with spaces. This is now used to make test 3 independent
2539 of link size. (Test 2 will be tidied up later.)
2540
2541 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2542 segmentation fault instead of returning the appropriate error message.
2543
2544 15. In multiline mode when the newline sequence was set to "any", the pattern
2545 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2546 This doesn't seem right; it now treats the CRLF combination as the line
2547 ending, and so does not match in that case. It's only a pattern such as ^$
2548 that would hit this one: something like ^ABC$ would have failed after \r
2549 and then tried again after \r\n.
2550
2551 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2552 in an attempt to make files that differ only in their line terminators
2553 compare equal. This works on Linux.
2554
2555 17. Under certain error circumstances pcregrep might try to free random memory
2556 as it exited. This is now fixed, thanks to valgrind.
2557
2558 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2559 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2560 was because its rules for how to advance for /g after matching an empty
2561 string at the end of a line did not allow for this case. They now check for
2562 it specially.
2563
2564 20. pcretest is supposed to handle patterns and data of any length, by
2565 extending its buffers when necessary. It was getting this wrong when the
2566 buffer for a data line had to be extended.
2567
2568 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2569 CRLF as a newline sequence.
2570
2571 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2572 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2573 I have nevertheless tidied it up.
2574
2575 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2576
2577 24. Added a man page for pcre-config.
2578
2579
2580 Version 7.0 19-Dec-06
2581 ---------------------
2582
2583 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2584 moving to gcc 4.1.1.
2585
2586 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2587 sys/time.h, which is documented as needed for this function. It doesn't
2588 seem to matter on Linux, but it showed up on some releases of OS X.
2589
2590 3. It seems that there are systems where bytes whose values are greater than
2591 127 match isprint() in the "C" locale. The "C" locale should be the
2592 default when a C program starts up. In most systems, only ASCII printing
2593 characters match isprint(). This difference caused the output from pcretest
2594 to vary, making some of the tests fail. I have changed pcretest so that:
2595
2596 (a) When it is outputting text in the compiled version of a pattern, bytes
2597 other than 32-126 are always shown as hex escapes.
2598
2599 (b) When it is outputting text that is a matched part of a subject string,
2600 it does the same, unless a different locale has been set for the match
2601 (using the /L modifier). In this case, it uses isprint() to decide.
2602
2603 4. Fixed a major bug that caused incorrect computation of the amount of memory
2604 required for a compiled pattern when options that changed within the
2605 pattern affected the logic of the preliminary scan that determines the
2606 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2607 that the computed length was too small. The symptoms of this bug were
2608 either the PCRE error "internal error: code overflow" from pcre_compile(),
2609 or a glibc crash with a message such as "pcretest: free(): invalid next
2610 size (fast)". Examples of patterns that provoked this bug (shown in
2611 pcretest format) are:
2612
2613 /(?-x: )/x
2614 /(?x)(?-x: \s*#\s*)/
2615 /((?i)[\x{c0}])/8
2616 /(?i:[\x{c0}])/8
2617
2618 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2619 is now done differently.
2620
2621 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2622 wrapper classes; (b) implement a new function in the C++ scanner that is
2623 more efficient than the old way of doing things because it avoids levels of
2624 recursion in the regex matching; (c) add a paragraph to the documentation
2625 for the FullMatch() function.
2626
2627 6. The escape sequence \n was being treated as whatever was defined as
2628 "newline". Not only was this contrary to the documentation, which states
2629 that \n is character 10 (hex 0A), but it also went horribly wrong when
2630 "newline" was defined as CRLF. This has been fixed.
2631
2632 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2633 was being set to -1 for the "end of line" case (supposedly a value that no
2634 character can have). Though this value is never used (the check for end of
2635 line is "zero bytes in current character"), it caused compiler complaints.
2636 I've changed it to 0xffffffff.
2637
2638 8. In pcre_version.c, the version string was being built by a sequence of
2639 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2640 string (as it is for production releases) called a macro with an empty
2641 argument. The C standard says the result of this is undefined. The gcc
2642 compiler treats it as an empty string (which was what was wanted) but it is
2643 reported that Visual C gives an error. The source has been hacked around to
2644 avoid this problem.
2645
2646 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2647 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2648 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2649 of them did).
2650
2651 10. Originally, pcretest opened its input and output without "b"; then I was
2652 told that "b" was needed in some environments, so it was added for release
2653 5.0 to both the input and output. (It makes no difference on Unix-like
2654 systems.) Later I was told that it is wrong for the input on Windows. I've
2655 now abstracted the modes into two macros, to make it easier to fiddle with
2656 them, and removed "b" from the input mode under Windows.
2657
2658 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2659
2660 12. Added -help and --help to pcretest as an official way of being reminded
2661 of the options.
2662
2663 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2664 and pcrecpp.cc because they annoy compilers at high warning levels.
2665
2666 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2667
2668 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2669 scripts are not C programs :-) and which was not noticed because it works
2670 on Linux.
2671
2672 16. pcretest is supposed to handle any length of pattern and data line (as one
2673 line or as a continued sequence of lines) by extending its input buffer if
2674 necessary. This feature was broken for very long pattern lines, leading to
2675 a string of junk being passed to pcre_compile() if the pattern was longer
2676 than about 50K.
2677
2678 17. I have done a major re-factoring of the way pcre_compile() computes the
2679 amount of memory needed for a compiled pattern. Previously, there was code
2680 that made a preliminary scan of the pattern in order to do this. That was
2681 OK when PCRE was new, but as the facilities have expanded, it has become
2682 harder and harder to keep it in step with the real compile phase, and there
2683 have been a number of bugs (see for example, 4 above). I have now found a
2684 cunning way of running the real compile function in a "fake" mode that
2685 enables it to compute how much memory it would need, while actually only
2686 ever using a few hundred bytes of working memory and without too many
2687 tests of the mode. This should make future maintenance and development
2688 easier. A side effect of this work is that the limit of 200 on the nesting
2689 depth of parentheses has been removed (though this was never a serious
2690 limitation, I suspect). However, there is a downside: pcre_compile() now
2691 runs more slowly than before (30% or more, depending on the pattern). I
2692 hope this isn't a big issue. There is no effect on runtime performance.
2693
2694 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2695 newline (only possible for the last line of a file) and it was a
2696 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2697
2698 19. Added additional timing features to pcretest. (1) The -tm option now times
2699 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2700 separate command line item, by a number that specifies the number of
2701 repeats to use when timing. The default is 50000; this gives better
2702 precision, but takes uncomfortably long for very large patterns.
2703
2704 20. Extended pcre_study() to be more clever in cases where a branch of a
2705 subpattern has no definite first character. For example, (a*|b*)[cd] would
2706 previously give no result from pcre_study(). Now it recognizes that the
2707 first character must be a, b, c, or d.
2708
2709 21. There was an incorrect error "recursive call could loop indefinitely" if
2710 a subpattern (or the entire pattern) that was being tested for matching an
2711 empty string contained only one non-empty item after a nested subpattern.
2712 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2713 incorrectly, because the \d was being skipped in the check.
2714
2715 22. The pcretest program now has a new pattern option /B and a command line
2716 option -b, which is equivalent to adding /B to every pattern. This causes
2717 it to show the compiled bytecode, without the additional information that
2718 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2719 is the same as /B/I).
2720
2721 23. A new optimization is now able automatically to treat some sequences such
2722 as a*b as a*+b. More specifically, if something simple (such as a character
2723 or a simple class like \d) has an unlimited quantifier, and is followed by
2724 something that cannot possibly match the quantified thing, the quantifier
2725 is automatically "possessified".
2726
2727 24. A recursive reference to a subpattern whose number was greater than 39
2728 went wrong under certain circumstances in UTF-8 mode. This bug could also
2729 have affected the operation of pcre_study().
2730
2731 25. Realized that a little bit of performance could be had by replacing
2732 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2733
2734 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2735
2736 27. Possessive quantifiers such as a++ were previously implemented by turning
2737 them into atomic groups such as ($>a+). Now they have their own opcodes,
2738 which improves performance. This includes the automatically created ones
2739 from 23 above.
2740
2741 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2742 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2743 the first matched character to be a colon. This applied both to named and
2744 numbered groups.
2745
2746 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2747
2748 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2749 building PCRE on Windows easier, so I have included it in the distribution.
2750
2751 31. There is now a check in pcretest against a ridiculously large number being
2752 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2753 loop, the loop is abandoned.
2754
2755 32. Forward references to subpatterns in conditions such as (?(2)...) where
2756 subpattern 2 is defined later cause pcre_compile() to search forwards in
2757 the pattern for the relevant set of parentheses. This search went wrong
2758 when there were unescaped parentheses in a character class, parentheses
2759 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2760
2761 33. "Subroutine" calls and backreferences were previously restricted to
2762 referencing subpatterns earlier in the regex. This restriction has now
2763 been removed.
2764
2765 34. Added a number of extra features that are going to be in Perl 5.10. On the
2766 whole, these are just syntactic alternatives for features that PCRE had
2767 previously implemented using the Python syntax or my own invention. The
2768 other formats are all retained for compatibility.
2769
2770 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2771 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2772 also .NET compatible.
2773
2774 (b) A recursion or subroutine call to a named group can now be defined as
2775 (?&name) as well as (?P>name).
2776
2777 (c) A backreference to a named group can now be defined as \k<name> or
2778 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2779 5.10, are also .NET compatible.
2780
2781 (d) A conditional reference to a named group can now use the syntax
2782 (?(<name>) or (?('name') as well as (?(name).
2783
2784 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2785 groups (named and numbered) that are never evaluated inline, but can be
2786 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2787 is always false. There may be only one alternative in such a group.
2788
2789 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2790 as the simple (?(R). The condition is true only if the most recent
2791 recursion is that of the given number or name. It does not search out
2792 through the entire recursion stack.
2793
2794 (g) The escape \gN or \g{N} has been added, where N is a positive or
2795 negative number, specifying an absolute or relative reference.
2796
2797 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2798 some "unreachable code" warnings.
2799
2800 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2801 things, this adds five new scripts.
2802
2803 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2804 There were also incompatibilities regarding the handling of \Q..\E inside
2805 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2806 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2807
2808 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2809 matches an empty string, and forcibly breaks the loop. There were bugs in
2810 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2811 against aaaa the result was just "a" rather than "aaaa", for example. Two
2812 separate and independent bugs (that affected different cases) have been
2813 fixed.
2814
2815 39. Refactored the code to abolish the use of different opcodes for small
2816 capturing bracket numbers. This is a tidy that I avoided doing when I
2817 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2818 The new approach is not only tidier, it makes it possible to reduce the
2819 memory needed to fix the previous bug (38).
2820
2821 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2822 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2823 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2824 mode.
2825
2826 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2827 report.
2828
2829 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2830 copy construction and assignment in the C++ wrapper.
2831
2832 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2833 couple of bugs that could have given wrong results in the "--newline=crlf"
2834 case.
2835
2836 44. Added a number of casts and did some reorganization of signed/unsigned int
2837 variables following suggestions from Dair Grant. Also renamed the variable
2838 "this" as "item" because it is a C++ keyword.
2839
2840 45. Arranged for dftables to add
2841
2842 #include "pcre_internal.h"
2843
2844 to pcre_chartables.c because without it, gcc 4.x may remove the array
2845 definition from the final binary if PCRE is built into a static library and
2846 dead code stripping is activated.
2847
2848 46. For an unanchored pattern, if a match attempt fails at the start of a
2849 newline sequence, and the newline setting is CRLF or ANY, and the next two
2850 characters are CRLF, advance by two characters instead of one.
2851
2852
2853 Version 6.7 04-Jul-06
2854 ---------------------
2855
2856 1. In order to handle tests when input lines are enormously long, pcretest has
2857 been re-factored so that it automatically extends its buffers when
2858 necessary. The code is crude, but this _is_ just a test program. The
2859 default size has been increased from 32K to 50K.
2860
2861 2. The code in pcre_study() was using the value of the re argument before
2862 testing it for NULL. (Of course, in any sensible call of the function, it
2863 won't be NULL.)
2864
2865 3. The memmove() emulation function in pcre_internal.h, which is used on
2866 systems that lack both memmove() and bcopy() - that is, hardly ever -
2867 was missing a "static" storage class specifier.
2868
2869 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2870 containing an extended class (one that cannot be represented by a bitmap
2871 because it contains high-valued characters or Unicode property items, e.g.
2872 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2873 pattern, but PCRE should not loop if you do not (it no longer does).
2874 [Detail: two cases were found: (a) a repeated subpattern containing an
2875 extended class; (b) a recursive reference to a subpattern that followed a
2876 previous extended class. It wasn't skipping over the extended class
2877 correctly when UTF-8 mode was not set.]
2878
2879 5. A negated single-character class was not being recognized as fixed-length
2880 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2881 compile error "lookbehind assertion is not fixed length".
2882
2883 6. The RunPerlTest auxiliary script was showing an unexpected difference
2884 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2885 write a Perl script that can interpret lines of an input file either as
2886 byte characters or as UTF-8, which is what "perltest" was being required to
2887 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2888 can't do is switch easily at run time between having the "use utf8;" pragma
2889 or not. In the end, I fudged it by using the RunPerlTest script to insert
2890 "use utf8;" explicitly for the UTF-8 tests.
2891
2892 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2893 the end of the subject string, contrary to the documentation and to what
2894 Perl does. This was true of both matching functions. Now it matches only at
2895 the start of the subject and immediately after *internal* newlines.
2896
2897 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2898 a pointer to an int instead of a pointer to an unsigned long int. This
2899 caused problems on 64-bit systems.
2900
2901 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2902 instance of the 'standard' template library not being so standard".
2903
2904 10. There was no check on the number of named subpatterns nor the maximum
2905 length of a subpattern name. The product of these values is used to compute
2906 the size of the memory block for a compiled pattern. By supplying a very
2907 long subpattern name and a large number of named subpatterns, the size
2908 computation could be caused to overflow. This is now prevented by limiting
2909 the length of names to 32 characters, and the number of named subpatterns
2910 to 10,000.
2911
2912 11. Subpatterns that are repeated with specific counts have to be replicated in
2913 the compiled pattern. The size of memory for this was computed from the
2914 length of the subpattern and the repeat count. The latter is limited to
2915 65535, but there was no limit on the former, meaning that integer overflow
2916 could in principle occur. The compiled length of a repeated subpattern is
2917 now limited to 30,000 bytes in order to prevent this.
2918
2919 12. Added the optional facility to have named substrings with the same name.
2920
2921 13. Added the ability to use a named substring as a condition, using the
2922 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2923 are numbers (not recommended). Forward references are permitted.
2924
2925 14. Added forward references in named backreferences (if you see what I mean).
2926
2927 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2928 pattern could run off the end of the subject. For example, the pattern
2929 "(?s)(.{1,5})"8 did this with the subject "ab".
2930
2931 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2932 PCRE_CASELESS was set when matching characters that were quantified with ?
2933 or *.
2934
2935 17. A character class other than a single negated character that had a minimum
2936 but no maximum quantifier - for example [ab]{6,} - was not handled
2937 correctly by pce_dfa_exec(). It would match only one character.
2938
2939 18. A valid (though odd) pattern that looked like a POSIX character
2940 class but used an invalid character after [ (for example [[,abc,]]) caused
2941 pcre_compile() to give the error "Failed: internal error: code overflow" or
2942 in some cases to crash with a glibc free() error. This could even happen if
2943 the pattern terminated after [[ but there just happened to be a sequence of
2944 letters, a binary zero, and a closing ] in the memory that followed.
2945
2946 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2947 over the years. Originally (before any Unicode support), just the bottom 8
2948 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2949 output from "man perlunicode" includes this:
2950
2951 The regular expression compiler produces polymorphic opcodes. That
2952 is, the pattern adapts to the data and automatically switches to
2953 the Unicode character scheme when presented with Unicode data--or
2954 instead uses a traditional byte scheme when presented with byte
2955 data.
2956
2957 Sadly, a wide octal escape does not cause a switch, and in a string with
2958 no other multibyte characters, these octal escapes are treated as before.
2959 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2960 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2961 Unicode string.
2962
2963 I have not perpetrated such confusion in PCRE. Up till now, it took just
2964 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2965 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2966 translate to the appropriate multibyte character.
2967
2968 29. Applied some refactoring to reduce the number of warnings from Microsoft
2969 and Borland compilers. This has included removing the fudge introduced
2970 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2971 a warning about an unused variable.
2972
2973 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2974 characters since release 4.0, because Perl (from release 5.004) does not.
2975 [Or at least, is documented not to: some releases seem to be in conflict
2976 with the documentation.] However, when a pattern was studied with
2977 pcre_study() and all its branches started with \s, PCRE still included VT
2978 as a possible starting character. Of course, this did no harm; it just
2979 caused an unnecessary match attempt.
2980
2981 22. Removed a now-redundant internal flag bit that recorded the fact that case
2982 dependency changed within the pattern. This was once needed for "required
2983 byte" processing, but is no longer used. This recovers a now-scarce options
2984 bit. Also moved the least significant internal flag bit to the most-
2985 significant bit of the word, which was not previously used (hangover from
2986 the days when it was an int rather than a uint) to free up another bit for
2987 the future.
2988
2989 23. Added support for CRLF line endings as well as CR and LF. As well as the
2990 default being selectable at build time, it can now be changed at runtime
2991 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2992 specify that it is scanning data with non-default line endings.
2993
2994 24. Changed the definition of CXXLINK to make it agree with the definition of
2995 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2996
2997 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2998 recursions. This makes a big different to stack usage for some patterns.
2999
3000 26. If a subpattern containing a named recursion or subroutine reference such
3001 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3002 the space required for the compiled pattern went wrong and gave too small a
3003 value. Depending on the environment, this could lead to "Failed: internal
3004 error: code overflow at offset 49" or "glibc detected double free or
3005 corruption" errors.
3006
3007 27. Applied patches from Google (a) to support the new newline modes and (b) to
3008 advance over multibyte UTF-8 characters in GlobalReplace.
3009
3010 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3011 difference for some implementation of PCRE in some Windows version.
3012
3013 29. Added some extra testing facilities to pcretest:
3014
3015 \q<number> in a data line sets the "match limit" value
3016 \Q<number> in a data line sets the "match recursion limt" value
3017 -S <number> sets the stack size, where <number> is in megabytes
3018
3019 The -S option isn't available for Windows.
3020
3021
3022 Version 6.6 06-Feb-06
3023 ---------------------
3024
3025 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3026 in pcreposix.h. I have copied the definition from pcre.h.
3027
3028 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3029 because pcre.h is no longer a built file.
3030
3031 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3032 not normally included in the compiled code.
3033
3034
3035 Version 6.5 01-Feb-06
3036 ---------------------
3037
3038 1. When using the partial match feature with pcre_dfa_exec(), it was not
3039 anchoring the second and subsequent partial matches at the new starting
3040 point. This could lead to incorrect results. For example, with the pattern
3041 /1234/, partially matching against "123" and then "a4" gave a match.
3042
3043 2. Changes to pcregrep:
3044
3045 (a) All non-match returns from pcre_exec() were being treated as failures
3046 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3047 error message is output. Some extra information is given for the
3048 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3049 probably the only errors that are likely to be caused by users (by
3050 specifying a regex that has nested indefinite repeats, for instance).
3051 If there are more than 20 of these errors, pcregrep is abandoned.
3052
3053 (b) A binary zero was treated as data while matching, but terminated the
3054 output line if it was written out. This has been fixed: binary zeroes
3055 are now no different to any other data bytes.
3056
3057 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3058 used to set a locale for matching. The --locale=xxxx long option has
3059 been added (no short equivalent) to specify a locale explicitly on the
3060 pcregrep command, overriding the environment variables.
3061
3062 (d) When -B was used with -n, some line numbers in the output were one less
3063 than they should have been.
3064
3065 (e) Added the -o (--only-matching) option.
3066
3067 (f) If -A or -C was used with -c (count only), some lines of context were
3068 accidentally printed for the final match.
3069
3070 (g) Added the -H (--with-filename) option.
3071
3072 (h) The combination of options -rh failed to suppress file names for files
3073 that were found from directory arguments.
3074
3075 (i) Added the -D (--devices) and -d (--directories) options.
3076
3077 (j) Added the -F (--fixed-strings) option.
3078
3079 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3080
3081 (l) Added the --colo(u)r option.
3082
3083 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3084 is not present by default.
3085
3086 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3087 items such as (?R) or (?1), when the recursion could match a number of
3088 alternatives. If it matched one of the alternatives, but subsequently,
3089 outside the recursion, there was a failure, the code tried to back up into
3090 the recursion. However, because of the way PCRE is implemented, this is not
3091 possible, and the result was an incorrect result from the match.
3092
3093 In order to prevent this happening, the specification of recursion has
3094 been changed so that all such subpatterns are automatically treated as
3095 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3096
3097 4. I had overlooked the fact that, in some locales, there are characters for
3098 which isalpha() is true but neither isupper() nor islower() are true. In
3099 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3100 and ordfeminine) are like this. This affected the treatment of \w and \W
3101 when they appeared in character classes, but not when they appeared outside
3102 a character class. The bit map for "word" characters is now created
3103 separately from the results of isalnum() instead of just taking it from the
3104 upper, lower, and digit maps. (Plus the underscore character, of course.)
3105
3106 5. The above bug also affected the handling of POSIX character classes such as
3107 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3108 permanent tables. Instead, the bit maps for such a class were previously
3109 created as the appropriate unions of the upper, lower, and digit bitmaps.
3110 Now they are created by subtraction from the [[:word:]] class, which has
3111 its own bitmap.
3112
3113 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3114 It is created by subtracting the vertical space characters (\x09, \x0a,
3115 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3116 subtraction was done in the overall bitmap for a character class, meaning
3117 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3118 be recognized. This bug has been fixed.
3119
3120 7. Patches from the folks at Google:
3121
3122 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3123 real life, but is still worth protecting against".
3124
3125 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3126 regular expressions".
3127
3128 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3129 have it.
3130
3131 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3132 "configure" and the latter not, in order to fix a problem somebody had
3133 with compiling the Arg class on HP-UX.
3134
3135 (e) Improve the error-handling of the C++ wrapper a little bit.
3136
3137 (f) New tests for checking recursion limiting.
3138
3139 8. The pcre_memmove() function, which is used only if the environment does not
3140 have a standard memmove() function (and is therefore rarely compiled),
3141 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3142 returning a result (though PCRE never actually uses the result).
3143
3144 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3145 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3146 returned instead of calling malloc() with an overflowing number that would
3147 most likely cause subsequent chaos.
3148
3149 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3150
3151 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3152 with this option is matched, the nmatch and pmatch options of regexec() are
3153 ignored.
3154
3155 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3156 provided in case anyone wants to the the POSIX interface with UTF-8
3157 strings.
3158
3159 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3160 C++ linking (needed for some HP-UX environments).
3161
3162 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3163 (unused parameter) and in the pcre_printint() function (omitted "default"
3164 switch label when the default is to do nothing).
3165
3166 15. Added some code to make it possible, when PCRE is compiled as a C++
3167 library, to replace subject pointers for pcre_exec() with a smart pointer
3168 class, thus making it possible to process discontinuous strings.
3169
3170 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3171 much the same function. They were added by different people who were trying
3172 to make PCRE easy to compile on non-Unix systems. It has been suggested
3173 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3174 for compiling on Windows systems. I have therefore replaced it with
3175 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3176 defaults to "extern" for C or "extern C" for C++, which works fine on
3177 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3178 SCOPE with something explicit in config.h. In addition:
3179
3180 (a) pcreposix.h still had just "extern" instead of either of these macros;
3181 I have replaced it with PCRE_DATA_SCOPE.
3182
3183 (b) Functions such as _pcre_xclass(), which are internal to the library,
3184 but external in the C sense, all had PCRE_EXPORT in their definitions.
3185 This is apparently wrong for the Windows case, so I have removed it.
3186 (It makes no difference on Unix-like systems.)
3187
3188 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3189 of recursive calls to match(). This is different to MATCH_LIMIT because
3190 that limits the total number of calls to match(), not all of which increase
3191 the depth of recursion. Limiting the recursion depth limits the amount of
3192 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3193 when PCRE is compiled, and changed at run time. A patch from Google adds
3194 this functionality to the C++ interface.
3195
3196 18. Changes to the handling of Unicode character properties:
3197
3198 (a) Updated the table to Unicode 4.1.0.
3199
3200 (b) Recognize characters that are not in the table as "Cn" (undefined).
3201
3202 (c) I revised the way the table is implemented to a much improved format
3203 which includes recognition of ranges. It now supports the ranges that
3204 are defined in UnicodeData.txt, and it also amalgamates other
3205 characters into ranges. This has reduced the number of entries in the
3206 table from around 16,000 to around 3,000, thus reducing its size
3207 considerably. I realized I did not need to use a tree structure after
3208 all - a binary chop search is just as efficient. Having reduced the
3209 number of entries, I extended their size from 6 bytes to 8 bytes to
3210 allow for more data.
3211
3212 (d) Added support for Unicode script names via properties such as \p{Han}.
3213
3214 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3215 matching that character.
3216
3217 20. When matching a repeated Unicode property with a minimum greater than zero,
3218 (for example \pL{2,}), PCRE could look past the end of the subject if it
3219 reached it while seeking the minimum number of characters. This could
3220 happen only if some of the characters were more than one byte long, because
3221 there is a check for at least the minimum number of bytes.
3222
3223 21. Refactored the implementation of \p and \P so as to be more general, to
3224 allow for more different types of property in future. This has changed the
3225 compiled form incompatibly. Anybody with saved compiled patterns that use
3226 \p or \P will have to recompile them.
3227
3228 22. Added "Any" and "L&" to the supported property types.
3229
3230 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3231 but give a compile time error if the value is greater than 0xff.
3232
3233 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3234 accidentally not being installed or uninstalled.
3235
3236 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3237 made were to insert the current release number. This seemed silly, because
3238 it made things harder for people building PCRE on systems that don't run
3239 "configure". I have turned pcre.h into a distributed file, no longer built
3240 by "configure", with the version identification directly included. There is
3241 no longer a pcre.h.in file.
3242
3243 However, this change necessitated a change to the pcre-config script as
3244 well. It is built from pcre-config.in, and one of the substitutions was the
3245 release number. I have updated configure.ac so that ./configure now finds
3246 the release number by grepping pcre.h.
3247
3248 26. Added the ability to run the tests under valgrind.
3249
3250
3251 Version 6.4 05-Sep-05
3252 ---------------------
3253
3254 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3255 "--" to be printed when multiple files were scanned, even when none of the
3256 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3257 consider it to be a bug, and have restored the previous behaviour.
3258
3259 2. A couple of code tidies to get rid of compiler warnings.
3260
3261 3. The pcretest program used to cheat by referring to symbols in the library
3262 whose names begin with _pcre_. These are internal symbols that are not
3263 really supposed to be visible externally, and in some environments it is
3264 possible to suppress them. The cheating is now confined to including
3265 certain files from the library's source, which is a bit cleaner.
3266
3267 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3268 file's purpose clearer.
3269
3270 5. Reorganized pcre_ucp_findchar().
3271
3272
3273 Version 6.3 15-Aug-05
3274 ---------------------
3275
3276 1. The file libpcre.pc.in did not have general read permission in the tarball.
3277
3278 2. There were some problems when building without C++ support:
3279
3280 (a) If C++ support was not built, "make install" and "make test" still
3281 tried to test it.
3282
3283 (b) There were problems when the value of CXX was explicitly set. Some
3284 changes have been made to try to fix these, and ...
3285
3286 (c) --disable-cpp can now be used to explicitly disable C++ support.
3287
3288 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3289 backslash in a target when C++ was disabled. This confuses some
3290 versions of "make", apparently. Using an intermediate variable solves
3291 this. (Same for CPP_LOBJ.)
3292
3293 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3294 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3295 necessary on certain architectures.
3296
3297 4. Added a setting of -export-symbols-regex to the link command to remove
3298 those symbols that are exported in the C sense, but actually are local
3299 within the library, and not documented. Their names all begin with
3300 "_pcre_". This is not a perfect job, because (a) we have to except some
3301 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3302 available (and never for static libraries). I have made a note to try to
3303 find a way round (a) in the future.
3304
3305
3306 Version 6.2 01-Aug-05
3307 ---------------------
3308
3309 1. There was no test for integer overflow of quantifier values. A construction
3310 such as {1111111111111111} would give undefined results. What is worse, if
3311 a minimum quantifier for a parenthesized subpattern overflowed and became
3312 negative, the calculation of the memory size went wrong. This could have
3313 led to memory overwriting.
3314
3315 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3316
3317 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3318 operating environments where this matters.
3319
3320 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3321 PCRE options from within the C++ wrapper.
3322
3323 5. Named capturing subpatterns were not being correctly counted when a pattern
3324 was compiled. This caused two problems: (a) If there were more than 100
3325 such subpatterns, the calculation of the memory needed for the whole
3326 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3327 back references of the form \12, where the number was greater than 9, were
3328 not recognized as back references, even though there were sufficient
3329 previous subpatterns.
3330
3331 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3332 versions of gcc, e.g. 2.95.4.
3333
3334
3335 Version 6.1 21-Jun-05
3336 ---------------------
3337
3338 1. There was one reference to the variable "posix" in pcretest.c that was not
3339 surrounded by "#if !defined NOPOSIX".
3340
3341 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3342 the cross-check on the old pcre_info() function, for the benefit of the
3343 cut-down version of PCRE that is currently imported into Exim.
3344
3345 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3346 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3347 patterns that start (?i) but I don't think that matters. The use of (?i) is
3348 just an example; this all applies to the other options as well.
3349
3350 4. Since libtool seems to echo the compile commands it is issuing, the output
3351 from "make" can be reduced a bit by putting "@" in front of each libtool
3352 compile command.
3353
3354 5. Patch from the folks at Google for configure.in to be a bit more thorough
3355 in checking for a suitable C++ installation before trying to compile the
3356 C++ stuff. This should fix a reported problem when a compiler was present,
3357 but no suitable headers.
3358
3359 6. The man pages all had just "PCRE" as their title. I have changed them to
3360 be the relevant file name. I have also arranged that these names are
3361 retained in the file doc/pcre.txt, which is a concatenation in text format
3362 of all the man pages except the little individual ones for each function.
3363
3364 7. The NON-UNIX-USE file had not been updated for the different set of source
3365 files that come with release 6. I also added a few comments about the C++
3366 wrapper.
3367
3368
3369 Version 6.0 07-Jun-05
3370 ---------------------
3371
3372 1. Some minor internal re-organization to help with my DFA experiments.
3373
3374 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3375 didn't matter for the library itself when fully configured, but did matter
3376 when compiling without UCP support, or within Exim, where the ucp files are
3377 not imported.
3378
3379 3. Refactoring of the library code to split up the various functions into
3380 different source modules. The addition of the new DFA matching code (see
3381 below) to a single monolithic source would have made it really too
3382 unwieldy, quite apart from causing all the code to be include in a
3383 statically linked application, when only some functions are used. This is
3384 relevant even without the DFA addition now that patterns can be compiled in
3385 one application and matched in another.
3386
3387 The downside of splitting up is that there have to be some external
3388 functions and data tables that are used internally in different modules of
3389 the library but which are not part of the API. These have all had their
3390 names changed to start with "_pcre_" so that they are unlikely to clash
3391 with other external names.
3392
3393 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3394 a different (DFA) algorithm. Although it is slower than the original
3395 function, it does have some advantages for certain types of matching
3396 problem.
3397
3398 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3399 including restarting after a partial match.
3400
3401 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3402 defined when compiling for Windows was sent to me. I have put it into the
3403 code, though I have no means of testing or verifying it.
3404
3405 7. Added the pcre_refcount() auxiliary function.
3406
3407 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3408 match before or at the first newline in the subject string. In pcretest,
3409 the /f option on a pattern can be used to set this.
3410
3411 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3412 would behave wrongly. This has been present in PCRE since release 4.0.
3413
3414 10. A number of changes to the pcregrep command:
3415
3416 (a) Refactored how -x works; insert ^(...)$ instead of setting
3417 PCRE_ANCHORED and checking the length, in preparation for adding
3418 something similar for -w.
3419
3420 (b) Added the -w (match as a word) option.
3421
3422 (c) Refactored the way lines are read and buffered so as to have more
3423 than one at a time available.
3424
3425 (d) Implemented a pcregrep test script.
3426
3427 (e) Added the -M (multiline match) option. This allows patterns to match
3428 over several lines of the subject. The buffering ensures that at least
3429 8K, or the rest of the document (whichever is the shorter) is available
3430 for matching (and similarly the previous 8K for lookbehind assertions).
3431
3432 (f) Changed the --help output so that it now says
3433
3434 -w, --word-regex(p)
3435
3436 instead of two lines, one with "regex" and the other with "regexp"
3437 because that confused at least one person since the short forms are the
3438 same. (This required a bit of code, as the output is generated
3439 automatically from a table. It wasn't just a text change.)
3440
3441 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3442 option but starts with a hyphen. Could be a pattern or a path name
3443 starting with a hyphen, for instance.
3444
3445 (h) "-" can be given as a file name to represent stdin.
3446
3447 (i) When file names are being printed, "(standard input)" is used for
3448 the standard input, for compatibility with GNU grep. Previously
3449 "<stdin>" was used.
3450
3451 (j) The option --label=xxx can be used to supply a name to be used for
3452 stdin when file names are being printed. There is no short form.
3453
3454 (k) Re-factored the options decoding logic because we are going to add
3455 two more options that take data. Such options can now be given in four
3456 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3457
3458 (l) Added the -A, -B, and -C options for requesting that lines of context
3459 around matches be printed.
3460
3461 (m) Added the -L option to print the names of files that do not contain
3462 any matching lines, that is, the complement of -l.
3463
3464 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3465 continue to scan other files.
3466
3467 (o) The -s option was incorrectly implemented. For compatibility with other
3468 greps, it now suppresses the error message for a non-existent or non-
3469 accessible file (but not the return code). There is a new option called
3470 -q that suppresses the output of matching lines, which was what -s was
3471 previously doing.
3472
3473 (p) Added --include and --exclude options to specify files for inclusion
3474 and exclusion when recursing.
3475
3476 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3477 Hopefully, it now does.
3478
3479 12. Missing cast in pcre_study().
3480
3481 13. Added an "uninstall" target to the makefile.
3482
3483 14. Replaced "extern" in the function prototypes in Makefile.in with
3484 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3485 world, but is set differently for Windows.
3486
3487 15. Added a second compiling function called pcre_compile2(). The only
3488 difference is that it has an extra argument, which is a pointer to an
3489 integer error code. When there is a compile-time failure, this is set
3490 non-zero, in addition to the error test pointer being set to point to an
3491 error message. The new argument may be NULL if no error number is required
3492 (but then you may as well call pcre_compile(), which is now just a
3493 wrapper). This facility is provided because some applications need a
3494 numeric error indication, but it has also enabled me to tidy up the way
3495 compile-time errors are handled in the POSIX wrapper.
3496
3497 16. Added VPATH=.libs to the makefile; this should help when building with one
3498 prefix path and installing with another. (Or so I'm told by someone who
3499 knows more about this stuff than I do.)
3500
3501 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3502 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3503 match everything, including newlines. This is not POSIX-compatible, but
3504 somebody wanted the feature. From pcretest it can be activated by using
3505 both the P and the s flags.
3506
3507 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3508
3509 19. libpcre.pc was being incorrectly installed as executable.
3510
3511 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3512 it now also looks for '\r' so that it will work unmodified on Windows.
3513
3514 21. Added Google's contributed C++ wrapper to the distribution.
3515
3516 22. Added some untidy missing memory free() calls in pcretest, to keep
3517 Electric Fence happy when testing.
3518
3519
3520
3521 Version 5.0 13-Sep-04
3522 ---------------------
3523
3524 1. Internal change: literal characters are no longer packed up into items
3525 containing multiple characters in a single byte-string. Each character
3526 is now matched using a separate opcode. However, there may be more than one
3527 byte in the character in UTF-8 mode.
3528
3529 2. The pcre_callout_block structure has two new fields: pattern_position and
3530 next_item_length. These contain the offset in the pattern to the next match
3531 item, and its length, respectively.
3532
3533 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3534 insertion of callouts before each pattern item. Added the /C option to
3535 pcretest to make use of this.
3536
3537 4. On the advice of a Windows user, the lines
3538
3539 #if defined(_WIN32) || defined(WIN32)
3540 _setmode( _fileno( stdout ), 0x8000 );
3541 #endif /* defined(_WIN32) || defined(WIN32) */
3542
3543 have been added to the source of pcretest. This apparently does useful
3544 magic in relation to line terminators.
3545
3546 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3547 for the benefit of those environments where the "b" makes a difference.
3548
3549 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3550 to know about it. I have put a hack into configure.in that adds in code
3551 to set GCC=yes if CC=icc. This seems to end up at a point in the
3552 generated configure script that is early enough to affect the setting of
3553 compiler options, which is what is needed, but I have no means of testing
3554 whether it really works. (The user who reported this had patched the
3555 generated configure script, which of course I cannot do.)
3556
3557 LATER: After change 22 below (new libtool files), the configure script
3558 seems to know about icc (and also ecc). Therefore, I have commented out
3559 this hack in configure.in.
3560
3561 7. Added support for pkg-config (2 patches were sent in).
3562
3563 8. Negated POSIX character classes that used a combination of internal tables
3564 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3565 [[:^ascii]]. Typically, they would match almost any characters. The other
3566 POSIX classes were not broken in this way.
3567
3568 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3569 to find the match, as PCRE was deluded into thinking that the match had to
3570 start at the start point or following a newline. The same bug applied to
3571 patterns with negative forward assertions or any backward assertions
3572 preceding ".*" at the start, unless the pattern required a fixed first
3573 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3574
3575 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3576 starting at the last subject character, bytes beyond the end of the subject
3577 string were read.
3578
3579 11. Renamed the variable "class" as "classbits" to make life easier for C++
3580 users. (Previously there was a macro definition, but it apparently wasn't
3581 enough.)
3582
3583 12. Added the new field "tables" to the extra data so that tables can be passed
3584 in at exec time, or the internal tables can be re-selected. This allows
3585 a compiled regex to be saved and re-used at a later time by a different
3586 program that might have everything at different addresses.
3587
3588 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3589 -R library as well as a -L library.
3590
3591 14. The debugging options of pcretest (-d on the command line or D on a
3592 pattern) showed incorrect output for anything following an extended class
3593 that contained multibyte characters and which was followed by a quantifier.
3594
3595 15. Added optional support for general category Unicode character properties
3596 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3597 support. It adds about 90K to the size of the library. The meanings of the
3598 inbuilt class escapes such as \d and \s have NOT been changed.
3599
3600 16. Updated pcredemo.c to include calls to free() to release the memory for the
3601 compiled pattern.
3602
3603 17. The generated file chartables.c was being created in the source directory
3604 instead of in the building directory. This caused the build to fail if the
3605 source directory was different from the building directory, and was
3606 read-only.
3607
3608 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3609 file. No doubt somebody will tell me if they don't make sense... Also added
3610 Dan Mooney's comments about building on OpenVMS.
3611
3612 19. Added support for partial matching via the PCRE_PARTIAL option for
3613 pcre_exec() and the \P data escape in pcretest.
3614
3615 20. Extended pcretest with 3 new pattern features:
3616
3617 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3618 write the compiled pattern to the file whose name is "rest-of-line".
3619 This is a straight binary dump of the data, with the saved pointer to
3620 the character tables forced to be NULL. The study data, if any, is
3621 written too. After writing, pcretest reads a new pattern.
3622
3623 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3624 compiled pattern from the given file. There must not be any
3625 occurrences of "<" in the file name (pretty unlikely); if there are,
3626 pcretest will instead treat the initial "<" as a pattern delimiter.
3627 After reading in the pattern, pcretest goes on to read data lines as
3628 usual.
3629
3630 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3631 and 16-bit fields in a compiled pattern, to simulate a pattern that
3632 was compiled on a host of opposite endianness.
3633
3634 21. The pcre-exec() function can now cope with patterns that were compiled on
3635 hosts of opposite endianness, with this restriction:
3636
3637 As for any compiled expression that is saved and used later, the tables
3638 pointer field cannot be preserved; the extra_data field in the arguments
3639 to pcre_exec() should be used to pass in a tables address if a value
3640 other than the default internal tables were used at compile time.
3641
3642 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3643 now diagnosed as an error. Previously, most of the time, a negative number
3644 would have been treated as zero, but if in addition "ovector" was passed as
3645 NULL, a crash could occur.
3646
3647 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3648 new versions from the libtool 1.5 distribution (the last one is a copy of
3649 a file called libtool.m4). This seems to have fixed the need to patch
3650 "configure" to support Darwin 1.3 (which I used to do). However, I still
3651 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3652 workstation).
3653
3654 24. Changed the PCRE licence to be the more standard "BSD" licence.
3655
3656
3657 Version 4.5 01-Dec-03
3658 ---------------------
3659
3660 1. There has been some re-arrangement of the code for the match() function so
3661 that it can be compiled in a version that does not call itself recursively.
3662 Instead, it keeps those local variables that need separate instances for
3663 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3664 needs to "recurse". Keeping track of where control must go is done by means
3665 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3666 hide most of the details from the main code, and operates only if
3667 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3668 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3669 operating.
3670
3671 To make it easier for callers to provide specially tailored get/free
3672 functions for this usage, two new functions, pcre_stack_malloc, and
3673 pcre_stack_free, are used. They are always called in strict stacking order,
3674 and the size of block requested is always the same.
3675
3676 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3677 PCRE has been compiled to use the stack or the heap for recursion. The
3678 -C option of pcretest uses this to show which version is compiled.
3679
3680 A new data escape \S, is added to pcretest; it causes the amounts of store
3681 obtained and freed by both kinds of malloc/free at match time to be added
3682 to the output.
3683
3684 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3685 what's available on my current Linux desktop machine.
3686
3687 3. When matching a UTF-8 string, the test for a valid string at the start has
3688 been extended. If start_offset is not zero, PCRE now checks that it points
3689 to a byte that is the start of a UTF-8 character. If not, it returns
3690 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3691 this is necessary because there may be backward assertions in the pattern.
3692 When matching the same subject several times, it may save resources to use
3693 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3694
3695 4. The code for checking the validity of UTF-8 strings has been tightened so
3696 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3697 containing "overlong sequences".
3698
3699 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3700 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3701 should have been "&", but it just so happened that all the cases this let
3702 through by mistake were picked up later in the function.
3703
3704 6. I had used a variable called "isblank" - this is a C99 function, causing
3705 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3706
3707 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3708 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3709 so the version is shown; (c) stop "make test" echoing "./RunTest".
3710
3711 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3712
3713 9. The prototype for memmove() for systems that don't have it was using
3714 size_t, but the inclusion of the header that defines size_t was later. I've
3715 moved the #includes for the C headers earlier to avoid this.
3716
3717 10. Added some adjustments to the code to make it easier to compiler on certain
3718 special systems:
3719
3720 (a) Some "const" qualifiers were missing.
3721 (b) Added the macro EXPORT before all exported functions; by default this
3722 is defined to be empty.
3723 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3724 that it reads its output file name as an argument instead of writing
3725 to the standard output and assuming this can be redirected.
3726
3727 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3728 class containing characters with values greater than 255, PCRE compilation
3729 went into a loop.
3730
3731 12. A recursive reference to a subpattern that was within another subpattern
3732 that had a minimum quantifier of zero caused PCRE to crash. For example,
3733 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3734 recursion. If the recursively-called subpattern itself had a zero repeat,
3735 that was OK.
3736
3737 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3738 buffer into which it was copied (for escape processing) was still set at
3739 1024, so long lines caused crashes.
3740
3741 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3742 "internal error: code overflow...". This applied to any character class
3743 that was followed by a possessive quantifier.
3744
3745 15. Modified the Makefile to add libpcre.la as a prerequisite for
3746 libpcreposix.la because I was told this is needed for a parallel build to
3747 work.
3748
3749 16. If a pattern that contained .* following optional items at the start was
3750 studied, the wrong optimizing data was generated, leading to matching
3751 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3752 matching string must start with a or b or c. The correct conclusion for
3753 this pattern is that a match can start with any character.
3754
3755
3756 Version 4.4 13-Aug-03
3757 ---------------------
3758
3759 1. In UTF-8 mode, a character class containing characters with values between
3760 127 and 255 was not handled correctly if the compiled pattern was studied.
3761 In fixing this, I have also improved the studying algorithm for such
3762 classes (slightly).
3763
3764 2. Three internal functions had redundant arguments passed to them. Removal
3765 might give a very teeny performance improvement.
3766
3767 3. Documentation bug: the value of the capture_top field in a callout is *one
3768 more than* the number of the hightest numbered captured substring.
3769
3770 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3771 in incorrectly linking with a previously installed version. They now link
3772 explicitly with libpcre.la.
3773
3774 5. configure.in no longer needs to recognize Cygwin specially.
3775
3776 6. A problem in pcre.in for Windows platforms is fixed.
3777
3778 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3779 pcretest, it used to include the size of the study block as part of its
3780 output. Unfortunately, the structure contains a field that has a different
3781 size on different hardware architectures. This meant that the tests that
3782 showed this size failed. As the block is currently always of a fixed size,
3783 this information isn't actually particularly useful in pcretest output, so
3784 I have just removed it.
3785
3786 8. Three pre-processor statements accidentally did not start in column 1.
3787 Sadly, there are *still* compilers around that complain, even though
3788 standard C has not required this for well over a decade. Sigh.
3789
3790 9. In pcretest, the code for checking callouts passed small integers in the
3791 callout_data field, which is a void * field. However, some picky compilers
3792 complained about the casts involved for this on 64-bit systems. Now
3793 pcretest passes the address of the small integer instead, which should get
3794 rid of the warnings.
3795
3796 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3797 both compile and run time, and gives an error if an invalid UTF-8 sequence
3798 is found. There is a option for disabling this check in cases where the
3799 string is known to be correct and/or the maximum performance is wanted.
3800
3801 11. In response to a bug report, I changed one line in Makefile.in from
3802
3803 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3804 to
3805 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3806
3807 to look similar to other lines, but I have no way of telling whether this
3808 is the right thing to do, as I do not use Windows. No doubt I'll get told
3809 if it's wrong...
3810
3811
3812 Version 4.3 21-May-03
3813 ---------------------
3814
3815 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3816 Makefile.
3817
3818 2. Some refactoring to improve the quality of the code:
3819
3820 (i) The utf8_table... variables are now declared "const".
3821
3822 (ii) The code for \cx, which used the "case flipping" table to upper case
3823 lower case letters, now just substracts 32. This is ASCII-specific,
3824 but the whole concept of \cx is ASCII-specific, so it seems
3825 reasonable.
3826
3827 (iii) PCRE was using its character types table to recognize decimal and
3828 hexadecimal digits in the pattern. This is silly, because it handles
3829 only 0-9, a-f, and A-F, but the character types table is locale-
3830 specific, which means strange things might happen. A private
3831 table is now used for this - though it costs 256 bytes, a table is
3832 much faster than multiple explicit tests. Of course, the standard
3833 character types table is still used for matching digits in subject
3834 strings against \d.
3835
3836 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3837 ending in _t are). So I've renamed it as ESC_tee.
3838
3839 3. The first argument for regexec() in the POSIX wrapper should have been
3840 defined as "const".
3841
3842 4. Changed pcretest to use malloc() for its buffers so that they can be
3843 Electric Fenced for debugging.
3844
3845 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3846 to read one or more bytes before the start of the subject string. Often this
3847 had no effect on PCRE's behaviour, but in some circumstances it could
3848 provoke a segmentation fault.
3849
3850 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3851 to try to read one or more bytes before the start of the subject string.
3852
3853 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3854 UTF-8 support could misbehave in various ways if the subject string
3855 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3856 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3857 back over UTF-8 characters.)
3858
3859
3860 Version 4.2 14-Apr-03
3861 ---------------------
3862
3863 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3864
3865 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3866 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3867 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3868 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3869 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3870 and BUILD_EXEEXT
3871 Note: automatic setting of the BUILD variables is not yet working
3872 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3873 compile-time but not at link-time
3874 [LINK]: use for linking executables only
3875 make different versions for Windows and non-Windows
3876 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3877 libraries
3878 [LINK_FOR_BUILD]: new variable
3879 [OBJEXT]: use throughout
3880 [EXEEXT]: use throughout
3881 <winshared>: new target
3882 <wininstall>: new target
3883 <dftables.o>: use native compiler
3884 <dftables>: use native linker
3885 <install>: handle Windows platform correctly
3886 <clean>: ditto
3887 <check>: ditto
3888 copy DLL to top builddir before testing
3889
3890 As part of these changes, -no-undefined was removed again. This was reported
3891 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3892 in any case.
3893
3894 3. Some tidies to get rid of compiler warnings:
3895
3896 . In the match_data structure, match_limit was an unsigned long int, whereas
3897 match_call_count was an int. I've made them both unsigned long ints.
3898
3899 . In pcretest the fact that a const uschar * doesn't automatically cast to
3900 a void * provoked a warning.
3901
3902 . Turning on some more compiler warnings threw up some "shadow" variables
3903 and a few more missing casts.
3904
3905 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3906 option, a class that contained a single character with a value between 128
3907 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3908
3909 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3910 option, a class that contained several characters, but with at least one
3911 whose value was between 128 and 255 caused PCRE to crash.
3912
3913
3914 Version 4.1 12-Mar-03
3915 ---------------------
3916
3917 1. Compiling with gcc -pedantic found a couple of places where casts were
3918 needed, and a string in dftables.c that was longer than standard compilers are
3919 required to support.
3920
3921 2. Compiling with Sun's compiler found a few more places where the code could
3922 be tidied up in order to avoid warnings.
3923
3924 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3925 first of these names is deprecated in the latest Autoconf in favour of the name
3926 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3927 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3928 analogy I have changed it to CFLAGS_FOR_BUILD.
3929
3930 4. Added -no-undefined to the linking command in the Makefile, because this is
3931 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3932 linking step for the pcreposix library.
3933
3934 5. PCRE was failing to diagnose the case of two named groups with the same
3935 name.
3936
3937 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3938 literal character that is needed in the subject for a match, and scans along to
3939 ensure that it is present before embarking on the full matching process. This
3940 saves time in cases of nested unlimited repeats that are never going to match.
3941 Problem: the scan can take a lot of time if the subject is very long (e.g.
3942 megabytes), thus penalizing straightforward matches. It is now done only if the
3943 amount of subject to be scanned is less than 1000 bytes.
3944
3945 7. A lesser problem with the same optimization is that it was recording the
3946 first character of an anchored pattern as "needed", thus provoking a search
3947 right along the subject, even when the first match of the pattern was going to
3948 fail. The "needed" character is now not set for anchored patterns, unless it
3949 follows something in the pattern that is of non-fixed length. Thus, it still
3950 fulfils its original purpose of finding quick non-matches in cases of nested
3951 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3952
3953
3954 Version 4.0 17-Feb-03
3955 ---------------------
3956
3957 1. If a comment in an extended regex that started immediately after a meta-item
3958 extended to the end of string, PCRE compiled incorrect data. This could lead to
3959 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3960
3961 2. Moved to autoconf 2.53 and libtool 1.4.2.
3962
3963 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3964 the special perltest8 script is no longer needed - all the tests can be run
3965 from a single perltest script.
3966
3967 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3968 by \s. It has now been removed in PCRE. This means it isn't recognized as
3969 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3970 class [:space:] *does* include VT, thereby creating a mess.
3971
3972 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3973 space and tab.
3974
3975 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3976 its new features into the main test script, reducing the number of scripts.
3977
3978 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3979 were backward compatible, and made the (?i) apply to the whole pattern, as if
3980 /i were given. Now it behaves more logically, and applies the option setting
3981 only to what follows. PCRE has been changed to follow suit. However, if it
3982 finds options settings right at the start of the pattern, it extracts them into
3983 the global options, as before. Thus, they show up in the info data.
3984
3985 8. Added support for the \Q...\E escape sequence. Characters in between are
3986 treated as literals. This is slightly different from Perl in that $ and @ are
3987 also handled as literals inside the quotes. In Perl, they will cause variable
3988 interpolation. Note the following examples:
3989
3990 Pattern PCRE matches Perl matches
3991
3992 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3993 \Qabc\$xyz\E abc\$xyz abc\$xyz
3994 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3995
3996 For compatibility with Perl, \Q...\E sequences are recognized inside character
3997 classes as well as outside them.
3998
3999 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4000 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4001 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4002 signed/unsigned warnings.
4003
4004 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4005 option for pcretest, so I've replaced it by a simple function that does just
4006 that job.
4007
4008 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4009 "pcregrep -".
4010
4011 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4012 Java package. This provides some syntactic sugar for simple cases of what my
4013 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4014 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4015 item, you can use this simplified notation. Note that only makes sense with
4016 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4017 greediness, whatever the setting of the PCRE_UNGREEDY option.
4018
4019 13. A change of greediness default within a pattern was not taking effect at
4020 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4021 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4022 was abstracted outside.
4023
4024 14. PCRE now supports the \G assertion. It is true when the current matching
4025 position is at the start point of the match. This differs from \A when the
4026 starting offset is non-zero. Used with the /g option of pcretest (or similar
4027 code), it works in the same way as it does for Perl's /g option. If all
4028 alternatives of a regex begin with \G, the expression is anchored to the start
4029 match position, and the "anchored" flag is set in the compiled expression.
4030
4031 15. Some bugs concerning the handling of certain option changes within patterns
4032 have been fixed. These applied to options other than (?ims). For example,
4033 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4034 been the other way round. Some of this was related to change 7 above.
4035
4036 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4037 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4038 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4039 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4040
4041 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4042 mode. Unlike ".", it always matches newline, whatever the setting of
4043 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4044 assertions. Perl allows it, but it doesn't (in general) work because it can't
4045 calculate the length of the lookbehind. At least, that's the case for Perl
4046 5.8.0 - I've been told they are going to document that it doesn't work in
4047 future.
4048
4049 18. Added an error diagnosis for escapes that PCRE does not support: these are
4050 \L, \l, \N, \P, \p, \U, \u, and \X.
4051
4052 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4053 reading past the end of the pattern in cases such as /[abcd/.
4054
4055 20. PCRE was getting more memory than necessary for patterns with classes that
4056 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4057
4058 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4059 compiling PCRE for use with Virtual Pascal.
4060
4061 22. Small fix to the Makefile to make it work properly if the build is done
4062 outside the source tree.
4063
4064 23. Added a new extension: a condition to go with recursion. If a conditional
4065 subpattern starts with (?(R) the "true" branch is used if recursion has
4066 happened, whereas the "false" branch is used only at the top level.
4067
4068 24. When there was a very long string of literal characters (over 255 bytes
4069 without UTF support, over 250 bytes with UTF support), the computation of how
4070 much memory was required could be incorrect, leading to segfaults or other
4071 strange effects.
4072
4073 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4074 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4075 there was a subsequent back reference to those brackets. This meant that, for
4076 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4077 possible to check for precisely this case. All we can do is abandon the
4078 optimization if .* occurs inside capturing brackets when there are any back
4079 references whatsoever. (See below for a better fix that came later.)
4080
4081 26. The handling of the optimization for finding the first character of a
4082 non-anchored pattern, and for finding a character that is required later in the
4083 match were failing in some cases. This didn't break the matching; it just
4084 failed to optimize when it could. The way this is done has been re-implemented.
4085
4086 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4087
4088 28. Added a new feature that provides some of the functionality that Perl
4089 provides with (?{...}). The facility is termed a "callout". The way it is done
4090 in PCRE is for the caller to provide an optional function, by setting
4091 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4092 global variable. By default it is unset, which disables all calling out. To get
4093 the function called, the regex must include (?C) at appropriate points. This
4094 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4095 This provides a means of identifying different callout points. When PCRE
4096 reaches such a point in the regex, if pcre_callout has been set, the external
4097 function is called. It is provided with data in a structure called
4098 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4099 matching continues; if it returns a non-zero value, the match at the current
4100 point fails. However, backtracking will occur if possible. [This was changed
4101 later and other features added - see item 49 below.]
4102
4103 29. pcretest is upgraded to test the callout functionality. It provides a
4104 callout function that displays information. By default, it shows the start of
4105 the match and the current position in the text. There are some new data escapes
4106 to vary what happens:
4107
4108 \C+ in addition, show current contents of captured substrings
4109 \C- do not supply a callout function
4110 \C!n return 1 when callout number n is reached
4111 \C!n!m return 1 when callout number n is reached for the mth time
4112
4113 30. If pcregrep was called with the -l option and just a single file name, it
4114 output "<stdin>" if a match was found, instead of the file name.
4115
4116 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4117 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4118 pcre_exec(). This saves a malloc/free per call. The default value of
4119 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4120 when configuring.
4121
4122 32. The default maximum size of a compiled pattern is 64K. There have been a
4123 few cases of people hitting this limit. The code now uses macros to handle the
4124 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4125 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4126 configuring. Tests 2 and 5 work only with 2-byte links because they output
4127 debugging information about compiled patterns.
4128
4129 33. Internal code re-arrangements:
4130
4131 (a) Moved the debugging function for printing out a compiled regex into
4132 its own source file (printint.c) and used #include to pull it into
4133 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4134 separate copies.
4135
4136 (b) Defined the list of op-code names for debugging as a macro in
4137 internal.h so that it is next to the definition of the opcodes.
4138
4139 (c) Defined a table of op-code lengths for simpler skipping along compiled
4140 code. This is again a macro in internal.h so that it is next to the
4141 definition of the opcodes.
4142
4143 34. Added support for recursive calls to individual subpatterns, along the
4144 lines of Robin Houston's patch (but implemented somewhat differently).
4145
4146 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4147 allow it to read and process whole directories in Win32. This code was
4148 contributed by Lionel Fourquaux; it has not been tested by me.
4149
4150 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4151 used to name a group. Names consist of alphanumerics and underscores, and must
4152 be unique. Back references use the syntax (?P=name) and recursive calls use
4153 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4154 numbers. The function pcre_fullinfo() can be used after compilation to extract
4155 a name/number map. There are three relevant calls:
4156
4157 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4158 PCRE_INFO_NAMECOUNT yields the number of entries
4159 PCRE_INFO_NAMETABLE yields a pointer to the map.
4160
4161 The map is a vector of fixed-size entries. The size of each entry depends on
4162 the length of the longest name used. The first two bytes of each entry are the
4163 group number, most significant byte first. There follows the corresponding
4164 name, zero terminated. The names are in alphabetical order.
4165
4166 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4167 case instead of 255. Making it the same both with and without UTF-8 support
4168 means that the same test output works with both.
4169
4170 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4171 calling malloc() with a zero argument.
4172
4173 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4174 optimization. I've improved things by keeping a bitmap of backreferences with
4175 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4176 fact referenced, the optimization can be applied. It is unlikely that a
4177 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4178 the match to follow \n) will appear inside brackets with a number greater than
4179 31, but if it does, any back reference > 31 suppresses the optimization.
4180
4181 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4182 of disabling numbered capturing parentheses. Any opening parenthesis that is
4183 not followed by ? behaves as if it were followed by ?: but named parentheses
4184 can still be used for capturing (and they will acquire numbers in the usual
4185 way).
4186
4187 41. Redesigned the return codes from the match() function into yes/no/error so
4188 that errors can be passed back from deep inside the nested calls. A malloc
4189 failure while inside a recursive subpattern call now causes the
4190 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4191
4192 42. It is now possible to set a limit on the number of times the match()
4193 function is called in a call to pcre_exec(). This facility makes it possible to
4194 limit the amount of recursion and backtracking, though not in a directly
4195 obvious way, because the match() function is used in a number of different
4196 circumstances. The count starts from zero for each position in the subject
4197 string (for non-anchored patterns). The default limit is, for compatibility, a
4198 large number, namely 10 000 000. You can change this in two ways:
4199
4200 (a) When configuring PCRE before making, you can use --with-match-limit=n
4201 to set a default value for the compiled library.
4202
4203 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4204 a different value is set. See 45 below.
4205
4206 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4207
4208 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4209 of things that can be changed at compile time. The first argument specifies
4210 what is wanted and the second points to where the information is to be placed.
4211 The current list of available information is:
4212
4213 PCRE_CONFIG_UTF8
4214
4215 The output is an integer that is set to one if UTF-8 support is available;
4216 otherwise it is set to zero.
4217
4218 PCRE_CONFIG_NEWLINE
4219
4220 The output is an integer that it set to the value of the code that is used for
4221 newline. It is either LF (10) or CR (13).
4222
4223 PCRE_CONFIG_LINK_SIZE
4224
4225 The output is an integer that contains the number of bytes used for internal
4226 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4227
4228 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4229
4230 The output is an integer that contains the threshold above which the POSIX
4231 interface uses malloc() for output vectors. See item 31 above.
4232
4233 PCRE_CONFIG_MATCH_LIMIT
4234
4235 The output is an unsigned integer that contains the default limit of the number
4236 of match() calls in a pcre_exec() execution. See 42 above.
4237
4238 44. pcretest has been upgraded by the addition of the -C option. This causes it
4239 to extract all the available output from the new pcre_config() function, and to
4240 output it. The program then exits immediately.
4241
4242 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4243 order to support additional features. One way would have been to define
4244 pcre_exec2() (for example) with extra arguments, but this would not have been
4245 extensible, and would also have required all calls to the original function to
4246 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4247 is used for passing in "extra" data from pcre_study().
4248
4249 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4250 contains the following fields:
4251
4252 flags a bitmap indicating which of the following fields are set
4253 study_data opaque data from pcre_study()
4254 match_limit a way of specifying a limit on match() calls for a specific
4255 call to pcre_exec()
4256 callout_data data for callouts (see 49 below)
4257
4258 The flag bits are also defined in pcre.h, and are
4259
4260 PCRE_EXTRA_STUDY_DATA
4261 PCRE_EXTRA_MATCH_LIMIT
4262 PCRE_EXTRA_CALLOUT_DATA
4263
4264 The pcre_study() function now returns one of these new pcre_extra blocks, with
4265 the actual study data pointed to by the study_data field, and the
4266 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4267 before. That is, this change is entirely upwards-compatible and requires no
4268 change to existing code.
4269
4270 If you want to pass in additional data to pcre_exec(), you can either place it
4271 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4272 block.
4273
4274 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4275 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4276 times with different match limits, until it finds the minimum value needed for
4277 pcre_exec() to complete. The value is then output. This can be instructive; for
4278 most simple matches the number is quite small, but for pathological cases it
4279 gets very large very quickly.
4280
4281 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4282 returns the size of the data block pointed to by the study_data field in a
4283 pcre_extra block, that is, the value that was passed as the argument to
4284 pcre_malloc() when PCRE was getting memory in which to place the information
4285 created by pcre_study(). The fourth argument should point to a size_t variable.
4286 pcretest has been extended so that this information is shown after a successful
4287 pcre_study() call when information about the compiled regex is being displayed.
4288
4289 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4290 because what follows is always an absolute path. (Later: it turns out that this
4291 is more than cosmetic for MinGW, because it doesn't like empty path
4292 components.)
4293
4294 49. Some changes have been made to the callout feature (see 28 above):
4295
4296 (i) A callout function now has three choices for what it returns:
4297
4298 0 => success, carry on matching
4299 > 0 => failure at this point, but backtrack if possible
4300 < 0 => serious error, return this value from pcre_exec()
4301
4302 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4303 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4304 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4305 use by callout functions. It will never be used by PCRE itself.
4306
4307 (ii) The pcre_extra structure (see 45 above) has a void * field called
4308 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4309 pcre_callout_block structure has a field of the same name. The contents of
4310 the field passed in the pcre_extra structure are passed to the callout
4311 function in the corresponding field in the callout block. This makes it
4312 easier to use the same callout-containing regex from multiple threads. For
4313 testing, the pcretest program has a new data escape
4314
4315 \C*n pass the number n (may be negative) as callout_data
4316
4317 If the callout function in pcretest receives a non-zero value as
4318 callout_data, it returns that value.
4319
4320 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4321 there were some redundant $(CFLAGS) in commands that are now specified as
4322 $(LINK), which already includes $(CFLAGS).
4323
4324 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4325 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4326 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4327 one-byte characters throughout. Note that case-insensitive matching applies
4328 only to characters whose values are less than 256. PCRE doesn't support the
4329 notion of cases for higher-valued characters.
4330
4331 (i) A character class whose characters are all within 0-255 is handled as
4332 a bit map, and the map is inverted for negative classes. Previously, a
4333 character > 255 always failed to match such a class; however it should
4334 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4335
4336 (ii) A negated character class with a single character < 255 is coded as
4337 "not this character" (OP_NOT). This wasn't working properly when the test
4338 character was multibyte, either singly or repeated.
4339
4340 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4341 mode, for example: \x{100}{2,3}.
4342
4343 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4344 singly or repeated) now correctly test multibyte characters. However,
4345 PCRE doesn't recognize any characters with values greater than 255 as
4346 digits, spaces, or word characters. Such characters always match \D, \S,
4347 and \W, and never match \d, \s, or \w.
4348
4349 (v) Classes may now contain characters and character ranges with values
4350 greater than 255. For example: [ab\x{100}-\x{400}].
4351
4352 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4353 PCRE in UTF-8 mode.
4354
4355 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4356 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4357 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4358 value.)
4359
4360 53. The single man page has become too large. I have therefore split it up into
4361 a number of separate man pages. These also give rise to individual HTML pages;
4362 these are now put in a separate directory, and there is an index.html page that
4363 lists them all. Some hyperlinking between the pages has been installed.
4364
4365 54. Added convenience functions for handling named capturing parentheses.
4366
4367 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4368 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4369 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4370 are faulted.
4371
4372 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4373 calling configure. These values are used when compiling the dftables.c program
4374 which is run to generate the source of the default character tables. They
4375 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4376 you will need to set these values.
4377
4378 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4379
4380
4381 Version 3.9 02-Jan-02
4382 ---------------------
4383
4384 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4385
4386 2. If --disable-static was given, the building process failed when trying to
4387 build pcretest and pcregrep. (For some reason it was using libtool to compile
4388 them, which is not right, as they aren't part of the library.)
4389
4390
4391 Version 3.8 18-Dec-01
4392 ---------------------
4393
4394 1. The experimental UTF-8 code was completely screwed up. It was packing the
4395 bytes in the wrong order. How dumb can you get?
4396
4397
4398 Version 3.7 29-Oct-01
4399 ---------------------
4400
4401 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4402 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4403 this didn't happen under Solaris 8, where I normally test things.
4404
4405 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4406 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4407 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4408 this fix an hour or so after the initial 3.7 release.)
4409
4410
4411 Version 3.6 23-Oct-01
4412 ---------------------
4413
4414 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4415 offsets passed as NULL with zero offset count.
4416
4417 2. The config.guess and config.sub files had not been updated when I moved to
4418 the latest autoconf.
4419
4420
4421 Version 3.5 15-Aug-01
4422 ---------------------
4423
4424 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4425 had been forgotten.
4426
4427 2. By using declared but undefined structures, we can avoid using "void"
4428 definitions in pcre.h while keeping the internal definitions of the structures
4429 private.
4430
4431 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4432 user point of view, this means that both static and shared libraries are built
4433 by default, but this can be individually controlled. More of the work of
4434 handling this static/shared cases is now inside libtool instead of PCRE's make
4435 file.
4436
4437 4. The pcretest utility is now installed along with pcregrep because it is
4438 useful for users (to test regexs) and by doing this, it automatically gets
4439 relinked by libtool. The documentation has been turned into a man page, so
4440 there are now .1, .txt, and .html versions in /doc.
4441
4442 5. Upgrades to pcregrep:
4443 (i) Added long-form option names like gnu grep.
4444 (ii) Added --help to list all options with an explanatory phrase.
4445 (iii) Added -r, --recursive to recurse into sub-directories.
4446 (iv) Added -f, --file to read patterns from a file.
4447
4448 6. pcre_exec() was referring to its "code" argument before testing that
4449 argument for NULL (and giving an error if it was NULL).
4450
4451 7. Upgraded Makefile.in to allow for compiling in a different directory from
4452 the source directory.
4453
4454 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4455 options bits, the pointer it was passed was to an int instead of to an unsigned
4456 long int. This mattered only on 64-bit systems.
4457
4458 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4459 generated) instead of pcre.in, which it its source. Also made the same change
4460 in several of the .c files.
4461
4462 10. A new release of gcc defines printf() as a macro, which broke pcretest
4463 because it had an ifdef in the middle of a string argument for printf(). Fixed
4464 by using separate calls to printf().
4465
4466 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4467 script, to force use of CR or LF instead of \n in the source. On non-Unix
4468 systems, the value can be set in config.h.
4469
4470 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4471 absolute limit. Changed the text of the error message to make this clear, and
4472 likewise updated the man page.
4473
4474 13. The limit of 99 on the number of capturing subpatterns has been removed.
4475 The new limit is 65535, which I hope will not be a "real" limit.
4476
4477
4478 Version 3.4 22-Aug-00
4479 ---------------------
4480
4481 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4482
4483 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4484
4485
4486 Version 3.3 01-Aug-00
4487 ---------------------
4488
4489 1. If an octal character was given, but the value was greater than \377, it
4490 was not getting masked to the least significant bits, as documented. This could
4491 lead to crashes in some systems.
4492
4493 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4494 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4495
4496 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4497 These just pass their arguments on to (pcre_free)(), but they are provided
4498 because some uses of PCRE bind it to non-C systems that can call its functions,
4499 but cannot call free() or pcre_free() directly.
4500
4501 4. Add "make test" as a synonym for "make check". Corrected some comments in
4502 the Makefile.
4503
4504 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4505 Makefile.
4506
4507 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4508 command called pgrep for grepping around the active processes.
4509
4510 7. Added the beginnings of support for UTF-8 character strings.
4511
4512 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4513 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4514 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4515 out for the ar command.)
4516
4517
4518 Version 3.2 12-May-00
4519 ---------------------
4520
4521 This is purely a bug fixing release.
4522
4523 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4524 of ZA. This was just one example of several cases that could provoke this bug,
4525 which was introduced by change 9 of version 2.00. The code for breaking
4526 infinite loops after an iteration that matches an empty string was't working
4527 correctly.
4528
4529 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4530 when matched against abbab (for example). After matching an empty string, it
4531 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4532 caused it to match further down the string than it should.
4533
4534 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4535 was there because it doesn't seem to be needed, and it causes trouble on some
4536 systems, as it is not a Standard C header. It has been removed.
4537
4538 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4539 were reported on the Macintosh. The changes were from
4540
4541 while ((c = *(++ptr)) != 0 && c != '\n');
4542 to
4543 while ((c = *(++ptr)) != 0 && c != '\n') ;
4544
4545 Totally extraordinary, but if that's what it takes...
4546
4547 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4548 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4549 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4550 assumes the way PCRE uses memmove() (always moving upwards).
4551
4552 6. PCRE is being used in one environment where strchr() is not available. There
4553 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4554 faster code anyway.
4555
4556
4557 Version 3.1 09-Feb-00
4558 ---------------------
4559
4560 The only change in this release is the fixing of some bugs in Makefile.in for
4561 the "install" target:
4562
4563 (1) It was failing to install pcreposix.h.
4564
4565 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4566
4567
4568 Version 3.0 01-Feb-00
4569 ---------------------
4570
4571 1. Add support for the /+ modifier to perltest (to output $` like it does in
4572 pcretest).
4573
4574 2. Add support for the /g modifier to perltest.
4575
4576 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4577 matches null strings.
4578
4579 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4580 pattern. Perl treats empty patterns specially - it reuses the most recent
4581 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4582 effect.
4583
4584 5. The POSIX interface was broken in that it was just handing over the POSIX
4585 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4586 required a bigger vector, with some working space on the end. This means that
4587 the POSIX wrapper now has to get and free some memory, and copy the results.
4588
4589 6. Added some simple autoconf support, placing the test data and the
4590 documentation in separate directories, re-organizing some of the
4591 information files, and making it build pcre-config (a GNU standard). Also added
4592 libtool support for building PCRE as a shared library, which is now the
4593 default.
4594
4595 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4596 09 are not valid octal constants. Single digits will be used for minor values
4597 less than 10.
4598
4599 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4600 existing programs that set these in the POSIX interface can use PCRE without
4601 modification.
4602
4603 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4604 return all that pcre_info() returns, plus additional data. The pcre_info()
4605 function is retained for compatibility, but is considered to be obsolete.
4606
4607 10. Added experimental recursion feature (?R) to handle one common case that
4608 Perl 5.6 will be able to do with (?p{...}).
4609
4610 11. Added support for POSIX character classes like [:alpha:], which Perl is
4611 adopting.
4612
4613
4614 Version 2.08 31-Aug-99
4615 ----------------------
4616
4617 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4618 trying to match at the startoffset position, but instead was moving forward to
4619 the next newline as if a previous match had failed.
4620
4621 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4622 and could get into a loop if a null string was matched other than at the start
4623 of the subject.
4624
4625 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4626 be distinguished at compile time, and for completeness also added PCRE_DATE.
4627
4628 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4629 in GnuWin32 environments.
4630
4631
4632 Version 2.07 29-Jul-99
4633 ----------------------
4634
4635 1. The documentation is now supplied in plain text form and HTML as well as in
4636 the form of man page sources.
4637
4638 2. C++ compilers don't like assigning (void *) values to other pointer types.
4639 In particular this affects malloc(). Although there is no problem in Standard
4640 C, I've put in casts to keep C++ compilers happy.
4641
4642 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4643 should be (const char *).
4644
4645 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4646 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4647 However, I haven't made this a standard facility. The documentation doesn't
4648 mention it, and the Makefile doesn't support it.
4649
4650 5. The Makefile now contains an "install" target, with editable destinations at
4651 the top of the file. The pcretest program is not installed.
4652
4653 6. pgrep -V now gives the PCRE version number and date.
4654
4655 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4656 causing the entire string to be ignored, instead of just the last character.
4657
4658 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4659 non-matching string, it can take a very, very long time, even for strings of
4660 quite modest length, because of the nested recursion. PCRE now does better in
4661 some of these cases. It does this by remembering the last required literal
4662 character in the pattern, and pre-searching the subject to ensure it is present
4663 before running the real match. In other words, it applies a heuristic to detect
4664 some types of certain failure quickly, and in the above example, if presented
4665 with a string that has no trailing " it gives "no match" very quickly.
4666
4667 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4668 other alternatives are tried instead.
4669
4670
4671 Version 2.06 09-Jun-99
4672 ----------------------
4673
4674 1. Change pcretest's output for amount of store used to show just the code
4675 space, because the remainder (the data block) varies in size between 32-bit and
4676 64-bit systems.
4677
4678 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4679 start matching at. This allows lookbehinds to work when searching for multiple
4680 occurrences in a string.
4681
4682 3. Added additional options to pcretest for testing multiple occurrences:
4683
4684 /+ outputs the rest of the string that follows a match
4685 /g loops for multiple occurrences, using the new startoffset argument
4686 /G loops for multiple occurrences by passing an incremented pointer
4687
4688 4. PCRE wasn't doing the "first character" optimization for patterns starting
4689 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4690 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4691 the letter 'x'. On long subject strings, this gives a significant speed-up.
4692
4693
4694 Version 2.05 21-Apr-99
4695 ----------------------
4696
4697 1. Changed the type of magic_number from int to long int so that it works
4698 properly on 16-bit systems.
4699
4700 2. Fixed a bug which caused patterns starting with .* not to work correctly
4701 when the subject string contained newline characters. PCRE was assuming
4702 anchoring for such patterns in all cases, which is not correct because .* will
4703 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4704 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4705 must be retried after every newline in the subject.
4706
4707
4708 Version 2.04 18-Feb-99
4709 ----------------------
4710
4711 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4712 computation of the store needed to hold the pattern was incorrect (too large).
4713 If such patterns were nested a few deep, this could multiply and become a real
4714 problem.
4715
4716 2. Added /M option to pcretest to show the memory requirement of a specific
4717 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4718
4719 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4720 compiled in such a way that the backtracking after subsequent failure was
4721 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4722 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4723
4724
4725 Version 2.03 02-Feb-99
4726 ----------------------
4727
4728 1. Fixed typo and small mistake in man page.
4729
4730 2. Added 4th condition (GPL supersedes if conflict) and created separate
4731 LICENCE file containing the conditions.
4732
4733 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4734 Perl, that is the internal \ allows the delimiter to be included in the
4735 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4736 the final delimiter, add \ to the end of the pattern (to test the error).
4737
4738 4. Added the convenience functions for extracting substrings after a successful
4739 match. Updated pcretest to make it able to test these functions.
4740
4741
4742 Version 2.02 14-Jan-99
4743 ----------------------
4744
4745 1. Initialized the working variables associated with each extraction so that
4746 their saving and restoring doesn't refer to uninitialized store.
4747
4748 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4749 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4750 fix the problem.
4751
4752 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4753 calls, and wasn't printing the correct value for compiling calls. Increased the
4754 default value of LOOPREPEAT, and the number of significant figures in the
4755 times.
4756
4757 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4758
4759 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4760 a building problem on Windows NT with a FAT file system.
4761
4762
4763 Version 2.01 21-Oct-98
4764 ----------------------
4765
4766 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4767 to character tables built by pcre_maketables() in the current locale. If NULL
4768 is passed, the default tables are used.
4769
4770
4771 Version 2.00 24-Sep-98
4772 ----------------------
4773
4774 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4775 it any more.
4776
4777 2. Allow quantification of (?>) groups, and make it work correctly.
4778
4779 3. The first character computation wasn't working for (?>) groups.
4780
4781 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4782 end of the subject) and add 5.005's \z, which really does match only at the
4783 very end of the subject.
4784
4785 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4786
4787 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4788 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4789 localized options. All options to pcre_study() were also removed.
4790
4791 7. Add other new features from 5.005:
4792
4793 $(?<= positive lookbehind
4794 $(?<! negative lookbehind
4795 (?imsx-imsx) added the unsetting capability
4796 such a setting is global if at outer level; local otherwise
4797 (?imsx-imsx:) non-capturing groups with option setting
4798 (?(cond)re|re) conditional pattern matching
4799
4800 A backreference to itself in a repeated group matches the previous
4801 captured string.
4802
4803 8. General tidying up of studying (both automatic and via "study")
4804 consequential on the addition of new assertions.
4805
4806 9. As in 5.005, unlimited repeated groups that could match an empty substring
4807 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4808 runtime if any iteration does actually match an empty substring.
4809
4810 10. Include the RunTest script in the distribution.
4811
4812 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4813 discrepancies, some of which were old and were also with respect to 5.004. They
4814 have now been fixed.
4815
4816
4817 Version 1.09 28-Apr-98
4818 ----------------------
4819
4820 1. A negated single character class followed by a quantifier with a minimum
4821 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4822 program crashes, or just wrong answers. This did not apply to negated classes
4823 containing more than one character, or to minima other than one.
4824
4825
4826 Version 1.08 27-Mar-98
4827 ----------------------
4828
4829 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4830
4831 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4832 latter must appear before anything that relies on it in the pattern.
4833
4834
4835 Version 1.07 16-Feb-98
4836 ----------------------
4837
4838 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4839 repeat of a potentially empty string).
4840
4841
4842 Version 1.06 23-Jan-98
4843 ----------------------
4844
4845 1. Added Markus Oberhumer's little patches for C++.
4846
4847 2. Literal strings longer than 255 characters were broken.
4848
4849
4850 Version 1.05 23-Dec-97
4851 ----------------------
4852
4853 1. Negated character classes containing more than one character were failing if
4854 PCRE_CASELESS was set at run time.
4855
4856
4857 Version 1.04 19-Dec-97
4858 ----------------------
4859
4860 1. Corrected the man page, where some "const" qualifiers had been omitted.
4861
4862 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4863 input syntax.
4864
4865 3. Fixed memory leak which occurred when a regex with back references was
4866 matched with an offsets vector that wasn't big enough. The temporary memory
4867 that is used in this case wasn't being freed if the match failed.
4868
4869 4. Tidied pcretest to ensure it frees memory that it gets.
4870
4871 5. Temporary memory was being obtained in the case where the passed offsets
4872 vector was exactly big enough.
4873
4874 6. Corrected definition of offsetof() from change 5 below.
4875
4876 7. I had screwed up change 6 below and broken the rules for the use of
4877 setjmp(). Now fixed.
4878
4879
4880 Version 1.03 18-Dec-97
4881 ----------------------
4882
4883 1. A erroneous regex with a missing opening parenthesis was correctly
4884 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4885 on some systems.
4886
4887 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4888 it was reported that one broken compiler failed on the former because "code" is
4889 also an independent variable.
4890
4891 3. The erroneous regex a[]b caused an array overrun reference.
4892
4893 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4894 fail on data ending with that character. (It was going on too far, and checking
4895 the next character, typically a binary zero.) This was specific to the
4896 optimized code for single-character negative classes.
4897
4898 5. Added a contributed patch from the TIN world which does the following:
4899
4900 + Add an undef for memmove, in case the the system defines a macro for it.
4901
4902 + Add a definition of offsetof(), in case there isn't one. (I don't know
4903 the reason behind this - offsetof() is part of the ANSI standard - but
4904 it does no harm).
4905
4906 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4907 most of the places where whitespace preceded '#'. I have given up and
4908 allowed the remaining 2 cases to be at the margin.
4909
4910 + Rename some variables in pcre to eliminate shadowing. This seems very
4911 pedantic, but does no harm, of course.
4912
4913 6. Moved the call to setjmp() into its own function, to get rid of warnings
4914 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4915
4916 7. Constructs such as \d{8,} were compiling into the equivalent of
4917 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4918 outcome, but in this particular case used more store than had been allocated,
4919 which caused the bug to be discovered because it threw up an internal error.
4920
4921 8. The debugging code in both pcre and pcretest for outputting the compiled
4922 form of a regex was going wrong in the case of back references followed by
4923 curly-bracketed repeats.
4924
4925
4926 Version 1.02 12-Dec-97
4927 ----------------------
4928
4929 1. Typos in pcre.3 and comments in the source fixed.
4930
4931 2. Applied a contributed patch to get rid of places where it used to remove
4932 'const' from variables, and fixed some signed/unsigned and uninitialized
4933 variable warnings.
4934
4935 3. Added the "runtest" target to Makefile.
4936
4937 4. Set default compiler flag to -O2 rather than just -O.
4938
4939
4940 Version 1.01 19-Nov-97
4941 ----------------------
4942
4943 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4944 like /([ab]*)*/, that is, for classes with more than one character in them.
4945
4946 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4947 as /((?>a*))*/ (a PCRE_EXTRA facility).
4948
4949
4950 Version 1.00 18-Nov-97
4951 ----------------------
4952
4953 1. Added compile-time macros to support systems such as SunOS4 which don't have
4954 memmove() or strerror() but have other things that can be used instead.
4955
4956 2. Arranged that "make clean" removes the executables.
4957
4958
4959 Version 0.99 27-Oct-97
4960 ----------------------
4961
4962 1. Fixed bug in code for optimizing classes with only one character. It was
4963 initializing a 32-byte map regardless, which could cause it to run off the end
4964 of the memory it had got.
4965
4966 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4967
4968
4969 Version 0.98 22-Oct-97
4970 ----------------------
4971
4972 1. Fixed bug in code for handling temporary memory usage when there are more
4973 back references than supplied space in the ovector. This could cause segfaults.
4974
4975
4976 Version 0.97 21-Oct-97
4977 ----------------------
4978
4979 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4980
4981 2. Optimized negated single characters not to use a bit map.
4982
4983 3. Brought error texts together as macro definitions; clarified some of them;
4984 fixed one that was wrong - it said "range out of order" when it meant "invalid
4985 escape sequence".
4986
4987 4. Changed some char * arguments to const char *.
4988
4989 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4990
4991 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4992 pcretest.
4993
4994
4995 Version 0.96 16-Oct-97
4996 ----------------------
4997
4998 1. Added a simple "pgrep" utility to the distribution.
4999
5000 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5001 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5002 where "ddd" means "one or more decimal digits".
5003
5004 3. Fixed serious bug. If a pattern had a back reference, but the call to
5005 pcre_exec() didn't supply a large enough ovector to record the related
5006 identifying subpattern, the match always failed. PCRE now remembers the number
5007 of the largest back reference, and gets some temporary memory in which to save
5008 the offsets during matching if necessary, in order to ensure that
5009 backreferences always work.
5010
5011 4. Increased the compatibility with Perl in a number of ways:
5012
5013 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5014 to request this handling. The option can be set at compile or exec time.
5015
5016 (b) $ matches before a terminating newline by default; an option
5017 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5018 mode). The option can be set at compile or exec time.
5019
5020 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5021 the same as Perl's. If the decimal number it represents is less than 10
5022 or there aren't that many previous left capturing parentheses, an octal
5023 escape is read. Inside a character class, it's always an octal escape,
5024 even if it is a single digit.
5025
5026 (d) An escaped but undefined alphabetic character is taken as a literal,
5027 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5028 escapes.
5029
5030 (e) {0} is now permitted. (The previous item is removed from the compiled
5031 pattern).
5032
5033 5. Changed all the names of code files so that the basic parts are no longer
5034 than 10 characters, and abolished the teeny "globals.c" file.
5035
5036 6. Changed the handling of character classes; they are now done with a 32-byte
5037 bit map always.
5038
5039 7. Added the -d and /D options to pcretest to make it possible to look at the
5040 internals of compilation without having to recompile pcre.
5041
5042
5043 Version 0.95 23-Sep-97
5044 ----------------------
5045
5046 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5047 \x20 at the start of a run of normal characters. These were being treated as
5048 real characters, instead of the source characters being re-checked.
5049
5050
5051 Version 0.94 18-Sep-97
5052 ----------------------
5053
5054 1. The functions are now thread-safe, with the caveat that the global variables
5055 containing pointers to malloc() and free() or alternative functions are the
5056 same for all threads.
5057
5058 2. Get pcre_study() to generate a bitmap of initial characters for non-
5059 anchored patterns when this is possible, and use it if passed to pcre_exec().
5060
5061
5062 Version 0.93 15-Sep-97
5063 ----------------------
5064
5065 1. /(b)|(:+)/ was computing an incorrect first character.
5066
5067 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5068 but not actually doing anything yet.
5069
5070 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5071 as Perl does (e.g. [-az] or [az-]).
5072
5073 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5074 all possible positions.
5075
5076 5. Split up into different modules to avoid including unneeded functions in a
5077 compiled binary. However, compile and exec are still in one module. The "study"
5078 function is split off.
5079
5080 6. The character tables are now in a separate module whose source is generated
5081 by an auxiliary program - but can then be edited by hand if required. There are
5082 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5083 toupper() in the code.
5084
5085 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5086 make them global. Abolish the function for setting them, as the caller can now
5087 set them directly.
5088
5089
5090 Version 0.92 11-Sep-97
5091 ----------------------
5092
5093 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5094 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5095
5096 2. Caseless matching was not working in character classes if the characters in
5097 the pattern were in upper case.
5098
5099 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5100
5101 4. Make PCRE_ANCHORED public and accept as a compile option.
5102
5103 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5104 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5105 pass them.
5106
5107 6. Give an error if bad option bits passed at compile or run time.
5108
5109 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5110 pcretest to cause it to pass that flag.
5111
5112 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5113 options, and the first character, if set.
5114
5115 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5116
5117
5118 Version 0.91 10-Sep-97
5119 ----------------------
5120
5121 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5122 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5123
5124 2. PCRE was looping on encountering an indefinitely repeated back reference to
5125 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5126 Perl does - treats the match as successful.
5127
5128 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5